Seminar Publication
Site Characterization for Subsurface Remediation
United StatesEnvironmental ProtectionAgency
Office of Research and DevelopmentWashington, DC 20460
EPA/625/4-91/026November 1991
EPA
Technology Transfer
EPA/625/4-91/026November 1991
Seminar Publication
Site Characterization for Subsurface Remediation
Center for Environmental Research InformationOffice of Research and Development
U.S. Environmental Protection AgencyCincinnati, OH 45268
Printed on recycled paper
Notice
This document has been reviewed in accordance with U.S. Environmental ProtectionAgency policy and approved for publication. Mention of trade names or commercialproducts does not constitute endorsement or recommendation for use.
Preface
Site characterization of contaminated sites has become an increasingly complexprocess as a result of rapid developments in (1) methods for observing the physical,chemical, and biological characteristics of the subsurface, and (2) methods for remediationof soil and ground water. Consideration of the possible methods that may be used to cleanup contaminated soils or ground water early in the site characterization process can ensurethat data collected are appropriate and possibly reduce the time it takes to initiate clean-upefforts.
This seminar publication provides a comprehensive approach to site characterizationfor subsurface remediation. Chapter 1 describes a methodology for integrating sitecharacterization with subsurface remediation. This introductory chapter of the handbookalso provides a guide for quickly and efficiently accessing information in the rest of thedocument for specific remediation applications through the use of summary tables,checklists, figures, and flow charts.
The rest of the handbook is divided into three parts. Part I covers methods forsubsurface characterization, Part II covers physical and chemical processes in the subsur-face that relate to the selection of remediation methods, and Part III covers methods for soiland ground-water remediation.
In Part I, Chapter 2 provides an overview of the site characterization process. The nextfour chapters cover physical aspects of site characterization: geologic and hydrogeologicaspects (Chapter 3), characterization of water movement in the unsaturated zone (Chapter4), characterization of the vadose zone (Chapter 5), and characterization of water move-ment in saturated fractured media (Chapter 6). The remaining three chapters in Part I covergeochemical aspects of site characterization: basic analytical and statistical concepts(Chapter 7), the geochemical variability of the natural and contaminated subsurface(Chapter 8), and geochemical sampling of soil and ground water (Chapter 9).
Part II contains three chapters on physiochemical processes affecting the transport ofmajor types of contaminants: organics in liquid and solid phases in the subsurface(Chapter 10), organic volatilization and gas-phase transport (Chapter 11), and inorganiccontaminants (Chapter 12). Chapter 13 focuses on abiotic and microbiological degradationand transformation processes in the subsurface.
Part III contains three chapters on remediation. Chapter 14 outlines basis approachesto remediation of contaminated soil and ground water. The concluding chapters providemore detailed information on specific techniques for cleaning up contaminated soil(Chapter 15) and ground water (Chapter 16).
iii
v
__________________________________________________________________________________
Contents
Notice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .x
Chapter 1 Integrating Site Characterization with Subsurface Remediation . . . . . . . . . . . . . . . . . 1
1.1 Approach for Integration of Site Characterization with Subsurface Remediation . . . . . . . . . . . . . 11.2 Subsurface Site Characterization for Remediation Technology Selection . . . . . . . . . . . . . . . . . . . 11.3 Site Reconnaissance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Part I: Methods for Subsurface Characterization
Chapter 2 Site Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.2 Flow System Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132.3 Contamination Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162.4 Techniques for Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172.5 Analysis of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Chapter 3 Geologic Aspects of Site Remediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.1 Stratigraphy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233.2 Lithology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243.3 Structural Geology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273.4 Hydrogeology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273.5 Hydrogeologic Investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.5.1 Geophysical Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283.5.2 Example - Hyde Park Landfill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Chapter 4 Characterization of Water Movement in the Saturated Zone . . . . . . . . . . . . . . . . . . 39
4.1 Review of Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394.2 Field Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.2.1 Drilling Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414.2.2 Methods to Measure Hydraulic Head . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454.2.3 Methods to Determine Aquifer Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 Analysis of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514.4 Remedial Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514.5 Example - Conservation Chemical Company Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
vi
_____________________________________________________________________________
Contents (continued)
Chapter 5 Characterization of the Vadose Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.1 Review of Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.2 Field Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.2.1 Precipitation and Infiltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.2.2 Evaporation and Evapotranspiration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.2.3 Moisture Content and Moisture Characteristics Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.2.4 Vadose-Zone Hydraulic Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645.2.5 Soil Gas Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.3 Analysis of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665.4 Remedial Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665.5 Example-Pepper’s Steel Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 685.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Chapter 6 Characterization of Water Movement in Saturated Fractured Media . . . . . . . . . . . . . . . . . . 73
6.1 Review of Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 736.2 Field Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
6.2.1 Fracture Trace Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746.2.2 Coring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 756.2.3 Aquifer Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 756.2.4 Tracer Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766.2.5 Geophysical Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766.2.6 Borehole Flowmeters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
6.3 Analysis of Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766.4 Remedial Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 776.5 Example-Marion County, Florida . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 776.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Chapter 7 Geochemical Characterization of the Subsurface: Basic Analytical and Statistical Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.1 Data Measurement Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 837.1.1 Deterministic versus Random Geochemical Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 837.1.2 Data Representativeness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 837.1.3 Measurement Bias, Precision, and Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 847.1.4 Sources of Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
7.2 Analytical and QA/QC Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 877.2.1 Instrumentation and Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887.2.2 Limit of Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887.2.3 Types of Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
7.3 Statistical Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 917.3.1 Statistical Approaches to Geochemical Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 917.3.2 Geostatistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
7.4 Interpretation of Geochemical and Water Chemistry Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 927.4.1 Analysis of Censored Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 937.4.2 Contaminant Levels versus Background Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
7.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Chapter 8 Geochemical Variability of the Natural and Contaminated Subsurface Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.1 Overview of Subsurface Geochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1038.1.1 Geochemical Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1038.1.2 Environmental Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
vii
__________________________________________________________________________________
Contents (continued)
8.1.3 The Vadose and Saturated Zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1068.2 Background Levels and Behavior of Chemical Constituents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1078.3 Spatial Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.3.1 Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1088.3.2 Physical Gradients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1088.3.3 Chemical Gradients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.4 Temporal Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1108.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Chapter 9 Geochemical Sampling of Subsurface Solids and Ground Water . . . . . . . . . . . . . . . . . . . . . . 123
9.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1239.1.1 Types of Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1239.1.2 Sampling Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1239.1.3 Sampling Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1259.1.4 Sampling Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1279.1.5 Sample Type and Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1289.1.6 Vadose versus Saturated Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
9.2 Sampling Subsurface Solids and Vadose Zone Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1299.2.1 Analyte Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1299.2.2 Sampling Devices and Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
9.3 Sampling Ground Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1309.3.1 Analyte Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1319.3.2 Well Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1329.3.3 Purging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1369.3.4 Well Construction and Sampling Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
9.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Part II: Physical and Chemical Processes in the Subsurface
Chapter 10 Physicochemical Processes: Organic Contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
10.1 Overview of Physicochemical Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15510.2 Dissolution of Nonaqueous Phase Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15610.3 Sorption Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
10.3.1 Adsorption Isotherms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15710.3.2 Determining Retardation Factors Using f and K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159oc oc
10.3.3 Determining Retardation Factors Using Batch Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16110.3.4 Determining Retardation Factors Using Column Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16210.3.5 Determining Retardation Factors From Field Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16310.3.6 Comparison of Methods for Estimation of Retardation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16310.3.7 Applicability and Limitations of Linear Partitioning and Retardation . . . . . . . . . . . . . . . . . . . . 164
10.4 Ionization and Cosolvation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16410.5 Expressions for Other Chemical Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16510.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Chapter 11 Physicochemical Processes: Volatilization and Gas-Phase Transport . . . . . . . . . . . . . . . . . 169
11.1 Volatilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17011.2 Gas-Phase Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
11.2.1 Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17011.2.2 Gas Phase Retardation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17111.2.3 Processes Affecting Gas-Phase Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
viii
_________________________________________________________________________________
Contents (continued)
11.3 Vapor Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17511.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Chapter 12 Physicochemical Processes: Inorganic Contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
12.1 Chemical Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18012.1.1 Speciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18012.1.2 Dissolution/Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18012.1.3 Oxidation/Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18112.1.4 Adsorption/Ion Exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
12.2 Particle Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18512.3 Organic-Inorganic Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18612.4 Computational Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
12.4.1 Mass Balance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18912.4.2 Chemical Speciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18912.4.3 Mass Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18912.4.4 Multicomponent Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
12.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
Chapter 13 Characterization of Subsurface Degradation Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
13.1 Abiotic Transformation Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19313.1.1 Hydrolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19313.1.2 Substitution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19413.1.3 Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19413.1.4 Oxidation-Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
13.2 Microbiological Transformations in the Subsurface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19413.2.1 Microbial Ecology of the Subsurface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19513.2.2 Relationship of Environmental Factors to Biodegradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19613.2.3 Microbial Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19613.2.4 Biological Reaction Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
13.3 Bioremediation of Organic Compounds in the Subsurface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 13.3.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
13.3.2 Compounds Appropriate to Consider for Bioremediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19813.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Part III: Soil and Ground-Water Remediation
Chapter 14 Soil and Ground-Water Remediation: Basic Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
14.1 Conceptual Approach to Soil and Ground-Water Remediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20314.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
14.2.1 Site Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20614.2.2 Assessment of Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20614.2.3 Treatment Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20914.2.4 Monitoring Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
14.3 Selection of Treatment Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21014.3.1 Utility of Mathematical Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21014.3.2 Treatability Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21014.3.3 Treatment Trains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
14.4 Measurement and Interpretation of Treatment Effectiveness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21114.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
ix
_________________________________________________________________________________
Contents (continued)
Chapter 15 Remediation Techniques for Contaminated Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
15.1 In Situ versus Prepared Bed Soil Remediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21515.2 In Situ Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
15.2.1 Soil Vacuum Extraction (SVE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21515.2.2 Bioremediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22915.2.3 Immobilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23215.2.4 Contaminant Mobilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
15.3 Prepared Bed Reactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23615.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
Chapter 16 Aquifer Restoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
16.1 Product Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24316.2 Pump-and-Treat Remediation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24416.3 Biorestoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
16.3.1 Example of the Use of Bioremediation: A Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25216.3.2 Advantages and Limitations in the Use of In Situ Bioremediation . . . . . . . . . . . . . . . . . . . . . . . 257
16.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Acknowledgments
This publication is based on the content of a series of U.S. Environmental ProtectionAgency (EPA) technology transfer seminars that were conducted in all ten EPA Regions,October 1989 through February 1990. This project was funded by the Office of SolidWaste and Emergency Response (OSWER) and the Office of Research and Development(ORD) to assist regulators and technical specialists in selecting the most appropriateremediation technologies for contaminated soils and ground water at Superfund sites.Seminar development was the responsibility of ORD staff in the Center for EnvironmentalResearch Information (CERI), Cincinnati, OH, and the Robert S. Kerr EnvironmentalResearch Laboratory (RSKERL), Ada, OK. Dominic DiGiulio, RSKERL, provided techni-cal direction for seminar development and publication review. Marion R. Scalf, RSKERL,and Carol Grove, CERI, were project managers. Seminars were held in October 1989(Chicago, IL; Kansas City, MO; Denver, CO; and Dallas, TX); November 1989 (Lowell,MA, and New York, NY); January, 1990 (Atlanta, GA, and Philalelphia, PA); andFebruary, 1990 (Seattle, WA, and San Francisco, CA).
Principal participants in the project include:
Michael Barcelona, Institute for Water Science, Western Michigan University,Kalamazoo, MI
J. Russell Boulding, Eastern Research Group, Inc., Arlington, MA William Fish, Department of Environmental Science and Engineering, Oregon
Graduate Institute of Science and Technology J. Michael Henson, RMT Engineering and Environmental Management Services,
Greenville, SC Richard Johnson, Department of Environmental Science and Engineering, Oregon
Graduate Institute of Science and Technology James Mercer, GeoTrans, Inc., Sterling, VA Carl Palmer, Department of Environmental Science and Engineering, Oregon
Graduate Institute of Science and Technology Judith Sims, Utah Water Research Laboratory, Utah State University, Logan, UT Ronald Sims, Department of Civil and Environmental Engineering, Utah State
University, Logan, UT Charles Spalding, GeoTrans, Inc., Sterling, VA
Eastern Research Group, Inc., Arlington, MA, provided technical, editorial, andproduction support for the project under Contract 68-C8-0014. Russell Boulding contrib-uted as author, editor, and reviewer of the document Trisha Hasch provided seminarcoordination; and Karen Ellzey, Susan Richmond, Heidi Schultz, and Denise Shortprovided editorial and production support.
x
Chapter 1Integrating Site Characterization with Subsurface Remediation
Ronald C. Sims and Judith L. Sims
This handbook on site characterization for subsurfaceremediation emphasizes processes and concepts (Parts I andII), characterization tools and analyses of data (Part I), andremediation decisions (Part III). Chapter 1 relates subsurfacesite characterization activities to the selection of subsurfaceremediation technologies. Chapters 2 through 16 each addressa specific aspect of site characterization or remediation tech-nology (e.g., geologic aspects, saturated zone, unsaturatedzone, remediation techniques for contaminated soils).
1.1 Approach for Integration of SiteCharacterization with SubsurfaceRemediation
Chapter 1 integrates the information presented in Chap-ters 2 through 16 so that the reader is guided through theHandbook and may access necessary interdisciplinary infor-mation quickly and efficiently for specific remediation appli-cations. The tables, checklists, figures, and flow charts in thischapter synthesize relevant terms, parameters, and conceptsrelating site characterization to specific subsurface remedia-tion techniques. Using this information to select subsurfacetreatment technologies requires specific information that isinterdisciplinary, thereby cutting across areas of specializa-tion, i.e., chapters. Therefore, this chapter not only providesan index to the Handbook, but also provides comments andguidance about the relationship between characterization pa-rameters and technology selection.
This chapter also discusses the importance of understand-ing the surface physical layout of a site, including culturalfeatures and industrial structures (e.g., buildings, lots, produc-tion units) and the evaluation of historical records of produc-tion and waste management within the context of sitecharacterization for subsurface remediation. Activities suchas making site visits and obtaining historical records of siteand waste management are an integral part of site character-ization. Information from these activities, which can providevaluable insights concerning limitations as well as applica-tions of remediation technologies at field scale, is referred tocollectively as site reconnaissance information.
1.2 Subsurface Site Characterization forRemediation Technology Selection
A methodology for integrating site characterization withsubsurface remediation is shown in Figure 1-1. The develop-
ment of information for a specific site progresses from charac-terization through monitoring (left to right as illustrated acrossFigure l-l). The figure presents characterization needs interms of waste interaction with unsaturated soil in the vadosezone or sediment or aquifer material in the saturated zone asinfluenced by site factors such as climate, topography, surfaceslope, etc. Information from site characterization is used toformulate, in qualitative and quantitative terms, the problem(s)in terms of pathways of migration, escape, and/or exposure ata contaminated site (problem assessment). This information isused for subsurface treatment technique evaluation, elimina-tion of unsuitable technique(s), and selection of an appropri-ate treatment (train). Monitoring provides feedback on rateand extent of remediation at field scale as well as informationfor modification of site management. Sections 14.1 and 14.2present this methodology in more detail.
Table 1-1 lists specific aspects of each step of the meth-odology, presents relevant concepts, and indicates referencesin the Handbook for additional information on each step of themethodology. Specific characterization parameters are relatedto problem assessment, treatment, and monitoring. For ex-ample, the distribution coefficient, Kd, will allow evaluationof the problem at a site with regard to migration. If soilflushing is selected as a treatment technique, it may be moni-tored effectively through pore-liquid phase sampling. Infor-mation on each aspect can be found in the sections in theHandbook listed under Text Reference (Section) on the table.
Subsurface-based waste characterization information needsare summarized in Table 1-2. Potential impacts of waste onground-water, vadose-zone, atmosphere, and surface-waterresources depend upon properties of the waste chemicals andproperties of the affected matrix. Information on these proper-ties is necessary to adequately assess the problem at a specificsite, as described above. Table 1-2 presents individual param-eters and text references for describing those parameters.
Figure 1-2 illustrates problem assessment in terms ofcompartments as well as pathways of migration for chemicalmigration, escape, and/or exposure. A mass balance concep-tual approach to the subsurface identifies chemicals that will(1) migrate upward (volatilization); (2) migrate downward(leaching, pure product); (3) migrate laterally (aqueous plumeand pure product); and (4) remain in place as persistentchemicals. A nonaqueous phase liquid (NAPL) may be fur-ther classified as a light NAPL (LNAPL) if the density of the
Figure 1-1. Methodology for integrating site characterization with subsurface remediation.
Table 1-1. Methodology for Relating Site Characterization to Subsurface Remediation
a(c) = core material; (1) = pore liquid phase; (g) = gas phase.bSpecies are determined primarily for inorganics (metals) and affect metal phase (aqueous, solid, gas).
2
liquid is less than water, or as a dense NAPL (DNAPL) if thedensity of the liquid is greater than water. Additional informa-tion on the compartments comprising the subsurface is pre-sented in Section 14.1. Figure 1-2 also indicates references foradditional information on each topic.
Subsurface remediation techniques that may be evaluatedbased upon site characterization and problem assessment, asoutlined above, are summarized in Table 1-3 and presentedfor each technology-and-environment combination in Tables1-4 through 1-9. The tables are organized according to treat-ment category (biological, physical/chemical, and contain-ment) and environment (vadose zone and saturated zone).Each table also is organized according to characterizationparameters, comments, and text reference by sections in theHandbook. These tables can be used to quickly locate infor-mation within the Handbook that relates treatment technolo-gies to specific site characterization parameters.
1.3 Site ReconnaissanceSite reconnaissance activities include gathering informa-
tion on site layout, history, and records of management.Aboveground natural and cultural features and industrial pro-cesses are important aspects of a site that may affect subsur-face processes and the application of subsurface remediationtechnologies. Table 1-10 lists important site conditions thatcan be used as part of site characterization for subsurfaceremediation. Identification of these features and processesprovides critical information concerning potential vadose-
Table 1-2. Subsurface-Baaed Waste Characterization
aOrganic (acid, base, polar neutral, nonpolar neutral), and inorganic.bMolecular weight, melting point, specific gravity, structure, volubility,
ionization, cosolvation.COxidation, reduction, hydrolysis, precipitation, dissolution,
polymerization.dAdsorption, desorption, ion exchange.eBiotic, abiotic.fHenry’s Law partitioning, soil gas analysis, vacuum extraction.gIncludes gas, inorganic mineral solid, organic matter solid, water,
and nonaqueous phases.
3
zone and ground-water quality as well as limitations for theapplication of subsurface remediation technologies. A sitevisit may reveal the industrial processes or waste sources thatcontribute to contamination at a site. Observations of topogra-phy, buildings, parking lots, and waste facilities provide valu-able information on accessibility for sampling, culturallyinduced flow of gases (e.g., beneath buildings), and limita-tions or constraints to the application of subsurface treatmenttechnologies (e.g., site size constraints or natural boundaries).
Information on past waste management practices thatdocuments conditions under which hazardous waste has beenmanaged is important to site characterization. Table 1-11 listsimportant waste management data and records that can beused in planning a site characterization effort. These recordsmay include available history of waste disposal and wastecomposition. This information may be used in conjunctionwith subsurface core and pore-liquid characterization data todetermine areas of contamination and areas of nonhomogeneity,to evaluate the areal and depth extent of contamination, and tomodify a site characterization plan.
Figure 1-3 presents a flow chart demonstrating an itera-tive approach for data collection from site characterizationactivities for subsurface remediation evaluation and selection,as well as field optimization of remediation technologies. Thisapproaches combines site reconnaissance information withsite characterization and sampling, utilizing the methodologypresented in Figure 1-1, for selecting, evaluating, and apply-ing the subsurface remediation techniques addressed in Tables1-4 through 1-9.
Figure 1-2.
NAPL (10) (16)
Problem assessment for site characterization baaed on mass balance approach (Chapters 2, 12, and 14).
1 Site reconnaissance activities at a site include gathering information such assite layout, history, and records of management.
2 Treatment = biological, physical/chemical, or containment (Table 1-3).
Figure 1-3. Flow chart for evaluation of site characterization for subsurface remediation.
4
Table 1-3. Summary of Tables of CharacterizationParameters for Subsurface RemediationTechnologies
Tabie 1-4. Characterization for Biologicai Treatment of Soil in the Vadose Zone*
Parameter Comments Text Reference (Section)
*Approaches and specific techniques for treatment are addressed in Chapters 14 and 15 and listed in Tables 15-3 and 15.4.
5
Table 1-5. Characterization for Bioiogical Treatment of Aquifer Materiai in the Saturated Zone*
Parameter Comments Text Reference (Section)
Physicaltemperature
permeability
geoiogy
geochemistry
pH
oxygen availability
interphase transferpotential
Chemicalindividual chemicals
redox potential
C:N:P ratio/nutrient
Biologicalkinetics/activity
enumeration
toxicity
metabolism
treatabiiity studies
adaptation
affects microbial activity/kinetics
affects nutrient suppiy and gas exchange
influences heterogeneity, day lensesinfluences waste distribution
influences microbial activity
affects chemical form (mobiiity) andmicrobial activity
affects aerobic/anaerobic metabolism, andactivity kinetics
used in mass balance to determine abioticremoval
affects rate and extent of degradation
often controlled by microorganisms andrelated to aerobic/anaerobic pathway
affects microbial growth
affects rate of degradation
related to population or mass of microorganisms
affects rate and extent of degradation
influences production of (toxic) intermediatesand indicates mechanism(s) of biodegradation
can indicate potential for degradation andimportant factors controlling rate and extent
ability of system to acclimate, indicated byincrease in rate and extent of degradation withincubation time and with repeated exposure
* Approaches and specific techniques for treatment are addressed in Chapters 14 and 16.
6
Table 1-6. Characterization Parameters for Physical/Chemical Treatment of the Vadose Zone*
Technique Parameter Comments Text Reference (Section)
Extraction
Oxidation/Reduction
Solidification/Stabilization
physicalparticle size distributionconductivity/permeabilityorganic mattermoisture content
heterogeneity/layeringdepth
soil gas
interphase transfer potential
chemicalindividual chemicalspH changes
chemicai characteristicsc
cation exchange capacity
organic and metal content
redox potential
individual chemicalsredox potential
individual chemicals/sitesporosity/permeability
affects volume reduction, sorption, extraction difficultyaffects flow velocity (time) for extractionaffects distribution and sorption of chemicalsaffects conductivity of air through soil for vacuum
extractionaffects relative rates of extraction for different layersalong with area, determines volume of contaminated
material and engineering strategies for extractionused along with soil core analysis to monitor extent
and rate of vacuum extractionused in mass balance to determine treatment
effectiveness
examples of chemicals that have been treatedmay indicate precipitation or dissolution that affects
ease of extraction (permeability)aids in selection of extraction fluid
determine cation sorption potential, related to claycontent
determine target and/or interfering constituents,pretreatment needs, extraction fluid
indicates mobile and immobile forms of chemicals
examples of chemicals that have been treatedstatus of the system before treatment
examples of chemcals/sites that have been treatedaffects delivery and mixing of chemicals
(15) (15.2.4)
(3.2)(5.2.4) (15.2.4)(3.2)(4.2.2) (5.1) (5.2.3)
(3.1) (3.3)(3.5.1) (5.1)
(5.2.5)
(5.2.5) (14.1) (14.4)
(5.4) (15)(8.1.2)
(5.1.5) (5.4) (7.4.2) (8.2)(10.3.1) (10.4) (11.1) (11.2.3)(12. 1) (12.3)
a Approaches and specific techniques for treatment are addressed in Chapters 14 and 15 and are listed in Tables 15-1, 15-2, 15-3, and 15-4.b Extraction techniques include aqueous, solvent, critical fiuid, vacuum (air/steam), and low temperature thermal stripping.C Chemical characteristics include vapor pressure, solubility, Henry’s Law constant, partition coefficient, boiling point, and specific gravity.
7
Table 1-7. Characterization Parameters for Physical/Chemical Treatment of the Saturated Zone
Tech- Parameternique
Comments Text Reference (Section)
Product (16)(16.1)Removal
physicalparticle size distribution affects amount of contaminant stored (3.2)(10.2)
in vadose zone in capillary fringe for LNAPLparticle size distribution affects permeability and product (3.2)(8.1.2)
in saturated zone retentionflow characterization affects direction, location, and (2.2)(2.4)(10.2)
extent of LNAPLgeology influences distribution of DNAPL (3)(6.1)
and LNAPLorganic matter affects distribution and sorption (3.2)(10.3.2)interphase transfer assists in determining phase(s) where (5.2.5)(14.1)(14.4)
potential more than one phase is involved
chemicalindividual chemicals, examples of contaminants that have (3.5.2)(4.4)(11.3)(14.3)
contaminants been treatedredox temporal and spatial variation may (8.1.2)(8.3.3)(8.4)(9.1.3)(13.2.2)
influence permeabilitysoil gas analysis assist in locating contamination (area) (5.2.5)(9.2.1)properties b assist in locating contamination (depth) (10.2)(10.4)(11.1)(11.2)
Pump-and-Treat
physicalparticle size distribution
in saturated zoneflow characterization
geologyorganic matterinterphase transferpotential
(16)(16.2)
affects pumping (recovery) rate ofwater and contaminant
affects direction, location, and (2.2)(2.3)(2.4)(4.1)(4.2.2)extent of contamination (4.2.3)(6.1)
influences distribution of contaminants (3)(6.4)affects distribution and sorption (3.2)(10.3.2.)assists in determining phase(s) where (14.1)(14.4)
contaminant is found
chemicalindividual chemicals, examples of contaminants that have (3.5.2)(4.4)(4.5)
contaminants been treatedredox temporal and spatial variation may (8.1.2)(8.3.3)(8.4)(9.1.3)(13.2.2)
influence permeabilitysoil gas analysis assist in locating contamination (area) (5.2.5)(9.2.1)propertied assist in locating contamination (depth) (10.2)(10.4)(11.1)(11.2)(12.1)(12.2)organic-inorganic affects design of systems (12.3)
interactionsa
Approaches and specific techniques for treatment are addressed in Chapters 14 and 16, Section 16.1.b
Properties include molecular weight, specific gravity, volubility, melting point, structure, ionization, and cosolvation.
8
Table 1-8. Characterization Parameters for lmmobilizationa/Containment b Techniques in the Vadose Zonef
Technique Parameter Comments Text Reference (Section)
Immobilization (15)(15.2.3)physicalparticle size distributionmoisture content
permeabilityorganic matterdepth
interphase transferpotential
chemicalindividual chemicalscontaminantsredox potential
pH
cation exchange capacityproperties
Containmentphysicalstratigraphy
interphase transferpotentialcontainment requirements
affects sorption, ion exchangeaffects efficiency, energy requirements,
and sorptionaffects delivery of chemicalsaffects distribution and sorptionalong with area, determines volume of
contaminated material and engineeringstrategies
affects extent of sorption and ionexchange
used in mass balance to evaluatesolution to solidphase transfer forimmobilization
examples of contaminants that havebeen treated
affects chemical speciation and thusimmobilization
affects chemical speciation and thusimmobilization
affects ion exchangeaffects affinity of chemicals for
surfaces and for precipitation
identify path ways and extent ofchemical migration
used in mass balance to evaluatesuccess of containment
evaluate containment of gas, liquid,and solid phases
(3.2)(5.1)(5.2.3)(14.1)(15)(15.2.3)
(5.2.4)(15.2.3)(3.2)(10.3.1)(14.1)(3.5.1)(5.1)
(3.2)
(5.2.5)(14.1)(14.2.2)(14.4)
(5.5)
(8.1.2)
(8.1.2)
(8.1.2)(8.1)(8.3)(10.3.1)(10.3.2)(10.4)(11.1)(12.1)(14.1)(14.2.1)
(3.1)
(5.2.5)(14.1)(14.2.2)(14.2.3)(14.4)
(14.2.3)
a immobilization techniques include sorption, ion exchange, precipitation, stabilization/solidification, and vitrification.
b Containment techniques include physical stuctures.
c Approaches and specific techniques for treatment are addressed in Chapters 14 and 15, Section 15.2.3.
d Properties include molecular weight, melting point, specific gravity, structure, ionization, solubility, and cosolvation.
9
Table 1-9. Characterization Parameters for Containment Technlquesa in the Saturated Zone b
Technique Parameter Comments Text Reference (Section)
Hydraulicphysicalflow system characterization
permeability
geology
advection
interphase transferpotential
fracture flow
physical gradients
chemicalcontaminants present
individual chemicals/contaminants
environmental parametersc
chemical gradients
propeties d
reactions e
PhysicalStructuresf
physicalflow system characterization
geology
fracture flow
chemicalcontaminants present
individual chemicals/contaminants
determine area and depth forcontainment
affects rate of movement and rate ofpumping
assists with flow systemcharacterization
generally primary transport (escape)path
used in mass balance to assess andevaluate containment
may exercise control on ground-waterflow
affects geochemistry, which may affectpermeability
identify chemicals of concern thatmight escape
examples of contaminants that havebeen contained
may change with pumping and affectrecovery and permeability
may affect geochemistry if reinfectedand affect permeability
affects affinity of chemicals forsurfaces and for precipitation, as wellas interphase transfer
may affect treatment/permeability whilepumping
determine area and depth forcontainment
assists with flow systemcharacterization
may exercise control on ground-water flow
identify chemicals of concern thatmight escape
examples of contaminants that havebeen contained
(2.2)
(4.4)
(3)
(4.1)
(5.2.5)(14.1)(14.2.2)(14.4)
(6.1)
(8.3.2)
(2.3)
(3.5.2)(4.4)(4.5)
(8.1.2)
(8.3.3)
(10.2)(10.3.1)(10.3.2)
(13.1)(13.2)
(2.2)
(3)
(6.1)
(2.3)
(3.5.2)
a Containment techniques may be temporary and used as part of a treatment train that includes product removal, pump-and-treat, pumpingand reinjection, and bioremediation.
b Approaches and specific techniques are addressed in Chapters 4 (section 4.4), 14 and 16 (Section 16.2).C Environmental parameters include pH, alkalinity, redox potential, salinity, temperature, and pressure.d Properties include molecular weight, melting point, specific gravity, structure, solubility, ionization, and cosolvation.e Reactions include hydrolysis, substitution, elimination, oxidation-reduction, and biodegradation.d Physical structures often are used in conjunction with hydraulic containment and withdrawal (e.g., clay cap to reduce recharge combined with
extraction wells to remove chemical) (refer to Section 3.5.2).
10
Table 1-10. Aboveground Features for Site Characterization
Item Specific Information
Site location Topographic map, including contours, map scale and date, floodplain areas, surface waters, springs andintermittent streams, and site legal boundaries.
Site map, including injection and withdrawal wells on site and off site; buildings and recreation areas, accessand internal roads; storm, sanitary, and process sewerage systems; loading and unloading areas; and firecontrol facilities.
Location of past ano/or present operation units and equipment cleaning areas, ground-water monitoring wells,delineation of waste management units, and site modifications.
Surrounding area land use patterns.
Vegetation (trees, shrubs, grasses).
Climatological data Precipitation/evaporation/humidity.
Site water budget.
Temperature (averages and extremes)
Wind rose.
Predicted storm events (e.g., 24-hour, 25-year, average number of days of rain and snow).
Frost action potential
Table 1-11. Waste Management Information for Site Characterization
Category Item Specific Information
History of waste application Years in operation and annual Records of measured annual waste quantity (weight/volume)quantity of waste generated over life of site. Include hazardous and nonhazardousand/or disposed. managed at same site.
Placement of waste. Records of quantity (weight/volume), and location of eachwaste disposal action.
Size of waste unit(s) Area and depth.
History of waste quality Waste analyses. Periodic analyses of hazardous. wastes.
Unit processes. History of unit processes employed in the generation andtreatment of wastes.
Disposal areas. Pits, ponds, lagoons, landfills, storage tanks, wastewatertreatment plant locations (present and historical).
11
PART I: METHODS FOR SUBSURFACE CHARACTERIZATION
Chapter 2Site Characterization Overview
James W. Mercer and Charles P. Spalding
2.1 Introduction
Characterization of a hazardous waste site involves gath-ering and analyzing data to describe the processes controllingthe transport of wastes from the site. It provides the under-standing to predict future site behavior based on past sitebehavior. It can encompass the characterization of the wasteitself as well as that of various transport pathways such as air,surface water, biota, and ground water. Ground water, thefocus of this discussion, is often the most significant and leastapparent transport pathway.
Site characterization follows the scientific method and isperformed in phases (see Figure 2-l). First, a hypothesis ismade concerning site or system behavior. Based on thishypothesis, a data collection program is designed, data arecollected, and an analysis or assessment is made. Using theresults of the analysis, the hypothesis is refined and additionaldata may be collected. As the knowledge of the site becomesmore detailed, the working hypothesis may take the form ofeither a numerical or analytical model. Data collection contin-ues until the hypothesis is proven sufficiently to form thebasis for decision making.
Because the ultimate goal of site characterization is tomake informed decisions, the first step is to define studyobjectives. A possible list of objectives, provided by Cartwrightand Shafer (1987), includes the following: (1) assess thebackground or “ambient” water quality (how was the waterbefore contamination?); (2) establish the impacts of certainfacilities, practices, or natural phenomena on water quality(what is the extent of contamination?); and (3) predict futureground-water quality trends under a variety of conditions(what would be the impact of various remedial actions?).
Whatever the objectives, ground-water site characteriza-tion has two major components: assessment of the ground-water flow system and assessment of the contamination in theground water. All too often, emphasis is placed on the lattercomponent, which involves ground-water quality monitoring,
13
Everett (1980) defines monitoring as a scientifically designedsurveillance system of continuing measurement and observa-tions. At many waste sites, ground-water quality data areabundant; however, water-level data used to determine advec-tive transport are limited. This is unfortunate because water-level data are equally important, and they are easier and lessexpensive to collect than water-quality data.
This chapter provides an overview of Part I of the Hand-book, which focuses on methods of site characterization. Thischapter covers the following topics (1) flow system charac-terization, (2) contamination characterization, (3) techniquesfor characterization, and (4) analysis of data.
2.2 Flow System CharacterizationFlow system characterization begins with an understand-
ing of controlling processes and of the data required to definethose processes (Table 2-l). Ground water is always in motionfrom areas of natural and artificial recharge to areas of naturaland artificial discharge. Natural recharge occurs from precipi-tation and surface water bodies; artificial recharge resultsfrom human-induced actions such as irrigation and well injec-tion. Ground water discharges naturally to springs and othersurface water bodies, e.g., rivers, lakes, and oceans. Undernatural conditions, ground water moves very slowly, its flowvelocity ranging from a fraction of a foot per year to severalfeet per day. In most cases, flow obeys Darcy’s law, whichstates that the velocity is proportional to both the hydraulicconductivity of the formation and the hydraulic gradient.. Theterm hydraulic conductivity is used to express the water-conducting capacity of the formation material. The hydraulicgradient is an expression of the slope of the ground-watersurface.
Shallow aquifers are usually important sources of groundwater. These upper aquifers are also the most susceptible tocontamination. Contaminants may enter an upper aquifer inone of the following ways: (1) artificial recharge or leakagethrough wells; (2) infiltration from precipitation or irrigation
and lakes, (4) spring discharge, and (5) evapotranspiration.Data required to assess these processes are shown in Table 2-2. In general, these data requirements include a geometricdescription of the site (layering and hydraulic boundaries);storage and transmissive properties; and source/sink informa-tion. such as wells. More specific lists of data with ranges ofvalues are provided in Mercer et al. (1982). For any particularsite, it is rare to have all this information. Data gaps can beaddressed by a field collection program, but to some extentmust be filled based on experience. In addition to physical andchemical data, other factors listed in Table 2-2 include regula-tory and legal issues such as water rights and future land use.
The first step in designing a field program is to reviewexisting data for the site or nearby locations. Sources ofinformation include the U.S. Geological Survey (USGS) (Mer-cer and Morgan, 1981); state geologic and water agencies;local water districts; and city, county, and state health depart-ments. Other federal agencies that may provide data includethe U.S. Environmental Protection Agency (EPA) (e.g., theSTORET computerized information storage system); U.S.Bureau of Reclamation; U.S. Army Corps of Engineers; and
Figure 2-1. Site characterization phasea (from Bouwer et al.,1988).
return flow through the vadose zone above the water table; (3)induced recharge from influent streams and lakes or othersurface water bodies; (4) inflow through aquifer boundariesand leakage from overlying or underlying formations; and (5)leakage or seepage from impoundments, landfills, or miscel-laneous spills.
Water and contaminants carried with it may leave anaquifer in the following ways: (1) ground-water leakage fromthe aquifer into adjacent strata, (2) ground-water withdrawalby pumping and drainage, (3) seepage into effluent streams
U.S. Soil Conservation Service. Additional inforation maybe available from consultants and universities. Several datasources are discussed below.
The U.S. Department of Agriculture, Soil ConservationService, has three soil geographic data bases the Soil SurveyGeographic Data Base (SSURGO), the State Soil GeographicData Base (STATSGO), and the National Soil GeographicData Base (NATSGO). Components of map units in eachgeographic data base are generally phases of soil series. TheSoil Conservation Service also maintains a soil interpretationsrecord data base, which encompasses more than 25 soil,physical, and chemical properties for the 15,300-plus soilseries recognized in the United States. Interpretations aredisplayed differently for each geographic data base to beconsistent with the level of detail expressed. Particle sizedistribution, bulk density, available water capacity, soil reac-tion, salinity, and organic matter are included for each majorlayer of the soil profile. Data on flooding, water table, bed-rock, and subsidence characteristics of the soil; and interpreta-tions for erosion potential, septic tank limitations, engineering,building and recreation development, and cropland, wood-land, wildlife habitat, and rangelands management also areincluded.
The U.S. Department of Interior Geological Survey cre-ated and maintains a central storage facility for water re-sources data, known as the National Water Data Storage andRetrieval System (WATSTORE), at its National Headquar-ters in Reston, Virginia. Included in this computerized storagefacility are representative ground-water data collected through-out the United States, This ground-water information residesin a computer data file, which is maintained by a databasemanagement system (DBMS) called SYSTEM 2000. Thename and acronym given this data base is the Ground-WaterSite-Inventory (GWSI) file. Although several field-collectedparameters of water-quality data (including temperature, con-ductance, and pH) are stored in the GWSI, the bulk of water-quality data reside in a nationwide file called Storage andRetrieval (STORET), a file maintained by EPA. The National
14
Table 2-1. A Summary of the Processes Associated with Dissolved Solute Transport and Their Impact
Process Definition Impact on Transport
Solute Transport
Advection Movement of solute as a consequence of ground-water flow. Most important way of transporting solute awayfrom source.
Diffusion Solute spreading due to molecular diffusion in response to An attenuation mechanism of second order inconcentration gradients. most flow systems where advection and
dispersion dominate.Dispersion Fluid mixing due to effects of unresolved heterogeneities in An attenuation mechanism that reduces solute
the permeability distribution. concentration in the plume. However, itspreads to a greater extent than a plumemoving by advection alone.
Solute Transfer
Radioactive decay
Sorption
Dissolutionprecipitation
Acid-basereactionsComplexation
Hydrolysis/substitution
Redox reactions(biodegradation)
Biologically MediatedMass Transfer
Biological transfor-mations
Irreversible decline in the activity of a radionuclide through a An important mechanism for attenuation when thenuclear reaction. half-life for decay is comparable to or less
than the residence time of the flow system.Also adds complexity in production of
daughter products.Partitioning of a solute between the ground water and mineral An important mechanism that reduces the rate at
or organic solids in the aquifer. which the solute is apparently moving. Makes itmore difficult to remove solute at a site.
The process of adding solutes to or removing them from solution Precipitation is an important attenuationby reactions dissolving or creating various solids. mechanism that can control the concentration in
solution. Solution concentration is mainlycontrolled either at the source or at a reactionfront.
Reactions involving a transfer of protons (I-P). Mainly an indirect control on solute transport bycontrolling the pH of ground water.
Combination of cations and anions to form more complexion. An important mechanism resulting in increasedvolubility of metals in ground water, if adsorptionis not enhanced. Major ion complexation will
increase the quantify of a solid dissolved insolution.
Reaction of a halogenated organic compound with water Often hydrolysis/substitution reactions make anor a component ion of water (hydrolysis) or with organic compound more susceptible toanother anion (substitution). biodegradation and more soluble.
Reactions that involve a transfer of electrons and An extremely important family of reactions ininclude elements with more than one oxidation state. retarding solute spread through precipitation of
metals.
Reactions involving the degradation of organic compounds Important mechanism for solute reduction, but canand whose rate is controlled by the availability of lead to undesirable daughter products.nutrients to adapted microorganisms and redox conditions.
From NRC, 1990
Water Data Exchange (NAWDEX) Local Assistance Centersare authorized users of the STORET file and may retrieveground-water quaIity data for subscribers.
A field program usually follows a data review of hydro-geologic investigation techniques (U.S. EPA, 1986 and Sisk,1981). Summaries of procedures for well installation andaquifer testing are described in Ford et al. (1984) and Aller etal. (1989). Kruseman and de Ridder (1976), Lawrence Berke-ley Laboratory (1977, 1978) discuss methods of analysis ofaquifer and slug tests. In general, as the scale of the observa-tion increases, the range of measured properties, such ashydraulic conductivity, tends to change because of the hetero-geneous nature of geologic materials. Particularly, ground-water flow rates estimated from measurements on cores may
underestimate ground-water flow rates in the area if flow is infractures or in other more permeable layers.
Because of seasonal changes in ground water, a minimumof one year should be devoted to characterization. As the sitecomplexity increases, this period will increase proportion-ately. Several factors influence the number of boreholes re-quired, the most important being heterogeneities in the aquifermaterials. Methods of quantifying ground-water networks arenot widely used but do exist. For example, van Geer (1987)shows how Kalman filters are used to design ground-watermonitoring networks. Another technique used to evaluateground-water networks is kriging (e.g., Olea, 1982); thistechnique is discussed further in Section 2.5 and in Chapter 7(Section 7.3.2).
15
Table 2-2. Data Pertinent to the Prediction of Ground-Water Flow
I. Physical Framework1. Hydrogeologic map showing areal extent and boundaries of aquifer.
2. Topographic map showing surface-water bodies.
3. Water-table, bedrock-configuration, and saturated-thickness maps.
4. Hydraulic conductivity map showing aquifer and boundaries.
5. Hydraulic conductivity and specific storage map of confining bed.
6. Map showing variation in storage coefficient of aquifer.
7. Relation of stream and aquifer (hydraulic connection).
Il. Stresses on System1. Type and extent of recharge areas (irrigated areas. recharge basins, recharge wells, impoundments, spills, tank leaks, etc.).
2. Surface water diversions.
3. Ground-water pumpage (distributed in time and space).
4. Stream flow (distributed in time and space).
5. Precipitation and evapotranspiration.
Ill. Observable Responses1. Water levels as a function of time and position.
IV. Other Factors1. Economic information about water supply.
2. Legal and administrative rules.
3. Environmental factors.
4. Planned changes in water and land use.
After Moore, 1979
2.3 Contamination CharacterizationAs with flow system characterization, contamination char-
acterization begins with understanding the processes control-ling transport and degradation (Table 2-1) and the data requiredto define those processes. These processes determine mini-mum data requirements needed to characterize a site.Nonreactive (conservative) dissolved contaminants in satu-rated porous media are controlled by the following factors:
1. Advection: This mechanism causes contaminants tobe transferred by the bulk motion of the groundwater. The term convection is sometimes used inplace of advection.
2. Mechanical (or kinematic) dispersion: This processinvolves meehanical mixing caused by three mecha-nisms. The first mechanism occurs in individual porechannels because molecules travel at different ve-locities depending on whether they are near the edgeor in the center of the channel. The second mecha-nism is triggered by differences in surface area androughness relative to the volume of water in indi-vidual pore channels, causing different bulk fluidvelocities in different pore channels. The third mecha-nism is related to the tortuosity, branching, andinterfingering of pore channels, causing the stream-lines to fluctuate with respect to the average flowdirection. Mechanical dispersion occurs in the direc-tion of the average flow velocity and in the planeorthogonal to the average flow direction, These ef-fects are called longitudinal dispersion and trans-verse dispersion, respectively, Longitudinal dispersion
is due to variations of the velocity component alongthe average flow direction, whereas transverse dis-persion is due to variations of the velocity compo-nents in the normal plane.
3. Molecular diffusion: Fickian diffusion causes thecontaminant molecules or ions to move from highconcentrations to lower concentrations. Movementalso is caused by the random kinetic motion of theions or molecules (Brownian diffusion).
The combined effect of mechanical dispersion and mo-lecular diffusion is known as hydrodynamic dispersion. Dis-persion causes the zone of contaminated ground water tooccupy a greater volume than if the contaminant distributionwere influenced only by advection. If a slug of contaminantenters the ground-water system, advection causes the slug tomove in the direction of ground-water flow. Hydrodynamicdispersion causes the volume of the contaminated zone toincrease and the maximum concentration in the slug to de-crease. Transverse dispersion may expand a contaminant plume10 to 20 percent beyond the width defined by convectivetransport (Lehr, 1988). Macroscopic variations in hydraulicconductivity and porosity are probably more significant fac-tors affecting solute transport than hydrodynamic 1 dispersionchanges (Wheatcraft, 1989).
Additional processes affect transport for reactive con-taminants. In addition to advection and hydrodynamic disper-sion, the migration of reactive contaminants is further controlledby adsorption, desorption, chemical reactions, and biologicaltransformation.
16
1. Adsorption or desorption: These processes involvemass transfer of contaminants. Adsorption is thetransfer of contaminants from the ground water tothe soil. Resorption is transfer of contaminants fromthe soil to the ground water.
2. Chemical reactions: These processes involve masstransfer of contaminants caused by various chemicalreactions (e.g., precipitation and dissolution, oxida-tion and reduction). For some contaminants, degra-dation is also an important process that may need tobe characterized.
3. Biological Transformation: These processes may re-move contaminants from the system by biologicaldegradation, or transform contaminants to other toxiccompounds that are subject to mass transfer by theother processes discussed above.
The processes of adsorption-desorption, chemical reac-tions, and biological transformation play important roles incontrolling the migration rate as well as concentration distri-butions. These processes tend to retard the rate of contaminantmigration and act as mechanisms to reduce concentrations.Because of their effects, the plume of a reactive contaminantexpands and the concentration changes more slowly thanthose of an equivalent nonreactive contaminant (see Figure 2-2). As discussed in subsequent chapters, however, resorptioncan require longer time periods to reach concentration cleanupstandards.
Table 2-3 shows data requirements for contaminationcharacterization, in addition to the requirements shown inTable 2-2. For example, to characterize advective transport,the flow system must first be understood. More specific listsof data with ranges of values are provided in Mercer et al.(1982). These data requirements provide a broad view of thefactors affecting contaminant transport from a site.
2.4 Techniques for CharacterizationFor site characterization, it is important to understand the
transport mechanisms and ground-water flow system at a site.Once these mechanisms and systems are understood, ground-water monitoring data can be interpreted to obtain informationfar more useful than simple information on contaminant levelsat specific points and times. The procedures used to obtainwater-quality data are of critical importance. Procedures fordrilling monitoring wells, taking samples, and having samplesanalyzed by a laboratory are discussed in this section.
Table 2-4 shows actions that were typically taken athazardous waste sites in the early 1980s. Two data gaps arethe vertical distribution of hydraulic head, as measured bywater levels in adjacent wells cased to different depths, andhydraulic conductivity values. Therefore, most guidance docu-ments now recommend the actions shown in Table 2-5. Atsites where conditions warrant (e.g., fractured media), addi-tional actions may be necessary to fully characterize the site(see Table 2-6).
A variety of common well drilling methods maybe usedto install monitoring wells at hazardous waste sites. Thesemethods include solid stem continuous flight and hollow stemcontinuous flight augering, cable tool drilling, mud and airrotary drilling, jetting, and driving well points. Detailed dis-cussions of the principles of operation of each of these meth-ods are available from numerous sources including Scalf et al.(1981), Driscoll (1986), and Campbell and Lehr (1973). Asummary of the advantages and disadvantages of variousdrilling methods relative to monitoring well construction isprovided in Scalf et al. (1981) and Larson (1981), as well as inChapter 4 of this Handbook (Section 4.2.1).
A variety of materials are available for use in casing,screening, and other structural and sampling components ofmonitoring wells. The most commonly used are mild steel,stainless steel, polyvinyl chloride (PVC), polypropylene, poly-ethylene, and Teflon®. Barcelona et al. (1983) summarizes thecharacteristics of several of these materials. These materialshave substantially different properties relative to strength,corrosion resistance, interference with specific contaminantmeasurements, expense, and availability. Consequently, theymust be selected carefully and demonstrated to be the mostappropriate for the particular monitoring program. Consider-ations should include all pertinent, site-specific factors suchas well installation method, depth, geochemical environment,and probable contaminants to be monitored. Well casingmaterials are discussed further in Section 4.2.1 (see especiallyTable 4-3) and Section 9.3.4.
Construction details for individual wells should be docu-mented and verifiable through the use of drilling logs. Thedrilling log should contain information about the texture,color, size, and hardness of the geologic materials encoun-tered during the drilling (Barcelona et al., 1985). Any use ofdrilling fluids, grouts, and seals also should be noted in therecord of well construction. Well casing materials should bedocumented because the type of well casing may have aneffect on the quality of the water samples (Barcelona et al.,1983). The same considerations that apply to well casingmaterials for newly constructed monitoring wells apply toevaluating the suitability of existing wells for ground-waterquality monitoring.
Guidance documents on ground-water monitoring em-phasize the need for depth-discrete data to determine thethree-dimensional flow field and chemical distribution(Barcelona et al., 1983; Barcelona et al., 1985; and U.S. EPA,1986). Shorter well screens and more nested wells are recom-mended where immiscible liquids (liquids that tend to floatabove water or sink to the bottom), heterogeneous conditions,or a thick flow zone are present (U.S. EPA, 1986). Barcelonaet al. (1983) recommend installing nested wells with shortwell screens (less than 5 ft long) where the potential flow zoneis more than 10 ft thick.
Once the wells are designed and drilled, accepted practiceis to remove fluid from the formation, with subsequent labora-tory analysis of the sample (Morrison, 1983; de Vera, 1980;USATHAMA, 1982 Guswa et al., 1984; and Everett et al.,1984). This approach results in a set of point data that repre-sent (depending on the type of well construction, the sampling
17
Distance from Continuous Contaminant Source
Distance from Slug-Release Contaminant Source
Figure 2-2. The influence of natural processes on levels ofcontaminants downgradient from continuous andslug-release sources (from Keety et al., 1986).
mechanism, laboratory procedure, and hydrodynamics of theground-water system), particular aspects of the in situ waterquality at a specific time. Much work (Gibb et al., 1981;Gillham et al., 1983; Keith et al., 1983; Nacht, 1983; Barcelonaet al., 1984; Olea, 1984; Barcelona et al., 1985) has focusedon improving this process (i.e., providing greater qualitycontrol and quality assurance). Chapter 9 discusses samplingof subsurface solids and ground water in more detail.
2.5 Analysis of DataAlthough this section emphasizes network design and
sampling considerations, no section on data analysis would becomplete without a discussion of database management sys-tems (DBMS) and geographic information systems (GIS). Athazardous waste sites, large amounts of data are generated. Totake full advantage of these data in the interpretation stage,they should be in electronic/magnetic format for use in a
Table 2-3. Data Pertinent to Prediction of the Pollutants inGround Water (in addition to those in Table 2-2)
l. Physical Framework1. Estimates of the parameters that comprise hydrodynamic
dispersion.2. Effective porosity distribution.3. information on natural (background) concentration
distribution (water quality) in the aquifer.4. Estimates of fluid density variations and relationship of
density to concentration (most important wherecontaminant is salt water or results in significantly higherconcentration of total dissolved solids compared to thenatural aquifer or where there are significant temperaturedifferences between the contaminant plume and thenatural aquifer).
ll. Stresses on System1. Sources and strengths of pollutants,
lll. Chemical/Biological Framework1. Mineralogy media matrix.2. Organic content of media matrix.3. Ground-water temperature.4. Solute properties.5. Major ion chemistry.6. Minor ion chemistry.7. Eh-pH environment.
lV. Observable Responses1. Areal and tamporal distribution of water quality in the
aquifer.2. Stream flow quality (distribution in time and space)
DBMS and/or GIS. Both systems can be used to manipulate,correlate, and display data, and this method of organizinglarge amounts of data facilitates the interpretation process.
The assessment of ground-water quality on any scaleinvolves the estimation of chemical variables distributed inthree-dimensional space. A key consideration in establishingan effective and efficient ground-water quality monitoringprogram is the spatial distribution of sampling locations. Caremust be taken in designing monitoring well networks to avoidbiasing any inferences made from the resulting data.
As pointed out, knowledge of the hydrodynamics of theground-water system(s) being monitored is also of criticalimportance for the design of monitoring networks. For certainground-water monitoring program objectives, an optimummonitoring network for a relatively homogeneous porousflow environment is different from that for a discretely frac-tured hydrogeologic medium. For other monitoring objec-tives, however, the fundamental differences between flowregimes may have very little impact on the design of anoptimum sampling network.
Proper ground-water sampling and analysis are equallyimportant for assuring effective ground-water monitoring. Aquality assurance program composed of well-conceived andeffectively implemented quality control procedures should befollowed (Cartwright and Shafer, 1987). Strict adherence to
18
Table 2-4. Acthons Typically Taken
1. Install shallow monitoring wells.2. Sample ground water numerous times for a range of pollutants
such as those constituents contained in the RCRA AppendixIX ground-water monitoring list.
3. Define geology primarily by drillers’ logs and drill cuttings.4. Evaluate local hytrology with water level contour maps of
shallow wells.5. Possibly obtain soil and core samples for chemical analyses.
Benefits1. Screening of the site problems is rapid.2. Costs of investigation are moderate to low.3. Field and laboratory techniques used are standard.4. Data analysis/interpretation is straightforward.5. Tentative identification of remedial alternatives is possible.
Shortcomings1. True extent of site problems may be misunderstood.2. Selected remedial alternatives may not be appropriate.3. Optimization of final remediation design may not be possible.4. Cleanup costs remain unpredictable, tend to excessive levels.5. Verification of compliance is uncertain and difficult.
Modified from Keely et al., 1986
Table 2-5. Recommended Actions
1. Install depth-specific clusters of monitoring wells.2. lnitially sample for a range of pollutants, but subsequently,
become more selective.3. Define geology by extensive coring/sediment samplings.4. Evaluate local hydrology wth well clusters and geohydraulic
tests.5. Perform limited tests on sediment samples (grain size, clay
content, etc.).6. Conduct surface geophysical surveys (resistivity, EM, ground-
penetrating radar).
Benefits1. Conceptual understanding of site problems is more complete.2. Prospects are improved for optimization of remedial actions.3. Predictability of remediation effectiveness is increased.4. Cleanup costs are lowered, estimates are more reliable.5. Verification of compliance is more soundly based.
Shortcomings1. Characterization costs are somewhat higher.2. Detailed understanding of site problems is still difficult.3. Full optimization of remediation is still not likely.4. Field tests may create secondary problems (disposal of pumped
waters).5. Demand for specialists is increased, shortage is a key limiting
factor.
Modified from Keely et al., 1986
quality assurance programs minimizes both systematic andrandom errors, and maximizes the likelihood of collectingground-water samples in a manner that ensures the reliabilityof analytical determinations. As with monitoring networkdesign, a detailed understanding of the overall objectives ofthe monitoring program is a key factor in determining sam-pling and analysis requirements. See Chapter 7 for further
19
Table 2-6. Additional Actions Where Conditions Warrant Them
1.2.3.
4.
5.
6.
7.
8.
9.
Assume Table 2.5 as starting point.Conduct soil vapor surveys for volatiles and fuels.Conduct tracer tests and borehole geophysical surveys (neutron
and gamma).Conduct karst stream tracing and recharge studies, if
appropriate to the setting.Conduct bedrock fracture orientation and interconnectivity
studies, if appropriate.Determine the percent organic carbon and cation exchange
capacity of solids.Measure redox potential, pH, and dissolved oxygen levels of
subsurface.Evaluate sorption-desorption behavior by laboratory column and
batch studies.Assess the potential for biotransformation of specific
compounds.
Benefits1. Thorough conceptual understandings of site problems are
obtained.2. Full optimization of the remediation is possible.3. Predictability of the effectiveness of remediation is maximized.4. Cleanup costs maybe lowered significantly, estimates are
reliable.5. Verification of compliance is assured.
Shortcomings1. Characterization costs may be much higher.2. Few previous applications of advanced theories and methods
have been completed.3. Field and Iaboratory techniques are specialized and are not
easily mastered.4. Availability of specialized equipment is low.5. Need for specialists is greatly increased (it may be the key
limitation overall).
Keely et al., 1986
discussion of sources of error in sampling and considerationsin the development of quality assurance programs.
The results of laboratory analyses are only as reliable asthe samples, field standards, and blanks received (Cartwrightand Shafer, 1987). Therefore, to assure that representativesamples are provided to the laboratory, careful thought andpractice must be part of any sampling program. A representa-tive sample accurately reflects in situ conditions in proximityto the sample point at the time the sample was collected.Maintaining representative samples requires consideration ofwell purging, sample collection, and sample preservation.Barcelona et al. (1985) have prepared an extensive guide tothe practical aspects of ground-water sampling. (See alsoChapters 7, 8, and 9 of this Handbook.)
Parameter selection is an important aspect of the designof a sampling program. The types of hydrochemical measure-ments to be made affect the choice of sampling equipment andthe sampling methodology. Barcelona et al. (1985) state that itis often wise to obtain slightly more chemical and hydrologicdata than immediately required in order to aid subsequentinterpretation. Sections 9.2.1 and 9.3.1 discuss further selec-tion of analytes for the vadose and saturated zones.
9
The frequency of sample collection is important in thedesign of an optimum ground-water quality monitoring pro-gram (Cartwright and Shafer, 1987). Sampling frequencyaffects the cost of the monitoring program and the appropri-ateness of any inference(s) made from the resulting data.Sample collection and analysis should not occur so often as toresult in redundant information that would increase costs withno marginal gain in useful information. Conversely, samplecollection should not be so infrequent as to detract from theability to accurately forecast trends in ground-water quality.Ground-water sampling frequency should be based on theobjectives of the monitoring program and the hydrodynamicsof the ground-water system being monitored. As discussed,since ground-water movement is relatively slow, there is littleneed to sample every few meters of the flow path. Samplingfrequency is discussed further in Section 9.1.4.
During the past decade, the use of geostatistical prin-ciples (i.e., structural analysis, kriging, and conditional simu-lation) to interpret ground-water data has increased.Geostatistical techniques are used to evaluate the spatial vari-ability of ground-water flow parameters, particularly hydrau-lic head and transmissivity. However, less work has beenconducted on the application of geostatistics to interprethydrochemical data and ground-water quality monitoring net-work design. Samper and Neuman (1985), who performed ageostatistical analysis of selected chemical variables, showedthat geostatistical approaches may be valid to evaluate ground-water chemical data, particularly on a regional scale (Cartwnghtand Shafer, 1987).
The principles of geostatistics may be appropriate forinterpolation of point data to estimate the spatial distributionof certain aspects of ground-water quality (Englund and Sparks,1988). Kriging measures the error of estimation, which can bemapped and used to select locations for additional samplingpoints. These error maps show where the interpolated valuesdeviated from the expected statistical structure, thus indicat-ing the best locations to place additional wells (Virdee andKottegoda, 1984). However, this information can only serveas a guide because of other constraints on well location suchas environmental concerns, political issues, and economiclimitations (see Table 2-2). Nevertheless, a near-optimal moni-toring network can be developed for a predetermined level ofreliability.
The use of geostatistics to design monitoring networksand interpolate data has limitations. Using kriging for ground-water investigations often may have a limited effectivenessbecause of lack of sufficient data to perform the structuralanalysis. Hughes and Lettenmaier (1981) suggest that a mini-mum sample size of 50 is required before kriging is superiorto more traditional interpolation schemes (e.g., the least squaresmethod). Even with sufficient data and suitable statisticalsupport, structural analysis is highly subjective. Further, thetheoretical basis for the application of geostatistics is theconcept of a regionalized variable, which is defined as aspatially correlated random variable. To date, there have beenno definitive studies of the validity of assuming that hydro-chemical properties of ground water behave as regionalizcdphenomena (Cartwright and Shafer, 1987). For a further dis-cussion of geostatistical methods, see Section 7.3.2.
2.6 ReferencesAller, L., T.W. Bennett G. Hackett, Rebecca J. Petty, J.H.
Lehr, H. Sedoris, D.M. Nielsen. 1989. Handbook ofSuggested Practices for the Design and Installation ofGround-Water Monitoring Wells. EPA/600/4-89/034(NTIS PB90-159807). Also published in NWWA/EPAseries, National Water Well Association, Dublin, OH.
Barcelona, M.J., J.P. Gibb, and R.A. Miller. 1983. A Guide tothe Selection of Materials for Monitoring Well Construc-tion and Ground Water Sampling. ISWS Contract Report327. Illinois State Water Survey, Champaign, IL.
Barcelona, MJ., J.A. Helfrich, E.E. Garske, and J.P. Gibb.1984. A Laboratory Evaluation of Ground Water Sam-pling Mechanisms. Ground Water Monitoring Review4(2):32-41.
Barcelona, MJ., J.P. Gibb, J.A. Helfrich, and E.E. Garske.1985. Practical Guide for Ground-Water Sampling. EPA600/2-85/104 (NTIS PB86-137304). Also published asISWS Contract Report 374, Illinois State Water Survey,Champaign, IL.
Bouwer, E., J.W. Mercer, M. Kavanaugh, and F. DiGiano,1988. Coping with Groundwater Contamination. J. WaterPollution Control Federation 60(8): 1414-1428.
Campbell, M.D. and J.H. Lehr. 1973. Water Well Technol-ogy. McGraw-Hill, New York, NY.
Cartwright, K. and J.M. Shafer. 1987. Selected TechnicalConsiderations for Data Collection and Intepretation-Ground Water. In: National Water Quality Monitoringand Assessment, National Academy Press, Washington,DC, pp. 33-56.
de Vera, E.R. 1980. Samplers and Sampling Procedures forHazardous Waste Streams. EPA-600/2-80-018 (NTISPB80-135353).
Driscoll, F.G. 1986. Groundwater and Wells, 2nd ed. JohnsonDivision, UOP, Inc., St. Paul, MN.
Englund, E. and A. Sparks. 1988 GEO-EAS (GeostatisticalEnvironmental Assessment Software) User’s Guide. EPA/600/4-88/033a (Guide NTIS PB89-151252; Software:NTIS PB89-151245).
Everett, L.G. 1980. Ground Water Monitoring. TechnologyMarketing Operation, General Electric Co., Schenectady,NY.
Everett, L.G., L.G. Wilson, and E.W. Hoylman. 1984. VadoseZone Monitoring for Hazardous Waste Sites. Noyes DataCorp., Park Ridge, NJ.
Ford, P.J., P.J. Turina, and D.E. Seely. 1984. Characterizationof Hazardous Waste Sites - A Methods Manual, II, Avail-able Sampling Methods, 2nd ed. EPA 600/4-84-076 (NTISPB85-521596). [The first edition was published in 1983as EPA/600/4-83-040 (NTIS PB84-126920)].
20
Gibb, J.P., R.M. Schuller, and R.A. Griffin. 1981. Proceduresfor the Collection of Representative Water Quality Datafrom Monitoring Wells. Cooperative Groundwater Re-port 7. Illinois State Water Survey and Illinois StateGeological Survey, Champaign, IL.
Gillham, R.W., M.J.L. Robin, J.F. Barker and J.A. Cherry.1983. Groundwater Monitoring and Sample Bias. APIPublication 4367. American Petroleum Institute, Wash-ington, DC.
Guswa, J.H., WJ. Lyman, A.S. Donigan, Jr., T.Y.R. Lo, andE.W. Shanahan. 1984. Groundwater Contamination andEmergency Response Guide. Noyes Publication, ParkRidge, NJ.
Hughes, J.P. and D.P. Lettenmaier. 1981. Data Requirementsfor Kriging: Estimation and Network Design. Water Re-sources Research 17(6): 1641-1650.
Keely, J.F., M.D. Piwoni, and J.T. Wilson. 1986. EvolvingConcepts of Subsurface Contaminant Transport. J. WaterPollution Control Federation 58(5):349-357.
Keith, S.J., M.T. Frank, G. McCarty, and G. Mossman. 1983.Dealing with the Problem of Obtaining Accurate Ground-water Quality Analytical Results. In: Proc. Third Nat.Symp. on Aquifer Restoration and Ground Water Moni-toring, National Water Well Association, Dublin, OH,pp. 272-283.
Kruseman, G.P. and N.A. de Ridder. 1976. Analysis andEvaluation of Pumping Test Data. International Institutefor Land Reclamation and Improvement, Bulletin 11.Wageningen, The Netherlands.
Larson, D. 1981. Materials Selection for Ground Water Moni-toring. Presented at the National Water Well AssociationShort Course entitled practical Considerations in the De-sign and Installation of Monitoring Wells, Columbus,OH, December 16-17.
Lawrence Berkeley Laboratory. 1977. Invitational Well-Test-ing Symposium Proceedings. LBL-7027. Lawrence Ber-keley Laboratory, Berkeley, CA.
Lawrence Berkeley Laboratory. 1978. Second InvitationalWell-Testing Symposium Proceedings. LBL-8883.Lawrence Berkeley Laboratory, Berkeley, CA.
Lehr, J. H. 1988. An Irreverent View of Contaminant Disper-sion. Ground Water Monitoring Review 8(4):4-6.
Mercer, J.W., S.D. Thomas, and B. Ross. 1982. Parametersand Variables Appearing in Repository Siting Models.NUREG/CR-3066. U.S. Nuclear Regulatory Commis-sion, Washington, DC.
Mercer, M.W. and C.O. Morgan. 1981. Storage and Retrievalof Ground-Water Data at the U.S. Geological Survey.Ground Water 19(5):543-551.
Moore, J.E. 1979. Contribution of Ground-Water Modeling toPlanning. J. Hydrology 43:121-128.
Morrison, R.D. 1983. Ground Water Monitoring Technology,Procedures, Equipment and Applications. TIMCO Manu-facturing, Inc., Prairie du Sac, WI.
Nacht, SJ. 1983. Monitoring Sampling Protocol Consider-ations. Ground Water Monitoring Review 3(3):23-29.
National Research Council (NRC). 1990. Ground Water Mod-els: Scientific and Regulatory Applications. NationalAcademy Press, Washington, DC.
Olea, R.A. 1982. Optimization of the High Plains AquiferObservation Network, Kansas. Groundwater Series 7.Kansas Geological Survey, Lawrence, KS.
Olea, R.A. 1984. Systematic Sampling of Spatial Functions.Series on Spatial Analysis No. 7. Kansas GeologicalSurvey, Lawrence, KS.
Samper, F.J. and S.P. Neuman. 1985. Gcostatistical Analysisof Hydrochemical Data from the Madrid Basin, Spain(Abstract). Eos (Trans. Am. Geophysical Union)66(46):905.
Scalf, M.R., J.F. McNabb, W.J. Dunlap, R.L. Cosby, and J.Fryberger. 1981. Manual of Ground-Water Quality Sam-pling Procedures. EPA/600/2-81/160, (NTIS PB82-103045). Also published in NWWA/EPA Series, NationalWater Well Association, Dublin OH.
Sisk, S.W. 1981. NEIC Manual for Groundwater/SubsurfaceInvestigations at Hazardous Waste Sites. EPA/330/9-81-002 (NTIS PB82-103755).
USATHAMA. 1982. Sampling and Chemical Analysis Qual-ity Assurance Program for U.S. Army Toxic and Hazard-ous Materials Agency, Aberdeen Proving Ground, MD.
U.S. Environmental Protection Agency (EPA). 1986. RCRAGround Water Monitoring Technical Enforcement Guid-ance Document. EPA OSWER-9950.1. Also published inNWWA/EPA Series, National Water Well Association,Dublin, OH.
van Geer, F.C. 1987. Applications of Kalman Filtering in theAnalysis and Design of Groundwater Monitoring Net-works. TNO Institute of Applied Geoscience, Delft, TheNetherlands.
Virdee, T.S. and N.T. Kottegoda. 1984. A Brief Review ofKriging and its Application to Optimal Interpolation andObservation Well Selection. Journal des SciencesHydrologiques 29(4):367-387.
Wheatcraft, S.W. 1989. An Altemate View of ContaminantDispersion. Ground Water Monitoring Review 9(3): 11-12.
21
Chapter 3Geologic Aspects of Site Remediation
James W. Mercer and Charles P. Spalding
This chapter addresses the geologic aspects of remedia-tion: (1) What geologic factors are significant? (2) How aregeologic data collected? and (3) How are geologic data inter-preted? To help answer these questions, this chapter includesinformation on stratigraphy, lithology, structural geology, andhydrogeology. However, this chapter does not include infor-mation on basic geology, but the reader may consult any ofnumerous textbooks on the subject. There is also a concisereview of basic geology in U.S. EPA (1987).
To support discussions of the geologic factors, means ofcollecting geologic data are also included. See Chapter 4 forspecific, detailed information on wells. This chapter coverssoil and rock coring, as well as various surface and boreholegeophysical techniques. A case history on the Hyde Parklandfill concludes the chapter.
3.1 StratigraphyStratigraphy is the study of the formation, composition,
sequence, and correlation of stratified rocks and unconsoli-dated materials (e.g., clays, sands, silts, and gravels). Strati-graphic data include formational designations, age, thickness,areal extent, composition, sequence, and correlations. In astratigraphic investigation, aquifers and confining formationsare identified so that units likely to transport pollutants can bedelineated, and lateral changes in formations (facies changes)are noted if present. In effect, the stratigraphy of a site definesthe geometry and framework of the ground-water flow sys-tem. Therefore, knowledge of the stratigraphy is necessary inorder to identify pathways of chemical migration, to estimateextent of migration, and to define the hydrogeologic framew-o r k .
The first step in conceptualizing a site is to study driller’slogs, well cuttings, and/or corings. While observations madeduring drilling activities can provide additional informationsuch as drilling rates and water losses, the primary goal ofthese observations is to characterize layers of like material.This layering can be differentiated based on material type, buta major consideration for characterization should be how wellthe material transmits water. The primary differentiation shouldbe based on whether the material has properties of an aquiferand readily transmits water or has properties of a confiningbed, prohibiting the movement of water.
Once the layering has been determined at each well, thenext task is to plot the wells at their relative locations to eachother and attempt to correlate the layers among the wells. Thiscorrelation involves interpreting well-log data and requiresknowledge of geological processes. At some sites, the correla-tion will be straightforward; at others, correlation may beimpossible, The ability to correlate also will depend on thescale of the correlation. To understand the stratigraphic con-trols of flow and chemical migration, only larger scale fea-tures may need to be correlated. The completed correlationresults in a figure called a fence diagram (see Figure 3-1). Asshown in the figure, a fence diagram is composed of intersect-ing geological cross-sections.
The elevations of where the layers connect can be con-toured to form structural maps representing either the top orbottom surface of various layers. Where dense immisciblefluids are a concern, structural maps on top of confining layersare valuable because such fluids will flow via gravity on topof the confining layer toward the lower elevations. Structuralmaps for adjacent units can be subtracted from each other toyield thickness or isopach maps. An isopach map may beused, for example, to show the overburden thickness of un-consolidated material overlying bedrock. Once completed,these maps, along with the fence diagram, will provide athree-dimensional picture of the subsurface system throughwhich the ground water and chemicals are moving.
In addition to wells and well cuttings, other means toobtain stratigraphic data include hand augers, split-spoonsamplers, shelby tubes, and rock-coring equipment. Handaugers are useful, particularly in sandy materials, for examin-ing soil profiles to shallow depths (a few meters) and forinstalling monitoring devices. Many types of hand augers areavailable, but all are limited to use in unconsolidated geologicmaterials and tend to be impractical in dense clays or stonymaterials (Gillham, 1988).
A split-spoon sampler consists of a metal cylinder that issplit longitudinally and threaded on both ends. A cutting headis threaded onto the lower end and a drill-rod attachmentthreaded onto the upper end. The sampler is driven into theformation at the bottom of an augered borehole, using adrilling rig with a 140-pound weight (ASTM, 1990a). Thenumber of blows required to penetrate a soil is a function ofthe compactability of the soil; thus, blow count can be used tocharacterize soil types. When withdrawn and opened,. the
23
Figure 3-1. Sample fence diagram construction (from Compton, 1962).
sample is relatively undisturbed and shows the natural stratifi-cation of the geologic material. Shelby tubes are thin-walledmetal tubes that are attached to drill rods and are driven intothe formation (Gillham, 1988). Samples can be sealed andstored in the tubes and later extruded for examination. How-ever, both shelby tubes and split-spoon samplers are limited tosampling of unconsolidated materials.
When formations are too hard to be sampled by soilsampling methods, core drilling can be used (ASTM, 1990b).The simplest core barrel consists of a hollow steel tube with acore catcher and a diamond or tungsten carbide core bit. Othercore barrels have a dual wall system with a floating innersleeve that remains stationary while the outer barrel rotatesand cuts the core. A wireline system is available that elimi-nates pulling the drill pipe from the hole to recover each core(Landau, 1987). In this system, the core material is retrievedthrough the annulus of the drill rods.
Analysis of cores is performed both in the field and in thelaboratory. Laboratory analysis includes determination of po-rosity; permeability; and saturation with respect to a specificfluid component e.g., nonaqueous phase liquids (NAPL); andlithology studies (Keelan, 1987). Field studies of cores in-clude determination of rock quality designation (RQD), corerecovery rate, fracture nature and frequency, presence ofchemical odors, and general core lithology. RQD representsthe amount of core greater than 4 in. in length divided by thelength of core run attempted. This parameter is related to thecompetence of the material core and the fracture density of thecore run, Because RQD often can be correlated to permeabili-ty, it is useful in characterization studies. Often cores arebroken during transport so all fracture-related analyses shouldbe performed as soon as possible after the core has beenretrieved,
Coring also provides opportunities to monitor drillingreturn fluids for both color changes related to lithology andvisual and olfactory evidence of contaminants. As coringproceeds, net drilling fluid loss or gain to the cored formationcan be determined by maintaining an accurate balance ofdrilling fluids used. Fluid losses to an interval may be theresult of fractures or solutioning within the rock matrix. Asrock of varying competence is encountered, drilling rate alsovaries and for a given drilling system, drilling rate can becharacteristic of the material penetrated.
Because the conceptualization of site conditions is basedon roughly correlated parameters of subsurface and unseenconditions, it is useful to construct a correlation chart ofselected parameters versus depth (see Figure 3-2). Additionalparameters that may have been included in this figure arepermeability, drilling rate per foot, water loss or gain, andpresence and type of contamination.
3.2 LithologyLithology is the study of the physical character and
composition of unconsolidated deposits or rocks. As dis-cussed in the Handbook, it includes (1) mineralogy, (2) or-ganic carbon content, (3) grain size, (4) grain shape, and (5)packing. The first two items affect sorption, whereas the lastthree items affect water storage and flow. Additiomlly, com-paction and cementation will reduce permeability based onprimary porosity, whereas solution channels will increasepermeability (Levorsen, 1967).
The mineral composition of rocks and unconsolidateddeposits can be used to determine the chemical composition.The chemical composition of the media affects chemicaltransport in ground water via a variety of chemical reactions.Such interactions primarily involve inorganics and include
24
9
Figure 3-2. Correlation chart of hydrogeologic features (from GeoTrans, 1989).
25
sorption, precipitation and dissolution, acid-base reactions,complexation, and redox reactions. Examples of chemicalsthat could be reduced to lower concentrations in ground waterthrough the formation of precipitates include arsenic (byreaction with iron, aluminum, or calcium), lead (by reactionwith sulfide or carbonate), and silver (by reaction with sulfideor chloride). Hydrolysis can lead to the precipitation of iron,manganese, copper, chromium, and zinc contaminants. Oxi-dation or reduction could favor the precipitation of chromium,arsenic, and selenium.
The tendency of an organic chemical to sorb is directlyrelated to the fraction of total organic carbon content in termsof grams of organic carbon per gram of soil. A typical value oforganic matter in mineral soils is 3.25 percent (Brady, 1974).The amount of organic matter is approximately 1.9 times theamount of organic carbon; therefore, a typical value for or-ganic carbon content is 1.7 percent. However, data will varyfrom site to site.
Although variation in sorption between different grain-size fractions is mostly a reflection of their organic carboncontent, other factors such as surface area have an effect. Ingeneral, the fine silt and clay fractions of soils have thegreatest tendency to sorb chemicals. Grain size also influ-ences water storage and movement. The amount of soil ineach of various size groups is one of the major factors used inanalyzing and classifying a soil. Various agencies define soil
groups in slightly different ways (see Figure 3-3). In general,coarser grained soil is more transmissive and has less storagecapacity than finer grained soil.
Grain shape also influences water storage and porositybecause grain shape affects the manner in which grains arearranged. Highly angular and irregularly shaped, noncementedgrains tend to result in a greater porosity than smooth, regu-larly shaped grains, although the difference may be slight
Grain-size analysis, conducted on samples from uncon-solidated formations, yields the proportion of material in eachspecified size range. Range distributions can be used to esti-mate permeabilities, design monitoring wells, and enablebetter stratigraphic interpretation. The results of a grain-sizeanalysis usually are plotted as shown in Figure 3-4. The sieve-opening size retaining 90 percent of the soil is called theeffective particle size (D90%), whereas the sieve-opening sizeretaining 50 percent is called the average particle size (D50%).Uniform soils consist of grains of predominantly one sizeyielding curves with steep slopes. Well-graded soils havegrains of many different sizes and, therefore, are characterizedby more gently sloping curves.
Soils composed of grains of nearly uniform particle sizehave a larger porosity than a well-graded soil because, in thewell-graded soil, small particles occupy a portion of thevolume between the larger particles. In the vadose zone,
* Colloids included in clay fraction in test reports.** The LL and PI of “Silt” plot below the “A” line of the plasticity chart, Table 4,
and the LL and PI for “Clay” plot above the “A” line.
Figure 3-3. Soil-group size limits of ASTM, AASHO, USDA, FAA, Corps of Engineers, and USBR (from Portland Cement Associa-tion, 1973).
26
Figure 3-4. Particle-size distribution for a uniform sand and awell-graded soil (from Bouwer, 1978).
uniform soils develop a well-defined capillary fringe, whereaswell-graded soils tend to have a higher, but less distinctcapillary fringe.
In summary, much qualitative information concerningproperties that affect flow and transport can be gained fromlithology. At many hazardous waste sites, this type of infor-mation may be all that is available in the early stages of fieldstudy. Thus, it may be used to help guide subsequent phases ofthe field work, such as well screen design. This type ofqualitative information may be very helpful in characterizingvadose-zone properties, where hydrologic testing is moredifficult to conduct and interpret.
3.3 Structural GeologyStructural geology includes studying and mapping fea-
tures produced by movement after deposition. Structural fea-tures include folds, faults, joints/fractures, and interconnectedvoids (i.e., caves and lava tubes). Highly vesicular tops andbottoms of basalt flows, for example, are often cited assources of significant permeability. Just as important to thedefinition of structural features is the more rapid cooling andmore intense fracturing of the top and bottom of flows (Huntley,1987). Deformed, inclined, or broken rock formations cancontrol topography, surface drainage, and ground-water re-charge and flow. Joints and fractures are commonly majoravenues of water transport (preferential pathways) and usu-ally occur in parallel sets.
Most fractures can be attributed to one of three causes(Lcworsen, 1967). Some fractures format depth as a result ofan increase in rock volume from the folding and bending ofstrata. Others are caused by the removal of overburden byerosion in the zone of weathering. As sediments are unloadedthrough erosion, the upper parts expand, and incipient weak-
27
nesses in the rocks become joints, fractures, and fissures.Therefore, an increase of fracturing below an unconformity isto be expected. Probably much of the initial solution channel-ing through which surface waters percolate results from thegradual increase in jointing and fracturing that accompaniesweathering. The third cause of fracturing is a reduction in thevolume of shales in the ground, due to diagenetic mineralchanges coupled with a loss of water during compaction.
Solution features, such as enlarged joints, sinkholes andcaves are common in limestone rocks and promote rapidground-water movement. Pertinent data on structural featuresnecessary to study and understand solution features includetype, compass orientation, dip direction and angle, and stratig-raphy. Chapter 6 discusses the influence of fractured media onground-water flow and how it is characterized.
3.4 HydrogeologyHydrogeology concerns the relationship of the movement
of subsurface waters with geology, and ties stratigraphy,!ithology, and structural geology to the theory of ground-water hydraulics. The main goal in studying hydrogeology isto determine directions and rates of ground-water flow. Thisinformation is essential to any ground-water remediation orground-water monitoring program. Although this topic isintroduced in this section, it is discussed in more detail inChapter 4.
Hydrologic factors that are important to hydrogedogyinclude surface drainage and surface water/ground-water rela-tionships. Surface drainage information includes tributaryrelationships, stream widths, depths, channel elevations, andflow data. In a hydrogeologic investigation, the nearest per-manent gaging station and period of record should be deter-mined. A U.S. Geologic Survey (USGS) 7 l/2-minutetopographic map will show some of the necessary informat-ion. Gaging stations and flow data can be identified andobtained through USGS data bases. Streams either can receiveground-water inflow or lose water by channel exfiltration. Aspart of the investigation, hydrologic literature should be re-viewed to determine if local streams are “gaining” or “losing.”Losing streams are common in areas of limestone bedrock andthose with arid climates and coarse-grained channel substrates. Potential ground-water recharge areas, sometimes in-dicated by flat areas or depressions noted on the landscape,also should be identified. Stereo-pair aerial photographs canalso be useful in these determinations (Ray, 1960). Irrigatedfields detected in aerial photographs suggest ground-waterrecharge areas; swampy, wet areas suggest areas of ground-water discharge.
Other important factors include aquifer delineation, back-ground water quality, and depth to ground water. As used inthis Handbook, depth to ground water refers to the verticaldistance from the ground surface to the standing water level ina well. In a confined aquifer, the depth to water represents apoint on a “piezometric” surface. The depths will limit theequipment that can be used for purging and sampling. Infor-mation should be collected to delineate aquifer type(unconfined, confined, or perched); composition; boundaries;hydraulic properties (permeability, porosity, transmissivity,
etc.); and interconnection with other aquifers (direction ofleakage). These data are generally available through geologi-cal survey publications.
Probable ground-water flow directions (both horizontaland vertical) are determined by comparing the elevation ofwater levels in different wells. The quality of ground andsurface water in an area should define, to a large extent,potential uses. Knowledge of natural or background waterquality and water uses is required to assess contaminantimpacts. The quality of surface waters is usually availablefrom U.S. EPA, USGS, and state records. Ground-water datawill probably be limited for any given area, but may bediscussed through USGS Water Resources Division offices,state geological surveys, and county health departments.
3.5 Hydrogeologic InvestigationsMuch of the data needed to understand site-specific
ground-water movement will be determined via hydrogeo-logic investigations. The purpose of these investigations is todetermine flow directions, pathways and rates of ground-water flow, potential receptors of ground water, potentialcontaminants, and the extent of contamination in the subsur-face. This information is required for selecting from altern-ative remedial strategies, and it provides the framework fordesign of a ground-water remedial program, if needed.
Some of the field methods used to obtain this informationinclude borehole exploration (including coring), mapping sur-face features, and geophysical methods (both surface anddownhole techniques). Much of the information gained fromthese methods will be helpful in interpreting the geology. Forground-water flow information, additional field methods in-clude (1) monitoring water elevations in wells and adjacentsurface waters, (2) performing aquifer tests (pumping and/orslug tests), and (3) using special methods such as laboratoryanalysis of cores and borehole flowmeters. For subsurfacechemistry, soil sample analysis must be performed, as well assampling and analysis of ground water. A typical monitoringwell for ground-water sampling is shown in Figure 3-5. Ifnonaqueous phase liquid (NAPL) is present, any free productthickness must be measured and sampling performed.
3.5.1 Geophysical TechniquesGeophysical techniques are used to better understand
subsurface conditions arid to delineate the extent of contami-nation. Common surface techniques used at hazardous wastesites include surface resistivity, electromagnetic surveys, seis-mic reflection method, ground-penetrating radar, and magne-tometer surveys (see Table 3-1).
In surface resistivity methods (Zohdy et al., 1974; Stewartet al., 1983), the geologic materials act as part of a directcurrent circuit. In general, there are two current electrodes andtwo electrodes for measuring voltage differences. The electri-cal potential measured between the electrodes depends on theelectrical properties of the geologic materials which, in turn,depend upon the resistivity of the pore water and the amountof pore water. Most soil and rock materials are highly resis-tive, while water is highly conductive. Porosity and local
28
stratigraphy, therefore, can be deduced from the measure-ments. Because of the concentrations of some solutes, con-taminant plumes frequently appear as a highly conductivelayer. Resistivity methods, therefore, can be useful in identi-fying and mapping certain plumes (Wish, 1983).
Electromagnetic instruments used in hydrogeologic in-vestigations consist of a transmitter and receiver (Stewart,1982). The transmitter produces an alternating magnetic fieldthat induces electrical currents within the ground. The in-duced currents vary with the electrical conductivity of thegeological materials and alter the magnetic field of the trans-mitter. This alteration is detected by a receiver. Generally,these devices are carried by one person, and do not require theinstallation of electrodes or geophones. They are likely to bemore cost-effective than resistivity methods because fieldwork can be completed more rapidly. They can be used todetect changes in subsurface conductivity related to contami-nant plumes or buried metallic waste such as drums (Green-house and Slaine, 1983).
In surjace seismic methods (Sverdrup, 1986), an impactis made at a particular point on the ground surface using amechanical hammer or an explosive device. The resultingsound waves are monitored by sensing devices (geophones)positioned at various distances from the impact. The time ofarrival of the sonic waves depends on velocity and densitycontrasts that occur as the wave passes through differentstratigraphic layers. By interpreting the sigml, the investiga-tion determines the geologic layering in the area.
In ground-penetrating radar (Koemer et al., 1981), radiowaves are transmitted into the ground and the reflected wavesare monitored and analyzed. Reflections occur as a result ofgeologic variations in porosity and water content. The methodis useful for determining stratigraphic variations and for locat-ing buried objects such as steel drums.
Magnetometer surveys (Gilkeson et al., 1986) measurethe strength of the earth’s magnetic field. A proton nuclearmagnetic resonance magnetometer is frequently used. Oneperson can rapidly perform a survey over a site of a few acresby using this hand-held instrument. The surveyor sets up agrid system and measures the magnetic field at each intersec-tion of the grid. Areas with large amounts of buried metal,such as steel drums, will have magnetic anomalies associatedwith them. The strength of the anomaly will vary with theamount and depth of the buried metal.
Borehole logging (Keys and MacCary, 1971; Keys, 1988)includes a variety of methods involving lowering a tool intothe borehole (see Table 3-2). The tool measures the physicalproperties of the geologic materials, or, alternatively, providesan impulse or disturbance to the natural system, and measuresthe response of the system to the disturbance. Common log-ging tools include caliper, resistivity, neutron, gamma, andsonic tools. Logging can proceed in both cased or uncasedboreholes, though most measurements can be made onlywhen the hole has not been cased. Most of the loggingmethods are effective in distinguishing between sand and clayand are, therefore, useful in locating zones of high permeabil-ity (Kwader, 1986),
Figure 3-5. A typical monitoring well design (from GeoTrans, 1989).
Induction logging can be used to identify soil and rocktypes, geologic correlations, soil and rock porosity, and porefluid conductivity. Resistivity logging is effective in identify-ing soil and rock types, geologic correlations, soil and rockporosity, pore fluid resistivity, and secondary permeabilitysuch as the locations of fractures and solution openings.Natural gamma logging can assist in positioning wells andcasings, by providing information on clay or shale content,grain size, pore fluid resistivity, and soil and rock identifica-tion. Gamma-gamma logging will help to position cementingfor the well casings and to determine total porosity or bulkdensity. Neutron logs can provide estimates of moisture con-tent above the water table, total porosity below the water
table, specific yield of confined aquifers, the location of thewater table outside the casing, chemical and physical proper-ties of the water, and the rate of moisture infiltration. Tem-perature logs help provide the chemical and physicalcharacteristics of the water source and movement of thewater in the well; and dilution, dispersion, and movement ofthe waste.
Video cameras also have been developed that can belowered down a 4-in. (l0-cm) diameter borehole. They can beused for visual inspection and to provide a visual record of thewall of the borehole. They are particularly useful for inspect-ing the casing for corrosion, damage, or leaks, and also are
29
Table 3-1. Summary of Surface Geophysical Methods
Surface GeophysicalSurvey Method Applications Advantages Limitations
REFRACTIONAND REFLECTION
Determines - Ground-water resourceIithological evaluations
changes in - Geotechnical profilingsubsurface - Subsurface stratigraphic
profiling including top ofbedrock
ELECTRICALRESISTIVITY
Delineates - Depth to water table estimatessubsurface - Subsurface stratigraphic profilingresistivity - Ground-water resource evaluationscontrasts due to - High ionic strength contaminatedIithology, ground ground-water studieswater, andchanges in ground-water qualify
ELECTROMAGNETICCONDUCTIVITY
Delineates - Subsurface stratigraphic profilingsubsurface - Ground-water contamination studiesconductivity - Landfill studiescontrasts due to - Ground-water resourcechanges in evaluationsground-water - Locating buried utilities,quality and tanks, and drumsIithology
GROUNDPENETRATINGRADAR
Provides contin- - Locating buried objectsuous visual profile - Delineation of bedrockof shallow sub- subsurface and structuresurface objects, - Delineation of karst featuresstructure, and - Delineation of physical integrity ofIithology manmade earthen structures
MAGNETICS
Detects presence - Location of buried ferrousof buried metallic objectsobjects - Detection of boundaries
of landfills containingferrous objects
- Location of iron-bearing rock
- Relatively easy accessibility- High depth of penetration
dependent on source ofvibration
- Rapid areal coverage
- Rapid areal coverage- High depth of penetration
possible (400-800 ft)- High mobility- Results can be approximated
in the field
- High mobility- Rapid resolution and data
interpretation- High accessibility- Effectiveness in analysis of very
high resistivity- Equipment readily accessible
- Great areal coverage- High vertical resolution in
suitable terrain- Visual picture of data
- High mobility- Data resolution possible in field- Rapid areal coverage
- Resolution can be obscured inlayered sequences
- Susceptibility to noise from urbandevelopment
- Difficult penetration in cold weather(depending on instrumentation)
- Operation restricted during wetweather
- Susceptibility to natural andartificial electrical interference
- Limited use in wet weather- Limited utility in urban areas- Interpretation that assumes a
layered subsurface- Lateral heterogeneity
not easily accounted for
- Data reduction less refined than withresistivity
- Use unsuitable in areas with surfaceor subsurface power sources,pipelines, utilities
- Less vertical resolution than withother methods
- Limited use in wet weather
- Limited depth of penetration(a meter or less in wet, clayeysoils; up to 25 meters in dry, sandysoils)
- Accessibility limited due to bulkinessof equipment and nature of survey
- Interpretation of data qualitative- Limited use in wet weather
- Detection dependent on size andferrous content of buried object
- Difficult data resolution in urban areas- Limited use in wet weather- Data interpretation complicated in
areas of natural magnetic drift
Source: Modified after O’Brien and Gere (1988)
30
Table 3-2. Summary of Borehole Log Applications
Required information on the Properties of Widely Available Logging Techniques That Might Be UtilizedRocks, Fluid Wells, or the Ground-Water System
Lithology and stratigraphic correlation of aquifers andassociated rocks
Total porosity or bulk density or gamma-gamma
Effective porosity or true resistivity
Clay or shale content
Permeability
fractures, solution openingsSecondary permeability -
Specific yield of unconfined aquifers
Grain size
Location of water level or saturated zones
Moisture content
infiltration
Direction, velocity, and path of ground-water flow
Dispersion, dilution, and movement of waste
Source and movement of water in a well
Chemical and physical characteristics of water,including salinity, temperature, density, and viscosity
Determining construction of existing wells, diameter andposition of casing, perforations, screens
Guide to screen setting
Cementing
Casing corrosion
Casing leaks and/or plugged screen
Source: Keys and MacCary (1971)
Electric, sonic, or caliper logs made in open holes. Nuclear logs made inopen or cased holes.
Calibrated sonic logs in open holes, calibrated neutron logs in open orcased holes.
Calibratad long-normal resistivity logs.
Gamma logs.
No direct measurement by logging. Maybe related to porosity, injectivity,sonic amplitude.
Caliper, sonic, or borehole televiewer or television logs.
Calibrated neutron logs.
Possible relation to formation factor derived from electric logs.
Electric, temperature, or fluid conductivity in open hole or inside casing. Neutronor gamma-gamma logs in open hole or outside casing.
Calibrated neutron logs
Time-interval neutron logs under special circumstances or radioactivetracers.
Single-well tracer techniques - point dilution and single-well pulse.Multiwell tracer techniques.
Fluid conductivity and temperature logs gamma logs for some radioactivewastes, fluid sampler.
infectivity profile. Flowmeter or tracer logging during pumping or injection.Temperature logs
Calibrated fluid conductivity and temperature in the well Neutron chloridelogging outside casing. Multielectrode resistivity.
Gamma-gamma, caliper, collar, and perforation locator, borehole television.
All logs providing data on the lithology, water-bearing characteristics, andcorrelation and thickness of aquifers.
Caliper, temperature, gamma-gamma. Acoustic for cement bond.
Under some conditions caliper or collar locator.
Tracer and flowmeter.
31
used in uncased rock holes for locating fractures and fracturezones (Gillharn, 1988).
3.5.2 Example-Hyde Park LandfillThis example, discussed in detail in Cohen et al. (1987),
concerns the Hyde Park landfill in Niagara Falls, New York(see Figure 3-6). Ground-water studies were initiated at thesite in 1978 when a shallow tile drain and clay cover wereinstalled at the landfill. Remedial investigations (RI), requiredby a settlement agreement, were conducted from 1982 to1984. A major component of the RI was a drilling programdesigned to determine the extent of chemical contamination inthe overburden and bedrock. Borings were cored and tested in15-ft sections to the top of the Rochester Shale along 10vectors radiating out from the landfill. Ground-water sampleswere taken for analysis from those 15-ft sections that yieldedsignificant amounts of water. If chemicals were present abovespecified levels, a new hole was drilled about 800 ft awayalong the vector. Some of these holes were used as observa-tion wells during aquifer tests prior to being grouted.
As a result of the drilling programs, the local geology isfairly well known. Approximately 15 to 30 ft of waste at the
landfill are underlain by O to 10 ft of silty clay sediments. AtHyde Park, the overburden lies unconformably on the LockportDolomite. Undulations in the bedrock surface were carved byprevious glaciation. The Lockport Dolomite ranges in thick-ness from 130 ft (200 ft southeast of the landfdl) to 65 ft at theNiagara Gorge. The Lockport Dolomite overlies the Roches-ter Shale and several lower units in a layer-cake sequence.
The hydrogeology of the Hyde Park area is unique be-cause of the Niagara River Gorge and the human-inducedchannels associated with a nearby pump storage reservoir (seeFigure 3-6). The Niagara Gorge (about 2,000 ft to the west),the forebay canal (about 4,000 ft to the north), and the buriedconduits (about 3,000 ft to the east) control ground-watermovement in the Hyde Park area.
The ground-water system can be conceptualized as aseries of slightly dipping, permeable zones sandwiched be-tween aquitards, all of which are bounded on three sides bydrains. Precipitation infiltrates the wastes and the low-perme-ability overburden before recharging the highly fracturedupper layer of the Lockport Dolomite. Where glacial sedi-ments are present beneath the landfill, downward ground-water flow and chemical migration are retarded. In areas
Figure 3-6. A generalized diagram showing the geologic formation and topographic features in the vicinity of the Hyde Parklandfiii (from Faust, 1985).
32
where these sediments are thin and/or absent, ground waterand chemicals move freely into the underlying rock. In thepermeable bedrock zones, much of the ground water flowslaterally toward the three boundaries. Between these zones,ground water moves slowly downward to the next lowerpermeable layer. Pumping tests suggest an anisotropic systemwhere hydraulic conductivities are greatly affected by prefer-ential flow along fractures (Figure 3-7). This conceptualizationis supported by the alignment and dip of joint systems ex-pressed at nearby outcrops.
Analyses of ground-water samples taken during the vec-tor well survey revealed that contamination had migratedmuch further than previously thought. In fact, Hyde Parkchemicals were found in seeps emanating from the LockportDolomite along the Niagara Gorge in July 1984. Dissolvedchemical and NAPL plumes in the overburden and in theLockport Dolomite were delineated during the RI as shown inFigure 3-8. Although the areal extent of contamination hasbeen defined, the depth of chemical migration was unknownbecause at many locations dissolved chemicals and NAPLwere observed all the way to the base of the Lockport Dolo-mite.
The distribution of chemicals in the overburden reflectsthe downward migration of contaminated surface runoff fromthe Hyde Park landfill, which is elevated relative to surround-ing properties. Lateral chemical transport through theoverburden has been limited because the potential for down-ward flow to bedrock exceeds that for outward flow throughthe low-permeability glacial sediments.
The contamination observed in the Lockport Dolomitereflects variations in the directions of ground-water flow thathave occurred since waste disposal began at Hyde Park and, toa lesser extent, at the dipping beds of the Lockport Dolomite.Chemical analyses indicate the past migration of chemicalsthrough the upper Lockport Dolomite in all directions. Presentground-water flow is primarily to the northwest, but thesouthern and eastern areas of contamination suggest that atone time ground water moved toward those areas. Ground-water flow prior to the construction of the forebay canal andburied conduits (from 1958 to 1962) was inferred to be towardthe southwest. Similarly, dewatering during the constructionof these conduits could have drawn contaminated groundwater toward the east. Chemicals have moved downward tothe base of the Lockport Dolomite by dissolution in groundwater and by dense NAPL flow.
The Hyde Park Stipulation requires several remedial ac-tions, focusing on source control, overburden remedies, bed-rock remedies, and control of seeps at the Niagara Gorge face.The application of a series of numerical models of ground-water flow and chemical transport facilitated these remedies.
The source control program is designed to reduce theamount of chemicals migrating from the landfill into theoverburden and bedrock. This reduction will be achieved by asynthetic cap to reduce recharge and by extraction wells toremove chemicals. During the prototype phase of the pro-gram, two large-diameter extraction wells will be installed inthe landfill. Exploratory boreholes were completed in the
33
landfill to characterize the overburden stratigraphy of thelandfill and to help determine stratigraphic controls on NAPLmovement. All exploratory boreholes were then converted toNAPL monitoring wells. The success of the prototype extrac-tion wells depends in part on the compatibility of the sandpackwith landfill materials. To test the selection of the wellsandpack, two sandpack materials were selected based onknown landfill constituents. If a reasonable amount of NAPLcan be removed with this method, an operational network ofsix extraction wells will be installed.
The remedial program specified for the overburden isdesigned to laterally contain the dissolved chemicals andNAPL and to maximize collection of NAPL. Mobile NAPLnot removed from the overburden will tend to sink downwardto the bedrock and will be addressed by the bedrock remedy.The overall approach of the program is to further define theboundary of the overburden NAPL plume with a series ofborings and then install a tile drain to collect mobile NAPL.The location and depth of the drain will be determined afterthe overburden plume boundaries have been refined by aseries of 44 overburden borings around the landfill. As thedrain is installed, additional stratigraphic information will beadded as soil is removed. The performance criteria for theoverburden system are:
An inward hydraulic gradient must be maintained towardthe drain or downward into the bedrock.
There must be no expansion of the NAPL plume towardthe drain or downward.
Remedial systems planned for the Lockport Dolomite aredesigned to contain both the NAPL and APL plumes. Specificobjectives of the bedrock remedial system are to containdissolved chemicals and NAPL within the NAPL plume,contain dissolved chemicals in the area near the gorge facethat is designated the remediated APL plume, and eliminatethe seepage of chemicals at the gorge face. However, portionsof the APL plume will not be remediated. As with the sourcecontrol system, a prototype system will be implemented firstand later refined into an operational system. The system willuse extraction and injection wells to maximize the collectionof both dissolved chemicals and NAPL. The locations ofpurge, injection, and monitoring wells, and a schematic cross-section of the containment concept, are shown in Figures 3-9and 3-10, respectively. The recirculation wells are added tothe NAPL plume containment system to speed up the recov-ery of contaminants and to maintain higher water levels forthe flushing of chemicals in the upper bedrock.
All prototype bedrock extraction/injection wells and re-lated Lockport monitoring wells will be completed in threeseparate hydrogeologic zones. The separation of the Lockportinto three zones allows optimization of the remedial systemthrough better characterization, monitoring, and pumpingschemes for the selected zones.
The main performance criteria for the bedrock system isthe maintenance of an inward hydraulic gradient at the NAPLplume boundary. In addition, the flux of certain chemicals to
Figure 3-7. Postulated ground-water drawdown contours during Hyde Park landfill pump test (from Conestoga-Rovers & Associ-ates Limited, 1984).
34
Figure 3-8. Boundaries of dissolved chemical (APL) and NAPL plumes of contaminated ground water emanating from the HydePark landfill through the overburden and Lockport Dolomite (from Faust, 1985).
the Niagara River must be below specified limits. The interimflux level for 2,3,7,8-TCDD is 0.5 g/yr. This level will bemodified based on a future study of TCDD in the NiagaraRiver and Lake Ontario.
Hyde Park is an excellent example of a remediation thatboth allows for better site characterization and does not makeitself obsolete as more data become available. The remediesdescribed in the stipulation include extensive monitoring pro-
grams that both ensure that performance goals are achievedand enhance the understanding of site hydrogeology. Thephased approach with initial prototype remedies allows forbetter initial site characterization that will ultimately lead tothe optimal remediation approach. The program is not limitedto current technologies, but can be modified should newinnovations be found. This flexibility is important because ofthe long cleanup times expected.
35
Figure 3-9. Locations of purge, injection, and monitoring wells to be installed for the prototype Lockport Dolomite hydrauliccontainment system at the Hyde Park site (from Faust, 1985).
Figure 3-10. A conceptual cross-section of the Lockport Dolomite hydraulic containment system at the Hyde Park site(from Faust, 1985).
37
3.6 ReferencesAmerican Society for Testing and Materials (ASTM). 1990a.
Standard Method for Diamond Core Drilling for SiteInvestigation. In: Annual Book of ASTM Standards, Vol.04.08, D2113-83. ASTM, Philadelphia, PA.
American Society for Testing and Materials (ASTM). 1990b.Standard Method for Penetration Test and Split-BarrelSampling of Soils. In: Annual Book of ASTM Standards,Vol. 04.08, D 1586-84.
Bouwer, H. 1978. Ground-Water Hydrology. McGraw-Hill,New York.
Brady, N.C. 1974. The Nature and Properties of Soils, 8th ed.MacMillan, New York.
Cohen, R.M., R.R. Rabold, C.R. Faust, J.O. Rumbaugh III,and J.R. Bridge. 1987. Investigation and Hydraulic Con-tainment of Chemical Migration: Four Landfills in NiagaraFalls. Civil Engineering Practice 2(1):33-58.
Compton, R.R. 1962. Manual of Field Geology. John Wiley& Sons, New York.
Conestoga-Rovers & Associates Limited. 1984. RequisiteRemedial Technology Study, Overburden & Bedrock,Hyde Park Remedial Program. Prepared for OccidentalChemical Corporation, Ref. No. 1069.
Faust C.R. 1985. Affidavit re: Hyde Park. U.S., N.Y. v.Hooker Chemicals and Plastics Corp. et al., Civil ActionNo. 79-989.
GeoTrans. 1989. Progress Report - Hydrogeological Charac-terization of the Bedrock Near the S-Area Landfill (NiagaraFalls, NY) in Support of Requisite Remedial Technology(RRT) Evaluation. GeoTrans, Inc., Hemdon, VA.
Gilkeson, R.H., P.C. Heigold, and D.E. Laymen. 1986. Prac-tical Application of Theoretical Models to MagnetometerSurveys of Hazardous Waste Disposal Sites—A CaseHistory. Ground Water Monitoring Review 6(1):54-61.
Gillham, R.W. 1988. Glossary of Ground Water MonitoringTerms. Water Well Journal 42(5):67-71.
Greenhouse, J.P. and D.J. Slaine. 1983. The Use of Recon-naissance Electromagnetic Methods to Map ContaminantMigration. Ground Water Monitoring Review 3(2):47-59.
Huntley, D. 1987. Some Fundamentals of Hydrogeology, 5thed. In: Subsurface Geology, L.W. LeRoy, D.O. LeRoy,S.D. Schwochow, and J.W. Raese (eds.), Colorado Schoolof Mines, Golden, CO, pp. 746-755.
Keelan, D.K. 1987. Core Analysis. In: Subsurface Geology,5th d., L.W. LeRoy, D.O. LeRoy, S.D. Schwochow, andJ.W. Raese (eds.), Colorado School of Mines, Golden,co, pp. 35-47.
Keys, W.S. and L.M. MacCay. 1971. Application of Bore-hole Geophysics to Water-Resources Investigations. U.S.Geological Survey Techniques of Water-Resources In-vestigations TWI-2E1.
Keys, W.S. 1988. Borehole Geophysics Applied to Ground-Water Investigations. U.S. Geological Survey Open-FileReport 87-539,303 pp. [Published in 1989 with the sametitle by National Water Well Association, Dublin, OH.]
Koemer, R.M., A.E. Lord, Jr., and J.J. Bowders. 1981. Utili-zation and Assessment of a Pulsed RF System to MonitorSubsurface Liquids. In: National Conference on Manage-ment of Uncontrolled Hazardous Waste Sites, HazardousMaterials Control Research Institute, Silver Spring, LD,pp. 165-170.
Kwader, T. 1986. The Use of Geophysical Logs for Determin-ing Formation Water Quality. Ground Water 24:11-15.
Landau, H.L. 1987. Coring Techniques and Applications. In:Subsurface Geology, 5th cd., L.W. LeRoy, D.O. LeRoy,S.D. Schwochow, and J.W. Raese, (eds.), Colorado Schoolof Mines, Golden, CO, pp. 395-398.
Levorsen, A.I. 1967. The Reservoir Pore Space. In: Geologyof Petroleum, 2nd ed, W.H. Freeman and Company, SanFrancisco, CA, Chapter 4, pp. 115, 119, 120.
O’Brien and Gere Engineers, Inc. 1988. Hazardous WasteSite Remediation. Van Nostrand Reinhold, New York,422 pp.
Portland Cement Association. 1973. PCA Soil Primer. Engi-neering Bulletin EBO07.045, Portland Cement Associa-tion, Skokie, IL, 39 pp.
Ray, R.G. 1960. Aerial Photographs in Geologic Interpreta-tion and Mapping. U.S. Geological Survey ProfessionalPaper 373,230 pp.
Stewart, M., M. Layton, and T. Lizanec. 1983. Application ofSurface Resistivity Surveys to Regional HydrogeologicReconnaissance. Ground Water 21:42-48.
Stewart, N.T. 1982. Evaluation of Electromagnetic Methodsfor Rapid Mapping of Salt-Water Interfaces in CoastalAquifers. Ground Water 20:538-545.
Sverdrup, K.A. 1986. Shallow Seismic Refraction Survey ofNear-Surface Ground Water Flow. Ground Water Moni-toring Review 6(1):80-83.
Urish, D.W. 1983. The Practical Application of Surface Elec-trical Resistivity to Detection of Ground-Water Pollution.Ground Water 21:144-152.
U.S. Environmental Protection Agency (EPA). 1987. Hand-book Ground Water. EPA/625/6-87/016, 212 pp.
Zohdy, A.A.R, G.P. Eaton, and D.R. Mabey. 1974. Applica-tion of Surface Geophysics to Ground-Water Investiga-tions. U.S. Geological Survey Techniques ofWater-Resources Investigations TWI-2D1.
38
Chapter 4Characterization of Water Movement in the Saturated
James W. Mercer and Charles P. Spalding
Advection is the primary transport mechanism for con-servative chemicals and for many nonconservative chemicals.It is the controlling process for chemicals moving away froma source area (e.g., a landfill or a spill) and for removingchemicals from the subsurface (e.g., pump-and-treat systems).Therefore, understanding advection is important to both sitecharacterization and remediation.
An understanding of the factors that control ground-watermovement is needed to understand advection. This chapterreviews concepts needed to determine and understand ground-water flow. This review is followed by a discussion of fieldtechniques used to obtain the data needed to characterizeground-water flow. As important as it is to collect data, it isjust as important to analyze and interpret the data. Therefore,this chapter also discusses different analysis techniques andground-water remedial actions. Finally, an example ties thediscussion together and illustrates the important points of thechapter.
Both general data requirements and characterization tech-niques are presented throughout this chapter. Each applicationof these techniques is unique and site specific. No subsurfacecharacterization tool provides perfect information; severaltechniques (e.g., geophysical and geochemical) should becombined, such that different types of data support the sameconclusion. Because the field work is completed in phases,remediation decisions often involve some uncertainty; there-fore, the importance of monitoring is stressed.
4.1 Review of ConceptsThere are numerous books that characterize and present
the principles and concepts of ground-water hydrology (e.g.,Bear, 1979; Bouwer, 1978; Davis and DeWiest, 1966; DeWiest,1969; Domenico, 1972; Freeze and Cherry, 1979; Todd, 1980and Walton, 1970). Other general references have been pub-lished by the U.S. Environmental Protection Agency (e.g.,U.S. EPA, 1987). This section specifically discusses contamin-ant hydrology and will not cover many of the general topicsincluded in these references.
At hazardous waste sites, the following questions need tobe addressed with respect to ground-water hydrology: (1)where is the water coming from? (2) where is the water going?and (3) what are the rates of movement? Answering thesequestions requires information on the local water balance, the
Zone
transmissive properties of the media, and the hydraulic headdistribution.
Hydraulic head is rhe elevation to which water rises in awell that is open to the surface (Figure 4-l). It is composed oftwo parts: (1) the pressure head that produces the column ofwater above the open interval; and (2) the elevation head,which is the elevation of the open interval relative to a datum,usually mean sea level. Depth to water normally is measuredfrom a reference point (e.g., top of the casing) that has beensurveyed. This information is used to compute water-level
Figure 4-1. A diagram of the relationships between hydraulichead, H, pressure head, h, and gravitatlonal head,Z. The pressure head is measured from the levelof termination of the piezometer or tensiometer inthe soil to the water level in the manometer and isnegative in the unsaturated soil.
39
elevations. Although depth to water is useful to know, withoutconverting it to a water-level elevation (i.e., hydraulic head),directions and rates of ground-water movement cannot bedetermined.
Hydraulic head data m often displayed in two dimen-sions as a potentiometric surface map (Figure 4-2). Such amap represents the elevation to which water would rise in anopen well placed in the interval of interest. It is analogous to a
topographic map with the direction of water flow from higherto lower elevations and generally running perpendicular to thecontours. However, ground-water flow directions may di-verge from the direction predicted by potentiometric contourswhen the aquifer is anisotropic (hydraulic conductivity is notthe same in all directions). Fetter (1981) describes techniquesfor determining the direction of ground-water flow in aniso-tropic aquifers. Again, using the analogy of the topographicmap, behavior of ground-water flow is similar to how surface
Figure 4.2. Potentlometric surface map (from EPA, 1988).
40
water runoff occurs via overland flow. Different subsurfaceunits or intervals may have different potentiometric surfacemaps. The uppermost potentiometric surface map, which is incontact with the atmosphere through the vadose zone (Chapter5), is the water table.
Although displayed on a two-dimensional surface, thehydraulic head distribution is generally a three-dimensionalphenomena that is, hydraulic head varies vertically as well asareally. To determine the vertical distribution of hydraulichead, wells must be drilled in the same vicinity, but must beopen to different depths (elevations). If hydraulic head in-creases with increasing depth, ground-water flow is upward;in general, this results in an area of discharge. If hydraulichead decreases with increasing depth, ground-water flow isdownward this is an area of ground-water recharge.
Often the stratigraphy supports multiple aquifers that areseparated by confining beds. In these cases, the aquifers aredominated by horizontal flow and the confining layers aredominated by vertical flow, i.e., leakage between adjacentaquifers. At hazardous waste sites, it is important to determinehow many aquifers are contaminated. As part of this determi-nation, the direction of leakage and the direction of flow in theaffected aquifers must be assessed. It is possible that flowdirection in one aquifer could differ from flow in an adjacentaquifer. The difference in hydraulic head over a given dis-tance is known as the hydraulic gradient. Hydraulic gradientsmust be known to determine rates and directions of ground-water movement.
Often, topographic highs are recharge areas and topo-graphic lows are discharge areas. For this reason, surfacewater bodies (such as lakes, rivers, springs, and seeps) areoften surface expressions of the water table. Therefore, thesesurface water bodies are useful for inferring watertable eleva-tion data where no wells exist.
As indicated, ground water generally flows from poten-tiometric highs to potentiometric lows, following a trace thatis perpendicular to the potentiometric contours. This trace issometimes referred to as a flow line. Unlike surface water,however, ground-water flow is resisted by the rock and soilthrough which it flows. This resistance is quantified by thetransmissive properties of the media. As these transmissiveproperties vary at different locations in the aquifer and indifferent directions from a given point, they cause the flowlines to change directions such that they may no longer beperpendicular to the apparent potentiometric contours. There-fore, in addition to hydraulic gradients, the transmissive prop-erties of the media must be known in order to determine ratesand directions of ground-water flow.
The transmissive properties of the media have been givendifferent but related terms, including intrinsic permeability,hydraulic conductivity, and transmissivity. Intrinsic perme-ability is a property of the porous medium and has dimensionsof length squared. It is a measure of the resistance to fluidflow through the medium; the greater the permeability, theless the resistance. Hydraulic conductivity is defined as thevolume of water that will move in unit time under a unithydraulic gradient through a unit area measured at right
41
angles to the direction of flow. It is a property of the fluid andmedium with dimensions of length per time. It is equal to theproduct of intrinsic permeability, density of water, and thegravitational acceleration constant divided by the dynamicviscosity of water. Finally, transmissivity is the rate of waterflow through a vertical strip of aquifer one unit wide, extend-ing the full saturated thickness of the aquifer, under a unithydraulic gradient. It is equal to the product of hydraulicconductivity and the aquifer thickness. Consequently, it hasdimensions of length squared per time.
All these properties can vary spatially and directionally ata given point. If the medium is homogeneous, the transmis-sive properties do not vary spatially. If the medium is isotro-pic, they do not vary when measured in different directionsfrom a given point. Most geologic materials are heteroge-neous and anisotropic.
The final information that is needed to answer the ques-tions about ground-water hydrology that were posed earlierconcerns the local water balance. At hazardous waste sites, itis generally not possible to accurately quantify the local waterbalance, primarily because of data limitations. One of themain goals at a hazardous waste site investigation is to definethe extent of contamination. Consequently, monitoring wellsare clustered near potential sources. Ground-water hydrolo-gists look at the bigger picture to determine what hydraulicboundaries control or influence flow at the site. Data coveringthe larger area are rarely available for hazardous waste sites.Regardless, it is important to attempt the mass balance and toestimate what regional factors control the local flow system.This exercise, while not highly quantitative, will provide avaluable qualitative understanding of the flow system control-ling contaminant migration.
4.2 Field TechniquesGround water is generally below the land surface and,
therefore, difficult to observe. One of the most effectivetechniques for observing ground water is to use point mea-surements made in wells. Wells must be designed, drilled, anddeveloped in order to measure water levels and to take waterquality samples. Tests are conducted to determine transmis-sive and storage properties. The following section discussesmethods used to drill wells, measure water levels, and deter-mine subsurface properties.
4.2.1 Drilling TechniquesTable 4-1 summarizes the advantages and disadvantages
of various drilling methods used for monitoring well construc-tion. In shallow unconsolidated deposits, a hollow stem con-tinuous flight auger is the preferred method. The use ofhollow stem augers (Figure 4-3) requires no fluid in theborehole and allows for installation of the casing and screensprior to removal of the augers, thereby eliminating problemsassociated with caving of the borehole. However, it may bedifficult to seal the annular space in wells constructed in thismanner, and other construction techniques may be more suit-able. In situations where borehole caving is not a problem, theuse of solid stem or bucket augers is equally suitable. Unfortu-nately, the use of augers becomes impractical when drilling
Table 4-1. Auger, Rotary, and Cable-Tool Drilling Techniques-Advantages and Disadvantagesfor Construction of Monitoring Wells
Type Advantages Disadvantages
Auger ● Minimal damage to aquifer Cannot be used in consolidated deposits● No drilling fluids required Limited to wells less than 150 feet in depth● Auger flights act as temporary casing, stabilizing hole for May have to abandon holes if boulders are encountered
well construction● Good technique for unconsolidated deposits● Continuous core can be collected by wireline method
Rotary ● Quick and efficient method ● Required drilling fluids which alter water chemistry● Excellent for large and small diameter holes ● Results in a mud cake on the borehole wall, requiring● No depth limitations additional well devopment, and potentially causing● Can be used in consolidated and unconsolidated deposits changes in chemistry● Continuous core can be collected by wireline method ● Loss of circulation can develop in fractured and high-
permeability material
Cable tool ● No limitation on well depth ● Limited rigs and experienced personnel available● Limited amount of drilling fluid required ● Slow and ineffcient● Can be used in both consolidated and unconsolidated ● Difficult to collect core
deposits● Can be used in areas where lost circulation is a problem● Good Iithologic control● Effective technique in boulder environments
From GeoTrans, 1989
deeper wells (100 to 150 ft) or when hard unconsolidateddeposits are encountered. Thick clay deposits that tend to bindaugers also may make the use of augers impractical. Whendrilling beneath the water table where cross-contaminationbetween water-bearing strata is considered problematic, au-gers may not be the optimum technique. If auger techniquesare used, it may not be possible to prevent fluid flow in theborehole between formations.
When drilling in deeper consolidated deposits, air rotarydrilling (Figure 4-3) is frequently the preferred method be-cause no drilling fluids are employed. However, oil from aircompressors may contaminate the borehole, and special filtersare required to minimize this effect. In some cases, drillersmay use foams to help lift cuttings to the surface and increasethe speed of drilling. Caving of unconsolidated material over-lying consolidated material can frequently limit the use of airrotary drilling. However, some air rotary rigs are equippedwith casing hammers that can drive a casing as drillingproceeds, similar to cable tool drilling techniques (see discus-sion below). Mud rotary techniques also can be used to drillthrough unconsolidated material, a casing can be set to holdthese deposits open, and the hole can be continued with airrotary.
Cable tool drilling methods (Figure 4-3) may be used forconstructing monitoring wells. However, cable tool drillingthrough unconsolidated material, particularly below the watertable, will probably require the simultaneous driving of acasing to prevent caving. Because casing driven in this man-ner may seal strata through which it is driven, this methodmay be used at sites when cross-contamination of waterbearing zones could be a problem. Completing a well casedduring drilling will probably require that the casing be pulled
42
back to expose the formation before setting the screens. Anadvantage of cable tool drilling is that it can be used to drill togreat depths, although a minimum borehole diameter of 3 in.is required. Another advantage is that it can penetrate throughconsolidated material, although frequently at a slow pace.
During drilling for any ground-water contamination in-vestigation, precautions must be taken to prevent cross-con-tamination of boreholes. Thoroughly cleaning the drilling rigand tools initially and after each borehole is drilled are ex-amples of specific precautions that should be taken. No uni-form procedure has been developed for all sites, but a soapwash followed by solvent and distilled water rinse is com-monly used. Proper drilling plans also can minimize potentialcross-contamination. If possible, drilling should progress fromthe least to most contaminated areas (Sisk, 1981).
Upon completion, the monitoring well must be devel-oped. Any contamination or formation damage from welldrilling and any fines from the natural formation must beremoved to provide a particulate-free discharge. A variety oftechniques are available to remove such contamination anddevelop a well (Table 4-2). To be effective, all these tech-niques require reversals or surges in flow to avoid bridging byparticles, which is common when flow is continuous in onedirection. These reversals or surges can be created by usingsurge blocks, air lifts, bailers, or pumps (see Scalf et al.,1981). Natural formation water should be used; use of otherwater is not recommended. The discharge from the wellshould be continuously monitored and development should becontinued until the discharge is particulate-free. Ideally, thewell should be developed so as to minimize the creation ofwater requiring disposal.
Figure 4-3. A conceptual comparison of the hollow-stem auger, the direct-rotary, and the cable-tool drilling methods (fromGeoTrans, 1989).
Table 4-2. Well Development Teohniques-Advantages and Disadvantages
Technique Advantages Disadvantages
Overpumping ● Minimal time and effort required ● Does not effectively remove fine-granted sediments● No new fluids introduced ● Can leave the lower portion of large screen intervals● Remove fluids introduced during drilling undeveloped
● Can result in a large volume of water to be contained anddisposed.
Backwashing ● Effectively rearranges filter pack ● Tends to push fine-grained sediments into filter pack● Breaks down bridging in filter pack ● Potential for air entrapment if air is used● No new fluids introduced ● Unless combined with pumping or bailing, does not
remove drilling fluids
Mechanical surging ● Effectively rearranges filter pack ● Tends to push fine-grained sediments into filter pack● Greater suction action and surging than ● Unless combined with pumping or bailing, does not
backwashing remove drilling fluids● Breaks down bridging in filter pack● No new fluids introduced
High velocity jetting ● Effectively rearranges filter pack ● Foreign water and contaminants introduced● Breaks down bridging in filter pack ● Air blockage can develop with air jetting● Effectively removes the mud cake around ● Air can change water chemistry and biology (iron bacteria)
screen near well● Unless combined with pumping or bailing, does not
remove drilling fluids
From GeoTrans, 1989
43
A variety of materials are available for use in casing,screenings and other structural and sampling components ofmonitoring wells (Table 4-3). Well materials must have suffi-cient strength to ensure the structural integrity of the wellduring installation and during protracted periods of monitor-ing. The materials should sufficiently resist deterioration thatmay result from long-term exposure to natural chemical orpollutant constituents in the ground water at each site. Thematerials also must be selected to minimize their interferencewith the measurement of specific constituents. The mostcommonly used materials are mild steel, stainless steel, poly-vinyl chloride (PVC), polypropylene, polyethylene, andTeflon®. These materials have substantially different proper-ties relative to strength, corrosion resistance, interference withspecific constituent measurements expense, and availability.Consequently, materials should be selected only after consid-eration of all pertinent, site-specific factors such as wellinstallation method, depth, geochemical environment, and
probable contaminants to be monitored. Larson (1981) andBarcelona et al. (1983) have summarized the chemical resis-tance of various casings and well materials to differing envi-ronments. These topics also are discussed in more detail insubsequent chapters of this Handbook.
There are three basic categories of monitoring well de-signs that are used to monitor vertical distribution of contami-nants at a specific location (Figure 4-4). The first type ofnested-sampler design consists of a series of multiple-portsamplers installed in a single borehole. The sampling ports areisolated from each other by inflatable packers or by otherannular seals. In some systems, a special tool is lowered intothe well to open ports at the specific location when a waterlevel or water quality sample is desired. In others, differentplastic (such as nylon) tubings are used for sampling eachzone where a vacuum is used to bring the sample to the
Table 4-3. Well Casing and Screen Material—Advantages and Disadvantages in Monitoring Wells
Type Advantages Disadvantages
Polytetrafluoroethylene(PTFE) or Teflon®
Polyvinylchloride(PVC)
Polyethylene
Polypropylene
Kynar
Stainless steel
Cast iron and low-carbon steel
Galvanized steel
.
.
.
.
.
●
✎
✎
✎
✎
✎
✎
✎
✎
✎
✎
✎
✎
Good chemical resistance to volatile organicsGood chemical resistance to corrosiveenvironments
F l u o r i n a t e d e t h y l e n epropylene (FEP)
LightweightHigh-impact strengthResistant to most chemicals
LightweightResistant to weak alkalis, alcohols, aliphatichydrocarbons, and oilsModerately resistant to strong acids and alkalis
Lightweight
LightweightResistant to mineral acids
Moderately resistant to alkalis, alcohols, ketones,and esters
High strengthResistant to most chemicals and solvents
High strengthGood chemical resistance to volatile organics
High strength
High strength
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Lower strength than steel and iron
Weaker than most plastic material
Weaker than steel and ironMore reactive than PTFE
Deteriorates when in contact with ketones,esters, and aromatic hydrocarbonsLow strengthMore reactive than PTFE, but less reactivethan PVCNot commonly available
Low strengthDeteriorates when in contact with oxidizingacids, aliphatic hydrocarbons, and aromatichydrocarbonsMore reactive than PTFE, but less reactivethan PVC
Not commonly available
Poor chemical resistance to ketones, acetoneNot commonly available
May be a source of chromium in low pHenvironmentsMay catalyze some organic reactions
Rusts easily, providing highly sorptive surfacefor many metalsDeteriorates in corrosive environments
May be a source of zincIf coating is scratched, will rust, providing ahighly sorptive surface for many metals
From GeoTrans, 1989
44
m
Figure 4-4. A conceptual comparison of three multilevel sampling designs (from GeoTrans, 1989).
surface. However, for deep wells and volatile organic chemi-cals, the vacuum may result in unacceptable chemical lossesfrom volatilization. The second configuration for nested-sam-plers consists of multiple well stings installed in one largeborehole (Figure 4-4). Individual zones are isolated from eachother using a low permeability material. Seals between zonesmay be difficult to obtain and maintain.
Finally, the third type of nested-sampler design consistsof drilling a separate borehole for each monitoring well (Fig-ure 4-4). This system is superior to the two previous systemsbecause the potential for cross- contamination from faultyseals is minimized, and smaller diameter holes can be drilled,thereby reducing the volume of water that needs to be pumpedprior to sampling. The additional costs associated with drill-ing multiple boreholes often is offset by technical problemsassociated with the installation of the two previous systems.Use of multiple piezometers and ports in a single boreholeshould be avoided according to U.S. EPA (1986), because thepotential for erroneous data is increased. (A piezometer is asmall-diameter well open to a point in the subsurface.) Table4-4 summarizes the advantages and disadvantages of thesethree multilevel sampling designs.
4.2.2 Methods to Measure Hydraulic HeadThere are a number of ways to measure hydraulic head in
the saturated or vadose zones (see Table 4-5). For conve-nience, both zones are discussed in the following section. Theaccuracy of depth-to-water measurements is discussed in aSuperfund ground-water issue paper (Thornhill, 1989). Whencomparing various methods of measurements, as indicated inTable 4-5, the steel tape method is the most precise. Althoughless precise, the air line method is useful in pumped wellswhere water turbulence exists. Pressure transducers can beused in either the saturated or vadose zones. They are usefulfor making frequent measurements, such as during a slug test.
In a saturated zone, the hydraulic head, H, is measured ata point using a piezometer (see Figure 4- 1) and is defined asthe elevation (pressure head) at which the water surface standsin an open piezometer tube terminated at a given point in theporous medium. Hydraulic head is a combination of pressurehead and elevation head (distance of the measuring pointabove a reference level [datum]). The reference level chosenfor measurement of H is arbitrary. The hydraulic head is apotential function, the potential energy per unit weight of theground water.
45
Table 4-4. Multilevel Monitoring Well Design-Advantages and Disadvantages in Monitoring Wells
Type Advantages Disadvantages
Multiple-port sampler ● Large number of sampling zones per borehole ● Potential for cross-contamination among ports● Smaller volume of water required for purging than ● Potential sampling ports becoming plugged
nested sampler/single borehole and multiple boreholes ● Special sampiing tools required● Lower drilling costs than nested sampler/multiple
boreholes
Nested sampler/single borehole
Nested sampler/multiple boreholes
● Lower drilling costs than multiple boreholes ● Potential for cross-contamination among screen● Low potential for screens becoming plugged intervals
● Number of sampling intervals limited to three orfour
● Larger volume of water required for purgingthan multiple-port campier or nested sampler/multiple boreholes
● Potential for cross-contamination minimized● Voliume of water required for purging smaller than
nested sampler/single borehole● Low installation costs● Low potential for screens becoming plugged
● Higher installation costs
● Higher drilling costs
From GeoTrans, 1989
Table 4-5. Summary of Methods to Measure Hydraulic Head
Method Application Reference
Steel tape
Electric probe
Air line
Mechanical floatrecorder
Pressuretransducer
Acoustic sounder
Tensiometry
Electricalresistivity
Thermocouplepsychrometry
Thermaldiffusivity
Saturated zone. Most precise method. Noncontinuous measurements. Slow
Saturated zone. Frequent measurements possible. Simple to use.Adequate precision
saturated zone. Continuous measurements. Useful for pumping tests.Limited accuracy
Saturated zone. Continuous measurements. Useful for long-term measurements.Permanent record can be delicate
saturated or vadose zone. Continuous or frequent measurements. Rapid responseto changing pressure. Permanent record. Expensive
Saturated zone. Fast; permanent record. Imprecise
Saturated or vadose zone. Laboratory or field method. Useful range is 0 to 0.85bars capillary pressure. Direct measurement. A widely used method
Vadose zone. Laboratory or field method. Useful range is 0 to 15 bars capillarypressure. indirect measurement. Prone to variable and erratic readings
Vadose zone. Laboratory or field method. Useful range 10 to 70 bars capillarypressure. interference from dissolved solutes likely in calcium-rich waste
Vadose zone. Laboratory or field method. Useful range O to 2.0 bars capillarypressure. indirect measurement
Garber and Koopman (1968)
Driscoll (1986)
Driscoll (1986)
USGS (1977)
Gerber and Koopman (1986)
Davis and DeWiest (1966)
Cassei and Klute (1986):Stannard (1986)
Campbell and Gee (1986);Rehm et al. (1987)
Rawlins and Campbell (1986)
Phene and Beale (1976)
Copyright® 1989 Electric Power Research Institute. EPRl EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
Modified from Thompson et al., 1989
46
These same concepts of hydraulic head, pressure head,and gravitational (or elevation) head may be applied to thevadose zone (Chapter 5). A common device used to measurethe hydraulic head in the vadose zone is a tensiometer. It isterminated in the soil by a porous cup permeable to water, butimpermeable to air, when the pores of the cup are filled withwater. The porous cup is necessary to establish hydrauliccontact between the water in the tensiometer and the soilwater. For the vadose zone, the pressure head is inherentlynegative, i.e., the free water surface in the open arm of themanometer will stand below the point of termination in thesoil.
Mercury often is used in the manometer, reducing ma-nometer size (Figure 4-5). Other measuring devices includevacuum gauges and pressure transducers. In areas subject tofreezing, a 40 percent ethylene-glycol solution can be used inthe tensiometer in place of water (Stephens and Knowlton,1986).
The effective pressure range of a standard tensiometer, Oto about -0.08 megapascals (MPa), is limited by the fact thatnegative pressures are measured with reference to atmospheric pressure. Peck and Rabbidge (1966; 1969) developedan osmotic tensiometer for field use that expands the effectivemeasurement range from O to as low as -1.5 MPa. Anotherinstrument that has a wide range of pressure measurements isthe thermocouple psychrometer (Table 4-5).
Hydraulic head can vary temporally at any given well.The variation may be the result of an aquifer’s response to aknown stress (e.g., a pumping well or seasonal changes inrecharge) and may demonstrate a temporal relationship be-tween hydraulic head and contamination concentrations. Forexample, an observation well, located adjacent to a ditch thatonly contains water during the growing season, exhibitschanges in hydraulic head that cause seasonal changes inuranium concentrations (Figure 4-6). This change highlightsthe importance of a sampling frequency sufficient to monitor
Figure 4-5. Schematic illustration of the essential parts of a tensiometer (from Hillel, 1980).
47
Figure 4-8. Hydrography versus uranium concentrations (modified from Goode and Wilder, 1987).
the range of contaminant concentrations that may occur at asite. At many sites, the seasonal variation of hydraulic headalready may be known from existing regional studies. At siteswhere these type of data do not exist, it may be necessary tomonitor water levels on a continuous basis to determine themeasurement frequency.
4.2.3 Methods to Determine Aquifer PropertiesThe aquifer properties considered here include storage
properties and hydraulic conductivity. In addition, methodsare considered for estimating the spatial variability of hydrau-lic conductivity. The methods used to measure or estimatestorage properties are listed in Table 4-6; methods to measureor estimate hydraulic conductivity are listed in Table 4-7; andmethods to measure or estimate spatial variability are listed inTable 4-8.
Determination of aquifer properties begins with identify-ing a known stress to the ground-water system, and thenmeasuring the response to that stress over space and/or time.Given the system geometry and boundary conditions repre-senting the stress, a mathematical description and correspond-ing solution (computed response) can be determined for arange of parameters. The observed aquifer response is matchedto a computed response, and the corresponding parameters are
determined. Thus, aquifer properties are not measured di-rectly, but instead are determined through this curve-matchingprocess. Using more than one method to determine aquiferproperties is recommended. Results then can be weightedtoward the best performed tests with the greatest stress to theaquifer system.
Stress on the ground-water system can result from tidal orriver stage fluctuations, from pumping, or from displacing thewater in a well by a known volume. According to Walton(1987), a pumping tests defined as a field in situ study aimedat obtaining controlled aquifer system response data. Usually,a well is pumped at several fractions of full capacity and/or ata constant rate, and water levels are measured at frequentintervals in the pumped well and nearby observation wells.Measurement of hydraulic head can be important in establish-ing equilibrium water-level conditions prior to pump tests.The influence of ground-water fluctuations external to a pumptest often can be eliminated prior to pump test analyses by thisprocess. Shallow watertable aquifers can exhibit centimeter-range changes in daily ground-water levels due to evapora-tion. Increases in barometric pressure can causecentimeter-range declines in hydraulic head in a well. Waterlevels in wells near coastal areas often respond to ocean tides.Centimeter-range changes in hydraulic head are typical forearth tide responses.
48
Table 4-6. Summary of Methods to Measure Storage Properties
Method Application Reference
Pumping test Can be used to measure storage values for unconfined or confined aquifers.Multiple-well tests are more accurate than single-well tests.Tests a relatively large volume of the aquifer.
Slug test Single-well tests for confined or unconfined aquifers. Test highly influenced by wellconstruction and borehole conditions.
Water-balance Measures specific yield only. Requires several observation wells around pumpingwell to accurately determine the cone of depression. Tests a relatively large volumeof the aquifer.
Laboratory Obtain a maximum long-term value. Fractures, macropores, and heterogeneitiesof geologic material may not be represented. Only specific yield can be determined.
Bureau of Reclamation(1985); Stallman (1971);Driscoll (1986);Lehman (1972)
Hvorslev (1951); Bouwer andRice (1976); Bouwer (1989):Lehman (1972);Cooper et al. (1967)
Nwankwor et al. (1984);Neuman (1972)
Nwankwor et al. (1984)
From Thompson et al., 1989
Copyright 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
Table 4-7. Summary of Methods to Measure Saturated Hydraulic-Conductivity Values in the Field and Laboratory (modified fromThompson et al, 1989)
Method Application Reference
Slug test Confined aquifers with fully penetrating wells screened along the entire aquifer Hvorslev (1951); Bouwer andthickness. Single-well test for wells. Rice (1976); Lehman (1972)
Pumping test Complex multiple-well tests for confined or unconfined aquifers with fully or partially Bureau of Reclamationpenetrating wells. Used for wide range of aquifer permeabilities. Test wells can be (1985); Stallman (1971);used for sampling. Tests a relatively large volume of the aquifer. Driscoll (1986):
Lehman (1972);
Steady-state Laboratory method to determine sample hydraulic conductivity within a range from Klute and Dirksen (1986)permaemeter 1.0 cm/sec to 10-5 crn/sec.
Falling-head Laboratory method to determine sample hydraulic conductivity within a range from Klute and Dirksen (1986)permeameter 10-3 cm/sec to 10-9 cm/sec.
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted with
Drawdown is defined as the drop in water level fromstatic-water level conditions as a result of pumping stress.Time-drawdown and distance-drawdown data are analyzedwith model equations and type-curve matching, straight-linematching, or inflection-point selection techniques. For ex-amples, see Bentall (1963); Ferns et al. (1962); Kruseman andDe Ridder (1976); Lehman (1972); Neuman (1974); Reed(1980); Stallman (1971); Walton (1962); and Walton (1970).One disadvantage of conducting pumping tests at hazardouswaste sites is the disposal of contaminated water. Pumpingtests are valuable, however, because a relatively large portionof the aquifer is stressed. Therefore, the hydraulic conductiv-ity determined from an aquifer test is more representative of
spatially averaged conditions. These type of data are requiredfor final design considerations of a pump-and-treat system.
The slug test method consists of causing a water-levelchange within a well and measuring the rate at which thewater level in the well returns to its initial level. The water-level change can be caused either by injecting or withdrawinga volume of water or weighted float in the well. The rate ofrecovery then can be related to the hydraulic conductivity ofthe surrounding aquifer material. For further information, seeCooper et al. (1967) or Bredehoeft and Papadopulos (1980).
As indicated in Table 4-7, a disadvantage of slug tests isthat only a small volume of aquifer material is tested. If the
49
Table 4-8. Summary of Methods to Measure Spatial Variability of Hydrogeologic Parameters
Method Application Reference
Piezometer slugtests
Localized measurement, influenced by well disturbed zone.Efficient and easy to conduct.
Hvorslev (1951); Bouwer andRice (1976);Lehman (1972)
Hydraulicconductivity fromgrain size
Surfacegeophysics
Boreholegeophysics
Large-scale aquifertests (pumping tests)
Geological mappingof sedimentologicalfacies
Continuous core
Borehole flowmeter
Samples of aquifer material required. Empirical and poor accuracy,especially for silt and day fractions.
Direct current resistivity, electromagnetic induction, streaming potential.Difficult to interpret and poor accuracy.
Natural gamma, gamma-gamma density, single-point resistance, neutron.K= (Ø), Accuracy?
Provides bulk parameters over relatively large region.
Problems with extrapolation-geological sections above water table andaway from site.
Split-spoon sampler, samples are disturbed. Grain size analysis,laboratory K.
Most promising. Equipment difficult to obtain.
Hazen (1982): Krumbein andMonk (1942); Masch andDenny (1966)
Zohdy et al. (1974); Sendleinand Yazicigal (1981);Yazicigal and Sandlein(1982)
Serra (1984); Wheatcraft et al.(1986); Wyllie (1963);Patten and Bennett (1963)
Bureau of Reclamation(1985);Stallman (1971);
Driscoll (1986);Lehman (1972)
Willis (1989); Leeder (1973);Matthews (1974);Turnbull et al. (1950)
Wolf (1988)
Rehfeldt et al. (1988);Hufschmied (1986);Guthrie (1986);Kerfoot (1964)
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprintad withpermission.
well has been damaged (such as from a skin effect fromdrilling mud), then the test may only determine the hydraulicconductivity of the skin (Faust and Mercer, 1984). However,at hazardous waste sites, slug tests offer many advantages,includlng (1) there are no contaminated water disposal prob-lems when a slug rod is used to displace the water, (2) apressure transducer can be used to measure the pressureresponse in wells so that data can be collected even in fairlypermeable material, and (3) decontamination is relativelysimple, allowing as many as a dozen wells to be slug tested ina day. The slug test method is very inexpensive and providesa considerable amount of data on the flow characteristics ofthe subsurface.
One method to determine hydraulic conductivity that islisted in Table 4-8 is grain size analysis. Since Hazen (1892),a number of formulas have been proposed that relate somemeasure of grain size to hydraulic conductivity (for example,Fair and Hatch, 1933; Krumbein and Monk, 1942; Masch andDenny, 1966; and Er-Hui, 1989). These formulas are empiri-cal with hydraulic conductivity proportional to a function ofrepresentative grain diameters. However, these formulas arenot very accurate, and the accuracy decreases when the samples
are predominantly silt or clay. In the early stages of a fieldinvestigation they may be very useful. They also may behelpful in estimating hydraulic conductivity in the vadosezone, which can be a difficult task (see Chapter 5).
Twenty years ago, when hydrologists were mainly inter-ested in water supply, one or two pumping tests were oftensufficient to design an adequate water supply system. With theadvent of contaminant hydrology, more information is re-quired to understand and remediate contamination distribu-tions in the subsurface. In general, the more detailed theinvestigation, the more heterogeneous the subsurface wasobserved to be. Recently, much research has focused on animproved capability to better define spatial variability and itsimpact on chemical transport. Methods used to determinespatial variability as depicted in Table 4-8 were developedfrom information in Waldrop et al. (1989). Another recentreference on this subject is Taylor et al. (1990), in which sixborehole methods are evaluated for determining the verticaldistribution of hydraulic conductivity.
Most of this discussion has focused on hydraulic conduc-tivity; however, many of the methods for determining hydrau-
50
I
lic conductivity also give an estimate of storage properties(Table 4-6). Hydraulic conductivity is needed to calculateground-water velocities and chemical travel times. Storageproperties are also important for the following reasons: (1)porosity is used in chemical travel-time calculations, (2) po-rosity is used to estimate mass in place, and (3) the storageproperties determine how rapidly the flow system will re-spond to pumpage. This latter factor is important for pumpand-treat systems where pulsed pumping is used because thestorage properties can be used to help determine the cycleduration of pumping.
4.3 Analysis of DataOnce ground-water data are collected, they must be ana-
lyzed and interpreted. Numerous analysis tools are available,including graphical methods, mathematical modeling, andgeostatistical techniques. Graphical methods have been usedfor years. However, with the increased use of microcomputersand software such as geographical information systems (GIS),database management systems (DBMS) and plotting pack-ages, it is now easy to view data via a variety of graphicaltechniques. The key is to have field data readily transferred ordirectly recorded on electronic/magnetic format instead ofpaper.
Mathematical models have been used extensively forground-water analysis since the mid-1960s. Models test hy-pothesized conceptualizations of site conditions. They oftenare enhanced by data acquisition and can test the relativeimportance of some information. Knowledge of the varyingimportance of data can help direct the data collection. Thus,where appropriate, using models in unison with active fieldinvestigations can aid in characterization efficiency. Once amodel has been properly calibrated, it can make limitedpredictions about future ground-water flow, contaminant trans-port, or the effectiveness of remedial activities. A large num-ber of models are available and are listed in van der Heijde etal. (1988). NRC (1990) also provides an overview of model-ing.
During the past decade, applying geostatistical principles(i.e., structural analysis, kriging, and conditional simulation)to interpret ground-water data has increased. Geostatisticaltechniques are used to evaluate the spatial variability ofground-water flow parameters, particularly hydraulic headand transmissivity. A code for performing geostatistical as-sessments is provided in Englund and Sparks (1988). Theprinciples of geostatistics may be appropriate for interpolatingpoint data to estimate the spatial distribution of certain aspectsof ground-water quality. Kriging provides a measure of theerror of estimation, which can be mapped and used to selectlocations for additional sampling points. Using this approach,a near-optimal monitoring network can be developed for apredetermined level of reliability.
4.4 Remedial ActionsPumping wells are part of a ground-water flow system. In
many cases, ground-water contamination is discovered be-cause a water-supply well has become affected. These wellscreate cones of depression in the potentiometric surface that
cause water to flow toward them. If that water is carryingcontaminants, they, too, will flow toward the well. Whencontamination is discovered, the immediate response is toshut the well down. This is the correct response, but doing sochanges the ground-water flow system. The potentiometricsurface adjusts to the change in source/sink term, usuallywithin a few days, and chemicals begin to slowly migrate toportions of the aquifer that perhaps were previously uncon-taminated. Therefore, an interim remedial action that shouldbe considered at such sites is well-head treatment. Suchtreatment will bring the well back into production, minimiz-ing the disruption to the water supply. It also will prevent thefurther spread of contamimnts within the aquifer, which,hopefully, will be consistent with any final remediation that isconducted at the site.
Final remedial actions at hazardous waste sites are dis-cussed in OTA (1984) and EPA (1988). Ground-water con-tainment/cleanup options include physical containment (e.g.,construction of low-permeability walls and caps/covers), insitu treatment (e.g., bioreclamation), and hydraulic contain-ment/cleanup (e.g., extraction wells and intercept trenches/drains). To effect complete cleanup, a treatment train combin-ing several methods may be formed.
When a pump-and-treat system is used for cleanup, con-taminated ground-water or mobile nonaqueous phase liquids(NAPLs) are captured and pumped to the surface for treat-ment. This process requires locating the ground-water con-taminant plume or NAPLs in three-dimensional space,determining aquifer and chemical properties, designing a cap-ture system, and installing extraction (and in some casesinjection) wells. Monitoring wells/piezometers, used to checkthe effectiveness of the pump-and-treat system, are an integralcomponent of the system. Injection wells are used to enhancethe extraction system by flushing contaminants (includingsome in the vadose zone) toward extraction wells or drains. Apump-and-treat system may be used in combination withother remedial actions, such as low-permeability walls, tolimit the amount of clean water flowing to the extractionwells, thus reducing the volume of water to be treated. Pumpand-treat technology also can be used as a hydraulic barrier toprevent offsite migration of contaminant plumes from land-fills or residual NAPLs. The basic principle of a barrier wellsystem is to lower ground-water levels near a line of wells,thus diverting ground-water flow toward the pumping wells.
Whether the objective of the pump-and-treat system is toreduce concentrations of contaminants to an acceptable level(cleanup) or to protect the subsurface from further contamina-tion (containment), the system components are
. A set of goals or objectives.
. Engineered components such as wells, pumps, and atreatment facility.
● Performance criteria and monitoring.
. Termination criteria.
51
Each of these components must be a part of the designand evaluation of a pump-and-treat technology.
Pump-and-treat technology is appropriate for manyground-water contamination problems (Ziegler, 1989). Forthis technology to be effective, the physical-chemical subsur-face system Must allow the contaminants to flow to theextraction wells. The subsurface must have sufficient hydrau-lic conductivity to allow fluid to flow readily and the chemi-cals must be transportable by the fluid. These requirementsmake the sure of pump-and-treat systems highly site specific.Cases in which contaminants cannot readily flow to pumpingwells include
✎ Heterogeneous aquifer conditions where low-perme-ability zones restrict contaminant flow toward ex-tinction wells.
l Presence of chemicals that are sorbed or precipitatedon the soil and slowly desorb or dissolve back intothe ground water as chemical equilibrium changes inresponse to the extraction process.
. Presence of immobile NAPMs that may contribute toa miscible contaminant plume by prolonged dissolu-tion (e.g., a separate phase gasoline at residual satu-ration).
In these cases modifications to pump-and-treat technol-ogy, such as pulsed pumping, maybe appropriate. Pump-and-treat technology also may be used in combination (treatmenttrain) with other remedial alternatives, such as vacuum extrac-tion and/or bioremediation. Under complex conditions, nosingle technology is a panacea for subsurface remediation.
The main limitation of pump-and-treat technology is thelong time that may be required to achieve an acceptable levelof cleanup. The length of time results from the “tailing” effectoften observed with this remedial action. Tailing is the asymp-totic decrease of contaminant concentration in water that isremoved in the cleanup process (Figure 4-7). Other potentiallimitations include (1) a design that fails to contain the con-taminant plume and allows continued migration of contami-mnts either horizontally or vertically or, (2) operational failuresthat allow the loss of containment. Typical operational prob-lems stem from the failure(s) of surface equipment or electri-cal and mechanical control systems; and chemical precipitationcausing plugging of wells, pumps, and surface plumbing.Limitations are discussed further in Mackay and Cherry (1989).
Physical containment involves low-permeability barrierssuch as slurry walls. Problems associated with slurry wallsmay involve a difficulty with achieving design permeabilityand underflow; such problems lead to loss of containment.Slurry walls also may be used to prevent the movement ofclean water into an area being remediated by a pump-and-treatsystem, thereby reducing the amount of water that needstreatment. Slurry walls also reduce the amount of fresh groundwater that is contaminated in a pump-and-treat system. Drainsalso can be used to create a hydraulic barrier. Factors thatmust be considered in drain construction include health andsafety during construction, maintenance access, disposal of
excavated soils, and expected volume of water produced.Generally, drains are used in shallow applications where low-permeability material discourages the use of wells. Usingdrains for deeper applications usually is not cost effective.Other ground-water remedial actions are discussedquent chapters.
4.5 Example-Conservation ChemicalCompany Site
in subse-
The Conservation Chemical Company (CCC) site is lo-cated over an alluvial aquifer about 1,000 ft from the MissouriRiver in Kansas City, Missouri (Figure 4-8). Formerly the sitewas used to treat, store, and dispose of hazardous waste. Asmay be seen, the Missouri River Valley is underlain bydeposits of alluvium with an average thickness of 90 to 95 ft.The alluvial sediments contain interbedded clays, silts, sands,and gravels. Although the composition varies locally, thereme some typical characteristics. Grain size increases withdepth, which reflects the depositional history of the MissouriRiver. In many locations, the increasing grain size createsthree layers: (1) the uppermost layer is composed of silts andclays; (2) the intermediate layer includes fine to mediumsands, and (3) the lowest layer is sands and gravels. The upperlayer is approximately 20-ft thick; the intermediate layer 40 to60-ft thick; and the lower layer 30-ft thick. These alluvialdeposits overlie interlayered shales and limestones.
The alluvial aquifer is highly productive and suppliesabout 500,000 gpd to a well located less than 2,000 ft from thesite. The aquifer is generally unconfined; however, short-termresponses to pumping tests and river-level variation indicatesemiconfined conditions. Various hydraulic tests conductedon and near the site indicate that hydraulic conductivityincreases with depth, as can be expected from the grain sizedistribution. Crabtree and Malone (1984) obtained hydraulicconductivity estimates from 0.51 to 2.35 ft/d for the shallowalluvium. Pumping tests at a nearby production well indicatedan overall transmissivity of the aquifer between 4,000 and16,700 ft2/d and a specific yield between 0.15 to 0.27. Slugtests were attempted but proved unsatisfactory because oflarge oscillations (see Chapter 6). Analysis of the response ofthe aquifer to changes in river levels suggests that the ratio ofhorizontal to vertical hydraulic conductivity is about 100:1 forthe site vicinity.
Water levels are from 5 to 15 ft below land surface(Crabtree and Malone, 1984). Water-level data indicate thatfor the area south of the river, ground-water discharges to theriver; however, during periods when the river is high, groundwater flows from the river into the aquifer. This variability isindicated in Figure 4-9 where the vector direction indicatesthe flow direction and its length indicates the gradient magni-tude. These data were collected over a l-year period. Clusterwells indicate a very small vertical hydraulic gradient.
The site was contaminated with metals and organiccompounds. The spatial distribution of concentrations forspecific contaminants did not define a meaningful “plume.”However, concentration of all contaminants tends to decreasewith distance from the site. Also, organic contaminants aregenerally located directly under, northeast, and southeast of
52
Water Filled Aquifer Volumes
Figure 4-7. Effects of tailing on pumping time (from Keeley et al., 1989).
N
Figure 4-8. Bock diagram showing the location of the CCC site and generalized geology.
53
0.004
Figure 4-9. Ground-water flow directions and gradients observed in various piezometers (from Larson, 1986).
the site; concentrations of metals are found north and west ofthe site. For the nearby offsite wells, the highest concentrationof organics generally are found in the deeper wells.
The design of a remedial pumping system at the CCC sitewas complicated by two factors-the impact of the MissouriRiver and the high productivity of the aquifer below the site.Changes in river stage cause significant variations in ground-water flow rates and directions. Consequently, the operatingsystem must be flexible enough to track these changes and tomodify pumping as necessary to meet design objectives.Pumping rates required to achieve design goals are relativelyhigh for all the alternatives considered because of the highproductivity of the aquifer. Even with high pumping rates, thearea of influence or control is difficult to verify because thechanges in water-level elevation, normally used to determineflow direction, are small and difficult to measure.
To evaluate optimal pumping and monitoring strategies,an analytical approach was embedded in a linear program.This approach accounts for variations in flow directions andprovides an analysis of pumping requirements under altern-ative performance criteria. Hydraulic gradients are of particularinterest because performance monitoring of site pumping isbased on the measurement of water-level elevation differ-ences between piezometer pairs. The amount of water pumpedhas been minimized while performance requirements con-tinue to be met. For a site pumping remedy, the quantifiableperformance requirement is a minimum inward hydraulicgradient at paired piezometers.
Numerous simulations were performed using gradientdata provided in Figure 4-9. These simulations were per-formed on both regional and local scales. The regional analy-sis was performed to study the influence of offsite pumpingcenters. Based on these simulations, a recovery system thatmet all the requirements is currently being implemented.
4.6 ReferencesBarcelona, MJ., J.P. Gibb, and R.A. Miller. 1983. A Guide to
the Selection of Materials for Monitoring Well Construc-tion and Ground-Water Sampling. ISWS Contract Report327. Illinois State Water Survey, Champaign, IL.
Bear, J. 1979. Hydraulics of Ground-water. McGraw-Hill,New York.
Bentall, R. (Compiler). 1963. Shortcuts and Special Problemsin Aquifer Tests. U.S. Geological Survey Water SupplyPaper 1545-C.
Bouwer, H. 1978. Ground-Water Hydrology. McGraw-Hill,New York.
Bouwer, H. 1989. The Bouwer and Rice Slug Test—AnUpdate. Ground Water 27(3):304309.
Bouwer, H. and P.C. Rice. 1976. A Slug Pest for DeterminingHydraulic Conductivity of Unconfined Aquifers withCompletely or Partially Penetrating Wells. Water Re-sources Research 12(3):423-428.
Bredehoeft, J.D. and S.S. Papadopulos. 1980. A Method forDetermining the Hydraulic Properties of Tight Forma-tions. Water Resources Research 16(1):233-238.
Bureau of Reclamation. 1985. Ground-Water Manual-AWater Resources Technical Publication, 2nd ed. U.S.Department of the Interior, Bureau of Reclamation, Den-ver, CO.
Campbell, G.S. and G.W. Gee. 1986. Water Potential: Miscel-laneous Methods. In: Methods of Soil Analysis, Part 1,2nd ed., A. Klute (ed.), Agronomy Monograph No. 9,American Society of Agronomy, Madison, WI, pp. 619-633.
Cassel, D.K. and A. Klute. 1986. Water Potential: Tensiom-etry. In: Methods of Soil Analysis, Part 1, 2nd ed., A.Klute (ed.), Agronomy Monograph No. 9, American So-ciety of Agronomy, Madison, WI, pp. 563-596.
Cooper, H. H., Jr., J.D. Bredehoeft, and I.S. Papadopulos.1967. Response of a Finite-Diameter Well to an Instanta-neous Charge of Water. Water Resources Research3(1):263-269
Crabtree, J.E. and P.G. Malone. 1984. Hydrogeologic Charac-terization Conservation Chemical Company Site, KansasCity, Missouri. U.S. Army Corps of Engineers Water-ways Experiment Station, Vicksburg, MS.
Davis, S .N. and R.J. DeWiest. 1966. Hydrogeology. JohnWiley & Sons, New York, 463 pp.
DeWiest, R.J.M. 1969. Flow through Porous Media. Aca-demic Press, New York.
Domenico, P.A. 1972. Concepts and Models in Ground-waterHydrology. McGraw-Hill, New York.
Driscoll, F.G. 1986. Ground-water and Wells, 2nd ed. John-son Division, St. Paul, MN, 1089 pp.
Englund, E. and A. Sparks. 1988. GEO-EAS (GeostatisticalEnvironmental Assessment Software) User’s Guide. EPA/600/4-88/033a (Guide: NTIS PB89-151252, Software:NTIS PB89-151245).
Er-Hui, Z. 1989. Experimental Research on Permeability ofGranular Media. Ground-Water, 27(6):848-854.
Fair, G.M. and L.P. Hatch. 1933. Fundamental Factors Gov-erning the Streamline Flow of Water through Sand. J.Am. Water Works Ass. 25:1551-1565.
Faust, C.R. and J.W. Mercer. 1984. Evaluation of Slug Testsin Wells Containing a Finite-Thickness Skin. Water Re-sources Research 20(4):504-506.
Ferris, J.G., D.B. Knowles, R.H. Brown, and R.W. Stallman.1962. Theory of Aquifer Tests. U.S. Geological SurveyWater Supply Paper 1536-E, 174 pp.
55
Fetter, Jr., C.W. 1981. Determination of the Direction ofGround-Water Flow. Ground-Water Monitoring Review1 (3):28-31.
Freeze, R.A. and J.A. Cherry. 1979. Ground-water. Prentice-Hall, Englewood Cliffs, NJ.
Garber, M.S. and F.C. Koopman. 1968. Methods of Measur-ing Water Levels in Deep Wells. U.S. Geological SurveyTechniques of Water-Resources Investigations TWI 8-A1.
GeoTrans, 1989. Ground-water Monitoring Manual for theElectric Utility Industry. Edison Electric Institute, Wash-ington, DC.
Goode, D.J. and R.J. Wilder. 1987. Ground-water Contamina-tion Near a Uranium Tailing Disposal Site in Colorado.Ground-Water 25(5):545-554
Guthrie, M. 1986. Use of a Geoflowmeter for the Determina-tion of Ground-Water Flow Direction. Ground-WaterMonitoring Review 6(1):81.
Hazen, A. 1892. Experiments upon the Purification of Sewageand Water at the Lawrence Experiment Station. In: 23rdAnnual Report, Massachusetts State Board of Health.
Hillel, D. 1980. Fundamentals of Soil Physics. AcademicPress, New York.
Hufschmied, P. 1986. Estimation of Three-Dimensional Sta-tistically Anisotropic Hydraulic Conductivity Field byMeans of Single Well Pumping Tests Combined withFlowmeter Measurements. Hydrogeologic 1986(2):163-174.
Hvorslev, M.J. 1951. Time Lag and Soil Permeability inGround-water Observations. U.S. Army Corps of Engi-neers Waterways Experiment Station, Bull. 36, Vicksburg,MS.
Keeley, J.W., D.C. Bouchard, M.R. Scalf, and C.G. Enfield.1989. Practical Limits to Pump and Treat Technology forAquifer Remediation. Submitted to Ground-Water Moni-toring Review.
Kerfoot, W.B. 1984. Darcian Flow Characteristics Upgradientof a Kettle Pond Determined by Direct Ground-WaterFlow Measurement. Ground-Water Monitoring Review4(4):188-192.
Klute, A. and C. Dirksen. 1986. Hydraulic Conductivity andDiffusivity: Laboratory Methods. In: Methods of SoilAnalysis, Part 1, 2nd ed., A. Klute (ed.), AgronomyMonograph No. 9, American Society of Agronomy, Madi-son, WI, pp. 687-734.
Krumbein, W.C. and G.D. Monk. 1942. Permeability as aFunction of the Size Parameters of Unconsolidated Sand.AIMME Petroleum Division Technical Pub. No. 1492(published in Petroleum Technology Vol. 5, No. 4).
Kruseman, G.P. and N.A. De Ridder. 1976. Analysis andEvaluation of Pumping Test Data. International Institutefor Land Reclamation and Improvement, Wageningen,The Netherlands, 200 pp.
Larson, D. 1981. Materials Selection for Ground-Water Moni-toring. Paper Presented at the National Water Well Asso-ciation Short Course entitled Practical Considerations inthe Design and Installation of Monitoring Wells, Colum-bus, OH, December 16-17.
Larson, S.P. 1986. Notes from November Meeting betweenSettling Defendants and EPA, Washington, DC.
Leeder, M.R 1973. Fluviatile Fining-Upward Cycles and theMagnitude of Paleochannels. Geology Magazine110(3):265-276.
Lohman, S.W. 1972. Ground-water Hydraulics. U.S. Geo-logical Survey Professional Paper 708.
Mackay, D.M. and J.A. Cherry. 1989. Ground-Water Con-tamination: Pump-and-Treat Remediation. Environ. Sci.Technol. 23(6):630-636.
Masch, F. and K. Denny. 1966. Grain Size Distribution and ItsEffect on the Permeability of Unconsolidated Sands.Water Resources Research 2(4):665-577.
Matthews, R.K. 1974. Dynamic Stratigraphy. Prentice Hall,Englewood Cliffs, NJ, Chapter 10, pp. 137-172.
National Research Council (NRQ. 1990. Ground-Water Mod-els Scientific and Regulatory Applications. National Acad-emy Press, Washington, DC, 303 pp.
Neuman, S.P. 1972. Theory of Flow in Unconfined AquifersConsidering Delayed Response of the Water Table. Wa-ter Resources Research 8(4): 1031-1045.
Neuman, S.P. 1974. Effect of Partial Penetration on Flow inUnconfined Aquifers Considering Delayed Gravity Re-sponse. Water Resources Research 10(2):303-312.
Nwankwor, G. I., J.A. Cherry, and R.W. Gillman. 1984. AComparative Study of Specific Yield Determinations fora Shallow Sand Aquifer. Ground-Water 22(6):764-772.
Office of Technology Assessment (OTA). 1984. Protectingthe Nation’s Ground-water from Contamination. OTA-O-233. U.S. Office of Technology Assessment, Washing-ton, DC.
Patten, Jr., E.P. and G.D. Bennett. 1963. Application of Elec-trical and Radioactive Well Logging to Ground-WaterHydrology. U.S. Geological Survey Water Supply Paper1544-D, 60 pp.
Peck, A.J. and R.M. Rabbidge. 1966. Soil Water Potential:Direct Measurements by a New Technique. Science151:1385-1386.
56
Peck, AJ. and R.M. Rabbidge. 1969. Design and Perfor-mance of an Osmotic Tensiometer for Measuring Capil-lary Potential. Soil Sci. Soc. Am. Proc. 33:196-202.
Phene, CJ. and D.W. Beale. 1976. High-Frequency Irrigationfor Water-Nutrient Management in Humid Regions. SoilSci. Soc. Am. J. 40:430-436.
Rawlins, S.L. and G.S. Campbell. 1986. Water Potential:Thermocouple Psychrometry. In: Methods of Soil Amly-sis, Part 1, 2nd ed., A. Klute (ed.), Agronomy MonographNo. 9, American Society of Agronomy, Madison, WI, pp.597-618.
Reed, J.E. 1980. Type Curves for Selected Problems of Flowto Wells in Confined Aquifers. U.S. Geological SurveyTechniques of Water-Resources Investigations TWI 3-B3.
Rehfeldt, K.R., P. Hufschmied, L.W. Gelhar, and M.E.Schaefer. 1988. The Borehole Flowmeter Technique forMeasuring Hydraulic Conductivity Variability. Draft topi-cal report prepared by MIT for Electric Power ResearchInstitute, Research Project 2485-5.
Rehm, B. W., B.J. Christel, T.R. Stolzenburg, D.G. Nichols,B. Lowery, and B.J. Andraki. 1987. Field Evaluation ofInstruments for the Measurement of Unsaturated Hydrau-lic Properties of Fly Ash. EPRI EA-5011. Electric PowerResearch Institute, Palo Alto, CA.
Scalf, M.R., J.F. McNabb, W.J. Dunlap, R.L. Cosby, and J.Fryberger. 1981. Manual of Ground-Water Quality Sam-pling Procedures. EPA/600/2-81/160, (NTIS PB82-103045). Also published in NWWA/EPA Series, NationalWater Well Association, Dublin OH.
Sendlein, L.V.A. and H. Yazicigal. 1981. Surface Geophysi-cal Techniques in Ground-Water Monitoring, Part 1.Ground-Water Monitoring Review 1(4):42-46.
Serra, O. 1984. Fundamentals of Well-Log Interpretation, 1:The Acquisition of Logging Data. In: Developments inPetroleum Science, Vol. 15A. Elsevier, New York, 423PP.
Sisk, S.W. 1981. NEIC Manual for Ground-water/SubsurfaceInvestigations at Hazardous Waste Sites. EPA/330/9-81-002 (NTIS PB82-103755).
Stallman, R.W. 1971. Aquifer-Test Design, Observation, andData Analysis. U.S. Geological Survey Techniques ofWater-Resources Investigations TWI 8-B1.
Stannard, D.I. 1986. Theory, Construction and Operation ofSimple Tensiometers. Ground-Water Monitoring Review6(3):70-78.
Stephens, D.B. and R. Knowlton, Jr. 1986. Soil Water Move-ment and Recharge through Sand at a Semiarid Site inNew Mexico. Water Resources Research 22(6):881-889.
Taylor, K., S. Wheatcraft, J. Hess, J. Hayworth, and F. Molz.1990. Evaluation of Methods for Determining the Verti-cal Distribution of Hydraulic Conductivity. Ground-Wa-ter 28(1):88-98.
Thompson, C.M., et al. 1989. Techniques to Develop Data forHydrogeochemical Models. EPRI EN-6637. ElectricPower Research Institute, Palo Alto, CA.
Thomhill, J.T. 1989. Accuracy of Depth to Water Measure-ments. EPA Superfund Ground-Water Issue Paper. EPA/540/4-89/002.
Todd, D.K. 1980. Ground-water Hydrology. John Wiley &Sons, New York.
Tumbull, W.J., E.S. Krinitsky, and L.J. Johnson. 1950. Sedi-mentary Geology of the Alluvial Valley of the Missis-sippi River and its Bearing on Foundation Problems. In:Applied Sedimentation, P.D. Trask, (ed.), John Wiley &Sons, New York, pp. 210-226.
U.S. Environmental Protection Agency (EPA). 1986. RCRAGround-Water Monitoring Technical Enforcement Guid-ance Document. EPA OSWER-9950. 1. Also published inNWWA/EPA Series, National Water Well Association,Dublin, OH.
U.S. Environmental Protection Agency (EPA). 1987. Hand-book Ground-water. EPA/625/6-87/016.
U.S. Environmental Protection Agency (EPA). 1988. Guid-ance on Remedial Actions for Contaminated Ground-Water at Superfund Sites. Advance Copy, OSWERDirective No. 9283.1-2.
U.S. Geological Survey (USGS). 1977. National Handbook ofRecommended Methods for Water Data Acquisition(Chapter 2-Ground-Water, updated January 1980).USGS Office of Water Data Coordination, Reston, VA.
van der Heijde, P.K M., A.I. E1-Kadi, and S.A. Williams.1988. Ground-water Modeling An Overview and StatusReport. EPA/600/2-89/028.
Waldrop, W.R., K.R. Rehfeldt, L.W. Gelhar, J.B. Southard,and A.M. Dasinger. 1989. Estimates of MacrodispersivityBased on Analyses of Hydraulic Conductivity Variabilityat the MADE Site. EPRI EN-6405. Electric Power Re-search Institute, Palo Alto, CA.
Walton, W.C. 1962. Selected Analytical Methods for Welland Aquifer Evaluation. ISWS Bulletin 49. Illinois StateWater Survey, Champaign, IL.
Walton, W.C. 1970. Ground-water Resource Evaluation.McGraw-Hill, New York.
Walton, W.C. 1987. Ground-water Pumping Tests Design andAnalysis. Lewis Publishers, Chelsea, MI, 201 pp.
57
Wheatcraft, S.W., K.C. Taylor, J.W. Hess, and T.M. Morris.1986. Borehole Sensing Methods for Ground-Water In-vestigations at Hazardous Waste Sites. EPA/6OO/2-86/l1(NTIS PB87-132783).
Willis, B J. 1989. Paleochannel Reconstruction from PointbarDeposits: A Three-Dimensional Perspective. Sedimen-tology 36:757-766.
Wolf, S. 1988. Spatial Variability of Hydraulic Conductivityin a Sand and Gravel Aquifer. Master of Science Thesis,Department of Civil Engineering, MIT, Cambridge, MA.
Yazicigal, H. and L.V.A. Sendlein. 1982. Surface Geophysi-cal Techniques in Ground-Water Monitoring, Part II.Ground-Water Monitoring Review 2(1):56-62.
Ziegler, G.J. 1989. Remediation through Ground-Water Re-covery and Treatment. Pollution Engineering 22(7):75-79.
Zohdy, A.A.R., G.P. Eaton, and D.R. Mabey. 1974. Applica-tion of Surface Geophysics to Ground-Water Investiga-tions. U.S. Geological Survey Techniques ofWater-Resources Investigations TWI 2-DI.
Wyllie, M.R.J. 1963. The Fundamentals of Well Log Interpre-tation. Academic Press, New York, 238 pp.
58
Chapter 5Characterization of the Vadose Zone
James W. Mercer and Charles P. Spalding
The vadose zone is the subsurface extending from landsurface to the water table. It also is called the zone of aeration,variably saturated zone, or the unsaturated zone. Use of thislatter term is discouraged, however, because since the vadosezone contains moisture up to 100 percent saturation, the termunsaturated could be misleading. Depth of the vadose zonecan vary greatly depending on the region of the site. Forexample, in the humid eastern portion of the United States, thevadose zone can be only a few feet thick, disappearing duringtimes of the year when the water table is high. In the arid west,the vadose zone can be several hundred feet thick.
Because the vadose zone overlies the saturated zone,chemical releases at or near the land surface must pass throughthe vadose zone before reaching the water table. Therefore, atmany contaminated sites, often both the vadose zone and thesaturated zone need to be characterized and remediated (i.e.,treatment trains must be applied). As discussed later in thischapter, the vadose zone can have more complex flow condi-tions than the saturated zone. These conditions can be difficultto characterize. On the other hand, because the vadose zone isnearer to the land surface, for remedial actions, the flowsystem may not need to be completely characterized undercertain site conditions and contaminants.
The main difference between the saturated and vadosezones is the presence of air/gas in the pore spaces of thevadose zone. The amount of water and air varies both spatiallyand temporally, which contributes to the complex nature ofthe vadose-zone flow system. However, the presence of soilgas also provides a valuable screening tool for locating vola-tile organic compounds (VOCs) In addition, there is thepotential for significant biological activity. The advantagesand disadvantages of characterizing and remediating the va-dose zone are discussed in the following sections: (1) Reviewof Concepts, (2) Field Techniques, (3) Analysis of Data, and(4) Remedial Actions. An example of the application oftechniques as discussed in the chapter follows these sections.
5.1 Review of ConceptsThe vadose zone can be divided into (1) the belt of soil
water, (2) the intermediate belt, and (3) the capillary fringe.The belt of soil water is the uppermost zone extending fromthe land surface to a depth where soil moisture changes areminimal. It contains the root zone of plants, and is the site ofmany active processes. Precipitation, for example, falls to the
land surface and runs off via overland flow or infiltrates intothe ground. Working against the infiltrating water are evapo-ration and transpiration. Evaporation is the process that con-verts the water at or near land surface to vapor. Transpirationis the process by which plant roots absorb water and releasewater vapor back to the atmosphere through their leaves andstems. Hydrologists combine these two processes into theterm evapotranspiration. Much of the infiltrating water isconsumed by evapotranspiration. The water that is not con-sumed and eventually makes it to the water table is recharge.It is important to understand and characterize these processesfor hazardous waste sites (1) to help understand rechargeevents and how contaminants may move through the vadosezone, and (2) to help design caps used to limit infiltration andrecharge to a contaminant source area.
The capillary fringe (at the base of the vadose zone)extends upward from the water table until there is a decreasein soil moisture. Portions of this zone can be at 100 percentsaturation. This zone also will change as recharge/dischargecauses the water table to fluctuate. The capillary fringe isformed due to a capillary rise caused by the surface tensionbetween air and water. Hydraulic head is made up of anelevation head and a pressure head. At the water table, thepressure head is zero. It increases below the water table anddecreases above the water table. That is, pressure head isnegative in the vadose zone, a phenomenon sometimes re-ferred to as soil tension or suction. The latter term refers to theeffect of water being sucked into a dry soil. The negativepressure head will pull water upward from the saturated zone,forming the capillary fringe. The height of the capillary fringedepends on the pore size of the soil (e.g., the capillary rise isgreater for smaller pores). Unfortunately, pore size is difficultto determine and is not directly related to grain size.
Hydraulic head in the vadose zone is defined the sameway as it is in the saturated zone-the sum of pressure headand elevation head. In the vadose zone, however, pressurehead is used for the saturation-dependent relationships. Capil-lary pressure, defined as the difference between the nonwettingfluid pressure and the wetting fluid pressure, also is used. Foran air-water system, the air pressure is assumed to be negli-gible, and capillary pressure is essentially equal to the nega-tive of the pressure head.
The moisture present in the vadose zone is quantified bya term called the volumetric water content or degree of
59
Figure 5-1. Moisture characteristic or specific retention curves for various soil types.
60
saturation. Saturation varies from zero to one and refers to theamount of volume of pore space filled with water. Volumetricwater content varies between zero and the porosity value. Forcomplete saturation, the volumetric water content is equal toporosity, and the degree of saturation is 100 percent or 1.0. Ifthe pore space is only half filled with water, then the satura-tion is 50 percent or 0.5 and volumetric water content is halfthe porosity.
In the vadose zone, a relationship called the moisturecharacteristic curve exists between volumetric water contentand pressure head (Figure 5-1). As the figure shows, thiscurve is nonlinear and generally is not a single-valued-func-tion relationship. That is, a different curve is used to describethe pressure-head-volumetric-water-content relationship de-pending on whether the soil is filling or draining. Dependingon the wetting history, an entire set of curves is needed. Thisphenomenon is called hysteresis, and is due in part to en-trapped air in the soil after wetting. This set of curves isnecessary to fully describe the flow conditions in the vadosezone.
Flow in the vadose zone is complicated further by thepresence of air. Because both air and water are in the porespace, each resists the flow of the other. This results in adecrease in fluid mobility, characterized by the term relativepermeability. Relative permeability varies between zero andone. It is a nonlinear function of saturation that also canexhibit hysteresis. Thus, to fully characterize flow in thevadose zone, the relative permeability function must be known,in addition to the saturated hydraulic conductivity.
5.2 Field TechniquesBased on the review of concepts, near-surface processes,
as well as other parameters that are functions of moisturecontent, need to be characterized. For hazardous waste reme-diation, vadose zone processes must be understood to designcaps and covers to minimize infiltration. Methods to measureor estimate these processes/parameters are discussed in thissection. Reviews of vadose zone monitoring are discussed inWilson (1980, 1981, 1982, 1983). Section 9.2 further dis-cusses sampling of subsurface solids and vadose zone water,and Table 9-5-identifies additionalcharacterization of the vadose zone.
references focusing on
5.2.1 Precipitation and InfiltrationPrecipitation is defined as the total amount of water that
reaches land surface, and is measured with gauges as a depthof water (see Table 5-l). Because weather stations are notgenerally set up at hazardous waste sites, precipitation infor-mation is obtained from nearby airports. Another source ofprecipitation data is the National Climatic Data Center inAsheville, North Carolina. Wind velocity and air temperaturealso are studied for remediation.
The maximum rate at which water can enter a soil is theinfiltration capacity or potential infiltration rate. The maxi-mum rate occurs when the water supply at the surface isunlimited. During precipitation events, all the water willinfiltrate if the rainfall intensity is less than the infiltrationcapacity. If this capacity is exceeded, the excess rain cannotinfiltrate and will produce surface runoff. Although this dis-cussion concerns water infiltration, it also applies to a chemi-cal spill infiltrating the subsurface. Infiltration capacity varieswith time; it is highest at the begiming of a precipitation eventand decreases as the soil becomes saturated. Table 5-2 listsmethods to measure or estimate infiltration rates. These meth-ods are discussed in Thompson et al. (1989) and in thereferences provided in the table.
Spatial variability is present in the vadose zone as well asthe saturated zone. Spatial variability produces a fingering offlow as it moves downward from the surface. This means thatthe wetting front does not move as a sharp front, but insteadmoves downward with an irregular shape where some zones(fingers) move more rapidly than other zones. Laboratorystudies by Stephens and Heermann (1988) suggest that thisvariability increases with decreasing soil moisture content.
5.2.2 Evaporation and EvapotranspirationEvaporation is the loss of water from the soil into the
atmosphere. In the absence of vegetative cover, the bare soilsurface is subject to radiation and wind effects, and soil waterevaporates directly from the soil surface. An associated pro-ccss is evaporation of water from plants, or transpiration. Forevaporation to occur (1) a continual supply of heat must meetthe latent heat requirements, (2) a vapor pressure gradientmust exist between the soil surface and the atmosphere, and(3) there must be a continual supply of water from and/orthrough the soil layers. The first two conditions determine theevaporative demand (Table 5-3) and are controlled by micro-meteorological factors such as air temperature, humidity,
Table 5-1. Summary of Methods to Measure Precipitation
Method Application Reference
Sacramento gage Accumulated precipitation. Manual recording. Finkelstein et al. (1989);National Weather Service (972)
Weighing gage Continuous measurement on precipitation. Mechanical recording. Finkelstein et al. (1989)
Tipping-bucket gage Continuous measurement of precipitation. Electronic recording. Finkelstein et al. (1989)Recommended.
From Thompson et al., 1989
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
61
Table 5-2 Summary of Methods to Measure or Estimate infiltration Rates
Method
Infiltrometers
Sprinklerinfiltrometer
Average infiltrationmethod
Empiricalrelations
infiltrationequations
Application Reference
Measures the maximum infiltration rate of surface soils. Useful for determining Dunne and Leopold (1978);relative infiltration rates of different soil types: however, infiltration rates Bouwer (1986)determined by this method tend to overestimate actual rates.
Measures the potential range of infiltration rates under various precipitation Dunne and Leopold (1978);conditions. Tends to be expensive and non-portable. Sprinkler infiltrometers Peterson and Bubenzerhave typically been used for long duration research studies. (1986)
Method for estimating the average infiltration rate for small watersheds. Dunne and Leopold (1978)Provides an approximate estimate of infiltration for specific precipitation eventsand antecedent moisture conditions.
Methods to approximate the infiltration for large watersheds. These methods can be Musgrave and Holtan (1964)useful when combined with limited infiltrometer measurements to obtain a grossapproximation of infiltration.
Analyticai equations for calculating infiltration rates. Parameters required in the Bouwer (1986);equations can be readily measured in the field or obtained from the literature. Green and Ampt (1911);Probably the least expensive and most efficient method for estimating infiltration. Philip (1957)
From Thompson et al., 1989
Copyright® 1969 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
Table 5-3. Summary of Methods to Measure Evaporation
Method Application Reference
Class-A pan Evaporation from surface of free liquid. Veihmeyer (1964);National Weather Service (1972)
Weighing lysimeter Direct measure of bare soil evaporation. USGS (1977) (updated 1982)
Remote sensing Currently in development. Useful for large areas. USGS (1977) (updated 1982)
Modified from Thompson et al., 1989
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted with
wind velocity, radiation, and crop cover. The third condition,which determines the rate of water supply to the evaporativesite (soil surface), is controlled by soil-water content, pressurepotential, and relative permeability of the soil. Thus, theactual evaporation rate is determined by evaporative demandand soil hydraulic properties.
Transpiration occurs in response to a vapor pressuredeficit between leaves and the atmosphere. To meet thisdemand, plants must extinct water from the root zone. Com-bined losses due to evaporation and transpiration are com-monly referred to as evapotranspimtion. When the soil surfaceis covered completely by a crop canopy, evaporation lossesare negligible, and transpiration is the principal process bywhich water is lost from the root zone. The same environmen-tal factors that control evaporation also control the potentialtranspiration. Table 5-4 summarizes methods to measure orestimate evapotranspiration.
5.2.3 Moisture Content and MoistureCharacteristic Curves
In the vadose zone, the void space is partly filled by airand partly by water. The moisture content or volumetric water
content represents the quantity of water present at a certaintime at a point in the porous media. The maximum value ofvolumetric water content occurs when all voids are filled; theminimum value occurs when all voids are empty (filled withair). Thus, moisture content varies between O to the value ofthe soil porosity.
Changes in moisture content are important to detect. Forexample, under a cap/cover, changes in moisture contentcould indicate leaks in the cover. By determining moisturecontent with depth, perched water zones can be located for usein water quality sampling. Several methods are used to mea-sure moisture content (see Table 5-5), but the recommendedtechniques are gravimetric and neutron scattering. Gravimet-ric moisture content measurements are made by weighingsoils before and after drying. The neutron scatter methodlowers the moisture meter, which contains a source of fastneutrons and a slow neutron detector, into the soil through anaccess tube (Figure 5-2). Neutrons are emitted by the source(e.g., radium or americium-beryllium) at a very high speed.When these neutrons collide with a small atom, such ashydrogen contained in soil water, their direction of movementis changed and they lose part of their energy. These “slowed”neutrons are measured by a detector tube and a scalar. This
62
Table 5-4. Summary of Methods to Measure or Estimate Evapotranspiration
Method Application Reference
WATER BALANCE METHODS
Pan Iysimeter
Soil moisture sampling
Potential evapotranspirometers
Cl tracer
Water-budget analysis
Ground-water fluctuation
Direct field method; accurate; moderate to low cost.
Direct field method; accurate; moderate to low cost.
Direct field method of PET Moderately accurate andlow cost.
Indirect combined field and laboratory method;moderate to high cost.
Indirect field estimate of ET; manageable to difficult;moderate to low cost.
Indirect field method; moderate to low cost.
MICROMETEOROLOGICAL METHODS
Profile method Indirect field method.
Energy budget/ Indirect field method; difficult; costly; requires dataBowen ratio which is often unobtainable; research oriented.
Eddy covariance method Indirect field method; costly measures water-vapor fluxdirectly; highly accurate; well accepted; research oriented.
Penman equation Indirect field method; ditficult; costly; very accurate;eliminates need for surface temperature measurements;research oriented.
Thornthwaite equation Empirical equation; most accepted for calculating PET:uses average monthly sunlight: moderate to low cost.
Blaney-Criddle equation Empirical equation; widely used; moderate to highaccuracy; low cost; adjusts for certain crops and vegetation.
Veihmeyer (1964);Sharma (1985)
Veihmeyer (1984)
Thornthwaite and Mather (1955)
Sharma (1985)
Davis and DeWiest (1966)
Davis and DeWiest (1966)
Sharma (1985)
Veihmeyer (1964);Shamra (1985)
Veihmeyer (1964); Sharma (1985)
Veihmeyer (1964); Sharma (1985)
Veihmeyer (1964); Sharma (1985)
Stephens and Stewart (1964)
From Thompson et al., 1989
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
Table 5-5. Summary of Methods for Measuring Moisture Content
Method Application Reference
Gravimetric Laboratory measurements of soils which should be dried at Gardner (1986):1100C. The standard method for moisture content determination. Radian Corporation (1988)Recommended.
Neutron scattering In situ measurements via installed access tubes. Widely used. van Bavel (1963)Requires calibration curves. Recommened.
Gamma ray In situ measurements via installed access tubes. Difficult to use. Gardner (1986)attenuation Not recommended for routine use.
Electromagnetic In situ measurements from implanted sensors. Not widely used. Schmugge et al. (1980)Not recommended for routine use.
Tensiometry In situ measurements inferred from moisture-matric potential relationship. Gardner (1986)Prone to error resulting from uncertainty of moisture-matric potentialrelationship. Not recommended.
From Thompson et al., 1989
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
63
Figure 5-2. Components of the neutron moisture meter (fromHillet, 1980).
reading is then related to the soil moisture content in thevadose zone and porosity in the saturated zone. These mea-surements are good indicators of relative changes in moisturecontent; absolute values of moisture content are difficult todetermine.
If the moisture characteristic curve is known (Figure 5-2),then pressure head can be measured using, for example, atensiometer, and then converted to moisture content using thecharacteristic curve. Because of the uncertainty involved,however, this approach is not recommended.
In a saturated soil at equilibrium with free water at thesame elevation, the matric potential or negative pressurepotential is atmospheric and hence equal to zero.Subatmospheric pressure (suction or tension) applied to soildraws water out of the soil, as the voids cannot retain wateragainst the applied suction. Thus, increasing matric potentialis associated with decreasing volumetric water content. Thesoil-water retention curve, also known as the soil-water char-acteristic curve, expresses this relationship. In most soils,drainage (drying) and infiltration (wetting) produce differentwater retention curves (Figure 5-3). This is because air that istrapped in the pores upon wetting decreases the water content.In this case, the soil-water characteristic curve is said todisplay hysteresis. Table 5-6 lists methods for determiningmoisture characteristic curves.
5.2.4 Vadose-Zone Hydraulic ConductivityThe hydraulic conductivity of a porous medium is largest
at saturation and decreases as the water content decreases. Thesaturated hydraulic conductivity in the vadose zone, as well asthe relationship between water content and hydraulic conduc-
tivity, must be determined. At relatively low water contents,the hydraulic conductivity decreases primarily because airoccupies more of the pore space, leaving less cross-sectionalarea available for water transport. The film of water coveringthe soil particles becomes thinner and thinner, until at lowwater contents, it becomes thin enough that attractive forcesbetween the water molecules and the soil particles becomestronger than other forces that might be acting to make watermove; at this point, the hydraulic conductivity approacheszero. Hence, in the vadose zone, hydraulic conductivity isexpressed as a function of moisture content or pressure head.
Measuring vadose-zone hydraulic conductivity values isdifficult because head gradients, flow rates, and moisturecontent or pressure head also must be measured. Factors thatinfluence these measurements include soil texture, soil struc-ture, initial water content, depth of water table, water tempera-ture, entrapped air, biological activity, entrained sediment inthe applied water, and chemistry of the applied water (Wilson,1982).
Relative permeability also must be determined. The rela-tive permeability is a normalized coefficient, which whenmultiplied by the saturated hydraulic conductivity, yields thevadose-zone hydraulic conductivity. It is typically presentedas either a function of capillary pressure or saturation. Rela-tive permeability ranges from one at 100 percent saturation tozero at residual saturation, the water saturation where thewater phase becomes disconnected and ceases to flow.
A number of empirical equations have been developedfor approximating the vadose-zone permeability of isotropicporous media. Three commonly used equations for estimatingthe vadose-zone hydraulic conductivity are those by Brooksand Corey (1964), Mualem (1976), and van Genuchten (1980).Methods to determine the vadose-zone hydraulic conductivityare listed in Table 5-7 and discussed in Thompson et al.(1989). Figure 5-4 shows typical relative permeability curvescomputed using van Genuchten (1980).
5.2.5 Soil Gas AnalysisAlthough not strictly flow related, soil gas analysis is an
important remote sensing tool for locating areas contaminatedby VOCs in the vadose zone. This method requires the drillingof a shallow hole or the injection of a sample tube into the soil.Volumes of soil gas are evacuated to the surface for collectionand analysis at a remote lab or measured on site by a lab-quality vapor analyzer. This method also can be used toanalyze cuttings from well drilling operations or in caseswhere installed wells yield no water. Soil gas analysis isdependent upon the pore spacing within the soil and is lessreliable in tightly packed soils such as clay. It also cannot beused to detect nonvolatile organic compounds and inorganiccompounds (see Table 5-8). Section 9.2.2 provides somefurther discussion of soil gas sampling techniques.
Using soil vapor monitoring wells to detect plumes ofground water contaminated with VOCs has been suggested asa cost-effective means of tracing ground-water contamination(e.g., Marrin and Kerfoot, 1988). Indeed, some success inusing this technique has been reported (Marrin and Thomp-
64
Water Content
Figure 5-3. Soil-water characteristic curve displaying hysteresis (modified from Hillei, 1980).
Table 5-6. Summary of Methods for Determining Moisture Characteristic Curves
Method Application Reference
Porous plate Standard laboratory method for measurement of soils. Klute (1986)Can be used to characterize both wetting and drying behavior.
Vapor equilibration 8est suited for matric potentials less than -15 bars. Klute (1986)
Osmotic Similar to porous plate method. Requires long equilibration times. Klute (1986)Not recommend.
From Thompson et al., 1989
Copyright@ 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models, Reprinted withpermission.
son, 1987; Marrin and Kerfoot 1988). See, also, the soil gassampling case studies summarized in Table 9-6.
If the source of the VOCs is below the water table, thenthe maximum concentration of the organics in the unsaturatedzone is the top of the capillary fringe. Once the contaminantshave reached the top of the capillary fringe they should diffusevery rapidly because of the large gas-phase diffusion coeffi-cients in the unsaturated zone. This rapid mass transfer fromthe water in capillary fringe to the soil air just above it shoulddeplete the capillary fringe of the volatile contaminant. Theconcentrations in the unsaturated zone, therefore, are morecontrolled by the rate of mass transfer from the ground waterto the top of the capillary fringe, a process controlled by thevery low solute diffusion coefficients. Laboratory studies ofmass transfer across the capillary fringe substantiate these
ideas. With the additional loss of mass by mass transfer acrossthe atmosphere-soil boundary and by biodegradation that alsomay be occurring (e.g., Huh et al., 1987), concentrations inthe unsaturated zone are expected to be very low. The bestopportunity for detecting VOC contaminants under these con-ditions is to use soil-gas monitoring wells installed just abovethe capillary fringe.
There are, of course, exceptions to this scenario. If thereis residual nonaqueous phase liquid (NAPL) in the unsatur-ated zone or product floating on the water table, then soil gasmonitoring would detect the volatiles. In the absence of anyNAPL, VOCs may be detected by soil-gas monitoring if thewater table fluctuates enough to bring the contaminated waterup into the unsaturated zone and leave it there as part of theresidual phase. The VOCs would then partition from the
65
Table 5-7. Summary of Methods to Measure Vadose-Zona Hydraulic-Conductivity Values in the Field and Laboratory
Method Application Reference
Constant-headboreholeinfiltration
Guelphpermeameter
Air-entrypermeameter
Instantaneousprofile
Crust-imposedsteady flux
Sprinkler-imposedsteady flux
Parameteridentification
Empirical equations
Field method in open or partially cased borehole. Most commonly usedmethod. Includes a relatively large volume of porous media in test.Amoozegar and Warrick (1986)
Field method in open, small-diameter borehole (> 5cm). Relatively fastmethod (5 to 60 minutes) requiring small volume of water. Ks, K ( and sorptivity are measured simultaneously. Many boreholes and testsmay be required to fully represent heterogeneities of porous media.
Field method. Test performed in cylinder which is driven into porousmedia. Small volume of material tested; hence, many tests maybeneeded. Fast, simple method requiring little water (-10 L).
Field or lab method. Field method measures vertical duringdrainage. Measurement of moisture content and hydraulic headneeds to be rapid and nondestructive to sample. Commonly usedmethod, reasonably accurate.
Field method. Measures vertical K( ) during wetting portion ofhysteresis loop. Labor and time intensive.
Field method. Larger sample area than for crust method. Usefulonly for relatively high moisture contents.
Results of one field or lab test are used by a numerical approximationmethod to develop over a wide range of Relatively fast method; however, unique solutions are notusually attained.
Each empirical equation has its own application based upon theassumptions of the equation. Relatively fast technique.
Bouwer (1978);Stephens and Neuman (1982 a,b,c);
Reynolds
Bouwer
and Elrick (1986)
(1966)
Bouma et al. (1974):Klute and Dirksen, (1986)
Green, et al. (1986)
Green, Ahuja, and Chong (1986)
Zachmann et al. (198Ia,b, 1982);Kool et al. (1985)
Brooks and Corey (1964);van Genuchten (1980); Mualem(1976)
From Thompson et al, 1989
Copyright® 1989 Electric Power Research Institute. EPRI EN-6637. Techniques to Develop Data for Hydrogeochemical Models. Reprinted withpermission.
residual phase into the gas phase where they could be de-tected.
Given the current understanding of the magnitude of theprocesses controlling the rate of migration of organic con-taminants in the gas phase, it may be more reasonable toreverse the argument. If there is some NAPL in the unsatur-ated zone, VOCs can travel significant distances in the gasphase. Provided that the Henry’s constants for these organiccontaminants are sufficiently small, these volatiles can parti-tion into the infiltrating water and be carried to the subsurfaceto form a shallow contaminant plume. So, the ground-watercontaminant plume results from the soil-gas contaminationrather than from the ground water.
If this describes the interaction between contaminatedsoil gas and contaminated ground water, then the greatest useof shallow soil-gas monitoring surveys is for locating poten-tial residuals of NAPL in the subsurface. The areas with thehighest gas-phase concentrations are most likely to be thoseclosest to any residual product. Thus, such a survey could bean effective guide for determining the optimal locations forsoil-gas extraction wells.
5.3 Analysis of DataThere are several programs used to evaluate flow in the
vadose zone, many of which are discussed in van der Heijde etal. (1988). Because of the nonlinear and hysteretic behavior ofvarious parameters, modeling vadose-zone flow is more diffi-cult than modeling saturated flow. There are additional prob-lems because of the atmospheric boundary conditionsassociated with seepage faces, infiltration, and evapotranspi-ration. Because of the research associated with pesticides,several programs that analyze the vadose zone are availablethrough the Center for Exposure Assessment Modeling inAthens, Georgia. Other vadose-zone programs are availablefrom the Robert S. Kerr Environmental Research Laboratoryin Ada, Oklahoma, and the International Ground Water Mod-eling Center, Holcomb Research Institute, Butler University,Indianapolis, Indiana.
5.4 Remedial ActionsWhen the vadose zone is shallow, excavation as a reme-
dial action is commonly considered. (An example of excava-tion with fixation is given in Section 5.5.). For volatilechemicals near or above the water table, vacuum extinction isanother technique that can remove contaminants from theresidual phase. During vacuum extraction, air is pulled through
66
Figure 5-4. Relative permeability curves for various soil types.
soils contaminated with VOCs. The resulting vapors movethrough the soil and are collected at extraction wells. Simpletechniques that have been developed to control subsurfacehydrocarbon vapors are discussed in O’Connor et al, (1984),Dunlap (1984), and Marley and Hoag (1984). In general, twoprincipal types of vapor management systems are available.The frost type, a positive differential pressure system, inducesvapor flow away from the control points, while the secondtype, a negative differential pressure system, induces vaporflow toward the control points. The vapor management meth-ods may be either passive or active. Passive methods usenaturally occurring differences in vapor pressures to inducethe required flow regime. Active methods require the artificialgeneration of differential vapor pressures to accomplish thesame flow pattern. Practical experience demonstrates thatactive generation of negative differential vapor pressures typi-cally provides the most favorable field results.
The air flow generates advective vapor fluxes that changethe vapor-liquid equilibrium, inducing volatilization of con-taminants. This method is advantageous because it is imple-mented in place, and, therefore, causes minimum disruption.This is especially important at active facilities or sites whereinvestigations are hindered by physical obstacles. Vacuum
extraction laboratory studies are descriked in Marley andHoag (1984), Thornton and Wootan (1982), and Texas Re-search Institute (1984). Crow et al. (1985, 1987) discusses afield-scale experiment. Agrelot et al. (1985), Regalbuto et al.(1988), Connor (1988), and Hutzler et al. (1989) show appli-cations to hazardous waste sites.
Vacuum extraction can effectively remove chemicals fromthe vadose zone. According to Hutzler et al. (1989), mostchemicals successfully extracted have a low molecular weightand high volatility. Most of the compounds have values ofHenry’s Law constants water than 0.01. If the water table islowered, vacuum extraction also can be used to remove re-sidual NAPL from below the original water table elevation.For example, ground-water pumping and vacuum extractionare being used together to clean up DNAPL contamination atthe Tyson’s Superfund site (Wassersug, 1989). Vacuum ex-traction also can increase natural biodegradation processes byintroducing additional oxygen into the subsurface. Finally,vacuum extraction generally is used in conjunction with otherremedial methods.
67
Table 5-8. Characteristics of Contaminants in ReIatlon to Soil Gas Surveying
Group/Contaminants Applicability of Soil-Gas Survey Techniques
Group A: Halogenated Methanes, Ethanes, and Ethenes
chloroform, vinyl chloride, Detectable in soil gas over a wide range of environmental conditions. Dense non-aqueous phasecarbon tetrachloride, liquid (DNAPL), wi// sink in aquifer if present as pure liquid.trichlorofluoromethane,TCA, EDB, TCE
Group B: Halogenated Propanes, Propenes and Benzenes
chlorobenzene, Limited value; detectable by soil-gas techniques oniy where probes can sample near contaminated soil ortrichlorobenzene, ground water. DNAPL.1,2-dichloropropane
Group C: Halogenated Polycyclic Aromatics
aldrin, DDT Do not partition into the gas phase adequate!v to be detected in soil gas under normal circumstances.chlordane, heptachlor, DNAPL.PCBs
Group D: Cl - C8 Petroleum Hydrocarbons
benzene, toluene, Most predictably detected in shallow aquifers or leaking underground storage tanks where probes can bexylene isomers, methane, driven near the source of contamination. Light nonaqueous phase liquids (LNAPLs), float as thin film onethane, cyclohexane, the water table. Can act as a solvent for DNAPLs, keeping them nearer the ground surface.gasoline, JP-4
Group E: Cg - C12 Petroleum Hydrocarbons
trimelhylbenzene, Limited value; detectable by soil gas techniques only where probes can sample near contaminatednaphthalene, decane, soil or ground water. DNAPL.diesel and jet A fuels
Group F: Polycylic Aromatic Hydrocarbons
anthracene, benzopyrene, Do not partition adequately into the gas phase to be detected in soil gas under normal circumstances.fluoranthene, chrysene, DNAPL.motor oils, coal tars
Group G: Low Molecular Weight Oxygenated Compounds
acetone, ethanol, LNAPLs, but dissolve readily in ground water. May be detected in soil gas if they result from aformaldehyde, leak or spill in relatively dry soil.methylethylketone
Source: Adapted from Marrin (1987)
5.5 Example-Pepper’s Steel SiteFixation technology is demonstrated in a case study of the
30-acre Pepper’s Steel cleanup site, located near Miami andMedley, Florida, where the Miami Canal borders the site(Figure 5-5). Ground water in the Biscayne aquifer is about 5to 6 ft below land surface. Soils above the aquifer werecontaminated as a result of prior business operations at thesite, and polychlorinated biphenyls (PCBs) and heavy metals(lead, arsenic) were found in concentrations significant towarrant action.
The two primary goals of site cleanup were
Collect and dispose of oils containing PCBs that areuncovered during site excavations.
Treat or dispose of soils that are contaminated withPCBs and heavy metals.
After reviewing several remedial options, investigators se-lected solidification/stabilization. In accordance with regula-tions, PCB-contaminated oils were removed and disposed ofat an approved facility off site. All the contaminated soilswere solidified on site with a proportioned mix of fly ash andcement. Solidification changes the physical characteristics ofthe waste and decreases the surface area of pollutants avail-able for leaching. Through stabilization, the wastes becomeless water soluble and less toxic. The PCBs are trapped in thecement mixture and the heavy metals (arsenic and lead)become insoluble metal silicates.
The amount of soil excavated for fixation was minimizedby using kriging on soil chemistry data. The kriged resultsindicated zones of contamination as well as a measure of theerror of estimation. Some details of the cleanup include (U.S.EPA, 1987):
. Approximately 60,000 cubic yards of contaminatedsoils were excavated.
68
Figure 5-5. Location of Pepper’s Steel and Alloy site monitoring wells.
All free oil uncovered during excavation was col-lected and sent for treatment or disposal.
Soils contaminated with PCBs and heavy metalswere stabilized solidified with a cement mixture.
Solidified materials were placed back on the Pepper’sSteel site and covered with 12 in. of crushed lime-stone.
Surface water was controlled by grading the site andplacing drains around the solidified material.
Ground water is monitored annually.
5.6 ReferencesAgrelot, J.C., J.J. Malot, and M.J. Visser. 1985. Vacuum:
Defense System for Ground Water VOC Contamination.In: Fifth National Symposium and Exposition on AquiferRestoration and Ground Water Monitoring, National WaterWell Association, Dublin, OH, pp. 485-494.
Amoozegar, A. and A.W. Warrick. 1986. Hydraulic Conduc-tivity of Saturated Soils: Field Methods. In: Methods ofSoil Analysis, Part 1, 2nd ed., A. Klute (ed.), Agronomy
Monograph No. 9, American Society of Agronomy, Madi-son, WI, pp. 735-770.
Bouma, J., F.G. Baker, and P.L.M. Veneman. 1974. Measure-ment of Water Movement in Soil Pedons above the WaterTable. University of Wisconsin-Extension, Geologicaland Natural History Survey, Information Circular No. 27.
Bouwer, H. 1966. Rapid Field Measurement Air Entry Valueand Hydraulic Conductivity of Soil as Significant Param-eters in Flow System Analysis. Water Resources Re-search 2(4):729-738.
Bouwer, H. 1978. Ground-Water Hydrology. McGraw-Hill,New York.
Bouwer, H. 1986. Intake Rate: Cylinder Infiltrometer. In:Methods of Soil Analysis, Part 1, 2nd ed., A. Klute (ed.),Agronomy Monograph No. 9, American Society ofAgronomy, Madison, WI, pp. 825-844.
Brooks, R.H. and A.T. Corey. 1964. Hydraulic Properties ofPorous Media. Hydrology Paper No. 3. Colorado StateUniversity, Fort Collins, CO.
Conner, J.R. 1988. Case Study of Soil Venting. PollutionEngineering 20(7):75-78.
69
Crow, W.L., E.P. Anderson, and E.M. Minugh. 1985. Subsur-face Venting of Hydrocarbon Vapors from an Under-ground Aquifer. In: Proc. NWWA/API Conf. on PetroleumHydrocarbons and Organic Chemicals in Ground Wa-ter-prevention, Detection and Restoration, NationalWater Well Association, Dublin, OH, pp. 536-554.
Crow, W.L., E.R. Anderson, and E. Minugh. 1987. Subsur-face Venting of Hydrocarbons Emanating from Hydro-carbon Product on Groundwater. Ground WaterMonitoring Review 7(1}51-57.
Davis, S.N. and R.J.M. DeWiest. 1966. Hydrogeology. JohnWiley & Sons, New York.
Dunlap, L.E. 1984. Abatement of Hydrocarbon Vapors inBuildings. In: Proc. NWWA/API Conf. on PetroleumHydrocarbons and Organic Chemicals in Ground Wa-ter—prevention, Detection and Restoration, NationalWater Well Association, Dublin, OH, pp. 504-518.
Dunne, T. and L.B. Leopold. 1978. Water in EnvironmentalPlanning. W.H. Freeman, San Francisco,818 pp.
Finkelstein, P.L., D.A. Mozzarella, T.A. Lockhart, WJ. King,and J.H. White. 1989. Quality Assurance Handbook forAir Pollution Measurement Systems, IV: MeteorologicalMeasurements, revised. EPA/600/4-90-O03.
Gardner, W.H. 1986. Water Content. In: Methods of SoilAnalysis, Part 1, 2nd ed., A. Klute (ed.), AgronomyMonograph No. 9, American Society of Agronomy, Madi-son, WI, pp. 493-544.
Green, R.E., L.R. Ahuja, and S.K. Chong. 1986. HydraulicConductivity, Diffusivity, and Sorptivity of UnsaturatedSoils: Field Methods. In: Methods of Soil Analysis, Part1, 2nd ed., A. Klute (ed.), Agronomy Monograph No. 9,American Society of Agronomy, Madison, WI, pp. 771-798.
Green, W.H. and C.A. Ampt. 1911. Studies on Soil Physics, I:Flow of Air and Water through Soils. J. AgriculturalScience 4:1-24.
Hillel, D. 1980. Fundamentals of Soil Physics. AcademicPress, New York, 413 pp.
Hult, M.F. and R.R. Grabbe. 1985. Permanent Gases andHydrocarbon Vapors in the Unsaturated Zone. In: Pro-ceedings, U.S. Geological Survey Second Toxic WasteTechnical Meeting, Cape Cod, MA, October 1985.
Hutzler, N.F., B.E. Murphy, and J.S. Gierke. 1989. State ofTechnology Review Soil Vapor Extraction Systems. U.S.EPA Cooperative Agreement CR-8143 19-01-1 (NTISPB89-195184), 36 pp.
Klute, A. 1986. Water Retention: Laboratory Methods. In:Methods of Soil Analysis, Part 1, 2nd ed., A. Klute (ed.),Agronomy Monograph No. 9, American Society ofAgronomy, Madison, WI, pp. 635-662.
Klute, A. and C. Dirksen. 1986. Hydraulic Conductivity andDiffusivity: Laboratory Methods. In: Methods of SoilAnalysis, Part 1, 2nd ed., A. Klute (ed.), AgronomyMonograph No. 9, American Society of Agronomy, Madi-son, WI, pp. 687-734.
Kool, J.B., J.C. Parker, and M.T. van Genuchten. 1985. Deter-mining Soil Hydraulic Properties from One-Step OutflowExperiments by Parameter Estimation: I. Theory andNumerical Studies. Soil Sci. Soc. Am. J. 49:1348-1354.
Marley, M.C. and G.E. Hoag. 1984. Induced Soil Venting forRecovery/ Restoration of Gasoline Hydrocarbons in theVadose Zone. In: Proc. NWWA/API Conf. on PetroleumHydrocarbons and Organic Chemicals in Ground Wa-ter—prevention, Detection and Restoration, NationalWater Well Association, Dublin, OH, pp. 473-503.
Marrin, D.L. 1987. Soil Gas Sampling Strategies: Deep vs.Shallow Aquifers. In: Proc. 1st Nat. Outdoor ActionConf. on Aquifer Restoration, Ground Water Monitoringand Geophysical Methods, National Water Well Associa-tion, Dublin, OH, pp. 437-454.
Marrin, D.L. and H.B. Kerfoot. 1988. Soil-gas SurveyingTechniques. Environ. Sci. Technol. 22(7):740-745.
Marrin, D.L. and G.M. Thompson. 1987. Gaseous Behaviorof TCE Overlying a Contaminated Aquifer. Groundwater25:21-27.
Mualem, Y. 1976. A New Model for Predicting the HydraulicConductivity of Unsaturated Porous Media. Water Re-sources Research 12:593-622.
Musgrave, G.W. and H.N. Holtan. 1964. Infiltration. In: Hand-book of Applied Hydrology, V.T. Chow (ed.), McGraw-Hill, New York, pp. 12-1 to 12-30.
National Weather Service. 1972. Observing Handbook No. 2.Data Acquisition Division, Office of Meteorological Opemtions, Silver Spring, MD.
O’Connor, M.J., J.G. Agar, and R.D. King. 1984. PracticalExperience in the Management of Hydrocarbon Vaporsin the Subsurface. In: Proc. NWWA/API Conf. on Petro-leum Hydrocarbons and Organic Chemicals in GroundWater-Prevention, Detection and Restoration, NationalWater Well Association, Dublin, OH, pp. 519-533.
Peterson, A.E. and G.D. Bubenzer. 1986. Intake Rate Sprin-kler Infiltrometer. In: Methods of Soil Analysis, Part 1,2nd ed., A. Klute (ed.), Agronomy Monograph No. 9,American Society of Agronomy, Madison, WI, pp. 845-870.
Philip, J.R. 1957. The Theory of Infiltration, I: The InfiltrationEquation and its Solution. J. Soil Science 83:345-357.
Radian Corporation. 1988. FGD Chemistry and AnalyticalMethods Handbook, 2 Chemical and Physical Test Meth-
70
I
ods, Revision 1. EPRI CS-3612. Electric Power ResearchInstitute, Palo Alto, CA. [Originally published in 1984].
Regalbuto, D.P., J.A. Barrera, and J.B. Lisiecki. 1988. In-situRemoval of VOCs by Means of Enhanced Volatilization.In: Proc. NWWA/API Conf. on Petroleum Hydrocarbonsand Organic Chemicals in Ground Water-Prevention,Detection and Restoration, National Water Well Associa-tion, Dublin, OH, pp. 571-590.
Reynolds, W.D. and D.E. Elrick. 1986. A Method for Simul-taneous In Situ Measurement in the Vadose Zone of FieldSaturated Hydraulic Conductivity, Sorptivity and the Con-ductivity-pressure Head Relationship. Ground WaterMonitoring Review 6(4):84-95.
Schmugge, T.J., T.J. Jackson, and H.L. McKim. 1980. Surveyof Methods for Soil Moisture Determination. Water Re-sources Research 16(6):961-979.
Sharma, M.L. 1985. Estimating Evapotranspiration. In: Ad-vances in Irrigation, 3. Academic Press, New York.
Stephens, D.B. and S.P. Neuman. 1982a. Vadose Zone Per-meability Tests: Summary. J. Hydraulics Division ASCE198(HY5):623-639.
Stephens, D.B. and S.P. Neuman. 1982b. Vadose Zone Per-meability: Steady State Results. J. Hydraulics DivisionASCE 198(HY5):640-659.
Stephens, D.B. and S.P. Neuman. 1982c. Vadose Zone Per-meability Unsteady Flow. J. Hydraulics Division ASCE198(HY5):660-677.
Stephens, D.B. and S. Heermann. 1988. Dependence of Anisot-ropy on Saturation in a Stratified Sand. Water ResourcesResearch 24(5):770-778.
Stephens, J.C. and E.H. Stewart. 1964. A Comparison ofProcedures for Computing Evaporation and Evapotrans-piration. Agricultural Research Service, Ft. Lauderdale,FL.
Texas Research Institute. 1984. Forced Venting to RemoveGasoline Vapors from a Large-Scale Model Aquifer.American Petroleum Institute, Washington, DC, 60 pp.
Thompson, C.M., et al. 1989. Techniques to Develop Data forHydrogeochemical Models. EPRI EN-6637. ElectricPower Research Institute, Palo Alto, CA.
Thomthwaite, C.W. and J.R. Mather. 1957. Instructions andTables for Computing Potential Evapotranspiration andWater Balance. Drexel Institute of Technology, Labora-tory of Climatology, X(3).
Thornton, J.S. and W.L. Wootan. 1982. Venting for the Re-moval of Hydrocarbon Vapors from Gasoline Contami-nated Soil. J. Environmental Science and HealthA17(1):31-44.
U.S. Environmental Protection Agency. 1987. Protecting theBiscayne Aquifers: Actions to be Taken at the Pepper’sSteel and Alloy Site. Prepared by CH2M Hill.
U.S. Geological Survey. 1977. National Handbook of Recom-mended Methods for Water Data Acquisition (Chapter8—Evaporation and Transpiration, updated June 1982).USGS Office of Water Data Coordination, Reston, VA.
van Bavel, C.H.M. 1963. Soil Moisture Measurement with theNeutron Method. USDA-ARS, ARS-41-70, U.S. Gov-ernment Printing Office, Washington, DC.
van Genuchten, M.T. 1980. A Closed Form Equation forPredicting the Hydraulic Conductivity of UnsaturatedSoils. Soil Sci. Soc. Am. J. 44:892-898.
van der Heijde, P. K. M., A.I. E1-Kadi, and S.A. Williams.1988. Groundwater Modeling: An Overview and StatusReport. EPA/600/2-89/028 (NTIS PB89-224497). Alsoavailable from International Ground Water ModelingCenter, Butler University, Indianapolis, IN.
Veihmeyer, F.J. 1964. Evapotranspiration. In: Handbook ofApplied Hydrology, V.T. Chow (ed.), McGraw-Hill, NewYork, pp. 11-1 to 11-38.
Wassersug, S.R. 1989. Policy Aspects of Current Practicesand Applications. In: Remediating Groundwater and SoilContamination, Report on a Colloquium, Water Scienceand Technology Board, National Academy Press, Wash-ington, DC.
Wilson, L.G. 1980. Monitoring in the Vadose Zone: A Re-view of Technical Elements and Methods. EPA/600/7-80-134 (NTIS PB81-125817), 168 pp.
Wilson, L.G. 1981. Monitoring in the Vadose Zone, Part I:Storage Changes. Ground Water Monitoring Review1 (3):32-41.
Wilson, L.G. 1982. Monitoring in the Vadose Zone, Part II:Ground Water Monitoring Review 2(4):31-42.
Wilson, L.G. 1983. Monitoring in the Vadose Zone, Part III:Ground Water Monitoring Review 3(4):155-166.
Zachmann, D.W., P.C. DuChateau, and A. Klute. 1981a. TheCalibration of the Richards Flow Equation for a DrainingColumn by Parameter Identification. Soil Sci. Soc. Am. J.45:1012-1015.
Zachmann, D.W., P.C. DuChateau, and A. Klute. 1981b. TheEstimation of Soil Hydraulic properties from Inflow Data.In: Proceedings, Symposium on Rainfall-Runoff Model-ing, V.V. Singh (ed.), Water Resources Publications,Littleton, CO, pp. 173-180.
Zachmarm, D.W., P.C. DuChateau, and A. Klute. 1982. Si-multaneous Approximation of Water Capacity and SoilConductivity by Parameter Identification. Soil Science134:157-163.
71
CharacterizationChapter 6
of Water Movement in Saturated Fractured MediaJames W. Mercer and Charles P. Spalding
Characterizing heterogeneity and anisotropy in the sub-surface is important, especially in fractured or karst media.Fracturing or caverns provide preferential flow paths forground water. Many of the characterization tools and tech-niques discussed for porous media also may be used forfractured media, if care is used to interpret the data. Tech-niques that are particularly helpful in understanding fractured/cavernous media include coring, aquifer tests, tracer tests,geophysical tools, geochemical techniques, and fracture traceanalysis. Most of these techniques are discussed in this chap-ter.
As in the preceding chapters, the discussion begins with areview of concepts. This review is followed by sections onfield techniques, analysis of data, and a case study. Thischapter draws freely upon material contained in a recent EPASuperfund ground water-issue paper on contaminant transportin fractured media (Schmelling and Ross, 1989).
6.1 Review of ConceptsMost fractured bedrock systems consist of rock bounded
by discrete discontinuities composed of fractures, joints, andshear zones, usually occurring in sets with similar geometries(Witherspoon et al., 1987). Figure 6-1 illustrates this type of
Figure 6-1. Conceptualization of discontinuities in a fracturedmedium (from Gale, 1982).
system, referred to as a dual-porosity system. In addition tothe discontinuities shown in the figure, bedding planes alsocan behave as discontinuities. Fractures may be open, min-eral-filled, deformed, or any combination thereof (Nelson,1985).
Open fractures may provide conduits for ground-waterand contaminant movement through a rock mass that is other-wise relatively impermeable. Fractures may be filled eitherpartially or completely by secondary cementing materials.such as quartz or carbonate minerals, which reduce or elimi-nate fracture porosity and permeability. The permeability ofdeformed fractures also may be reduced by gouge, a finelyabraded material produced by the cataclasis of grains incontact across a fault plane during displacement of the rockmasses. Slickensides, striated surfaces formed by frictionalsliding along a fault plane, also are a deformed-fracture fea-ture. Slickensides reduce permeability perpendicular to thefracture plane, but the mismatch of fracture surfaces mayincrease permeability along the fracture plane. Very littledisplacement is necessary to produce gouge or slickensides.Another factor that may reduce permeability is the depositionof a thin layer of low- permeability material called a fractureskin. This skin prevents the free exchange of fluids betweenthe rock matrix and the fracture (Moench, 1984).
The concept of fracturing presented so far is one elementof a more complicated hierarchy of multiple-porosity systems.In soluble bedrock like limestone, dolostone, or evaporates,conduit flow can develop as original fracture systems areenlarged by solution. The important feature of conduit flow,when it is able to develop, is the integration of the drainagenetwork (Quinlan and Ewers, 1985). In many ways, the net-work is analogous to a river system with smaller tributariessupplying water to a succession of larger and larger conduits.As a result of the integration, both the conduit system and theindividual conduits can become large. For example, the karstsystem at Mammoth Cave, Kentucky, has over 330 miles ofconnected passages.
Major factors affecting ground-water flow through frac-tured rock include (1) fracture density, (2) orientation, (3)effective aperture width, and (4) the nature of the rock matrix.Fracture density (number of fractures per unit volume of rock)and orientation are important determinants of the degree ofinterconnection of fracture sets, which is a critical featurecontributing to the hydraulic conductivity of a fractured rock
73
system (Witherspoon et al., 1987). Only interconnected frac-tures provide pathways for ground-water flow and contami-nant transport, although the flow network may be a subset ofthe fracture network. Fractures oriented parallel to the hydrau-lic gradient are more likely to provide effective pathways thanfractures oriented perpendicular to the hydraulic gradient.Flow in fractured rock systems can be similar to flow inporous media when (1) the fracture apertures are constant, (2)the fracture orientations are randomly distributed, and (3) thefracture spacing is small relative to the scale of the system(Long et al., 1982).
The cross-sectional area of a fracture will have an impor-tant effect on flow through the fracture. Under certain condi-tions, fracture-flux is generally proportional to the cube of thefracture aperture (distance between rock blocks) when aper-tures exceed 10 microns (Witherspoon et al., 1987). Fractureapertures and, therefore, flow through fractures are highlystress-dependent and generally decrease with depth (Gale,1982).
The nature of the rock matrix affects the movement ofwater and contaminants through fractured rock systems. Meta-morphic and igneous rocks generally have very low primaryporosity and permeability. Fractures may account for most ofthe permeability in such systems, and the movement of waterand contaminants into and out of the rock matrix may beminimal. On the other hand, sedimentary rocks generally havehigher primary porosity and varying permeability. Examplesinclude coarse-grained materials, such as sandstone, whichhave relatively high primary porosity and significant matrixpermeability, and fine-grained materials, such as shale, whichhave high primary porosity and low permeability.
Fractures may enhance the permeability of all types ofmaterials. High porosity allows significant storage of waterand contaminants in the rock matrix. Authigenic clays formedduring the weathering on certain rock-forming minerals maysignificantly reduce the porosity and permeability of the frac-tures and rock matrix. Rates of contaminant migration intoand out of the rock matrix will depend on the permeability ofthe matrix, the presence of low-permeability fracture skins,and the matrix diffusion coefficient of the contaminant (Fig-ure 6-2).
A complete description of a contaminated fractured rocksystem would include data on (1) the dimensions of thesystem; (2) individual fracture length, aperture width, loca-tion, and orientation; (3) the hydraulic head throughout thesystem; (4) the porosity and permeability of the rock matrix;(5) the sources of water and contaminants; (6) the nature andconcentrations of the contaminants throughout the system;and (7) the chemical interactions between the contaminantsand rock matrix. Presently, collection of such detailed infor-mation is neither technically possible nor economically fea-sible at the scale of most contaminated sites.
6.2 Field TechniquesHydrogeologic characterization methods usually are most
successful when used in conjunction with one another. Thesemethods may include coring, aquifer tests, tracer tests, surface
Figure 6-2. Flow through fractures and diffusion of contami-nants from fractures into the rock matrix of a dual-porosity medium (from Anderson, 1984).
and borehole geophysical techniques, and use of boreholeflowmeters, or other tools. Important information may begathered before, during, and after drilling operations.
6.2.1 Fracture Trace AnalysisGround-water flow in bedrock is generally concentrated
in the upper weathered zone of the rock and in fractures atdepth. A well penetrating a zone of subsurface fractures,therefore will yield more water than a well drilled in an areawith relatively few fractures. Such zones are also pathwaysfor contaminant migration. Selecting drill sites by examiningaerial photographs stereoscopically for surficial expressionsof linear zones of subsurface fractures will increase the prob-ability of high yields and locating contaminants. This type ofstudy is known as fracture trace analysis (Ray, 1960; Fetter,1980). Figure 6-3 shows the relationship between fracturetraces and zones of fractures. In general, higher yields can beexpected in topographic low areas because (1) swales andvalleys tend to be cut into less-resistant, more highly fracturedand more-permeable rock; and (2) ground-water flow usuallyconverges in stream valleys.
Although fracture traces, fault planes, and other linea-ments are often identifiable on aerial photographs, they mustbe field-verified to distinguish anthropogenic features such asfences and buried pipelines from geologic features. The orien-tation of all fractures (e.g., outcrops and excavations) identi-fied from aerial photographs and field observations should bemeasured and plotted on maps as well as on rose diagrams(where the frequency of fracture orientation is plotted) toidentify major fracture trends. Such trends are usually relatedto the geologic (tectonic) history of a site. A basic understand-ing of a site’s tectonic history and subsequent fracture orienta-tion allows a better understanding of potentiaI contaminantpathways.
74
Figure 6-3. Relationship between fracture traces end zones of fracture concentration (after Lattman and Parizek, 1964).
6.2.2 CoringCore material obtained during drilling operations can
yield information on the density, location, and orientation offractures and provide samples for physicaI and chemicaltesting. Core samples also may provide information concer-ning fracture roughness and mineral precipitation on fracturesurfaces. Information collected during coring operations withopen hole completions should include (1) the location ofmajor water-bearing fractures, (2) changes in hydraulic headwith depth, and (3) changes in the ground-water geochemis-try, Water loss to a fracture zone, drilling rates, and thepresence of contaminants also are useful active drilling data(this information is discussed in detail in Chapter 4). In certaininstances, cores may be taken diagonally to intercept nearvertical fractures and determine fracture azimuth. While amajor drawback of coring can be the relatively high cost, theinformation obtained often makes this characterization tech-nique cost effective.
6.2.3 Aquifer TestsAquifer tests, including constant rate pumping tests and
slug tests, can provide hydraulic conductivity and informationon anisotropy for fractured formations. These tests also allowthe estimation of average fracture apertures of a medium. Thesame tests commonly used for unconsolidated porous mediacan be used for fractured media, but the test results willgenerally be more difficult to interpret. Barker and Black(1983) note that transmissivity values will always be overesti-
mated by applying standard type curve analysis to fissuredaquifers.
Other more complex tests, such as cross-hole packertests, are particularly applicable to fractured media. For ex-ample, Hsieh and Neuman (1985) and Hsieh et al. (1985)describe a method of determining the three-dimensional hy-draulic conductivity tensor. The method consists of injectingfluid into, or withdrawing fluid out of, selected intervalsisolated by inflatable packers and monitoring the transientresponse in isolated intervals of neighboring wells.
This method is applicable to situations where the princi-pal directions of the hydraulic conductivity tensor are notnecessarily vertical and horizontal. A minimum of six cross-hole tests is required to determine the six independent compnents of the hydraulic conductivity tensor. In practice, scatterin the data is likely to be such that more than six cross-holetests will be required. Hsieh and his coworkers concluded thatfailure to fit data to an ellipsoidal representation indicated thatthe rock under study could not be represented by an equiva-lent, continuous, uniform, anisotropic medium the scale of thetest. Depending on the application to be made, the test may berepeated on a larger scale or the data may be interpreted interms of discrete fractures of the system.
While aquifer tests can provide information on aquiferanisotropy, heterogeneity, and boundary conditions, they donot provide information on the range of fracture apertures or
75
surface roughness. One of the major drawbacks associatedwith long-term aquifer testing is the necessity to store andtreat the large volume of water discharged during the test.
Results of aquifer tests in fractured media often demon-strate S-shaped response curves. Early in the pumping test, thefractures control the yield to the well; therefore, the fractureproperties control the aquifer response. Once the fracturesdrain, there is a transition period followed by a time periodduring which the porous block properties control the aquiferresponse (see Streltsova, 1988).
6.2.4 Tracer TestsTracer tests can provide information on effective poros-
ity, dispersion, and matrix diffusion generally unobtainablefrom other hydrogeologic methods. Tracer tests either can beconducted under natural-gradient or forced-gradient condi-tions. The primary disadvantages of tracer tests are the time,expense, number of necessary sampling points, and difficul-ties associated with data interpretation. However, the impor-tant information provided by tracer tests is difficult to obtainby any other means. Davis et al. (1985) provide an introduc-tion to the use of tracers in ground-water investigations (seealso discussion of this report by Quinlan, 1986, and reply byDavis, 1986). Tracers, most commonly fluorescent dyes, alsoare used to help map karst areas (LaMoreaux et al., 1989;Mull et al., 1988; Quinlan, 1986, 1989).
Graphical geochemical techniques commonly used inporous media may provide valuable information at fracturedrock sites. Hem (1985) and Lloyd and Heathcote (1985)provide overviews of methods used to identify the sources andextent of ground-water mixing. Environmental isotopes, suchas tritium, also are used to interpret pathways and travel times(LaMoreaux et al., 1989).
6.2.5 Geophysical ToolsBoth surface and borehole geophysical methods can be
used to characterize fractured rock systems. Application ofsurface geophysical methods such as direct current electricalresistivity, electromagnetic induction methods, ground-pen-etrating radar, magnetometer surveys, and seismic and remotesensing techniques should be evaluated before a drilling pro-gram is initiated. These techniques may provide insight forlocating potential monitoring wells by identifying the locationof contaminant plumes or the orientation of major fracturesystems. However, the correlation of major surface geophysi-cal features with contaminant transport processes in fracturedmedia has yet to be thoroughly characterized.
Borehole walls are usually less susceptible than cores tofractures induced during drilling operations. Borehole geo-physical techniques can usually provide a more reliable esti-mate of fracture density than can cores. However, as indicatedby Nelson (1985) in a review of down-hole techniques, re-sponses used to detect fractures on well logs are nonuniqueand require detailed knowledge of the tool and the variousrock property effects that could cause fracture-like responses.Borehole geophysical methods include acoustic, electricalresistivity, caliper, gamma, and other high-energy logging
techniques. The acoustic televiewer presents a continuousimage of the acoustic response of the borehole face and candetect fracture apertures as small as one millimeter. Thisoriented tool also allows the determination of fracture orienta-tions. Caliper logs are best suited for determining relativefracture intensity in continuous, competent rock. Advances inelectronic miniaturization have led to the development ofdown-hole cameras, capable of providing in situ viewing offractures in the subsurface (Morahan and Dorrier, 1984).
6.2.6 Borehole FlowmetersFlowmeters have been used for many years in industry.
Only recently, however, has instrumentation been developedthat can accurately measure very low flow rates. Boreholeflowmeters measure the incremental discharge along screenedor open-hole portions of wells during small-scale pumpingtests. The three major types of flowmeters currently beingdeveloped are impeller, heat-pulse, and electromagnetic. Heat-pulse and electromagnetic flowmeters have no moving partsthat may deteriorate over time; they also are more sensitivethan impeller flowmeters (Young and Waldrop, 1989). Thisgreater sensitivity may allow the detection of the verticalmovement of water within the borehole under nonpumpingconditions. Under pumping conditions, fracture zones con-tributing ground water to a borehole may be identified.
6.3 Analysis of DataFlow in fractured media has been modeled using one of
three possible conceptualizations: (1) an equivalent porouscontinuum, (2) a discrete fracture network, and (3) a dual-porosity medium (National Research Council, 1990). The firstof these approaches assumes that the medium is fractured tothe extent that it behaves hydraulically as a porous medium.The actual existence of fractures is reflected in the choice ofvalues for the material coefficients (e.g., hydraulic conductiv-ity, storativity, or relative permeability). Often these param-eters take on values significantly different from those used formodeling a porous medium (Shapiro, 1987). Examples of thisapproach as cited by Shapiro (1987) are presented in Elkins(1953), Elkins and Skov (1960), and Grisak and Cherry(1975).
With the discrete fracture approach, most or all of theground water moves through a network of fractures. Thisapproach assumes that the geometric character of each frac-ture (e.g., position in space, length, width, and aperture) aswell as the pattern of connection among fractures are knownexactly. In the simplest theoretical treatment, the blocks areconsidered to be impermeable. Figure 6-4a is an idealizationof a two-dimensional network of fractures consisting of twodifferent sets. Note how each fracture, represented on thefigure by a line segment, has a definite position in space,length, and aperture. The hydraulic characteristics of thefracture system develop as a consequence of the intersectionof the individual fractures. In three dimensions, the networkcan be described in terms of intersecting planes that could berectangular (Figure 6-4b) or circular (Figure 6-4c). Examplesof the discrete fracture treatment of flow in networks areincluded in Long et al. (1982), Long (1985), Robinson (1984),Schwartz et al. (1983), and Smith and Schwartz (1984).
76
The dual-porosity conceptualization of a fractured me-dium considers the fluid in the fractures and the fluid in theblocks as separate continua. Unlike in the discrete approaches,no account is taken of the specific arrangement of fractureswith respect to each other-there is simply a mixing of fluidsin interacting continua (Shapiro, 1987). In the most generalformulation of the dual-porosity model, the possibility existsfor flow through both the blocks and the fractures, with atransfer function describing the exchange between the twocontinua. Thus, a loss in fluid from the fracture represents again in fluid in the blocks (Shapiro, 1987).
Although modeling tools exist to deal with fracturedmedia, at present, results should be interpreted with caution.Systems are often complex and extraordinarily difficult tocharacterize, especially with the level of effort considerednormal for most site investigations. The state of the art in fieldtesting provides a relatively rudimentary estimate of valuesfor some parameters like hydraulic conductivity, while otherparameters, like storativity, must be established through fit-ting simple theoretical models (usually of the porous mediumtype).
6.4 Remedial ActionsIn principle, the remedial actions discussed for porous
media apply to fractured media. However, the remediation forfractured media is usually more difficult to design and imple-ment. For example, there are two major problems associatedwith pump-and-treat technologies: (1) hydraulic conductivityreduction with stress; and (2) matrix diffusion.
Fractures are difficult to work with because aperturesand, hence, hydraulic conductivity, depend on the stress withinthe medium. A fracture can be opened or closed simply byreducing or increasing the forces applied to it, For example,pumping a well in a fractured medium reduces the porepressure, effectively decreasing the fracture aperture. Gale(1982) describes a number of empirical-theoretical approachesdesigned to model the stress coupling to hydraulic conductiv-ity.
For heterogeneous conditions such as fractured media,advected water will sweep through the higher permeablezones (fractures), removing contamination from those zones.Movement of contaminants out of the less-permeable zones isa slower process than advective transport in the higher perme-ability zones. The contaminants either are slowly exchangedby diffusion with the flow water present in the larger pores ormove at relatively slower velocities in the smaller pores. Arule of thumb is that the longer the site has been contaminatedand the more lenticular (layered) the geologic material, thelonger will be the tailing effect. The water and contaminantsresiding in the more permeable zones are those first mobilizedduring pumping. Thus, pump-and-treat technologies work inheterogeneous media, but cleanup times will be longer andmore difficult to estimate than for similar systems in morehomogeneous media.
6.5 Example-Marion County, FloridaThis example involves site characterization in Marion
County, Florida, at a site located approximately 10 mi west of
Flgure 6-4, Three different oonceptuallzatlons of fracture networks: (a) a two-dlmensional system of line segments (from Shimoand Long, 1987); (b) a three-dimensional system of rectangular fractures (from Smith et al., 1985); and (c) a three.dimensional syetem of “penny-shaped” crecks (from Long, 1986).
77
Figure 6-5. Fracture-trace expreasions based on photo interpretation.
Ocala. The work, performed for the Southwest Florida WaterManagement District (Ward et al., 1989), concerned waterresource assessment of the Floridan aquifer however, manyof the steps and techniques used to characterize the site aresimilar to those that would be used at a hazardous wastefacility overlying fractured media. Some of the work is de-scribed in Giffin and Ward (1989).
The first step of the assessment was to perform a fracture-trace analysis using aerial photographs. Photolinears wereclassified as I, II, or III depending on the strength and continu-ity of their linear patterns on the photo. Class I photolinearshave the strongest, most continuous expression; Class III havethe weakest. Figure 6-5 shows the fmcture-trace map and thelocation of Regional Observation Monitoring Program (ROMP)Well 120.
After field checking the mapped fractures, the next stepwas to confirm them using surface geophysics. The tri-poten-tial method was used (Ogden and Eddy, 1984; Habberjam,1969), and the results of this geophysical survey were used topinpoint two lineaments within a few hundred feet of asite where aquifer testing would be performed.
To help locate monitoring wells for the aquifer testing,numerical modeling was performed using a fracture flowcode. Data typical for that part of Florida were used toestimate the response to pumping. Based on the field workand the modeling, the wells were located as shown in Figure6-6. The locations of some wells were modified due to accessdifficulties; three wells were located to penetrate fracture orsolution channel zones; and one well was sited within thelimestone matrix.
After drilling the wells, the investigators performed bore-hole geophysics tests including caliper, gamma-gamma, andneutron. In general, cavernous zones are located using thecaliper log, whereas shalely zones that are less likely to formcavernous zones are located using the gamma-gamma log.The neutron log is used to indicate porous zones, whichshould correspond to caverns. Unfortunately, the geophysicallogs were not useful in differentiating between areas of solu-tion features (OW1, 0W2, and 0W3) and rock matrix (OW4).
The final step in the characterization of this site was toperform hydraulic testing. Both slug tests and an aquifer testperformed at the site demonstrated an underdamped response(see Figure 6-7). In this type of response, the water level in thewell oscillates due to inertial effects, which are common inhighly permeable aquifers. Vart der Kamp (1976) presents amethod for analyzing underdamped responses to slug tests.Pumping tests were analyzed using classical Theis analysisand anew approach based on early-time deviations (Ward andGiffin, 1989, and Shapiro, 1989).
As a result of site data analysis, dual-porosityconceptualization, thought to be appropriate for this site, didnot need to be observed in the field testing. The site was usedto develop a regional dual-porosity and discrete fracture model,which was then calibrated with transient response at wells andmajor spring discharges. A conceptual composite of the siteand model response (Figure 6-8) demonstrates the dramaticdifference in the site-scale storage as compared to the re-gional-scale matrix response. This difference is evidenced bya four order of magnitude shift in time forming the dual-porosity envelope.
78
* Coordinates of type curve overlay and graph
Figure 6-7. Pump test Interpretation using the deep transducer at monitor well OW1 using Theis method.
79
Figure 6-6. Conceptual composite of aquifer test and dual-porosity model response.
6.6 ReferencesAnderson, M.P. 1984. Movement of Contaminants in Ground
Water Ground Water Transport-Advection and Disper-sion. In: Ground-Water Contamination, National Acad-emy Press, Washington, DC, pp. 37-45.
Barker, J.A. and J.H. Black. 1983. Slug Tests in FissuredAquifers. Water Resources Research 19:1558-1564.
Davis, S.N. 1986. Reply to the Discussion by James F. Quinlanof Ground-Water Tracers. Ground Water 24(3):398-399.
Davis, S.N., D.J. Campbell, H.W. Bentley, and T.J. Flynn.1985. Introduction to Ground-Water Tracers. EPA 600/2-85/022 (NTIS PB86-100591). Also published under thetitle Ground-Water Tracers in NWWA/EPA Series, Na-tional Water Well Association, Dublin, OH.
Elkins, L.F. 1953. Reservoir Performance and Well Spacing,Spraberry Trend Area Field of West Texas. Trans. Ameri-can Institute of Mining Engineers 198:177-196.
Elkins, L.F. and A.M. Skov. 1960. Determination of FractureOrientation from Pressure Interference. Trans. AmericanInstitute of Mining Engineers 219:301-304.
Fetter, Jr., C.W. 1980. Applied Hydrogeology. Charles E.Merrill, Columbus, OH, pp. 406-411.
Gale, J.E. 1982. Assessing the Permeability Characteristics ofFractured Rock. GSA Special Paper 189. GeologicalSociety of America, Boulder, CO, pp. 163-181.
Giffin, D.A. and D.S. Ward. 1989. Analysis of Early-TimeOscillatory Aquifer Response. In: Proc. Conf. on NewField Techniques for Quantifying the Physical and Chemi-cal Properties of Heterogeneous Aquifers (Dallas, TX),National Water Well Association, Dublin, OH, pp. 187-211.
Grisak, G.E. and J.A. Cherry. 1975. Hydrologic Characteris-tics and Responses of Fractured Till and Clay Confining aShallow Aquifer. Canadian Geotechnical Journal 12:23-43.
Habberjam, G.M. 1969. The Location of Spherical CavitiesUsing a Tri-Potential Technique. Geophysics 34(5):780-784.
Hem, J.D. 1985. Study and Interpretation of the ChemicalCharacteristics of Natural Water, 3rd ed. U.S. GeologicalSurvey Water-Supply Paper 2254,263 pp.
Hsieh, P.A. and S.P. Neuman. 1985. Field Determination ofthe Three-Dimensional Hydraulic Conductivity Tensorof Anisotropic Media, 1. Theory. Water Resources Re-search 21:1655-1665.
Hsieh, P.A., S.P. Neuman, G.K. Stiles, and E.S. Simpson.1985. Field Determination of the Three-Dimensional Hy-draulic Conductivity Tensor of Anisotropic Media, 2.
Methodology and Application to Fractured Rocks. WaterResources Research 21:1667-1676.
LaMoreaux, P.E., E. Prohic, J. Zoetl, J.M. Tanner, and B.N.Roche. 1989. Hydrology of Limestone Terranes Anno-tated Bibliography of Carbonate Rocks, Volume 4. Inter-national Association of Hydrogeologists, InternationalContributions to Hydrogeology, Vol. 10, Verlag HeinzHeise GmbH, Hannover, Germany, 267 pp.
Lattman, L. H., and R.R. Parizek. 1964. Relationship betweenFracture Traces and the Occurrence of Ground Water inCarbonate Rocks. J. Hydrology, 2:73-91.
Lloyd, J.W. and J.A. Heathcote. 1985. Natural InorganicHydrochemistry in Relation to Ground Water. ClarendonPress, Oxford, 296 pp.
Long, J. C. S., J.S. Remer, C.R. Wilson, and P.A. Witherspoon.1982. Porous Media Equivalents for Networks of Discon-tinuous Fractures. Water Resources Research 18:645-658.
Long, J.C.S. 1985. Verification and Characterization of Frac-tured Rock at AECL Underground Research Laboratory.BMI/OCRD- 17. Office of Crystalline Repository Devel-opment, Battelle Memorial Institute, 239 pp.
Long, J. C. S., P. Gilmour, and P.A. Witherspoon. 1985. AModel for Steady Fluid Flow in Random Three-Dimen-sional Networks of Disc-Shaped Fractures. Water Re-sources Research 21:1105-1115.
Moench, A.F. 1984. Double-Porosity Models for a FissuredGround Water Reservoir with Fracture Skin. Water Re-sources Research 20831-846.
Morahan, T. and R.C. Dorrier. 1984. The Application ofTelevision Borehole Logging to Ground-Water Monitor-ing Programs. Ground-Water Monitoring Review4(4): 172-175.
Mull, D.S., T.D. Lieberman, J.L. Smoot, and L.H. Woosely,Jr. 1988. Application of Dye-Tracing Techniques forDetermining Solute-Transport Characteristics of GroundWater in Karst Terranes. EPA 904/6-88-001, Region 4,Atlanta, GA.
National Research Council. 1990. Ground-Water Models:Scientific and Regulatory Applications. National Acad-emy Press, Washington, DC, 303 pp.
Nelson, R.A. 1985. Geologic Analysis of Naturally FracturedReservoirs. Contributions in Petroleum Geology and En-gineering, Vol. 1. Gulf Publishing Company, Houston,TX, 320 pp.
Ogden, A. and P.S. Eddy, Jr. 1984. The Use of Tri-PotentialResistivity to Locate Fractures and Caves for High YieldWater Wells. In: NWWA/EPA Conf. on Surface andBorehole Geophysical Methods in Ground Water Investi-
81
gations (San Antonio, TX), National Water Well Asso-ciation, Dublin, OH, pp. 130-149.
Quinlan, J.F. 1986. Discussion of “Ground Water Tracers” byDavis et al. (1985) with Emphasis on Dye Tracing, Espe-cially in Karst Terranes. Ground Water 24(2):253-259and 24(3):396-397 (References).
Quinlan, J.F. 1989. Ground-Water Monitoring in Karst Ter-ranes: Recommended Protocols and Implicit Assump-tions. EPA 600/X-89/050, EMSL, Las Vegas, NV.
Quinlan, J.F. and R.O. Ewers. 1985. Ground Water Flow inLimestone Terrains: Strategy Rationale and Procedurefor Reliable, Efficient Monitoring of Ground Water Qual-ity in Karst Areas. In: Proc. Fifth National Symposiumand Exposition on Aquifer Restoration and Ground Wa-ter Monitoring, National Water Well Association, Dublin,OH, pp. 197-234.
Ray, R.G. 1960. Aerial Photographs in Geologic Interpreta-tion and Mapping. Geological Survey Professional Paper373, 230 pp.
Robinson, P.C. 1984. Connectivity Flow and Transport inNetwork Models of Fractured Media. DP 1072. Theoreti-cal Physics Division, AERE, Harwell, U.K.
Schmelling, S.G. and R.R. Ross. 1989. Contaminant Trans-port in Fractured Media Models for Decision Makers.EPA Superfund Ground Water Issue Paper. EPA/540/4-89/004.
Schwartz, F.W., L. Smith, and A.S. Crowe. 1983. A Stochas-tic Analysis of Macroscopic Dispersion in Fractured Me-dia. Water Resources Research 19:1253-1265.
Shapiro, A.M. 1987. Transport Equations for Fractured Po-rous Media. In: Advances in Transport Phenomena inPorous Media, J. Bear and M.Y. Corapcioglu (eds.),NATO Advanced Study Institutes Series E, Vol. 128,Martinus Nijhoff Publishers, Dordrecht, The Netherlands,pp. 407-471.
Shapiro, A.M. 1989. Interpretation of Oscillatory Water-LevelResponses in Observation Wells During Aquifer Tests in
Fracture Rock. Water Resources Research 25(10>2129-2138.
Shimo, M. and J.C.S. Long. 1987. A Numerical Study ofTransport Parameters in Fracture Networks. In: Flow andTransport through Unsaturated Fractured Rock, D.D.Evans and T.J. Nicholson (eds.), AGU Monograph 42,American Geophysical Union, Washington, DC, pp. 121-131.
Smith, L. and F.W. Schwartz. 1984. An Analysis of FractureGeometry on Mass Transport in Fractured Media. WaterResources Research 20:1241-1252.
Smith, L., C.W. Mase, and F.W. Schwartz. 1985. A StochasticModel for Transport in Networks of Planar Fractures. In:Greco 35 Hydrogeologie, Ministere de la Recherche et laTechnologies Centre Nationale de la RechercheScientifique, Paris.
Streltsova, T.D. 1988. Well Testing in Heterogeneous Forma-tions. John Wiley & Sons, New York, 413 pp.
van der Kamp, G. 1976. Determining Aquifer Transmisivityby Means of Well Response Tests: The UnderdampenedCase. Water Resources Research 12(1):71-77.
Ward, D.S., D.C. Skipp, D.A. Giffin, and M.D. Barcelo. 1989.Dual-Porosity and Discrete Fracture Simulation of GroundWater Flow in West-Central Florida. In: NWWA Confer-ence on Solving Ground Water Problems with Models(Indianapolis, IN), National Water Well Association,Dublin, OH, pp. 385-408.
Witherspoon, P.A., J.C.S. Long, E.L. Majer, and L.R. Myer.1987. A New Seismic Hydraulic Approach to ModelingFlow in Fractured Rocks. In: Proceedings, NWWA/IGWMC Conference on Solving Ground-Water Prob-lems with Models (Denver, CO), National Water WellAssociation, Dublin, OH, pp. 793-826.
Young, S.C. and W.R. Waldrop. 1989. An ElectromagneticBorehole Flowmeter for Measuring Hydraulic Conduc-tivity Variability. In: Proc. Conf. on New Field Tech-niques for Quantifying the Physical and ChemicalProperties of Heterogeneous Aquifers (Dallas, TX), Na-tional Water Well Association, Dublin, OH, pp. 463-474.
82
GeochemicalChapter 7
Characterization of the Subsurface: Basic Analytical and StatisticalConcepts
J. Russell Boulding and Michael J. Barcelona
This chapter presents basic analytical and statistical con-cepts related to the measurement and interpretation of geo-chemical data on the natural and contaminated subsurfaceenvironment. Many expensive geochemical investigations suf-fer because analytical and statistical variability may have beenignored or not fully appreciated in the sample design andcollection phase. Consequently, these analytical and statisticalconcepts are covered here before the chapters on the subsur-face geochemical variability (Chapter 8), and the best meth-ods for sampling the subsurface to characterize this variability(Chapter 9). In the normal sequence of events, laboratoryamlysis and data interpretation come after sample collection.However, because they should be carefully considered in thedesign of geochemical investigations they are presented herefirst.
7.1 Data Measurement and Reliability
7.1.1 Deterministic versus RandomGeochemical Data
Observation or measurement of physical phenomena canbe broadly classified as either deterministic or nondeterministic.Deterministic data can be described by an explicit mathemati-cal relationship. Nondeterministic or random data, must bedescribed in terms of probability statements and statisticalaverages rather than by the use of explicit equations. Figure 7-1 summarizes a classification scheme for deterministic andrandom data from Bendat and Piersol (1986). The classifica-tion of physical data as deterministic or nondeterministic isnot always clear-cut in the real world. In fact, most geochemi-cal data probably fail in a gray area between the two types ofdata. For example, the total dissolved solids in an aquifer is afunction of the chemical composition of the aquifer solids andresidence time of the flowing ground water. Consequently, thedistribution of sample values over space and time will not becompletely random. On the other hand, the factors that deter-mine the precise value of a given sample are sufficientlycomplex and variable that the distribution often cannot bepredicted by an explicit mathematical equation.
The transient, nonperiodic data box in Figure 7-la is aresidual category that includes all data not included in theother boxes, This nonperiodic characteristic of geochemicaldata allows modeling of the distribution of geochemical spe-
cies using thermodynamic principles. Essentially all geo-chemical modeling of the subsurface is done deterministi-tally. The difficulty in accurately modeling the geochemistryof the subsurface can, however, be attributed to large randomelements (see Figure 7-lb). Depending on the geochemicalparameter, and the time frame of sampling, data may bestationary, where characteristics of the population beingsampled do not vary over time, or nonstationary, where therandom process varies with time. Typically, geochemicalsubsurface data invoIving contamination are nonstationary,but are not fully random (i.e., the value of one sample mayshow some correlation with the value of an adjacent sample).This creates special considerations in statistical analysis thatare discussed in Section 7.3. Subsurface physical parameterssuch as hydraulic conductivity, porosity, and soil particle sizedistribution do not typically change with time, at least not ona time scale of human concern. These parameters, however,are not fully random.
7.1.2 Data RepresentativenessIn measuring environmental parameters, there is no “true”
value, but rather a distribution of values. A representative unitor sample is one selected for measurement from a targetpopulation so that it, in combination with other representativesamples, will give an accurate picture of the phenomena beingstudied (Gilbert, 1987). Failure to take samples from locationsand to use methods that yield samples that are “representa-tive” of a site will result in the collection, at some expense, ofanalytical data that may be worthless. Representativenessdetermines whether accurate analysis of the samples will yieldresults that are close to actual conditions. Quality assurancdquality control systems (QA/QC) in the laboratory or fieldmay be useless if even greater emphasis isn’t placed on QA/QC in selecting locations and procedures for sampling.
Thorough site characterization of soils, hydrology, andgeology, as described in the previous chapters, is an essentialprerequisite to geochemical sampling. This information pro-vides the basis for developing sampling strategies that willprovide some assurance that geochemical samples accuratelyreflect what is happening in the field. Sample representative-ness is essentially knowledge-based. For example, samplinglocations selected by someone with a rudimentary under-standing of sampling theory may yield less accurate results
83
Figure 7-1. Classifications of (a) deterministic and (b) randomdata (from Bendat and Piersol, 1986).
than locations chosen by an individual thoroughly groundedin this theory. At the same time, sampling locations selectedwithout careful site characterization will yield less representa-tive samples than locations selected with thorough site charac-terization even with equally sophisticated application ofsampling theory. See Section 9.1 for general considerations indesigning sampling plans.
In contamination investigations, obtaining samples thatcan be considered representative for assessing one or moreparticular kinds of environmental exposure is a primary objec-tive. This requires selecting not only the right place, but theright type of sample (see discussions of analyte selection inSections 9.2.1 and 9.3.1).
7.1.3 Measurement Bias, Precision, andAccuracy
A measured value that is close to the estimate of the trueaverage value is an unbiased or accurate value. This averageor mean can only be estimated by a number of repeat determi-nations. Biased measurements will consistently under- oroverestimate the true values in sampled population units.Precision is a measure of how closely individual measure-ments agree and is influenced principally by random measure-ment uncertainties. Both bias and precision influence accuracyas illustrated in Figure 7-2. The center of each target in the
figure represents that true value. Both low bias and highprecision are required for high accuracy.
Accuracy is largely technologically based. In other words,accuracy can be improved by better drilling and monitoringwell installation procedures and better sampling devices andprocedures. Pennino (1988) has suggested that “there is nosuch thing as a representative ground water sample” becauseof geochemical biases inherent in well installation, purging,and sample collection. However, a good understanding ofboth potential sources of error (see next section) and the wayalternative sampling methods may bias results (see Section9.3) minimizes sample disturbance. The final evaluation ofthe results should be done with full consideration of theunavoidable disturbances involved in subsurface investiga-tions.
7.1.4 Sources of ErrorRandom error results from slight differences in the execu-
tion of the same sampling procedure. Systematic error resultsfrom procedures that alter the properties of the sample. Ran-dom error is unavoidable, but must be evaluated to determineits effect on accuracy. For example, Figure 7-2b shows datawith no systematic bias, but accuracy is low because randomerror is high. Systematic errors can be minimized by carefulselection and conduct of sampling techniques.
Figure 7-3 illustrates five possible sources of error inground-water sampling: (1) site selection, (2) sampling,
Figure 7-2. Shots on a target analogy for illustrating influenceof bias and precision on accuracy (after Jessen,1 9 7 8 ) .(a): high bias + low precision= low accuracy; (b):low bias + low precision = low accuracy; (c): highbias + high precision = low accuracy; (d) low bias+ high precision = high accuracy.
84
Figure 7-3. Sources of error involved in ground-water monitoring programs contributing to total variance (from Barcelona et al.,1983).
(3) measurement methods, (4) reference samples for calibra-tion, and (5) data handling. Both random and systematicerrors may be involved in each stage. Errors at each stage arecumulative, but are not of equal significance or magnitude.Total variance in geochemical data results from the combina-tion of natural geochemical variability and the cumulativeerror. The percentage of variance attributable to natural vari-ability may often be greater than either field or laboratoryerror. Natural variance cannot be reduced; however, varianceresulting from field and laboratory error can be reduced sothat the actual variance closely approximates the natural vari-ance.
Table 7-1 shows estimates of the relative contribution ofnatural variability, field error, and laboratory error to totalvariance at three sites of ground-water investigations. Formost chemical constituents, at the three sites, natural variabil-ity accounted for more than 90 percent of the variance. Formost inorganic constituents where field and laboratory errorcould be estimated, field error contributed a larger percentageof total variance. Table 7-1 also shows that organic contami-nant indicators (TOC and TOX) showed typically much higherpercentages of variance due to field and laboratory error thandid the inorganic indicators. Both field sampling and labora-tory analyses were subject to strict QA/QC procedures at thesites shown in Table 7-1, so variance due to field and labora-tory error during routine ground-water investigations willcommonly be greater than shown in the table.
Field Error. Figure 7-4 identifies specific possible sourcesof error at various steps in ground-water sampling. The largestsources of error are unrepresentative sample locations (hencethe importance of hydrogeologic site characterization prior togeochemical sampling design) and disturbances caused bydrilling and well construction. Sample collection is the nextlargest source of error. Major sources of systematic samplingerror include (1) well construction and screen design prevent-ing representative samples, and (2) improper purging. All ofthese large sources of systematic error are related to thehydrology of the site over which there is often little QA/QC.
Table 7-2 lists potential contributions of sampling meth-ods and materials to error in ground-water chemical results.
85
This table shows that well purging procedures can result inlarge variations in pH, TOC, Fe(II), and VOCs (also seeTable 9-11 for variations in other constituents). Table 7-2shows well casing to be the next largest source of error,followed by sampling mechanisms and grouting/sealing.Poorly grouted or cemented wells can greatly alter the pH ofwater (as much as pH 12). Sampling tubing can result inerrors in VOC measurement. Sections 9.3.3 (Purging) and9.3.4 (Well Construction and Sampling Devices) discussselection criteria for minimizing error from these sources.
Other possible sources of systematic error in samplinginclude (1) changing sampling procedures, (2) changingsampling personnel without a strictly defined sampling pro-tocol, and (3) failure to document unavoidable deviationsfrom the sampling protocols, such as no water in the well.Another source of water quality error is mixing from mul-tiple aquifers. Mixing is most common with public watersupply wells that penetrate several hydrological unconnectedaquifers. Improper sealing of ground-water monitor wellsalso may bias results by mixing water from distinct subsur-face formations.
Analytical Error. Figure 7-5 identifies possible sourcesof error during water sample analysis. Analysis, includingmeasurement methods and reference samples, is typicallysubject to the most stringent QA/QC procedures, and conse-quently analytical errors tend to be relatively minor compo-nents of total error (see Table 7-l). Failure to analyze blanks,standards, and samples by exactly the same procedures mayresult in either a biased blank correction or a biased calibra-tion (Kirchmer, 1983). Porter (1986) examined in detail thesources of random analytical error for measurement near thelimit of detection and how to incorporate this observationerror into data analysis procedures. Sources of analyticalerror are discussed further in Section 7.3.
Einarson and Pei (1988) and Rice et al, (1988), inseparate studies of laboratory performance, concluded thatthe reliability of laboratory analyses should not be taken forgranted. Both studies also concluded that the cost of analysisdid not necessarily correlate with analytical accuracy. Themost expensive of the 10 laboratories evaluated by Einarson
Table 7-1. Percentage of Variance Attributable to Laboratory Error, Field Error, and Natural Variability by Chemical and Site
TOC 15.4 84.6 29.9 70.1TOX
40.6 59.50.0 100.0 12.5 87.5 24.6 75.4
a NA indicated that the number of observations on which the estimated variance was based was less than 5, or the estimated variance wasnegative.
b True field spiked standards not available for these costituents demanding combined estimates of laboratory and field variability.
Source: Barcelona et al. 1989
Figure 7-4. Steps in ground-water sampling and sources of error (from Barcelona et al., 1985).
86
Table 7-2. Potential Contributions of Sampling Methods and Materials to Error” in Ground-Water Chemicail Results
a Bias values exceeding >± 100% denoted as gross errors ( + or -): other values expressed as percent of reported mean.b No data available on the type and extent of error for this parameter.c Concentration range 0.5-15 µg/L (from Barcelona and Helfrich, 1984).d Concentration range 80-8000 µg/L (from Barcelona et al., 1984; Ho, 1983).
1 Barcelona and Helfrich (1984)2 Barcelona et al. (1983)3 Barcelona et al. (1984)4 Barcelona et al. (1988)5 Gibbet al. (1981)6 HO (1983)7 Schuller et al. (1981)
Source: Adapted from Barcelona et al. (1988)
and Pei (1988) tied for the worst ranking, while the four leastexpensive laboratories included the top ranked and otherbottom ranked laboratory. Both studies describe criteria andprocedures for choosing laboratories that will provide goodanalytical results. Section 7.3 discusses analytical and QA/QCconcepts further.
Data Handling Error. There is probably no large body ofscientific records free from human or machine errors. Faultyrecording of observations in field or laboratory notebooks orincorrect coding for computer analysis are examples of datahandling errors. Misrecorded values that are much larger orsmaller than the range of the actual population are calledoutliers and may distort the results of statistical analysis.Statistical techniques are available for analyzing such datasets (Gilbert, 1987), but prevention of data handling error isalways better than a cure. Censoring of analytical measure-ments below the limit of detection (see Section 7.4.1) isanother serious error introduced by data handling.
Webster (1977) suggests some of the following methodsto reduce data handling errors: (1) write neatly, formingcharacters well; (2) distinguish ambiguous digits and lettersby a firm convention; (3) restrict the digit O to mean zero anduse other notations for “missing” or “inapplicable”; (4) elimi-nate or minimize transcription of field notes (5) record dataon forms designed for the purpose of the investigation withclear headings and ample space; and (6) double-check anytranscribed data against the original
7.2 Analytical and QA/QC ConceptsQuality assurance and quality control are accomplished
by (1) selecting the best methods for the program purpose, (2)clearly defining protocols or procedures to be followed, and(3) carefully documenting adherence or departures from theprotocols. Figure 7-6 shows the relationship of program pur-pose and protocols to the scientific method. Both field sam-pling and laboratory analyses require protocols for good QA/QC. Campbell and Mabey (1985) have summarized key ele-
87
Figure 7-5. Steps in water sample analysis and sources of error (from Barcelona et al., 1985).
Figure 7-6. Relationship of program purpose and protocols tothe scientific method (from Barcelona, 1988).
ments of data evaluation systems applicable to both field andlaboratory measurements. Provost and Elder (1985) have pro-vided guidance for choosing cost-effective QA/QC programsfor chemical laboratories. Evans (1986) reviews data qualityobjectives for remedial site investigations, and Starks andFlatman (1991) discuss the use of industrial quality controlmethods as a model for evaluating RCRA ground-water moni-toring decision procedures.
7.2.1 Instrumentation and Analytical MethodsA bewildering array of methods are available for analyz-
ing geochemical constituents. Table 7-3 lists the major signalsand analytical methods based on signal measurement. Mostmethods used for geochemical analysis involve either emis-sion or adsorption of radiation. The fine points of instrumenta-tion and analysis are the province of the analytical chemist,but the field scientist can benefit from a general understand-ing. Skoog (1985) and Willard et al. (1988) are two goodgeneral references on this topic. Analytical techniques forspecific constituents of geochemical interest may be specifiedby regulation or, if not so specified, determined by the instru-mentation that is most readily available. Table 7-4 lists sevenmajor sources of information describing analytical techniquesfor specific chemical constituents.
7.2.2 Limit of DetectionGround-water detection monitoring commonly involves
measurement of contaminants that are either at or below thedetection limit of analytical procedures. The statistical con-cept of detection limit includes accurately reporting and ana-lvzing data including measurement near or below the detectionlimit (McNichols and Davis, 1988).
88
Table 7-3. Major Analytical Signals and Methods
Signal Analytical Methods Based onMeasurement of Signal
Emission of
radiation
Absorption ofradiation
Scattering ofradiation
Refraction ofradiation
Rotation ofradiation
Electricalpotential
Electricalcurrent
Mass-to-chargeratio
Rate of reaction
Thermalproperties
Mass
Volume
Emission spectroscopy (X-ray, UV, visibleelectron auger); fluorescence andphosphorescence spectroscopy (X-ray, UV,visible); radiochemistry
Spectrophotometry (X-ray, UV, visible, IR);photoacoustic spectroscopy; nuclearmagnetic resonance and electron spinresonance spectroscopy
Turbidimetry; nephelometry; RamanSpactroscopy
Refractometry; interferometry
Polarimetry; optical rotatory dispersion;circular dichroism
Potentiometry; chronopotentiometry
Polarography; amperometty; coulometty
Mass Spectrometry
Kinetic methods
Thermal conductivity and enthalpy methods
Gravimetric analysis
Volumetric analysisa Source: Skoog (1985)
Figure 7-7 and Table 7-5 illustrate the definitions of limitof detection and regions of analyte measurement recom-mended by the Subcommittee on Environmental AnalyticalChemistry of the American Chemical Society’s (ACS) Com-mittee on Environmental Improvement (1980). The zero analytesignal for measuring the limit of detection comes from thefield blank (see Section 7.2.3). If the actual field blank mea-surement gives a positive signal, this means that analyticalmeasurements on other samples with a lower signal will berecorded as a negative concentration. For example, a lowconcentration standard (typically 1 part per billion (ppb) fororganic constituents) is made in the laboratory for the con-taminant of interest. The standard deviation for analyticalmeasurement of the 1 ppb standard is commonly plus orminus 100 percent or 1 µg/L. The detection limit for acontaminated sample is defined as three standard deviations(3 µg/L) above the mean for the standard, or six standarddeviations above the zero point defined by the field blank (seeFigure 7-7). The limit of detection should be defined everyday of analysis. The detection limit is probably the mostimportant kind of laboratory quality assurance data and shouldbe reported with the analytical results for each constituent.
Table 7-5 lists the regions of analyte measurement. Fol-lowing the above example, signals below three standard de-viations are considered below the limit of detection. The
region of detection is between 3 and 10 standard deviations (5standard deviations by some rules) and is where the constitu-ent can be said to be present but the precise concentrationcannot be stated with certainty. Analyte signals above thelimit of quantification (plus 10 standard deviations) can beinterpreted quantitatively.
The above-described definition reaffirms the model forlimit of detection calculations adopted by the InternationalUnion of Pure and Applied Chemistry (IUPAC) in 1975(IUPAC, 1978). However, considerable confusion still sur-rounds the definition of the limit of detection. This is because(1) acceptance of the above definition by the general analyti-cal community has been slow, and (2) different statisticalapproaches to calculating limits of detection for constituentscan easily vary by an order of magnitude (Long andWinefordner, 1983). This is particularly true for chemicalconstituents at the ppb level.
A major problem with failure to understand the statisticalnature of the limit of detection is negative censoring of data.Negative censoring involves reporting analyte concentrationsthat are below the limit of detection as zero, “less than”values, or “not detected.” Since 1983 the American Societyfor Testing and Materials (ASTM) has recommended that datashould not be routinely censored by laboratories (ASTM,1983). Nevertheless, censoring of water quality analyticaldata remains a problem (Porter et al., 1988). Section 7.4.1examines this issue further.
Laboratories should be asked to provide uncensored dataon all water samples with measurements near or below thelimit of detection. Measurement data should not be discardedunless the lack of statistical control in the measurement pro-cess is clearly demonstrated. The general public, and even theuninformed scientist, may find the concept of a negativeconcentration difficult to understand, so it is prudent to reportless than zero values as “trace.” Remediation decisions, how-ever, should be based on concentrations at or above the limitof quantification, not the limit of detection.
The limit of detection is both a site- (as a result of thefield blank) and instrument/operator-specific value. Conse-quently, the precision and accuracy for low standards must bereported on the analytical report forms. The instrumentmanufacturer’s definition of detection is based normally oncarefully controlled conditions (e.g., distilled water solutions)that may not be achievable in routine analyses of complexsamples. Consequently, actual limits of detection in contami-nated ground water are often higher.
7.2.3 Types of SamplesField scientists tend to consider QA/QC requirements and
procedures to be primarily the responsibility of the laboratory.However, QA/QC procedures are equally, if not more impor-tant in the field. Chapter 8 examines methods to minimizeerror in selecting sample location and collecting samples.Field personnel also should be familiar with the differenttypes of samples that may be taken, and their importance forinterpreting the analytical results.
89
Table 7-4. Major Compilations of Analytical Procedures for Constituents of Geochemical interest
Reference Description
American Public HealthAssociation (1990)
ASTM, annual
Fresenius et al. (1988)
Klute (1986), Page et al.(1982)
Kopp and McKee (1983)
Longbottom andLichtenberg (1982)
Mueller et al. (1991)
Noblett and Burke (1990)Radian Corporation(1988)
Rainwater andThatcher (1960)
Smith (1991)
Thompson et al. (1989)
Comprehensive compilation of analytical methods for measurement of metals, inorganic nonmetallic, andorganic constituents in water samples.
Published annually by the American Society for Testing and Materials, Water and Environmental TechnologyVolumes 11.01and 11.02 cover analytic methods for water.
A guide to physico-chemical, chemicaland microbiological analysis of water and qualify assurance procedures.
Part 1 (Klute, 1986) contains 50 chapters covering a range of physical and mineralogical methods and Part 2(Page et al., 1982) contains 54 chapters covering methods for analyzing chemical and microbiologicalproperties of soils.
This third edition contains the chemical analytical procedures used in U.S. EPA laboratories for examiningground and surface water, domestic and industrial waste effluents and treatment process samples.
Describes tests for 15 groups of organic chemicals and includes an appendix defining procedures fordetermining the detection limit of an analytic method. The test procedures in this manual are cited inTable IC (organic chemical parameters) and ID (pesticide parameters) in 40 CFR 136.3(a).
Compilation of summary information on more than 150 EPA-approved, and a total of 650, sampling andanalysis methods for industrial chemicals, pesticides, elements, and water quality parameters. Associateddata base is available on diskette.
Handbook on flue gas desulfurization (FGD) chemistry and analytical methods. Volume 1 (Noblett andBurke, 1980) covers sampling, measurement, laboratory, and process performance guidelines. Volume 2(Radian Corporation, 1988) presents 54 physical-testing and chemical-analysis methods for FGD reagents,slurries, and solids.
Describes types of methods, choice of analytical methods for water samples, and specific analyticalprocedures for over 40 inorganic water parameters.
Edited volume with 14 chapters on instrumental techniques for soil analysis.
Contains summary description of methods for elemental analysis, analysis of anionic species, inorganic andorganic carbon, redox sensitive species and other chemical parameters along with recommendations formethods best suited for obtaining data for hydrochemical modeling.
U.S. EPA (1988)
U.S. Geological SurveyTechniques of Water-Resource Investga-tions
Westerman (1990)
Guide for selection of instrumental methods for field screening of inorganic and organic contaminants. Covers26 specific field screening methods. Also available as a computerized information retrieval system.
USGS’s TWI series includes manuals describing procedures for planning and conducting specialized work inwater-resources investigations. Wood (1976) covers field analysis of unstable constituents; Skougstad et al.(1979) cover methods for analyzing inorganic constituents in water and fiuvial sediment; Barnett and Mallory(1971) describe determination of minor e!ements in water by emission spectroscopy: Wershaw et al (1987)cover methods for determination of organic substances in water and fiuvial sediments(revision of Goerlitz et al., 1972).
Edited volume on methods for analysis of soil and plants focussing on use for assessing nutritionalrequirements of crops, efficient fertilizer use, saline-sodic conditions, and toxicity of metals.
A field blank is a sample of distilled or deionized watertaken from the laboratory out into the field, poured into asampling vial at the site, closed, and returned as if it were asample. The level of contamination of the field blank is thezero analyte signal for determining the limit of detection,
A rinse or cleaning blank is a sample of the final rinse ofa sampling mechanism before it is put in a new well. This typeof sample is used to evaluate whether a sample may have beencontaminated from material taken in the previous sample.
Field samples are those samples that are taken in the fieldas “representative” of conditions at the site and analyzed inthe laboratory for constituents of interest. If sampling pointsor locations are unrepresentative, or biased sampling proce-dures are used, no amount of care in QA/QC in subsequent
stages will salvage an accurate picture of actual field condi-tions.
Duplicate samples are collected and not analyzed unlessit is later determined that they contain additional useful infor-mation. Soil samples are commonly duplicated.
Replicate samples are subsamples of the same samplethat are labeled separately to estimate the precision of labora-tory analytical results.
Split samples are field samples that are split between twostorage vessels or cut in half in the field. One subsample maybe analyzed by one laboratory and the other subsample maybe archived or given to another laboratory,
90
Table 7-5. Regions of Analyte Measurement
Figure 7-7. The ACS recommended definition of limit ofdetection (from ACS Committee on EnvironmentalImprovement, 1980).
Spiked samples are field samples that may be split withone aliquot receiving a spike volume of a reference standardto estimate the recovery of the analyte in the laboratory.Spiked samples allow estimates of accuracy and detect pos-sible matrix interference problems.
Laboratory blanks are similar to field blanks except thatthe distilled deionized water used in the laboratory at the timeeach batch of samples is received is analyzed in the samemanner as other samples. This type of sampling may detectcontamination that occurs in the laboratory.
Standard reference samples have been analyzed previ-ously by outside laboratories. These samples are availablefrom the National Institute of Standards and Technology orthe EPA to detect either instrument calibration error or the useof inappropriate laboratory analytical methods (Keith et al.,1983).
7.3 Statistical Techniques
7.3.1 Statistical Approaches to GeochemicalVariability
Virtually all soil sampling and most ground-water sam-pling that has been done at a high enough level of resolutionhave shown that chemical constituent concentrations are nei-ther normally distributed nor independent (i.e., noncorrelated).This creates special challenges for statistical analysis of geo-chemical sampling data because many of the traditional statis-tical techniques for analyzing sample data, such as linearregression and t-testing, assume that the population sampledhas the symmetric, bell-shaped Gaussian (normal) distribu-tion. Linear regression is probably the most frequently mis-
Source: ACS Committee on Environmental Improvement (1980)
used statistical technique in this context (Mann, 1987; Kite,1989).
The first step in analyzing geochemical data is to deter-mine whether they are normally distributed. If they are, tradi-tional techniques described in standard textbooks on statisticscan be used. If not, one or more of the following methodsmust be used (1) data transformations such as logarithmicconversions to create data sets that are normally distributedand hence amenable for analysis by conventional methods(Wilson et al., 1990, discuss how to evaluate bias that may beintroduced by this manipulation); (2) nonparametric or distri-bution-free statistical techniques that do not require indepen-dent data observations and (3) geostatistical techniques thatfacilitate differentiation of correlated and noncorrelated datasets and interpolation of values between sample points. Thetechnique of “fuzzy” linear regression maybe useful in hydro-logic situations where the relationship between variables isimprecise, data are inaccurate, and/or sample sizes are insuffi-cient (Bardossy et al., 1990). Subsurface contamination inves-tigations typically involve measurements of concentrationchanges in geochemical parameters over time. Consequently,statistical techniques designed specifically for analysis oftrends in time-series data are important (Harris et al., 1987;Montgomery et al., 1987).
Alhajjar et al. (1990) describe use of the median-polishstatistical methods of exploratory data analysis developed byTukey (1977) for analyzing highly variable geochemical datacollected during a study of chemical pollution from septicsystems. This technique is especially well suited for analyzingdata in two-way tables (multiple rows and columns) in whicheach data value is related simultaneously to two factors.
7.3.2 GeostatisticsGeostatistical techniques such as use of correlograms,
semivariograms, and kriging have gained increasing popular-ity in evaluating spatially distributed hydrologic and geo-chemical data in the last 10 years. Using empirical gold-oreevaluation techniques developed by D.C. Krige in SouthAfrica (hence the term kriging), the French mathematician G.Matheron developed the theory of regionalized variables inthe late 1960s (Matheron, 1971). This general theory ofsampling and estimating spatially dependent (autocorrelated)variables is well suited to analysis of hydrologic and geo-chemical parameters, which tend to be nonrandom in theclassical Gaussian statistical sense.
91
Geostatistical techniques have three main applications forcharacterization of subsurface variability: (1) they can assistin reducing spatial sampling intensity, and hence reduce sam-pling and analytical costs; (2) they can be used to differentiatesample data that are autccorrelated or noncorrelated, elucidat-ing trends for selecting the appropriate statistical analysis ofsampling analytical results; and (3) they can be used tointerpolate values at locations where measurements have notbeen made. The last application is done by kriging, a weightedmoving-averaging technique, that in most situations will provide the most accurate way of contouring data on physical andgeochemical parameters. Furthermore, a kriging standard de-viation map that provides a clear indication of the reliability ofcontours can be readily created from kriged contour data.
One of the first steps in geostatistical analysis is tocalculate the nonsampling variance (gamma) of samples atdifferent distance spacings. Gamma is a statistical measure ofthe difference between sample values. For example, if sampleswere taken from a 50-m grid, gamma would be calculated forthe samples spaced at 50 m, 100 m, 150 m, 200 m, and so on.Next, a semivariogram is plotted on a XY plot, where X isdistance and Y is the nonsampling variance. Figure 7-8 showsan “ideal” semivariogram. Samples within a certain range ofinfluence, also called the range of correlation (distance a inFigure 7-8), show an approximately linear correlation (areautocorrelated). At some spacing distance, if there is no trendin the data, a sill (C on Figure 7-8) marks a plateau that limitsthe range of correlation. The nonsampling variance betweensamples will equal C as long as the distance is greater than a.
From a sampling perspective, samples spaced closer thandistance a in Figure 7-8 will yield redundant, correlated data,which results in both unnecessary expense and complicationsin statistical analysis. The minimum distance at which samples
Figure 7-8. The ‘“ideal” shape for a semivariogram-sphericalmodel (from Clark, 1979).
are independent (distance a in Figure 7-8) is the optimumsampling distance.
Figure 7-9 shows a semivariogram of lead values in soilsampled by Flatman (1986) on a systematic 750-ft grid. Thediagram shows that samples for lead that are closer to eachother than about 1,200 ft are correlated. In other words, thesame information could be obtained by cutting the number ofsamples almost in half. Figure 7-10 shows a kriged contourmap of lead concentrations in the vicinity of the smelter, andFigure 7-11 shows contours of the standard deviations of thelead concentrations.
Table 7-6 summarizes ranges of influence (in meters) thathave been estimated for a variety of soil physical and chemi-cal parameters. Direct comparisons between different studiesare difficult, however, because definitions and the methodolo-gies for determining the range vary somewhat. Commonly,however, the range is scale-dependent, i.e., as the sample areaincreases, the range increases. For example, at the same siteGajem et al. (1981) found ranges of 1.5,21, and 260 m for pHvalues of 100-member transects spaced at 0.2, 2, and 20 m.
Semivariograms may exhibit a variety of correlation struc-tures other than the one shown in Figure 7-8, and correctinterpretation requires an understanding of the various modelsthat are available for describing semivariogram plots. Whendata are not normally distributed, such as when a spatial trendis present, estimating the correlation structure is difficult. Inthese cases, some of the techniques for transforming log-normal data for conventional statistical analysis can be used(Gilbert, 1987).
Most basic texts on geostatistics are still oriented towardsmining. Clark (1979) provides a good introduction togeostatistics and kriging, while more comprehensive treat-ments (all oriented toward mining) can be found in the follow-ing sources: David (1977), Isaaks and Srivastava (1989),Matheron (1971), and Journal and Huijbregts (1978). Olea(1974, 1975) provide a good introduction to the use of geosta-tistics in contour mapping of data. Gilbert and Simpson (1985)provide a good review of potentials and problems with usingkriging for estimating spatial pattern of contaminants
Trangmar et al. (1985) and Warrick et al. (1986) re-viewed specific geostatistical methods applied to spatial stud-ies of soil properties. Use of geostatistics in sampling for soilcontaminants is discussed by Flatman (1984), Flatman andYfantis (1984), and Flatman (1986). Delhomme (1978, 1979)reviewed the use of geostatistics in the characterization ofground-water variability, and Hughes and Lettenmaier (1981)and Sophocleous et al. (1982) discuss applications for ground-water monitoring network design.
7.4 Interpretation of Geochemical and WaterChemistry Data
Table 7-7 indexes some sources of information on (1)basic statistical approaches to data analysis, (2) methods foranalysis of soil data, and (3) methods for analysis of waterquality data. The general references on soil and water chemis-try listed in Table 7-1 provide a framework for interpreting
92
Figure 7-9. A semivariogram of Iead samples taken systematically on a 230-m (750-foot) grid (from Flatman, 1986).
background geochemistry. Hem (1985) is an especially goodsource for the interpretation of water quality data.
Gilbert (1987) presented probably the best systematictreatment of statistical methods for environmental pollutionmonitoring. Bury (1975) provides a comprehensive treatmentof basic statistical concepts and models oriented toward theapplied scientist. Hollander and Wolfe (1973), Lehmann andD’Abrera (1975), and Seigel (1956) offer more in-depth dis-cussion of nonparrametric statistical methods. Bury (1975)presents a table that is a useful guide for finding the appropri-ate nonparametric procedure for particular topics or problems.Chatfield (1984) is a good source on techniques for analysis oftime series.
7.4.1 Analysis of Censored DataTable 7-8 illustrates the effect of two types of censoring
of analytical results near and below the limit of detection.Data reported as less than the limit of detection are heavilycensored and yield an average concentration of 3.5 µg/Lsince only two values are quantified. Reporting of negativeconcentrations as zero is called negative censoring; in Pable7-8 negative censoring yields an average of 1.2 µg/L. Theuncensored data average 0.5 µg/L. The averages of the heavilyand negatively censored data would appear to indicate con-tamination, but the 95 percent confidence interval for theuncensored data is at best equivocal.
Gilliom et al. (1984) found that any censoring of trace-level water quality data, even when the censored data werehighly unreliable, reduced the ability to detect trends in the
data. Unfortunately, censored data continues to be routinelyreported by laboratories. The following references containdiscussions of statistical techniques for analyzing censoreddata: Gilbert (1987), Gilliom and Helsel (1986), Gilliom et al.(1984), Helsel and Gilliom (1986), McBean and Rovers (1984)and Porter et al. (1988).
7.4.2 Contaminant Levels versus BackgroundConditions
Numbers on a standard list from an analytical laboratoryarc useful only to the extent that they can be compared toknown or estimated background conditions before contamina-tion. Using such numbers effectively requires both data onbackground conditions and the use of appropriate techniquesto detect statistically significant departures from backgroundlevels. An analytical result from a rinse or cleaning blankbetween the limit of detection and the limit of quantificationmay indicate that more careful decontamination proceduresshould be followed, but does not add to the information onwhich to base remediation decisions.
Crustal and natural background abundances of metallicelements must be considered when evaluating analyses forinorganic contaminants. See the listing under “background”for soil chemical parameters and water chemistry in Table 8-1, which identifies some sources of background data on minorand trace elements in the United States. For organics, there isalways some background of total inorganic carbon and or-ganic carbon, which should be determined in some samples toidentify natural background levels. The amount of organicmatter may vary considerably in soil, but dissolved organic
93
carbon in ground water does not vary greatly. There aredefinite analytical difficulties in achieving reliable analyses inthe range of 0.1 to 0.5 percent organic carbon in the solidfraction.
Equilibrium calculations based on thorough chemicalanalysis may be useful for interpreting water quality data(Jenne, 1979; Melchior and Bassett 1990 Summers et al.,1985). For example, reducing or suboxic conditions, indicatedby low Eh (i.e., measured oxidation-reduction potential), lackof detectable dissolved oxygen, and presence of ferrous iron,may indicate conditions favorable for movement of elementssuch as manganese, mercury, chromium, and arsenic. Arsenic(V) under oxidizing conditions may be considered immobile,but under reducing conditions, arsenic (III) is often the pre-dicted “stable” species of arsenic and is frequently more toxicand more mobile than As(V) due to higher volubility (Holmand Curtis, 1984).
Figure 7-10. Kriged contour map of lead concentration in ppmaround a smelter (from Flatman, 1988).
94
Figure 7-11. Kriging standard deviation map for lead concentrations around a smelter (from Flatman, 1986).
95
Table 7-6. Reported Values of Ranges of Correlation of Soil Physical and Chemical Properties
Source Parameter Range or SiteScale (m)
96
Table 7-7. Sources of Information on Techniques for Anailzing Soil and Water-Quality Data
Topic References
Basic Statistical Approaches
General Bandat and Pierson (1986), Bury (1975), Gilbert (1987), Jessen (1978), tin (1966), Ott (1984)
Nonparametrics Hollander and Wolfe (1973), Lehmann (1975), Seigel (1956)
Time series Chatfield (1984)
Exploratory data Tukey (1977), Velleman and Hoaglin (1981), Alhajjar et al (1990)(Median-Polish)
Geostatistics (basic) Clark (1979), Englund and Sparks (1988), Gilbert and Simpson (1965), Journaf (1984),Olea (1974, 1975), Yates and Yates (1990)
Geostatistics (adv.) David (1977), Journal and Huijbregts (1978), Isaaks and Srivastava (1989), Matheron (1971)
Soil Data Analysis
Population properties Butler (1980), Sinclair (1986), Webster (1977)
Geostatistics Sinclair (1986), Trangmar et al. (1985), Warrick et al. (1986). See also Table 7-6
Contaminated soils Flatman (1964), Flatman and Yfantis (1984), and Flatman (1986)
Soil Gas Data See Table 9-5
Water Quality Data
General Beck and van Stratten (1983), Gillham et al (1983), U.S. EPA (1989)
Contaminant detection Chapman and El-Shaarawi (1989), Davis and McNichols (1988), Gibbons (1987a,b; 1990),McBean and Rovers (1990), McNichols and Davis (1988)
Geostatistics Delhomme (1978, 1979), Hughes and Lettenmaier (1981), Samper and Neuman (1985),Sophocleous et al (1982)
Population properties Harris et al. (1987), Montgomery et al. (1987)
Spatial data Lawrence and Upchurch (1976), McBean et al. (1988)
Time series data Close (1989), Harris et al. (1987), McBean et al. (1988), Montgomery et al (1987),Sgambat and Stedinger (1981), Yevjevich and Harmancioglu (1989)
Table 7-8. Effects of Censoring Analyte Signals at andBelow the Limit of Detection
Source: ASTM (1987)
97
7.5 ReferencesACS Committee on Environmental Improvement. 1980. Guide-
lines for Data Acquisition and Data Quality Evaluation inEnvironmental Chemistry. Analytical Chemistry 52:2242-2249.
Alhajjar, B.J., G. Chesters, and J.M. Harkin. 1990. Indicatorsof Chemical Pollution from Septic Systems. Ground Wa-ter 28(4):559-568.
American Public Health Association. 1990. Standard Meth-ods for the Examination of Water and Wastewater, 17thed. APHA, Washington, DC.
American Society for Testing and Materials (ASTM). AnnualBooks of ASTM Standards. Water and EnvironmentalTechnology, Volumes 11.01 and 11.02 (Water). ASTM,Philadelphia, PA.
Barcelona, M.J., D.P. Lettenmaier, and M.R. Schock. 1989.Network Design Factors for Assessing Temporal Vari-ability in Ground-Water Quality. Environmental Moni-toring and Assessment 12:149-179.
Bardossy, A., I. Bogardi, and L. Duckstein. 1990. FuzzyRegression in Hydrology. J. Hydrology 26(7):1497-1508.
Bamett, P.R. and E.C. Mallory, Jr. 1971. Determination ofMinor Elements in Water by Emission Spectroscopy.U.S. Geological Survey TWI 5-A2.
Beck, B.F. and G. Van Strateen (eds.). 1983. Uncertainty andForecasting of Water Quality. Springer-Verlag, New York.
Bendat, J.S. and A.G. Piersol. 1986. Random Data, Analysisand Measurement Procedures, 2nd ed. Wiley-Interscience,New York.
American Society for Testing and Materials (ASTM). 1987. Burgess T.M. and R. Webster. 1980. Optimal InterpolationStandard Practice for Intralaboratory Quality Control Pro- and Isarithmic Mapping of Soil Properties I. Thecedures and a Discussion on Reporting Low-Level Data. Variogram and Punctual Kriging. J. Soil Sci. 31:315-331.In: Annual Book of ASTM Standards, Vol. 11.01, D4210-83. ASTM, Philadelphia, PA. Bury, K.V. 1975. Statistical Models in Applied Science. John
Wiley & Sons, New York.American Society for Testing and Materials (ASTM), Sub-
committee 019.02.1983. Annual Book ASTM Standards, Butler, B.E. 1980. Soil Classification for Soil Survey, ChapterVolume 11.01, Chapter D, pp. 4210-4283. 2. Oxford University Press, New York.
Barcelona, M.J. 1988. Overview of the Sampling Process. In:Principles of Environmental Sampling, L.H. Keith (ed.),ACS Professional Reference Book, American ChemicalSociety, Washington, DC, pp. 1-23.
Barcelona, M.J. and J.A. Helfrich. 1986. Effects of WellConstruction Materials on Ground Water Samples.Environ. Sci. Technol. 20(11):1179-1184.
Barcelona, M.J., J.P. Gibb, and R.A. Miller. 1983. A Guide tothe Selection of Materials for Monitoring Well Construc-tion and Ground-Water Sampling. ISWS Contract Report327. Illinois State Water Survey, Champaign, IL.
Barcelona, M.J., J.A. Helfrich, E.E. Garske, and J.P. Gibb.1984. A Laboratory Evaluation of Ground-Water Sam-pling Mechanisms. Ground Water Monitoring Review4(2):32-41.
Barcelona, M.J., J.P. Gibb, J.A. Helfrich, and E.E. Garske.1985. Practical Guide for Ground-Water Sampling. EPA600/2-85/104 (NTIS PB86-137304). Also published asISWS Contract Report 374, Illinois State Water Survey,Champaign, IL.
Barcelona, M.J., J.A. Helfrich, and E.E. Garske. 1988. Verifi-cation of Sampling Methods and Selection of Materialsfor Ground-Water Contamination Studies. In: Ground-Water Contamination: Field Methods, A.G. Collins andA.I. Johnson (eds.), ASTM STP 963, American Societyfor Testing and Materials, Philadelphia, PA, pp. 221-231.
Campbell, J.A. and W.R. Mabey. 1985 A Systematic Ap-proach for Evaluating the Quality of Ground Water Moni-toring Data. Ground Water Monitoring Review 5(4):58-22.
Campbell, J.B. 1978. Spatial Variability of Sand Content andpH within Continuous Delineations of Two MappingUnits. Soil Sci. Soc. Am. J. 42:46044.
Chapman, D.T. and A.H. El-Shaarawi. 1989. Statistical Meth-ods for the Assessment of Point Source Pollution. Envi-ronmental Monitoring and Assessment 13(2/3):1-467.[Special issue with 21 papers].
Chatfield, C. 1984. The Analysis of Time Series: Theory andPractice, 3rd ed. Chapman and Hall, London.
Clark, I. 1979. Practical Geostatistics. Applied Science Pub-lishers, London.
Clifton, P.M. and S.P. Neuman. 1982. Effects of Kriging andInverse Modeling on Conditional Simulation of the AvraValley in Southern Arizona. Water Resources Research18:1215-1234.
Close, M.E. 1989. Effect of Serial Correction on GroundWater Quality Sampling Frequency. Water ResourcesBulletin 25(3):507-515.
David, M. 1977. Geostatistical Ore Reserve Estimation.Elsevier, New York.
98
1
Davis, C.B., and R.J. McNichols. 1988. Discussion of “Statis-tical Prediction Intervals for the Evaluations of Ground-Water Quality.” Ground Water 25(1):90-91.
Delhomme, J.P. 1978. Kriging in the Hydrosciences. Adv.Water Resources 1:251-266.
Delhomme, J.P. 1979. Spatial Variability and Uncertainty inGroundwater Flow Parameters: A Geostatistical Approach.Water Resources Research 15:269-280.
Einarson, J.H. and P.C. Pei. 1988. A Comparison of Labora-tory Performances. Environ. Sci. Technol. 22:1121-1125.
Englund, E.J. and A.R. Sparks. 1988. Geo-EAS (GeostatisticalEnvironmental Assessment Software) User’s Guide. EPA/600/4-88/033a (Guide: NTIS PB89-151252, Software:PB89-151245).
Evans, R.B. 1986. Ground-Water Monitoring Data QualityObjectives for Remedial Site Investigations. In: QualityControl in Remedial Site Investigation: Hazardous andIndustrial Solid Waste Testing, Fifth Volume, C.L. Perket(ed.), ASTM STP 925, American Society for Testing andMaterials, pp. 21-33.
Flatman, G.T. 1984. Using Geostatistics in Assessing LeadContamination Near Smelters. In: Environmental Sam-pling for Hazardous Wastes, G.E. Schweitzer and J.A.Santolucito (eds.), ACS Symp. Ser. 267, American Chemi-cal Society, Washington, DC, pp. 43-52.
Flatman, G.T. 1986. Design of Soil Sampling Programs:Statistical Considerations. In: Quality Control in Reme-dial Site Investigation: Hazardous and Industrial SolidWaste Testing, 5th volume, C.L. Perket (ed.), ASTMSTP 925, American Society for Testing and Materials,Philadelphia, PA, pp. 43-56.
Flatman, G.T. and A.A. Yfantis. 1984. Geostatistical Strategyfor Soil Sampling: The Survey and the Census. Environ-mental Monitoring and Assessment 4:335-350.
Folorunso, O.A. and D.E. Rolston. 1984. Spatial Variabilityof Field Measured Denitrification Gas Fluxes. Soil Sci.Soc. Am. J. 48:1214-1219.
Fresenius, W., K.E. Quentin, and W. Schneider (eds.). 1988.Water Analysis: A Practical Guide to Physico-Chemicaland Microbiological Water Examination and Quality As-surance. Springer-Verlag, New York.
Gajem, Y. M., A. W. Warrick, and D.E. Myers. 1981. SpatialDependence of Physical Properties of a Typic TornfluventSoil. Soil Sci. Soc. Am. J. 46:709-715.
Gibb, J.P, R.M. Schuller, and R.A. Griffin. 1981. Proceduresfor the Collection of Representative Water Quality Datafrom Monitoring Wells. ISWS/IGS Cooperative GroundWater Report 7. Illinois State Water Survey, Champaign,IL.
Gibbons, R.D. 1987a. Statistical Prediction Intervals for theEvaluation of Ground-Water Quality. Ground Water25(4):455-465.
Gibbons, R.D. 1987b. Statistical Models for the Analysis ofVolatile Organic Compounds in Waste Disposal Facili-ties. Ground Water 25:572-580.
Gibbons, R.D. 1990. A General Statistical Procedure forGround-Water Detection Monitoring at Waste DisposalFacilities. Ground Water 28(2):235-243.
Gilbert, R.O. 1987. Statistical Methods for EnvironmentalPollution Monitoring. Van Nostrand Reinhold, New York.
Gilbert, R.O. and J.C. Simpson. 1985. Kriging from Estimat-ing Spatial Pattern of Contaminants: Potential and Prob-lems. Environmental Monitoring and Assessment5:113-135.
Gillham, R. W., M.J.L. Robin, J.F. Barker and J.A. Cherry.1983. Groundwater Monitoring and Sample Bias. APIPublication 4367. American Petroleum Institute, Wash-ington, DC.
Gilliom, R.J. and D.R. Helsel. 1986. Estimation of Distribu-tional Parameters for Censored Trace Level Water Qual-ity Data: 1. Estimation Techniques. Water ResourcesResearch 22:135-146.
Gilliom, R.J., R.M. Hirsch, and E.J. Gilroy. 1984. Effect ofCensoring Trace-Level Water-Quality Data on Trend-Detection Capability. Environ. Sci. Technol. 18:530-536.
Goerlitz, D.F. and E. Brown. 1972. Methods for Analysis ofOrganic Substances in Water. U.S. Geological SurveyTWI 5-A3. (updated by Wershaw et al., 1987).
Harris, J., J.C. Loftis, and R.H. Montgomery. 1987. StatisticalMethods for Characterizing Ground-Water Quality.Ground Water 25(2): 185-193.
Helsel, D.R. and R.J. Gilliom. 1986. Estimation of Distribu-tional Parameters for Censored Trace Level Water Qual-ity Data: 2. Verification and Applications. WaterResources Research 22:146-155.
Hem, J.D. 1985. Study and Interpretation of the ChemicalCharacteristics of Natural Water. U.S. Geological SurveyWater-Supply Paper 2254.
Ho, J.S-Y. 1983. Effect of Sampling Variables on Recoveryof Volatile Organics in Water. J. Am. Water Works Ass.12:583-586.
Hollander, M. and D.A. Wolfe. 1973. Nonparametric Statisti-cal Methods. John Wiley & Sons, New York.
99
Holm, T.R. and C.D. Curtis. 1984. A Comparison of Oxida-tion-Reduction Potentials Calculated from the As(V)/As(III) and Fe(III)/Fe(II) Couples with Measured Plati-num Electrode Potentials in Ground Water. J. Contami-nant Hydrology 5:67-81.
Hughes, J.P. and D.P. Lettenmaier. 1981. Data Requirementsfor Kriging: Estimation and Network Design. Water Re-sources Research 17:1641-1650.
International Union of Pure and Applied Chemistry (IUPAC).1978. Nomenclature, Symbols, Units and their Usage inSpectrochemical Analysis—II. Data Interpretation.Spectrochimica Acts B 33B:242-245.
Isaaks, E.H. and R.M. Srivastava. 1989. Applied Geostatistics.Oxford University Press, New York.
Jenne, E.A. (ed.). 1979. Chemical Modeling in AqueousSystems: Speciation, Sorption, Volubility, and Kinetics.ACS Symp. Series 93. American Chemical Society, Wash-ington, DC.
Jessen, R.J. 1978. Statistical Survey Techniques. John Wiley& Sons, New York.
Journal, A.G. 1984. New Ways of Assessing Spatial Distribu-tion of Pollutants. In: Environmental Sampling for Haz-ardous Wastes, G.E. Schweitzer and J.A. Santolucito(eds.), ACS Symp. Ser. 267, American Chemical Society,Washington, DC, pp. 109-118.
Journal, A.G. and C.J. Huijbregts. 1978. Mining Geostatistics.Academic Press, New York.
Kachonoski, R.G., D.E. Rolston, and E. deJong. 1985. Spatialand Spectral Relationships of Soil Properties andMicrotopography: I. Density and Thickness of A Hori-zon. Soil Sci. Soc. Am. J. 49:804-812.
Keith, S.J., M.T. Frank, G. McCarty, and G. Massman. 1983.Dealing with the Problem of Obtaining Accurate Ground-Water Quality Analytical Results. In: Proc. 3rd Nat.Symp. on Aquifer Restoration and Ground Water Moni-toring, National Water Well Association, Dublin, OH,pp. 272-283.
Kirchmer, C.J. 1983. Quality Control in Water Analyses.Environ. Sci. Technol. 17(4):174A-181A.
Kite, G. 1989. Some Statistical Observations. Water Re-sources Bulletin 25(3):483-490. [See also 1990 discus-sion by Kirby et al., and reply by Kite in Water ResourcesBulletin 26(4)693-698].
Klute, A. (ed.). 1986. Methods of Soils Analysis, Part l-Physical and Mineralogical Methods, 2nd ed. ASA Mono-graph 9. American Society of Agronomy, Madison, WI.
Kopp, J.F. and G.D. McKee. 1983. Methods for ChemicalAnalysis of Water and Wastes. EPA 600/4-74-020 (NTISPB84-128677). [supersedes report with the same titledated 1979].
Lawrence, F.W. and S.B. Upchurch. 1976. Identification ofGeochemical Patterns in Groundwater by NumericalAnalysis. In: Advances in Groundwater Hydrology,AWRA Proc. Ser. No. 21. American Water ResourcesAssociation, Bethesda, MD.
Lehmann, E.L. 1975. Nonparametrics: Statistical MethodsBased on Ranks. Holden Day, San Francisco, CA, andMcGraw-Hill, New York.
Lin, P.C.L. 1986. User-Friendly IBM PC Computer Programsfor Solving Sampling and Statistical Problems. EPA/600/4-86/023 (NTIS PB86-203783).
Long, G.L. and J.D. Winefordner. 1983. Limit of Detection, aCloser Look at the IUPAC Definition. Analytical Chem-istry 55(7):712A-724A.
Longbottom, J.E. and J.J. Lichtenberg. 1982. Methods forOrganic Chemical Analysis of Municipal and IndustrialWastewater. EPA/600/4-82-057 (NTIS PB83-201798).
Mann, J.C. 1987. Misuses of Linear Regression in EarthSciences. In: Use and Abuse of Statistical Methods in theEarth Sciences, W.B. Size (ed.), Oxford University Press,New York, pp.74-106.
Matheron, G. 1971. The Theory of Regionalized Variablesand Its Applications. Cahiers du Centre de MorphologicMathematique de Fontainebleau. No. 5.
McBean, E.A. and F.A. Rovers. 1984. Alternatives for Han-dling Detection Limit Data in Impact Assessments. GroundWater Monitoring Review 4(2):42-44.
McBean, E.A. and F.A. Rovers. 1990. Flexible Selection ofStatistical Discrimination Tests for Field-Monitored Data.In: Ground Water and Vadose Zone Monitoring, ASTMSTP 1053, D.M. Nielsen and A.I. Johnson (eds.), Ameri-can Society for Testing and Materials, Philadelphia, PA,pp. 256-265.
McBean, E. A., M. Kompter, and F. Rovers. 1988. A CriticalExamination of Approximations Implicit in Cochran’sProcedure. Ground Water Monitoring Review 8(1):83-87.
McBratney, A.B. and R. Webster. 1981. Spatial Dependenceand Classification of the Soil Along a Transect in North-east Scotland. Guoderma 26:63-82.
McNichols, R.J. and C.B. Davis. 1988. Statistical Issues andProblems in Ground Water Detection Monitoring at Haz-ardous Waste Facilities. Ground Water Monitoring Re-view 8(4):135-150.
100
Melchior, D.C. and R.L. Bassett (eds.). 1990. Chemical Mod-eling of Aqueous Systems II. ACS Symp. Series 416.American Chemical Society, Washington, DC.
Montgomery, R.H., J.C. Loftis, and J. Harris. 1987. StatisticalCharacteristics of Ground Water Quality Variability.Ground Water 25(2): 176-184.
Mueller, W., D.L. Smith, and L.H. Keith. 1991. Compilationof EPA’s Sampling and Analysis Methods. Lewis Pub-lishers, Chelsea, MI, 465 pp. [On diskette EPA’s Sam-pling and Analysis Methods Database, Vol. 1 (IndustrialChemicals), Vol. 2 (Pesticides, Herbicides, Dioxins andPCBS), and Vol. 3 (Elements and Water Quality Param-eters)].
Noblett, J.G. and J.M. Burke. 1990. FGD Chemistry andAnalytical Methods Handbook, 1: Process Chemistry—Sampling, Measurement, Laboratory, and Process Perfor-mance Guidelines, Revision 1. EPRI CS-3612. ElectricPower Research Institute, Palo Alto, CA. [Originallypublished in 1984, see Radian Corporation (1988) forVolume 2].
Olea, R.A. 1974. Optimal Contour Mapping Using UniversalKriging. J. Geophysical Research 79(5):696-702.
Olea, R.A. 1975. Optimum Mapping Techniques Using Re-gionalized Variable Theory. KGS Series on Spatial Analy-sis No. 2. Kansas Geological Survey, Lawrence, KS.
Ott, L. 1984. An Introduction to Statistical Methods and DataAnalysis, 2nd ed. Duxbury Press, Boston, MA, 775 pp.
Page, A.L., R.H. Miller, D.R. Keeney (eds.). 1982. Methodsof Soils Analysis, Part 2-Chemical and MicrobiologicalProperties, 2nd ed. ASA Monograph 9. American Societyof Agronomy, Madison, WI.
Pennino, J.D. 1988. There’s No Such Thing as a Representa-tive Ground Water Sample. Ground Water MonitoringReview 8(3):4-9.
Porter, P.S. 1986. A Description of Measurement Error NearLimits of Detection. In: Monitoring to Detect Changes inWater Quality Series, D. Lerner (ed.), Int. Ass. of Hydro-logical Sciences Pub. No. 157.
Porter, P.S., R.C. Ward, and H.F. Bell. 1988. The DetectionLimit. Environ. Sci. Technol. 22:856-861.
Provost L.P. and R.S. Elder. 1985. Choosing Cost-EffectiveQA/QC Programs for Chemical Analysis. EPA/600/4-85/056 (NTIS PB85-241461).
Radian Corporation. 1988. FGD Chemistry and AnalyticalMethods Handbook, Vol. 2: Chemical and Physical TestMethods, Revision 1. EPRI CS-3612. Electric PowerResearch Institute, Palo Alto, CA. [Originally publishedin 1984, see Noblett and Burke (1990) for Volume 1].
Rainwater, F.H. and L.L. Thatcher. 1960. Methods for Collec-tions and Analysis of Water Samples. U.S. GeologicalSurvey Water-Supply Paper 1454.
Rice, G., J. Brinkman, and D. Muller. 1988. Reliability ofChemical Analyses of Water Samples-The Experienceof the UMTRA Project. Ground Water Monitoring Re-view 8(3):71-75.
Russo, D. and E. Bresler. 1981a. Effect of Field Variability inSoil Hydraulic Properties on Solutions and UnsaturatedWater and Salt Flows. Soil Sci. Soc. Am. J. 45:675-681.
Russo, D. and E. Bresler. 198lb. Soil Hydraulic Properties asStochastic Processes I. An Analysis of Field SpatialVariability. Soil Sci. Soc. Am J. 45:682-687.
Samper, F.J. and S.P. Neuman. 1985. Geostatistical Analysisof Hydrochemical Data from the Madrid Basin, Spain(Abstract). Eos (Trans. Am. Geophysical Union)66(46):905.
Schuller, R.M., J.P. Gibb, and R.A. Griffin. 1981. Recom-mended Sampling Procedures for Monitoring Wells.Ground Water Monitoring Review 1(1):42-46.
Seigel, S. 1956. Nonparametric Statistics for the BehavioralSciences, McGraw-Hill, New York.
Sgambat, J.P. and J.R. Stedinger. 1981. Confidence in Ground-Water Monitoring. Ground Water Monitoring Reviewl(Spring):62-69.
Sinclair, A.J. 1986. Statistical Interpretation of Soil Geo-chemical Data. In: Exploration Geochemistry, Designand Interpretation of Soil Surveys. Reviews in EconomicGeology 3:97-115.
Skoog, D.A. 1985. Principles of Instrumental Analysis, 3rded. Saunders College Publishing, Philadelphia, PA.
Skougstad, M.W. et al. (eds.). 1979. Methods for Determina-tion of Inorganic Substances in Water and Fluvial Sedi-ments. U.S. Geological Survey TWI 5-Al.
Sisson, J.B. and P.J. Wierenga. 1981. Spatial Variability ofSteady-State Infiltration Rates as a Stochastic Process.Soil Sci. Soc. Am. J. 45:699-704.
Smith, K.A. (ed.). 1991. Soil Analysis: Modem InstrumentalMethods, 2nd ed. Marcell Dekker, New York.
Sophodeous, M. J.E. Paschetto, and R.A. Olea. 1982. Gmund-Water Network Design for Northwest Kansas, Using theTheory of Regionalized Variables. Ground Water 20:48-58.
Starks, T.H. and G.T. Flatman. 1991. RCRA Ground-WaterMonitoring Decision Procedures Viewed as Quality Con-trol Schemes. Environmental Monitoring and Assess-ment 16:19-37.
101
Summers, K.V., G.L. Rupp, G.F. Davis, and S.A. Gherini.1985. Ground Water Data Analysis at Utility WasteDisposal Sites. EPRI EA-4165. Electric Power ResearchInstitute, Palo Alto, CA.
Thompson, C.M., et al. 1989. Techniques to Develop Data forHydrogeochemical Models. EPRI EN-6637. ElectricPower Research Institute, Palo Alto, CA.
Trangmar, B.B., R.S. YOSt and G. Uehara. 1985. Applicationof Geostatistics to Spatial Studies of Soil Properties.Advances in Agronomy 38:45-93.
Tukey, J.W. 1977. Exploratory Data Analysis. Addison-Wesley, New York, 506 pp.
U.S. Environmental Protection Agency (EPA). 1988. FieldScreening Methods Catalog: User’s Guide. EPA/540/2-88/005. FSMC System Coordinator, OERR, AnalyticalOperations Branch (WH-548-A), U.S. EPA, Washington,DC 20460.
U.S. Environmental Protection Agency (EPA). 1989. Guid-ance Document on Statistical Analysis of Ground-WaterMonitoring Data at RCRA Facilities-Interim Final Guid-ance. Office of Solid Waste Management Division (NTISPB89-151047).
van Kuilenburg, J., J.J. DeGruijter, B.A. Marsma, and J.Bourna. 1982. Accuracy of Spatial Interpretation Be-tween Point Data on Soil Moisture Supply CapacityCompared with Estimations from Mapping Units.Geoderma 27:311-325.
Vauclin, M., S.R. Vieira, R. Bernard, and J.L. Hatfield. 1982.Spatial Variability of Surface Temperature Along toTransects on a Bare Soil. Water Resources Research18:1677-1686.
Vauclin, M., S.R. Vieira, G. Vachaud, and D.R. Nielsen.1983. The Use of Co-Kriging with Limited Field SoilObservations. Soil Sci. Soc. Am. J. 47:175-184.
Velleman, P.F. and D.C. Hoaglin. 1981. Applications, Basics,and Computing of Exploratory Data Analysis. DuxburyPress, Boston, MA, 354 pp.
Vieira, S.R., D.R. Nielsen, and J.W. Biggar. 1981. SpatialVariability of Field Measured Infiltration Rate. Soil Sci.Soc. Am. J. 45:1040-1048.
Warrick, A.W., D.E. Myers, and D.R. Nielsen. 1986. Geo-statistical Methods Applied to Soil Science. In: Methodsof Soil Analysis, Part I—Physical and MineralogicalMethods, 2nd ed., A. Klute (ed.), ASA Monograph No. 9,American Society of Agronomy, Madison, WI, pp. 53-82.
Webster, R. 1977. Quantitative and Numerical Methods inSoil Classification and Survey. Oxford University Press,New York.
Wershaw, R.L., M.J. Fishman, R.R. Bragge, and L.E. Lowe(eds.). 1987. Methods for the Determination of OrganicSubstances in Water and Fluvial Sediments. U.S. Gee-logical Survey TWI 5-A3. (Revision of Goerlitz andBrown, 1972).
Westerman, R.L. 1990. Soil Testing and Plant Analysis, 3rded. Soil Science Society of America, Madison, WI, 812PP.
Willard, H. H., L.L. Merritt, Jr., J.A. Dean, and F.A. Settle, Jr.1988. Instrumental Methods of Analysis, 7th ed.Wadsworth Publishing Co., Belmont, CA.
Wilson, B.G., B.J. Adams, and B.W. Karney. 1990. Bias inLog-Transformed Frequency Distributions. J. Hydrology118:19-37.
Wollum II, A.G. and D.K. Cassel. 1984. Spatial Variability ofRhizobium japonicum in Two North Carolina Soils. SoilSci. Soc. Am. J. 48:1082-1086.
Wood, W.W. 1976. Guidelines for Collection and Field Analy-sis of Ground-Water Samples for Selected Unstable Con-stituents. U.S. Geological Survey TWI 1-D2
Yates, S.R. and M.V. Yates. 1990. Geostatistics for WasteManagement: A User’s Manual for the GEOPACK (Ver-sion 1.0) Geostatistical Software System. EPA/600/8-90/004 (NTIS PB90-186420/AS).
Yevjevich, V. and N.B. Harmancioglu. 1989. Description ofPeriodic Variation in Parameters of Hydrologic TimeSeries. Water Resources Research 25(3):421-428.
Yost, R.S., G. Uehara, and R.L. Fox. 1982. GeostatisticalAnalysis of Soil Chemical Properties of Large LandAreas. I. Variograms. Soil Sci. Sco. Am. J. 46:1028-1032.
102
Chapter 8Geochemical Variability of the Natural and Contaminated Subsurface Environment
J. Russell Boulding and Michael J. Barcelona
This chapter focuses on subsurface geochemical pro-cesses and environmental parameters that may significantlyaffect the accuracy of geochemical sampling to characterizethe natural and contaminated subsurface. Subsequent chaptersexamine in more detail subsurface physiochemical and deg-radation processes that affect the fate and transport of con-taminants. Table 8-1 indexes references on topics covered inthis chapter.
8.1 Overview of Subsurface GeochemistryA basic assumption in performing remediation is that
one cannot remediate what is not observed. Consequently,complete geochemical characterization of the subsurface re-quires an understanding of what to observe and how to goabout making the observations. Elements and compounds inthe subsurface may exist in one or more of three phases (solid,liquid, or gas). Within a phase, a substance may exist asseveral forms or species (e.g., ions, neutral molecules, andcomplex molecules in water). The partitioning of naturalconstituents and contaminants between solid, liquid, and gasor their transformation to other chemical forms is dependenton both the thermodynamics and kinetics of different types ofchemical processes. Thermodynamic prediction and reactionkinetics may be strongly influenced by subsurface environ-mental conditions. Information on indicators of ground-waterconditions, such as pH, Eh, temperature, and pressure, there-fore, is essential for interpreting geochemical data.
8.1.1 Geochemical ProcessesMajor geochemical processes in the subsurface include
(1) acid-base equilibria (also called ionization); (2) sorption-desorption; (3) precipitation-dissolution; (4) oxidation-reduc-tion (redox reactions); and (5) hydrolysis (see Chapters 10,12, and 13). Microorganisms frequently are the catalysts orpromoters of reactions in the subsurface. Volatilization isanother important process affecting contaminants that readilymove into the gas phase. Interactions between these variousprocesses are typically complex and must be understood interms of both thermodynamic and kinetic controls.
Thermodynamically, a chemical system is in equilibriumwhen its free energy is minimized; thus, thermodynamicprinciples define the stability of substances within the system
and whether a reaction will tend to occur. Thermodynamiccalculations can predict whether a chemical reaction is likelyto occur under specified conditions but give no indication ofhow fast the reaction will occur. Kinetics describe the rate ofchemical reactions. Some reactions, such as the reaction thatoccurs when a strong acid is added to water, will occur almostinstantaneously; other reactions, such as the hydrolysis ofcyanides at low pH, may take tens of thousands of years.
In nonequilibrium systems, chemical processes act toalter the chemical composition and/or phase of the system,and the system may tend to approach equilibrium. Simplesystems, such as dilute mixtures of sodium chloride andwater, attain solution equilibrium quickly, whereas complexsystems may only tend towards equilibrium. For example,geochemical modeling by Apps et al. (1988) suggests thatGulf Coast brines are not in equilibrium after tens of thou-sands of years with respect to magnesium and sulfate concen-tration. Lindberg and Runnells (1984) have suggested thatground water is rarely, if ever, in complete equilibrium withrespect to redox reactions.
Equilibrium implies that as long as no significant changesin environmental factors or phases occur within the system,the chemical composition of the system will be predictable.An equilibrium state does not imply that chemical reactionscease, rather that the rates of forward and reverse reactionscompensate one other.
8.1.2 Environmental ParametersThe act of sampling the subsurface tends to alter its
chemical equilibrium and results in reactions that may removeor release some of the chemical constituents being measured.The potential geochemical effects of drilling methods, materi-als used for well construction and sampling devices, andsampling methods all must be considered when developing asampling protocol. The sensitivity of a chemical system todisturbance depends on a number of physical and chemicalenvironmental parameters. Some of the most important ofthese parameters are discussed below, along with examples ofhow sampling may bias the results of laboratory analyses.
The major geochemical parameters that characterize thesubsurface include (1) water content, (2) hydrogen ion con-centration (pH), (3) redox potential (Eh), (4) microbial popu-
103
Table 8-1. Sources of Information on Natural and Contaminant Variability of Geochemical Parameters In the Subsurface
TOpiC References
Soil Chemical Parameters
General chemistry Bohn et al. (1985), Bolt and Bruggenwert (1978), Dragun (1988), Fairbridge and Finkl (1979), Sparks (1986,1989), Sposito (1984, 1989)
Background levels Connor and Shacklette (1975), Ebens and Shacklette (1982), Shacklette et al. (1971a,b, 1973, 1974)
Redox chemistry Brookins (1988), Ponnamperuma (1972), Ransom and Smeck (1986)
Contaminants Loehr et al. (1986)
Soil gases Barber et al. (1990), van Cleemput and El-Sebaay (1985)
Soil Physical Parameters
Variability
Flow channels
Vadose Zone
General
Water movement
Water Chemistry
General
Background levels
Redox Chemistry
Biochemical Changes
Corrosion/scaling
Variability
Jury (1985). See also Table 7-6
Bouma et al (1983), Miller (1975), Simpson and Cunningham (1982), White (1985)
Arnold et al. (1982), Evans and Nicholson (1987), Rijtema and Wassink (1969), Yaron et al. (1984), Zimmie andRiggs (1979)
Barnes (1989), Diment and Watson (1985), Hill and Parlange (1972), Raats (1973)
Drew (1989), Eriksson (1985), Faust and Aly (1981), Garrels and Christ (1965), Hem (1985), Lloyd andHeathcote (1985), Morel (1983), Pagendorf (1978), Stumm and Morgan (1981)
Durum and Haffty (1961), Durum et al. (1971), Ebens and Shacklette (1982), Ledin et al. (1989), Leenheer et al.(1974), Thurman (1985), White et al. (1963)
Baas-Backing et al. (1960), Back and Barnes (1965), Barcelona et al. (1989a), Champ et al. (1979), Edmunds(1973), Hem and Cropper (1959), Lindberg and Runnells (1984), Smith et al. (1991), Zehnder and Stumm(1988), ZoBell (1946). (See also, Tables 8-9 and 8- 10.)
Bouwer and McCarty (1984), Ghiorse and Wilson (1988), Smith et al. (1991), Wood and Bassett (1973)
Barnes and Clarke (1969), Langelier (1936), Larson and Buswell (1942), Ryzner (1944), Singley et al. (1985), Stiffand Davis (1952).
Back and Hanshaw (1988), Montgomery et al. (1987), Schmidt (1977), Seaber (1965), van Beek and van Puffelen(1987). (See also Tables 7-9 and 7-10.)
lation, (5) salinity and dissolved constituents, (6) physical andchemical character of solids, (7) temperature, and (8) pres-sure. Eh, pH, and pressure are probably the most importantparameters affecting sampling of near-surface aquifers; thesefactors strongly influence microbial population. Dissolvedconstituents and the physical and chemical character of sub-surface solids are highly site specific and influenced primarilyby geologic and soil-forming processes. Salinity, temperature,and solution composition gain increasing importance as thedepth of sampling increases.
pH and Alkalinity. The pH and alkalinity are mastervariables that help to describe solution composition and po-tential for preeipitation reactions. For example, pump-and-treat operations using air stripping to remove volatile organiccompounds (VOCs) can increase pH by 0.5 to 1 pH unitthrough removing carbon dioxide, with subsequent precipita-tion of calcium carbonate and iron oxides. Table 7-2 identifieschanges in pH that may result from sampling methods andmaterials. Table 8-2 identifies the effects of pH on a numberof subsurface geochemical processes,
Alkalinity indicates the buffer capacity or resistance tochange in pH, A solution with high buffer capacity has a large
resistance to change in pH, requiring the addition of a propor-tionally large amount of acid or base to change the solutionpH condition in the water. Since carbonate buffering is com-mon to most natural waters, the solution pH may be quitesensitive to volatilization of CO2 during sampling operations.
Redox Potential. The oxidation-reduction potential, orEh, is an expression of the intensity of redox conditions in asystem. It is measured in volts or millivolts (mV) as thepotential difference between a working electrode and thestandard hydrogen electrode. Positive readings in natural wa-ter generally indicate oxidizing conditions, and negative read-ings indicate reducing conditions. Ponnamperuma (1972)suggests that Eh values of +200 mV or lower indicate reduc-ing conditions in near-surface soils and sediments. Surfacewater bodies are generally around 400 to 600 mV becausethey are often in equilibrium with oxygen in the atmosphere.Principal oxidizing species in ground-water systems are oxy-gen and perhaps some hydrogen peroxide (the intermediatespecies in the reduction of oxygen to water). Other oxidizingspecies in ground water include nitrate and manganese (IV)and Fe(III). Under reducing conditions, Fe(III) species willtend to be reduced to Fe(II), sulfate is reduced to sulfide, and
104
Table 8-2. Effects of pH on Subsurface Geochemical Processes and Other Environmental Factors
Process/Factor pH Effect
Acid-base .
Adsorption-desorption
Precipitation-dissolution
Complexation
Oxidation-reduction
biodegradation
Eh
Salinity
Reservoir matrix
Temperature
Pressure
Measures acid-base reactions. Strong acids (bases) will tend to change pH: weak acids (bases) Will buffersolutions to minimize pH changes.
Strongly influences adsorption, because hydrogen ions play an active role in both chemical and physical bondingprocesses. Abbility of heavy metals is strongiy influenced by pH. Adsorption rates of organics are also PHdependant.
Strongly influences precipitation-dissolution reactions. Mixing of solutions with different pH often results inprecipitation reactions. Sea also reservoir matrix below.
Strongly influences positions of equilibria involving complex ions and metal chelate formation.
Redox systems generally become more reducing with increasing pH (ZoBell, 1946).
in combination with Eh, strongly influences the types of bacteria that will be present. High-to medium-pH, low-Ehenvironments will generally restrict bacterial populations to sulfate reducers and heterotrophic anaerobes
(Baas-Becking et al., 1960).
increasing pH generally lowers Eh.
pH-induced dissolution increases salinity: pH-induced precipitation decreases salinity.
Acidic solutions tend to dissolve carbonates and clays; highly alkaline solutions tend to dissolve silica and clays.Greater pH generally increases cation-exchange capacity of clays.
pH-driven exothermic (heat-releasing) reactions will increase fluid temperature: pH-driven endothermic (heat-consuming) reactions will decrease fluid temperature.
Will not infiuence pressure unless pH-induced reactions result in a significant change in the volume of reactionproducts.
Source: Adapted from U.S. EPA (1989)
carbon dioxide to methane. Oxidation/reduction processes arediscussed further in Section 12.1.3.
Most redox reactions in the subsurface are microbiallymediated. The measurement of the major by-products of thesereactions may be a better indicator of the strength of thereducing environment than Eh measurements or calculatedequilibrium potentials. A sequence of redox reactions underincreasing reducing conditions may be (1) denitrification re-actions which deplete nitrate and produce nitrogen gas, (2)sulfate reduction which depletes sulfate and produces hydro-gen sulfide, and (3) methanogenic reactions which depletecarbon dioxide and produce methane. Microbially mediatedredox processes are discussed further in Section 12.2.3.
Redox potential measurements or calculated potentialsare only measures of intensity. Reduction capacity measuresthe resistance to change in the redox potential, and is analo-gous to buffer capacity for pH in water. Reduction capacity ismeasured by how much oxidizing or reducing constituentmust be added to change redox conditions. Ground-watersystems tend to have some natural reduction capacity due tothe presence of organic carbon in aquifer solids. The introduc-
tion of organic contaminants, which serve as an energy sourcefor microorganisms to ground water, increases the tendency toshift towards more reducing conditions. In contrast bias caneasily be introduced into analytical results by the addition ofoxygen during the sampling process. Increases in dissolvedoxygen, resulting in decreased Fe(II) concentrations in samples(see Table 7-2), and precipitation of iron oxides are commonbiases introduced by the exposure of ground-water samples tothe atmosphere.
The concept of biologically mediated redox zones isuseful for evaluating the biodegradation of organic contami-nants in ground water. Table 8-3 shows how the degradationof various organic micropollutants might occur with increas-ing distance from a point of injection.
When organic contaminants are present in relatively lowconcentrations, as with artificial recharge of treated sewageeffluent, oxygen is present near the zone of injection andcompounds susceptible to aerobic biodegradation will decom-pose. As the redox potential declines at a greater distancefrom the point of injection, denitrifying conditions develop,and compounds such as carbon tetrachloride, which are not
105
susceptible to aerobic degradation, may be degraded. If redoxpotential declines further and conditions favorable for sulfate-reducing bacteria exist, cresols and chlorophenols may bedegraded. Finally, where methanogenic bacteria predominate,halogenated aliphatics that may have passed through thedenitrification zone may be degraded.
Implicit in this redox zone model is that compounds thatpass through the zone in which they are susceptible to biodeg-radation will persist in ground water unless immobilized oraltered by inorganic chemical processes. In heavily contami-nated ground water, this sequence may be reversed, with thegreatest reducing conditions closest to the point of contaminat-ion grading to mildly oxygenated conditions (in shallowaquifers, at least) at the outside edge of the contaminantplume.
Salinity and Dissolved Constituents. Total dissolved sol-ids (TDS) content can be qualitatively estimated in the fieldby measuring specific conductance. The major dissolved con-stituents in ground water may be near equilibrium with condi-tions at their location, although subject to seasonal fluctuations(see Section 8.4). During well development, if purging orsampling-process ground water is mixed with water of differ-ing salinity or chemical composition, the result may be pre-cipitation-dissolution and redox reactions that significantlychange the inorganic chemistry of a sample. Geochemicalsampling of water wells that tap multiple aquifers is especiallyproblematic because of these effects. The more saline thewater, or the more different in chemical composition the twowaters, the greater the bias that can be introduced to geo-chemical samples.
Soil/Aquifer Matrix. The mineralogy and particle sizedistribution of the unsaturated and saturated zones stronglyinfluence geochemistry of subsurface waters. As particle sizedecreases, the surface area increases, providing more opportu-nities for chemical reactions between solids and water. Aparticularly important chemical parameter of solids is thecation exchange capacity (CEC). CEC is a function of miner-alogy, particle size, and previous geochemical history. It maybe a good measure of the potential attenuation of pollutants byion exchange or sorption reactions. The CEC of clays isstrongly dependent on crystalline structure, with the highshrink-swell smectite group (80 to 150 meq/100 g) having thehighest CEC and the nonswelling clays such as kaolinite thelowest (3 to 15 meq/100 g). Characterization of clay mineral-ogy can provide considerable insight into subsurface geo-chemistry.
Temperature and Pressure. Temperature and pressuredirectly influence the rate of chemical reactions. As pressureincreases, the amount of dissolved gases in solution tend toincrease. Consequently, sampling methods that allow gasesand VOCs to degas to the atmosphere at the land surface maytend to underestimate concentrations. The deeper the sam-pling, the greater the potential for errors resulting from pres-sure changes.
Microbial Activity. Virtually all ground waters containdiverse populations of microorganisms. The main limitation
to microbial growth in the subsurface is low levels of nutrientand dissolved organic carbon. Microorganisms exist that arecapable of adapting to transform many types of organic con-taminants. Unfortunately, most organic contaminants are morereadily degraded under aerobic conditions, and any contami-nant loading that adds more than traces of contaminants willrapidly deplete the available natural oxygen supply. As shownin Table 8-3, halogenated aliphatic hydrocarbons and bromi-nated methanes may be degraded under anaerobic conditions.Phenols, alkyl phenols, and chlorophenols also may be de-graded under these conditions (Wilson and McNabb, 1983).
Tetra- and trichloroethylene are readily degraded underanaerobic conditions to intermediate daughter products, in-cluding 1,2-dichloroethenes and 1,1 -dichloroethene, until vi-nyl chloride is formed. Unfortumtely, vinyl chloride is resistantto anaerobic degradation, although it readily degrades underaerobic conditions. Other anaerobic degradation sequencesthat end in relatively resistant compounds include carbontetrachloride to chloroform to methylene chloride and 1,1,1-trichloroethane to 1,1-dichloroethane to chloroethane (Woodet al., 1985).
Whether a specific contaminant will be degraded dependson geochemical conditions and on the presence of microor-ganisms that are capable of adaptation. Redox potential andwater chemistry can provide considerable insight into subsur-face microbial activity even when samples are not taken formicroorganisms. Nitrogen, ammonia, hydrogen sulfide, andmethane in ground water are all indicators of microbial activ-ity. Carbon dioxide also may indicate microbial activity;however, its presence is more difficult to interpret becausecarbon dioxide also may come from inorganic sources such ascalcium carbonate and dolomite. Section 13.2 discusses mi-crobiological transformations in the subsurface in more detail.
8.1.3 The Vadose and Saturated ZonesThe vadose and saturated zones have distinct geochemi-
cal differences that must be considered when sampling toevaluate contamination. The vadose zone is a dynamic envi-ronment with gases moving across the surface, the presence ofabundant organic matter, and solutes moving in and out of thesaturated zones. Gas transfers of interest include oxygengoing in, carbon dioxide moving out, and gases like nitrousoxide or nitrogen being generated by bacteria. Organic matteraccumulation, weathering of minerals in the soil profile toform clays, and the presence of air create a chemically reac-tive environment.
The vadose zone also is characterized by considerableheterogeneity in hydraulic conductivity. Macropores such asold root channels, animal burrows, and channels between soilstructural units allow much more rapid movement of waterand associated contaminants than the aggregated soil particles(see references in Table 8-l). These variations make represen-tative sampling of soluble contaminants in the vadose zoneextremely difficult.
106
Table 8-3. Redox Zones for Biodegradation of Organic Micropollutants
Increasing Distance from Injection Point —>
Biological Conditions
Aerobic Denitrification Sulfateheterotrophic
Methanogenesisrespiration
respiration
Organic Pollutants Transformed
Chlorinated Carbon tetrachloride Phenol C1 and C2
benzenes Bromodichloromethane CresolsEthylbenzene
HalogenatadDibromochloromethane Chiorophenols aliphatics
Styrene Bromoform Naphthalene
Source: Adapted from Bouwer and McCarty (1984)
Table 8-4. Dissolved Solids in Potabie Water -a Tentative 8.2 Background Levels and Behavior ofClassification of Abundance Chemical ConstituentsMajor Constituents (1.0 to 1000 ppm)
Sodium BicarbonateCalcium SulfateMagnesium ChlorideSilica
Secondary Constituents (0.01 to 10.0 ppm)
iron CarbonateStrontium NitratePotassium FluorideBoron
Minor Constituents (0.0001 to 0.1 ppm)
Antimony*AluminumArsenicBariumBromideCadmium*Chromium*cobaltCopperGermanium*iodide
LeadLithiumManganeseMolybdenumNickelPhosphateRubidium*SeleniumTitanium*UraniumVanadiumZinc
Trace Constituents (generally <0.001 ppm)
BerylliumBismuthCerium*CesiumGalliumGoldIridiumLanthanumNiobium*PlatinumRadium
Ruthenium*Scandium*SilverThallium*Thorium*TinTungsten*YtterbiumYttrium*Zirconium*
* Element which occupies an uncertain position in the list.
Source: Adapted from Davis and DeWiest (1966)
Interpretation of subsurface geochemical data requiressome knowledge of background levels as a baseline for evalu-ating possible contamination and the chemical behavior ofindividual constituents. Tables 8-4 and 8-5 show two classifi-cation schemes for the abundance of dissolved species inground water. The first for potable water, includes onlydissolved solids and has four classes: major (1.0 to 1,000ppm), secondary (0.1 to 10 ppm), minor (0.0001 to 0.1 ppm),and trace (generally less than 0.001 ppm). The second schemeis for highly mineralized water (>1,000 mg/L), and includesgases and organic acids. The classification of the organicacids is based on data from the petroleum-bearing Frio forma-tion in Texas (Kreitler et al., 1988). Organic acids fornonpetroleum-bearing reeks would typically be in the minorcategory.
Table 8-1 lists some sources of information on back-ground levels of trace constituents in soils and ground water.The U.S. Geological Survey is a good source of backgroundinformation on elemental composition of soils (Connor andShacklette, 1975; Ebens and Shacklette, 1982 and Shackletteet al. 197 la,b, 1973, 1974) and water (Durum and Haffty,1961; Durum et al. 1971; Ebens and Shacklette, 1982; Whiteet al., 1963). Thurman (1985), using data primarily fromLeenheer et al. (1974), reported the following median concen-trations of organic carbon in various types of aquifers: sandand gravel and limestone and sandstone - 0.7 mg/L; igneous -0.5 mg/L; oil shales - 3.0 mg/L; organically rich rechargewaters - 10.0 mg/L, and petroleum associated wastes - 100mg/L.
Table 8-1 also lists a number of general references on soiland water chemistry and sources of information on morespecific geochemical topics such as redox chemistry, soilgases, biochemical changes, and corrosion and sealing inground water. Tables 8-6 and 8-7 describe sources of informa-tion on the chemical behavior of inorganic and organic naturalconstituents and contaminants in the subsurface, respectively.
107
Table 8-5. Classlfication of Dissolved Species in Deep- Water Injection Zones
Abundance Cations Anions Gases Organic Acids
MajorSodium Chloride Carbon DioxideCalcium
AcetateBicarbonate Propionate
Magnesium Sulfate
Intermediate
Silica Nitrate Nitrogen ButyrateBarium Nitrite Hydrogen Sulfide b
Potassium Orthophosphate MethaneStrontium BromideBoron lodideIronb
Minor
Aluminun b
Manganes b
ArsenicBerylliumCadmiumChromiumcobaltCopperLeadLithiumMolybdenumNickelSeleniumZinc
Fluoride
a Abundance classification criteria (mg/L): Major: 103-105 Intermediate: 10’-103; Minor: <10’.b Of possible special significance in assessing reactivity with injected wastes.
Source: U.S. EPA (1989)
8.3 Spatial VariabilitySpatial variability of the subsurface is a result of scale
effects and physical and chemical gradients, which generallyexist both horizontally and vertically. Table 8-8 summarizestypical ranges of subsurface environmental parameters thatmay be found at a site. In general, contaminated sites have agreater range of geochemical variation for all parameters thando undisturbed sites. Spatial gradients for individual param-eters are discussed below.
8.3.1 ScaleSoil and ground-water geochemistry vary regionally pri-
marily as a function of changes in climate and geology. Animportant factor affecting ground-water chemistry is distancefrom the recharge zone. In recharge zones, ground water tendsto be less mineralized than in areas of discharge. Regional-scale changes in ground water are characterized byhydrochemical facies (Seaber 1965); dominant chemical con-stituents change with a shift in facies. Regional-scale patternsin ground-water chemistry (e.g., Back and Hanshaw, 1971, oncarbonate equilibria) may not apply on the site scale. This isparticularly true with respect to oxygen-sensitive species,because of disturbed land surface and substantial variability oflocal recharge in surficial aquifers at the site level, whichinfluences oxygen concentration.
The maximum transport distance for contaminants de-pends on the source and the medium of transport. Soil con-tamination from atmospheric sources of heavy metals (lead,zinc, cadmium) from smelters can extend from hundreds ofmeters to kilometers. Contamination from underground stor-age tanks (hydrocarbons and nonaqueous phase liquids[NAPLs]) can have a radius of influence of about 50 to 2,000m. NAPLs can migrate vertically 50 to 100 m.
8.3.2 Physical GradientsTemperature Gradients. Temperature gradients affect
mixing, reaction paths and rate, and volubility. Vertical tem-perature gradients can vary greatly, being very steep in geo-thermal areas, but a good rule of thumb is that temperatureincreases 1oF for every 50 to 60 feet of depth. Ground waterdowngradient from a landfill may exhibit temperatures 8 to12°F higher than water upgradient from a landfill.
Pressure Gradients. Vertical pressure gradients are onthe order of an atmosphere every 30 ft. Sampling mechanismsused effectively at or near the land surface may not be validwhen used at 2 to 5 atmospheres (pressure at depths in excessof 60 ft). Volatiles are in greater danger of being lost duringsampling when brought to the surface where the pressure islower.
108
Table 8-6. Sources of Information on Chemical Behavior of Natural Inorganic Constituents and Contaminants in the Subsurface
Reference Description
Velocity Gradients. Velocity gradients are a function ofpressure differences and hydraulic conductivity. Ground wa-ter may flow at a rate of 10 to 100 m a day in the vicinity ofpumping wells. Increased velocity resulting from pumpingmay have pronounced geochemical effects on ground-waterquality. Evidence of chemical zonation tends to be morepronounced when water movement is rapid in relation to therate of chemical reactions (Baedecker and Back, 1979).
8.3.3 Chemical GradientsA factor of 10 gradient in chemical concentrations over
vertical distances of less than 10 m is possible. Smith et al.(1991) observed a 27-fold increase in bacterial abundance in a9-m interval where an aquifer contained nitrate and organiccontaminants, Figure 8-1 shows changes in Eh, pH, oxygen,and hydrogen sulfide in an aquifer from its point of outcrop-ping where recharge occurs to about 24 km downdip. Dis-solved oxygen dropped to zero about 11 km from the outcrop.At the point that oxygen disappeared, Eh dropped signifi-
cantly from 400 mV to about 100 mV and continued todecline slowly to around 0 mV at 24 km. Ground-water pHshowed a general upward trend. Once reducing conditionsprevailed in the aquifer, sulfate reduction, as evidenced byhydrogen sulfide, was observed in 4 of the 10 samplingpoints.
At the site level, redox potential can vary by a factor of 5or 10 from the surface of a sand and gravel aquifer to a depthof 100 feet in the same aquifer. Figure 8-2 shows verticalchanges in Eh, oxygen, and Fe(II). As in Figure 8-1, whenoxygen drops to zero at around 30 m, Eh drops and theconcentration of reduced Fe(II) increases dramatically. Tables8-9 and 8-10 summarize examples of horizontal and verticalrcdox gradients at the site and regional scales, respectively.An uncontaminated aquifer may have a gradient in redoxpotential of 30 or 50 mV/m vertically. In contaminated situa-tions, redox potential may show a gradient of 150 mV/mhorizontally.
109
Table 8-7. Sources of lnformation on Chemical Behavior of Natural Organic Constituents and Contaminants in the Subsurface
Reference Description
Changes of a factor of four or five in pH, alkalinity, orredox potential can mean magnitude changes in many chemi-cal constituents. For example, in oxidizing conditions there isvirtually no dissolved iron in ground water. In anoxic groundwater reduced ferrous iron (Fe+2) can commonly approach 3 to4 mg/L.
Samples from large-screen intervals in ground-water moni-toring wells may give a misleading picture of subsurfacegeochemistry as a result of mixing chemically different groundwaters. For example, Cowgill (1988) sampled a 10-m screenby taking discrete grab samples from the top, middle, and
bottom of the screen interval and found that some metalconstituents differed by as much as a factor of 10.
8.4 Temporal VariabilityVariations by a factor of two to five in the concentration
of the major ionic constituents (mg/L) in ground water canoccur for no apparent reason over the course of a hydrologicyear. Very little data are available for µ/L level natural con-stituents in ground water.
Shallow aquifers are particularly sensitive to changes inpH and Eh in response to recharge events. Recharge at one
110
Table 8-8. Ranges of Geochemically Significant Physical, Biological, and Chemical Values of Natural and Disturbed Near-SurfaceGround Water
Variable Effects Natural Disturbed
time of the year may result in a set of chemical reactionsaffecting chemical composition, whereas 6 months later anentirely different set of reactions may occur. Thus, “represen-tative” concentrations of background constituents may varyseasonally.
Tables 8-11 and 8-12 summarize data on short-term(minutes to days) and long-term (seasons to decades) varia-tions of ionic constituents and several contaminants, respec-tively. In general, both short-and long-term temporal variationsare less than an order of magnitude, with nitrate sometimesshowing a greater than order of magnitude variation (13X)and Fe2+ showing up to two orders of magnitude variation(110X). Short-term variations generally result from individualground-water recharge events or well pumping and purging.Seasonal variability generally results from variations in pre-cipitation or irrigation, and multiyear trends typically resultfrom human activities such as salt-water intrusion from pump-
ing, irrigation, and fertilizer applications and nonagriculturalcontamination.
Table 8-13 shows subjective estimates of strength ofseasonality or trend in 28 chemical constituents at threedifferent sites. The Sand Ridge site, which is far removedfrom any sources of contamination, shows strong seasonaltrends in temperature and weak seasonal trends in alkalinity,calcium, and magnesium concentrations. At the Beardstownsite, monitoring wells are located up- and down-gradient froman anaerobic treatment lagoon for hog processing waste. Thecontaminated downgradient wells at the Beardstown site ex-hibited seasonality or trends for 16 constituents. The upgradientground water showed seasonality or trends for 12 constitu-ents, an intermediate value between the pristine and contami-nated ground water. For further information on references thatlist methods for analyzing time-series water quality data forseasonality and trend, see Table 7-7.
111
Figure 8-1. Horizontal gradients in uncontaminated oxidation-reduction conditions, Lincolnshire limestone (from Champ et al.,1979, after Edmunds, 1973).
112
Figure 8-2. Vertical gradients in uncontaminated oxidation-reduction conditions, Sand Ridge State Forest, Illinois (fromBarcelona et al., 1989a).
113
Table 8-9. Spatial Gradients in Subsurface Oxidation-Reduction Conditions, Site Scale
Redox Gradient
Source: Barcelona et al., 1989a
Tabie 8-10. Spatial Gradients in Subsurface Oxidation-Reduction Conditions, Large Scale
114
Table 8-11. Observations of Temporal Variations in Ground-Water Quality: Short-Term Variations
Nature of variability
*Denotes variations observed in water supply production wells, PCE = perchloroethylene, TCE = trichloroethylene, 1,2-t-DCE = 1,2 trans-dichloroethylene
Source: Barcelona et al., 1989b
115
Table 8-12. Observations of Temporal Variations in Ground- Water Quality: Long-Term Variations
Nature of variability
* Denotes variations observed in water supply production wells, PCE = perchloroethylene, TCE = trichloroethylene
Source: Barcelona et al., 1989b
116
Table 8-13. Subjective Estimate of Strength of Seasonality or Trend Ground- Water Constituents In Uncontaminated (Sand Ridgeand Upgradient Beards town) and Contaminated (Downgradient Beardstown) Sites
+ Indicates strongly seasonal. o Indicates apparent trend or possible seasonality.
TOC = VOC + NVOC; Total Organic Carbon = Volatile Organic Carbon + Nonvolatile Organic Carbon.
Source: Barcelona et al., 1989b
8.5 References Back, W. and I. Barnes. 1965. Relation of Electrochemical
Allen H., E.M. Perdue, and D. Brown (eds.). 1990. Metal Potentials and Iron Content to Ground Water Flow Pat-
Speciation in Groundwater. Lewis Publishers, Chelsea, terns. U.S. Geological Survey Professional Paper 498-C.
MI. Back, W. and B. Hanshaw. 1971. Rates of Physical and
Apps, J., L. Tsao, and O. Weres. 1988. The Chemistry of Chemical Processes in a Carbonate Aquifer. In: Non-
Waste Fluid Disposal in Deep Injection Wells. In: Second Equilibrium Concepts in Natural Water Chemistry, ACS
Berkeley Symposium on Topics in Petroleum Engineer- Adv. in Chemistry Series 106, American Chemical Soci-
ing, LBL-24337, Lawrence Berkeley Laboratory, Berke- ety, Washington, DC, pp. 77-93.ley CA, pp. 79-82.
Baedecker, MJ. and W. Back. 1979. Modem Marine Sedi-Arnold, E. M., G.W. Gee, and R.W. Nelson (eds.). 1982. ments as a Natural Analog to the Chemically Stressed
Proceedings of the Symposium on Unsaturated Flow and Environment of a Landfill. J. Hydrology 43:393-414.
Transport Modeling. NUREG/CP-0030. U.S. NuclearRegulatory Commission, Washington, DC. Barber, C., G.B. Davis, D. Briegel, and J.K. Ward. 1990.
Factors Controlling the Concentration of Methane and
Aubert, H and M. Pints. 1978. Trace Elements in Soils. Other Volatiles in Groundwater and Soil-Gas Around a
Elsevier, New York, 396 pp. Waste Site. J. Contaminant Hydrology 5:155-169.
Baas-Becking, L.G.M., I.R. Kaplan, and D. Moore. 1960. Barcelona, M.J. and J.A. Helfrich. 1986. Effects of Well
Limits of the Natural Environment in Terms of pH and Construction Materials on Ground Water Samples.
Oxidation-Reduction Potentials. J. Geology 68(3):243- Environ. Sci. Technol. 20(1 1):1179-1184.
284.
117
Barcelona, M.J. and T.G. Naymik. 1985. Dynamics of aFertilizer Contaminant Plume in Ground Water. Environ.Sci. Technol. 18(4):257-261.
Barcelona M.J., T.R. Helm, M.R. Schock, and G.K. George.1989a. Spatial and Temporal Gradients in Aquifer Oxida-tion-Reduction Conditions. Water Resources Research25(5):991-1003.
Barcelona, M.J., D,P. Lettenmaier, and M.R. Schock. 1989b.Network Design Factors for Assessing Temporal Vari-ability in Ground-Water Quality. Environmental Moni-toring and Assessment 12:149-179.
Barnes, C.J. 1989. Solute and Water Movement in Unsatur-ated Soils. Water Resources Research 25(2):38-42.
Barnes, I. and F.E. Clarke. 1969. Chemical Properties ofGround Water and Their Corrosion and EncrustationEffects on Wells. U.S. Geological Survey ProfessionalPaper 498-D.
Bar-Yosef, B., N.J. Barrow, and J. Goldschmid (eds.), 1989.Inorganic Chemicals in the Vadose Zone. Springer-Verlag,New York.
Bitton, G. and C.P. Gerba (eds.). 1984. Groundwater Pollu-tion Microbiology. Wiley-Interscience, New York.
Bohn, H.L., B.L. McNeal, and G.A. O’Connor. 1985. SoilChemistry, 2nd ed. Wiley-Interscience, New York.
Bolt, G.H. and M.G.M. Bruggenwert (eds.). 1978. Soil Chem-istry. A. Basic Elements. Elsevier, New York, 282 pp.
Bouma, J., C. Belmans, L.W. Dekker, and W.J.M. Jeurissen.1983. Assessing the Suitability of Soils with MacroPoresfor Subsurface Liquid Waste Disposal. J. Environ. Qual.12(3):305-311.
Bouwer, E.J. and P.L. McCarty. 1984. Modeling of TraceOrganics Biotransformation in the Subsurface. GroundWater 22:433-440.
Brookins, D.G. 1988. Eh-pH Diagrams for Geochemistry.Springer-Verlag, New York, 176 pp.
Callahan, M.A. et al. 1979. Water-Related EnvironmentalFate of 129 Priority Pollutants (2 Volumes). EPA/440/4-79-029a-b.
Champ, D.R., J. Gulens, and R.E. Jackson, 1979. Oxidation-Reduction Sequences in Ground Water Flow Systems.Can. J. Earth Sci. 16:12-23.
Cheng, H.H. (ed.). 1990. Pesticides in the Soil Environment:Processes, Impacts and Modeling, Soil Science Societyof America, Madison, WI, 554 pp.
Colchin, M.P., LJ. Turk, and MJ. Humenick. 1978. Sam-pling of Ground Water Baseline and Monitoring Data forIn Situ Processes. Water Resources Research Center Re-port EHE 78-01, University of Texas, Austin, TX.
Connor, J.J. and H.T. Shacklette. 1975. Background Geo-chemistry of Some Rocks, Soils, Plants, and Vegetablesin the Conterminous United States. U.S. Geological Sur-vey Professional Paper 574-F.
Copenhaver, E.D. and B.K. Wilkinson. 1979. Movement ofHazardous Substances in Soil: A Bibliography, Vol. I:Selected Metals, Municipal. EPA/600/9-79-024a (NTISPB80-113103).
Cowgill, U. 1988. Sampling Waters: The Impact of SampleVariability on Planning and Confidence Levels. In Prin-ciples of Environmental Sampling, L.H. Keith (ed.), ACSProfessional Reference Book, American Chemical Soci-ety, Washington, DC, Chapter 11.
Crist, M.A. 1974. Selenium in Waters in and Adjacent toKendrick Project, Natrona County, WY. U.S. GeologicalSurvey Water-Supply Paper 2023.
Davis, S.N. and RJ. DeWiest. 1966. Hydrogeology. JohnWiley & Sons, New York.
Diment, G.A. and K.K. Watson. 1985. Stability Analysis ofWater Movement in Unsaturated Porous Materials 3.Experimental Studies. Water Resources Research21(7):979-984.
Dragun, J. 1988. The Soil Chemistry of Hazardous Materials.Hazardous Materials Control Research Institute, SilverSpring, MD, 447+ pp.
Drever, J.I. 1989. The Geochemistry of Natural Water, 2nded. Prentice-Hall, Englewood Cliffs, NJ.
Durum, W .H. and J. Haffty. 1961. Occurrence of MinorElements in Water. U.S. Geological Survey Circular 445.
Durum, W.H., J.D. Hem, and S.G. Heidel. 1971. Reconnais-sance of Selected Minor Elements in Surface Waters ofthe United States, October 1970. U.S. Geological Surveycircular 643.
Ebens, R.J. and H.T. Shacklette. 1982. Geochemistry of SomeRocks, Mines Spoils, Stream Sediments, Soils, Plants andWaters in the Western Energy Region of the Contermi-nous United States. U.S. Geological Survey ProfessionalPaper 1238.
Eccles, L. A., J.M. Klein, and W.F. Hardt. 1978. USGS Scien-tists Bring California Water Supply into Compliance withFederal Regulations. Water Well Journal 13(2):42-45.
Edmunds, W.M. 1973. Trace Element Variations Across anOxidation-Reduction Barrier in a Limestone Aquifer. InProc. Symp. on Hydrogeochemistry and Biogeochemis-tg (Tokyo, 1970), E. Ingerson (ed.), Clarke Company,Washington, DC, pp. 500-528.
Edmunds, W.M., D.L. Miles, and J.M. Cook. 1984. A Com-parative Study of Sequential Redox Processes in ThreeBritish Aquifers. In: Hydrochemical Balances of Fresh-
118
water Systems, Int. Ass. of Hydrological Sciences Pub.No. 150, pp. 55-70.
Eriksson, E. 1985. Principles and Applications ofHydrochemistry. Chapman and Hall, New York.
Evans, D.D. and T.J. Nicholson (eds.). 1988. Flow and Trans-port Through Unsaturated Fractured Rock. AGU Geo-physical Monograph 42. American Geophysical Union,Washington, DC.
Evenson, R.E. 1965. Suitability of Irrigation Water andChanges in Ground-Water Quality in the Lompoc Subareaof the Santa Ynez River Basin, Santa Barbara, CA. U.S.Geological Survey Water-Supply Paper 1809-S.
Fairbridge, R.W. and C.W. Finkl, Jr. (eds.). 1979. TheEncylcopedia of Soil Science, Part 1: Physics, Chemistry,Biology, Fertility, and Technology. Dowden, Hutchinson& Ross, Stroudsburg, PA, 646 pp.
Faust, S.D. and O.M. Aly. 1981. Chemistry of Natural Wa-ters. Ann Arbor Science Publishers, Ann Arbor, MI.
Faust, S.D. and J.V. Hunter (eds.). 1971. Organic Compoundsin Aquatic Environments. Marcel Dekker, New York.
Feulner, A.J. and R.G. Schupp. 1963. Seasoml Changes in theChemical Quality of Shallow Ground Water in North-western Alaska. U.S. Geological Survey ProfessionalPaper 475-B, pp. B189-B191.
Förstner, U. and G.T.W. Wittmann. 1979. Metal Pollution inthe Aquatic Environment. Springer-Verlag, New York.
Fuller, W.H. 1978. Movement of Selected Metals, Asbestosand Cyanide in Soils: Applications to Waste DisposalProblems. EPA/600/2-77-020 (NTIS PB 266905).
Garrels, R.M. and C.L. Christ. 1965. Solutions, Minerals andEquilibria. Freeman, Cooper & Co., San Francisco, CA.
Gerstl, Z., Y. Chen, U. Mingelgrin, and B. Yaron (eds.). 1989.Toxic Organic Chemicals in Porous Media. Springer-Verlag, New York.
Gherini, S. A., K.V. Summers, R.K. Munson, and W.B. Mills.1988. Chemical Data for Predicting the Fate of OrganicCompounds in Water, Vol. 2: Database. EPRI EA-5818.Electric Power Research Institute, Palo Alto, CA.
Gherini, S.A., K.V. Summers, R.K. Munson, and W.B. Mills.1989. Chemical Data for Predicting the Fate of OrganicCompounds in Water, Vol. 1: Technical Basis. EPRI EA-5818. Electric Power Research Institute, Palo Alto, CA.
Ghiorse, W.C. and J.T. Wilson. 1988. Microbial Ecology ofthe Terrestrial Subsurface. Adv. Appl. Microbiol. 33:107-172.
Gibson, D.T. (ed.). 1984. Microbial Degradation of OrganicCompounds. Marcel Dekker, New York.
Goring, C.A.I. and J.W. Hamaker. 1972. Organic Chemicalsin the Soil Environment, 2 Volumes. Marcel Dekker,New York.
Handy, A.H., R.W. Mower, and G.W. Sandberg. 1969. Changesin the Chemical Quality of Ground Water in Three Areasin the Great Basin, UT. U.S. Geological Survey Profes-sional Paper 650-D, pp. D228-D234.
Hem, J.D. 1985. Study and Interpretation of the ChemicalCharacteristics of Natural Water, 3rd ed. U.S. GeologicalSurvey Water-Supply Paper 2254.
Hem, J.D. and W.H. Croper. 1959. Survey of Ferrous-FerricChemical Equilibria Redox Potentials. U.S. GeologicalSurvey Water-Supply Paper 1959-A.
Hill, D.E. and J.-Y. Parlange. 1972. Wetting Front Instabilityin Layered Soils. Soil Sci. Sco. Am. Proc. 36(5):697-702.
Howard, P. H., W.F. Jarvis, W.M. Meylan, and E.M.Mikalenko. 1991. Handbook of Environmental Degrada-tion Rates. Lewis Publishers, Chelsea, MI, 700+ pp.
Huang, P.M. and M. Schnitzer. 1986. Interactions of SoilMinerals with Natural Organics and Microbes. SSSA Sp.Pub. No. 17. Soil Science Society of America, Madison.WI, 606 pp.
Humenick, M.J., L.J. Turk, and M.P. Colchin. 1980. Method-ology for Monitoring Ground Water at Uranium SolutionMines. Ground Water 18(3):262-273.
Jackson, R.E. and R.J. Patterson. 1982. Inteq)retation of pHand Eh Trends in a Fluvial-Sand Aquifer System. WaterResources Research 18(4):1255-1268.
Jackson, R.E., R.J. Patterson, B.W. Graham, J. Bahr, D.Belanger, J. Lockwood, and M. Priddle. 1985. Contami-nant Hydrogeology of Toxic Organic Chemicals at aDisposal Site, Gloucester, Ontario. 1. Chemical Conceptsand Site Assessment. National Hydrologic Research In-stitute Paper 23. Canadian Department of Environment,Inland Water Branch, Ottawa, Ontario.
Jacobs, L.W. (ed.). 1989. Selenium in Agriculture and theEnvironment. SSSA Sp. Pub. No. 23. Soil Science Soci-ety of America, Madison, WI, 233 pp.
Jury, W.A. 1985. Spatial Variability of Soil Physical Param-eters in Solute Migration: A Critical Literature Review.EPRI EA-4228. Electric Power Research Institute, PaloAlto, CA.
Kabata-Pendias, A. and H. Pendias. 1984. Trace Elements inSoils and Plants. CRC Press, Boca Raton, FL, 336 pp.
Keely, J.F. and F. Wolf. 1983. Field Applications of ChemicalTime-Series Sampling. Ground Water Monitoring Re-view 3(4):26-33.
119
Kobayashi, H. and B.E. Rittmann. 1982. Microbial Removalof Hazardous Organic Compounds. Environ. Sci. Technol.16:170A-183A.
Kramer, C.J.M. and J.C. Duinker. 1984. Complexation ofTrace Metals in Natural Waters. Martinus Nijhoff/Dr W.Junk Publishers, Boston.
Kreitler, C.W., M.S. Akhter, and A.C.A. Donnelly. 1988.Hydrogeologic-Hydrochemical Characterization of TexasGulf Coast Formations Used for Deep-Well Injection ofChemical Wastes. Bureau of Economic Geology, Univer-sity of Texas-Austin, TX.
Langelier, W .F. 1936. The Analytical Control of Anti-Corro-sion Water Treatment. J. Am. Water Works Ass. 28:1500-1521.
Larson, T.E. and A.M. Buswell. 1942. Calcium CarbonateSaturation Index and Alkalinity Interpretations. J. AmWater Works Ass. 34:1667-1684.
Ledin, A., C. Pettersson, B. Allard, and M. Aastrup. 1989.Background Concentration Ranges of Heavy Metals inSwedish Groundwaters from Crystalline Rocks: A Re-view. Water, Air, and Soil Pollution 47:419-426. In-cludes: Cr. Cu, Zn, Cd, Pb.
Leenheer, J.A., R.L. Malcolm, P.W. McKinley, and L.A.Eccles. 1974. Occurrence of Dissolved Organic Carbonin Selected Groundwater Samples in the United States. J.Res. U.S. Geological Survey 2:361-369.
Libra, R.D., G.R. Hallberg, B.E. Hoyer, and L.G. Johnson.1986. Agricultural Impacts on Ground Water Quality:The Big Spring Basin Study, Iowa. In: Proc. of Agricul-tural Impacts on Ground Water (Omaha, NB), NationalWater Well Association, Dublin, OH, pp. 252-273.
Lindberg, D.K. and D.D. Runnells. 1984. Ground WaterRedox Reactions: An Analysis of Equilibrium State Ap-plied to Eh Measurements and Geochemical Modeling.Science 225:925-928.
Lisk, D.J. 1972. Trace Metals in Soils, Plants and Animals.Advances in Agronomy 24:267-325.
Lloyd, J. W, and J.A. Heathcote. 1985. Natural InorganicHydrochemistry in Relation to Groundwater. Oxford Uni-versity Press, New York.
Loehr, R.C., J.H. Martin, Jr., E.F. Neuhauser. 1986. SpatialVariation of Characteristics in the Zone of Incorporationat an Industrial Waste Land Treatment Site. In: Hazard-ous and Industrial Solid Waste Testing: Fourth Sympo-sium, ASTM STP 886, J.K. Petros, Jr., W.J. Lacy, andR.A. Conway (eds.), American Society for Testing andMaterials, Philadelphia, PA, pp. 285-298.
Lyman, W.J., W.F. Reehl, and D.H. Rosenblatt (eds.). 1982.Handbook of Chemical Property Estimation Methods:Environmental Behavior of Organic Compounds.McGraw-Hill, New York.
Mabey, W.R., et al. 1982. Aquatic Fate Process Data forOrganic priority Pollutants. EPA/440/4-81-O14 (NTISPB87-169090).
Maki, A.W., K.L. Dickson, and J. Cairns, Jr. (eds.). 1980.Biotransformation and Fate of Chemicals in the AquaticEnvironment. American Society for Microbiology, Wash-ington, DC.
McBride, M.A. 1989. Reactions Controlling Heavy MetalVolubility in Soils. In Advances in Soil Science, B.A.Stewart (ed.), Spnnger-Verlag, New York, Vol. 10.
McReynolds, L. 1986. Monitoring Organic Contaminants inLos Angeles San Fernando Valley Ground Water Basin.In: Proc. of the 15th Biennial Conference on GroundWater, J.J. Devries (ed.), California Water ResourcesCenter, San Diego, CA, pp. 53-60.
Miller, E.E. 1975. Physics of Swelling and Cracking Soils. J.Colloid and Interface Science 52(3):434-443.
Montgomery, J.H. 1991. Ground Water Chemical Desk Ref-erence, Vol. 2. Lewis Publishers, Chelsea, ML
Montgomery, J.H. and L.M. Welkom. 1989. Ground WaterChemicals Desk Reference. Lewis Publishers, Chelsea,MI.
Montgomery, R.H., J.C. Loftis, and J. Harris. 1988. StatisticalCharacteristics of Ground Water Quality Variability.Ground Water 25(2):176-184.
Moore, J.W. and S. Ramamoorthy. 1984a. Heavy Metal inNatural Waters: Applied Monitoring and Impact Assess-ment. Springer-Verlag, New York.
Moore, J.W. and S. Ramamoorthy. 1984b. Organic Chemicalsin Natural Waters: Applied Monitoring and Impact As-sessment. Springer-Verlag, New York.
Morel, F.M.M. 1983. Principles of Aquatic Chemistry. WileyInterscience, New York, 435+ pp.
Morril, L.G., B. Mahalum, and S.H. Mohiuddin. 1982. Or-ganic Compounds in Soils: Sorption, Degradation andPersistence. Ann Arbor Science/The Butterworth Group,Wobum, MA, 326 pp.
National Research Council Canada, 1976. Effects of Chro-mium in the Canadian Environment, NRCC Report No.15018. Ottawa, Ontario.
National Research Council Canada, 1978a. Effects of Arsenicin the Canadian Environment. NRCC Report No. 15391.Ottawa, Ontario.
National Research Council Canada, 1978b. Effects of Lead inthe Environment - 1978: Quantitative Aspects. NRCCReport No. 16736. Ottawa, Ontario.
120
National Research Council Canada, 1979a. Effects of Mer-cury in the Canadian Environment. NRCC Report No.16739. Ottawa, Ontario.
National Research Council Canada, 1979b. Effects of Cad-mium in the Canadian Environment. NRCC Report No.16743. Ottawa, Ontario.
National Research Council Canada, 1981. Effects of Nickel inthe Canadian Environment. NRCC Report No. 18568(Reprint). Ottawa, Ontario.
National Research Council Canada. 1982. Data Sheets onSelected Toxic Elements. NRCC Report No. 19252. Ot-tawa, Ontario. Includes: Sb, Ba, Be, Bi, B, Cs, Ga, Ge, In,Mo, Ag, Te, Tl, Sn (inorganic and organic), U, Zr.
Nelson, D.W., D.E. Elrick, and K.K. Tanji. 1983. ChemicalMobility and Reactivity in Soil Systems. SSSA Sp. Pub.No. 11. Soil Science Society of America, Madison, WI,262 pp.
Nicholson, R. V., J.A. Cherry and E.J. Reardon. 1983. Migra-tion of Contaminants in Ground Water at a Landfill, aCase Study: 6. Hydrogeochemistry. J. Hydrology 63:131-176.
Overcash, M.R. (ed.). 1981. Decomposition of Toxic andNontoxic Organic Compounds in Soil. Ann Arbor Sci-ence/The Butterworth Group, Wobum, MA, 375 pp.
Pagendorf, G.K. 1978. Introduction to Natural Water Chemis-try. Marcel Dekker, New York.
Perlmutter, N.M. and E. Koch. 1972. Preliminary Hydrogeo-logic Appraisal of Nitrate in Ground Water and Streams,Southern Nassau County, Long Island, NY. U.S. Geo-logical Survey Professional Paper 800-B, pp. B225-B235.
Pettyjohn, W.A. 1976. Monitoring Cyclic Fluctuations onGround-Water Quality. Ground Water 14(6):472480.
Pettyjohn, W.A. 1982. Cause and Effect of Cyclic Changes inGround-Water Quality. Ground Water Monitoring Re-view 2(1):43-49.
Ponnamperuma, F.N. 1972. The Chemistry of SubmergedSoils. Advances in Agronomy 24:29-98.
Purves, D. 1978. Trace-Element Contamination of the Envi-ronment. Elsevier, New York.
Raats, P.A.C. 1973. Unstable Wetting Fronts in Uniform andNonuniform Soils. Soil Sci. Sco. Am. Proc. 37:681-685.
Rai, D. and J.M. Zachara. 1984. Chemical Attenuation Rates,Coefficients and Constants in Leachate Migration. Vol.1: A Critical Review. Vol. 2: Annotated Bibliography,EPRI EA-3356. Electric Power Research Institute, PaloAlto, CA.
Rai, D. and J.M. Zachara. 1988. Chromium Reactions inGeologic Materials. EPRI EA-5741. Electric Power Re-search Institute, Palo Alto, CA.
Rajagopal, R. and R.L. Talcott. 1983. Patterns in Groundwa-ter Quality Selected Observations in Iowa. Environmen-tal Management 1(5):465-474.
Ransom, M.D. and N.E. Smeck. 1986. Water Table Charac-teristics and Water Chemistry of Seasonally Wet Soils ofSouthwestern Ohio. Soil Sci. Sco. Am. J. 50:1281-1290.
Rijtema, P.E., and H. Wassink (eds.). 1969. Water in theUnsaturated Zone, 2 Volumes. Studies and Reports inHydrology No. 2, UNESCO, Paris.
Ryzner, J.W. 1944. A New Index for Determining Amount ofCalcium Carbonate Scale Formed by Water. J. Am. Wa-ter Works Ass. 36:472-486.
Rose, S. and A. Long. 1988. Dissolved Oxygen Systematic inthe Tucson Basin Aquifer. Water Resources Research24(1):127-136.
Sablic, A. 1988. On the Prediction of Soil Sorption Coeffi-cients of Organic Pollutants by Molecular Topology.Environ. Sci. Technol, 21(4):358-366.
Sawhney, B.L. and K. Brown (eds.). 1989. Reactions andMovement of Organic Chemicals in Soils. SSSA SpecialPubl. No. 22. American Society of Agronomy, Madison,WI.
Schmidt, K.D. 1978. Water Quality Variations for PumpingWells. Ground Water 15(2):130-138.
Schwarzenbach, R.P., W. Giger, E. Hoehn, and J.K. Schneider.1983. Behavior of Organic Compounds During Infiltra-tion of River Water to Ground Water Field Studies.Environ. Sci. Technol. 17(8):472-479.
Seaber, P.R. 1965. Variations in ChemicaI Character of Waterin the Englishtown Formation, New Jersey. U.S. Geo-logical Survey Professional Paper 498-B.
Shacklette, H.T. et al. 1971a. Elemental Composition of Surfi-cial Materials in the Conterminous United States. U.S.Geological Survey Professional Paper 574-D. Includes:Al, Ba, Be, Bo, Ca, Ce, Cr, Co, Cu, Ga, Fe, La, Pb, Mg,Mo, Ne, Ni, Nb, P, K, Sc, Na, Sr, Ti, V, Y, Yb, Zn, Zr.
Shacklette, H.T. et al. 1971b. Mercury in the Environment—Surficial Materials of the Conterminous United States.U.S. Geological Survey Circular 644.
Shacklette, H.T. et al. 1973. Lithium in Surficial Materials ofthe Conterminous United States and Partial Data on Cad-mium. U.S. Geological Survey Circular 673.
Shacklette, H.T. et al. 1974. Selenium, Fluorine, and Arsenicin Surficial Materials of the Conterminous United States.U.S. Geological Survey Circular 692.
121
Simpson, T.W. and RL. Cunningham. 1982. The Occurrence Tryon, C.P. 1976. Ground-Water Quality Variations in Phelpsof Flow Channels in Soils. J. Environ. Qual. 11(1):29-30. County, Missouri. Ground Water 14(4):214-223.
Singer, P.C. 1973. Trace Metals and Metal Organic Interac- U.S. Environmental Protection Agency (EPA). 1989. Assess-tions in Natural Waters. Ann Arbor Science, Ann Arbor, ing the Geochemical Fate of Deep-Well-Injected Hazard-MI. ous Waste A Reference Guide. EPA/625/6-89-025a.
Singley, J.E., R.A. Pisigan, Jr., A. Ahmadi, P.O. Pisigian, and Van Beek, C.G.E.M. and J. Van Puffelen. 1988. Changes inT-Y. Lee. 1985. Corrosion and Calcium Carbonate Satu- the Chemical Composition of Drinking Water after Wellration Index in Water Distribution Systems. EPA/600/2- Infiltration in an Unconsolidated Sandy Aquifer. Water85/079 (NTIS PB85-228112) Resources Research 23(1):69-76.
Smith, R.L., R.W. Harvey, and D.R. LeBlanc. 1991. Impor-tance of Closely Spaced Vertical Sampling in DelineatingChemical and Microbiological Gradients in GroundwaterStudies. J. Contaminant Hydrology 7:285-300.
Spalding, R.F. and M.E. Exner. 1980. Areal, Vertical andTemporal Differences in Ground Water Chemistry: I.Inorganic Constituents. J. Env. Quality 9(3):466-479.
Sparks, D.L. (ed.). 1986. Soil Physical Chemistry. CRC Press,Boca Raton, FL, 320 pp.
Sparks, D.L. 1989. The Kinetics of Soil Chemical Processes.Academic Press, New York.
Sposito, G. 1984. The Surface Chemistry of Soils. OxfordUniversity Press, New York.
Sposito, G. 1989. The Chemistry of Soils. Oxford UniversityPress, New York.
Stiff, H.A. and L.E. Davis. 1952. A Method for Predicting theTendency of Oil Field Water to Deposit Calcium Carbon-ate. AIME Trans. Petroleum Div. 195:213-216
Stumm, W. and JJ. Morgan. 1981. Aquatic Chemistry, 2nded. Wiley Interscience, New York.
Tabak, H.H. et al. 1981. Biodegradability Studies with Or-ganic Priority Pollutant Compounds. J. Water PollutionControl Federation 53(10):1503-1518.
Tenorio, P. A., R.H.F. Young, and H.C. Whitehead. 1969.Identification of Return Irrigation Water in the Subsur-face. Water Resources Research Center Tech. Report 30,University of Hawaii, Honolulu, HI.
Thornton, I. (ed.). 1983. Applied Environmental Gochemis-try. Academic Press, New York.
Thorstenson, D.C., D.W. Fisher, and M.G. Croft. 1979. TheGeochemistry of the Fox Hills-Basal Hell Creek Aquiferin Southwestern North Dakota and Northwestern SouthDakota. Water Resources Research 15(6):1479-1498.
Thurman, E.M. 1985. Humic Substances in Groundwater. In:Humic Substances in Soil, Sediment, and Waten Geo-chemistry, Isolation, and Characterization, Aiken, G.R.,D.M. McKnight R.L. Wershaw, and P. MacCarthy (eds.),John Wiley & Sons, New York, pp. 87-103.
Van Cleemput, O. and A.S. El-Sebaay. 1985. Gaseous Hydro-carbons in Soil. Advances in Agronomy 38:159-181.
White, R.E. 1985. The Influence of MacroPores on the Trans-port of Dissolved and Suspended Matter Through Soil.In: Advances in Soil Science, B.A. Stewart (ed.), Springer-Verlag, New York, Vol. 3.
White, D.E., J.D. Hem, and G.A. Waring. 1963. ChemicalComposition of Subsurface Waters. U.S. Geological Sur-vey Professional Paper 440-F.
Wilson, J.T. and J.F. McNabb. 1983. Biological Transforma-tion of Organic Pollutants in Groundwater. Eos (Trans.Am. Geophysical Union) 20:997-1002.
Wilson, L.C. and J.V. Rouse. 1983. Variations in WaterQuality During Initial Pumping of Monitoring Wells.Ground Water Monitoring Review 3(1):103-109.
Wood, W.W. and R.L. Bassett. 1973. Chemical Quality ofRecharge Water as a Function of Bacterial Activity Be-neath a Recharge Basin. Eos (Trans. Am. GeophysicalUnion) 54:261.
Wood, P.R., R.F. Lang, and I.L. Payan. 1985. AnaerobicTransformation, Transport and Removal of Volatile Chlo-rinated Organics in Ground Water. In: Ground WaterQuality, C.H. Ward, W. Giger, and P.L. McCarty, (eds.),Wiley Interscience, New York, pp. 493-511.
Yamn, B., G. Dagan, and J. Goldschmid (eds.). 1984. Pollut-ants in Porous Media: The Unsaturated Zone BetweenSoil Surface and Groundwater. Springer-Verlag, NewYork.
Zimmie, T.F. and C.O. Riggs (eds.) 1979. Permeability andGroundwater Contaminant Transport. ASTM STP 746.American Society for Testing and Materials, Philadel-phia, PA.
Zehnder, A.J.B. (ed.). 1988. Biology of Anaerobic Microor-ganisms. Wiley-Interscience, New York.
Zehnder, AJ.B. and W. Stumm. 1988. Geochemistry andBiogeochemistry of Anaerobic Habitats. In: Biology ofAnaerobic Microorganisms, A.J.B. Zehnder (ed.), Wiley-Interscience, New York, pp. 1-38.
ZoBell, C.E. 1946. Studies on Redox Potential of MarineSediments. Bull. Am. Ass. Petroleum Geol. 30(4):477-513.
122
Chapter 9Geochemical Sampling of Subsurface Solids and Ground Water
J. Russell Boulding and Michael J. Barcelona
9.1 General Considerations
9.1.1 Types of MonitoringA complete sampling program for subsurface site charac-
terization includes several types of monitoring, each with itsown goal. The goal of detection monitoring is generally todetermine the presence of contaminated conditions. Unfortu-nately, drinking water wells have been among the most com-mon detective monitoring systems historically. Assessmentmonitoring seeks to identify the extent and magnitude ofcontamination. If assessment monitoring results indicate adegree of contamination requiring remediation, evaluationmonitoring is used to provide data necessary to design theremediation system. Performance monitoring is designed toevaluate the success of remediation efforts. Each stage ofmonitoring often requires the placement of additional moni-toring wells and piezometers for water level measurements.Other types of monitoring include litigation monitoring inresponse to legal actions at contaminated sites and researchmonitoring aimed at specific scientific objectives.
9.1.2 Sampling ProtocolThe goal of a sampling program with an overall monitor-
ing design is often to avoid underestimating a particularimpact either in terms of concentration or spatial distribution.Characterization of geochemical variability also is necessaryto identify potential chemical problems that may affect selec-tion and design of systems for ground-water treatment.
The field sampling protocol is often the weakest link insoil and ground-water sampling programs. Most initial effortand fiscal resources should be spent on characterizing basicsite geology and hydrology. An optimal program may call forthe placement of three or four times as many piezometers thanwells for water-quality sampling. Initial selection of locationsfor sampling must be based on a good preliminary character-ization of the geology and hydrogeology of the site. This mayrequire spending more of the available financial resources onhydrogeologic characterization than on chemical samplingand analyses. Additional sample locations should be added asunderstanding of the site evolves.
As discussed in Section 7.1.4, sample location and fre-quency are among the most critical aspects of sampling be-cause sample collection and sample analysis sometimes can
give entirely erroneous results even when approached andexecuted carefully. Good vertical and horizontal resolution ofhydrogeologic conditions are essential before choosing samplelocations. Uncertainty, hydrogeologic variability, and quality-assurance decision-making need to be addressed from theinitial design stage. Later, an effective sampling strategy andwritten protocols should be prepared. These measures canimprove confidence in subsequent chemical results. Docu-mentation of all sampling procedures is essential, becausedata collected for a particular purpose may end up being usedand interpreted for other objectives.
Sampling protocols should leave room for evolutionarydevelopment of the network design. For example, samplingexperiments can be used to determine spatial correlation forsolid samples. A large number of surface samples or split-spoon samples can be collected but it may only be necessaryto analyze a certain percentage (20 to 50 percent) to achieveadequate spatial coverage. If the initial sample groups indicatesufficient sampling resolution, the other samples need not beanalyzed. If necessary, additional samples can be analyzeduntil geostatistical analysis indicates an adequate samplingintensity has been achieved. Samples should not be thrownaway if there is any possibility that somebody may use themin the future and if adequate preservative measures are fea-sible.
Many references thoroughly cover one or more aspects ofdeveloping a sampling program and protocol for subsurfacesolids and/or ground-water. Table 9-1 lists and summarizes 29of these major reference sources. Rehm et al. (1985) probablycontains the most comprehensive review of the literature onmethods for hydrogeologic investigations up to 1985. The restof this chapter focuses on developments since that time,although particularly relevant pre-1985 references are occa-sionally cited.
Table 9-2 lists sources of information on four aspects ofgeneral sample design: (1) general theory, (2) soil sampling,(3) vadose zone sampling, and (4) ground-water sampling.General aspects of selecting sample location, frequency, andsize are discussed in the remainder of this section. Section 9.2reviews sampling of subsurface solids and vadose zone waterfurther, and Section 9.3 covers sampling of ground water.
123
Table 9-2. Sources of Information on General SampleDesign
Topic References
9.1.3 Sample LocationTable 9-3 summarizes major types of sampling designs
and when they should be used for characterizing subsurfacegeochemistry. In general, haphazard water-quality or solidsampling is not an appropriate approach to designing sam-pling for subsurface geochemical characterization, even thoughprofessional judgment alone, is probably the most frequentlyused method for siting ground-water monitoring wells. Figure9-1 illustrates some two-dimensional probability samplingdesigns for spatial characterization. The trends or patterns thatcommonly exist in subsurface contamination mean that simplerandom sampling will not give as accurate an estimate ofpopulation characteristics as stratified random and grid sam-pling designs.
Hydrogeologic characterization, initially using surfacegeophysical techniques followed by piezometers and prelimi-nary well tests to estimate the distribution of hydrogeologicparameters, should come before the location and installationof monitor wells, Good vertical resolution is essential insampling to characterize distribution of oxidized and reducedspecies, contaminants, and microbiota. Achieving this resolu-tion requires more discrete well completions with short screens.
Table 9-3. Summary of Sampling Designs and Condltions forTheir Use
Type of Sampling Conditions When the SamplingDesign Design is Useful
Source: Adapted from Gilbert (1987)
In most cases, 5-ft to 1.5-m well screens should give adequatevertical resolution.
The spatial distribution of contamination is a major con-cern with sampling solids. The intensity and number of samplesdepends on the nonsampling variance, which is the variabilityof concentration that is unrelated to sampling procedures.Spatial structure determines the distance between samplesthat have essentially the same concentration, called the rangeof correlation, to avoid oversarnpling (see Section 7.3.2).
125
Figure 9-1. Some two-dimensional probability sampling designsfor sampling over space (from Gilbert, 1987). SeeTable 9-3 for description of when these samplingdesigns are useful.
There are two broad designs for soil sampling: (1) gridsin which samples are taken from a matrix of squares orquadrants at a site, and (2) transects in which samples aretaken at specified intervals along a line. Figure 7-10 in Chap-ter 7 shows contours of lead concentration in soil drawn fromgrid sampling. Grids presume an aerial or dispersed source ofsome kind, and transects presume a preferential source. Forexample, Starks et al. (1986) established sampling transectswhere the length was proportional to the frequency withwhich wind blew in a particular direction to characterizemetal contamination from a smelter near Palmerton, Pennsyl-vania (see Figures 9-2 and 9-3). Flatman (1986) describes useof geostatistics for determining sampling intensity. Grids canbe used to estimate short-range correlation. Transects alongthe path of ground-water or contaminant movement providethe best way to look at long-range correlation. The combina-tion of the two strategies coupled with the initial analysis ofselected solid samples at alternate grid or transect locationscan be quite effective.
The combined strategy also can avoid the potential col-lection of redundant information. Using geostatistical analysistechniques of successive analytical subsets minimizes the
Figure 9-2. Palmerton wind rose, 1978-1979 data (from Starkset al., 1986).
number of samples actually analyzed. Transects could be bothparallel and perpendicular to the axis of ground-water move-ment, along with some random samples from a grid, as shownin Figure 9-1 (f). Analysis of samples from four equally spacedlocations on a transect or grid within the area of influence is agood starting point to estimate the distance of short-rangecorrelation. For soils, at least 5 percent of sampling pointsshould be duplicated to help determine the sampling variabil-ity, so it can be analyzed with geostatistical techniques. Atleast 5 percent of the samples should be split as well.
Preliminary efforts that can help guide the location ofinitial wells for ground-water sampling include (1) surfacegeophysical techniques for mapping extent of contaminantplumes; (2) soil gas sampling techniques; (3) Hydropunch®sampling; and (4) selective sampling of piezometers for simpleconstituents such as pH, conductance, and possibly iron ordissolved oxygen concentrations.
Soil gas monitoring (see Section 5.2.5) and Hydropunch®ground-water sampling (see Section 9.3.4) probably give thebest pictures of short-range variability in three dimensions.Sampling from monitoring wells usually gives some sort ofintegrated value depending on the relative width or thicknessof the hydrogeologic formation of interest and the length ofthe screen. Disadvantages of soil gas concentrations include(1) lack of the ability to directly calibrate, because all valuesare relative and difficult to reproduce, (2) decontamination,and (3) short circuiting of air from the surface, which can .distort results.
126
Figure 9-3. Sample pattern for the initial Palmerton survey (1”= 4250’) (from Starks et al., 1986).
9.1.4 Sampling FrequencyTable 9-4 shows estimated ranges of sampling frequency
in months necessary to maintain information loss at less than10 percent for selected types of chemical parameters. Formany chemical constituents, quarterly sampling is adequatefor characterizing short-term (i.e., monthly to 1 or 2 years)changes over time. For some reactive constituents such as ironand other redox-sensitive constituents, bimonthly samplingmay be required.
With intermittent sources of contamination, it is espe-cially important that the frequency of sampling not allow acontaminant to be missed. Barcelona et al. (1985a) describe aprocedure for estimating sample frequency to detect contami-nant plumes based on the type of plume (slug, intermittent, orcontinuous) and hydrogeologic parameters of gradient, hy-draulic conductivity, effective porosity, and distance alongthe flow path. Figure 9-4 shows a nomograph that can be usedwhen these parameters are known. When the contaminantplume is a slug source or intermittent, sampling frequency
should probably be more frequent to ensure that the plume isnot missed. One advantage to the slow movement of groundwater is that if there are questions about a sample, resampling,a week later will yield roughly the same ground water.
Precise estimation of optimum sampling frequency isprobably impractical for most investigations. For example,Bell and DeLong (1988) found that tetrachloroethylene atconcentrations of 200 to 300 µg/L exhibited variations of afactor of one or two over the course of a year. Their workpoints out that data collection may be required for 4 years ormore in order to estimate the optimal sampling frequency todetermine seasonal variability. Therefore, it is important toselect sampling frequency on the basis of an initial period ofmonitoring in the context of the duration of the program.
It should not be necessary to sample all monitoring wellsevery time samples are taken. Sampling selected wells candevelop a preliminary picture, with additional follow-up sam-pling at additional wells rounding out the picture.
127
9.1.5 Sample Type and SizeSoil sampling must take into account fractures in earth
materials and the fact that the subsurface is heterogeneous (asscales ranging from centimeters to meters). If the soil hasobvious fractures and channels in the subsurface, samplingshould sample both affected and apparently nonfractured ar-eas for comparison. Soil sample quantities of less than 100 gtend to be unrepresentative even of the areas where the sampleis taken. In the laboratory, the sample can be mixed, subsampledprior to analysis.
Compositing samples is often beneficial for soil investi-gations. However, where volatile constituents are involved,compositing is not practical because handling samples in theair for compositing will result in the loss of the contaminant.One way to get around this problem is to take two or threesamples within each identifiable core segment and put theminto a sealed glass vial immediately after sampling. In thiscase, a volume of methyl alcohol in the sealed vial canimprove volatile recovery and expedite analysis. However, itis possible that the sampling variance from potential loss of
Table 9-4. Estimated Ranges of Sampling Frequency (inMonths) to Maintain information Loss at <10% forSelected Types of Chemical Parameters
Figure 9-4. Sampling frequency nomograph (from Barcelona et al., 1985).
128
volatiles involved in handling the sample may far exceed theactual variability in the field.
Williams (1989) compares the results of one 500-g sample,twenty 25-g composite samples, and ten 50-g compositesamples. He found that a single 25-g composite sample wasthe most accurate and precise technique for determining ra-dium concentrations in contaminated surface soil. Initial soilsamples of 100-g are about the best size for such compositeanalyses.
9.1.6 Vadose versus Saturated ZoneCareful sampling of gases and solids in the vadose zone
can provide information for better locating monitoring wellsin the saturated zone. The mass of contaminant, or at least themost persistent contaminants, are often associated with thesolids.
9.2 Sampling Subsurface Solids and VadoseZone Water
9.2.1 Analyte SelectionHalocarbons, chlorinated hydrocarbon solvents (e.g., tetra-
and trichloroethylene), and fuel constituents (e.g., toluene,benzene, ethyl benzene, and xylenes) are amenable to prelimi-nary delineation by soil gas methods. Soil gas samples forcarbon dioxide, methane, oxygen, and nitrogen can provideadditional insights into subsurface chemistry, particularly mi-crobiological activity.
In addition to examining chemical constituents, analyz-ing solid samples for grain-size distribution and correlationwith permeability can be helpful.
9.2.2 Sampling Devices and TechniquesTable 9-5 lists sources of information on sampling soil
and vadose zone solids, water, and gases.
Simple techniques for surface sampling of soils includethe hand auger, brace and bit, and posthole diggers. The mostcommonly used core sampling devices are split spoons orShelby tubes that provide a continuous or driven core duringdrilling operations. Sampling continuously or ahead of hol-low-stem- drilling augers are good ways to obtain uncontami-nated and minimally disturbed soil samples. Section 3.1provides some additional discussion of these sampling meth-ods for obtaining information on subsurface stratigraphy.Where the surface layer of soil is known to be heavilycontaminated, as with sites involving smelters and uraniummills, the surface should be scraped away before sampling atlower levels so the sampler is not contaminated as it passesthrough the contaminated surface.
Figure 9-5 shows a soil core sampling apparatus de-scribed by Myers et al. (1989) that can obtain undisturbedcores for laboratory leaching experiments. A variety of sam-plers are available that advance in front of an auger. The betterdevices have a plunger or cylinder that maintains a partialvacuum to prevent the soil material from falling out of the
core (Munch and Killey, 1985; and Zapico et al., 1987). Thisvacuum is particularly important for saturated sands thatsimply flow out of normal sampling tubes. Figure 9-6 shows amodified wireline piston design for sampling cohesionlesssediments and Figure 9-7 shows how this device can be usedto take samples through a hollow-stem auger. In careful use,the more sophisticated devices can achieve a 50 percent corerecovery. Heaving sands create special problems. Filling theauger with water sometimes helps prevent clogging fromheaving sands by maintaining hydrostatic pressure.
Suction lysimeters can be used to sample pore water inthe vadose zone. Extremely variable transmissive propertiesof surface soils make accurate interpretation of soil pm waterconcentrations very difficult. Virtually all of the water move-ment and associated contaminant transport may occur in about5 percent of the soil profile. The zone of sampling influencewith a suction lysimeter is about 10 or 20 cm for a 24-hourperiod (Morrison and Lowery, 1990). In some instances,longer suction sampling periods may extend the influence to50 cm.
Sampling for microbiological parameters requires boththe collection of soil samples and the paring of any outsideportion that may have been in contact with the samplingapparatus. This operation should be done before placing thesamples in sterile glass vials.
Table 9-5. Sources of Information on Sampling Soil andVadose Zone Solids, Solutes, and Gases
Topic References
129
Figure 9-5. Undisturbed soil core sampling apparatus (from Myers et aI., 1989).
Soil gas sampling generally involves driving a probe intothe subsurface. Typically, the probes are driven by hand orwith some kind of pneumatic or electric hammer. Soil gas isobtained by applying a vacuum that brings the soil gas into thevicinity of the tip of the probe. Samples are collected influorocarbon bags or syringes and analyzed on site or in alaboratory. Analysis techniques can be as simple and nonspe-cific as a hand-held gas survey meter, and as detailed andspecific as an analytic laboratory’s instrumentation allows.Mobile laboratories provide an intermediate level of analyti-cal detail; they provide semiquantitative results with precisionon the order of plus or minus 100 percent. At least 5 percent ofair-filled porosity is required to pull a vacuum to obtainsamples.
Table 9-6 summarizes information from 14 soil gas in-vestigations. Soil gas samples for areal characterization areusually taken at a uniform depth with the specific depth
typically from 1 to 6 ft below the surface, although Glaccumet al. (1983) sampled immediately above the water table,which was as much as 10 m deep. Vertical profiles mayprovide additional insight into contaminant behavior. Figure9-8 shows six types of vertical concentration profiles thatdevelop under different subsurface conditions. Special careshould be taken to identify any underground utility lines toavoid accidental puncture with the probe. Buried sewers orproduct lines may be the source of soil gas contamination, andother utility lines may provide a directional component tocontamination (Marrin and Thompson, 1987). See Section5.2.5 for additional discussion of soil gas sampling methods.
9.3 Sampling Ground WaterFigure 9-9 shows a generalized flow diagram of ground-
water sampling steps, and Table 9-7 lists additional sources ofinformation on various aspects of ground-water sampling.
130
Figure 9-8. Modified wireline piston design (from Armstrong et al., 1988).
References that provide good general coverage of ground-water sampling include Barcelona et al. (1985), GeoTrans(1989), Gillham et al. (1983), Rehm et al. (1985), and Scalf etal. (1981).
9.3.1 Analyte SelectionTables 9-8 and 9-9 identify chemical constituents of
interest for various types of ground-water monitoring activi-ties. In hazardous waste site investigations, regulations willgenerally specify the contaminants to be tested for. Focusingon priority pollutants alone, however, may not provide acomplete geochemical picture of contamination. The sourceof contamination may involve a large number of individualcontaminants that are not classified as hazardous. Also, deter-mination of redox-sensitive constituents (dissolved oxygenand dissolved iron), pH, and conductance, may provide valu-able insight into subsurface contaminant geochemistry.
Highly mineralized ground water, commonly encoun-tered in formations being evaluated for deep-well injection ofwastes, may require more complete analyses for natural inor-ganic and organic constituents. Table 9-10 lists analyticalresults for ground-water samples from four deep-well injec-
tion sites and the Frio formation in Texas (which has receivedmore deep-well injected wastes than any other formation inthe United States). Not all of the studies analyzed the sameconstituents in all samples, but an examination of this tablemay give some guidance for analyte selection.
Iron, an inexpensive constituent to determine analyti-cally, can be used as an indicator of redox conditions andpotential mobility for heavy metals. Dissolved gases are ex-cellent indicators of redox conditions and microbial activity.For example, Leenheer and Malcolm (1973) analyzed for H2
N2, CH4, CO2, and H2S in serial samples from a well throughwhich a plume of deep-well injected wastes passed. They usedchanges in the relative percentages of the different gases asindicators of changing microbial activity. Malcolm andLeenheer (1973) suggest that separate analysis for dissolvedorganic carbon DOC) and suspended organic carbon (SOC)can yield more complete analytical results.
Calcium carbonate and iron/manganese concentrationsare especially important parameters if remediation involvesair stripping. Air-stripping towers are particularly susceptibleto fouling by calcium carbonate and metal oxide precipitates.
131
Figure 9-7. Wireline piston core barrel sampling operation (from Zapico et al., 1987).
When trichloroethylene (TCE) is involved as a contami-nant, it is important also to analyze for biotransformationproducts (e.g., 1,2-dichloroethene, 1,1-dichloroethene, andvinyl chloride), The vinyl chloride monomer metabolic prod-uct is more toxic than TCE and resistant to degradation underanaerobic conditions.
Battista and Connelly (1989) found that inorganic param-eters such as chemical oxygen demand, specific conductance,chloride, alkalinity, and hardness were reasonably good indi-cators for predicting VOC contamination from landfills. Whenthe inorganic parameters were detected above backgroundlevels in monitoring wells, VOCs were also usually present.Out of 49 ground-water samples at landfill sites in Wisconsin,VOCs and elevated inorganic parameters were detected atabout the same frequency in 20 (41 percent), elevated inor-ganic parameters without VOCs were detected in 11 (22
percent), and VOCs without elevated inorganic parameterswere detected in 3 wells (6 percent). The remaining 15 wellsin the study showed neither VOCs nor elevated inorganicparameters.
9.3.2 Well DevelopmentWell development, which involves the removal of fines
created during the drilling process, is essential before sam-pling begins. Pumping rates generally used for well develop-ment are 5 to 10 gpm. Bailing, swabbing, pumping, andair-lifting are common methods used for development. Table4-2 compares the advantages and disadvantages of the mostcommonly used well development techniques. Air develop-ment may increase the possibility of environmental exposureto workers at the surface where volatiles are involved.
132
Table 9-6. Soil Gas Sampling Case Studies
(A) Homogeneous Porous Material with Sufficient Air-filled Porosity(B) Impermeable Subsurface Layer (e.g., Clay or Perched Water)(C) Impermeable Surface Layer (e.g., Pavement)(D) Zone of High Microbiological Activity (Circles and Wavy Lines Indicate Different Compounds)(E) VOC Source in the Vadose Zone
Figure 9-8. Soil-gas ooncentrations under a variety of conditions (from Marrin and Kerfoot, 1988).
133
Denotes samples that should be filtered to determine dissolvad constituents. Filtration should be accomplished preferably with in-line filtersand pump pressure or by N2 pressure methods. Samples for dissolved gases or volatile organics should not be filtered. in instances wherewell development procedures do not allow for turbidity free samples and may bias analytical results. split samples should be spiked withstandards before filtration. Both spiked samples and regular samples should be analyzed to determine recoveries from both types of handling.
Denotes analytical determinations that should be made in the field. See Puls and Barcelona (1989).
Figure 9-9. Generalized flow diagram of ground-water sampling steps (adapted from Barcelona et al., 1985).
134
Table 9-7. Sources of Information on Various Aspects of Ground-Water sampling
Topic References
Analyte identification Barcelona (1983), Battista and Connelly (1989), Spruill (1988)
Well construction
Purging
Sample devices
Chemical changes
Comparisons
Packer samplers
Hydropunch
Discrete point
Sampling procedures
Decontamination
Metals
Volatiles
Oil-water mixtures
Field measurement
Aller et al. (1989), Cohen and Rabold (1988), Hackett (1988), Palmer at al. (1987), Pennino (1988), Perry and Hart(1985), Sykes et al. (1986). See also Table 9-12
Barcelona and Helrich (1986), Barber and Davis (1987), Gibs and Imbrigiotta (1990), Herzog et al. (1988),Oliveros et al. (1988), Palmer et al. (1987), Panko and 8arth (1988), Pennino (1988), Robbins (1989), Robinand Gillham (1987), Smith et al. (1988), Unwin and Maltby (1988). See also, Table 9-11
Barcelona et al. (1985b), Barker and Dickhout (1988), Holm et al. (1988), Pannino (1988), Stolzenburg andNichols (1985), Schalla et al. (1988), Rose and Long (1988)
Barcelona et al. (1984), 8arcelona et al. (1988), Pohlmann and Hess (1988), Nielsen and Yeates (1985)
Anderson (1979)
Cordry (1986), Edge and Cordry (1989)
McPherson and Pankow (1988)
Meade and Ellis (1985), Mickam et al. (1989)
Puls and Barcelona (1989)
Barker and Dickhout (1988), Schalla et al. (1988), Unwin and Maltby (1988)
Borst (1987)
Garner (1988), Garske and Schock (1986), Holm et al. (1987)
Table 9-8. Chemical Constituents of Interest in Ground- Water Monitoring
135
Table 9-9. Recommended Analytical Parameters for Detective Monitoring
Source: Barcelona et al. (1985a)
9.3.3 PurgingPurging involves removing stagnant water from a moni-
toring well before taking a sample for analysis. Once monitor-ing well locations have been selected, inadequate purgingprocedures probably account for more sampling error thanany other step of the sampling process (see Table 7-2), Thereis no universally correct purge volume. Monitoring wellsfinished in materials of widely varying hydraulic conductivitymay require different purge volumes since chemical constitu-ents are likely to migrate towards a pumped well at differentrates (Gibs and Imbriggiotta, 1990).
Recommended rules of thumb such as using 3 to 5volumes (Fenn et al., 1977) should be treated only as a startingpoint. Consistent estimation of purge volume requires know-ing (1) well yield, determined from a slug or pumping testand (2) the stagnant volumes of both the well casing and thesand pack. Pumping rates for purging (i.e., generally 1 to 5gpm) should be below the rate used for development (gener-ally 5 to 20 gpm) to avoid well damage, which could inducethe migration of fines into the screened interval. The length oftime required to remove the stagnant water at the plannedpumping rate can readily be estimated from the well yield andstagnant volume calculations. In most cases, it is important tominimize the purge requirement to avoid dealing with largevolumes of contaminated water.
Monitoring pH, conductance, and temperature duringpurge pumping can provide indications of background chem-istry. After the stagnant water has been removed or isolated,these indicators should continue to be monitored until theyreach a consistent end point (no upward or downward trend)before sampling. Even after stagnant water has been removed,some constituents may show increasing or decreasing trends.Table 9-11 summarizes the results of observations in sevenstudies where concentrations were measured as a function ofwell volumes pumped, Increasing or decreasing concentrationtrends usually will reach a constant level, although volatileconstituents may show considerable variance (see below).
The site- and constituent-specific nature of concentrationtrends with purging is evident from the fact that bicarbonate,nitrate, and specific conductance exhibited both increasingand decreasing trends in different studies.
In studies by Smith et al. (1988), measurements of tri-chloroethylene ranged from O to 250 mm as a function ofpurge volumes from O to 25 volumes inside the well casingand the volume inside the sand pack. After two to three wellvolumes, trichloroethylene concentrations reached 100 to 125µg/L. Five to ten well volumes averaged 150 to 175 µg/L, soat least five well volumes was required to obtain samples nearthe average. Concentrations dropped quickly after purgingstopped, and purging a day later yielded similar results. Thiseffect is probably the result of volatile losses from the stag-nant water.
9.3.4 Well Construction and Sampling DevicesThe Hydropunch® collects one-time ground-water samples
in unconsolidated material (see Figure 9-10). It is driven intothe soil and when the bottom of the probe is at least 5 ft belowthe water table, the outer cylinder can be pulled back exposinga perforated stainless steel sample entry barrel covered witheither a nylon or polyethylene filter material (see Figure 9-1la). Hydrostatic pressure forces ground water that is rela-tively free of turbidity into the sample compartment (seeFigure 9-1lb). About 6 to 10 water samples of between 500and 1,000 mL each often can be obtained in this manner if nomajor problems occur. Geologic materials that can be auguredor sampled with a split spoon are suitable for sampling withthe Hydropunch®.
All decisions preceding monitoring well construction andsample collection have to be quality-assured and documented.Among the references listed in Table 9-1, Barcelona et al.(1983) and Aller et al. (1989) focus primarily on monitoringwell design and construction. Screen slot size selection shouldbe justified, preferably by a quick sieve analysis in the field.
136
Table 9-10. Chemical Constituents of Formation Waters Analyzed in Studies Related to Deep-Well Injection
Wilmington Pensacola Belle Glade Marshall FrioConstituent NC FL FL IL TX
137
Table 9-11. Observed Trends in Measured Concentrationswith Well Volumes Pumped
References Indicating Trend inMeasured Concentration
Parameter Increasing Constant Decreasing
ArsenicAlkalinityAmmoniumBicarbonateBoronCadmiumCalciumCarbonateChlorideChromiumCopperDOCHardnessIronFluorideMagnesiumManganeseNitratepHPotassiumSeleniumSodiumSpecific
ConductanceSulfateTDSTemperatureZinc
2
5
1 131,3,51,3,5
2,321,2,3,5
3,571,3,5
52,5
3123,51,2,3,5211,32,3,5
12
2
1,6,7
2
7
54,61,5
1,2,3,51,3 1,3,4,6
1,2,51,3
51
7
Figure 9-10. HydroPunch® schematic (from Edge and Cordry,1989).
2,5
References:1 Chapin (1981)2 Gibb et al. (1981)3 SIawson et al. (1982)4 Schmidt (1982)5 Marsh and Lloyd (1980)6 Nightingale and Bianchi (1980)7 Keith et al. (1982)
Source: Adapted from Rehm et al. (1985)
Common rigid well-casing materials that might be used in-clude polyvinyl chloride, stainless steel, and polytetrafluoro-ethylene. Table 4-3 summarizes the advantages anddisadvantages of these and other well casing and screenmaterials. Figure 9-12 shows a sample decision tree for theselection of rigid materials for casing.
organics to some extent. The impact of sorptive losses orleaching contamination can be expected to be different withaged materials than with the virgin material.
Figure 9-13 summarizes recommended sampling meth-ods for various parameters for detective monitoring programsand Figure 9-14 shows a decision tree for selection of sam-pling mechanisms. With sampling devices, pressure changesand the loss of volatiles are the main concern. Samplingwithin 30 feet of the surface involves little pressure changeand most samplers may be expected to perform similarly forvolatile and gas-sensitive species. Sampling at depths in ex-cess of 60 ft (two or more atmospheres) can be expected toyield differences in sampling devices. Teflon®, polypropylene,and polyethylene are the best tubing materials for sampling.Polyvinyl chloride, Tygon®, and silicone rubber tubing should
Table 9-12 summarizes data on the leaching and sorptioncharacteristics of well casing materials. Stainless steel may bethe best overall metal easing and screening material, but it isstill susceptible to microbiological corrosion. In most in-stances, casing and screen materials should last for at least 30years without corrosion closing down the effective area of thescreen. Teflon® and polyvinyl chloride have structural prob-lems for emplacement in deeper holes. All common easingand tubing materials may be expected to sorb hydrophobic
138
Figure 9-11. HydroPunch® sampling operation.(a) HydroPunch® is pushed Into target ares with acone-penetrometer rig. (b) Once exposed, groundwater flows through the intake tube and into thesample chamber (from Edge and Cordry, 1989).
be avoided, particularly if VOCs are involved, due to docu-mented major losses of these species. Dedicated samplingdevices can greatly increase the cost efficiency of takingsamples.
A bladder pump is a cylinder with an internal bladder thatcan be compressed and expanded under the influence of a gas.The squeezing and release of pressure can be controlled with afrequency that will give virtually pulseless flow. Bladderpumps operate on air or nitrogen and air compressors areavailable that are relatively easy to move around for supplyingthem. Bladder pumps provide precise flow rates at givenoperating pressures and frequencies of pressure/release. Theyhave worked reliably with continuous submersion in the samewell for extended periods. Any malfunction such as a leak in abladder pump is immediately apparent because it will stopworking. Repair in the field is also relatively easily accom-plished. Bladder pumps are best adapted for purging small-diameter monitoring wells (less than 4 in.) and their depthrange is limited to about 450 ft.
Bailers are commonly used sampling devices, but have anumber of disadvantages compared to bladder pumps. Thebasic performance difficulties with bailers are that virtually allindividuals bail differently, and in-line determination of pH,conductance, temperature, and dissolved oxygen are not pos-sible. Also, sample transfer can be inconsistent, which createsvariability that shouldn’t be in the sample data set. Anothermajor problem with bailers is the difficulty of determiningwhere a sample is actually retrieved. In this case, bailers maymalfunction without the operator knowing when the checkvalve actually sealed. Bailers or grab samplers can minimizevolatilization losses, and are probably the best way to sampleNAPLs at the water table surface. Newer bailer designs allowfiltration in the field and transfer of volatile samples withoutcontact with the atmosphere, but to not address the problem ofinconsistency in bailing. Bailers should not be used for purg-ing because all they do is homogenize the volume within thewell bore.
Electric submersibles can be useful for purging large-diameter deep wells with high volume purging requirements,particularly when flow rates can be controlled. They may notbe good for sampling unless the flow rate for sampling can becontrolled or diverted from the main pumping stream. Ingeneral the accuracy is poor for gas-sensitive parameters, notonly volatile organics, but also oxygen and carbon dioxide.
Suction pumps, venturi mechanism pumps, and somegrab-driven mechanisms create turbulence that puts negativepressure on the sample for volatiles. Flow is generally diffi-cult to control, particularly to obtain preferable low flows(i.e., 100 mL to 2 L/min) for sampling.
Sampling devices and sample handling should be ex-ecuted so as to minimize temperature and pressure changes.Reproducible flow rates and freedom from operator-inducederrors tend to yield the most precise results. Figure 9-15contains a matrix rating the suitability of different chemicaldevices for different chemical constituents. Figure 9-16 ratessuitability of 12 devices (described in Table 9-13) for use with12 types of ground-water parameters. If VOCs are sampled
139
Figure 9-12. Decision tree for recommended welI-casing/screen materials. Adapted by Barcelona and Gibb (1888) from Barcelonaet al. (1985).
140
Table 9-12. Effects of Well Caslng Materiail on TraceConcentration in Well Water
Parameter Leaches From Adsorbed By
Arsenic ABS*Cadmium Steel, Galvanized 2
Chromium SteelCopper SteelDissolved OrganicCarbon ABS, PVC**iron Steel, Galvanized 1
Manganese Steel, Galvanizedz
Total Organic Carbon ABS, PVC ABSZinc Steel, Galvanized 1 ABSLithium ABSMercury ABSMolybdenum ABSSelenium ABS
Source: Adapted from Houghton and Berger (1984) Acrylonitrile-butadiene-styrene copolvmer Polyvinyl chloride I2
Suggested by data from Gibb et al. (1981)Barcelona et al. (1983)
effectively, the results with most other constituents may beexpected to be reproducible and accurate.
Studies of sampling errors associated with the samplingmechanisms alone have found that bladder pumps and bailerscome out with sampling error consistently less than theanalytical error. However, most comparisons of bladder pumpsand bailers have been conducted at shallow lifts. At depths upto 200 or 300 m, bladder pumps are probably superior.Vacuum devices, peristaltic and suction pumps, on the otherhand, yield a sampling error on about the same order as thelevel of sorptive losses or handling errors. Where sensitiveconstituents are involved, bladder pumps and bailers are themost frequently used devices. A bladder pump is probablythe best overall sampling device and will probably provide50 to 100 percent better recovery and far better precision forvolatiles than a bailer.
Once samples are collected, procedures for the handlingand preservation of samples should be carefully followed tominimize errors from this stage of the sampling process.Table 9-14 summarizes recommended sample handling andpreservation procedures for a comprehensive detective moni-toring program.
141
*Materials in order of preference include: Teflon® (T); stainless steel (S): PVC, polypropylene, polyethylene (P); borosilicate glass (G); othermaterials: silicone, polycarbonate, mild steal, etc. (0)
Figure 9-13. Recommended sample collection methods for detective monitoring programs (from Barcelona et al., 1985).
142
Figure 9-14. Decision tree for recommended purge and sampling mechanlsm. Adapted by Barcelona and Gibb (1988) fromBarcelona et al. (1985).
143
Figure 9-15. Matrix of sensitive chemical constituents and various sampling mechanisms (from Barcelona et al., 1985).
144
* Sampling devices on this chart are divided into two categones: (1) portable devices for sampling existing monitoringwells; and (2) in situ monitoring devices (often multilevel) that are permanently installed. Sampling device constructionmaterials (including tubing, haul lines, etc.) should be evaluated for suitability in analyzing specific ground waterparameters. It is assumed on this chart that existing monitoring wells are properly installed and constructed of materialssuitable for detection of the parameters of interest. See references for additional information.
† Sample delvivery rates and volumes are average ranges based on typical field conditions. Actual delivery rates areafunction of diameter of monitoring installation, size and capacity of sampling device, hydrogeologic conditions, and depthto sampling point. For all devices, delivery rate should be carefully controlled to prevent aeration or degassing of thesample.
Ž Indicates device is generally suitable for application (assuming device is cleaned and operated properly and isconstructed of suitable materials).
Figure 9-16. Generalized ground-water sampling device matrix (from Pohlmann and Hess, 1988).
Table 9-13. Description of Ground-Water Sampling Devices and Construction Materials Commonly Used in Ground- WaterMonitoring (see also Figure 8-16)
Sample Device Description
Open bailer Open top. Bottom sealed or fitted with foot valve. Available in wide range of rigid materials.Point-source bailer Check valve at both top and bottom. Valves are opened by cable operated from ground surface.
Available in wide range of rigid materials.Syringe sampler Sample container is pressurized or evacuated and lowered into sampling installation. Opening the
container and/or releasing the pressure allows sample to enter the device. Materials may includestainless steel 316, Teflon®, polyethylene, glass.
Gear-drive pump Electric motor rotates a set of Teflon gears, which drives the sample up the discharge line. Constructedof stainless steel 304, Teflon, and Viton®.
Bladder pump Flexible bladder within device has check valves at each end. Gas from ground surface is cycled betweenbladder and sampler wall, forcing sample to enter bladder and then be driven up the discharge line. Gasdoes not contact sample. Materials may include stainless steel 316, Teflon, Viton, polyvinyl chloride (PVC),silicone, Neoprene®, polycarbonate, Delrin®.
Helical-rotor pump Water sample is forced up discharge line by electrically driven rotor-stator assembly. Materials may includestainless steel 304, ethylene propylene rubber (EPDM), Teflon, Viton, polypropylene.
Gas-driven pump Piston is driven up and down by gas pressure controlled from the surface. Gas does not contact sample.Materials may include stainless steel 304, Teflon, Delrin, polypropylene, Viton, acrylic, polyethylene.
Centrifugal pump Electrically driven rotating impeller accelerates water within the pump body, building up pressure and forcingthe sample up discharge line. Commonly constructed of stainless steel, rubber, and brass.
Peristaltic pump Self priming vacuum pump is operated at ground surface and is attached to tubing, which is lowered to thedesired sampilng depth. Sample contacts vacuum. Materials may include Tygon®, silicone, Viton, Neoprene,rubber, Teflon.
Gas-lift devices Gas emitted from gas line at desired depth forces sample to surface through sampling installation.Another method utilizes gas to reduce effective specific gravity of water, causing it to rise.Wide variety of materials available for tubing.
Gas-drive devices Positive gas pressure applied to water within device’s sample chamber forces sample to surface. Materialsmay include polyethylene, brass, nylon, aluminum oxide, PVC, polypropylene.
Pneumatic in situ device generally utilizes the same operating principles as syringe samplers: a pressurized or evacuatedsample container is lowered to the sampling port and opened, allowing the sample to enter. Materials mayinclude PVC, stainless steel, polypropylene, Teflon.
Source: Pohlmann and Hess (1988)
146
Table 9-14. Recommended Sample Handling and Preservation Procedures for a Detective-Monitoring Program
VolumeParameters
MaximumRequired (mL) Containers Preservation Holding
(Type) 1 Samplea (Material) Method Period
a It is assumed that at each site, for each sampling date, replicates, a field blank, and standards must be taken at equal volume to those of thesamples.
b Temperature correction must be made for reliable reporting. Variations greater than ±10% may result from a longer holding period.c In the event that HNOa cannot be used because of shipping restrictions, the sample should be refrigerated to 4°C, shipped immediately, and
acidified on receipt at the laboratory. Container should be rinsed with 1:1 HNOa and included with sample.d 28-day holding time if samples are preserved (acidified).e Longer holding times in EPA (1986b).f Filtration is not recommended for samples intended to indicate the mobile substance lead. See Puls and Barcelona (1989) for more specific
recommendations for filtration procedures involving samples for dissolved species.Note: T= Teflon; S = stainless steal; P = PVC, polypropylene, polyethylene;G = borosilicate glass.Source: Adapted from Scalf et al. (1981) and U.S. EPA (1986b)
147
9.4 ReferencesAller, L., T.W. Bennett G. Hackett, Rebecca J. Petty, J.H.
Lehr, H. Sedoris, D.M. Nielsen. 1989. Handbook ofSuggested Practices for the Design and Installation ofGround-Water Monitoring Wells. EPA/600/4-89/034(NTIS PB90-159807). Also published in NWWA/EPAseries, National Water Well Association, Dublin, OH.
Andersen, L.J. 1983. Sampling Techniques of Groundwaterfrom Water Wells. In: Proc. UNESCO Symp. Methodsand Instrumentation for the Investigation of GroundwaterSystems, Committee for Hydrological Research, CHO-TNO, The Hague, The Netherlands, pp. 521-527.
Armstrong, J.M., W. Korreck, L.E. Leach, R.M. Powell, S.V.Vandegrift, and J.T. Wilson. 1988. Bioremediation of aFuel Spill: Evaluation of Techniques for Preliminary SiteCharacterization. In: Proc. 5th NWWA/API Conf. Petro-leum Hydrocarbons and Organic Chemicals in GroundWater-Prevention, Detection and Restoration, NationalWater Well Association, Dublin, OH, pp. 931-948.
Barber, C. and G.B. Davis. 1987. Representative Sampling ofGround Water from Short-Screened Boreholes. GroundWater 25(5):581-587.
Barcelona, M.J. 1983. Chemical Problems in Ground-WaterMonitoring Programs. In: Proc. 3rd Nat. Symp. on Aqui-fer Restoration and Ground Water Monitoring, NationalWater Well Association, Dublin, OH, pp. 263-271.
Barcelona, M.J. and J.P. Gibb. 1988. Development of Effec-tive Ground-Water Sampling Protocols. In: Ground-Wa-ter Contamination: Field Methods, A.G. Collins and A.LJohnson (eds.), ASTM STP 963, American Society forTesting and Materials, Philadelphia, PA, pp. 17-26.
Barcelona, M.J. and J.A. Helfrich. 1986. Well Constructionand Purging Effects on Ground-Water Samples. Environ.Sci. Technol. 20:1179-1184.
Barcelona, M.J., J.P. Gibb, and R.A. Miller. 1983. A Guide tothe Selection of Materials for Monitoring Well Construc-tion and Ground-Water Sampling. ISWS Contract Report327. Illinois State Water Survey, Champaign, IL.
Barcelona, M.J., J.A. Helfrich, E.E. Garske, and J.P. Gibb.1984. A Laboratory Evaluation of Ground Water Sam-pling Mechanisms. Ground Water Monitoring Review4(2):32-41.
Barcelona, M.J., J.P. Gibb, J.A. Helfrich, and E.E. Garske,1985a. Practical Guide for Ground-Water Sampling. EPA/600/2-85/104 (NTIS PB86-137304). Also published asISWS Contract Report 374, Illinois State Water Survev,
Barcelona, MJ., J.A. Helfrich, and E.E. Garske. 1988. Verifi-cation of Sampling Methods and Selection of Materialsfor Ground-Water Contamination Studies. In Ground-Water Contamination: Field Methods, A.G. Collins andA.I. Johnson (eds.), ASTM STP 963, American Societyfor Testing and Materials, Philadelphia, PA, pp. 221-231.
Barcelona, M.J., D.P. Lettenmaier, and M.R. Schock. 1989.Network Design Factors for Assessing Temporal Vari-ability in Ground-Water Quality. Environmental Moni-toring and Assessment 12:149-179.
Barker, J.F. and R. Dickhout. 1988. An Evaluation of SomeSystems for Sampling Gas-Charged Ground Water forVolatile Organic Analysis. Ground Water MonitoringReview 8(4):112-119.
Barth, D.S. and T.H. Starks. 1985. Sediment Sampling Qual-ity Assurance User’s Guide. EPA/600/4-85&8 (NTISPB85-233542).
Barth, D.S., B.J. Mason, and T.H. Starks. 1989. Soil SamplingQuality Assurance User’s Guide, 2nd ed. EPA/600/8-89/046 (NTIS PB89-189864), 225 pp.
Battista, J.R. and J.P. Connelly. 1989. VOC Contamination atSelected Wisconsin Landfills-Sampling Results andPolicy Implications. WDNR PUBL-W-094 89. Wiscon-sin Department of Natural Resources, Madison, WI.
Bell, H.F. and H.P. DeLong. 1988. Data Characteristics GroundWater Monitoring’s Catch 22. ACS Abstracts 28(2):20-24.
Berg, E.L. 982. Handbook for Sampling and Sample Preser-vation of Water and Wastewater. EPA/600/4-82-029(NTIS PB83-124503).
Borst, M. 1987. Sampling Oil-Water Mixtures at OHMSEIT.EPA/600/2-87/073 (NTIS PB88-102892).
Brown, K.W. 1987. Efficiency of Soil Core and Soil-PoreWater Sampling Systems. EPA/OO/2-86/083 (NTIS PB87-106100).
Bumb, A.C., C.R. McKee, R.B. Evans, and L.A. Eccles. 1988.Design of Lysimeter Leak Detector Networks for SurfaceImpoundments and Landfills. Ground Water MonitoringReview 8(2):102-114.
Cameron, R.E., et al. 1966. Sampling and Handling of DesertSoils. Technical Report No. 32-908. Jet Propulsion Labo-ratory, California Institute of Technology, Pasadena, CA,37 pp.
Champaign, IL. . Chapin, R.I. 1981. Short-Term Variations, Sampling Tech-. - .
niques and Accuracy of Analysis of the Concentrations ofBarcelona, MJ., J.A. Helfrich, and E.E. Garske. 1985b. Sam- Nitrate in Produced Municipal Ground Waters—North
pling Tubing Effects on Groundwater Samples. Anal. Texas. M.A. Thesis, University of Texas, Austin.Chem. 57:460-464.
148
Close, M.E. 1989. Effect of Serial Correction on GroundWater Quality Sampling Frequency. Water ResourcesBulletin 25(3):507-515.
Cochran, W.G. 1977. Sampling Techniques, 3rd ed. JohnWiley & Sons, New York.
Cohen, R.M. and R.R. Rabid. 1988. Simulation of Samplingand Hydraulic Tests to Assess a Hybrid Monitoring WellDesign. Ground Water Monitoring Review 8(1):51-59.
Collins, A.G. and A.I. Johnson (eds.). 1988. Ground-WaterContamination: Field Methods. ASTM STP 963. Ameri-can Society for Testing and Materials, Philadelphia, PA.
Cordry, K. 1986. Ground Water Sampling Without Wells. In:Proc. 6th Nat. Symp. on Aquifer Restoration and GroundWater Monitoring, National Water Well Association,Dublin, OH, pp. 266-271.
Deming, W.E. 1950. Some Theory of Sampling. John Wiley& Sons, New York. (available as a Dover reprint, 1966).
Devitt, D. A., R.B. Evans, W.A. Jury, T.H. Starks, and B.
Di,
Eklund. 1987. Soil Gas Sensing for Detection and Map-ping of Volatile Organics. EPA/600/8-87/036 (NTISPB87-228516).
H.J., B.B. Trangmar, and R.A. Kemp. 1989. Use ofGeostatistics in Designing Sampling Strategies for SoilSurvey. Soil Sci. Soc. Am. J.53:1163-1167.
Diem, D.A., B.E. Ross and H.B. Kerfoot. 1988. Field Evalua-tion of a Soil-Gas Analysis Method for Detection ofSubsurface Diesel Fuel Contamination. In: Proc. 2nd Nat.Outdoor Action Conf. on Aquifer Restoration, GroundWater Monitoring and Geophysical Methods, NationalWater Well Association, Dublin, OH, pp. 1015-1031.
Dunlap, W.J., J.F. McNabb, M.R. Scalf, and R.L. Cosby.1977. Sampling for Organic Chemicals and Microorgan-isms in the Subsurface. EPA/600/2-77/176 (NTIS PB272679).
Edge, R.W. and K. Cordry. 1989. The Hydropunch: An InSitu Sampling Tool for Collecting Ground Water fromUnconsolidated Sediments Ground Water MonitoringReview 9(3):177-183.
Everett, L.G., K.D. Schmidt, R.M. Tinlin, and D.K. Todd.1976. Monitoring Groundwater Quality: Methods andcosts. EPA/600/4-76/023 (NTIS PB257133).
Everett, L.G. and L.G. McMillion. 1985. Operational Rangesfor Suction Lysimeters. Ground Water Monitoring Re-view 5(3):5 1-60.
Everett, L.G., L.G. Wilson, and L.G. McMillion. 1982. Va-dose Zone Monitoring Concepts for Hazardous WasteSites. Ground Water 20(3):312-324,
Everett, L.G., L.G. Wilson, and E.W. Hoylman. 1983. VadoseZone Monitoring for Hazardous Waste Sites. EPA/600/X-83/064 (NTIS PB84-212752). Also published in 1984by Noyes Data Corporation, Park Ridge, NJ.
Everett, N. G., E.W. Hoylman, L.G. Wilson, and L.G.McMillion. 1984. Constraints and Categories of VadoseZone Monitoring Devices. Ground Water MonitoringReview 4(1):26-31.
Fenn, D., E. Cocozza, J. Isbister, O. Braids, B. Yare, and P.Roux. 1977. Procedures Manual for Ground Water Moni-toring at Solid Waste Disposal Facilities.EPA/530/SW611(NTIS PB84-174820).
Flatman, G.T. 1986. Design of Soil Sampling Programs:Statistical Considerations. In: Quality Control in Reme-dial Site Investigation: Hazardous and Industrial SolidWaste Testing, 5th volume, C.L. Perket (ed.), ASTMSTP 925, American Society for Testing and Materials,Philadelphia, PA, pp. 43-56.
Ford, P.J., P.J. Turina, and D.E. Seely. 1984. Characterizationof Hazardous Waste Sites—A Methods Manual: Vol. II.Available Sampling Methods, 2nd ed. EPA/600/4-84-076(NTIS PB85-521596).
Garner, S. 1988. Making the Most of Field-Measurable GroundWater Quality Parameters. Ground Water MonitoringReview 8(3):60-66.
Garske, E.E. and M.R. Schock. 1986. An Inexpensive Flow-Through Ceil and Measurement System for MonitoringSelected Chemical Parameters in Ground Water. GroundWater Monitoring Review 6(3):78-84.
GeoTrans, 1989. Groundwater Monitoring Manual for theElectric Utility Industry. Edison Electric Institute, Wash-ington, DC.
Gibb, J.P, R.M. Schuller, and R.A. Griffh. 1981. Proceduresfor the Collection of Representative Water Quality Datafrom Monitoring Wells. ISWS/lGS Cooperative GroundWater Report 7. Illinois State Water Survey, Champaign,IL.
Gibs, J. and T.E. Imbrigiotta. 1990. Well-Purging Criteria forSampling Purgeable Organic Compounds. Ground Water28(1):68-78.
Gilbert, R.O. 1987. Statistical Methods for EnvironmentalPollution Monitoring. Van Nostrand Reinhold, New York.
Gillham, R.W., M.J.L. Robin, J.F. Barker and J.A. Cherry.1983. Groundwater Monitoring and Sample Bias. APIPublication 4367. American Petroleum Institute, Wash-ington, DC.
Glaccum, R., M. Noel, R. Evans, and L. McMillion. 1983.Correlation of Geophysical and Organic Vapor AnalyzerData Over a Conductive Plume Containing Volatile Or-ganics. In Proc. 3rd Nat. Symp. on Aquifer Restoration
149
and Ground Water Monitoring, National Water WellAssociation, Dublin, OH, pp. 421-427.
Goodwin, C.D. and R.M. Burger. 1989. Inexpensive but Ef-fective Soil Gas Screening—A Case Study. In: Proc. 3rdNat. Outdoor Action Conf. on Aquifer Restoration, GroundWater Monitoring and Geophysical Methods, NationalWater Well Association, Dublin, OH, pp. 263-271.
Goolsby, D. A. 1972. Geochemical Effects and Movement ofInjected Industrial Waste in a Limestone Aquifer. In:Symposium on Underground Waste Management andEnvironmental Implications (Houston, TX), T.D. Cook(ed.), Am. Ass. Petr. Geol. Mere. 18, pp. 355-368.
Hackett G. 1988. Drilling and Constructing Monitoring Wellswith Hollow-Stem Augers: Part 2. Monitoring Well In-stallation. Ground Water Monitoring Review 8(1)60-68.
Hansen, M.H. W.N. Hurwitz, and W.G. Madow. 1953. SampleSurvey Methods and Theory, Vols. I and H. John Wiley& Sons, New York.
Hendricks, W.A. 1956. The Mathematical Theory of Sam-pling. Scarecrow Press, New Brunswick, NJ.
Henzog, B.L., S.-F. Chou, J.R. Valkenberg, and R.A. Griffin.1988. Changes in Volatile Organic Chemical Concentra-tions After Purging Slowly Recovering Wells. GroundWater Monitoring Review 8(4):93-99.
Hoffman, S.J. 1986. Soil Sampling. In: Exploration Geo-chemistry, Design and Interpretation of Soil Surveys.Reviews in Economic Geology 3:39-71.
Holden, P.W. 1984. Primer on Well Water Sampling forVolatile Organic Compounds. Water Resources ResearchCenter, University of Arizona, Tucson, AZ.
Holm, T.R., G.K. George, and M.J. Barcelona. 1987. Fluoro-metric Determination of Hydrogen Peroxide in Ground-water. Anal. Chem. 59:582-586.
Holm, T.R., G.K. George and M.J. Barcelona. 1988. OxygenTransfer through Flexible Tubing and Its Effects onGround Water Sampling Results. Ground Water Moni-toring Review 8(3):83-89.
Houghton, R.L. and M.E. Berger. 1984. Effects of Well-Casing Composition and Sampling Method on ApparentQuality of Ground Water. In: Proc. 4th Nat. Symp. onAquifer Restoration and Ground Water Monitoring, Na-tional Water Well Association, Dublin, OH, pp. 203-213.
Howsam, P. (ed.). 1990. Proceedings of International Ground-water Engineering Conference on Water Wells: Monitor-ing, Maintenance, and Rehabilitation. Chapman and Hall,London, 422 pp.
Hsueh, Y-W. and R. Rajagopal. 1988. Modeling GroundWater Quality Sampling Decisions. Ground Water Moni-toring Review 8(4):121-133.
Hughes, J.P. and D.P. Lettenmaier. 1981. Data Requirementsfor Kriging: Estimation and Network Design. Water Re-sources Research 17:1641-1650.
Jessen, R.J. 1978. Statistical Survey Techniques. John Wiley& Sons, New York.
Johnson, T.M. and K. Cartwright. 1980. Monitoring ofLeachate Migration in the Unsaturated Zone in the Vicin-ity of Sanitary Landfills. ISGS Circular 514. Illinois StateGeological Survey, Champaign, IL.
Kaufman, M. I., D.A. Goolsby, and G.L. Faulkner. 1973.Injection of Acidic Industrial Waste into a Saline Carbon-ate Aquifer Geochemical Aspects. In: Symposium onUnderground Waste Management and Artificial Recharge,J. Braunstein (ed.), Int. Ass. of Hydrological SciencesPub. No. 110, pp. 526-551.
Keith, S.J., L.G. Wilson, H.R. Fitch, and D.M. Esposito.1982. Sources of Spatial-Temporal Variability in Ground-Water Quality Data and Methods of Control: Case Studyof the Cortaro Monitoring Program, Arizona. In: Proc.2nd Nat. Symp. on Aquifer Restoration and GroundWater Monitoring, National Water Well Association,Dublin, OH, pp. 217-227.
Kerfoot, H.B. and J. Soderberg. 1988. Three-DimensionalCharacterization of a Vadose Zone Plume in IrregularlyInterbedded Silt and Sand Deposits. In: Proc. 2nd Nat.Outdoor Action Conf. on Aquifer Restoration, GroundWater Monitoring and Geophysical Methods, NationalWater Well Association, Dublin, OH, pp. 1071-1087.
Kerfoot, H.B. and L.J. Barrows. 1987. Soil Gas Measurementfor Detection of Subsurface Organic Contamination. EPA/600/2-87/027 (NTIS PB87-174884).
Kirschner, Jr., F.E. and G.L. Bloomsburg. 1988. Vadose ZoneMonitoring: An Early Warning System. Ground WaterMonitoring Review 8(2):49-50.
Kish, L. 1965. Survey Sampling. John Wiley & Sons, NewYork.
Kreitler, C.W., M.S. Akhter, and A.C.A. Donnelly. 1988,Hydrologic-Hydrochemical Characterization of TexasGulf Coast Formations Used for Deep-Well Injection ofChemical Wastes. University of Texas at Austin, Bureauof Economic Geology.
Leenheer, J.A. and R.L. Malcolm. 1973. Case History ofSubsurface Waste Injection of an Industrial Organic Waste.In: Symposium on Underground Waste Management andArtificial Recharge, J. Braunsteirt (ed.), Int, Ass. of Hy-drological Sciences Pub. No 110, pp. 565-579.
Leenheer, J. A., R.L. Malcolm, and W.R. White. 1976, Physi-cal, Chemical and Biological Aspects of Subsurface Or-ganic Waste Injection near Wilmington, North Carolina.U.S, Geological Survey Professional Paper 987.
150
Litaor, M.I. 1988. Review of Soil Solution Samplers. WaterResources Research 24(5):727-733.
Loaiciga, H.A. 1989. An Optimization Approach for Ground-water Quality Monitoring Network Design. Water Re-sources Research 25(8):1771-1782.
Loehr, R.C., J.H. Martin, Jr., E.F. Neuhauser. 1986. SpatialVariation of Characteristics in the Zone of Incorporationat an Industrial Waste Land Treatment Site. In: Hazard-ous and Industrial Solid Waste Testing: Fourth Sympsium,ASTM STP 886, J.K. Petros, Jr., W.J. Lacy, and R.A.Conway (eds.), American Society for Testing and Materi-als, Philadelphia, PA, pp. 285-297.
Mackay, D. and W.Y. Shiu. 1981. A Critical Review ofHenry’s Law Constants for Chemicals of EnvironmentalInterest. J. Phy. Chem. Ref. Data 10(4):1 175-1199.
MacPhemon. Jr., J.R. and J.F. Pankow. 1988. A Discrete Point
tions in the Vadose Zone. Ground Water MonitoringReview 8(2):125-136.
McNichols, R.J. and C.B. Davis. 1988. Statistical Issues andProblems in Ground Water Detection Monitoring at Haz-ardous Waste Facilities. Ground Water Monitoring Re-view 8(4):135-150.
Meade, J.P. and W.D. Ellis. 1985. Decontamination Tech-niques for Mobile Response Equipment Used at WasteSites (State-of-the-Art Survey). EPA/600/2-85/105 (NTISPB85-247021).
Mickam, J.T., R. Bellandi, and E.C. Tifft, Jr. 1989. EquipmentDecontamination Procedures for Ground Water and Va-dose Zone Monitoring Programs: Status and Prospects.Ground Water Monitoring Review 9(2):100-121.
Morrison, R.D. and B. Lowery. 1990. Sampling Radius of aPorous Cup Sampler Experimental Results. Ground Wa-
Sampler for Ground Water Monitoring Wells. Ground ter 28(2):262-267.Water Monitoring Review 8(3):160-164.
Malcolm, R.L. and J.A. Leenheer. 1973. The Usefulness ofOrganic Carbon Parameters in Water Quality Investiga-tions. In: Proc. of the Inst. of Env. Sciences 1973 AnnualMeeting (Anaheim, CA), pp. 336-340. Available fromJ.A. Leenheer, USGS MS 408, Box 25046, Federal Cen-ter, Denver CO, 80225.
Marrin, D.L. 1987. Soil Gas Sampling Strategies: Deep vs.Shallow Aquifers. In: Proc. 1st Nat. Outdoor ActionConf. on Aquifer Restoration, Ground Water Monitoringand Geophysical Methods, National Water Well Associa-tion, Dublin, OH, pp. 437-454.
Marrin, D.L. and W.B. Kerfoot. 1988. Soil Gas SurveyingTechniques. Environ. Sci. Technol. 22(7)-740-745.
Marrin, D.L. and G.M. Thompson. 1984. Remote Detectionof Volatile Organic Contaminants in Ground Water ViaShallow Soil Gas Sampling. In: Proc. 1st NWWA/APIConf. Petroleum Hydrocarbons and Organic Chemicalsin Ground Water-Prevention, Detection and Restora-tion, National Water Well Association, Dublin, OH, pp.172-187.
Marrin, D.L. and G.M. Thompson. 1987. Gaseous Behaviorof TCE Overlying a Contaminated Aquifer. Ground Wa-ter 25(1):21-27.
Marsh, J.M. and J.W. Lloyd. 1980. Details of HydrochemicalVariations in Flowing Wells. Ground Water 18(4):366-373.
Mason, B.J. 1983. Preparation of Soil Sampling Protocol:Techniques and Strategies. EPA/600/4-83-020 (NTISPB83-206979).
Munch, J.H. and R.W.D. Killey. 1985. Equipment and Meth-odology for Sampling and Testing Cohesionless Sedi-ments. Ground Water Monitoring Review 5(1):38-42.
Myers, R.G., C.W. Swallow, and D.E. Kissel. 1989. A Methodto Secure, Leach and Incubate Undisturbed Soil Cores.Soil Sci. Soc. Am. J. 53:467-471.
National Council of the Paper Industry for Air and StreamImprovement (NCASI). 1982. A Guide to GroundwaterSampling. Technical Bulletin 362, NCASI, New York,NY.
New Jersey Department of Environmental Protection (NJDEP).1988. Field Sampling Procedures Manual. HazardousWaste Program, NJDEP, Trenton, NJ.
Newman, W., J.M. Armstrong, and M. Ettenhofer. 1988. AnImproved Soil Gas Survey Method Using AdsorbentTubes for Sample Collection. In: Proc. 2nd Nat. OutdoorAction Conf. on Aquifer Restoration, Ground Water Moni-toring and Geophysical Methods, National Water WellAssociation, Dublin, OH, pp. 1033-1049.
Niaki, S. and J.A. Broscious. 1986. Underground Tank LeakDetection Methods: A State-of-the-Art Review. EPA/600/2-86/001 (NTIS PB86-137155).
Nielsen, D.M. (ed.). 1991. Practical Handbook of GroundWater Monitoring. Lewis Publishers, Chelsea, MI (pub-lished in cooperation with National Water Well Associa-tion, Dublin, OH), 717 pp.
Nielsen, D.M. and G.L. Yeates. 1985. A Comparison ofSampling Mechanisms Available for Small-DiameterGround Water Monitoring Wells. Ground Water Moni-toring Review 5(2):83-99.
McKee, C.R. and A.C. Bumb. 1988. A Three-DimensionalAnalytical Model to Aid in Selecting Monitoring Loca-
151
Nightingale, H.I. and W.C. Bianchi. 1979. Influence of WellWater Quality Variability on Sampling Decisions andMonitoring. Water Resources Bulletin 15(5):1394-1407.
Nightingale, H.I. and W.C. Bianchi. 1980. Well Water Qual-ity Changes Correlated with Well Pumping Time andAquifer Parameters-Fresno, California. Ground Water18:275-280.
Olivems, J.P., D.A. Vroblesky, and M.M. Lorah. 1988. In-creasing Purging Efficiency Through the Use of Inflat-able Packers. In: Proc. 2nd Nat. Outdoor Action Conf. onAquifer Restoration, Ground Water Monitoring and Geo-physical Methods, National Water Well Association,Dublin, OH, pp. 457-469.
Oudjik, G. and K. Mujica. 1989. Handbook for Identification,Location and Investigation of Pollution Sources Affect:ing Ground Water. National Water Well Association,Dublin, OH.
Palmer, C.D., J.F. Keely, and W. Fish. 1987. Potential forSolute Retardation on Monitoring Well Sand Packs andIts Effect on Purging Requirements for Ground WaterSampling. Ground Water Monitoring Review 7(2):40-47.
Panko, A.W. and P. Barth. 1988. Chemical Stability Prior toGround-Water Sampling: A Review of Current WellPurging Methods. In: Ground-Water Contamination: FieldMethods, A.G. Collins and A.I. Johnson (eds.), ASTMSTP 963, American Society for Testing and Materials,Philadelphia, PA, pp. 232-239
Pennino, J.D. 1988. There’s No Such Things as a Representa-tive Ground Water Sample. Ground Water MonitoringReview 8(3):4-9.
Perry, C.A. and R.J. Hart. 1985. Installation of ObservationWells on Hazardous Waste Sites in Kansas Using aHollow-Stem Auger. Ground Water Monitoring Review5(4):70-73.
Peterson, R.G. and L.D. Calvin. 1986. Sampling. In: Methodsof Soil Analysis, Part I—Physical and MineralogicalMethods, 2nd ed., A. Klute (ed.), ASA Monograph No. 9,American Society of Agronomy, Madison, WI, pp. 33-51.
Pitard, F.F. 1989. Pierre Gy’s Sampling Theory and Samplingpractice, Vol. I Heterogeneity and Sampling, Vol. IISampling Correctness and Sampling. CRC Press, BocaRaton, FL.
Pohhnann, K.F. and J.W. Hess. 1988. Generalized GroundWater Sampling Device Matrix. Ground Water Monitor-ing Review 8(4):82-84.
Puls, R.W, and M.J. Barcelona. 1989. Ground Water Sam-pling for Metals Analyses. Superfund Ground WaterIssue Paper. EPA/548/4-89/001.
Rajagopal, R. 1986. The Effect of Sampling Frequency onGround Water Quality Characterization. Ground WaterMonitoring Review 6(4)65-73.
Rajagopal, R. and L.R. Williams. 1989. Economics of SampleCompositing as a Screening Tool in Ground Water Qual-ity Monitoring. Ground Water Monitoring Review9(1):186-192.
Rehm, B.W., T.R. Stolzenburg, D.G. Nichols. 1985. FieldMeasurement Methods for Hydrogeologic Investigations:A Critical Review of the Literature. EPRI AA-4301.Electric Power Research Institute, Palo Alto, CA.
Robbins, G.A. 1989. Influence of Using Purged and PartiallyPenetrating Monitoring Wells on Contaminant Detection.Mapping, and Modeling. Ground Water 27(2):155 -162.’
Robin, M.J.L. and R.W. Gillham. 1987. Field Evaluation ofWell Purging Procedures. Ground Water Monitoring Re-view 7 4):85-93.
Rose, S.R. and A. Long. 1988. Monitoring Dissolved Oxygenin Ground Water Some Basic Considerations. GroundWater Monitoring Review 8(1):93-97.
Roy W.R., S.C. Mravik, I.G. Krapac, D.R. Dickerson, andR.A. Griggin. 1989. Geochemical Interactions of Hazard-ous Wastes with Geological Formations in Deep-WellSystems. ISGS Environmental Geology Notes 130. Illi-nois State Geological Survey, Champaign, IL.
Scalf M.R., J.F. McNabb, W.J. Dunlap, R.L. Cosby, and J.Fryberger. 1981. Manual of Ground-Water Quality Sam-pling Procedures. EPA/600/2-81/160, (NTIS PB82-103045). Also published in NWWA/EPA Series, NationalWater Well Association, Dublin OH.
Schalla, R., D.A. Myers, M.A. Simmons, J.M. Thomas, andA.P. Toste. 1988. The Sensitivity of Four MonitoringWell Sampling Systems to Low Concentrations of ThreeVolatile Organics. Ground Water Monitoring Review8(3):90-96.
Schmidt, K.D. 1982. How Representative are Water SamplesCollected from Wells? In: Proc. 2nd Nat. Symp. onAquifer Restoration and Ground Water Monitoring, Na-tional Water Well Association, Dublin, OH, pp. 117-128.
Sgambat, J.P. and R. Stedinger. 1981. Confidence in Ground-Water Monitoring. Ground Water Monitoring Reviewl(Spring):62-69.
Shangraw, T.C., D.P. Michaud, T.M. Murphy. 1988. Verifica-tion of the Utility of a Photovac Gas Chromatography forConduct of Soil Gas Surveys. In: Proc. 2nd Nat. OutdoorAction Conf. on Aquifer Restoration, Ground Water Moni-toring and Geophysical Methods, National Water WellAssociation, Dublin, OH, pp. 1089-1108.
Simmons, M.S. 1991. Hazardous Waste Measurements. LewisPublishers, Chelsea, MI, 315 pp.
152
Sisk, S.W. 1981. NEIC Manual for Groundwater/SubsurfaceInvestigations at Hazardous Waste Sites. EPA/330/9-81-002 (NTIS PB82-103755).
Slater, J.P. and F.R. McLaren. 1983. Sampling and Analysisof Soil for Volatile Organic Compounds. In: Proc.NWWA/EPA Conf. on Characterization and Monitoringof the Vadose (Unsaturated) Zone (Las Vegas, NV),National Water Well Association, Dublin, OH, pp. 623-658.
Slawson, Jr., G.C., K.E. Kelly, and L.G. Everett. 1982. Evalu-ation of Ground Water Pumping and Bailing Methods—Application in the Oil Shale Industry. Ground WaterMonitoring Review 2(3):27-32.
Slonim, M.J. 1957. Sampling in a Nutshell. J. Am. StatisticalAss. 52:143-161.
Smith, J. S., D.P. Steele, M.J. Malley, and M.A. Bryant. 1988.Groundwater Sampling. In: Principles of EnvironmentalSampling, L.H. Keith (ed.), ACS Professional ReferenceBook, American Chemical Society, Washington, DC, pp.255-260.
Sophocleous, M., J.E. Paschetto, and R.A. Olea. 1982. Ground-Water Network Design for Northwest Kansas, Using theTheory of Regionalized Variables. Ground Water 20:48-58.
Spittler, T.M., W.S Clifford, and L.G. Fitch. 1985. A NewMethod for Detection of Organic Vapors in the VadoseZone. In: Proc. 2nd Conf. on Characterization and Moni-toring in the Vadose (Unsaturated) Zone, National WaterWell Association, Dublin, OH, pp. 236-246.
Spruill, T.B. 1988. Use of Total Organic Carbon as an Indica-tor of Contamination from an Oil Refinery, South-Cen-tral Kansas. Ground Water Monitoring Review 8(3)76-82.
Starks, T.H, K.W. Brown, and N.J. Fisher. 1986. PreliminaryMonitoring Design for Metal Pollution in Palmerton,Pennsylvania. In: Quality Control in Remedial Site In-vestigation: Hazardous and Industrial Solid Waste Test-ing, 5th Volume, C.L. Perket (ed.), ASTM STP 925,American Society for Testing and Materials, Philadel-phia, PA, pp. 57-66.
Steel, T.D. 1986. Converting Water Quality Information Goalsinto Statistical Design Criteria. In: Monitoring to DetectChanges in Water Quality Series, D. Lemer (ed.), Int.Ass. of Hydrological Sciences Pub. No. 157, pp. 71-79,
Stevenson, C.D. 1978. Simple Apparatus for Monitoring LandDisposal Systems by Sampling Percolating Soil Waters.Environ. Sci. Technol. 12:329-331.
Stolzenburg, T.R. and D.G. Nichols. 1985. Preliminary Re-sults on Chemical Changes in Groundwater Samples Dueto Sampling Devices. EPRI EA-4118. Electric PowerResearch Institute, Palo Alto, CA,
Sukhatme, P.V. and B.V. Sukhatme. 1970. Sampling Theoryof Surveys with Applications, 2nd ed. Iowa State Univer-sity Press, Ames, IA.
Summers, K.V. and S.A. Gherini. 1987. Sampling Guidelinesfor Groundwater Quality. EPRI EA-4952. Electric PowerResearch Institute, Palo Alto, CA.
Sykes, A.L., R.A. McAllister, and J.B. Homolya. 1986. Sorp-tion of Organics by Monitoring Well Construction Mate-rials. Ground Water Monitoring Review 6(4):44-47.
Tanur, J.M., (ed.). 1978. Statistics: A Guide to the Unknown.Holden-Day, San Francisco.
Tinlin, R.M. 1976. Monitoring Groundwater Quality: Illustra-tive Examples. EPA/600/4-76-036 (NTIS PB257 936).
Todd, D.K., R.M. Tinlin, K.D. Schmidt, and L.G. Everett.1976. Monitoring Ground-Water Quality: MonitoringMethodology. EPA/600/4-76-026 (NTIS PB256-068).
UNESCO. 1983. proceedings of the Symposium-Methodsand Instrumentation for the Investigation of GroundwaterSystems. Committee for Hydrological Research, CHO-TNO, The Hague, The Netherlands.
Unwin, J. and V. Maltby. 1988. Investigations of Techniquesfor Purging Ground-Water Monitoring Wells and Sam-pling Ground Water for Volatile Organic Compounds. In:Ground-Water Contamination: Field Methods, A.G.Collins and A.I. Johnson (eds.), ASTM STP 963, Americ-an Society for Testing and Materials, Philadelphia, PA,pp. 240-252.
U.S. Environmental Protection Agency (EPA). 1985. RCRAGround-Water Monitoring Compliance Order Guidance.EPA Office of Solid Waste and Emergency Response(NTIS PB87-19371O).
U.S. Environmental Protection Agency (EPA). 1986a. RCRAGround Water Monitoring Technical Enforcement Guid-ance Document. EPA OSWER-9950.1. Also published inNWWA/EPA Series, National Water Well Association,Dublin, OH.
US. Environmental Protection Agency (EPA). 1986b. TestMethods for Evaluating Solid Waste, 3rd ed., Vol. IIField Manual Physical/Chemical Methods. EPA/530/SW-846 (NTIS PB88-239223) First update, 3rd ed. EPA/530/SW-86.3-l (NTIS PB89-148076).
U.S. Environmental Protection Agency (EPA). 1986c. PermitGuidance Manual on Unsaturated Zone Monitoring forHazardous Waste Land Treatment Units. EPA/530/SW-86-040.
U.S Geological Survey. 1977+. National Handbook of Rec-ommended Methods for Water Data Acquisition. USGSOffice of Water Data Coordination, Reston, VA.
153
van Duijvenbooden, W. and H.G. van Waegeningh (eds.).1987. Vulnerability of Soil and Groundwater to Pollut-ants. Committee for Hydrological Research, CHO-TNO,The Hague, The Netherlands.
Voorhees, K.J., J.C. Hickey, and R.W. Klusman. 1984. Analy-sis of Groundwater Contamination by a New SurfaceStatic Trapping/Mass Spectrometry Technique. Anal.Chem. 56:2602-2604.
Williams, L.R., R.W. Leggett, M.L. Espegren, and C.A. Little.1989. Optimization of Sampling for the Determination ofMean Radium-226 Concentration in Surface Soil. Envi-ronmental Monitoring and Assessment 12:83-96.
Williams, W.H. 1978. A Sampler on Sampling. John Wiley&Sons, New York.
Wilson, L.G. 1980. Monitoring in the Vadose Zone: A Re-view of Technical Elements. EPA/600/7-80-134 (NTISPB81-125817).
Wilson, L.G. 1981. Monitoring in the Vadose Zone: Part I.Ground Water Monitoring Review 1(3):32-41.
Wilson, L,G. 1982. Monitoring in the Vadose Zonti Part II.Ground Water Monitoring Review 2(1):3 1-42.
Wilson, L.G. 1983. Monitoring in the Vadose Zone Part III.Ground Water Monitoring Review 3(2):155-166.
Wittmann, S. G., K.J. Quinn, and R.D. Lee. 1985. Use of SoilGas Sampling Techniques for Assessment of GroundWater Contamination. In: Proc. NWWA/API Conf. Pe-troleum Hydrocarbons and Organic Chemicals in GroundWater—Prevention, Detection and Restoration, 1985,National Water Well Association, Dublin, OH, pp. 291-309.
Wood, W .W. 1976. Guidelines for Collection and Field Analy-sis of Groundwater Samples for Selected Unstable Con-stituents. U.S. Geological Survey TWI 1-D2.
Yates, F. 1980. Sampling Methods for Censuses and Surveys,4th ed. MacMillan, New York.
Zapico, M.M., W. Vales, and J.A. Cherry. 1987. A WirelinePiston Core Barrel for Sampling Cohesionless Sand andGravel Below the Water Table. Ground Water Monitor-ing Review 7(3); 74-82.
154
PART II: PHYSICAL AND CHEMICAL PROCESSES IN THE SUBSURFACE
Chapter 10Physiochemical Processes: Organic Contaminants
Carl D. Palmer and Richard L. Johnson
10.1 Overview of Physiochemical Processes
The characterization of hazardous waste sites to designremediation strategies requires a broad range of backgroundinformation. As discussed in Chapter 9, good sampling meth-ods and strategies are required to determine the contaminationlevel and the extent to which contaminants have moved withinthe subsurface. Understanding of the physical processes dis-cussed in Chapters 4 and 5 allows determination of the rateand direction in which contaminated ground water is flowing.This information also can be used to determine whether thecontaminants will be captured and removed by extractionwells. However, this information by itself is not sufficient foroptimal choice of remediation schemes. Critical questionssuch as how much water must pass through a section of anaquifer to remove the contaminants or how much time isrequired for contaminants to diffuse out of low-permeabilityzones also must be answered. The answers to these questionsdepends on the physiochemical processes occurring withinthe subsurface.
The next three chapters address the physiochemicalprocesses that recur within the subsurface, the parametersrequired for their characterization, and the implications ofthese processes for remediation design. In this chapter, thediscussion is limited to processes occurring below the watertable that affect the concentration, transport, and hence re-moval of organic contaminants. Chapter 11 addresses thetransport of volatile organic compounds through the unsatur-ated zone, and Chapter 12 discusses inorganic contaminants.
The design of optimal remediation schemes often re-quires some “prediction” of the distribution of contaminantswithin the subsurface over time. These predictions then can beused to evaluate different remediation scenarios. The basis formaking such predictions is generally the application of theconcepts of mass balance. A common method for applyingmass balance concepts to dissolved chemical constituents inground-water systems is the advection-dispersion equation,which is written in its one-dimensioml form as:
[10-1]
where v is the ground-water velocity (L/T), D is the dispersioncoefficient (L2/T), C is the concentration of the dissolvedconstituent (M/L3), t is time, and RXN represents a generalchemical reaction term. The frost term in eq. 10-1 describesthe net advective flux of the contaminant in and out of avolume of the aquifer (Figure 10-1). The second term de-scribes the net dispersive flux of the contaminant. The firstterm on the right-hand side of the equation describes thechange in concentration of the contaminant in the watercontained within the volume of aquifer. The second term onthe right-hand side represents the amount of contaminant thatmay be added or lost to the ground water by some chemical orbiological reaction. If there is no reaction term, then theequation describes the transport of a conservative, nonreactingtracer such as chloride or bromide. More detailed informationabout the development and derivation of eq. 10-1 is found inPalmer and Johnson (1989), Gillham and Cherry (1982),Freeze and Cherry (1979), or Bear (1979, 1969).
Some understanding of this mass balance equation isuseful even to the individual who is not directly responsiblefor making mathematical representations of the distribution ofcontaminants within the subsurface. The equation is an ex-ample of the current understanding of the processes control-ling the fate and transport of contaminants in the subsurface.The equation lists the parameters that should be quantifiedeither by performing appropriate field or laboratory measure-ments or by using the best known values. The results of theapplication of this modeling are unlikely to ever exactly“predict” how the contaminants behave at a particular fieldsite but they can provide a general set of expectations that areuseful in the design of a remedial system. These results alsocan be used to compare aquifer remediation performance.
According to eq. 10-1, two parameters that must bedetermined are the ground-water velocity, v, and the disper-sion coefficient, D. These parameters are described in Chapters 4 and 5 as well as in other sources (e.g., Palmer andJohnson, 1989 a,b). Chemical processes that can affect thefate and transport of organic contaminants below the watertable include (1) abiotic degradation, (2) biotic degradation,(3) dissolution nonaqueous phase liquids (NAPLs), (4) sorp-tion reactions, and (5) ionization. Both abiotic and biotic
155
Figure 10-1. Mass balance equation for the transport of reactive solutes through porous media.
degradation are discussed in Chapter 13. The discussion inthis chapter is limited to the three latter processes.
10.2 Dissolution of Nonaqueous Phase LiquidsMany of the organic chemicals of environmental concern
enter the subsurface in the nonaqueous phase. How thesesolvents move through the soil depends on the grain size ofthe aquifer material, the degree of water saturation in the porespace, and the density and viscosity of the solvent relative towater (Palmer and Johnson, 1989c; Schwille, 1988). Forexample, if there is a spill of nonaqueous phase liquid that hasa density greater than water (DNAPL), as it flows through theunsaturated zone, because the water is in the wetting phase, itwill pass through the center of the pores. If there is residualwater within the unsaturated zone then the combination ofhigher density and lower viscosity of the DNAPL relative towater results in unstable flow or significant fingering of theDNAPL as it moves through the porous media. If the spill islarge enough so that the DNAPL can penetrate the capillaryfringe and move below the water table, this fingering contin-ues to occur. The transport of the DNAPLs is also verysensitive to small changes in permeability. Therefore, theDNAPL tends to spread laterally as it encounters lenses offiner grained material in the subsurface. This combination ofviscous fingering and lateral flow results in a series of fingersand pools of DNAPL. The DNAPL in the fingers tends todrain to some residual saturation while the pools containDNAPL above the residual saturation.
As ground water flows through the fingers, the DNAPL isdissolved by the passing ground water. Laboratory experi-ments (Anderson, 1988; Anderson et al., 1987) using a 15-cm-diameter cylindrical finger of tetrachloroethylene (TeCE)(Figure 10-2) demonstrate that the ground water passingthrough the fingers can quickly reach saturation with theTeCE. This was found to be true for ground-water velocitiesranging from 10 to 100 cm/day (Figure 10-3). However, theseresults do not imply that where a DNAPL spill has occurredthe sampled ground water is saturated with the solvent. In-deed, sampling results usually indicate that most waters arehighly undersaturated with respect to the DNAPLs. Although
Figure 10-2. Cylindrical source of tetrachloroethylene (TeCE)used in the experiments by Anderson (1988).
the water that passes through the fingers or very close to thepools of DNAPL within the subsurface is saturated with theDNAPL, mass transfer of the dissolved DNAPL to the areasfurther from these fingers and pools is predominantly bymolecular diffusion. As a result, many areas within the aquiferthat lie between the pools and fingers contain little or nodissolved solvent. While the distance between such fingersand pools is generally unknown, it is probably at least as greatas the mean distance between the small-scale beds within theaquifer. For the Borden aquifer in Ontario, Sudicky (1986)found this distance to be about 10 cm in the vertical direction.A typical monitoring well would have an intake length of atleast 2 m. Thus, the water saturated with the solvent is mixed
156
10-3.
Distance from Plume Center (cm)
Concentration of TeCE across the flow field at the end tank in the sand box experiments conducted(1988).
with the uncontaminated ground water resulting in measuredconcentrations that are substantially below saturation.
Estimating the time required to remove the nonaqueousphase liquid from the subsurface is difficult. Estimates requireknowledge of the amount that was spilled and the distributionof the solvent within the aquifer. While the former piece ofinformation is often difficult to obtain, the latter is virtuallyimpossible. If the solvent is assumed to be uniformly distrib-uted (a residual saturation, S1) within the aquifer, and theground water flowing through the aquifer instantaneouslyequilibrates with the solvent, then the time required to removethe solvent by dissolution, tr, is
[10-2]
where q is the porosity of the aquifer, L is the length of theaquifer containing the solvent through which the ground waterflows, Ccq is the equilibrium concentration of the contaminantin the ground water, and q is the ground-water flux. Estimatesof removal times based on eq. 10-2, however, underestimatethe actual removal time because the equation does not accountfor the role of soil heterogeneity, the differential times theground water takes to flow along different flowlines, or thelimitations in mass transfer of pools of NAPL that are aboveresidual saturation. If a pump-and-treat remediation scheme isalready in place, remediation time can be roughly estimatedby dividing the total mass of solvent in the aquifer by the massbeing removed per unit time by extraction wells.
10.3 Sorption Phenomena
10.3.1 Adsorption IsothermsOnce an organic compound has been dissolved into the
ground water, it will be transported away from the source areaby ground-water flow. The contaminants do not travel at the
byAnderson
same velocity as the ground water but can be slowed by theirinteraction with the soil matrix. This interaction with the soilis often described graphically as an adsorption isotherm. Anadsorption isotherm is simply a plot of the concentration ofthe contaminant on the soil versus the concentration of thecontaminant in solution. Isotherms are so named because theyare conducted at constant temperature. Different types ofadsorption isotherms are defined according to their generalshape and mathematical representation. For a Langmuir iso-therm, the concentrations on the soil increase with increasingground-water concentrations until a maximum concentrationon the soil is reached (Figure 10-4). The isotherm can berepresented by the equation
[10-3]
Aqueous Concentration
Figure 10-4. Lndmuir adsorption isotherm.
157
Solution Concentration
Figure 10-5. Freundlich adsorption isotherm.
where S (M/M) is the concentration on the soil, Smax (M/M) isthe maximum concentration on the soil, K(L3/M) is theLangmuir adsorption constant, and C (M/L3) is the concentra-tion in the ground water. A Freundlich (or Küster) isotherm isgiven by the equation:
[104]
where K is the Freundlich adsorption constant and a is apositive parameter. The shape of a Freundlich isotherm de-pends on the value of a. If a is greater than 1.0, the isothermbecomes steeper with increasing concentrations in the groundwater. If a is less than 1.0, the isotherm becomes steeper atlower concentrations (Figure 10-5).
A linear isotherm is a special case of the Freundlichisotherm where the parameter a is equal to unity. Linearisotherms are of particular interest because (1) many nonpo-lar, hydrophobic organic compounds tend to follow linearisotherms (Figure 10-6) over a wide range of conditions and(2) the application of a linear isotherm simplifies the math-ematical model used to simulate the rate of contaminantmovement in the subsurface and reduces the number of pa-rameters that need to be obtained during characterization.
Another way of representing the partitioning between thesoil and the ground water is by a “partition coefficient,” Kp.The partition coefficient is the ratio of the change in concen-tration of the contaminant on the soil to the change in concen-tration of the contaminant in the ground water or more simply,the slope of the isotherm. When the isotherm for a particularsoil is linear, the partition coefficient is constant.
Figure 10-6. Linear sorption isotherms obtained for several priority pollutants (after Chlou et al., 1979).
158
The partition coefficient of an organic chemical is notconstant for every soil. In general, KP increases as the fractionof organic carbon, foc, increases in the soil (Karickhoff, 1981).In other words, the sorption of nonpolar, hydrophobic organiccompounds in soils is primarily an equilibrium partitioningprocess into soil organic matter. KP can be represented by
[10-5]
where Koc is the slope of the experimentally determined Kp
versus foc curves like those in Figure 10-7. Alternatively, Koc:can be considered to be the partition coefficient for theorganic compound into an hypothetical pure organic carbonphase.
If sorption is the primary reaction occurring in the subsur-face, the right-hand side of eq. 10-1 represents the change inthe total mass of contaminant within a volume of the aquifer.The total change in mass in the volume of the aquifer is equalto the change in mass in the ground water plus the change inmass on the solid phase. The reaction term in eq. 10-1 is then
and volumetric water content of the soil, respectively. Substi-tuting
[10-6]
into this reaction term and recognizing that is equal toKP for a linear adsorption isotherm, eq. 10-1 can now bewritten as
[10-7]
Figure 10-7. Partition coefficients for pyrene and phenanthreneversus the fraction of organic carbon in the soil(after Karickhoff, 1981).
where the constant
[10-8]
is known as the “retardation factor.” The general form of theequation only changes by the constant R. All of the math-ematical solutions that are used to solve the transport ofnonreacting tracers can be used to solve for the transport ofnonpolar hydrophobic organic compounds if the ground-wa-ter velocity and dispersion coefficient are divided by R.
The retardation factor can be interpreted in slightly dif-ferent but equally valid ways. It is the ratio of the ground-water velocity, v, to the solute velocity, vs, (i.e., R = v/vs). It isalso the ratio of the time for the solute to travel from a sourceto an observation point divided by the time for the groundwater to travel that same path. The retardation factor also canbe thought to represent the number of pore volumes that mustbe flushed through a soil to remove the contaminant. All ofthese definitions assume that the only process occurring islinear sorption.
Application of the new expression (eq. 10-7) requiresknowledge of the additional parameter R. This parameter canbe obtained by several methods including (1) calculation fromeq. 10-8, where Kp is obtained from correlation techniques;(2) calculation from eq. 10-8, with Kp, obtained from batchsorption tests; (3) measurement from column tests and (4)estimation from field data. The other parameters in eq. 10-8(porosity and dry bulk density) are physical parameters thatcan be obtained using common techniques (see Chapter 4 andPalmer and Johnson, 1989c).
10.3.2 Determining Retardation Factors UsingfOC and KOC
The relationship between the Koc value and other knownproperties of organic contaminants has been examined bynumerous researchers (Kenaga and Goring, 1980; Karickhoff,1981, Schwartzbach and Westall, 1981; Chiou et al., 1982 and1983), For example, some research has revealed linear rela-tionships between the log of the volubility of the contaminantand the log (Koc) (Figure 10-8). Similarly, Karickhoff sug-gested that the partitioning of organic contaminants into soilorganic matter must be analogous to the partitioning of thosecontaminants into other organic compounds such as octanol.He found a linear relationship (Figure 10-9) between log (Koc)and log (KOW), where KOW is the octanol-water partition coeffi- cient. Several regression equations relating the properties oforganic compounds to the Koc have been derived (Table 10-1).Thus, by knowing the name of the compound of interest, theseproperties can be found in tables of chemical properties(Mabey et al., 1982) and the regression equations used toapproximate Koc. The goal, however, is to determine thepartition coefficient and ultimately the retardation factor. Todo this, eq. 10-5 must be applied and a measurement of thefraction of organic carbon must be obtained.
The many methods of measuring the amount of organiccarbon in the soil can be broadly classified as either wetcombustion or dry combustion techniques. Wet combustiontechniques involve the addition of a strong oxidizing agent
159
Figure 10-8. Log Koc versus logarithm of the volubility of the compound in water (after Kenaga and Goring, 1980).
Figure 10-9. Log Koc versus the octanol-water partition coefficient. Data from Karickhoff (1981).
such as bichromate to the soil. There are several such wetcombustion techniques including the Walkley-Black methodand the modified Mebius procedure; these procedures arediscussed in detail by Nelson and Sommers (1982). In spite ofsome limitations, these methods can provide a relatively rapidand inexpensive method for obtaining estimates of foc.
Dry combustion methods generally involve heating thesoil sample in the presence of oxygen. The oxygen reacts withthe soil carbon to form carbon dioxide that can be detected bya variety of techniques.
To estimate the linear retardation factor, the Koc obtainedfrom one (or more) of the regression equations given in Table10-1 is multiplied by the fraction of organic carbon to yieldthe partition coefficient (eq. 10-1). The retardation factor isobtained from the KP, by eq. 10-8.
There are several limitations to the use of the correlationtechniques described above. The linear relationship betweenfoc and KP is not always easy to determine. In particular, therelationship is most likely to fail when (1) the foc is very low(<0,001), (2) when there are large amounts of swelling clayspresent, and (3) the organic compound is polar (e.g., com-pounds that contain amine or carboxylic acid groups) (Pankow,
160
Table 10-1. Some Reported Correlation Equations
a Number in parentheses refer to the number of compounds in data base.b SW is the solubility of the compound in water in ppm.c Derived from the original equation assuming Ka = 1.7 Komd Xs is the mole fraction solubulity at 25°C.
After Pankow, 1984
1984). There are also several reasons why the relationshipbetween log (Koc and log (KoW) may not always be linear(Pankow, 1984). If mechanisms other than simple partitioninginto soil organic carbon are contributing to the adsorption ofthe organic contaminant, then the Koc value, computed as theratio Kp/foc , will be in error. Also, If the molecule is large itmay not fit into the soil organic matter to the same extent as itwould in octanol (steric limitations). Finally, if the adsorptionis strong, a contaminant may take a substantial period of timeto equilibrate with the soil organic carbon.
10.3.3 Determining Retardation Factors UsingBatch Tests
Retardation factors also can be measured with batch tests.These tests are, in principle, easy to perform, and the methodis outlined in Figure 10-10. A known volume of solution, Vw,containing an initial concentration, CO, of a contaminant isplaced into a container. A known mass of soil, Ms, is thenadded and the mixture is shaken and allowed to equilibrate.The soil then is separated from the solution by centrifuging,and an aliquot of the supernatant is sampled. The concentra-tion of the contaminant in this aliquot, C, is measured and theconcentration on the soil, S, is calculated by
[10-9]
Figure 10-10. Batch adsorption tests.
161
This test can be run several times with different initialconcentrations or different masses of soil. The result is aseries of contaminant concentrations with corresponding aque-ous phase concentrations that yield an isotherm when they areplotted. If the isotherm is linear, the slope, or partition coeffi-cient, can be easily determined. The retardation factor thencart be calculated from KP, pb, and O using eq. 10-8.
Prior to conducting such batch adsorption tests, the soil isprepared by drying and then sieving through a 2-mm sieve.The sieving is to ensure that aggregated soil particles arerelatively small, thus reducing the time for the contaminant todiffuse into the particles and equilibrate with the soil. Anotherimportant preparatory step is to estimate the KP using, forexample, the correlation methods described in Section 10.3.2.This is important in choosing the proper amount of soil to usein the tests. If Kp is large and too much soil is added to thereaction vessel, then most of the contaminant is partitioned tothe soil and the concentration in solution cannot be accuratelydetermined. Similarly, if KP is small and too little soil is addedto the reaction vessel, then the measured contaminant concen-tration falls within the analytical error of the initial concentra-tion and an accurate estimate of the contaminant concentrationof the soil cannot be obtained. Both of these cases lead to poormeasures of the partition coefficient.
There are some problems that complicate the use of batchtests for determining Kp. For example, batch tests assume thatequilibrium is established between the soil and the solution,but some contaminants may take a very long period of time toequilibrate. Experiments on the resorption of hexachloro-benzene from soils (Karickhoff and Morris, 1985) indicatedthat even after 35 days equilibrium was not obtained (Figure10-11).
Another problem involves nonsettling particles. The sepa-ration of the soil and the water is assumed to be completebefore sampling of the supernatural however, very fine, col-loidal-size particles may remain in suspension. The contami-nants attached, to these particles are stripped during the analysisof the water, which causes overestimation of the aqueousphase concentration. This results in underestimation of thepartition coefficient (e.g., Gschwend and Wu, 1984). Themagnitude of the effect depends on the concentration ofnonsettling particles (NSPS) and the true partition coefficientonto those particles (Figure 10- 12). If the partition coefficientis small, then most of the mass of the contaminant is insolution and the error caused by the NSPS is negligible. If thepartition coefficient is large, then a significant mass of thecontaminant is really partitioned onto the soil particles caus-ing significant errors in the aqueous phase concentration andhence C
p
A third problem arises from the loss of contaminant byvolatilization during equilibration, sampling, and analysis.This problem can be minimized by eliminating head-spaceand using properly sealed reaction vessels.
Uncontaminated background soils are recommended forbatch adsorption tests. If the soils contain any NAPLs, thecontaminant being investigated will partition into the NAPL,yielding a potentially large and incorrect partition coefficient.Once Kp is determined in the batch test, the retardation factor,R, can be estimated by using eq. 10-8.
10.3.4 Determining Retardation Factors fromColumn Tests
A third method for estimating linear repartition factors iswith column tests. In these tests, a column of soil is prepared,
Figure 10-11. The fraction of hexachlorobenzene sorbed to two soils versus time during desorption teats (after Karickhoff andMorris, 1985).
162
Figure 10-12. The effects of nonsettling particles on theobserved partition coefficient (after Pankow,1984).
Figure 10-13. Column tests for determination of retardation
and a solution containing a nonadsorbing tracer and the con-taminant of interest is run through the column (Figure 10-13).The concentrations of the tracer and contaminant can bemeasured in the water that has passed through the column.The retardation factor is then the ratio of the time (or volume)for the center of mass of the contaminant to break through thecolumn to the time (or volume) for the center of mass of thenonreactive tracer to break through the column. This tech-nique provides a direct measure of R; however, it is onlywell suited for those contaminants that have a relativelysmall (< 10) retardation factor. Contaminants with retardationfactors much greater than 10 require too much time to mea-sure to be practical. Other disadvantages of using column testsinclude the slow flow rates in fine-grained material, the de-struction of soil structure by soil repacking, and the difficultyin distinguishing kinetic behavior from the heterogeneouspacking within the column.
10.3.5 Determining Retardation Factors fromField Data
Site-specific field information obtained during the Reme-dial Investigation/Feasibility Study (RI/FS) can, in some cases,be used to estimate contaminant retardation. While in prin-ciple retardation factors can be back-calculated from break-through curves obtained at monitoring wells or through thespatial distribution of the contaminants in the subsurface, inpractice, only the latter is likely to be obtained. The retarda-tion factors can be estimated by dividing the velocity ofground water by the velocity of the contaminant. The ground-water velocity can be estimated from Darcy’s Law and theporosity, or alternatively by the distance some nonadsorbingsolute travels after the release. The solute velocity can be
estimated by dividing the mean distance the contaminant hastraveled by the time since its release into the subsurface. Oneof the potential disadvantages of this method is that otherprocesses that are not included in the data analysis are occur-ring within the aquifer. Ignoring these processes can result inpoor estimates of the retardation factor.
10.3.6 Comparison of Methods for Estimation ofRetardation
Each of the methods for estimating the retardation factorhas advantages and disadvantages. One of the key questions,however, is how do these different methods for estimatingretardation compare. The best technique for comparison is tolook to large-scale field tracer experiments where very accu-rate field values have been obtained. This has been done forthe Stanford-Waterloo tracer experiment that was conductedin the sandy aquifer on Canadian Forces Base Borden inOntario, Canada. Details of the experiment and analysis of theresults can be found in Mackay et al. (1986); Roberts et al.(1986); Curtis et al. (1986); Freyburg (1986); and Sudicky(1986).
A summary of the retardation factors obtained for fivedifferent compounds using a correlation method, batch tests,and temporal and spatial data from the field experiment isgiven in Table 10-2. The batch tests agree closely with thefield data. The correlation technique tends to consistentlyunderestimate the retardation factors. The underestimation ofthe retardation factors may be the result of poor estimates ofthe fraction of organic carbon (e.g., Powell et al., 1989) orerrors in the assumptions in eq. 10-5, or they may be the result
163
Table 10.2 Comparison of Methods for Retardation Factors
of the inherent error in the regression equations. Recall thatthe regression equations are based on the logarithms of thevalues; therefore, the best estimates of the Koc and hence theretardation factor may be a factor of 2 or 3 from the “true”value. Nonetheless, the correlation techniques do provide thecorrect order of magnitude estimate of the retardation factor atvery little expense. Such values would be appropriate for thepreliminary design of the remedial strategies. If more accurateestimates are required, then the more expensive batch orcolumn tests should be used. Enough samples should betested, however, to estimate the uncertainty of the retardationfactor for each of the important geologic units.
10.3.7 Applicability and Limitations of LinearPartitioning and Retardation
Most of the emphasis in this section has been on thelinear partitioning and retardation model for the adsorption ofneutral, hydrophobic organic compounds in the environment.While this model is adequate for many situations, it is impor-tant to recognize the limitations in the assumptions so that it isnot applied to situations where it is imppropriate.
Retardation only describes the process of the partitioningof the contaminant between the ground water and soil organicmatter. If the nonaqueous solvent phase is dissolving or theorganic compounds are degrading, then these additional pro-cesses also must be taken into account. However, for describ-ing the partitioning process, the linear retardation model isreasonable for many compounds if the concentration of thecontaminant is less than 10-5 molar or less than half thevolubility, whichever is lower (Karickhoff et al., 1979;Karickhoff, 1984). At high or low concentrations the linearisotherm may deviate. Some data on the adsorption of TCE toglacial till suggest that the partition coefficient is not constantbut may vary by as much as 50-fold over range in ground-water concentrations from 10 to 10,000 parts per billion (ppb)(Figure 10-14). This variation occurs even though the parti-tion coefficient is approximately constant over the range from100 to several thousand ppb.
The linear retardation model assumes that equilibrium isachieved quickly. In some circumstances, the rate of adsorp-tion and resorption can bean important factor. As mentionedin Section 10.3.3, Karickhoff and Morris (1985) found thatduring the resorption of hexachlorobenzene, equilibrium wasnot achieved even after 35 days of reaction time (Figure 10-11).
10.4 Ionization and CosolvationAnother important reaction that can affect sorption and
hence the rate of removal of organic contaminants from thesubsurface is ionization. Acidic compounds such as phenols,catechols, quinoline, and organic acids can lose or gain pro-tons (H+) depending upon the pH. The resultant ions are muchmore soluble and less hydrophobic than the uncharged forms.Therefore, the ionized forms have much lower Koc values thanthe uncharged forms, The pH at which this reduction in Koc
becomes substantial can be predicted based on the acidity ofthe compound. This acidity is often represented as the pKa ofthe compound, which is the pH at which 50 percent of themolecules are ionized.
Table 10-3 lists pKa’s for a number of environmentallysignificant ionizing compounds. For example, trichlorophenolionizes to a phenolate (Figure 10-15). The trichlorophenol hasa relatively large Koc value (2330) and readily partitions intothe soil organic matter. The ionized form is not as hydropho-bic and its KOC value is substantially smaller than the Koc of thetrichlorophenol. As the pH increases, the fraction of thephenol that is ionized increases and the Koc decreases (Figure10-16). Therefore, the Koc value based on the total concentra-tion of the phenolic compound is dependent on the degree ofionization of the compound. While the phenolate compoundmay be retarded mainly by anion adsorption to oxide surfacesin low carbon soils, there is evidence that the phenolate alsopartitions into the soil organic carbon Schellenberg et al.,1984).
Studies with other compounds also have indicated therelative importance of ionization of organic compounds. Stud-ies of quinoline in low carbon soils suggest that the mainmechanism for sorption is primarily by ion adsorption (Zacharaet al., 1986; Ainsworth et al., 1987).
It often is assumed that water at hazardous waste sites hasabout the same chemical properties as pure water and that thesolubilities of hydrophobic organic contaminants are rela-tively constant within a very narrow range. However, many ofthe chemical properties of mixtures of solvents, such as waterand methanol, can change as the fraction of the cosolvent inthe mixture changes. The thermodynamic basis for some ofthese cosolvation effects is described by Rao et al. (1985) andWoodburn et al. (1986). Of particular interest is that thevolubility of many organic compounds can be increased byorders of magnitude within mixtures of water and other mis-cible solvents (Nkedi-Kizza et al., 1985; Fu and Luthy, 1986aand 1986b; Zachara et al., 1988). For example, the partitioncoefficient of anthracene decreases more than an order ofmagnitude as the fraction of methanol (the cosolvent) isincreased from O to 50 percent (Figure 10-17).
Such cosolvation effects may be either advantageous ordisadvantageous depending on the specific problem. If thesemiscible liquid cosolvents have been codisposed with prioritypollutants on site and the main concern is compliance moni-toring, then the lower partition coefficient results in highertransport rates to the compliance boundary. If the focus,however, is on remediation, then the cosolvation effect mayallow a technology such as pump-and-treat to be considered a
164
Figure 10-14. Partition coefficients for TCE on glacial till.
Table 10-3. Acid Dissociation Constants for Several PriorityPollutants
viable option. Alternatively, the addition of cosolvents to thesubsurface for the express purpose of enhancing the removalof these organic contaminants in a timely and cost-effectivemanner may be a possibility; however, such technology hasyet to be demonstrated in the field.
10.5 Expressions for Other Chemical ProcessesThe emphasis in the discussion above centered mostly on
the dissolution of the NAPL phases and equilibrium adsorp-tion with linear partitioning. These processes are emphasizedbecause under many conditions they are the more importantprocesses controlling the rate of transport and removal of
Figure 10-15. Ionization of trichlorophenol to trichlorphenolate.
organic contaminants from the subsurface. However, otherchemical processes may be taking place within the subsurfaceand equilibrium may not always be a reasonable assumption.These other equilibrium and nonequilibrium processes alsocan be represented in the general expression given by eq. 10-1. A few of the expressions for different chemical processesare given in Table 10-4. If one of these other expressions isrequired to describe the reactions that are occurring within thesubsurface, then other parameters must be measured or esti-mated. For example, if adsorption/desorption for a particularcompound is rate-controlled rather than equilibrium-controlled,then the rates of adsorption and resorption should be deter-mined. These rates can be inferred from batch or column testssimilar to those described above, but they require measure-ments over time and a more sophisticated level of interpreta-tion and analysis. Such models should be called upon ifrequired for understanding the processes at a particular site.
165
Figure 10-16.
I
Percent of ionization of three different chlorophenolic compounds versus pH. Based on data from Schellenberget aI. (1984).
Table 10-4. Reaction Terms for Various Chemical Processes
Figure IO-17. Partition coefficient of anthracene on threedifferent soils versus fraction of methanol presentas a cosolvent (adapted from Nkedi-Kizza et al.,1985).
166
10.6 References cent Trends in Hydrogeology, T.N. Narasimhan, (ed.),
Ainsworth, C.C., J.M. Zachara, and R.L. Schmidt. 1987. Geological Society of America Special Paper 189, pp.
Quinoline Sorption on Na-Montmorillonite: Contribu- 31-62.
tions of the Protonated and Neutral Species. Clays andClay Minerals 35:121-128. Gschwend, P.M. and S-C. Wu, 1984. On the Constancy of
Sediment-Water Partition Coefficients of HydrophobicAnderson, M.A. 1988. Dissolution of Tetrachloroethylene Organic Pollutants. Environ. Sci. Technol. 19:90-96.
into Ground Water. PhD Dissertation, Oregon GraduateCenter, Beaverton, OR. Johnson, R.L., S.M. Brillante, L.M. Isabelle, J.E. Houck, and
J.F. Pankow. 1985. Migration of Chlorophenolic Com-
Anderson, M.A., J.F. Pankow, and R.L. Johnson. 1987. The pounds at the Chemical Waste Disposal Site at AlkaliDissolution of Residual Dense Non-Aqueous Phase Liq- Lake, OR-2. Contaminant Distributions, Transport, anduid (DNAPL) from a Saturated Porous Medium. In: Proc. Retardation. Ground Water 23(5):652-666.
NWWA/API Conf. on Petroleum Hydrocarbons and Or-ganic Chemicals in Ground Water-Prevention, Detec- Johnson, R. L., J.A. Cherry, and J.F. Pankow, 1989. Diffusive
tion and Restoration, National Water Well Association, Contaminant Transport in Natural Clay: A Field Example
Dublin, OH, pp. 409-428. and Implications for Clay-Lined Waste Disposal Sites.Environ. Sci. Technol. 23:340-349.
Bear, J. 1969. Hydrodynamic Dispersion. In: Flow ThroughPorous Media, R.J.M. De Wiest (ed.), Academic Press, Karickhoff, S.W. and K.R. Morris. 1985. Sorption Dynamics
New York, pp. 109-199. of Hydrophobic Pollutants in Sediment Suspensions.Environ. Toxicol. Chem. 4:469-479.
Bear, J. 1979. Hydraulics of Groundwater. McGraw-Hill,New York. Karickhoff, S.W., D.S. Brown and T.A. Scott. 1979. Sorption
of Hydrophobic Pollutants on Natural Sediments. WaterChiou, C.T., D.W. Schmedding and M. Manes. 1982. Parti- Research 13:241-248.
tioning of Organic Compounds on Octanol-Water Sys-tems. Environ. Sci. Technol. 16:4-10. Karickhoff, S.W. 1984. Organic Pollutant Sorption in Aquatic
Systems. J. Hydraulic Engineering ASCE 110:707-735.
Chiou C.T., L.J. Peters and V.H. Freed. 1979. A PhysicalConcept of Soil-Water Equilibria for Nonionic Organic Karickhoff, S.W. 1981. Semi-Empirical Estimation of Sorp-
Compounds. Science 206:831-832. tion of Hydrophobic Pollutants on Natural Sediments andSoils. Chemosphere 10:833-846.
Chiou, C.T., P.E. Porter and D.W. Schmedding. 1983. Parti-tion Equilibria of Nonionic Organic Compounds Be- Kenaga, E.E. and C.A.I. Goring, 1980. Relationship betweentween Soil Organic Matter and Water. Environ. Sci. Water Volubility, Soil-Sorption, Octanol-Water Partition-Technol. 17:227-231. ing, and Bioconcentration of Chemicals in Biota. In:
Aquatic Toxicology (Proc. 3rd Annual Symp. on Aquatic
Curtis, G.P, P.V. Roberts, M. Reinhard, 1986. A Natural Toxicology), ASTM STP 707, American Society for
Gradient Experiment on Solute Transport in a Sand Aqui- Testing and Materials, Philadelphia, PA, pp. 78-115.fer, 4, Sorption of Organic Solutes and Its Influence onMobility. Water Resources Research, 22(13):2059-2067. Mabey, W.R., et al. 1982. Aquatic Fate Process Data for
Organic Priority Pollutants. EPA/440/4-81-014 (NTISFreeze, R.A. and J.A. Cherry. 1979. Groundwater. Prentice-
Hall, Englewood Cliffs, NJ.
Freyberg, D.L. 1986. A Natural Gradient Experiment onSolute Transport in a Sand Aquifer. 2. Spatial Momentsand the Advection and Dispersion of Nonreactive Trac-ers. Water Resources Research 22(13):2031-2046.
Fu, J.K. and R.G. Luthy. 1986a. Aromatic Compound Solu-bility in Solvent/Water Mixtures. J. Environ. Eng. 112:328-345.
PB87-169090), Chapter 4,
Mackay, D. and B. Powers. 1987. Sorption of HydrophobicChemicals From Water A Hypothesis for the Mechanismof the Particle Concentration Effect. Chemosphere 16:745-757.
Mackay, D. M., D.L. Freyberg, P.V. Roberts, and J.A. Cherry.1986. A Natural Gradient Experiment on Solute Trans-port in a Sand Aquifer, 1. Approach and Overview ofPlume Movement. Water Resources Research22(13):2017-2030.
Fu, J.K. and R.G. Luthy. 1986b. Effect of Organic Solvent onSorption of Aromatic Solutes onto Soils. J. Environ. Eng. McKay, L.D. and M.R. Trudel. 1987. Sorption of Trichloro-112346-366. ethylene in Clayey Soils at the Tricil Waste Disposal Site
near Sarnia. Ontario. Unpublished report. University ofGillham R.W. and J.A. Cherry. 1982. Contaminant Migration Waterloo Institute for Ground Water Research.
in Saturated Unconsolidated Geologic Deposits. In: Re-
167
Myrand, D., 1987. Diffusion of Volatile Organic Compoundsin Natural Clay Deposits. M. SC. Thesis, Department ofEarth Sciences, University of Waterloo, Waterloo, Ontario.
Nelson, D.W. and L.E. Sommers. 1982. Total Carbon, Or-ganic Carbon, and Organic Matter. In: Methods of SoilAnalysis, Part 2, Chemical and Biological Properties,A.L. Page, R.H. Miller, and D.R. Keeney (eds.), ASAMonograph No. 9, American Society of Agronomy, Madi-son, WI, pp. 539-580.
Nkedi-Kizza, P., P.S.C. Rao, and A.G. Hornsby. 1985. Influ-ence of Organic Cosolvents on Sorption of HydrophobicOrganic Chemicals by Soils. Environ. Sci. Technol.19:975-979.
Palmer, C.D. and R.L. Johnson. 1989a. Physical ProcessesControlling the Transport of Contaminants in the Aque-ous Phase. In: Transport and Fate of Contaminants in theSubsurface, EPA/625/4-89/O19, pp. 5-22.
Palmer, C.D and R.L. Johnson. 1989b. Physical ProcessesControlling the Transport of Non-Aqueous Phase Liquidsin the Subsurface. In: Transport and Fate of Contamin-ants in the Subsurface, EPA/625/4-89/019, pp. 23-27.
Palmer, C.D. and R.L. Johnson. 1989c. Determination ofPhysical Transport Parameters. In: Transport and Fate ofContaminants in the Subsurface, EPA/625/4-89/019, pp.29-40.
Pankow, J. 1984. Groundwater Contamination by OrganicCompounds: Principles of Contaminant Migration andDetermimtion. Short Course Notes, Oregon GraduateInstitute, Beaverton, Oregon.
Powell, R.M., B.E. Bledsoe, G.P. Curtis, and R.L. Johnson,1989. Interlaboratory Methods Comparison for the TotalOrganic Carbon Analysis of Aquifer Materials. Environ.Sci. Technol. 23(10):1246-1249.
Rae, P.S.C., A.G. Hornsby, D.P. Kilcrease, and P. Nkedi-Kizza. 1985. Sorption and Transport of Toxic OrganicSubstances in Aqueous and Mixed Solvent Systems. J.Environ. Quality 14:376-383.
Roberts, P. V., M.N. Goltz, and D.M. Mackay. 1986. A Natu-ral Gradient Experiment on Solute Transport in a SandAquifer 3. Retardation Estimates and Mass Balances forOrganic Solutes. Water Resources Research 22(13):2047-2058.
Schellenberg, K.C., C. Leuenberger, and R.P. Schwarzenbach.1984. Sorption of Chlorinated Phenols by Natural Sedi-ments and Aquifer Materials. Environ. Sci. Technol.18:1360-1367.
Schwarzenbach, R., and J. Westall. 1981. Transport of Non-polar Organic Compounds from Surface Water to GroundWater Laboratory Sorption Studies. Environ. Sci. Technol.15:1360-1367.
Schwille, F. 1988. Dense Chlorinated Solvents in Porous andFractured Media: Model Experiments. Lewis Publishers,Chelsea, MI.
Sudicky, E.A. 1986. A Natural Gradient Experiment in a SandAquifer Spatial Variability of Hydraulic Conductivityand Its Role in the Dispersion Process. Water ResourceResearch 22(13):2069-2082.
Witkowski, P.J., P.R. Jaffe, and R.A. Ferrara. 1988. Sorptionand Resorption Dynamics of Aroclor 1242 to NaturalSediment. J. Contaminant Hydrology 2:249-269.
Woodbum, K.B., 1986. Solvaphobic Approach for PredictingSorption of Hydrophobic Organic Chemicals on Syn-thetic Sorbents and Soils. J. Contaminant Hydrology1:227-241.
Wu, S.-C. and P.M. Gschwend. 1986. Sorption Kinetics ofHydrophobic Organic Compounds to Natural Sedimentsand Soils. Environ. Sci. Technol. 20:717-725.
Zachara, J.M., C.C. Ainsworth, L.J. Felice, and C.T. Resch.1986. Quinoline Sorption to Subsurface Materials: Roleof pH and Retention of the Organic Cation. Environ. Sci.Technol. 20:620-627.
Zachara, J.M., C.C. Ainsworth, C.E. Cowan, and B.L. Tho-mas. 1987. Sorption of Binary Mixtures of AromaticNitrogen Heterocyclic Compounds on Subsurface Mate-rials. Environ. Sci. Technol. 21:397-402.
Zachara, J.M., C.C. Ainsworth, R.L. Schmidt, and C.T. Resch.1988. Influence of Cosolvents on Quinoline Sorption bySubsurface Materials and Clays. J. Contaminant Hydrol-ogy 2:343-364.
168
Chapter 11Physiochemical Processes: Volatilization and Gas-Phase Transport
Carl D. Palmer and Richard L. Johnson
Many nonaqueous phase liquids (NAPLs) are volatileorganic compounds of environmental concern (e.g., chlori-nated solvents, gasoline). They frequently enter ground-watersystems after they have been spilled on the surface and passthrough the unsaturated zone (Figure 1l-l). As these NAPLsflow through the unsaturated zone, a portion of the liquidremains behind in fingers at residual saturation, in pools ofmaterial on small heterogeneities, or above the capillary fringe(e.g., Palmer and Johnson, 1989a Feenstra and Cherry, 1987;Schwille, 1967 and 1988). The NAPL that remains in theunsaturated zone is an important source of contaminationbecause it is dissolved by (1) the passing recharge water, and(2) the passing ground water as the water table rises. Such
sources of contamination can last for many years and con-taminate large volumes of ground water. However, in additionto these pathways, contaminants also can be transported throughthe unsaturated zone in the gas phase. This transport pathwaymay spread the contaminants over a much broader area of theaquifer. Another complicating factor is the mass transfer ofthe contaminants across the atmosphere-soil boundary. Ofgreater interest is the implication that these sources of ground-water contamination can be quickly remediated by activelypumping the soil gas and removing the volatile organic con-taminants from the unsaturated zone to the surface where theymay be treated. As with the transport of dissolved contami-nants, the design of optimal remediation schemes requires
Figure 11-1. Transport of a DNAPL into the subsurface illustrating the distribution of the DNAPL, the dense vapors, and thedissolved chemical plume (after Feenstra and Cherry, 1988).
169
knowledge of the physiochemical processes that controltransport pathways. Mercer and Cohen (1990) provide a re-cent review of the literature on properties, models, character-ization, and remediation of NAPLs.
11.1 VolatilizationNear a NAPL spill, a four-phase system exists. The
phases include (1) the aquifer matrix, (2) the residual soilwater, (3) the NAPL, and (4) the air-filled pore space (Figure11-2). A volatile NAPL partitions from the NAPL phase intothe gas phase where it then can be transported to otherportions of the unsaturated zone. The vapor pressure of aparticular contaminant, Pk, in the gas phase can be calculatedfrom Raoult’s Law:
where Xk is the mole
[11-1]
e fraction of component k in the NAPLand P0
k iS the vapor pressure above the pure component. Forexample, if the NAPL is gasoline, the partial pressure onbenzene, one of the many components of gasoline, is the molefraction of benzene in the gasoline times the ideal vaporpressure above pure benzene. The concentration of the gas inthe soil atmosphere then can be calculated from the ideal gaslaw:
[11-2]
where n is the number of mole of component k, V is thevolume of gas, T is the kelvin temperature, and R is the gasconstant (0.082057 Iiter-atm mole-l ddg-1).
11.2 Gas-Phase TransportThe movement of the contaminants in the gas phase of
the unsaturated zone can be described by performing a massbalance on a volume of aquifer (Figure 11-3) in a mannersimilar to the approach taken in Chapter 10 (see Section 10.1).
11.2.1 DiffusionUnder nonpumping conditions, Fickian diffusion is the
prime process for gas-phase transport. The mass balance ortransport equation can then be written in its one-dimensionalform as:
[11-3]
where G is the concentration of the contaminant in the gasphasa, 01, is the air-filled porosity, D is the free air diffusioncoefficient and 1 is the air-phase tortuosity factor. The
Figure 11-2. Four-phase system consisting of soil matrix, water, NAPL and air (after Schwille, 1988).
170
Figure II-3. Mass balance equation for the transport of contaminants in the gas phase.
term accounts for the diffusion taking place in a porousmedium rather than in an open air space such as a room. Theindividual molecules must travel around the sand grains andwater films that are present in the porous medium. The termon the right-hand side of eq. 11-3 represents the net diffusiveflux per unit time of the contaminant in and out of the volumeof soil. The term on the left-hand side of eq. 11-3 representsthe change in mass of gas within the volume per unit time.Again, this equation is a useful example, listing the minimumparameters that must be obtained if vapor transport is to bedescribed.
The air-filled porosity is a physical parameter that can beobtained by finding the difference of the porosity and thevolumetric water content by using the methods referred to inChapter 5 and in Palmer and Johnson (1989b). The air tortuosityfactor can be obtained from empirical equations that are Figure 11-4. The air tortuosity factor based on the Millington-provided from detailed studies of gas-phase transport. For Quirk equation (Millington, 1959) as a function ofexample, one such equation is the Millington-Quirk air porosity for four different total porosities.(Millington, 1959) equation:
where 6 is the total porosity, which is equal to the sum of the
So the air-phase tortuosity can be calculated from physicalparameters that are already obtained. The actual value of theair-phase tortuosity factor varies from O when the entire porespace is occupied by water (saturated conditions) to about 0.8when the porosity is high and the medium is dry (Figure 11-4).
The third important parameter, the free air diffusioncoefficient, sometimes can be found for the specific com-pound of interest in reference tables. If the diffusion coeffi-cient for the specific compound (Dl) cannot be found, it canbe estimated from the diffusion coefficient (D2) and molecular
weight (M2) of another compound and the molecular weightof the compound of interest (M l) by:
[11-5]
11.2.2 Gas-Phase RetardationThe example given in eq. 11-3 describes the diffusive
transport of a gas-phase contaminant moving through the air-filled pores of the unsaturated zone, but it does not account forany chemical interactions with either the soil water or the soilmatrix. The expression should be modified by adding a gen-eral reaction term, RXN, to account for these processes. Atequilibrium, partitioning of the contaminant between the gas
171
phase and the water phase is defined by the dimensionlessHenry’s constant, KH:
G = KHC [11-6]
where G is the gas concentration and C is the water-phaseconcentration. For most neutral, nonpolar hydrophobic or-ganic compounds, the partitioning between the water phaseand the soil organic carbon can be described by a linearisotherm (S = KPC) that can be written in terms of the gas-phase concentrations as:
[11-7]
where S is the concentration of contaminant on the soil and KP
is the linear partition coefficient from the water phase to thesoil organic carbon. Including the reaction term, RXN,
the addition of these two sinks (sources) of contaminantwithin the volume of aquifer (Figure 11-2) results in a massbalance equation (paradigm):
coefficient, KP, are discussed in Chapter 10. The remainingparameter, KH, can be obtained from tables of chemical prop-erties (Mabey et al., 1982).
A more physical interpretation to this air-phase retarda-tion factor, Ra: can be given. As the contaminated gas diffusesthrough the air-filled pores, the rate of diffusion of the con-taminant in the air phase is less than that of the air itselfbecause of the loss of mass from the air phase. This mass islost from the residual water contained within the pore spaceand/or the soil organic carbon that is part of the soil matrix.The retardation factor, therefore, can be defined as the ratio ofthe rate of diffusion of the air to the rate of diffusion of thecontaminant front in the soil atmosphere. Ra. is also the mini-mum number of pore volumes that must pass through a three-phase contaminated soil system (soil, water, and air) to removethe contaminants. It is a minimum because the approachignores the effects of mass transfer limitations between phases.The effects result from heterogeneity and kinetics and unequaltravel times along flow lines from the edge of the contami-nated area to the vapor extraction well.
Another, more direct, method for obtaining effective diffusion coefficient) is through column tests such asthose used by Johnson et al. (1987). These column tests(Figure 11-5) use a dead-end column with a mixture ofnitrogen and the organic contaminants maintained at one end.The only process that can carry the contaminant into thecolumn is molecular diffusion. If a sampling line is fitted tothe interior of the column, then samples can be obtained overtime and the concentration breakthrough curve obtained. Thiscurve can be fitted to a one-dimensional analytical solution tothe diffusion equation to obtain a fitted, effective diffusionequation.
11.2.3 Processes Affecting Gas-Phase TransportSome insight into the migration of contaminants in the
vapor phase can be attained by considering the differentprocesses included in eq. 11-9. If there is no partitioning of thecontaminant between the gas phase and the soil (KP is zero),and if the Henry’s constant, KH, is large (i.e., there is no
where R1 is a gas-phase retardation factor that is defined by
[11-11]
where rb is the dry bulk density of the soil. This retardationfactor is a constant if the water content of the soil does notchange and is analogous to the retardation factor, R, for themovement of organic contaminants in the saturated zone. Thesecond term in eq. 11-11 represents the partitioning of thecontaminant from the gas phase to the water phase. The thirdterm represents the partitioning from the gas phase, throughthe water phase, to the solid phase. In this modified example,the retardation factor, Ra, must be determined. Methods fordetermining the physical parameters have been identified. Methods for obtaining the partition Figure 11-5. Column for measuring effective vapor phase
diffusion coefficients.
172
significant mass loss to the water phase), then the retardationfactor is close to unity and the contaminants move through theporous medium with the air. As the value of KH becomessmaller, the retardation factor at a given porosity and watercontent becomes larger because of the partitioning into thewater phase (Figure 11-6). For example, the amount of retar-dation for benzene increases with increasing water content(Figure 11-7) because it partitions into the water phase (i.e.,KH is small). In contrast, the retardation factor for pentane isinsensitive to changes in volumeric water content because itdoes not significantly partition into the water phase (Figure11-7).
These effects are also seen in column tests (Johnson et al.1987). The breakthrough curves for methane, trichloroethene(TCE), and chlorobenzene were obtained in a sand-filledcolumn under both dry and wet conditions (Figure 11-8).Methane, with the largest Henry’s constant and smallest Koc
values of the three compounds is observed to be the frost tobreak through. TCE and chlorobenzene break through laterbecause of the larger Koc and smaller KH values. The differ-ence between the damp sand and the dry sand reflects thedifferences in the Henry’s constants for the compounds. Inanother test, two columns were prepared, one containingvirtually no soil organic carbon (SOC) and another containingapproximately 1 percent SOC. The breakthrough curves formethane, octane, and benzene in these two columns (Figures11-9 and 11-10) demonstrate the role of SOC in the retarda-tion of the compounds. The differences in the breakthroughcurves for the three compounds in the column containing noSOC (Figure 11-9) can be attributed to the differences in theHenry’s constants. The column containing 1 percent SOCrequires more pore volumes to achieve breakthrough of theoctane and benzene, and the differences between the com-pounds are much greater (Figure 11-10). This increased retar-dation is the result of the greater partitioning of the contaminantfrom the gas phase to the SOC with the larger Koc values forthese compounds.
Temperature can have a significant influence on the rateof migration of gas-phase volatile organic contaminants. Thediffusion coefficients increase with increasing temperature.The effect of this temperature dependence can be calculatedfrom:
[11-12]
where T is the kelvin temperature. The exponent, m, shouldtheoretically be 1.5; however, experimental data yield valuesbetween 1.75 and 2.0 (Hamaker, 1972). The temperature alsoaffects the vapor pressure of the compounds (Figure 11-11)and, therefore, the concentration in the gas phase (eq. 11-2).The Henry’s constant also shows a temperature dependenceby increasing with increasing temperature (Figure 11-12).From the definition of the gas-phase retardation factor (eq. 11-11), the increased Henry’s constant is reflected as a decreasein Ra (Figure 11-13). Thus, fewer pore volumes of air need tobe moved through a contaminated soil to remove the contami-nant at 35°C than at 10°C.
The concentration of volatile organic contaminants in thegas phase of the unsaturated zone is influenced by the pres-
Figure 11-6. Retardation factor as a function of dimensionlessHenry's constant when there is no adsorption(after Johnson et al., 1987).
Figure 11-7. Retardation factor versus water content forbenzene and pentane (after Johnson et al., 1987).
173
Figure 11-8. Relatlve gas phaae concentration versus time in a column experiment (after Johnson et al., 1987).
Figure 11-9. Relative gas phase concentration versus the number of pore volumes of air moved through an unsaturated columnwith no soil organic carbon (after Johnson et al., 1987).
ence of boundaries that can impede the rate of migration.Spills of solvents and hydrocarbons often occur in industrialand urban areas where parking lots, roads, and foundationscan act as low permeability boundaries that limit the masstransfer of the contaminated gases from the unsaturated zoneto the atmosphere. In the absence of such barriers, concentra-tions of the contaminants in the soil-gas phase should remainlow very near the surface. When these barriers are present, theconcentrations in the soil-gas phase can be much greater. Theeffect of these impermeable caps is illustrated in the numeri-cal simulations by Baehr (1987). The mass of total hydrocar-bon that is in the soil-gas phase is about 2.5 times greaterwhen there is a cap present than when there is no cap (Figure11-14).
Measuring permanent gases such as O2 and CO2 in addi-tion to the priority pollutants can provide insight into pro-cesses that are occurring in the subsurface. Measurements ofthe distribution of total hydrocarbons in the unsaturated zonenear an oil spill in Bemidji, Minnesota, (Hult and Grabbe,1985) show that the concentrations are greatest near thesource and decrease with greater distance from the pooledmaterial (Figure 11-15a). O2 is near atmospheric values farfrom the source and nearly depleted near the spill area (Figure11-15b). CO2 distributions are opposite to those of oxygen,with the highest concentrations being near the source (Figure11-15c). While the diffusion and retardation processes dis-cussed above have an important role in the distribution of thetotal hydrocarbons, the depletion of O2 and the generation of
174
Figure 11-10. Relative gas phase concentration versus thenumber of pore volumes of air moved through anunsaturated column with 1% soil organic carbon(SOC) (after Johnson et al., 1987).
Figure 11-11. The vapor pressure of TCE as a function oftemperature.
CO2 near the source area suggest that biodegradation of thehydrocarbons is also Occurring (see Chapter 13).
There are other factors that affect the migration of or-ganic vapors. Cultural factors such as underground utilityconduits, trenches, sewers, and pipes can act as preferentialpathways along which these gases may travel. The type ofbackfill used around underground storage tanks affects thewater content and retardation of the gas-phase contaminants.Other environmental factors such as variations in atmosphericpressure, fluctuations in the elevation of the water table, andthe amount of infiltration in the contaminated area also have asignificant influence on transport of contaminants in soil gasat certain sites.
11.3 Vapor ExtractionVapor phase extraction is an important method for re-
moving residual volatile organic solvents from the subsurface.
I
Figure 11-12. Dimensionless Henry’s constant for TCE versustemperature.
Figure 11-13. Vapor retardation factor for TCE versus tempera-ture.
In principle, the technique works by removing the volume ofcontaminated air from the subsurface. As more air moves intothe contaminated area, the contaminants partition from theNAPL to the air phase. The extraction is continued untilsufficient pore volumes of air have passed through the con-taminated zone to remove the entire mass of the NAPL fromthe subsurface.
Many of the same problems that are encountered inground-water pump-and-treat systems also are expected invapor extraction systems. As the contaminated air is extractedfrom the unsaturated zone, the highly contaminated soil gasthat was initially present is removed. The concentrations maythen begin to decrease and remain at some concentration thatis substantially lower than the initial concentration but signifi-cantly higher than the target level (Figure 11-16). This “tail-ing” is the result of several processes. One factor is the rate ofresorption of the organic contaminants from the soil organicmatter (Figure 11-17). Although little is known about theserates, some data on hexachlorobenzene (Karickhoff and Morris,1985) show that the rate is initially rapid and decreases with
175
Figure 11-14. Comparison of total hydrocarbons present in asoil with a cap versus without a cap (after Baehr,1987).
time, and the equilibration may take more than 30 days (e.g.,Figure 10-11, Chapter 10); this time scale is significant com-pared to the rate of movement of the air. Another consider-ation is the form of the isotherm itself. If the isotherm isnonlinear (Langmuir- or Freundlich-type isotherm with theexponent less than unity), there can be tailing in the concen-tration versus time curves.
If the NAPL in the porous media is locally surrounded bywater, then the concentrations in the air that is being advected through the adjacent pores may be limited by the rate ofdiffusion through the water (Figure 11-18). In the air, velocity is low relative to the rate of diffusion (this is the case when there is no extraction), and the concentrations are limited by the vapor pressure of the compound. If the air velocities are large relative to the rate of diffusion, then the concentration of the contaminant is limited by diffusion through the water. Ananalogous situation may arise when thick pools of NAPL arebeing removed by vapor extraction. The more volatile compo-nents from the upper surfaces of the NAPL are removed first.If the air velocity is large relative to the rate of diffusion ofthose volatile components through the NAPL, then the con-centrations in the gas phase are limited by this rate of diffu-sion through the NAPL.
Another important aspect of NAPLs that are composed ofmore than one component is that as the more volatile compo- nents are removed, their concentration in the NAPL (as a mole fraction) decreases. This decrease in the mole fraction de- creases the vapor pressure (eq. 11-2) and hence the gas-phase concentration.
Soil heterogeneity plays a major role in controlling theconcentration of contaminants in gases extracted from theunsaturated zone. As NAPLs infiltrate into the subsurfacethey spread into pools on top of lenses of finer grainedmaterial within the aquifer. The NAPL also may be drawninto the fine-grained zones by capillary action: As air isadvected through the contaminated soil (Figure 11-19a), thoseparcels that pass through the fingers or very close to the poolsare close to saturation with respect to the NAPLs. The concen- Figure 11-15.tration of the volatiles in the parcels of air between the fingersand pools is controlled by the rate of vapor diffusion from the
Distrlbutlon of gases near an oil spill inBemidjl, MN (after Hult and Grabbe, 1985).
176
Tailing in the vapor concentration versus timecurve for a vapor extraction well.
Figure 11-18. The gas phase concentration controlled byeither the vapor pressure or the rate of diffusionof the volatile organic through the water phase.P denotes the residual product.
pools and fingers. During vapor extraction, the fingers arelikely to be removed before the pools (Figure 1l-19b), be-cause there is generally less mass of NAPL in the fingers thanin the pools. Also, a greater surface area of the NAPL exposedin the fingers facilitates the mass transfer to the advected air.As the residual fingers are removed, the concentrations in theextracted air should decrease as the less contaminated airbetween the pools is mixed with the more contaminated airfrom just above the pools. Removal from the pools mayrequire a substantial period of time (years) to complete. TheNAPL that moved into the finer grained zones is effectively
Figure 11-17. The concentration in the gas phase controlled byremoved from the advective air flow. Under these conditions,
the rate of desorption. the removal of NAPL is limited by the relatively slow diffu-
177
Figure 11-19. The transport of air through a heterogeneous media with fingers and pools of NAPL present (A) and at a later timewhen only pools are present (B).
sion of the NAPL out of the finer grained zone. If a substantialmass of NAPL is trapped in this way, remediation couldrequire many years to complete.
11.4 ReferencesBaehr, A.L. 1987. Selective Transport of Hydrocarbons in the
Unsaturated Zone Due to Aqueous and Vapor PhasePartitioning. Water Resources Research 23(10):1926-1938.
Feenstra, S. and J.A. Cherry. 1987. Dense Organic Solvents inGround Water An Introduction. In: Dense ChlorinatedSolvents in Ground Water, Progress Report 0863985,Institute for Ground Water Research, University of Wa-terloo, Waterloo, Ontario.
Hamaker, J.W. 1972. Diffusion and Volatilization. In: Or-ganic Chemicals in the Soil Environment Vol. 1., C.A.I.Goring and J.W. Hamaker (eds.), Marcel Dekker, NewYork, Chapter 5.
Huh, M.F. and R.R. Grabbe. 1985. Distribution of Gases andHydrocarbon Vapors in the Unsaturated Zone. In: Pro-ceedings, U.S. Geological Survey Second Toxic WasteTechnical Meeting, Cape Cod, MA, October 1985, U.S.Geological Survey Open File Report 86-0481, pp. C21-C25.
ter-prevention, Detection and Restoration, NationalWater Well Association, Dublin, OH, pp. 493-507.
Karickhoff, S.W. and K.R. Morris. 1985. Sorption Dynamicsof Hydrophobic Pollutants in Sediment Suspensions.Environ. Toxicol. Chem. 4:469-479.
Mabey, W.R., et al. 1982. Aquatic Fate Process Data forOrganic Priority Pollutants. EPA 440/4-81/014 (NTISPB87-169090), Chapter 4.
Mereer, J.W. and R.M. Cohen. 1990. A Review of ImmiscibleFluids in the Subsurface Properties, Models, Character-ization and Remediation. J. Contaminant Hydrology 6:107-163.
Millington, R.J. 1959. Gas Diffusion in Porous Media. Sci-ence 130100-102.
Palmer, C.D. and R.L. Johnson. 1989a. Physical ProcessesControlling the Transport of Non-Aqueous Phase Liquidsin the Subsurface. In: Transport and Fate of Contami-nants in the Subsurface, EPA/625/4-89/019, pp. 23-27.
Palmer, C.D. and R.L. Johnson. 1989b. Determination ofPhysical Transport Parameters. In: Transport and Fate ofContaminants in the Subsurface, EPA/625/4-89/O19, pp.29-40.
Johnson, R.A., C.D. Palmer, and J.F. Keely. 1987. Mass Schwille, F. 1967. Petroleum Contamination of the Subsoil -Transfer of Organics Between Soil, Water, and Vapor A Hydrological Problem. In: The Joint Problems of thePhases: Implications for Monitoring, Biodegradation and Oil and Water Industries, Peter Hepple (ed.), Elsevier,Remediation. In Proc. NWWA/API Symp. on Petroleum New York, pp. 23-53.Hydrocarbons and Organic Chemicals in Ground Wa-
Schwille, F. 1988. Dense Chlorinated Solvents in Porous andFractured Media. Lewis Publishers, Chelsea, ML
178
Chapter 12Physiochemical Processes: Inorganic Contaminants
Carl D. Palmer and William Fish
Inorganic compounds are a common and widespreadclass of contaminants at many hazardous waste sites. Therelative importance of these inorganic contaminants is illus-trated in the Records of Decision (RODS) signed by EPAbetween 1982 and 1986 (Booz-Allen and Hamilton, Inc.,1987). Of the 108 RODS, 56 percent involved Superfund siteswhere inorganic compounds were designated as a potentialthreat or problem (Palmer et al., 1988). While organics/volatile organic compounds (VOCs) are the most frequentlyreported contaminants, heavy metals and inorganic are thesecond and third most frequently reported categories of haz-ardous substances (Figure 12-1). Inorganic waste problems
Figure 12-1.
Primary Hazardous Substances Detected
Primary hazardous substances detected athazardous waste sites. Based on data from Booz-Allen and Hamilton, Inc. (1987), from Palmer et al.(1988).
can be particularly severe because they often occur at sitesthat cover many square miles. Remediation of such sites isoften difficult simply because of the size of the affected area.
The most common inorganic constituents of concern arethe 13 priority pollutant metals (Table 12-1). However, otherinorganic substances such as nitrate, phosphate, cyanide, andradionuclides may be found at levels far exceeding drinkingwater standards. Iron, manganese, aluminum, calcium, silica,and carbonates are not priority pollutants but also can contrib-ute to the overall cost of a remediation scheme by increasingmaintenance and disposal costs. For example, air strippers atseveral sites have been temporarily disabled by iron precipita-tion problems.
The behavior and toxicity of inorganic compounds areaffected by chemical and physical processes. Understandingthese processes may lead to more cost-effective restoration ofcontaminated sites and can reveal how inorganic substancesmay affect the cleanup of sites, even at sites where organiccontaminants are the main concern.
Table 12-1. The 13 Priority Metals
179
12.1 Chemical ProcessesSeveral chemical processes affect the concentration, spe-
cific form, rate of transport, and ease of removal of inorganicsubstances from the subsurface. These processes include (1)chemical speciation, (2) oxidation/reduction, (3) dissolutionand precipitation of solid phases, (4) ion exchange and ad-sorption onto the soil matrix, and (5) transport of particles inthe subsurface. One of the difficulties in working with inor-ganic contaminants is that all of these processes can beoperating simultaneously. Therefore, it is sometimes difficultto determine which is the most important at a particular site.The relative importance of these processes not only variesfrom site to site but may vary from one area to another withina given site. The relative importance of these processes alsomay change during the cleanup operation as subsurface chemi-cal conditions are altered. Each of these processes is discussedbelow.
12.1.1 SpeciationWhen a water sample is sent to the laboratory for metals
analysis, the results are usually returned as total concentra-tions. In reality, the metals interact with anions (or ligands)that exist in the ground water to form different chemicalspecies or “aqueous complexes.” For example, cadmium may
complexes. In addition to Cl- and OH-, petal ions can com-
phates. Complex formation of transition metals is moreextensive than for other metals. The relative tendency ofmetals to form complexes is in the order
(Hanzlik, 1976). The concentration of aqueous complexesdepends upon the concentration of the metal ion of interest,the ligand with which it forms the complex (Figure 12-2), and
Figure 12-2. Fraction of total Cd in various chemical com-plexes versus the Cl- concentration (after Mooreand Ramamoorthy, 1984).
the concentration of the other metal ions that may compete forthe ligand. Because chemical reactions usually are governedby the amount of free ion rather than by the total metal,knowledge of the concentration of complexes is important toproperly identify the role of processes such as adsorption,mineral dissolution, and precipitation.
The formation of inorganic aqueous complexes is often arapid process that can be described by equilibrium expres-sions. For example, the formation of the mercuric chloridecomplex, HgCl+, can be written as the chemical reaction:
Hg2+ + Cl- + HgCl+
which can then be written in terms of an equilibrium constantas:
Ka= {Hg2+} {C1-}/{HgCl +} [12-1]
where the braces represent the thermodynamic concentrationor “activity” of the chemical species. The activity of thespecies is simply the product of the concentration of thatchemical species and an “activity coefficient,” y. For example,the activity of Hg2+ can be written as:
[12-2]
where Ka, is an equilibrium (or stability) constant for the abovereaction and the brackets indicate the concentration of thechemical species. The activity coefficient accounts for thechange in chemical reactivity due to electrostatic interactionsamong ions in solution. Several methods are used to calculateactivity coefficients. See Stumm and Morgan (1981), Morel(1983), or Sposito (1986) for more detailed discussions ofthese methods.
Calculation of the concentrations of each chemical spe-cies in solution requires the total concentration of each metaland each ligand in solution as well as the equilibrium constantfor the formation of each complex. Total concentrations areobtained from chemical analysis of the water sample. Equilib-rium constants (or more fundamental thermodynamic data)for inorganic complex formation have been studied by manyresearchers and are tabulated in a variety of references (e.g.,Ball et al., 1980; Felmy et al., 1984; Smith and Marten, 1976).An extensive list of thermodynamic data sources is given byNordstrom and Munoz (1985).
While in principle it is easy to calculate the concentrationof these aqueous complexes, it does not take too many combi-nations of metals and ligands to result in a series of equationsthat are unmanageable to solve by hand. Fortunately, compu-tational algorithms can quickly perform these calculations, sotime can be better invested in interpreting the results of thecalculations.
12.1.2 Dissolution/PrecipitationGround water passing through an aquifer may be in direct
contact with a wide variety of mineral phases. Dissolution orweathering of those mineral phases contributes greatly to thenatural chemical composition of the ground water. Generally,
180
dissolution refers to a reaction where all of the chemicalspecies that comprise the mineral come into solution. Somecommon minerals that undergo dissolution in shallow aquifersystems include gypsum (CaS04• H2O), calcite (CaCO3) andquartz (Si02). Weathering is a partial dissolution process inwhich certain ions come into solution while others remain aspart of secondary mineral phases. During the initial stages ofweathering of aluminosilicates, Ca2+, Mg2+, Na+, K+, and someH4SiO go into solution while the remaining ions become clayssuch as kaolinite and montmorillonite, If the concentrations ofcertain ions are sufficiently high, they may be removed fromsolution by the formation of a solid phase (i.e., by precipita-tion). These precipitated minerals may dissolve later if physi-ochemical conditions within that portion of the aquifer change.
These types of reactions can have a great effect on theconcentrations of priority pollutants within the aquifer. Forexample, the weathering of fly-ash piles can yield selenium,arsenate, lithium, and heavy metals (Honeyman et al., 1982;Murarka and Macintosh, 1987). BaCrO4 may precipitate in anaquifer contaminated with Cr04 and later it may dissolveduring remediation, thus prolonging the time required torestore the aquifer. Pb can be removed from solution byprecipitation as PbCO3. In addition, dissolution of naturallyoccurring minerals can neutralize acid or alkaline waters andthus enhance the adsorption or precipitation of priority metals.The precipitation of calcite or hydrous ferric oxides mayreduce the permeability of the aquifer, clog well screens, andincrease the cost of treatment and disposal. Therefore, it isadvantageous to be aware of the potential for these types ofreactions and to include their effects in the cost and design ofremediation.
Equilibrium between the ground water and a solid phasecan be expressed in terms of an equilibrium constant (orvolubility product). For example, the dissolution or precipita-tion of BaCrO4, which involves two priority metals, is writtenas:
and at equilibrium this can be expressed as:
[12-3]
where Ksp is the volubility product. As with the aqueouscomplexes, the terms in the braces refer to the activity of theparticular species and not the total concentration of the ele-ments. The equilibrium constants for many solid phases canbe found in the same references given for the aqueous com-plexes. The ability of a ground water to dissolve or precipitatea solid phase is sometimes expressed as a saturation index(SI), which is defined as:
[12-4]
where the IAP or “ion activity product” is the same expressionof ion activities used for the volubility product but at theconcentrations found in the ground water rather than at equi-librium. For example, the ion activity product for bariumchromate is {Ba2+) {Cro2-
4}. If the ground water is in equilib-rium with the solid phase, then the IAP is equal to the KSP and
the saturation index is equal to zero. If the saturation index isless than zero, the water is undersaturated with that solidphase and the ground water has the potential to dissolve thatphase if it is present. If the saturation index is greater thanzero, then the water is supersaturated with respect to the solidphase and has the potential to precipitate that phase providedthat the reaction is fast enough to occur with the time scales ofinterest. Calculation of the saturation indices for mineralphases requires knowledge of the concentration of the aque-ous complexes. Therefore, the computational tools used tocalculate the concentration of aqueous complexes also can beused to calculate saturation indices.
Saturation indices at waste sites can be useful for identi-fying potential sources and sinks for metal ions. If calculatedsaturation indices for PbCO3 are close to zero, then Pb islikely being removed from solution and may not move veryfar from the source. If the saturation indices for BaCrO4 areclose to zero, there may be a large reserve of solid phase in theaquifer that could prolong a pump-and-treat remediationscheme for the removal of the CrO4. If the waters of interestare undersaturated or supersaturated with respect to solidphases of interest, the amount of the solid phase that must bedissolved or precipitated per liter of water to reach equilibrium can be calculated. Such theoretical calculations may beparticularly useful in evaluating the potential for mineralprecipitation as a result of mixing contaminated and uncon-taminated waters in extraction wells. Note that saturationindices alone do not prove the presence or absence of amineral phase. However, saturation indices are relatively easyto obtain and are valuable for identifying possible mineralphases.
12.1.3 Oxidation/ReductionThe number of electrons associated with an element
dictates its oxidation state. Elements can exist in severaloxidation states. For example, iron commonly exists in the +2or +3 state, arsenic as +3 or +5, and chromium as +3 or +6.Oxidation-reduction (redox) reactions involve a transfer ofelectrons and, therefore, a change in the oxidation state ofelements. Redox reactions are important to studies of subsur-face contamination because the chemical properties for theelements can change substantially with changes in the oxida-tion state. For example, in slightly acidic to alkaline environ-ments, Fe(III) is fairly insoluble and precipitates as a solidphase (hydrous ferric oxide) that has a large adsorption capac-ity for metal ions. In contrast, Fe(II) is fairly soluble and itsoxides have a much lower adsorption capacity. As the Fe(III)solid phase is reduced, not only is the Fe(II) brought intosolution but so are any contaminants that may have beenadsorbed onto it (Evans et al., 1983; Sholkovitz, 1985). An-other environmentally important redox-active element is chro-mium. Hexavalent chromium, Cr(VI), exists in ground wateras the relatively mobile and toxic anions HCrO-
4 and CrO2-
4.The reduced form of chromium, Cr(III), is a cation that underslightly acidic to alkaline conditions is fairly insoluble, readilyadsorbed, and much less toxic than Cr(VI). Selenium alsoundergoes important redox transformations. Selenate (Se(VI))is more mobile and less toxic than selenite (Se(IV)).
181
Because redox reactions involve the transfer of electrons,the change in oxidation state of one element necessitates achange in the oxidation state of another. For example, asCr(Vl) is reduced to Cr(III), it must gain three electrons fromanother element. One possible electron donor is ferrous iron(Fe(II)):
Redox reactions cannot occur unless there is both a suitableelectron donor and a suitable electron acceptor.
The expected form of an element at equilibrium dependson the pH and the redox state of the water. The redox state ismeasured by an electrical potential (in volts (v) or millivolts(mV)) at a standard electrode. This potential is called the EHof the ground water. Alternatively, the redox conditions aresometimes reported in terms of the “pe,” which is the negativelogarithm of the activity of the electron. This is a directanalogy to pH, which is the negative logarithm of the activityof H+. The EH and pe of a watar measure the same property,but due to differences in definition they are not numericallyequivalent. Redox conditions within natural aquifers varyfrom highly oxidizing conditions (high EH, - +800-900 mV)to very reducing conditions (low EH,~ -200 mV). Variationwithin contaminated aquifers is at least as great as this range,but often is marked by abrupt transitions over much smallerscales than is typical of uncontaminated aquifers (see Sections8.3.3 and 8.4).
The conditions of pH and EH for which a particular redoxspecies is theoretically stable are represented graphically onan EH-pH diagram. These diagrams are also known as pe-pHand Pourbaix diagrams. For example, the EH-pH diagram forFe (Figure 12-3) illustrates the predominance of the ironhydroxide solid at slightly acidic to alkaline conditions andhigh EH conditions, whereas aqueous Fez+ predominates atlow EH and slightly alkaline to acid conditions. Methods forthe construction of such diagrams can be found in Garrels andChrist (1965), Drever (1989), and Stumm and Morgan (1981).A collection of EH-pH diagrams for many metals was preparedby Brookins (1988).
In theory, knowledge of the EH, pH, and total elementalconcentrations in a ground water allows the quantitative pre-diction of the concentration of each redox-active species insolution. However, many redox reactions are microbiallycatalyzed, nonreversible, and, therefore, not found in a state ofmutual equilibrium. Redox species that are not at equilibriumwith each other often have been observed to occur together(Lindberg and Runnels, 1984). Except for a very few situa-tions, it is impossible to predict general redox behavior inaquifers using equilibrium concepts. Nonetheless, there areobservable and consistent trends in redox conditions in naturalaquifers that suggest that at least qualitative estimates ofbehavior are possible. Because of the importance of redoxreactions it is important to know at least the possible changesin the redox state. Some idea of the transformations that maybe taking place within the aquifer can be obtained in certaincases (e.g., for Cr, Se, As) by directly measuring the concen-trations of different oxidation states of the contaminants.More must be learned about the rates of redox reactions if
Figure 12-3. pe-pH diagram for the Fe-H2O system.
these reactions are to be put into proper perspective withregard to the transport and removal of inorganic contaminantsfrom the subsurface. A review by Fish (1990) summarizesthese concepts, problems, and some possible solutions.
12.1.4 Adsorption/Ion ExchangeIon exchange and adsorption can exert a great influence
on the concentrations of ions in solution. Clay minerals areimportant ion-exchangers in subsurface systems. During ionexchange, ions in certain layers of the three-dimensionalstructure of clays are replaced by ions in solution, while aconstant total charge within the clay layer is maintained. Forexample, Ca-Na exchange can be written as:
and using the Vanselow (1932) convention can be expressedin terms of a “selectivity coefficient”, Ks, as:
[12-5]
where NaX and CaX represent the Na and Ca on the clay andXNaX and XCaX are the mole fractions of exchangeable sitesoccupied by Na and Ca, respectively. The selectivity coeffi-cients are empirical and vary with the concentration of thecations in the ground water (Reichenberg, 1966) so thatlocation-specific values must be used.
182
Knowledge about ion exchange is important to under-stand the binding of alkali metals and the alkaline earths andsome anions to c!ays and condensed humic matter (Sposito,1984; Helfferich, 1962). However, ion exchange does notadequately describe the interaction of many transition metalswith mineral surfaces. These interactions are better describedby adsorption processes. Ionic adsorption involves the coordi-nation bonding of metal cations and anions to surfaces ofminerals exposed in the pore space of the aquifer. In ionexchange, the total electrostatic charge of the solid phase isconstant, whereas for adsorption, the charge of the surfacevaries with solution pH and the amount of ions adsorbed. Atthis time, no generally accepted model for the adsorption ofinorganic ions exists, so several approaches are discussedbelow.
As with organic solutes, adsorption isotherms for inor-ganic compounds can be constructed and Langmuir- andFreundlich-type isotherms can be utilized. In general, anionstend to follow Langmuir isotherms while cations tend tofollow Freundlich isotherms (Dzombak, 1986). For the or-ganic contaminants considered in Chapter 10, linear isothermswere found for many compounds allowing a constant retarda-tion factor to be applied in transport calculations. There is nomodel of comparable simplicity that can properly describe thetransport of inorganic substances in the subsurface. One of thekey reasons why an analogous model does not work forinorganic constituents is the strong dependence of the amountof adsorption on the pH of the ground water. This is illustratedin a “pH-edge” which is a plot of the fraction of the total massof a metal adsorbed versus the pH of the solution (Figure 12-4).
Figur
For cations, very little of the metal is adsorbed onto theaquifer material at low PH. As the pH increases, the fractionadsorbed onto the soil increases until virtually all of the metalis adsorbed or all metal-binding sites are occupied. The exactposition of the pH-edge on the diagram depends on thespecific ion considered, the concentration of the metal, theamount of adsorbent, and the concentration of other ions insolution, These additional ions may compete for adsorptionsites or form complexes with the metal ion and shift the edge(Figure 12-5) (Benjamin and Leckie, 1981). The pH-edges foranions are approximately mirror images of those for cations,with maximum adsorption Occurring at low pH and decreas-ing with increasing pH. Any useful model of ionic adsorptionof metal ions should account for the pH dependence ofadsorption. The so-called “surface complexation models” forion adsorption meet this criterion (e.g., Stumm et al., 1976Schindler, 1981; Schindler and Stumm, 1987; Dzombak andMorel, 1990). These models have a foundation in chemicaltheory and if used for a well-defined system, may be appliedover a somewhat wider range of conditions than the specificexperiments used to determine the particular model param-eters. Surface complexation models are based on the conceptthat ions form complexes with solid-phase atoms at the oxide/solution interface. These complexes are analogous to com-plexes formed in solution between metals and ligands. Thereis, however, an added difficulty due to the formation of asurface charge at the oxide-water interface.
An oxide can be viewed as an array of metal ions andoxygen atoms (Figure 12-6a). When the oxide is immersed inan aqueous solution, water molecules arrange themselvesaround the surface metal ions (Figure 12-6b). Some of theseadsorbed water molecules then dissociate, and the resultinghydrogen ions bind to the adjacent surface oxygen atoms(Figure 12-6c). Adsorption of anions and cations to the oxidesurface can be described as an exchange of the metal ions forthe H+ and the ligands for the OH groups on the surface of the
183
Figure 12-6. Formation of a hydroxylated oxide surface inwater (after Schindler, 1981).
oxide (Figures 12-7 and 12-8). Such adsorption reactions canbe written in a manner similartions discussed above:
where the svmbol = denotes
to the solution complex reac-
the oxide surface, M2+ is adivalent cation, and L- is a monovalent anion. Equilibriumconstant expressions for these reactions are
Several surface complexation models have been devel-oped. The most important feature that distinguishes one sur-face complexation model from another is the treatment of theelectrostatic term. Computation of the electrostatic effects onadsorption requires postulation of a particular arrangement ofelectrical charges near the surface. These charge distributionsare developed by hypothesizing one or more layers of chargesnear the surface. The three types of surface complexationmodels often used are two-layer models, Stem-layer models,and triple-layer models. More detailed discussions of thesemodels are given by Dzombak (1986) and Dzombak andMorel (1990). These models represent increasing complexityin the geometric view of the oxide-water interface and requirean increasing number of fitted parameters in addition to theequilibrium constants. Despite apparent differences in theirsophistication, all three of these surface complexation modelsequally describe acid-base titration data for oxide surfaces(Westall and Hohl, 1980; Dzombak, 1986). In addition,Dzombak (1986) found that the simpler two-layer model wasquite capable of modeling anion adsorption and cation adsorp-tion if the cation concentrations were not extremely high.Therefore, the choice of model is best based on which is themost parsimonious; the obvious choice in many cases is thetwo-layer model.
One of the major disadvantages to using this type ofadsorption model is a lack of knowledge about the equilib-rium constants. The problem in using constants reported in theliterature is that they are specific to the adsorption model usedto fit the experimental data. Therefore, reported constants arequite dissimilar from those appropriate for a different adsorp-tion model. To overcome this limitation, Dzombak (1986) andDzombak and Morel (1990) reinterpreted the raw data fromadsorption experiments for different ions on hydrous ferricoxides (HFO) using the basic two-layer model. This effortprovides a set of consistent constants based on a commonadsorption model and makes more widespread use of suchmodels feasible. If the main adsorbent in the aquifer is HFO,the derived constants provided by Dzombak and Morel (1990)should be adequate. However, natural porous media maycontain many oxides and other surfaces to which metal ions
These equations are analogous to the equilibrium expressions
describe the effect of the electrostatic charge near the surfaceof the oxide. Figure 12-7. Adsorption of a divalent cation on a hydroxylatad
oxide surface (from Palmer et al., 1988).
184
Figure 12-8. Adsorption of a monovalent anion on a hydroxy-Iatad oxide surface (from Palmer et al., 1988).
can bind (Figure 12-9), and a consistent set of equilibriumadsorption constants has yet to be derived for other types ofsurfaces. It also is unclear whether mixtures of surfaces arelinearly additive.
12.2 Particle TransportA potentially important mechanism for the migration of
inorganic substances is particle transport. These particles maybe inorganic, organic, or biological and may include bacteria,viruses, natural organic matter, inorganic precipitates, asbes-tos, and clay. Inorganic ions may migrate as integral constitu-ents of the particles or they maybe adsorbed onto the surfacesof the particles. The distance these particles move dependsupon the size of the particles relative to the size of the poresthrough which they must pass as well as the chemical condi-tions in the subsurface.
Particles can be removed from solution by three majormechanisms: (1) surface filtration, (2) straining, or (3) physi-cal-chemical processes (McDowell-Boyer et al., 1986). If theparticles are larger than the largest pores within the aquifer,they cannot penetrate the aquifer and they are filtered out atthe interface between the medium and the source of theparticles (surface filtration). If the particles are smaller thanthe largest pores but larger than the smallest pores, the par-ticles can travel some distance into the aquifer before theyencounter a pore through which they cannot pass; they willthen be strained from solution (straining). If the particles aresmaller than the smallest pores in the medium, then they cantravel great distances. However, the particles still can beremoved from solution by adhering to the sand grains becauseof collision. Collision with the sand grains occurs as a resultof sedimentation, interception, and Brownian motion. Par-ticles in the subsurface also can aggregate if chemical condi-tions such as pH or ionic strength change significantly(physical-chemical processes). The particle aggregates thencan be removed from the water by straining.
Particle transport is not likely to be an important factor inevery environment. Therefore, it is useful to target thosesituations where particle transport is most likely to be impor-tant Such situations include environments where there arehigh concentrations of organic carbon, dissolved solids, orsuspended solids. Movement of particles may be induced in
areas where the flow rates are very high, either because ofnatural flow conditions or more commonly because of highpumping rates. Any time there is an abrupt transition in pH orrcdox conditions within the aquifer subsurface, there is anopportunity for the precipitation of colloidal-size particlesthat can travel through the aquifer. When a water appears tobe supersaturated with common mineral phases that are nor-mally expected to be at equilibrium, then particle transportshould be suspected. A simple example of the latter would behigh iron concentrations in the presence of oxygen undermildly acidic to alkaline conditions. The iron is likely toprecipitate as a fine colloid and to be included in the total ironanalysis of the water (Fish, in press). If the particle transport isbelieved to be important, there are several techniques forparticle detection. These techniques include filtration, micros-copy, electrophoresis, and light scattering. A review of light-scattering techniques is provided by Rees (1987).
Particle transport has been documented for at least twodifferent ground-water contamination sites. It has been ob-served in a contaminant plume emanating from rapid infiltra-tion beds used to recharge treated sewage to a sand and gravelaquifer at Otis Air Force Base on Cape Cod, Massachusetts(Gschwend and Reynolds, 1987). Particles near the sourcewere relatively small (< 6nm) but down-gradient from theinfiltration beds, apparently mobile particles about 100 nm indiameter were observed (Figure 12-10). Chemical analysis ofthe particles indicated that they were composed of Fe(II) andPO4 and may be the mineral vivianite (Fe,(PO4)2). As thetreated sewage, which was high in organic carbon, was re-charged to the aquifer, reducing conditions were created withinthe plume. Reductive dissolution of the naturally existingHFO or other iron-containing minerals resulted in elevatedlevels of Fez+ in solution. Phosphate entered the aquifer inwastewater percolating through the infiltration basins. Thephosphorus originated from detergents used prior to the mid-1970s. As phosphate entered the aquifer with the rechargewater and mingled with Fe(II), the water eventually reachedsaturation and began to precipitate the Fe-phosphate solidphase (Figure 12-11). However, the precipitate remained insolution as fine particles that migrated through the sand andgravel aquifer.
Another documented example of particle transport in-volves the migration of radionuclides at the Nevada Test Site(Buddemeir and Hunt, 1988). Large volumes of water werepassed through a series of ultrafilters to measure the concen-tration of radionuclides in different size fractions of particles.The results (Figure 12-12) indicate that some nuclides such as125Sb are almost entirely in solution (<3 nm), while nuclidessuch as 54Mn are transported on relatively large particles. Yetother nuclides, such as 105Ru, are not associated with anyparticular particle size but are evenly distributed over thedifferent particle sizes.
The definition of what is a molecule in solution and whatis a colloidal particle is arbitrary. For many years the 450-nmpore size was used as the standard break between what is insolution and what is a particle. However, this pore sizecorresponds to molecular weights in the range of severalhundred thousand atomic-weight units. Materials of this sizeare now thought to be more properly defined as colloidal
185
Figure 12-9. Natural porous media containing many different types of adsorption surfaces (from Palmer et al., 1988).
Figure 12-10. Particle sizes measured in the subsurface at theOtis Air Force Base, Cape Cod, MA (modified fromGschwend and Reynolds, 1987).
particles. Consequenty, smaller pore-diameter filters are in-creasingly used to distinguish “particles” from “solutes.” How-ever, there is no objective standard of what constitutes theproper dividing size.
Particle transport is particularly important for sites wherethe contaminant is highly toxic and the general expectation isthat the contaminant is not mobile because of its high affinityfor adsorption. However, if the material to which it is ad-sorbed is fairly mobile, then the sorbate may move rapidlybeyond the site.
For remediation efforts, particle transport can be a benefitor a liability. If particle transport is significant, it may bepossible to remove a significantly greater mass of contamin-ant per unit time (hence per unit cost) than if the contaminantwere adsorbed onto immobile particles. However, particlescan plug injection wells or aggregate in the subsurface andreduce the permeability of the formation near the extractionwells. These effects may increase the overall cost of aquiferrestoration if filter presses and longer pumping schedules arerequired to overcome these problems.
12.3 Organic-Inorganic InteractionsMixtures of many types of wastes are found in landfills,
dumps, and ground-water contamination sites. Consequently,it is not unusual to find both inorganic and organic contami-nants together. For example, an analysis of leachate from amunicipal landfill in Brookhaven Town, Long Island, NewYork, revealed 660 µg/L Cr (VI), 127 µg/L lead, 151 µg/L
186
Figure 12-11. Formation of particles at the Otis Air Force Base site, Cape Cod, MA.
xylenes, 40 µg/L methylene chloride, 27 µg/L naphthalene,and 25 µg/L benzene (Black and Heil, 1982). At Woburn,Massachusetts, Cr (VI) levels of over 2,000 µg/L were mea-sured in ground-water samples that contained high levels of p-chloro-m-cresol, phenol, p-nitrophenol, N-nitroso-diphenylamine, phthalate esters, and 35 other organic com-pounds (Cook and DiNitto, 1982). The behavior of inorganicconstituents in such waste mixtures can be dramatically dif-ferent from their behavior when the inorganic contaminantsare found by themselves.
The interactions among organic and inorganic compoundscan be classified as either direct or indirect. Direct interactionsinclude processes such as complexation (chelation) of metalions with organic solutes, oxidation-reduction reactions be-tween organic and inorganic constituents, and the competitionbetween organic and inorganic solutes for adsorption sites.Indirect interactions refer to the changes in pH and redoxconditions as a consequence of degradation of organic con-taminants in the subsurface. Most of the research on directinteractions between organic and inorganic materials has fo-cused on finding or characterizing synthetic pathways for thecommercial production of chemicals, for example, the oxida-tion of alcohols with Cr(VI) to produce aldehydes or carboxy-lic acids. Often this research has been conducted under extremeconditions of concentrations and pH that are of little environ-mental significance. The few studies that are of environmentalinterest indicate that organic-inorganic reactions are impor-tant in several situations. Stone (1986) found that phenols canbe oxidized in the presence of MnOz. Voudrias and Reinhard(1986) reviewed several investigations of the oxidation oforganic compounds by metal-substituted clays. Laha and Luthy
(1990) studied the oxidation of aniline and other aromaticamines by MnOz. Fish and Elovitz (1990) observed the reduc-tion of hexavalent chrome by cresols. They found that the rateof reduction was strongly dependent on the pH and theparticular isomer involved in the reaction. While the implica-tions of these results for remediation have yet to be seriouslyconsidered, the results may have implications on the design ofsystems where waters may potentially mix in extraction wellsand treatment trains.
Even at sites contaminated only by organic compounds,inorganic constituents cannot be ignored. While elevated con-centrations of inorganic ions may result directly from wasteleachate, they also may result from mobilization of naturallyoccurring ions in response to the changing pH and redoxconditions induced by organic contamimnts. Such changingconditions are typical consequences of biodegradation (Figure12-13). Biodegradation consumes oxygen, thereby decreasingthe EH (pe) within the contaminant plume. COZ, a by-productof biodegradation, forms carbonic acid and decreases the pHwithin the plume. Organic acid by-products also may decreasethe pH. These processes can result in the resorption of metalions and the dissolution of hydrous ferric oxide (an importantadsorbent).
An example of such conditions is found at a creosoteplume in Pensacola, Florida (Cozzarelli et al.( 1987). Elevatedconcentrations of Fe, COZ, and CH4 and depleted con-centrations of dissolved oxygen and NO-
3 are associ-ated with the biodegrading creosote plume. Barium,molybdenum, manganese, nickel, and strontium are as muchas two orders of magnitude greater than background levels.
187
Figure 12-14. Application of mass balance computationalmodels.
These chemical alterations are found over a much wider areaboth horizontally and vertically than the organic contaminantsthemselves. Similar conditions also have been described at acrude oil spill in Bemidji, Minnesota, (Siegel, 1987) whereelevated concentrations of iron, aluminum, and silica arereported. While many of these elements are not toxic they cannevertheless pose costly problems of scaling and cloggingduring pump-and-treat remediation.
12.4 Computational ToolsThe sections above discussed the importance of complex-
ation to understanding the controlling processes of site reme-diation with inorganic contaminants. Determining theconcentration of each of the complexes is a computationallycomplex task that is left to computational algorithms. Inaddition, other types of computationally intensive chemicalcalculations would be useful, and there are a variety of com-putational tools to assist in such calculations. In general, thesetools can assist in the calculation of (1) mass balance, (2)chemical speciation, (3) mass transfer, and (4) multicompo-nent transport. While some of these algorithms are readilyavailable at little or no cost, others are still classified asresearch tools and are not likely to be available for general usefor several years.
12.4.1 Mass BalanceMass balance calculations can be applied to a system
such as that illustrated in Figure 12-14. If the chemical com-position of the water is known at locations A and B along theflow path, then the change in concentration of each of theelements along the flow path is known. If the reactions thattake place between the two wells are known, then the amountof each reaction can be calculated. Reactions such as (1)mineral dissolution/precipitation, (2) gas exchange, (3) ionexchange, (4) simple isotope balances, (5) oxidation-reduc-tion, and (6) the mixing of waters can be included in suchcalculations. The code BALANCE (Parkhurst et al., 1982) is areadily available FORTRAN code that can run on personalcomputers. So far this code has been used to study thegeochemical evolution of natural waters (e.g., Plummer and
Back, 1980), yet it has not been applied to the transport ofcontaminants or the performance of remediation activities. Amass balance model such as BALANCE should not be usedby itself but should be used in conjunction with chemicalspeciation and mass transfer tools as well as practical knowl-edge of chemical systems.
12.4.2 Chemical SpeciationChemical speciation algorithms are used to calculate the
concentration and activities of each of the chemical speciesthat are in solution. The data requirements for the proper useof such models include accurate field pH, temperature, andalkalinity. In addition, a complete inorganic chemical analysisis required. A complete chemical analysis requires the con-centration of all of the major anions and cations and thepriority metals and anions under investigation. Some knowl-edge of the redox conditions within the aquifer is useful,particularly the total concentration of each of the redox statesof the metals of concern. Most chemical speciation modelsalso calculate and print out the mineral saturation indices.
There are several chemical speciation models available.WATEQ4F and SOLMNEQ88 are versions of models thatwere originally published in the mid- 1970s by the UnitedStates Geological Survey (e.g., Kharaka et al., 1988; Balletal., 1979; Kharaka and Barnes, 1973; Plummer et al., 1976Truesdell and Jones, 1974). Chemical speciation also is per-formed by mass transfer models (see below); it maybe morepractical to have a single program for all such calculations.
12.4.3 Mass TransferMass transfer models allow calculation of how much of a
given mineral phase must react for the water to reach equilib-rium with that phase and achieve the pH and EH of theequilibrated solution. The basic data requirements for the useof this type of model are similar to those for chemical specia-tion models. There are several available mass transfer models,including PHREEQE (Parkhurst et al., 1980), EQ3/6 (Wolery,1979, 1983), and MINTEQ (Felmy et al., 1984). MINTEQ hasan extensive data base that includes many of the prioritymetals that are of interest at waste sites. MINTEQ is also theonly model that includes choices for adsorption processes,including (1) ion exchange, (2) Langmuir isotherms, (3)Freundlich isotherms, (4) double-layer model, (5) Stem-layermodel, and (6) triple-layer model.
12.4.4 Multicomponent TransportThe ultimate tool for assisting in the design of aquifer
remediation strategies is a computational algorithm that ac-counts for the physical process of advection as well as all ofthe chemical processes discussed above. While progress hasbeen made in this area (Jennings et al., 1982; Yeh andTripathi, 1989), these models are not generally available.Although these models are still considered to be researchtools, there is much work being done to complete models thatsoon will be available for general use.
189
12.5 ReferencesBall, J.W., E.A. Jenne, and D.K Nordstrom. 1979. WATEQ2
— A Computerized Chemical Model for Trace and MajorElement Speciation and Mineral Equilibria of NaturalWaters. In: Chemical Modeling in Aqueous Systems:Speciation, Sorption, Volubility, and Kinetics, E.A. Jenne(ed.), ACS Symp. Series 93, American Chemical Society,Washington, DC, pp. 815-836.
Ball, J. W., D.K. Nordstrom, and E.A. Jenne. 1980. Additionaland Revised Thermochemical Data and Computer Codefor WATEQ2: A Computerized Chemical Model forTrace and Major Element Speciation and Mineral Equi-libria of Natural Waters. U.S. Geological Survey WaterResources Investigations No. 78-116.
Benjamin, M.M. and J.O. Leckie. 1981. Multiple-Site Ad-sorption of Cd, CU. Zn, and Pb on Amorphous IronOxyhydroxides. J. Coll. Interface Sci. 79(2):209-221.
Black, J.A. and J.H. Heil. 1982. Municipal Solid WasteLeachate and Scavengerwaste: Problems and Prospects inBrookhaven Town. In: Proceedings of the Northeast Con-ference on the Impact of Waste Storage and Disposal onGround-Water Resources, R.P. Voitski and G. Levine(eds.), U.S. Geological Survey and Cornell University,2.1:1-12.
Booz-Allen and Hamilton, Inc. 1987. ROD (Record of Deci-sion) Annual Report FY 1986. U.S. Environmental Pro-tection Agency (NTIS PB87-199550), 182 pp.
Brookins, D.G. 1988. Eh-pH Diagrams for Geochemistry.Springer-Verlag, New York, 176 pp.
Buddemeir, R.W. and J.R. Hunt. 1988. Transport of ColloidalContaminants in Groundwater Radionuclide Migrationat the Nevada Test Site. Applied Geochemistry 3:535-548.
Cook, D.K. and R.G. DiNitto. 1982. Evaluation of Groundwa-ter Quality in East and North Wobum, Massachusetts. In:Proceedings of the Northeast Conference on the Impactof Waste Storage and Disposal on Ground-Water Re-sources, R.P. Voitski and G. Levine (eds.), U.S. Geologi-cal Survey and Cornell University, 4.2:1-20.
Cozzarelli, I.M., M.J. Baedecker, and J.A. Hopple. 1987.Effects of Creosote Products on the Aqueous Geochemis-try of Unstable Constituents in a Surficial Aquifer. In:U.S. Geological Survey Program on Toxic Waste-Ground-Water Contamination: Proceeding of the ThirdTechnical Meeting, Pensacola, Florida, March 23-27,1987, B.J. Franks (ed.), U.S. Geological Survey Open-File Report 87-109, pp. A15-A16.
Drever, J.I. 1989. The Geochemistry of Natural Waters, 2nded. Prentice-Hall, Englewood Cliffs, NJ. [First edition1982].
Dzombak, D.M. 1986. Towards a Uniform Model for Sorp-tion of Inorganic Ions Hydrous Oxides. PhD Dissertation,Department of Civil Engineering, Massachusetts Instituteof Technology.
Dzombak, D.A. and F.M.M. Morel. 1986. Sorption of Cad-mium on Hydrous Ferric Oxide at High Sorbate/SorbentRatios: Equilibrium, Kinetics, and Modelling. J. ColloidInterface Sci. 112 2):588-598.
Dzombak, D.A. and F.M.M. Morel. 1990. Surface Complex-ation Modelling, Hydrous Ferric Oxide. John Wiley &Sons, New York, 393 pp.
Dzombak, D.A., W. Fish, and F.M.M. Morel. 1986. Metal-Humate Interaction. 1. Discrete Ligand and ContinuousDistribution Models. Environ. Sci. Technol. 20:669-675.
Evans, D. W., J.J. Alberts, and R.A. Clark. 1983. ReversibleIon-Exchange of Cesium-137 Leading to Mobilizationfrom Reservoir Sediments. Geochimica et CosmochimicaActa 47(ll):1041 -1049.
Felmy, A.R., D.C. Girvin, E.A. Jenne. 1984. MINTEQ AComputer Program for Calculating Aqueous Gemchemi-cal Equilibria. EPA/600/3-84-032 (NTIS PB84-157148).
Fish, W. 1990. Subsurface Redox Chemistry: A comparisonof Equilibrium and Reaction-Based Approaches. In: MetalSpeciation in Groundwater, H. Allen E.M. Perdue, and D.Brown (eds.), Lewis Publishers, Chelsea, MI.
Fish, W. (in press). Subsurface Transport of Gasoline-DerivedLead. Ground Water.
Fish, W. and M.S. Elovitz. 1990. Redox and Solvation Inter-actions between Hexavalent Chromium and Hydroxy-lated Organic Compounds. U.S. EPA Contract Report90-R-8 14136-01-0.
Garrels, R.M. and C.L. Christ. 1965. Solutions, Minerals, andEquilibria. Harper and Row, New York, 450 pp.
Gschwend, P.M. and M.D. Reynolds. 1987. MonodisperseFerrous Phosphate Colloids in an Anoxic GroundwaterPlume. J. Contaminant Hydrology 1:309-327.
Hanzlik, R.P. 1976. Inorganic Aspects of Biological andOrganic Chemistry. Academic Press, New York.
Helfferich, F. 1962. Ion Exchange. McGraw-Hill, New York.
Honeyman, B. D., K.F. Hayes, and J.O. Leckie. 1982. Aque-ous Chemistry of As, B, Cr, Se, and V with ParticulmReference to Fly-ash Transport Water. EPRI-91O-1. Elec-tric Power Research Institute, Palo Alto, California.
Jennings, A. A., D.J. Kirkner, and T.L. Theis. 1982. Multi-component Equilibrium Chemistry in Ground Water Qual-ity Models. Water Resources Reach 18: 1089 -l096.
190
Kharaka, Y.K. and I. Barnes. 1973. SOLMINEQ: Solution-Mineral Equilibrium Computations. U.S. Geological Sur-vey, Menlo Park, CA (NTIS PB215-899).
Kharaka, Y.K., W.D. Gunter, P.K. Aggarwal, E.H. Perkins,and J.D. DeBraal. 1988. SOLMINEQ.88: A ComputerProgram for Geochemical Modeling of Water-Rock In-teractions. U.S. Geological Survey Water-Resources In-vestigations Report 88-4227, 420 pp.
Laha, S. and R.G. Luthy. 1990. Oxidation of Aniline andOther Primary Aromatic Amines by Manganese Dioxide.Environ. Sci. Technol. 24:363-373.
Lindberg, R.D., and D.D. Runnels. 1984. Ground Water Re-dox Reactions: An Analysis of Equilibrium State Appliedto Eh Measurements and Geochemical Modeling. Sci-ence 225:925-927.
McDowell-Boyer, J.R. Hunt, and N. Sitar. 1986. ParticleTransport Through Porous Media. Water Resources Re-search 22:1901-1921.
Moore, J.W. and S. Ramamoorthy. 1984. Heavy Metals inNatural Waters. Springer-Verlag, New York, 268 pp.
Morel, F.M.M. 1983. Principles of Aquatic Chemistry. WileyInterscience, New York.
Murarka, I.P. and D.A. McIntosh. 1987. Solid-Waste Envi-ronmental Studies (SWES): Description, Status, and Avail-able Results. EPRI EA-5322-SR. Electric Power ResearchInstitute, Palo Alto, CA.
Nordstrom, D.K. and J.L. Munoz. 1985. Geochemical Ther-modynamics. Benjamin/Cummings Publishing, MenloPark, CA, 477 pp.
Palmer, C. D., W. Fish, and J.F. Keely. 1988. Inorganic Con-taminants: Recognizing the Problem. In: Proc. SecondNat. Outdcmr Action Conf. on Aquifer Restoration, GroundWater Monitoring, and Geophysical Methods, NationalWater Well Association, Dublin, OH, pp. 555-579.
Parkhurst, D.L., D.C. Thorstenson, and L.N. Plummer. 1980.PHREEQ — A Computer Program for Geochemical Cal-culations. U.S. Geological Survey Water-Resources In-vestigations Report 76-13, 81 pp.
Parkhurst, D.L., L.N. Plummer, and D.C. Thorstenson. 1982.BALANCE-A Computer Program for Calculating MassTransfer for Geochemical Reactions in Ground Water.U.S. Geological Survey Water-Resources InvestigationsReport 82-14 (NTIS PB82-255902), 29 pp.
Plummer, D.L. and W. Back. 1980. The Mass Balance Ap-proach: Application to Interpreting the Chemical Evolu-tion of Hydrologic Systems. Am. J. Science 280:130-142.
Plummer, L.N., B.F. Jones, and A.H. Truesdell. 1976.WATEQF — A FORTRAN IV Version of WATEQ, aComputer program for Calculating Chemical Equilib-
19
rium of Natural Waters. U.S. Geological Survey Water-Resources Investigations Report 75-61, 73pp.
Rees, T.F. 1987. A Review of Light-Scattering Techniquesfor the Study of Colloids in Natural Waters. J. Contami-nant Hydrology 1:431-439.
Reichenbcrg, D. 1966. Ion Exchange Selectivity. In: IonExchange and Solvent Extraction, Vol. 1, J.A Marinsky(ed.), Marcel Dekker, New York, pp. 227-276.
Schindler, P.W. 1981. Surface Complexes at Oxide-WaterInterfaces. In: Adsorption of Inorganic at Solid-LiquidInterfaces, M.A. Anderson and A.J. Rubin (eds.), AnnArbor Science, Ann Arbor, MI, pp. 1-49.
Schindler, P.W. and W. Stumm. 1987. The Surface Chemistryof Oxides, Hydroxides, and Oxide Minerals. In: AquaticSurface Chemistry, W. Stumm (ed.), John Wiley & Sons,New York, pp. 83-110.
Sholkovitz, E.R. 1985. Redox-related Geochemistry in Lakes:Alkali Metals, Alkaline Earth Metals, and Cesium-137,In: Chemical Processes in Lakes, W. Stumm (ed.), Wiley-Interscience, New York.
Siegel, D.I. 1987. Geochemical Facies and Mineral Dissolu-tion, Bemidji, Minnesota, Research Site. In: U.S. Geo-logical Survey Program on Toxic Waste-Ground-WaterContamination: Proceeding of the Third Technical Meet-ing, Pensacola, Florida, March 23-27, 1987, B.J. Franks(ed.), U.S. Geological Survey Open-File Report 87-109,pp. C13-C15.
Smith, R.L. and A.E. Marten. 1976. Critical Stability Con-stants. Plenum, New York.
Sposito, G. 1984. The Surface Chemistry of Soils. OxfordUniversity Press, New York.
Sposito, G. 1986. Sorption of Trace Metals by Humic Materi-als in Soils and Natural Waters. CRC Critical Reviews inEnvironmental Control 16:193-229.
Stone, A.T. 1986. Adsorption of Organic Reductants andSubsequent Electron Transfer on Metal Oxide Surfaces.In: Geochemical Processes at Mineral Surfaces, J.A. Davisand K.F. Hayes (eds.s), ACS Symp. Series 323, Ameri-can Chemical Society, Washington, DC, pp. 446-461.
Stumm, W. and J.J. Morgan. 1981. Aquatic Chemistry, 2nded. Wiley Interscience, New York.
Stumm, W., H. Hohl, F. Dalang. 1976. Interaction of MetalIons with Hydrous Oxide Surfaces. Croat. Chem. Acts.48(4):491-504.
Truesdell, A.H. and B.F. Jones. 1974. WATEQ, A ComputerProgram for Calculating Chemical Equilibria of NaturalWaters. J. Research U.S. Geological Survey 2:233-248.
Vanselow, A.P. 1932. Equilibria of the Base Exchange Reac-tions of Bentonites, Permutites, Soil Colloids and Zeo-lites. Soil Science 33:95-113.
Voudrias, E.A. and M. Reinhard. 1986. Abiotic Organic Re-actions at Mineral Surfaces. In: Geochemical Processesat Mineral Surfaces, J.A. Davis and K.F. Hayes (eds.),ACS Symp. Series 323, American Chemical Society,Washington, DC, pp. 462-486.
Westall, J.C. and H. Hohl. 1980. A Comparison of Electro-static Models for the Oxide/Solution Interface. AdvancesColl. Interface Sci. 12(2):265-294.
Wolery, T.J. 1979. Calculation of Chemical Equilibrium Be-tween Aqueous Solutions and Minerals: The EQ3/6 Soft-ware Package. Report UCRL 52658. Lawrence LivermoreNational Laboratory, Livermore, CA.
Wolery, P.J. 1983. EQ3NR, a Computer Program for Geo-chemical Aqueous Speciation-Volubility Calculations.Report UCRL 53414. Lawrence Livermore National Labo-ratory, Livermore CA.
Yeh, G.T. and V.S. Tripathi. 1989. A Critical Evaluation ofRecent Developments in Hydrogeochemical TransportModels of Reactive Multichemical Components. WaterResources Research 25:93-108.
192
Chapter 13Characterization of Subsurface Degradation Processes
J. Michael Henson
When chemical constituents enter the subsurface envi-ronment, they are subjected to physical, chemical, and bio-logical processes that ultimately determine their fate andtransport characteristics. Knowledge of the degradation pro-cesses that determine the fate of organic compounds in thesubsurface can be used to guide remediation efforts at sitesthat have been affected. The physical processes that controlthe transport of constituents in the subsurface are discussed inprevious chapters, This chapter describes biological andnonbiological processes that may control the fate of organicchemicals once they have entered the subsurface. An under-standing of these principles will aid in the efficient and cost-effective remediation of releases of organic constituents.Objectives of this chapter are to:
Present information about abiotic degradation pro-cesses.
Present information about biological degradation pro-cesses.
Provide a basis for site evaluation to determine thepotential for biological remediation.
Build the foundation for the discussion of bioreme-diation of soils (Section 15.2.2) and ground water(Section 16.3).
This chapter discusses two classes of transformations thatmay occur in the subsurface-abiotic and biologic transfor-mation. Abiotic reactions are those reactions that do notinvolve metabolically active organisms, a product of a livingcell, or a product of a previously living organism. Someexamples of products of cells are extracellular enzymes, he-moprotein, iron porphyrins, cytochromes, flavins, and re-duced pyridine nucleotides.
13.1 Abiotic Transformation ReactionsHydrolysis, substitution, elimination, and oxidation-re-
duction are the abiotic reactions that will be discussed in thischapter. These reactions produce a variety of end-productswhose presence may play a role in decisions made to selectcompounds for the remedial investigation phase. The resultsof an abiotic reaction may enhance the biological degradabilityof a compound and provide possible treatment of the parent
compound. Dragun (1988) provides an excellent presentationof abiotic reactions.
13.1.1 HydrolysisHydrolysis reactions are those reactions where an organic
chemical reacts with either water or a hydroxide ion to pro-duce an alcohol. The following equations represent thesereactions:
R-X + H2O —> R-OH + H+ + X
R-X + OH —> R-OH + X
In these reactions, either H2O or OH act as a nucleophile andattack the electrophile, RX, to displace the leaving group, X.This type of reaction is referred to as a nucleophilic displace-ment reaction and in this example results in the formation of adaughter product that is an alcohol. For a more detaileddiscussion of this nucleophilic displacement reaction mecha-nism, see Dragun (1988). The rate of hydrolysis reactions istypically first order with respect to the concentration of thecompound. The rate of a first-order reaction increases as theconcentration of the organic compound increases. The first-order rate constant k can be calculated as:
k = (2.303/t) log[C0/(CO - Ct)]
where t is time, C0 is initial concentration, and Ct is concentra-tion at t.
The time required for half of the concentration of the com-pound to degrade is known as the half-life, t½, and is calcu-lated as:
t ½= 0.693/k
Some examples of hydrolysis half-lives for some organiccompounds are presented in Table 13-1. A more extensivelisting of hydrolysis half-lives can be found in Dragun (1988).
The rates of hydrolysis vary from compound to com-pound and can be on the order of hours to years. The rates ofhydrolysis also indicate the susceptibility of the compounds tohydrolysis. Some examples of organic chemicals that aresubject to hydrolysis are alkyl halides, carbamates, chlori-nated amides, esters, and epoxides. Examples of chemicals
193
.
Table 13-1. Selected Hydrolysis Half-Lives for a Variety ofOrganic Compounds
that are more resistant to hydrolysis are aldehydes, alkanes,alkenes, and compounds with carboxy - or nitro-substituents.
Once an organic compound enters the subsurface, envi-ronmental factors can decrease or increase the hydrolysis half-life that might be projected from the results of a laboratoryevaluation. One effect that soil can have on the hydrolysishalf-life is on localized pH differences. The pH at the surfaceof the soil particles can be very different from the overall soilpH. These localized effects may alter the half-life by enhanc-ing or inhibiting the hydrolysis reaction. Another effect of soilon half-life results from metal ions that are present as normalcomponents of the soil. These metals can serve as catalysts fororganic reactions. A third environmental factor is adsorptionof the organic compound to the soil particles, which can affectthe rate of hydrolysis reactions. By adsorbing to the soilparticle, the compound is in effect removed from the water.Other factors such as soil water content and the type of soilmatrix also can affect the rate of hydrolysis.
13.1.2 SubstitutionHydrolysis reactions are classified as a type of substitu-
tion reaction but they are presented first because of thepredominance of water, which causes the reactions to occur.Other chemicals in the subsurface can cause substitutionreactions to occur. An example of a substitution reactioninvolves hydrogen sulfide acting as the nucleophilic agent toattack organic compounds, which result in the production ofsulfur-containing compounds.
13.1.3 EliminationElimination reactions cause the loss of two adjacent
groups from within the molecule resulting in the formation ofa double bond. The reaction occurs as:
One example of an elimination reaction is the formationof 1,1 -dichloroethene (1,1-DCE) from 1,1,1-trichloroethane(1,1,1-TCA). An additional formation product of an abioticreaction was the detection of acetic acid formed as a result ofsubstitution. The ratio of acetic acid to 1, 1-DCE was about 3:1(Cline et al., 1988). Elimination also can result in the forma-tion of bromoethene from 1,2-dibromoethane andbromopropene from l,2-dibromopropane (Dragun, 1988).
13.1.4 Oxidation-ReductionOxidation is the net removal of electrons from an organic
compound, while reduction is the net gain of electrons by anorganic compound. These reactions are coupled by the trans-fer of electrons from one compound to another. The oxida-tion-reduction couples in soil systems are complex and multiple.In many instances, if a biological response to an organiccompound occurs, the biological system will tend to becomepredominant. Inorganic redox reactions are discussed furtherin Section 12.1.3.
Abiotic reactions may occur in the subsurface by a vari-ety of mechanisms and at varying rates. The use of abioticreactions as a remediation technology has not received a lot ofattention, but may provide an alternative treatment in someinstances. Abiotic reactions may occur in conjunction withbiological reactions and make some compounds more suscep-tible to biodegradation. Abiotic reactions may not alwaysprovide extensive treatment of the organic compound but thetreatment that does occur may produce a compound of lessenvironmental concern.
13.2 Microbiological Transformations in theSubsurface
Microbiological transformations are the second class ofprocesses that have an impact on the fate of organic com-pounds once they enter the subsurface. This class of processescan result in either partial or complete degradation of theorganic compounds to detoxify or remove them from thesubsurface. The knowledge of biological responses to variousorganic compounds can be utilized during the site investiga-tion process to collect data that will aid in evaluating potentialremediation alternatives. These remediation alternatives caninclude biological remediation.
When addressing biological transformations, biodegrada-tion is typically used to mean complete degradation. How-ever, biodegradation specifically refers to the biologicaltransformation of an organic compound without regard to theextent of transformation. Mineralization specifically refers tothe conversion of an organic compound to carbon dioxide (ormethane in anaerobic environments), water, and a halogenatom, if the parent compound was halogenated.
Knowledge of biological responses to organic compoundsthat may occur under different microbial growth conditionsprovides an understanding of the metabolic potential by whichmicroorganisms may transform these compounds. For ex-ample, if partial degradation of an organic compound occurs,the daughter products formed may or may not be of environ-mental concern. The observation of microbial intermediates
194
of metabolism indicates that a biological response to theparent compounds has occurred and that the potential for siteremediation through biological processes exists.
The use of microorganisms for remediation of sites af-fected with organic compounds is gaining increasing interest.This process of bioremediation requires an integrated ap-proach involving the disciplines of microbiology, hydrogeol-ogy, and engineering. The relationship of these three disciplinesis analogous to a “three-legged stool,” in that if one of the legsis weak, the stool cannot support much weight. These threedisciplines also must be augmented with an awareness of theprinciples of the physical-chemical interactions that are thesubject of previous chapters, and an understanding of theregulatory requirements in which the application of bioreme-diation will take place. This section will provide a basicunderstanding of the principles of microbial ecology as re-lated to the subsurface. This understanding can aid in theevaluating of sites affected with organic compounds andprovide a basis for the following chapters where examples ofbioremediation will be presented.
13.2.1 Microbial Ecology of the SubsurfaceAlthough it is now known that significant numbers of
microorganisms are distributed throughout the subsurface(Back, 1989; Ghiorse and Wilson, 1988), it was once sug-gested that numbers of microorganisms in soil decreased withdepth (Waksman, 1916). More recent investigations, how-ever, routinely detect microorganisms in aquifers. These in-vestigations include aquifers not known to have been affectedwith organic compounds and aquifers that have receivedinputs of organic compounds.
The development of techniques to investigate water tableaquifers was instrumental to the elucidation of the microbialecology of the subsurface. McNabb and Mallard (1984) de-scribed sampling techniques designed to prevent the micro-bial contamination of samples retrieved from the terrestrialsubsurface. These techniques rely on the collection of coresfrom the depth to be investigated. After collection, the outer-most layer can be removed in the field with alcohol-sterilizeddevices designed to strip away the soils that were in contactwith drilling equipment. These techniques produce a subcoreof the original core in an aseptic manner. Subcores can beobtained in the laboratory by a variety of mechanisms, as long
as aseptic techniques are used. For field or laboratory condi-tions, the subcore can be collected under anaerobic (Beemanand Suflita, 1987) as well as aerobic conditions.
The collection of subsurface samples using aseptic tech-niques to prevent intrusion of microorganisms not representa-tive of the subsurface has yielded considerable informationabout the microbial ecology of the subsurface. For example,Wilson et al. (1983) and Balkwill and Ghiorse (1985) reportedthe presence of between 1 and 10 million microorganisms pergram of sediment using the Acridine Orange Direct Count(AODC) staining technique to count the microorganisms. Thesame authors, using a plate count assay to count viable micro-organisms, detected between 200,000 and 2.5 million micro-organisms per gram of sediment in two aquifers that were notknown to have received input of organic compounds. Similarranges of counts for microorganisms for shallow aquifers notreceiving organic chemicals arc shown in Table 13-2.
Beeman and Suflita (1987) reported a range of 11 to 17x106 cells (g dry wgt) measured by AODC in a sand aquiferreceiving landfill leachate in Norman, Oklahoma. Similarranges of microbial counts by AODC were observed by Erlichet al. (1983) and Webster et al. (1985) for two differentaquifers that were affected with creosote compounds.
The results of these investigations indicate that the terres-trial subsurface whether pristine or not is populated by micro-organisms. These numbers of microorganisms are relativelyhigh and were detected in a variety of geologic environmentsand depths. Analysis of subsurface samples indicates that themicroorganisms are predominantly attached to the subsurfacesoil particles (Harvey et al., 1984). Evidence also is accumu-lating that even deeper geologic environments are inhabitedby microorganisms (Updegraff, 1982).
Biochemical diversity of microorganisms present in thesubsurface is evidenced by the variety of organic compoundsreported to be metabolized. Petroleum hydrocarbons, includ-ing fuels, creosote constituents, and products of coal gasifica-tion, are reported to be substrates for subsurfacemicroorganisms under a variety of growth conditions. Table13-3 presents examples of organic compounds metabolizedby subsurface microorganisms.
Table 13-2. Microbial Cell Counts for Selected Aquifers That Were Not Receiving Known Inputs of Organic Compounds
195
13.2.2 Relationship of Environmental Factors toBiodegradation
Microorganisms require a suitable set of environmentalfactors in order to grow. These factors include the chemicaland physical parameters of pH, available water or osmoticpressure, temperature, and absence of toxic conditions.
The pH of the environment is an easily measured param-eter and indicates the potential for microbial activity. Manymicrooganisms grow best in the pH range of 6 to 8. Microbiallife at extremes of pH does occur and, therefore, a pH outsideof the 6 to 8 range does not exclude microbial growth. Growthof microorganisms can raise or lower the pH by producingend-products that affect pH or by removing the parent com-pounds, thus, affecting the pH. The measurement of pH inground water could indicate the potential for microorganismsto grow in the aquifer.
Temperature generally affects microbial growth in that anincrease in temperature results in an increase in microbiologi-cal growth. Microorganisms have lower, upper, and optimumtemperature limits for growth. Many microorganisms in thesoil have an optimum temperature for growth between 10°and 30°C. Temperatures of ground waters within the UnitedStates are within this range (Dragun, 1988).
Microorganisms require water for active growth. Theavailability of water depends on the number of moleculespresent in the solution. An increase in the number of mol-ecules, relative to the number of molecules within the micro-bial cells, results in the movement of water from the cell intothe surrounding environment as a result of osmosis. Theopposite effect occurs when the number of molecules outsidethe microbial cell is less than the number inside the microbialcell. The soil moisture content is sometimes critical to thegrowth of microorganisms. If the moisture content is nearsaturation, transfer of oxygen may become a growth-limitingfactor. If the soil is dry, growth of the microorganisms will bevery limited.
13.2.3 Microbial MetabolismThe ability of communities of microorganisms to me-
tabolize many types of organic compounds including syn-thetic organic compounds is well documented (Alexander,1981; Gibson, 1984). A number of these organic compounds
are utilized by microorganisms as a source of carbon andenergy. The degradation of the compounds may not occur atthe initial time of release to the environment. A period of timemay elapse before an increase in the rate of degradation isobserved. This period of time is referred to as an adaptation oracclimation period. The adaptation period may vary with thecompound and the environmental conditions into which thecompound is released. For example, under anaerobic condi-tions, adaptation periods may be as long as several months.Once the adaptation occurs, however, the rate of degradationbecomes a function of the processes controlling the availabil-ity of nutrients to the microorganisms and not of an intrinsicmetabolic property of the microorganisms. In addition, oncethe microbial community adapts to a particular organic com-pound or compounds, the compound or compounds can con-tinue to be added without re-adaptation. The microbialcommunity, thus, becomes enriched in members that canmetabolize the organic compounds.
An additional opportunity for microbial degradation isthrough a process of nongrowth metabolism. In this process,the microorganisms do not use the organic compound as asource of carbon and energy, which results in growth. Instead,the microorganisms cometabolize a substance that cannot beutilized for growth in the presence of a compound that can beutilized for growth. The cometabolized compound is oftentransformed into an intermediate that can undergo transforma-tion by other microorganisms. A specific example, to bediscussed in more detail later, is the degradation of trichloro-ethene and dichloroethene by microorganisms that are grow-ing on methane and fortuitously react with the halogenatedcompounds.
The ability of microorganisms to degrade organic com-pounds depends on the presence of a terminal electron accep-tor (TEA), as well as other nutrients. The TEA receiveselectrons from a series of oxidation-reduction reactions withinthe cell that generate energy allowing the microorganism togrow. Some microorganisms can use several TEAs whereasother microorganisms can use only one. If more than one TEAis present when an organic compound enters the environment,the one that results in the highest energy transfer will be usedfirst. Next, the TEA with the second highest energy transferwill bc used, and so on until either the organic compound isremoved or the TEAs have been consumed.
Table 13-3. Representative Examples of the Diversity of Organic Compounds Metabolized by Subsurface Microorganisms
195
Table 13-4 presents the relative energy charge associatedwith the consumption of various TEAs. The succession ofmetabolic events will proceed from the reactions with TEAsthat can transfer the most energy, which are denoted by themost negative values in Table 13-4. The succession also isrelated to the toxicity of TEAs for groups of bacteria. Forexample, methanogenic bacteria are inhibited by oxygen;therefore, the development of active methanogenesis usuallydoes not occur until oxygen is removed from the environmentand reducing conditions are established.
Aerobic respiration is the process of consuming organiccompounds with oxygen serving as the TEA. The end-productof the respiration of oxygen is water. The degradation ofhydrocarbons also requires oxygen as a cosubstrate where theoxygen is inserted into the hydrocarbon molecule.
Once the metabolic demand for oxygen exceeds the rateof supply, the anoxic conditions that are established allowother TEAs to be utilized. The next TEA that, if present,would be used is nitrate. The respiration of nitrate is referredto as denitrification and results in the production of nitrogengas (Knowles, 1982). The transfer of energy is similar to thatof the respiration of oxygen. Many of the organisms that usenitrate as an electron acceptor also use oxygen so that accli-mation of a new population of microorganisms may not berequired.
Once the nitrate has been consumed and the oxidation-reduction state becomes reducing, the respiration of sulfatecan begin as a process known as sulfate reduction (Postgate,1979). Sulfate reduction results in the production of hydrogensulfide, which can be corrosive to equipment and potentiallytoxic to humans. Sulfate reduction does not yield as muchenergy, only about one-fourth, as does the respiration ofnitrate or oxygen.
If nitrate is present in a reducing environment, its respira-tion does not result in the production of nitrogen gas; instead,ammonia is produced (Caskey and Tiedje, 1980). The respira-tion of nitrate under reducing conditions does not transfer asmuch energy as the respiration of oxygen.
As the conditions become more reducing and alternativeTEAs are consumed, the respiration of carbonate will result inthe production of methane. The microorganisms that carry out
Table 13-4. Comparison of Free Energy Values for Metabolismof Glucose in the Presence of Various ElectronAcceptors
this reaction are known as methanogenic bacteria. The energytransferred during methane production is about one-fourththat of the respiration of oxygen or nitrate.
The ability of microorganisms to carry out a variety ofrespirations provides the opportunity to collect data during thesite investigation phase that indicate whether a microbiologi-cal response to organic compounds has occurred. If accuratemeasurements of dissolved oxygen (DO) in ground waterindicate that oxygen is present outside a plume of organiccompounds and DO is not detected within the plume, then abiological response may have occurred to consume the oxy-gen. If methane or another of the respiratory end-products isdetected within the plume, the results suggest that a biologicalresponse has occurred and that reducing conditions may exist.
In addition, the range of metabolic capabilities of micro-organisms extends beyond the respiration of oxygen. Nitrateis more water soluble than oxygen and may be less costly touse in the treatment of some affected aquifers. Reducingconditions allow some biological transformations to occurthat do not occur under oxidizing conditions. One example ofthis type of biological transformation is reductive dechlorina-tion. This microbiological process removes chlorines fromchlorinated compounds (discussed in Section 13.3.2).
13.2.4 Biological Reaction KineticsThe rate at which microorganisms can remove organic
compounds from the subsurface can be expressed mathemati-cally to approximate the time required for remediation. Thefirst-order rate constant is based on the observation that as theconcentration of the organic compound increases, the rate ofdegradation increases. The first-order rate constant k is calcu-lated as:
k = (2.303/t) log[C0/(CO - Ct)]
where t is time, C0 is initial concentration, and Ct is concentra-tion at t. The time required for half of the concentration of thecompound to degrade is known as the half-life, t½, and iscalculated as follows:
t½ = 0.693/k
However, metabolism in microorganisms occurs via enzymesthat become saturated; the substrates are degraded when theconcentration of the substrate continues to increase. Once theenzymes become saturated, the rate of degradation cannotincrease and the degradation rate curve becomes hyperbolic.
The use of the first-order rate kinetics provides a generalexpression of the rate of biodegradation for many compounds.Dragun (1988) provides a compilation of first-order degrada-tion rates that should not be used without comparing theenvironments from which these samples were taken. A directextrapolation of results obtained from one environment toanother environment is typically not useful.
197
13.3 Bioremediation of Organic Compounds inthe Subsurface
13.3.1 General ConsiderationsThe basic premises of microbial ecology are related to
bioremediation in that many organic compounds can be usedby microorganisms as a source of carbon and energy. Many ofthe compounds that are considered hazardous can be degradedin the subsurface if the concentrations are not toxic to themicroorganisms and the appropriate environmental param-eters can be established. Bioremediation is based on theunderstanding of the carbon cycle and extrapolation of com-pound mineralization in other environments to the subsurface.Environmental factors, such as pH, oxidation-reduction po-tential, and temperature, may play a role in determining thepotential for bioremediation. However, the rate at whichnutrients, especially a TEA, can be delivered to the microor-ganisms may determine whether bioremediation is feasible.There are several reviews that provide detailed discussions ofbioremediation (Lee et al., 1988; Thomas and Ward, 1989;Wilson et al., 1986; and Wilson and Ward, 1987).
Certain information is required before design of the bio-remediation system can begin. An assessment of the site toevaluate history, geology, and hydrology can provide infor-mation valuable for bioremediation design. The delivery ofnutrients to subsurface microorganisms for in situ remediationis dependent on the site hydrology. Sites with low permeabil-ity, such as those with clays, may not allow the delivery ofnutrients in an efficient manner.
A thorough laboratory assessment of the microbiologyalso provides information to indicate whether bioremediationis an appropriate treatment technology. Some components ofthis assessment are:
. Evaluate the presence of requisite microorganisms.
. Assess potential toxicity to the microorganisms.
● Evaluate nutrient requirements to enhance degrada-tion activity.
. Evaluate the compatibility of the site geochemistrywith the nutrient solution proposed for addition.
Requisite microorganisms are the ones that are capable ofdegrading the organic compounds present at the site. Formany sites, these microorganisms are naturally occurring andjust need some nutrients to stimulate their growth. The pres-ence of these microorganisms at the site is evaluated insamples representative of the environment to be remediatcd.If the remediation is an in situ aquifer remediation, then thesamples should be collected from the aquifer. The microor-ganisms are predominantly attached to the soils; therefore,samples of the soils below the water table should be collected.Several methods exist for collecting the samples. Principlesfor collection are discussed in McNabb and Mallard (1984).
Microorganisms present in the samples should be enu-merated in a manner to indicate the presence of viable micro-
organisms. Staining techniques exist, such as the AODC, butthis technique is limited because it does not indicate theviability of the microorganisms. The results of viable countssuggest the environment that was sampled was not so toxic asto completely inhibit the presence of microorganisms. Tech-niques such as standard plate counts can be used to detect thenumber of general microorganisms present. Plate counts usinga microbial medium containing the compound of interest alsocan be used to enumerate the bacteria present in the samplecapable of growth on that compound. The numbers can becompared before and after treatment to assess whether thetreatment resulted in an increase in the number of microbescapable of growth on the compound of interest. An increase inthe observed number of bacteria would suggest an effectiveprocess.
The nutrients required to enhance microbial growth areassessed primarily on the nitrogen and phosphorous require-ments of the microorganisms. However, the microorganismsmay require other nutrients such as potassium, magnesium,manganese, and iron. The site’s geochemistry may providemany of these necessary nutrients. The nutrient solution se-lected should be compatible with the geochemistry of the siteto prevent possible precipitation of minerals, which mightdecrease the permeability of the aquifer. In addition, an evalu-ation of the compatibility of the TEA chosen with the site’sgeochemistry should indicate whether undesired reactions canoccur.
The laboratory assessment for the removal of the parentcompound can measure the disappearance of the compound,the rate of removal, and the production of daughter products.The rate of rcmoved usually reflects the laboratory conditions,however, and cannot be extrapolated directly to the rate ofremoval that would be expected in the field. Disappearance ofthe parent compound may, by itself, not always indicate thatmineralization has occurred.
13.3.2 Compounds Appropriate to Consider forBioremediation
During the initial evaluations for bioremediation of a site,existing information should be considered. Information aboutthe volubility of the compound to be degraded indicates thepotential availability of the compound to the microorganisms.Previous evaluations of the biodegradation of the compoundoften can be found in the scientific literature. These studiescan provide information about the inherent degradability ofthe compound as well as the potential products of degradation.Information about the environmental factors that upon stimu-lation were critical to degradation also may be available.Dragun (1988), for example, contains a list of organic com-pounds and provides information about the conditions used inthe evaluations to develop the rates of biodegradation pre-sented.
In general, hydrocarbons are good candidates for biore-mediation. The review paper by Atlas (1981) and the booksedited by Gibson (1984) and Atlas (1984) provide an over-view of the microbiological degradation of petroleum hydro-carbons. Many components of fuel hydrocarbons, such asbenzene, toluene, and xylenes are degraded by a variety of
198
microorganisms. Creosote, which is a by-product of the pro-duction of coke from coal, is composed of a number of highmolecular weight hydrocarbons referred to as polycyclic aro-matic hydrocarbons (PAHs). This complex mixture of hydro-carbons has components that are biodegradable with thedegradation rate decreasing as the molecular weight of thehydrocarbons increases. Generally, the PAHs with three ringsor less degrade at a greater rate than do the more complexPAHS.
A variety of organic compounds can biodegrade in thesubsurface if the environmental conditions are appropriate.For example, alcohols, glycols, ketones, phenols, chlorinatedphenols, and other organic compounds have the potential tobiodegrade. Some factors that may enhance biodegradationare the water volubility and molecular weight of the com-pounds. Increasing volubility enhances the potential for bio-degradation assuming the concentration does not reach levelstoxic to the microorganisms. Increasing molecular weight orbranching of organic compounds may tend to slow the rate ofdegradation.
Halogenated compounds generally tend to persist in aero-bic environments, but continued research is providing evi-dence that biological alternatives to these compounds mayexist. Under anaerobic conditions, several chlorinated com-pounds have been shown to undergo transformation. Forexample, tetrachloroethene (PCE) has been shown to be de-chlorinated under environmental conditions that support thegrowth of anaerobic bacteria. This process is known as reduc-tive dechlorination and is given as follows:
PCE —> TCE + Cl —> DCE + Cl —> CE + Cl—> C02 + Cl
The compounds produced are trichloroethene (TCE), the iso-mers of dichloroethene (DCE), and chloroethene (CE). Theremoval of the chlorine atoms enhances the potential foraerobic microorganisms to degrade the daughter products.DCE has a greater potential for aerobic degradation than doesPCE.
A method to enhance the aerobic degradation of DCE isto create an environment for the growth of methane-utilizingbacteria. The addition of methane to soils and aquifers typi-cally results in the growth of these bacteria within severaldays to a few weeks. These bacteria have been shown todegrade a variety of halogenated compounds including TCE,cis-DCE, trans-DCE, chloroform, dichloromethane, and 1,2-dichloroethane (Henson et al., 1989). It seems plausible thatthe series of reaction processes that enhances anaerobic reduc-tive dechlorination of highly chlorinated compounds and yieldsthe less chlorinated compounds that can undergo aerobicdegradation may be a good mechanism to remove compoundsfrom the subsurface environment. The value for this treatmentprocess is further enhanced when the increased sorptive ca-pacity of the higher chlorinated compounds is considered.Utilizing the microorganisms in an in situ treatment processcan significantly expedite remediation.
Bioremediation of other halogenated compounds such aspolychlorinated biphenyls (PCBs) also can be considered. The
reductive dechlorination of PCBS was detected in the environ-ment (Brown et al., 1987) and confirmed in the laboratory(Quensen et al., 1988). The anaerobic reductive dechlorina-tion process removes chlorines from the PCBS, thus reducingpotential toxicity and enhancing the aerobic degradability ofthe compounds. Anaerobic biological treatment followed byaerobic biological treatment is a technology that could removethese chlorinated compounds from the environment in a cost-effective and environmentally acceptable manner.
Bioremediation in the subsurface can remove a variety oforganic compounds. The evaluation of the bioremediationprocess should include observation of the removal of theorganic compound(s) in a manner so as to provide a massbalance. In the laboratory, mass balances can be approximatedwith the use of proper abiotic controls. The use of abioticcontrols in the laboratory evaluation cannot be overempha-sized. In the field, a mass balance can be approximated withthe collection of samples prior to remediation to evaluate theamount of organic compound present. Samples collected sub-sequent to the initiation of bioremediation can be evaluatedrelative to the initial concentrations. If the bioremediationeffort is succeeding, a reduction in the concentration of theorganic compound should be observed. In areas not undergo-ing bioremediation, the concentration of the organic com-pound should remain relatively unchanged. If TEAs are added,removal of these compounds also suggests biological activity.The presence of metabolic intermediates also indicates thatbiological processes are occurring. Other observations, suchas adaptation or acclimation or an increase in microbial activ-ity of the compound being degraded, are positive indicators ofthe enhancement of naturally occurring bacteria to achievebioremediation.
13.4 ReferencesAlexander, M. 1981. Biodegradation of Chemicals of Envi-
ronmental Concern. Science 211:132-138.
Atlas, R.M. 1981. Microbial Degradation of Petroleum Hy-drocarbons: an Environmental Perspective. Appl. Environ.Microbiol. Vol. 45, pp. 180-209.
Atlas, R.M. 1984. Petroleum Microbiology. Macmillan, NewYork.
Back, W. 1989. Early Concepts of the Role of Microorgan-isms in Hydrogeology. Ground Water 27:618-622.
Balkwill, D.L. and W.C. Ghiorse. 1985. Characterization ofSubsurface Bacteria Associated with Two Shallow Aqui-fers in Oklahoma. Appl. Environ. Microbiol. 50:560-588.
Beeman, R.E. and J.M. Suflita. 1987. Microbial Ecology of aShallow Unconfined Ground-water Aquifer Polluted byMunicipal Landfill Leachatc. Microb. Ecol. 14:39-54.
Brown, J.F., R.E. Wagner, H. Feng, DL. Bedard, M.J. Brennen,J.C. Carnahan, and R.J. May. 1987. Environmental De-chlorination of PCBS. Environ. Toxicol. Chem. 6579-593,
199
Caskey, W.H. and J.M. Tiedje. 1980. The Reduction of Ni-trate to Ammonium by a Clostridium sp. Isolated fromSoil. J. Gen. Microbiol. 119:217-223.
Cline, P.V., J.J. Delfino, and T. Potter. 1988. Degradation andAdvection of 1,1,1-Tricholorethane in the Saturated ZoneContaining Residual Solvent. In: Superfund ’88, Hazard-ous Materials Control Research Institute, Silver Spring,MD.
Dragun, J. 1988. The Soil Chemistry of Hazardous Materials.Hazardous Materials Control Research Institute, SilverSpring, MD.
Ehrlich, G.G., E.M. Godsy, D.F. Goerlitz, and M.F. Hult.1983. Microbial Ecology of a Creosote-ContaminatedAquifer at St, Louis Park, Minnesota. Dev. Ind. Microbiol.24:235-245.
Ehrlich, G.G., R.A. Schroeder, and P. Martin. 1985. MicrobialPopulations in a Jet-Fuel Contaminated Shallow Aquiferat Tustin, California. U.S. Geological Survey Open-FileReport 85-335.
Genthner, B.R.S., W.A. Price, and H.P. Pritchard. 1989.Anaerobic Degradation of Chloroaromatic Compoundsin Aquifer Sediments under a Variety of EnrichmentConditions. Appl. Environ. Microbiol. 55:1466-1471.
Ghiorse, W.C. and D.L. Balkwill. 1983. Enumeration andMorphological Characterization of Bacteria Indigenousto Subsurface Environments. Dev. Ind. Microbiol. 24:213-224.
Ghiorse, W.C. and D.L. Balkwill. 1985. Microbiological Char-acterization of Subsurface Environments. In: Ground-Water Quality, C.H. Ward, W. Giger, and P.L. McCarty(eds.), John Wiley & Sons, New York, pp. 536-556.
Ghiorse, W.C. and J.T. Wilson. 1988. Microbial Ecology ofthe Terrestrial Subsurface. Adv. Appl. Microbiol. 33:107-172.
Gibson, D. T. (ed.). 1984. Microbial Degradation of OrganicCompounds. Marcel Dekker, New York.
Grbic-Galic, D. and T.E. Vogel. 1987. Transformation ofToluene and Benzene by Mixed Methanogenic Cultures.Appl. Environ. Microbiol, 53:254-260.
Harvey, R.W., R.L. Smith, and L. George. 1984. Effect ofOrganic Contaminants upon Microbial Distribution andHeterotrophic Uptake in a Cape Cod, Massachusetts Aqui-fer. Appl. Environ. Microbiol, 48:1197-1202.
Henson, J.M., M.V. Yates, and J.W. Cochran. 1989. Metabo-lism of Chlorinated Methanes, Ethanes, and Ethylenes bya Mixed Bacterial Culture Growing on Methane. J.Indust. Microbiol. 4:29-35.
Humenick, M.J.H., L.N. Bitton, and C.F. Maddox. 1982.Natural Restoration of Ground Water in UCG. In Situ6:107-125
Jamison, V.W., R.L. Raymond, and J.O. Hudson. 1975. Bio-degradation of High-Octane Gasoline in Groundwater.Dev. Ind. Microbiol. 16:305-312.
Knowles, R. 1982. Denitrification. Microbiol. Rev. 46:43-70.
Ladd, T.I., et al. 1982. Heterotrophic Activity and Biodegra-dation of Labile and Refractory Compounds by GroundWater and Stream Microbial Populations. Appl. Environ.Microbiol. 44:321-329.
Lee, M.D. and C.H. Ward. 1985. Biological Methods for theRestoration of Contaminated Aquifers. Environ. Toxicol.Chem. 4:721-726.
Lee, M.D., J.M. Thomas, R.C. Borden, P.B. Bedient, J.T.Wilson, and C.H. Ward. 1988. Biorestoration of AquifersContaminated with Organic Compounds. CRC Crit. Rev.Environ. Control 18:29-89.
McCarty, P.L. 1975. Energetics of Organic Matter Degrada-tion. In: Water Pollution Microbiology, Vol. 1, R. Mitchell(ed.), Wiley-Interscience, New York, pp. 91-110.
McGinnis, G. D., H. Borazjani, L.K. McFarland, D.F. Pope,and D.A. Strobcl. 1988. Characterization and LaboratorySoil Treatability Studies for Creosote and Pentachloro-phenol Sludges and Contaminated Soil. EPA/600/2-88-055 (NTIS PB89-10992O).
McNabb, J.F. and G.E. Mallard. 1984. Microbiological Sam-pling in the Assessment of Groundwater Pollution. InGroundwater Pollution Microbiology, G. Bitton and C. P.Gerba (eds.), John Wiley & Sons, New York, pp. 235-260.
Postgate, J.R. 1979. The Sulphate-Reducing Bacteria. Cam-bridge University Press, Cambridge.
Quensen, J.F., J.M. Tiedje, and S.A. Boyd. 1988. ReductiveDechlorination of Polychlorinated Biphenyls by Anaero-bic Microorganisms from Sediments. Science 242:752-754.
Raymond, R. L., V.W. Jamison, and J.O. Hudson. 1976. Ben-eficial Stimulation of Bacterial Activity in Ground WaterContaining Petroleum Products. AICE Symposium Se-ries 73:390-404.
Roberts, P. V., L. Semprini, G.D. Hopkins, D. Grbic-Galic,P.L. McCarty, and M. Reinhard. 1989. In-Situ Restora-tion of Chlorinated Aliphatics by Methanotrophic Bacte-ria. EPA/600/2-89-033 (NTIS PB89-219992).
Smolenski, W.J. and J.M. Suflita. 1987. Biodegradation ofCresol Isomers in Anoxic Aquifers. Appl. Environ.Microbiol. 53:710-716.
200
Suflita, J.M., S.A. Gibson, and R.E. Beeman. 1988. Anaero- Wilson, B.H. and J.F. Rees. 1985. Biotransformation of Gaso-bic Biotransformation of Pollutant Chemicals in Aqui-fers. J. Ind. Microbiol. 3:179-194.
Thomas, J.M. and C.H. Ward. 1989. In Situ Biorestoration ofOrganic Contaminants in the Subsurface. Environ. Sci.Technol. 23:760-766.
Updegraff, D.M. 1982. Plugging and Penetration of Petro-leum Reservoir Rock by Microorganisms. In: Proc. 1982Int. Conf. on Microbial Enhancement of Oil Recovery.
Ventullo, R.M. and R.J. Larson, 1985. Metabolic Diversityand Activity of Heterotrophic Bacteria in Ground Water,In Environ. Toxicol. Chem. 4:759-771.
Vogel, T.E. and D. Grbic-Galic. 1986. Incorporation of Oxy-gen into Toluene and Benzene During Anaerobic Fer-mentative Transformation. Appl. Environ. Microbiol.52:200-202.
Waksman, S.A. 1916. Bacterial Numbers in Soil, at DifferentDepths, and in Different Seasons of the Year. Soil Sci-ence 1:363-380.
Webster, J.J., G.J. Hampton, J.T. Wilson, W.C. Ghiorse, andF.R. Leach. 1985. Determination of Microbial Numbersin Subsurface Environments. Ground Water 23:17-25.
line Hydrocarbons in Methanogenic Aquifer Material. In:Proc. NWWA/API Conf. on Petroleum Hydrocarbonsand Organic Chemicals in Ground Water-Prevention,Detection and Restoration, National Water Well Associa-tion, Dublin, OH, pp. 128-139.
Wilson, J.T. and C.H. Ward. 1987. Opportunities for Biorec-lamation of Aquifers Contaminated with Petroleum Hy-drocarbons. Dev. Indust. Microbiol. 27:109-116.
Wilson, J.T., J.F. McNabb, D.L. Balkwill, and W.C. Ghiorse.1983. Enumeration and Characterization of Bacteria In-digenous to a Shallow Water-Table Aquifer. GroundWater 21:134-142.
Wilson, J.T. J.F. McNabb, J.W. Cochran, T.H. Wang, M.B.Tom son, and P.B Bedient. 1985a. Influence of MicrobialAdaptation on the Fate of Organic Pollutants in GroundWater. Environ. Toxicol. Chem. 4:721-726.
Wilson, J.T., M.J. Noonan, and J.F. McNabb. 1985b. Biodeg-radation of Contaminants in the Subsurface. In: Ground-Water Quality, C.H. Ward, W. Giger, and P.L. McCarty,(eds.), John Wiley & Sons, New York, pp. 483-498.
Wilson, J.T., L.E. Leach, M. Henson, and J.N. Jones. 1986. InSitu Biorestoration as a Ground Water Remediation Tech-nique. Ground Water Monitoring Review 6(4):56-64.
201
PART III: SOIL AND GROUND-WATER REMEDIATIONChapter 14
Soil and Ground-Water Remediation: Basic ApproachesRonald C. Sims and Judith L. Sims
Subsurface remediation includes identifying, quantify-ing, and controlling contaminant source(s); considering cleanuplevels required for each medium (air, soil, and ground water)to protect human health and the environment; and selectingtreatment technologies based on information obtained con-cerning source(s) and cleanup levels. The challenge is toeffectively relate site characterization activities to selectingthe most appropriate remediation technologies for contami-nated soils and ground water at hazardous waste sites. Effec-tively relating these activities with technology selectionimproves the efficiency, purpose, and results of both sitecharacterization and remediation technique selection. Thischapter addresses specific subsurface physical, chemical, andbiological processes that have been discussed in previouschapters within the context of(1) site characterization require-ments, (2) evaluation and selection of remediation techniquesand treatment trains utilizing several techniques, and (3) de-sign of monitoring programs.
There is currently a lack of methods and approaches forevaluating and selecting remedial technologies for site-spe-cific scenarios in the area of subsurface remediation, includ-ing soil and ground-water remediation. This chapter presentsa rational approach for addressing soil and ground-waterremedial technologies, including evaluating and selecting newtechnologies as they become available to the user community.Specific soil and aquifer remediation techniques, includingapplications and limitations, also are discussed.
14.1 Conceptual Approach to Soil and Ground-Water Remediation
A conceptual framework for soil remediation techniqueevaluation, selection, and monitoring, based on current infor-mation and activities employed at hazardous waste sites isproposed. The conceptual framework is the chemical massbalance, the cornerstone of science and engineering researchand industry. The concept of a chemical mass balance isfamiliar to professionals trained in the physical or life sci-ences or in engineering. It provides a rational and fundamentalbasis for asking specific questions and obtaining specificinformation that is necessary for determining fate and behav-ior, for evaluating and selecting treatment options, and formonitoring treatment effectiveness at both laboratory-scale
and field-scale. A mass balance approach also meets the goalof obtaining quantitative accuracy about the amount of con-taminants initially present at an uncontrolled site. While amass balance, or materials balance, is routinely conducted onaboveground treatment processes (Bailey and Ollis, 1986;Benefield et al., 1982; Corbitt, 1989; Metcalf and Eddy, Inc.,1979), and for ground-water processes (Willis and Yeh, 1987;Wilson et al., 1989), a mass balance approach has generallynot been applied to the soil environment or to the subsurface/surface system to link characterization activities and treat-ment technology selection. The information needed to con-struct a mass balance for contamination at a site simultaneouslyaddresses site characterization and remediation evaluationand selection.
The conceptual approach for the soil and ground-watersubsurface environment at a contaminated site is illustrated inFigure 14-1. The contaminated subsurface is a system gener-ally consisting of two phases (solid and fluid) and five com-partments (gas, an inorganic mineral solid compartment anorganic matter solid compartment, water, and oil [NAPL])(Sims et al., 1989). Generally NAPLs are subdivided into twoclasses: those that are lighter than water (LNAPLs), and thosewith a density greater than water (DNAPLs). LNAPLs in-clude hydrocarbon fuels, such as gasoline, heating oil, kero-sene, jet fuel, and aviation gas. DNAPLs include chlorinatedhydrocarbons, such as 1,1,1 -trichloroethane, carbon tetrachlo-ride, chlorophenols, chlorobenzenes, tetrachloroethylene, andpolychlorinated biphenyls (PCBs).
Specific subsurface processes concerning water move-ment, sampling, sorption and reaction, and degradation arediscussed in the previous chapters. The processes and termi-nology described in the previous chapters will be used in thischapter for the discussion of the components of a massbalance and the mass balance approach to evaluation andselection of soil remediation techniques.
Interphase transfer potential for waste constituents amongoil (waste or NAPL), water, air, and solid (organic and inor-ganic) phases of a subsurface system is affected by the relativeaffinity of waste constituents for each phase shown in Figure14-1, and may be quantified through calculation of distribu-tion coefficients (Loehr, 1989; Sims et al., 1988; U.S. EPA,
203
Figure 14-1. Mass balance conceptual framework for the soil end ground-water subsurface environment at a contaminated site.
1986). Distribution coefficients are calculated as the ratio ofthe concentration of a chemical in the soil (or aquifer materi-als), oil, or gas phases to the concentration of a chemical in thewater phase. A waste chemical, depending on its tendency tobe associated with each phase, will distribute itself among thephases, and can be quantified in terms of distribution coeffi-cients. Distribution coefficients are available for a variety ofchemicals and can be expressed as ratios of the concentrationsof a chemical between two phases in the subsurface
Kd = Concentration in solid phase/Concentration in aque-ous phase
KO = Concentration in oil phase/Concentration in aqueousphase
Kh = Concentration in air phase/Concentration in aqueous
When distribution coefficients are not available, they can beestimated using structure-activity relationships (SARs) or canbe determined in laboratory tests (Sims et al., 1988). Foradditional detail concerning these processes, see Chapters 10and 11.
Distribution coefficients have been used most success-fully with organic chemicals. However, since metals distrib-ute among the phases of the subsurface systems describedpreviously, distribution coefficients also may be used, along
with multiphase metal speciation information (Sims et al.,1984), to evaluate metal distribution in a contaminated sub-surface system. For additional detail concerning these pro-cesses see Chapter 12.
Knowledge of migration and distribution of chemicalsand chemical intermediates among the phases and compart-ments of a contaminated subsurface system (illustrated inFigure 14-2) provides fundamental information about the fateand behavior of contaminants, which can be used for selectingand evaluating subsurface remedial techniques. Retardationof the downward transport (leaching potential) and upwardtransport (volatilization potential) is referred to as immobili-zation of waste constituents, and has been related to thesubsurface organic matter content, especially for hydrophobicchemicals (Nkedi-Kizza et al., 1983), soil moisture (Mahmood,1989), and presence and concentration of organic solvents(Mahmood and Sims, 1986; Rao et al., 1985).
In summary, subsurface processes described above, com-bined with information about the movement of fluids asdiscussed in Chapters 4, 5, and 6 (gases, aqueous phase, andpure product flow) in the unsaturated and saturated zones,provide the inputs into the chemical mass balance that can beused for (1) characterizing a site; (2) assessing the problem ofmobility; (3) evaluating treatment techniques; and (4) identi-
204
Figure 14-2. Interphase transfer potential of chemicals in thesubsurface (from Sims et al., 1990).
fying chemicals in specific phases for monitoring treatmenteffectiveness.
14.2 MethodologyUsing the chemical mass balance approach, the authors of
this Handbook developed a methodology for integrating datacollection activities at CERCLA sites to address simultaneoussite characterization and remediation technique selection, Theproposed methodology consists of four elements: (1) charac-terization, (2) assessment of the problem, (3) treatment (train)selection, and (4) monitoring treatment performance (Figure14-3). The first element involves characterization in the con-text of waste/subsurface/site interactions to address the ques-tion, “Where is the contamination and in what form(s) does itexist?” The second element, assessment of the problem, uti-lizes subsurface fate and behavior information to address thequestion “Where is the contamination going under the influ-ence of natural processes?” The problem can be define in thecontext of mobility versus degradation for chemicals at a site.Using mathematical models or other tools, the chemicals canbe ranked in order of their relative tendencies to leach, tovolatilize, to move in a NAPL phase and to remain in-placeunder site-specific conditions. Containment and/or treatmentoptions then can be selected that are chemical-specific andthat address specific escape and attenuation pathways (third
Methodology for Integrating Site Characterization with Subsurface Remediation
Problem Assessment Treatment (train) Monitoring
Figure 14-3. Methodology using mass balance approach for integrating data collection activities at a contaminated site.
205
element). Therefore, treatment trains can be selected to ad-dress specific waste phases at specific times during remedia-tion (volatile, leachate, solid phase, and pure product), withthe selection based upon results of a mass balance evaluationthrough time to identify the fate of each waste phase. Finallymonitoring programs can be designed for specific chemicalsin specific phases in the subsurface at specific times (fourthelement).
The approach for using the methodology described aboveconsists of applying a mass balance for each element of themethodology. This approach assists in the collection of spe-cific information that is transferable among all four elementsof the methodology, and also addresses the technical issues ofsoil remediation within the context of regulatory goals.
14.2.1 Site CharacterizationIdentifying waste sources by subsurface phases, i.e., iden-
tification and amount (if possible) of waste constituents asso-ciated with solid and fluid phases (Figure 14-1), allowsassessment of the magnitude (mass) and physical form(s) ofwaste that must be treated. This assessment comprises the firststep in the mass balance characterization of waste sources at asite.
Wastewater historically has been characterized and sub-sequently treated in terms of its interaction and potentialimpact of the assimilative capacity of surface water receiversystems, generally rivers or lakes (e.g., requiring measure-ment of characteristics such as oxygen-demanding substances,nutrients, and levels of substances toxic to aquatic organisms).However, a waste characterization program at a hazardouswaste site addresses the vadose zone and ground water, inaddition to surface water, as the receiver systems (e.g., requir-ing measurement of characteristics that reflect individualchemical mobility and destruction in the subsurface environ-ment and those that affect human health as well as character-istics that affect environmental toxicity). Also, it describes thebehavioral interaction of waste chemicals in each surface andsubsurface phase. Thus, hazardous waste is more appropri-ately characterized in terms of the interaction and potentialimpact on the subsurface assimilative capacity.
Specific site characteristics important for describing andassessing the environmental behavior and fate of organicconstituents in the soil and subsurface are listed in Table 14-1.For each chemical, or chemical class, required informationincludes (1) characteristics related to potential leaching (e.g.,water solubility, octanol/water partition coefficient, solid sorp-tion coefficient); (2) characteristics related to potential vola-tilization (e.g., vapor pressure, relative volatilization index);(3) characteristics related to potential degradation (e.g., half-life, degradation rate, degradability index); and (4) character-istics related to chemical reactivity (e.g., hydrolysis half-life,soil redox potential) (Sims et al., 1984). The informationpresented in Table 14-1 also is used to assess problem(s)concerning migration potential at a site and to evaluate andselect containment- and treatment-management options.
If the distribution of waste chemicals among phases thatcomprise the soil and subsurface at a site are determined, then
potential pathways of transport, or escape, from a site can beindicated. Therefore, exposure pathways for human healthand the environment may be evaluated, i.e., risk assessmentcan be made. Through a determination of subsurface flowconditions as part of site characterization activities (aqueous,gas, and pure product flow in the vadose zone and aqueousplume and pure product movement in the saturated zone), themass of material moving through a site and potential move-ment off site can be assessed:
concentration (mass/vol) X rate of flow (vol/time)= mass flow at site (mass/time)
This information is combined with additional information,discussed in the next section, that is needed to assess theproblem(s) with respect to treatment technique selection.
The U.S. Environmental Protection Agency (EPA’s) Rob-ert S. Kerr Environmental Research Laboratory, as part of itsSuperfund Technology Support Center Program activities,provides assistance to EPA regional offices and state regula-tory agencies about appropriate site characterization activitiesat Superfund sites and other uncontrolled hazardous wastesites to support selection of effective remediation tcchnol~gies. Table 14-2 presents examples of recommended siteevaluation and characterization actions as related to the use ofsoil and the subsurface as the receiver system at uncontrolledhazardous waste sites (Scalf and Draper, 1989).
14.2.2 Assessment of ProblemAssessment of the contamination involves organizing the
information obtained from site characterization activities toevaluate the transport and degradation behavior of each chemi-cal of concern at a site under consideration. Specifically, therate of transport can be compared with the rate of degradationto determine if transport is significant relative to degradation.This approach to problem(s) assessment will allow chemicalsto be prioritized individually according to (1) magnitude andrate of transport (escape) from a site, (2) persistence, and (3)pathway(s) of migration from a site. Treatment techniqueevaluation and selection then can be based upon specificcombinations of chemical and physical phase-migrationpathway.
Interfacing subsurface-based behavioral characteristicsof specific contaminants (Table 14-1) with specific site andsubsurface properties allows an assessment of the problem(s)related to contamination of other media (due to mobility),including the ground water under the contaminated area, theatmosphere over the site or at the site boundaries, surfacewaters, and/or persistence of chemicals at a site. Pathways ofmovement and potential mechanisms of removal of contami-nants at a specific site are illustrated in Figure 14-2. Thiselement of the methodology functions to identify chemicalsthat will (1) migrate upward (volatilization), (2) migrate down-ward (leaching), (3) migrate laterally (aqueous plume andpure product), (4) degrade, and (5) remain at the site aspersistent chemicals. By ranking the chemicals in the order inwhich lhey migrate or persist, chemicals can be prioritizedwith regard to urgency for treatment and for monitoring.
206
Table 14-1. Subsurface-Based Waste Characterization
Source: Sims et al., 1984
Waste characteristics identified in Table 14-1, includingpotential sorption, degradation, and volatilization at a site, canbe determined in laboratory mass balance tests, using waste/soil mixtures from a site. These characteristics can be used toevaluate the fate of the waste at the site, and to generatespecific data that can be used to develop treatment approaches.Figure 14-4 illustrates a laboratory flask apparatus that can beused to develop a chemical mass balance by measuring inter-phase transfer potential of chemicals as well as degradationpotential at a site (Park et al., 1990).
The contaminated material is placed in a flask, which isthen closed and incubated under controlled conditions for aperiod of time. During the incubation period, air is drownthrough the flask and then through a sorbent material. Volatil-ized materials are collected by the sorbent and are measured toestimate volatilization loss of the constituents of interest. Atthe end of the incubation period, a portion of the contaminatedsoil is treated with an extracting solution to determine theextent of loss of the constituents in the soil matrix. This losscan be attributed to degradation and possible immobilizationin the soil materials. It is necessary to select an appropriateextracting solution and procedure to maximize constituentrecovery from a soil-waste mixture (Coover et al., 1987).Another portion of the soil is leached with water to determineleaching potential of the remaining constituents. Abiotic andbiological processes involved in removal of the parent com-
pound are evaluated by comparing microbially active soil/waste mixtures with mixtures that have been treated with amicrobial poison, e.g., mercuric chloride or propylene oxide.Samples generated from the different phases of the system inmicrocosm mass balance studies identified above can beanalyzed for intermediate degradation products and used inbioassay studies to provide information concerning transfor-mation and detoxification processes.
The use of a procedure incorporating features illustratedby the use of this microcosm (Figure 14-4) is crucial to obtaina materials balance of waste constituents in the subsurfacesystem, Examples of such protocols may be found in EPAguidance documents and research reports (Loehr, 1989; Simset al., 1988; EPA, 1986; and Park et al., 1990). Contaminatedmaterials also can be spiked with radiolabeled chemicals;tracking the fate of the chemicals as they move through themultiple phases of the soil system also provides a materialsmass balance.
The mass balance approach identified above usuaIly rep-resents optimum conditions with respect to mixing, contact ofsol id materials with waste constituents, and homogeneousconditions throughout the laboratory microcosm; therefore, itdoes not incorporate site nonhomogeneity in the evaluation.This aspect must be defined during site characterization ac-tivities and evaluated with regard to potential effect on fateand behavior regarding migration and persistence at the site(problem assessment).
In addition to the laboratory tests described, bench-scalereactors, pilot-scale reactors and/or field-scale plots may beused to generate mass balance information for problem as-sessment. The set of experimental conditions (e.g., tempera-ture, moisture, waste concentration) under which the studieswere conducted and experimental results should be presented.
Information from the performance of site characterizationand experimental mass balance studies may be integrated withthe use of comprehensive mathematical modeling to aid inproblem assessment. In general, models are used to analyzethe behavior of an environmental system under both current(or past) and anticipated (or future) conditions (Donagian andRao, 1986). A mathematical model provides a tool for (1)integrating degradation and partitioning processes with site-,soil-, and waste-specific characterization; (2) simulating thebehavior of waste constituents in a contaminated soil; and (3)predicting the pathways of migration through the contami-nated area, and therefore pathways of exposure to humans andto the environment. DiGiulio and Suffet (1988) and Weaver etal. (1989) have presented guidance on the selection of appro-priate subsurface zone models for site-specific applications,focusing on recognition of limitations of process descriptionsof models and difficulties in obtaining input parameters re-quired by these process descriptions.
The Regulatory and Investigative Treatment Zone Model(RITZ), developed at the EPA’s Robert S. Kerr Environmen-tal Research Laboratory by Short (1986) is an example of avadose zone model that has been used to describe the potentialfate and behavior of organic constituents in a contaminatedsoil system (U.S. EPA, 1988a). The RITZ ModeI is based on
207
an approach developed by Jury et al. (1983). An expandedversion of RITZ, the Vadose Zone Interactive Processes (VIP)model, incorporates predictive capabilities for the dynamicbehavior of organic constituents in unsaturated soil systemsunder conditions of variable precipitation, temperature, andwaste concentrations (McLean et al., 1988; Stevens et al.1988, 1989; Symons et al., 1988; U.S. EPA, 1986). Both theRITZ and VIP models simulate vadose zone processes, in-cluding volatilization, degradation, sorption/desorption, ad-vection, and dispersion (Grenney et al., 1987).
For example, the VIP model was used to evaluate therelative tendencies for a group of pesticides to volatilize andto leach under specific waste-soil conditions (McLean et al.,1988). Information input into the model included half-life(measured in laboratory tests), distribution coefficients(Kd,Kh,Ko) (calculated), soil texture and moisture (measured),and site-specific climatic data (rainfall and temperature). Re-sults are presented in Table 14-3. The ranking of pesticidesprovided by the model indicated that the tendency of thepesticides to volatilize was not similar to their tendency toleach (McLean et al. 1988). This information can be used toassess which chemicals are likely to volatilize first, whichchemicals are likely to leach first, and which chemicals arepersistent under site-specific conditions. In addition to assist-ing in the problem assessment step of the methodology,mathematical models also can be used to design studies forevaluation and selection of treatment options for these chemi-cals, as well as to design monitoring strategies (i.e., whichchemicals to monitor in which media).
With regard to ground-water models that can be used aspart of the problem assessment, the International GroundWater Modeling Center (IGWMC), through EPA, publishedinformation about the kinds and availability of models, theirspecific characteristics, and the information, data, and techni-cal expertise needed for their operation (U.S. EPA, 1988 b).Ground-water models also have been addressed within thecontext of scientific and regulatory applications, with selectedcase studies, by the National Research Council (NAS, 1990).
Table 14-3. Ratios of Concentration of Pesticides BetweenWater/Soil and Air/Soil at 15 cm After 81 Days(Ranked in Order from Greatest Potential forLeaching and Volatilization to Least Potential)
A numerical model, BIOPLUME, was developed to simu-late oxygen-limited biodegradation in ground-water environ-ments. BIOPLUME simulates advection, dispersion, andretardation processes as well as the reaction between oxygenand the contaminants under steady, uniform flow (Rifai et al.,1989). BIOPLUME was applied to an aviation gasoline spillsite at Traverse City, Michigan. Model predictions for therates of mass loss closely matched calculated rates from fielddata.
14.2.3 Treatment ApproachesInformation obtained from an integrated assessment (mod-
eling) of the problem (migration and persistence), based upona thorough characterization of waste/soil/site interactions, canbe used to select treatment approaches for further evaluationwith respect to technical and cost-effectiveness factors. Re-sults of characterization and assessment efforts can aid in theidentification of constituents that will require treatment in thefollowing phases: (1) air (volatile) phase, (2) leachate phase,and (3) solid (soil) phase. This approach allows evaluationand comparison of different treatment systems identified pre-viously (in situ and prepared bed). Specifically, if treatment isrequired, the information is used to (1) determine containmentrequirements to prevent contamination of offsite receiversystems; (2) develop techniques to maximize mass transfer ofchemicals affecting a process (e.g., affecting microbial activ-ity through addition of mineral nutrients, oxygen, additionalenergy sources, pH control products, or removal of toxicproducts in order to enhance bioremediation); and (3) design acost-effective and efficient monitoring program to evaluateeffectiveness of treatment.
Containment Requirements. If the major pathway of trans-port is volatilization, containment and treatment to controlvolatilization is required. An inflatable plastic dome erectedover a contaminated site is a containment method that hasbeen used to control escape of volatile constituents at hazard-ous waste silts (St. John and Sikes, 1988). Volatiles are drawnfrom the dome through a conduit and treated in an abovegroundtreatment system. If leaching has been identified as an impor-tant factor, control of soil water movement should be imple-mented. For example, if contaminated materials are expectedto leach downward from the site, run-on and run-off controlscan be implemented, or the contaminated materials can betemporarily removed from the site and a plastic or clay linercan be placed under the site (Lynch and Genes, 1989; Ross etal., 1988). When downward as well as upward migration aresignificant, both volatilization and leaching containment sys-tems can be installed. Some hydrophobic chemicals do nottend to volatilize or to leach but are persistent within the soilsolid phase; therefore, containment efforts may not be re-quired. With regard to the saturated zone, containment isgenerally accomplished by physical barriers (e.g., slurry walls,sheet pilings, grout curtains) or hydraulic barriers (e.g., pumping systems, french drains).
Maximizing Chemical Mass Transfer. An area of sig-nificant research concerns delivery and recovery technologiesfor maximizing mass transfer of chemicals that affect the rateand/or extent of treatment. Murdoch et al. (1988) discussed
209
delivery and recovery technologies, many of which are de-rived from the petroleum and mining industries. While aliquid phase is usually employed for delivery of chemicals,some technologies utilize vapor and solid phases for delivery.Principal recovery technologies involve hydraulic, thermal,and chemical systems. Delivery and recovery techniques areimportant in influencing the success of technologies, includ-ing bioremediation, vapor extraction, and solidification/stabi-lization. Specific delivery and recovery systems for in situtreatment systems identified by EPA include hydraulic frac-turing, radial well drilling, ultrasonic methods, kerfing, jet-induced slurry methods, carbon dioxide injection, hot brineinjection, and cyclic pumping (U.S. EPA, 1990).
14.2.4 Monitoring ProgramA mass balance approach to monitoring, the fourth ele-
ment in the methodology (Figure 14-3), can be performed atlaboratory, pilot, and field scales. Monitoring efforts can befocused on the appropriate environmental phase to evaluatetreatment effectiveness for specific chemicals. If a compre-hensive and thorough evaluation of a specific contaminatedsystem has been conducted, not all chemicals may need to bemonitored in each phase. Specific chemicals wall be associ-ated with specific phases; therefore, a monitoring plan can bedesigned that is chemical/phase specific. This approach alsofocuses analytical efforts so that methods of development arechemical- and phase-specific.
The level of contamination associated with a particulartreatment technology requires monitoring. In addition, thetreatment system components, including delivery and recov-ery systems, maintenance, and structures such as infiltrationgalleries must be monitored.
14.3 Selection of Treatment Methods
14.3.1 Utility of Mathematical ModelsA critical and cost-effective use of modeling in treatment
(train) selection and evaluation is for analysis of proposed oralternative future conditions i.e., the model is used as amanagement or decision-making tool to help answer “what if"questions (Donagian and Rao, 1986). Models also may beused to approximate and estimate the rates and extent oftreatment that may be expected at the field-scale under vary-
ing conditions. Attempting to answer such questions throughdata collection programs would be expensive and practicallyimpossible in many situations. For example, information canbe generated to evaluate the effects of using different ap-preaches for enhancing microbial activity and for acceleratingbiodegradation and detoxification of the contaminated area byaltering environmental conditions that affect microbial activ-ity. Therefore, modeling may be used to assist in the design oftreatability studies for considering and evaluating the applica-tion of different treatment technologies, and therefore to assistin focusing available resources (time and money). Section14.2.2 (Assessment of Problem) provides more informationon the existence, applications, and limitations of mathemati-cal models for vadose zone and ground-water analysis andmanagement.
14.3.2 Treatability StudiesTreatability studies can be used for evaluating and com-
paring rate and extent of remediation among several technolo-gies and also to provide specific information about the potentialapplication of treatment technologies at field scale. Treatabil-ity studies can be conducted in laboratory microcosms orbench-scale reactors, pilot-sale facilities, or in the field. Labo-ratory treatability studies are generally screening studies usedto (1) establish the validity of a technology, (2) generate datathat can be used as indicators of potential to meet performancegoals, and (3) identify parameters for investigation duringbench- or pilot-scale testing. Laboratory treatability studiesare generally not appropriate for generating design or costdata (U.S. EPA, 1989). Pilot-scale testing is conducted togenerate information on quantitative performance, cost, anddesign information. Three proposed categories of treatabilitytesting and associated descriptions are included in Table 14-4(U.S. EPA, 1989 b).
Treatability study results are commonly used to provideinformation on rates and extent of treatment of hazardousorganic constituents when mass transfer rates of potentiallimiting substances are not limiting the treatment. Treatabilitystudies also usually represent optimum conditions with re-spect to mixing, contact of soil solid materials with wasteconstituents and with microorganisms, and homogeneous con-ditions throughout the microcosm. Therefore, treatability stud-ies provide information concerning potential levels oftreatment. Rates and extent of remediation in a prepared bed
Table 14-4. General Comparison of Laboratory Screening, Bench-Scale Testing, and Pilot-Scale Testing
Type Usual Wasteof data Critical No. of process stream lime
Tier generated parameters replicates Study size type volume required Cost, $
Laboratory screening Qualitative several Single/ Jar tests or Batch Small Hours/duplicate
10,000-beaker studies day 50,000
Bench-scale testing Quantitative Few Duplicate/ Bench-top Batch or Medium Days/ 50,000-triplicate (some larger) continuous week 250,000
Pilot-scale testing Quantitative Few Triplicate Pilot-plant Batch or Large Weeks/ 250,000or more (onsite or offsite) continuous month I,000,000
Source: U.S. EPA, 1989
210
or in situ system are generally limited by accessibility and rateof mass transfer of chemical substances to the contaminatedsoil and removal of inhibitory microbial degradation products(Symons and Sims, 1988).
Information from mass balance treatability studies, in-cluding laboratory screening-, bench- and pilot-scale studies,is combined with information about site and waste character-istics to determine applications and limitations of each tech-nology. Information obtained from treatability studies shouldbe focused on identifying ultimate limitations to the use of aremediation technology at a specific site. Limitations areusually related to (1) time required for cleanup, (2) level ofcleanup attainable, and (3) cost of cleanup (Sims et al., 1989).
14.3.3 Treatment TrainsThe use of treatment trains also is important to consider
in an engineering approach for using treatment techniques forsubsurface site remediation. For example, vacuum extractionis known to be applicable to unsaturated sites characterized bypermeable materials containing volatile chemicals. Vacuumextraction also can be used for the degradation of moresemivolatile chemicals. This degradation is accomplished byproviding a source of oxygen (air) to the subsurface environ-ment microorganisms where anoxic conditions exist due torelative slow replenishment of oxygen through atmosphericdiffusion. This is an example of the use of one technology forthe treatment of both volatile and semivolatile chemicals inthe subsurface.
Another example of the use of a treatment train forcreosote-contaminated soil and ground water involves (1)product removal using a pumping system, (2) flushing withwater and surfactants using pump-and-treat technology, and(3) in situ biodegradation of the residual contamination (Kuhnand Piontek, 1989). Each technology is employed in the orderof ease of removal of creosote from the subsurface. Thetreatment train selected was based on a site characterization toidentify where the creosote was located and the mass ofcreosote (including pure product) associated with subsurfacephases, i.e., the vadose zone and aquifer materials. The prob-lem assessment identified the following areas of concern: (1)potential offsite migration of pure product; (2) slow leachingof low levels of creosote contaminants sorbed to soil, subsur-face, and aquifer materials; and (3) presence of high molecu-lar weight polycyclic aromatic compounds that are toxic tohuman health, are nonvolatile, and have very low watersolubilities. Each technology was evaluated in laboratory-scale treatability tests for treatment effectiveness and for caseof application to contaminated materials obtained from thesite. Engineering design and implementation was based onresults of site characterization, mass balance determinations atthe site, and treatability studies,
Information from treatability studies is used to prepare anapproach to the engineering design and implementation of aremediation system at a specific site that combines the treat-ment techniques evaluated to construct an appropriate treat-ment train. The formulation of a treatment train for a sitegenerally is based upon information from simulations (e.g.,
mathematical modeling) generated from mass balance stud-ies, treatability studies, and site/soil characterization data.
14.4 Measurement and Interpretation ofTreatment Effectiveness
Typically, subsurface samples are taken from a treatabil-ity reactor (in situ or prepared bed) from laboratory-, bench-,or pilot-scale studies, or from a field site. Waste constituentsare extracted from the samples with a solvent or are thermallydesorbed. Compound concentration is usually measured in thesolvent extract or the thermal resorption stream using chemi-cal instrumentation (e.g., gas or liquid chromatography withappropriate detectors). This information is termed the “appar-ent loss” of the compound and refers to the observation thatthe compound only has disappeared from the solvent orextraction phase, but does not necessarily represent a chemi-cal mass balance (Park et al., 1990). The change in concentra-tion of the compound in the solvent with time often is used tocalculate rate and extent of decrease in concentration of thecompound in soil. This information is commonly used tointerpret treatment effectiveness for different technologies aswell as to determine engineering strategies and managementapproaches, including (1) time required to attain cleanuptarget concentrations; and (2) effects of environmental factorsor experimental variables (chemical, physical, or biological)on treatment effectiveness.
However, additional information is needed to accuratelymeasure and interpret treatment effectiveness. In order tounderstand treatment mechanisms and to base the selection oftreatment technologies on a rational approach, identificationand measurement of distribution among the physical phasesthat comprise a subsurface system is necessary. In addition,the mechanisms by which a compound may be chemicallyaltered in a subsurface system must be identified and differen-tiated (Dupont and Reineman, 1986; Goring et al., 1975;Guenzi, 1974; Park et al., 1988, 1990; Sims et al., 1988;Stevens et al. 1989; Unterman et al., 1988).
Information obtained about the rate of apparent loss ofchemicals from a subsurface extract can be enhanced withinformation about the (1) interphase transfer potential be-tween solid and gas phases of the subsurface, and (2) knowl-edge of mechanisms of interactions of compounds withsubsurface phases. This information then provides the basisfor a more rational approach to subsurface remediation. Evalu-ation of remediation technology effectiveness also can bebased upon specific media (solid, air) and upon specificmechanisms, such as recovery of the air phase or enhance-ment of abiotic destruction or biological degradation, to im-prove treatment. Evaluation of interphase transfer also allowscharacterization of routes by which chemicals may migratefrom the subsurface to the multimedia environment that thenmay lead to human exposure. Thus, measuring treatmenteffectiveness based upon interphase transfer potential (a massbalance approach) is also valuable for determining risk reduc-tion and implementing risk management strategies (Park etal., 1990). The laboratory flask apparatus used for massbalance determinations (Figure 14-4) also can be used tomeasure and compare potential effectiveness for differenttreatment scenarios.
211
14.5 References Loehr, R. 1989. Treatability Potential for EPA Listed Hazard-Bailey, J.E. and D.F. Ollis. 1986. Biochemical Engineering ous Chemicals in Soil. EPA/600/2-89/011 (NTIS PB89-
Fundamentals, 2nd ed. McGraw-Hill, New York, NY. 166581/AS).
Benefield, L. D., J.F. Judkins, and D.L. Weand. 1982. ProcessChemistry for Water and Wastewater Treatment. Prentice-Hall, Englewood Cliffs, NJ.
Coover, M.P., R.C. Sims, and W.J. Doucette. 1987. Extrac-tion of Polycyclic Aromatic Hydrocarbons from SpikedSoil. Journal of the Association of Official AnalyticalChemists 70:1018-1020.
Corbitt, R.A. 1989. Standard Handbook of EnvironmentalEngineering. McGraw-Hill, New York, NY.
DiGiulio, D.C. and I.H. Suffet. 1988. Effects of Physical,Chemical, and Biological Variability in Modeling Or-ganic Contaminant Migration through Soil. In: Superfund’88, Hazardous Materials Control Research Institute, Sil-ver Spring, MD, pp. 132-137.
Donigian, A. S., Jr. and P.S.C. Rao. 1986. Overview of Terres-trial Processes and Modeling. In: Guidelines for FieldTesting Soil Fate and Transport Models, S.C. Hern andS.M. Melancon, (eds.), EPA/600/4-86/020 (NTIS PB86-209400), pp. 1-32.
Dupont, R.R. and J.A. Reineman. 1986. Evaluation of Volatil-ization of Hazardous Constituents at Hazardous WasteLand Treatment Sites. EPA/600/2-86/071 (NTIS PB86-233939).
Goring, C.A.I., D.A. Laskowski, J,W. Hamaker, and R.W.Miekle. 1975. Principles of Pesticide Degradation in Soil.In: Environmental Dynamics of Pesticides, R. Haque andW. H. Freek, (eds.), Plenum Press, New York, NY.
Grenney, W.J., C.L. Caupp, R.C. Sims, and T.E. Short. 1987.A Mathematical Model for the Fate of Hazardous Sub-stances in Soil: Model Description and ExperimentalResults. Hazardous Wastes & Hazardous Materials 4:223-239.
Guenzi, W.D. (ed). 1974. Pesticides in Soil and Water. SoilScience Society of America, Madison, WI.
Jury, W.A., W.F. Spencer, and W.J. Farmer. 1983. BehaviorAssessment Model for Trace Organics in Soil: ModelDescription. Journal of Environmental Quality 12:558-564.
Kuhn, R.C. and K.R. Piontek. 1989. A Site-Specific In SituTreatment Process Development Program for a WoodPreserving Site. Paper presented at EPA Technical Pro-gram on Oily Waste Fate, Transport, Site Characteriza-tion, and Remediation Seminar, Denver, CO, May 17-18(Organized by John Matthews, EPA Robert S. Kerr Labo-ratory, Ada, OK).
Lynch, J. and B.R. Genes. 1989. Land Treatment of Hydro-carbon Contaminated Soils. In: Petroleum ContaminatedSoils, Vol. 1: Remediation Techniques, EnvironmentalFate, and Risk Assessment, P.T. Kostecki and E.J.Calabrese (eds.), Lewis Publishers, Chelsea, MI, pp. 163-174,
Mahmood, R.J. 1989. Evaluation of Enhanced Mobility ofPAHs in Soil Systems. Ph.D. Dissertation, Department ofCivil and Environmental Engineering, Utah State Univer-sity, Logan, UT.
Mahmood, R.J. and R.C. Sims. 1986. Mobility of Organics inLand Treatment Systems. Journal of Environmental En-gineering (ASCE) 112:236-245.
McLem, J.E., R.C. Sims, W.J. Doucette, C.L. Caupp, andW.J. Grcnney. 1988. Evaluation of Mobility of Pesticidesin Soil using U.S. EPA Methodology. Journal of Environ-mental Engineering (ASCE) 114:689-703.
Metcalf and Eddy, Inc. 1979. Wastewater Engineering: Treat-ment, Disposal, and Reuse. McGraw-Hill, New York,NY.
Murdoch, L., B. Patterson, G. Losonsky, and W. Harrar. 1988.Innovative Technologies of Delivery or Recovery: AReview of Current Research and a Strategy for Maximiz-ing Future Investigations. EPA/600/2-89/066 (NTIS PB90156225/AS).
National Academy of Sciences (NAS). 1990. Ground WaterModels: Scientific and Regulatory Applications. NationalAcademy Press, Washington, DC.
Nkedi-Kizza, P., P.S.C. Rao, and J.W. Johnson. 1983. Ad-sorption of Diuron and 2,4,5-Ton Soil Particle Separates.Journal of Environmental Quality 12:195-197,
Park, K. S., R.C. Sims, W.J. Doucette, and J.E. Matthews.1988. Biological Transformation and Detoxification of7,12-Dimethylbenz(a) anthracene in Soil Systems. Jour-nal Water Pollution Control Federation 60:1822-1825.
Park, K. S., R.C. Sims, R.R. Dupont, W.J. Doucette, and J.E.Matthews. 1990. Fate of PAH Compounds in Two SoilTypes: Influence of Volatilization, Abiotic Loss and Bio-logical Activity. Environ. Toxicol. Chem. 9:187-195.
Rao, P. S. C., A.G, Hornsby, D.P. Kilcrease, and P. Nkedi-Kizza. 1985. Sorption and Transport of HydrophobicOrganic Chemicals in Aqueous and Mixed Solvent Sys-tems: Model Development and Preliminary Evaluation.Journal of Environmental Quality 14:376-383.
Rifai, H. S., P.B Bedient, R.C. Bordon, and J.F. Haasbeek.1989. BIOPLUME II-Computer Model of Two-Dimen-sional Contaminant Transport Under the Influence of
212
Oxygen Limited Biodegradation in Ground Water (User’sManual Version 1.0; Preprocessor Source Code Version1.0; Source Code Version 1.0). EPA/600/8-88/093 (NTISPB89-151 120/AS).
Ross, D., T.P. Marziarz, and A.L. Bourquin. 1988. Bioreme-diation of Hazardous Waste Sites in the USA: CaseHistories. In: Superfund ’88, Hazardous Materials Con-trol Research Institute, Silver Spring, MD, pp. 395-397.
Scalf, M.R. and D.C. Draper. 1989. RSKERL-Ada SuperfundTechnology Support Center The First Two Years. U.S.Environmental Protection Agency (Robert S. Kerr Envi-ronmental Research Laboratory, Ada, OK).
Short, T.E. 1986. Modeling Processes in the UnsaturatedZone. In: Land Treatment A Hazardous Waste Manage-ment Alternative, R.C. Loehr and J. F. Malina (eds.),Water Resources Symposium No. 13, University of TexasPress, Austin, TX, pp. 211-240.
Sims, R.C., D.L. Sorensen, J.L. Sims, J.E. McLean, R.Mahmood, and R.R. Dupont. 1984. Review of In-PlaceTreatment Technologies for Contaminated Surface Soils-Volume 2: Background Information for In-Situ Treat-ment. EPA-540/2-84-O03b (NTIS PB85-124899).
Sims, R.C., W.J. Doucette, J.E. McLean, W.J. Grenney, andR.R. DuPont. 1988. Treatment Potential for 56 EPAListed Hazardous Chemicals in Soil. EPA/600/6-88/001(NTIS PB88-174446).
Sims, J.L., R.C. Sims, and J.E. Matthews. 1989. Bioremedia-tion of Contaminated Soils. EPA/600/9-89/073 (NTISPB90-164047).
Sims J.L.. R.C. Sims, and J.E. Matthews. 1990. Approach to
St.
Symons, B. D., R.C. Sims, and W.J. Grenney. 1988. Fate andTransport of Organics in Soil: Model Predictions andExperimental Results. Journal Water Pollution ControlFederation 60:1684-1693.
Unterman, R., D.L. Bedard, M.J. Brennan, L.H. Bopp, F.J.Mondcllo, R.E. Brooks, D.P. Bobley, J.B. McDerrnotqC.C. Schwartz, and D.K. Dietrich. 1988. Biological Ap-proaches for Polychlorinated Biphenyl Degradation. InEnvironmental Biotechnology - Reducing Risks fromEnvironmental Chemicals through Biotechnology, G.S.Omenn (ed.), Plenum Press, New York, NY, pp. 253-269.
U.S. Environmental Protection Agency (EPA). 1984. Reviewof In-Place Treatment Techniques for Contaminated Sur-face Soils. EPA-540/2-84-O03a (NTIS PB85-124881).
U.S. Environmental Protection Agency (EPA). 1986. PermitGuidance Manual on Hazardous Waste Land TreatmentDemonstrations. EPA-530/SW-86-032 (NTIS PB86-229184).
U.S. Environmcntal Protection Agency (EPA). 1988a. Inter-activc Simulation of the Fate of Hazardous Chemicalsduring Land Treatment of Oily Wastes: RITZ User’sGuide. EPA/600/8-88-OOl (NTIS PB88-195540).
U.S. Environmental Protection Agency (EPA). 1988b. Ground-watcr Modeling: An Overview and Status Report. EPA/600/2-89/028 (NTIS PB89-229497). Also available fromInternational Ground Water Modeling Center, Butler Uni-versity, Indianapolis, IN.
U.S. Environmental Protection Agency (EPA). 1989. Guidefor Conducting Treatability Studies under CERCLA. EPA/540/2-89/058.
Bioremediation of Contaminated Soils. Hazardous Waste U.S. Environmental Protection Agency (EPA). 1990. Hand-and Hazardous Materials 7(2): 117-149. book on In Situ Treatment of Hazardous Waste-Contami-
nated Soils. EPA/540/2-90-W2 (NTIS PB90-155607).John. W.D. and D.J. Sikes. 1988. Complex IndustrialWaste Sites. In: Environmental Biotechnology - Reduc-ing Risks from Environmental Chemicals through Bio-technology, G.S. Omenn (ed.), Plenum Press, NcwYork, NY, pp. 237-252.
Stevens, D.K., W.J. Grenney, and Z. Yan. 1988. User’s Manual:Vadose Zone Interactive Processes Model. Utah StateUniversity, Logan, UT.
Stevens, D.K., W.J. Grenney, Z. Yan, and R.C. Sims. 1989.Sensitive Parameter Evaluation for a Vadose Zone Fateand Transport Model. EPA/600/2-89/039 (NTIS PB89-213987/AS).
Symons, B.D. and R.C. Sims. 1988. Assessing Detoxificationof a Complex Hazardous Waste Using the MicrotoxTM
Bioassay. Archives of Environmental Contamination andToxicology 17:497-505.
Weaver, J., C.G. Enfield, S. Yates, D. Kreamer, and D. White.1989. Predicting Subsurface Contaminant Transport andTransformation: Considerations for Model Selection andField Validation. EPA/600/2-89/045 (NTIS PB90-155615).
Willis, R. and W. W-G. Yeh. 1987. Groundwater SystemsPlanning and Management. Prentice Hall, EnglewoodCliffs, NJ.
Wilson, J.T., L.E. Leach, J. Michalowski, S. Vandegrift andR. Callaway. 1989. In Situ Bioremediation of Spills fromUnderground Storage Tanks: New Approaches for SiteCharacterization, Project Design, and Evaluation of Per-formance. EPA/600/2-89/042 (NTIS PB89-219976/AS),
213
Chapter 15Remediation Techniques for Contaminated Soils
Ronald C. Sims and Judith L. Sims
Currently, many remedial techniques are being used andevaluated for cleanup of contaminated soils. Tables 15-1 and15-2 list participants in the U.S. Environmental ProtectionAgency (EPA) SITE program that are testing and evaluatingremedial technologies applicable to contaminated soils (U.S.EPA, 1989f). Table 15-3 summarizes technologies applicableto contaminated soils that are currently being demonstratedand evaluated in the NATO/CCMS Pilot Study, Demonstra-tion of Remedial Action Technologies for Contaminated Landand Groundwater (U.S. EPA, 1989d).
Selected physical, chemical, biological, thermal, and fixa-tion/ encapsulation soil remediation techniques were catego-rized as in situ and prepared bed and are summarized in Table15-4 (Rich and Cherry, 1987; U.S. EPA, 1987, 1988c, 1989b).Each soil remediation technique also was evaluated withrespect to function (separation, detoxification, etc.); potentialfor formation of residuals/transformation products; applica-tions; and limitations. This chapter presents a subset of thesetechniques, evaluated at pilot or field scale, that were selectedfor additional description.
15.1 In Situ versus Prepared Bed SoilRemediation
The vadose zone is the region extending from the groundsurface to the upper surface of the principal water-bearingformation. It is divided into three characteristic areas or belts.The uppermost belt consists of soil and other materials that lienear to the surface and discharge perceptible quantities ofwater into the atmosphere. The water is discharged by theaction of plants or by soil evaporation and convection. Thelowest belt, the capillary fringe, is located immediately abovethe water table and contains water drawn up from the zone ofsaturation by capillary action. The intermediate belt lies be-tween the belt of soil water and the capillary fringe (Lehr,1988). In this chapter, soil remediation techniques address thevadose zone and situations where the saturated zone is engi-neered to become unsaturated, e.g., when ground water ispumped to create an unsaturated zone.
The two soil treatment processes discussed in this chapterare in situ treatment and prepared bed treatment. In situtreatment consists of treating contaminated soil in place, i.e.,the contaminated soil is not moved from the ground. Mile-stone publications that should be consulted for scientific andengineering information specifically addressing in situ treat-
ment include Sims et al. (1984); U.S. EPA (1984); U.S. EPA(1990); Sims et al. (1989); and Dupont et al. (1988).
In a prepared bed system, the contaminated soil may beeither (1) physically moved from its original site to a newlyprepared area, which has been designed to enhance treatmentand/or to prevent transport of contaminants from the site; or(2) removed from the site to a storage area while the originallocation is prepared for use, then returned to the bed, wheretreatment is accomplished. Preparation of the bed may includeplacement of a clay or plastic liner to retard transport ofcontaminants from the site or addition of uncontaminated soilto provide additional treatment medium. Treatment may beenhanced with biological and/or physical/chemical methods,as with in situ systems (Sims and Sims, 1986; Sims et al.,1989). Prepared bed treatment approaches are based on modi-fications of principles developed in the areas of land applica-tion of solid and liquid wastes and in land treatment ofhazardous wastes (Sims et al., 1989, U.S. EPA, 1983, U.S.EPA, 1986).
15.2 In Situ TechniquesIn situ treatment techniques addressed include (1) soil
vacuum extraction, (2) bioremediation, (3) immobilization,and (4) mobilization.
15.2.1 Soil Vacuum Extraction (SVE)Referred to as soil vacuum extraction (SVE), forced air
venting, or in situ air stripping, this technique involves extrac-tion of air and contaminants from unsaturated soil. In contrastto a static equilibrium soil system where evaporation of achemical is equal to the condensation of the chemical (Figure15-1), with SVE, clean air is injected or passively flows intothe unsaturated zone. Volatile chemicals then partition fromsoil water into soil air, with relative partitioning based on theair/water partition coefficient (Kh) or Henry’s Law constant(Figure 15-2) and the vapor-laden air is removed using vacuumextraction wells.
Typically, components of SVE consist of vacuum extrac-tion wells (Figure 15-3), air inlet wells, and vapor monitoringwells distributed across a contaminated site, and a blower(s)to control air flow. Extraction wells may be placed verticallyor horizontally, although vertical alignment is typical fordeeper contamination zones and for residues in radial flow
215
Table 15-1. SITE Demonstration Program Participants with Technoiogies Appiicabie to Remediation of Contaminated SoilsApplicabie Wastea
Deveioper Technology Inorganic Organic
American Combustion Technologies, Inc.Norcross, GA
American Toxic Disposal Inc.Waukegan, IL
AWD Technologies, Inc.Burbank, CA
Biotrol, Inc.Chaska, MN
CF Systems CorporationWaltham, MA
Chemfix Technologies, Inc.Metairie, LA
Chemical Waste Management, Inc.Oakbrook, IL
Dehydro-Tech CorporationEast Hanover, NJ
Ecova CorporationRedmond, WA
EPOC Water, Inc.Fresno, CA
Exxon Chemicals, Inc./RioLinda Chemical Co.Long Beach, CA
GeoSafe CorporationKirkland, WA
HAZCON, Inc.Brookshire, TX
Horsehead Resources Development Co., Inc.Monaca, PA
international Waste Technologies/Gee-Con, Inc.Wichita, KS
MoTec, Inc.Austin, Tx
Ogden Environmental ServicesSan Diego, CA
Ozonics Recycling Corp.Boca Raton, FL
Resources Conservation Co.Bellevue, WA
Retech, Inc.Ukiah, CA
S.M.W. Seiko, Inc.Redwood City, CA
Shirco Infrared Systems, Inc.
Silicate Technology Corp.Scottsdale, AZ
Soliditech, Inc.Houston, TX
Solvent Services, Inc.San Jose, CA
Terra Vac, Inc.San Juan, PR
Toxic Treatments (USA) Inc.San Francisco, CA
Wastach, Inc.Oak Ridge, TN
Pyreton oxygen burner
Vapor extraction system
integrated vapor extractionand steam vacuum stripping
Soil washing system
Solvent extraction
Solidification/stabilization
X*TRAX” low temperaturethermal resorption
Carver-Greenfield processfor extraction of oily waste
in situ biological treatment
Leaching and micro filtration
Chemical oxidation/cyanidedestruction
in situ vitrification
Solidification/stabilization
Flame (slagging) reactor
in situ solidification/stabilization
Liquid/solid contact digestion
Circulating fluidized bedcombustor
Soil washing, catalytic/ozoneoxidation
Solvent extraction (BEST)
Plasma reactor
in situ solidificationlstabilization
infrared thermal destruction
Solidification/stabilizationwith silicate compounds
SoIidification/stabilization
Steam injection and vacuumextraction (SIVE)
in situ vacuum extraction
in situ steam/air stripping
Solidification/stabilization
NA
Volatile
NA
Metals
NA
Heavy metals
NA
NA
NA
Specific forheavy metals
Cyanide
Non-specific
Heavy metals
Heavy metals
Non-specific
NA
NA
Cyanide
NA
Metals
Metals
NA
Metals, cyanide,ammonia
Metals
NA
NA
NA
Non-specificradioactive
Non-specific
Volatile and semivolatileorganics includng PCBs,PAHs, PCPs, some pesticides
Volatile organic compounds
High molecular weight organics
PCBs, volatile, and semivolatileorganic compounds, petroleumbyproducts
High molecular weight organics
Volatile and semivoiatileorganics, PCBs
PCBs, dioxin, oil-solubleorganics
Chlorinated solvents, non-chlorinatad organic compounds
NA
NA
Non-specific
Not an inhibitor
NA
PCBs, other non-specificorganic compounds
Halogenated and non-halogenated organiccompounds, pesticides
Halogenatad and non-halogenated organiccompounds
Semivolatiles, pesticides, PCBsPCP, dioxin
Specific for high molecularweight organics
Non-specific
Semivolatile organiccompounds
Non-specific
High molecular weight organics
Non-specific
Volatile and semivolatileorganic compounds
Volatile and semivolatileorganic compounds
Volatile organic compoundsand hydrocarbons
Non-specific
a NA = non applicable Source: U.S. EPA, 1989f
216
Table 15-2. SITE Emerging Technology Program Participants with Technologies Applicable to Remediation of Contaminated Soils
Applicable Waste
Developer Technology Inorganic Organic
Babcock & Wilcox Co. Cyclone combustor Non-specific Non-specificAlliance, OH
Battelle Memorial Institute, In situ electroacoustic Specific for heavy NAColumbus Division decontamination metalsColumbus, OH
Enviro-Sciences, Inc. Low energy solvent NA PCBs, other non-specificRandolph, NJ extraction organic compounds
Harmon Environmental Services, Inc. Soil washing NA Heavy organic compounds(formerly Envirite Field Services, Inc.)Auburn, AL
IT Corporation Batch steam distillation/ Non-specific Non-specificKnoxville, TN metal extraction
Western Research Institute Contained recovery of oily NA Coal tar derivatives,Laramie, WY wastes (CROW) petroleum byproducts
NA = non applicable
Source: U.S. EPA, 1989f
patterns (Hutzler, 1990). Schematics of a gas extraction welland a gas monitoring well are presented in Figures 15-4 and15-5, respectively.
Important system variables that may affect the perfor-mance of SVE include properties of the chemical, such asvapor pressure and volatilization, and properties of the site,such as soil moisture content, soil texture, and distribution ofcontaminants. Vapor pressure is important when a chemicaloccurs in a pure phase in the subsurface. Vapor pressuresabove 14 mm Hg at 20°C are desirable for application of SVE.Vapor pressure values for selected subsurface contaminantsare given in Table 15-5. When chemicals are distributed in thewater phase in the soil, the Henry’s Law constant is important,and a dimensionless Henry’s constant above 0.01 (mg/L/mg/L) desirable for use of SVE. Table 15-6 gives Henry’s Lawconstants for a set of selected organic chemicals where theapplication of SVE would be appropriate.
Since movement of volatile organic chemicals (VOCs) isgenerally 10,000 times faster in a gas phase than in a waterphase, VOC removal is expected to be enhanced by decreas-ing soil moisture. However, when soil is very dry, which mayoccur when dry air is drawn through soil, VOCs may adsorbdirectly onto mineral surfaces, where the magnitude of sorp-tion is increased and consequently volatilization is decreased(Figure 15-6). Henry’s Law constant is not appropriate underthese conditions, since partitioning is between air and soilphases only. When moisture is added to soil, the effect isreversible. The moisture content at which a decrease in vapordensity becomes apparent is often termed the critical moisturecontent and generally is equivalent to approximately a mono-layer of water molecules coating the soil particles (Spencer etal., 1969, 1973). The effect of soil water content on dieldrinvapor pressure is illustrated in Figure 15-7. Johnson andSterrett (1988) noted that dichloropropane concentrations werecorrelated with ambient air moisture during the use of SVE ata site in Benson, Arizona.
If contaminated soil contains immiscible fluids in theform of oils, (e.g., petroleum hydrocarbons), the four-com-
partment system discussed previously is operative (water, air,oil, and soil as discussed in Chapter 14). In this system,chemical volatility will be affected by the chemical vaporpressure and mole fraction within the immiscible oil fluid, andgoverned by Raoult’s Law:
[15-1]
where Pa= vapor pressure of solvent over solution (mm Hg),Xa = mole fraction of solvent in solution, and Pa° = vaporpressure of pure solvent (mm Hg).
For contamination by hydrocarbons with multiple com-ponents, volatilization will proceed such that lower molecularweight chemicals will volatilize before higher molecular weightcompounds. Through this process of weathering of the waste/soil mixture, SVE extraction efficiency is observed to de-crease to less than 10 percent when the fraction of gasolineremaining is approximately 40 percent (Figures 15-8 and 15-9) (Johnson, 1989). Therefore, measuring general parameterssuch as total hydrocarbons is not sufficient to indicate theremoval efficiency of individual constituents.
Soil texture has been evaluated as it influences air perme-ability (DiGiulio et al., 1990). In less permeable media, suchas glacial till and clayey soils, secondary permeability orporosity (fractures) will dominate air flow. There will be rapidremoval of VOCs in fractures and slow removal in the soilmatrix. In more permeable media, such as sands, sandy loams,and loamy sands, SVE is appropriate (see Figures 15-10 and15-11). Pneumatic pump tests in the field are recommendedfor site-specific evaluation of SVE application.
Due to release of VOCs from the soil matrix, whenextraction wells are temporarily turned off, concentrations ofVOC increase in soil air (referred to as “VOC rebound ef-fect”), with an equilibrium concentration that is determinedby Henry’s Law constant. When blowers are turned on, anincrease in the concentration of extracted vapor from the soilwill be observed. Diffusive release from subsurface stratigra-
217
Table 15-3. NATO/CCMS Projects for the Remediation of Contaminated Soils
Treatable Contaminants
TreatmentSpecfic
Aliphatic Aromatic Halogenated Heavy Petroleum ContaminantsOrganization/site
Treatment Status ofHydrocarbons Hydrocarbons Hydrocarbons Metals Fuels, Oil Treated Location Technology
BiologicalEnhanced aerobic restoration Jet fuel In-situ Experimental
U.S. Air Force, BattelleEglin Air Force Base, FLUnited States
Microbial treatment Polycyclic aromatic On-site, DemonstrationFormer gas works hydrocarbons, in-situFredensborg, Denmark phenols, cyanides
Chemical/PhysicalK-PEG process
U.S. EnvironmentalProtection AgencyWide Beach, NY United States
High pressure soil washingScrap metal & copper refineryBerlin, Federal Republic of Germany
High pressure soil washing and oxidationGoldbeck Haus, HamburgFederal Republic of Germany
Soil vapor extractionU.S. Environmental Protection AgencyVerona Well FieldBattle Creek, Ml, United States
Stabilization/SolidificationIn-situ vitrification
Parsons Chemical SiteMichigan, United States
ElectrokineticElectro-reclamation
LoppersumThe Netherlands
PCBs, dioxin On-site, Demonstratedmobile
Lead, PAHCs On-site Commercialmobile
Phenol, kresol In-situ Demonstration
Halogenated and In-situ Demonstratedaromatic hydrocarbons
Mercury
Arsenic
In-situ Experimental
In-situ Commercial
ThermalThermal resorption and destruction
(radiation heating)Chlorobenzenes, On-site Experimental
Dekonta GmbH, HamburgChlorophenols,
Federal Republic of GermanyHexachlorocyclohexane,
dioxins, furans
Source: U.S. EPA, 1989d
I
Table 15-4. Selected Remediation Techniques Possibly Suitable for Cleanup of Contaminated Soils
Remediation Techniques
Type of Treatment Treatment Function Possible Residuals/ Possible Applications Possible LimitationsTechnology Category Transformation Products
Phyaical/ChemicalTreatmentLow Temperature In-tank Separation
Thermal Stripping In situ(including radiofrequency heating)
Soil Washing
Soil Flushing
In-tank
In situ
Separation;volumereduction
Off-gas; spent carbon orash from afterburner;processed soil; hazardousemissions from in situapplications
Extracted materials;water/flushing agentmix
Separation; Extracted materials;volume water/washing agentreduction mix
Soil Vacuum Extraction In situ Separation(SVE) Prepared bed
Glycolate In-tank DetoxificationDechlorination In situ
Volatile organics andvolatile toxic metals
Water/reagent mix;reaction products
Compounds of low watervolubility and high volatility
Organics and inorganic;most suitable for soilscontaminated with only a fewspecific chemicals
Organics and inorganic;most suitable for soilscontaminated with only afew specific chemicals
Volatile organics and toxicmetals; may be enhancedby the use of steam
Dehalogenation of aromatichalide compounds
Limited to organics with Henry’s Lawconstant greater than 3.0 x 10 -3 atm-m -3/mole and boiling points less than 800°;more effective for soils with low contentsof organic matter and moisture
Unfavorable contaminant separationcoefficients; less effective with complexmixtures of waste types and variationin waste imposition; unfavorable soilcharacteristics include:- high humiccontent, soil/solvent reactions, highsilt and clay content, and clay soilscontaining semivolatiles; unfavorablewashing fluid characteristics include:difficult recovery of solvent or surfactant,poor treatability of washing fluid,reduction of soil permeability, andhigh toxicity of washing fluid
Unfavorable contaminant separationcoefficients: less effective with complexmixtures of waste types and variationin waste composition; unfavorable soilcharacteristics include: variable soilconditions, high organic mattercontent, soil/solvent reactions, highsilt and clay content, and clay soilscontaining semivolatiles; unfavorableflushing fluid characteristics include:difficult recovery of solvent or surfactant,poor treatability of washing fluid,reduction of soil permeability, andhigh toxicity of washing fluid; requirescontainment of Ieachate and ground waterto prevent off-site groundwatercontamination
Soil heterogeneity (e.g., permeability,texture); not applicable to saturatedmaterials or miscible compounds
Heat and excess reagent required forsoils with greater than 20%. moistureand contaminant concentrations greaterthan 5%, and that contain competingreactive metals (e.g., aluminum)
(continued)
Table 15-4. (Continued)
Remediation Techniques
Type of Treatment Treatment Function Possible Residuals/ Possible Applications Possible LimitationsTechnology Category Transformation Products
Neutralization In situPrepared bedIn-tank
Detoxification;immobilization
Detoxification
Detoxification
Separation;volume reduction;immobilization
Detoxification
Separation;immobilization
Separation;immobilization
Volumereduction;detoxification
Volumereduction:detoxification
Volumereduction;detoxification
Precipitated salts
Oxidized reaction products
Reaction products
Precipitated metals
Reduced reaction products
Processed soil
Processed soil
Off-gases (possibly acidicand with incomplete com-bustion products); treatedmaterials with residual metals;fly ash: scrubber water
Off-gases (possibly acidic andwith incomplete combustionproducts); treated materials withresidual metals; fly ash;scrubber water
Waste acids and alkaliesto reduce reactivity andcorrosiveness
Compatibility of waste andtreatment chemical to preventformation of more toxic orhazardous compounds
Oxidation In situPrepared bedIn-tank
Prepared bed
Cyanides andoxidizable organics
Possible explosive reactions;production of more toxic orhazardous products; non-selective
Dioxins: nitrated wastes Inability of light to penetratesoil
Photolysis
Precipitation In situPrepared bedIn-tank
In situPrepared bedIn-tank
In situPrepared bed
Metals; certain anions Unfavorable effects on soil permeability;long-term stability unknown
Reduction Chromium, silver, andmercury
Possible explosive reactions;production of more toxic orhazardous products; non-selective
Carbon Adsorption Organic wasteswastes with high molecularweight and boiling point andlow volubility and polarity
Metal contaminants
Long-term stability unknown
Ion Exchange
Thermal TreatmentFluidized Bed
In situPrepared bed
Selectivity/competitionlimitations; pH requirements
In-tank Halogenated and non-halogenated organics;inorganic cyanides
High maintenance requirements;waste size and homogeneityrequirements; applicableto wastes with low sodium andmetal contents
Infrared In-tank Halogenated and non-halogenated organics;inorganic cyanides
Limited particle sizes, so mayrequire size reduction equipment
pyrolysis In-tank Nonvolatile char and ash (metals, Wastes not conducivesalts, and particulates) to conventional incineration;
wastes with volatile metals orrecoverable residues
Small capacity
Table 15-4. (Continued)
Remediation Techniques
Type of Treatment Treatment FunctionTechnology Category
Possible Residuals/ Possible Applications Possible LimitationsTransformation Products
High particulate emissions;limited particle sizes, so mayrequire size reductionequipment
Rotary Kiln In-tank Volumereduction;detoxification
Off-gases (possibly acidic andwith incomplete combustionproducts); treated materials withresidual metals; fly ash; scrubberwater
Halogenatad and non-halogenated organics;inorganic cyanides
Biological TreatmentAerobic In-tank; Detoxification
bioremediation prepared bed;In situ
Ability to control environmentalfactors conducive to biodegradation;formation of more toxic or hazardoustransformation products; prepared bed:areal limitations due to cost of bedpreparation
May require long treatmentperiods; incomplete treatment,possibly requiring aerobic conditionsto complete degradation process
Hazardous volatile emissions;incomplete and possiblyhazardous degradation products;Ieachates in soil systems
Biodegradable organicwastes
Anaerobicbioremediation
In-tank; Detoxificationprepared bedIn situ
Hazardous volatile emissions;carbon dioxide, methane andother gases; incomplete andpossibly hazardous degradationproducts; Ieachates in soilsystems
Certain halogenatedorganics
Survival and activity of organisms inintroduced environment (affected byenvironmental factors and competitionwith native species)
Maintenance of optimum environmentalconditions for biological activity;requires large amounts of compost materialsmixed with only about 10% wastes
Activity and stability of introduced enzymesin natural systems
In-tank; Detoxificationprepared bed;in situ
Hazardous volatile emissions;incomplete and possiblyhazardous degradation products;Ieachates in soil systems
Many biodegradable organicwastes
Biological Seeding
In-tank; Detoxificationprepared bed
Hazardous volatile emissions;incomplete and possiblyhazardous degradation products;Ieachates and runoff water
Biodegradable organicwastes
Comporting
Certain biodegradableorganic wastes
Enzyme addition In-tank; Detoxificationprepared bed;In situ
Hazardous volatile emissions;incomplete and possiblyhazardous degradation products;Ieachates in soil systems
Fixation/EncapsulationCement solidification In-tank
In situLeachates; hazardous volatileemissions; solidified wastematerials
Metal cations, latex andsolid plastic wastes
Incompatible with large amounts ofdissolved sulfate salts or metallicanions such as arsenates or borates;setting time increased by presenceof organic matter, lignite, silt, orclay: requires complete anduniform mixing of soils and reagents;long term stability unknown;may reduce soil permeability andincrease run-off
Storage;immobilization
(continued)
Table 15-4. (Continued)
Remediation Techniques
Type of Treatment Treatment Function Possible Residuals/ Possible Applications Possible LimitationsTechnology Category Transformation Products
Fixation/EncapsulationClassification/ In-tank Storage; Leachates; hazardous volatile
vitification In situ immobilization emissions; glassifie orvitrified waste materials; aqueousscrub solution
Lime Solidification In-tank(Silicate) In situ
Thermoplastic In-tankMicroencapsulation In situ
Storage; Leachates, hazardous volatileimmobilization emissions; solidified waste
materials
Volumereduction;storage;immobilization
Leachates, hazardous volatileemissions; encapsulated wastematerials
Inorganic and someorganics in liquidsand contaminatedsoils
Metals, waste oils, andsolvents
Complex, difficult totreat hazardous wastes
Long-term stability unknown; highenergy requirements, especially withhigh soil water contents and lowpermeability; electrical shortingcaused by buried metal drums;possible underground fire from combustiblematerials; volatile metals nearsurface may volatilize; site may requirerun-off controls
Long-term stability unknown;incompatible with borates, sulfates,carbohydrates; requires complete anduniform mixing of soils and reagents;may reduce soil permeability andincrease run-off
Wastes not treatable: wastes withhigh water content; stronglyoxidizing contaminants: anhydrousinorganic salts, tetraborates, ironand aluminum salts, and organicswith low molecular weights andhigh vapor pressures; long-termstability unknown; requires completeand uniform mixing of soils and reagents;may reduce soil permeability andincrease run-off
Sources: Rich and Cherry, 1987; U.S. EPA, 1987, 1988b, 1989b
Unventilated Soil
Figure 15-1. Static soil system in equilibrium (modified from Valsaraj and Thibodeaux, 1988).
Ventilated Soil
Figure 15-2. Enhancement of volatilization through application of soil vacuum extraction (modified from Valsaraj and Thibodeaux,1988).
223
Figure 15-3. Typical components of a soil vacuum extraction system (from Hutzler et al., 1990).
Figure 15-4. Schematic of a gas extraction weii used in a soil vacuum extraction system (from DiGiulio, 1989).
224
Figure 15-5. Schematic of a gas monitoring well used in a soil vacuum extraction system (from DiGiulio, 1989).
Table 15-5. Comparative Vapor Pressures and Henry’sConstants
Table 15-6. Oxygen Supply
phy of less permeability will cause the slow continual releaseof chemicals into the soil-gas phase (Figure 15-12).
Design considerations that affect SVE include extractionwell spacing and extraction well depth. As permeability de-creases, well spacing decreases; typical well spacings of 10 mto 30 m are common (Figure 15-13). Also, air circulation
generally is not significant below the screened interval forextraction wells. Where contamination is deep and permeabil-ity is high throughout the soil profile, the slotted (screened)interval should be extended to the maximum depth possible tomaximize treatment, rather than slotted fully vertically (Fig-ure 15-14).
A promising application of SVE is for enhancement ofbiodegradation of volatile and semivolatile chemicals in soils.SVE provides air to the vadose zone, and thus carries oxygenthat can be used as the terminal electron acceptor (TEA) bysoil microorganisms to biodegrade chemicals (Figure 15-15).Air has a much greater potential than water for deliveringoxygen to soil on a weight-to-weight and volume-to-volumebasis (Table 15-6). Oxygen provided by air is more easilydelivered since the fluid is less viscous than water higheroxygen concentrations in air also provide a large driving forcefor diffusion of oxygen into less permeable areas within a soilformation (Miller, 1990).
Hinchee (1989) and Hinchee and Downey (1990) suc-cessfully applied SVE to enhance biodegradation of petro-leum hydrocarbons in JP-4 jet fuel at Hill Air Force Base,Ogden, Utah, by increasing subsurface oxygen concentra-tions. Soil moisture was found to be a sensitive variableaffecting biodegradation, with increased soil moisture (from20 percent to 75 percent field capacity) related to increasedbiodegradation (Figure 15-16). Monitoring carbon dioxideand oxygen concentrations, as well as estimating the mass ofVOC biodegraded, is recommended for evaluating potentialenhancement of biodegradation using SVE.
225
Figure 15-6. Volatile organic carbon adsorption to soil surface in the presence of two soil moisture regimes (from Reible, 1989).
Figure 15-7. Effect of soil water content on dieldrin vaporpressure (modified from Spencer and Claith,1989).
Figure 15-8. Volatilization of different hydrocarbon compo-nents in gasoline (from Johnson, 1989).
226
Figure 16-9. Soii vacuum extraction efficiency based on total hydrocarbon vapors (from Johnson, 1989).
Vacuum Extraction Application Related to Soil Texture
Stop and Evaluate Carefully Clay, Silty Clay, Silty Clay LoamSome Difficulty Sandy Clay, Clay LoamGood Sandy Clay Loam, Silt LoamVery Good Sand, Loamy Sand, Sandy Loam,
Loam
Figure 15-10. Soil texture trilinear diagram (modified from DiGiulio, 1989).
227
Figure 15-11. Effect of geologic stratification on velocity and resuitant dominant flow process (from Keely et ai., in press).
Figure 15-12. Chemicai concentration in the vapor phase versus time for a soil vacuum extraction system where the system istemporarily discontinued, then restarted (from DiGiulio et al., 1990).
228
Figure 15-13. Effect of weil spacing on totai solute massremaining in soii with vacuum extraction time(from Wilson et al., 1989).
Figure 15-14. Effect of weil depth on total soiute mass remain-ing in soil with vacuum extraction time (fromWilson et aI., 1989).
In situ vacuum extraction has been demonstrated in Mas-sachusetts as part of the Superfund SITE program (U.S. EPA,1989c, and 1989f), in Michigan and Puerto Rico (U.S. EPA,1988a), and at several other locations in the United States(U.S. EPA, 1990).
15.2.2 BioremediationBiotic reactions in the subsurface, including definitions
and mechanisms, are addressed in Chapter 13. Wilson (1983)identified biological processes, including microbial degrada-tion, as important mechanisms for attenuating contaminantsduring transport through the vadose zone to the ground water.In situ soil remedial measures using biological processes canreduce or eliminate continuing or potential ground-water con-tamination, thus reducing the need for extensive ground-watermonitoring and treatment requirements (Wilson, 1981, 1982,1983).
In situ biological remediation of soils contaminated withorganic chemicals is also an alternative treatment technologyfor achieving a permanent cleanup remedy at hazardous waste
Figure 15-15. Aerobic biodegradation using hydrocarbon as theeiectron donor and oxygen as the electronacceptor (from Hinchee, 1989).
sites, as encouraged by the EPA for implementation of theSuperfund Amendments and Reauthorization Act (SARA) of1986. Information for design of in situ bioremediation isbased on land treatment systems designed for hazardous wastes(Overcash and Pal, 1979; U.S. EPA, 1983, 1986). These landtreatment designs provide a significant information base fordesigning in situ soil remediation systems.
In situ bioremediation involves the use of naturally oc-curring microorganisms (in contrast to genetically engineeredmicroorganisms) to degrade and/or detoxify hazardous con-stituents in the soil at a contaminated site to protect publichealth and the environment. Bioremediation techniques forcontaminated soils have been addressed at several scientificmeetings and conferences (AWMA/U.S. EPA, 1989, 1990;HMCRI, 1989; Omenn, 1988; Lewandowski et al., 1989; U.S.EPA, 1989a). The use of bioremediation techniques in con-junction with chemical and physical treatment processes, i.e.,the use of a “treatment train,” is an effective means forcomprehensive site-specific remediation (Ross et al., 1988).
Components of soil bioremediation systems generallyinclude (1) delivery systems, such as injection nozzles, plows,and irrigation systems, deliver water, nutrients; oxygen; or-ganic matter, specialized microorganisms, and/or other amend-ments, as required; and (2) run-on and run-off controls formoisture control and waste containment (U.S. EPA, 1984,1990).
Four approaches are generally used for in situ biologicaltreatment: (1) enhancement of biochemical mechanisms fordetoxifying or degrading chemicals, (2) augmentation withexogenous acclimated or specialized microorganisms origi-nating from uncontaminated or contaminated environments,(3) application of cell-free enzymes, and (4) vegetative uptake(U.S. EPA, 1990). Enhancement of biochemical mechanismsmay involve (1) control of soil factors such as contaminantconcentrations that do not severely inhibit microbial activity,soil moisture, pH, nutrients, and temperature in order tooptimize microbial activity; (2) addition of organic amend-ments to stimulate cooxidation or cometabolism; (3) controlof soil oxygen by moisture control to accomplish aerobic oranaerobic biodegradation; and (4) addition of colloidal gasaphrons (microscopic bubbles of gas) to increase the concen-tration of terminal electron acceptors (oxygen) in the soil and
229
Figure 15-16. Enhancement of bioremediation of gasoline components using vacuum extraction of soil amended with nutrients andmoisture (from Hinchse, 1989).
thereby enhance aerobic biodegradation (Keck et al., 1989;Sims et al., 1989; U.S. EPA 1989a, 1990).
The soil contaminant concentration effect on rate andextent of detoxification of contaminated soil is illustrated inFigure 15-17. Detoxification of the soil/waste mixture wasmeasured using the MicrotoxTM bioassay. The MicrotoxTM
assay is an aqueous general toxicity assay that measures thereduction in light output produced by a suspension of marineluminescent bacteria in response to an environmental sample(Bulich, 1979). Bioluminescence of the test organism dependson a complex chain of biochemical reactions. Chemical inhi-bition of any of the biochemical reactions causes a reductionin bacterial luminescence. Therefore, the MicrotoxTM testconsiders the physiological effect of a toxicant, not just mor-tality. Matthews and Bulich (1984) described a method ofusing the MicrotoxTM assay to predict the land treatability ofhazardous organic wastes. Matthews and Hastings (1987)developed a method using the MicrotoxTM assay to determinean appropriate range of waste application loading for soil-based waste treatment systems. Symons and Sims (1988)utilized the assay to assess the detoxification of a complexpetroleum waste in a soil environment. The assay also wasincluded as a recommended bioassay in the EPA’s PermitGuidance Manual on Hazardous Waste Lund Treatment Dem-onstrations (1986). Comparison of results presented in Figure15-18 for a clay loam soil with results for the sandy loam soilshown in Figure 15-17 indicates that detoxification rate andextent for a waste is a function of soil type. Implications formanagement of heavily contaminated soils, therefore, mayinclude the incorporation of additional treatment medium(uncontaminated soil) into contaminated soil. This incorpora- Figure 15-17. Detoxification of sandy loam soil measured by
Microtox TM assay (from Symonsand Sims, 1988.
230
tion will decrease the concentration of contaminant to levelsthat are less inhibitory to soil microbial processes, therebyrendering treatment more rapidly and completely.
Acclimation of a soil to the presence of a waste is shownfor a fossil fuel-contaminated soil in Table 15-7. The accli-mated soil was exposed to the fossil fuel waste for one yearbefore a repeat application of the waste. Results presented inTable 15-7 indicate that a higher percentage of waste wastreated in the acclimated soil. Treatment also occurred morerapidly compared to treatment in unacclimated soil. Manage-ment of contaminated soil, therefore, may include the additionof lightly contaminated, preexposed, soil to more heavilycontaminated and/or newly contaminated soil to increase therate and extent of treatment..
The effect of soil moisture on treatment of contaminatedsoil is illustrated in Table 15-8 and Figure 15-19. The chemi-cal degradation rates given in Table 15-8 indicate more rapiddegradation at a soil moisture content of 60 to 80 percent offield capacity than at a soil moisture content of 20 to 40percent. MicrotoxTM assay results for evaluation of the changesin toxicity of four wastes (Figure 15-19), two petroleum andtwo wood preserving, incubated in relatively dry sandy loamsoil (20 to 40 percent field capacity) over a period of 360 daysindicated little change in toxicity for three wastes and anincrease in toxicity for one waste. Comparison of resultsobtained for lower soil moisture (Figure 15-19) with those forhigher soil moisture (Figure 15-17) for petroleum wastes insandy loam soil indicate the importance of soil moisture ininfluencing microbial activity in waste/soil mixtures.
The effect of temperature on apparent loss of polycyclicaromatic hydrocarbon (PAHs) compounds in a sandy loamsoil is summarized in Table 15-9 (Coover and Sims, 1987).Temperature has an important effect on the fate and behaviorof PAHs and, therefore, has implications for seasonal effectson the rate of biological remediation of soil contaminated withthese chemicals. Microbial ecologists have identified rangesof critical environmental conditions that affect aerobic activ-ity of soil microorganisms (Table 15-10). Many of theseconditions are controllable and can be modified to enhanceactivity (Huddleston et al., 1986; Paul and Clark, 1989;Rochkind et al., 1986; Sims et al., 1984).
The application of cooxidation processes for the biodeg-radation of high molecular weight PAHs present in oil (NAPL)phases in soil has been investigated by Keck et al. (1989). Incertain cases, PAH degradation may be limited by the rate ofprimary substrate (oil) degradation, which is limited by therate of supply of terminal electron acceptors (oxygen) to thesubsurface. In the study by Keck et al., aerobic conditionswere not sufficient to stimulate biodegradation of high mo-lecular weight PAHs present as a synthetic mixture in soil;however, when PAHs were present in an oily matrix in thesoil, and the soil was supplied with oxygen, PAHs wereobserved to exhibit faster degradation kinetics (Figure 15-20).Results indicated that oxygen may limit the rate and extent ofbiodegradation in soil environments, in addition to saturatedenvironments. Supplying oxygen to the contaminated vadosezone may allow biodegradation of oily components of soil
wastes, which may result in simultaneous cooxidation ofresistant PAHs present in the oily waste.
There is also increasing evidence that some halogenatedcompounds may be degraded under methanogenic conditionsthrough a process of reductive dehalogenation (Suflita et al.,1982, 1983, 1984). Kobayashi and Rittmann (1982) deter-mined that the redox potential of the environment must bebelow 0.35 V for significant reductive dechlorination to oc-cur. Reductive reactions may be catalyzed by both abiotic andbiochemical means in anaerobic environments.
Oxygen may be consumed faster than it can be replacedby diffusion from the atmosphere, and the soil may becomeanaerobic. Clay content of soil and the presence of organicmatter also may affect oxygen content in soil. Clayey soilstend to retain a higher moisture content, which restricts oxy-gen diffusion, while organic matter may increase microbialactivity and deplete available oxygen. Loss of oxygen as ametabolic electron acceptor induces a change in the activityand composition of the soil microbial population. Obligateanaerobic organisms and facultative anaerobic organisms,which use oxygen when it is present or switch to alternativeelectron acceptors such as nitrate or sulfate in the absence ofoxygen, become the dominant populations. Additional infor-mation concerning in situ anaerobic bioremediation can befound in the document, Handbook on In Situ Treatment ofhazardous Waste-Contaminated Soils (U.S. EPA, 1990).
The use of plants for stimulating microbial activity in soilresults in increased biodegradation of target organic chemi-cals in contrast to the possibility of vegetative accumulationof chemicals for harvesting and removal from a site. Thismethod is currently being investigated by Walton and Ander-son (1990) and Aprill and Sims (1990). In soils with lowlevels of contamination, plant roots may stimulate the biodeg-radation of toxic chemicals by providing exudates that serveas carbon and energy substrates for soil microorganisms. Theeffects of prairie grasses on soil PAH concentrations aresummarized i n Table 15-11. For soil with initial concentra-tions of PAHs of approximately 10 to 50 mg/kg, the presenceof vegetation in the soil (prairie grasses) resulted in a statisti-cally significant reduction in PAHs, compared with nonveg-etated soil.
The environmental factors presented in Table 15-10, aswell as waste and soil/site characteristics identified in Chapter14, interact to affect microbial activity at a specific contami-nated site. Computer modeling techniques are useful designand evaluation tools to describe these interactions and theireffects on bioremediation treatment techniques for organicconstituents in a specific situation.
Measurement of physical abiotic loss mechanisms (dis-cussed in Chapter 13) and partitioning of organic substancesinto air and soil phases (discussed in Chapters 10 and 11)should be used in degradation studies to ensure that generatedinformation is related to disappearance mechanisms of theconstituents in the soil system (Abbott and Sims, 1989;Armstrong and Konrad, 1974). This type of information isneeded to more accurately evaluate and select treatment tech-niques. For example, for organophosphorus pesticides, sorp-
231
Table 15-7 Acclimation of Soil to Complex Foesil Fuel Waste
Unacclimated Soil Acclimated Soil
PNA Initial Soil Reduction in Soil ConcentrationConstituent
Reduction inConcentration 40 days (%) after First Reapplication 22 days (%)(mg/kg-dry Wt) of Waste (after 168 days
incubation at initial level)(mglkg-dry wt)
Naphthalene 38 38 100Phenanthrene 30 30 83Anthracene 38 58 38 99Fluoranthene 154 51 159 8 2Pyrene 177 4 7 160 86Benz(a)anthracene 30 42 40Chrysene
702 7 25 33 61
Benz(a)pyrene 10 40 12 50
Source: Sims, 1986
Figure 15-18.
30 60 90 120 150 180
Time (days)
Detoxification of clay loam soil measured byMicrotoxTM assay (from Symons and Sims, 1988).
Tabie 15-8. Effect of Soil Moisture on PNA Degradation(Reauits Presented as Half-Life in Days)
Phenan- Fluoran-Moisture Anthracene threne thene
20-40% field capacity 43 61 559
60-80% field capacity 37 54 231
Source: Sims, 1986
tion-catalyzed hydrolysis of ester linkages is known to be animportant influence on soil degradation. An understanding ofabiotic reactions as influenced by sorption and pH of thesystem may allow the design of a more effective remediationstrategy. If abiotic controls are not used, the disappearance ofchemicals may be attributed solely to biological activity,though biological activity may not play the major role in thedegradation of the chemical. Therefore, knowledge of thereaction mechanism is directly related to efficiency and effec-tiveness in remediation strategy design and remediation tech-nique selection.
15.2.3 ImmobilizationOne way to predict and control the rate of transport of a
constituent through a subsurface system is to describe itsmobility (or relative immobility) by predicting its retardation(Borden and Bedient, 1987; Mahmood and Sims, 1986). Re-tardation describes the relative velocity of the constituentcompared to the rate of movement of water through thesubsurface (see Section 10.3 for more information). Retarda-tion in unsaturated soil can be represented as:
R = 1 + (rKd/q) [15-2]
where p = soil bulk density; Kd = soil/water partition coeffi-cient, which describes the partitioning between the soil solid
For additional detail about this process, see Section 10.3.
This information can be used to evaluate treatment tech-niques for a contaminated soil system (e.g., techniques tomodify the soil/water partition coefficient, such as control ofsoil moisture, changes in bulk density, or addition of amend-ments to the soil). Constituents can be “captured” or containedwithin the system by using these techniques, thus allowingtime for degradation at the site or for engineering implementa-tion and performance of other remediation treatment tech-niques, such as soil washing (Sims et al., 1989).
Linear retardation of chemicals in the vapor phase isdiscussed in Chapter 11. Variables in the equations given in
232
Figure 15-19. MicrotoxTM assay results for various materials (from Aprill et al., 1990).
that chapter can be used by professionals involved in treat-ment technique selection to determine site conditions (pb, Kp, that may influence the effectiveness of specifictreatment technologies. For additional detail about these pro-cesses, see Section 11.2.2.
Constituents in in situ and prepared bed treatment sys-tems are generally immobilized through sorption, ion ex-change, and/or precipitation reactions. These techniques reducethe rate of contaminant release from the soil environment sothat concentrations along exposure pathways are held withinacceptable limits. The effects of moisture and distributioncoefficient, Kd, on immobilization are illustrated in Figure 15-21. Results indicate that for chemicals with Kd, values less than10, management of soil moisture is important with regard toimmobilizing chemicals; for chemicals with Kd values greaterthan 10, management of soil moisture is less important. Ap-preaches for controlling soil moisture include run-on and run-off controls, temporary capping or covering, and irrigationscheduling.
The cation exchange capacity (CEC) of soil also can beevaluated with regard to organic as well as metal immobiliza-tion. Positively charged organic chemicals and metals willgenerally readily attach to soil materials with negativelycharged functional groups and negatively charged clay par-ticles. Addition of clays, synthetic resins, and zeolites willincrease the CEC of soils and increase immobilization ofchemicals sensitive to CEC characteristics of a soil (Sims etal., 1984; U.S. EPA, 1984). For inorganic chemicals that arenegatively charged in soil systems and can exist in several
oxidation states (e.g., chromium, selenium, and arsenic), im-mobilization, as well as the toxic form of the chemical, maypotentially be controlled by managing the redox and pH of thesoil system. Management of redox and pH may be short-termor long-term, depending upon the goals of site management(e.g., temporary immobilization while delivery and recoverysystems are designed and implemented, followed by soilflushing with aqueous or surfactant solutions for removal andrecovery of the contaminants) (Sims et al., 1984; U.S. EPA,1984).
Solidification and stabilization are additional immobili-zation techniques that are applicable to in situ and preparedbed systems. These techniques are designed to accomplishone or more of the following: (1) production of a solid from aliquid or semisolid waste, (2) reduction of contaminant volu-bility, and/or (3) a decrease in the exposed surface area acrosswhich transfer may occur. Solidification may involve encap-sulation of fine waste particles (microencapsulation) or largeblocks of waste (macroencapsulation).
Stabilization refers to the process of reducing the hazard-ous potential of waste materials by converting contaminantsinto their least soluble, mobile, or toxic form (U.S. EPA,1990). A milestone publication providing additional detail onthis technique is the Handbook for Stabilization Solidifica-tion of Hazardous Wastes (Cullinane et al., 1986).
Systems for delivering reagents to the contaminated areainclude (1) injection systems; (2) soil surface applicators; and/or (3) delivery and application of electrical energy for melting
233
Figure 15-20. Persistence in soil of PAH compounds as afunction of number of fused benzene rings (fromKeck et sI., 1989).
Table 15-11.
soils and rocks that contain hazardous materials. Equipmentrequired for preparing, mixing, and applying reagents de-pends upon the reagent process, and depth of contamination(U.S. EPA, 1990).
Important parameters identified by Truett et al. (1983) forsolidification and stabilization of hazardous wastes include(1) reagent viscosity; (2) permeability of soils; (3) porosity ofwaste materials and soil; (4) distribution of waste in surround-ing material (rocks, soils, etc.); and (5) rate of reaction. Themost significant challenge in applying solidification/stabiliza-tion treatment in situ is achieving uniform mixing of addedchemical agent(s) with the contaminated soils (U.S. EPA,1990).
Design factors involve delivery and mixing systems toobtain complete and uniform distribution of added reagentthroughout the contaminated soil (U.S. EPA, 1990).
In situ solidification/stabilization was applied and evalu-ated under the Superfund SITE program for treatment ofpolychlorinatcd biphenyl (PCB) contaminated soils (U.S. EPA,1989e). Eight additional application sites have been summa-rized in U.S. EPA (1990).
Apparent Disappearance of PAH Compounds in Vegetated and Unvegetated Soii Using the TissumizerTM ExtractionMethod
235
Figure 15-21.
1
Sorption of chemicals to soil as functions of soil.moisture content and partition coefficient Kd(Sims et al., 1986).
15.2.4 Contaminant MobilizationMobilization of organic ant/or inorganic contaminants
from soil may be accomplished using soil flushing and recov-ery and treatment of the elutriate (U.S. EPA, 1984, 1990).Flushing solutions generally include water, acidic and basicsolutions, surfactants, and solvents. The solutions partition acontaminant into the liquid phase through the volume ofadded liquid or by decreasing the distribution coefficientbetween the soil and the flushing phase (Sims et al., 1984;Raghavan et al., 1990). A schematic of a soil flushing systemis shown in Figure 15-22 (U.S EPA, 1984). Componentsconsist of (1) the flushing solution, and (2) delivery andrecovery systems, which may include injection and recoverywells, equipment for surface applications, and holding tanksfor storing elutriate for reapplication (U.S. EPA, 1984, 1990).
Variables affecting application of the technique include(1) concentration and volume of contamination; (2) distribu-tion coefficients of waste constituents; (3) interactions offlushing solutions with soil; and (4) suitability of site forinstallation of wells, drains, etc., for delivery and recovery.Design factors include sizing the delivery and recovery sys-tems to ensure complete recovery of elutriate. Problems withrespect to flushing of bulk fluids, or NAPLs, from soil sys-tems are due to the following characteristics of bulk fluids: (1)low water volubility, (2) high interracial tension, and (3) poorrelative permeability. Relative permeability is defined as:
[15-3]
where M = mobility ratio; Kd = fluid permeability (water); KO
= oil permeability; Ud = viscosity of fluid (water); and UO =viscosity of oil. Strategies for flushing of bulk liquids fromsoil generally involve control of one or more of the variablesaffecting the mobility ratio through adding chemicals to de-crease mobility of water or increase mobility of oil (e.g.,adding surfactants or steam to decrease UO or adding polymersto increase Ud).
Use of soil flushing in a treatment train with bioremedia-tion has been evaluated by Dworkin et al. (1988) and by Kuhnand Piontek (1989) for wood preserving contaminated sites.Flushing using surfactant/polymer combinations was used toremove high concentrations of PAH compounds; residual lowconcentrations were treated using biological processes.
The effect of adding a solvent on the partitioning ofPAHs between soil and solution (solvent) phases of a soilsystem is illustrated in Figure 15-23. When methanol wasused as the solvent in a soil system to flush PAHs from a soil,the resultant concentration of the PAHs in the solution phasewas severaI orders of magnitude higher than the concentrationof the PAHs in water.
15.3 Prepared Bed ReactorsIn a prepared bed system, the contaminated soil may be
either (1) physically moved from its original site to a newlyprepared area, which has been designed to enhance remedia-tion and/or to prevent transport of contaminants from the site;or (2) removed from the site to a storage area while theoriginal location is prepared for use, then returned to the bed,where the treatment is accomplished. Preparation of the bedmay include placement of a clay or plastic liner to retardtransport of contaminants from the site or addition of uncon-taminated soil to provide additional treatment medium.
Possible prepared bed reactor technologies are identifiedin Table 15-4 and are evaluated for function as well asapplication and limitations. Treatment of contaminants with aprepared bed may be based on the techniques previouslyidentified and described for in situ treatment.
An example of the use of a prepared bed reactor for soilremediation was described by Lynch and Genes (1989). Pre-pared bed treatment of creosote-contamimted soils from ashallow, unlined surface impoundment was demonstrated at adisposal facility for a wood-preserving operation in Minne-sota. The contaminated soils contained creosote constituentsconsisting primarily of PAHs at concentrations ranging from1,000 to 10,000 ppm. Prior to implementation of the full-scaletreatment operation, bench-scale and pilot-scale studies simu-lating proposed full-scale conditions were conducted to defineoperation and design parameters. Over a 4-month period, 62to 80 percent removal of total PAHs was achieved in all testplots and laboratory reactors. Two-ring PAH compoundswere reduced by 80 to 90 percent, 3-ring PAHs by 82 to 93percent, and 4+-ring PAHs by 21 to 60 percent.
The full-scale system involved preparation of a treatmentarea within the confines of the existing impoundment. A linedwaste pile for temporary storage of the sludge and contami-
236
Figure 15-22. Schematic of soil flushing and recycle system (U.S. EPA, 1990).
nated soil from the impoundment was constructed. All stand-ing water from the impoundment was removed, and the slud-ges were excavated and segregated for subsequent free oilrecovery. Three to five feet of “visibly” contaminated soil wasexcavated and stored in the lined waste pile. The bottom of theimpoundment was stabilized as a base for the treatment area.The treatment area was constructed by installation of a poly-ethylene liner, a leachate collection system, 4 feet of cleanbackfill, and addition of manure to achieve a carbon: nitrogenratio of 50:1. A sump for collection of storm water andleachate and a center pivot irrigation system also were in-stalled. The lined treatment area was required because naturalsoils at the site were highly permeable. A cap also was neededfor residual contaminants left in place below the lincr. Con-taminated soil was periodically applied to the treatment facil-ity and rototilled into the treatment soil. Soil moisture wasmaintained near field capacity with the irrigation systcm.During the first year of operation, greater than 95 percentreductions in concentration were obtained for 2- and 3-ringPAHs. Greater than 70 percent of 4- and 5-ring PAH com-pounds were degraded during the first year. Comparison ofhalf-lives of PNAs in the full-scale facility were in the lowend of the range of half-lives reported for the test plot units.
Only two PNA compounds were detected in drain tile watersamples, at concentrations near analytical detection limits.
Prepared bed treatment of a Texas oilfield site withstorage pit backfill soils contaminated with styrene, still bot-tom tars, and chlorinated hydrocarbon solvents was demon-strated on a pilot scale (St. John and Sikes, 1988). Theremediation efforts included biological, chemical, and physi-cal treatment strategies. The pilot-scale, solid-phase biologi-cal treatment facility consisted of a plastic film greenhouseenclosure, a lined soil treatment bed with an underdrain, anoverhead spray system for distributing water, nutrients, andinocula, an organic vapor control system consisting of acti-vated carbon absorbers, and a fermentation vessel for prepar-ing microbial inoculum or treating contaminated leachatefrom the backfill soils. Soils were excavated from the con-taminated area and transferred to the treatment facility. Aver-age concentrations of volatite organic compounds (VOCs)were reduced by more than 99 percent during the 94-dayperiod of operation of the facility; most of the removal wasattributed to air stripping. Biodegradation of semivolatilecompounds reduced average concentrations by 89 percentduring the treatment period.
237
Figure 15-23. PAH adsorption isotherms with methanol and clay loam soil (from Mahmood and Sims, 1985).
15.4 ReferencesAbbott, C. and R.C. Sims. 1989. Use of Bioassays to Monitor
Polycyclic Aromatic Hydrocarbon Contamination in Soil.In: Superfund ’89, Hazardous Materials Control ResearchInstitute, Silver Spring, MD, pp. 23-26.
Aprill, W. and R.C. Sims. 1990. Evaluation of the Use ofPrairie Grasses for Stimulating Polycyclic Aromatic Hy-drocarbon Treatment in Soil. Chemosphere 20:253-265.
Aprill, W., R.C, Sims, J.L. Sims, and J.E. Matthews. 1990.Assessing Detoxification and Degradation of Wood Pre-serving and Petroleum Wastes in Contaminated Soil.Waste Manag. and Res. 8:45-65.
Armstrong, D.E. and J.G. Konrad. 1974. Nonbiological Deg-radation of Pesticides. In: Pesticides in Soil and Water,W. D. Guenzi (ed.), Soil Science Society of America,Madison, WI, Chapter 7.
AWMA/EPA. 1989. Proceedings of the International Sympo-sium on Hazardous Waste Treatment: Biosystems forPollution Control (Cincinnati, OH). Air and Waste Man-agement Association, Pittsburgh.
AWMA/EPA. 1990. Proceedings of the International Sympo-sium on Hazardous Waste Treatment: Treatment ofContaminated Soils (Cincinnati, OH). Air and WasteManagement Association, Pittsburgh.
Borden, R.C., P.B. Bedient, M.D. Lee, C.H. Ward, and J.T.Wilson. 1986. Transport of Dissolved Hydrocarbons In-fluenced by Reaeration and Oxygen Limited Biodegrada-tion. II. Field Application. Water Resources Research22:1983-1990.
Toxicology, L.L. Markings and R.A. Kimerle, (eds.),American Society for Testing and Materials, Philadel-phia, PA, pp. 98-106.
Coovcr, M.P. and R.C. Sims. 1987. The Effect of Tempera-ture on Polycyclic Aromatic Hydrocarbon Persistence inan Unacclimated Soil. Hazardous Waste and HazardousMaterials 4:69-82.
Cullinane, Jr., M.J., L.W. Jones, and P.G. Malone. 1986.Handbook for Stabilization/ Solidification of HazardousWastes. EPA/540/2-86/OOl (NTIS PB87-116745/REB).
DiGiulio, D.C. 1989. U.S. EPA Robert S. Kerr EnvironmentalResearch Laboratory, Ada, OK, personal communica-tion.
DiGiulio, D.C., J.S. Cho, R.R. DuPont, and M.W. Kemblowski.1990. Conducting Field Tests for Evaluation of SoilVacuum Extraction Application. In: Proc. 4th Nat. Out-door Action Conf. on Aquifer Restoration, Ground WaterMonitoring and Geophysical Methods, National WaterWell Association, Dublin, OH, pp. 587-601.
Dupont, R. R., R.C. Sims, J.L. Sims, and D.L. Sorensen. 1988.In Situ Biological Treatment of Hazardous Waste-Con-taminated Soils. In: Biotreatment Systems, D. L. Wise(ed.), CRC Press, Boca Raton, FL, Chapter 2.
Dworkin, D., D.J. Messinger, and R.M. Shapot. 1988. In SituFlushing and Bioreclamation Technologies at a Creosote-Based Wood Treatment Plant. In: Proc. 5th NationalConference on Hazardous Wastes and Hazardous Materi-als, Hazardous Materials Control Research Institute, Sil-ver Spring, MD, pp. 67-78.
Bulich, A.A. 1979. Use of Luminescent Bacteria for Deter- Hinchee, R. 1989. Enhanced Biodegradation Through Soilmining Toxicity in Aquatic Environments. In: Aquatic Venting. Presented at Workshop on Soil Vacuum Extrac-
238
tion held at US. EPA Robert S. Kerr EnvironmentalResearch Laboratory, Ada, OK, April 27-28 (DominicDiGiulio, Technical Coordinator).
Hinchee R., and D. Downey. 1990. In Situ Enhanced Biodeg-radation of Petroleum Distillates in the Vadose Zone. In:Proceedings of the International Symposium on Hazard-ous Waste Treatment: Treatment of Contaminated Soils(Cincinnati, OH). Air and Waste Management Associa-tion, Pittsburgh, PA.
Hazardous Materials Control Research Institute (HMCRI).1989. Symposium on Use of Genetically Altered orAdapted Organisms in the Treatment of Hazardous Wastes.HMCRI, Silver Spring, MD.
Huddleston, R.L., C.A. Bleckman, and J.R. Wolfe. 1986.Land Treatment Biological Degradation Processes. In:Land Treatment A Hazardous Waste Management Alter-native, R.C. Loehr and J.F. Malina (eds.), Water Re-sources Symposium No. 13, University of Texas Press,Austin, TX, pp. 41-61.
Hutzler, N.J., B.E. Murphy, and J.S. Gierkc. 1990. State ofTechnology Review: Soil Vapor Extraction Systems. EPA/600/2-89/024 (NTIS PB89-195 184).
Johnson, R.L. 1989. Soil Vacuum Extraction: Laboratory andPhysical Model Studies. Presented at Workshop on SoilVacuum Extraction held at U.S. EPA Robert S. KerrEnvironmental Research Laboratory, Ada, OK, April 27-28 (Dominic DiGiulio, Technical Coordinator),
Johnson, J.J. and R.J. Sterrett. 1988. Analysis of In Situ AirStripping Data. In: Proc. 5th National Conference onHazardous Waste and Hazardous Materials, HazardousMaterials Control Research Institute, Silver Spring MD,pp. 451-455.
Keck, J., R.C. Sims, M. Coorer, K. Park, and B. Symons.1989. Evidence of CoOxidation of Polynuclear AromaticHydrocarbons in Soil. Water Research 23(12):1467-9476.
Keely, J.W., D.C. Bouchard, M.R. Scalf, and C.G. Enfield.Practical Limits to Pump and Treat Technology for Aqui-fer Remediation. Ground Water Monitoring Review. Inpress.
Kobayashi, H. and B.E, Rittmann. 1982. Microbial Removalof Hazardous Organic Compounds. Environ. Sci. Tcchnol.16:170A-183A.
Kuhn, R.C. and K.R. Piontek. 1989. A Site-Specific In SituTreatment Process Development Program for a WoodPreserving Site. Presented at Seminar on Oily WasteFate, Transport, Site Characterization, and Remediation,Denver, CO, May 17-18 (John Matthews, Technical Co-ordinator, Robert S. Kerr Environmental Research Labo-ratory, Ada, OK).
Lehr, J. 1988. The Misunderstood World of UnsaturatedFlow. Ground Water Monitoring Review 8(2):4-6.
Lewandowski, G., P. Armenante, and B. Baltzis (eds.). 1989.Biotechnology Applications in Hazardous Waste Treat-ment. Engineering Foundation, New York, NY.
Lynch, J, and B.R. Genes. 1989, Land Treatment of Hydro-carbon Contaminated Soils. In: Petroleum ContaminatedSoils, Vol. I: Remediation Techniques, EnvironmentalFate, and Risk Assessment, P.T. Kostecki and E.J.Calabrese (eds.), Lewis Publishers, Chelsea, MI, pp. 163-174.
Mahmood, R.J. and R.C. Sims. 1985. Enhanced Motility ofPolynuclear Aromatic Compounds in Soil Systems. In:Proc. 1985 Environmental Engineering Specialty AnnualConference (Boston, MA), American Society of CivilEngineers, pp. 128-135.
Mahmood, R.J. and R.C. Sims. 1986. Mobility of Organics inLand Treatment Systems. Journal of Environmental En-gineering (ASCE) 112:236-245.
Matthews, J.E. and A.A. Bulich. 1984. A Toxicity ReductionTest System to Assist Predicting Land Treatability ofHazardous Wastes. In: Hazardous and Industrial SolidWaste Testing: Fourth Symposium, J.K. Petros, Jr., W.J.Lacy, and R.A. Conway (eds.), ASTM STP-886, Ameri-can Society of Testing and Materials, Philadelphia, PA,pp. 176-191.
Matthews, J.E. and L. Hastings. 1987. Evaluation of ToxicityTest Procedure for Screening Treatability Potential ofWaste i. Soil. Toxicity Assessment: International Quar-terly 2:265-21.
Miller, R. 1990. A Field Scale Investigation of EnhancedPetroleum Hydrocarbon Biodegradation in the VadoseZone Combining Soil Venting as an Oxygen Source withMoisture and Nutrient Addition. Ph.D. Dissertation, De-partment of Civil and Environmental Engineering, UtahState University, Logan, UT.
Omenn, G.S. (ed.). 1988. Environmental Biotechnology—Reducing Risks from Environmental Chemicals throughBiotechnology. Plenum Press, New York, NY, 505 pp.
Ovcrcash, M.R. and D. Pal. 1979. Design of Land TreatmentSystems for Industrial Wastes—Theory and Practice. AnnArbor Science, Ann Arbor, MI.
PACE. 1985. The Persistence of Polynuclear Aromatic Hy-drocarbons in Soil. PACE Report No. 85-2. PetroleumAssociation for Conservation of the Canadian Environ-ment, Ottawa, Ontario, Canada.
Paul, E.A. and F.E. Clark. 1989. Soil Microbiology andBiochemistry. Academic Press, San Diego, CA.
Raghavan, R., E. Coles, and D. Dietz. 1990. Cleaning Exca-vated Soil Using Extraction Agents: A State-of-the-ArtReview. EPA/600/2-89/034 (NTIS PB89-212757/AS).
239
Reible, D.D. 1989. Introduction to Physicochemical Processes Water Content, Temperature and Dieldrin Concentration.Influencing Enhanced Volatilization. Presented at Work- Soil Sci. Soc. Am. Proc. 33:509-511.shop on Soil Vacuum Extraction held at U.S. EPA RobertS. Kerr Environmental Research Laboratory, Ada, OK, Spencer, W .F., W.J. Farmer, and M.M. Cliath. 1973. PesticideApril 27-28 (Dominic DiGiulio, Technical Coordinator). Volatilization. Residue Reviews 49:1-47.
Rich, G. and K. Cherry. 1987. Hazardous Waste TreatmentTechnologies. Pudvan Publishing, Northbrook, IL, Chap-ter 7.
Rochkind, M.L., J.W. Blackbum, and G. Sayler. 1986. Micro-bial Decomposition of Chlorinated Aromatic Compounds.EPA/600/2-86/090 (NTIS PB87-116943/REB).
Ross, D., T.P. Marziarz, and A.L. Bourquin. 1988. Bioreme-diation of Hazardous Waste Sites in the USA: CaseHistories. In: Superfund ’88, Hazardous Materials Con-trol Research Institute, Silver Spring, MD, pp. 395-397.
Sims, R.C. 1982. Land Treatment of Polynuclear AromaticCompounds. Ph.D Dissertation, Departmcnt of Biologi-cal and Agricultural Engineering, North Carolina StateUniversity, Raleigh, NC.
Sims, R.C. 1986. Loading Rates and Frequencies for LandTreatment Systems. In: Land Treatment—A HazardousWaste Management Alternative, R.C. Loehr, and J.F.Malina (eds.), Water Resources Symposium No. 13, Uni-versity of Texas Press, Austin, TX, pp. 151-170.
Sims, R.C., and M.R. Overcash. 1983. Fate of PolynuclcarAromatic Compounds (PNAs) in Soil-Plant Systems. Resi-due Reviews 86:1-68.
Sims, R.C. and J.L. Sims. 1986. Cleanup of ContaminatedSoil. In: Utilization, Treatment, and Disposal of Waste onLand, K.W. Brown, B.L. Carlile, R.H. Miller, E.M.Turledge, and E.C.A. Runge (eds.), Soil Science Societyof America, Madison, WI, pp. 257-277.
Sims, R.C., D.L. Sorensen, J.L. Sims, J.E. McLean, R.Mahmood, and R.R. DuPont. 1984. Review of In-PlaceTreatment Technologies for Contaminated Surface Soils-Volume 2: Background Information for In-Situ Treat-ment. EPA/540/2-84-O03b (NTIS PB85-124899).
Sims, R.C., D. Sorensen, J.L. Sims, J. McLean, R.J Mahmood,R. Dupont, and J. Jurinak. 1986. Contaminated SurfaceSoils In-Place Treatment Techniques. Pollution Technol-ogy Review No. 132. Noyes Publications, Park Ridge,NJ, 536 pp.
Sims, J. L., R.C. Sims, and J.E. Matthews. 1989. Bioremedia-tion of Contaminated Soils. EPA/600/9-89/073 (NTISPB90-164047).
Spencer, W.F. and M.M. Cliath. 1969. Vapor Density ofDieldrin. Environ. Sci. Technol. 3:670-674.
Spencer, W.F., M.M. Cliath, and W.J. Farmer. 1969. VaporDensity of Soil-Applied Dieldrin as Related to Soil-
St. John, W. D., and D.J. Sikes. 1988. Complex IndustrialWaste Sites. In: Environmental Biotechnology-Reduc-ing Risks from Environmental Chemicals through Bio-technology, G.S. Omenn (ed.), Plenum Press, New York,NY, pp. 237-252.
Suflita, J. M., A. Horowitz, D.R. Shelton, and J.M. Tiedje.1982. Dehalogenatiom a Novel Pathway for the Anaero-bic Biodegradation of Haloaromatic Compounds. Sci-ence 218:ll15-1117.
Suflita, J.M., J.A. Robinson, and J.M. Tiedje. 1983. Kineticsof Microbial Dchalogenation of Haloaromatic Substratesin Methanogenic Environments. Appl. Environ. Microbiol.45:1466-1473.
Suflita, J. M., J. Stout, and J.M. Tiedje. 1984. Dechlorinationof (2,4,5-Trichlorophenoxy) Acetic Acid by AnaerobicMicroorganisms. Journal of Agricultural and Food Chem-istry 32:218-221.
Symons, B.D. and R.C. Sims. 1988. Detoxification of a Com-plex Hazardous Waste Using the MicrotoxTM Bioassay.Archives of Environmental Contamination and Toxicol-ogy 17:497-505.
Truett, J. B., R.L. Holbergcr, and K.W. Barrett, 1983. Feasibil-ity of In Situ Solidification/Stabilization of LandfilledHazardous Wastes. EPA/600/2-83/088 (NTIS PB83-261099).
U.S. Environmental Protection Agency (EPA). 1983. Hazard-ous Waste Land Treatment & EPA SW-874.
U.S. Environmental Protection Agency (EPA). 1984. Reviewof In-Place Treatment Techniques for Contaminated Sur-face Soils. EPA/540/2-84-O03a (NTIS PB85-124881).
U.S. Environmental Protection Agency (EPA). 1986. PermitGuidance Manual on Hazardous Waste Land TreatmentDemonstrations. EPA/530/SW-86-032 (NTIS PB86-229 184).
U.S. Environmental Protection Agency (EPA). 1987. A Com-pendium of Technologies Used in the Treatment of Haz-ardous Wastes. EPA/625/8-87/O14 (Available from Centerfor Environmental Research Information, Cincinnati, OH).
U.S. Environmental Protection Agency (EPA). 1988a. Cleanupof Releases from Petroleum USTS: Selected Technolo-gies. EPA/530/UST-88/OOl (NTIS PB88-241856).
U.S Environmental Protection Agency (EPA). 1988b. Tech-nology Screening Guide for Treatment of CERCLA Soilsand Sludges. EPA/540/2-88/O04 (NTIS PB89-132674/REB).
240
U.S. Environmental Protection Agency (EPA). 1989a. Biore-mediation of Hazardous Waste Sites Workshop: SpeakerSlide Copy and Supporting Information. EPA CERI-89-11 (NTIS PB89-169205/REB).
U.S. Environmental Protection Agency (EPA). 1989b. Cor-rective Action: Technologies and Applications. EPA/625/4-89/020 (Available from Center for EnvironmentalResearch Information, Cincinnati, OH).
U.S. Environmental protection Agency (EPA). 1989c. Dem-onstration Bulletin: In Situ Vacuum Extmct.ion, TerraVac, Inc. EPA/540/M5-89/O03 (NTIS PB90-126665/GAR).
U.S Environmental Protection Agency (EPA). 1989d. Dem-onstration of Remedial Action Technologies for Con-taminated Land and Groundwater. In: Procccdings andAppendices, Third International Conference, NATO Com-mittee on Challenges of Modern Society (CCMS)(Montreal, Canada), pp. v-vii.
U.S. Environmental Protection Agency (EPA). 1989e. Tech-nology Evaluation Report: SITE Program DemonstrationTest, International Waste Technologies, In Situ Stabiliza-tion/Solidification, Hialeah, Florida, Volume 1. EPA/540/5-89/004a (NTIS PB89-194161/AS).
U.S. Environmental Protection Agency (EPA). 1989f. TheSuperfund Innovative Technology Evaluation Program:
Technology Profiles. EPA/540/5-89/O13 (NTIS PB90-249756/A07).
U.S. Environmental Prelection Agency (EPA). 1990. Hand-book on In Situ Treatment of Hazardous Waste-Contami-nated Soils. EPA/540/2-90-O02 (NTIS PB90-155607).
Valsaraj, K.T. and L.J. Thibodcaux. 1988. Equilibrium Ad-sorption of Chemical Vapors on Surface Soils, Landfills,and Landfarms-A Review. J. Hazardous Materials 19:79-99.
Wahon, B.T., and T.A. Anderson. 1990. Microbial Degrada-tion of Trichlorocthylene in the Rhizosphere: PotentialApplication to Biological Remediation of Waste Sites.Appl. Environ. Microbiol. 56:1012-1016.
Wilson, L.G. 1981. Monitoring in the Vadose Zone: Part I,Storage Changes. Ground Water Monitoring Review1(3):32-41.
Wilson, L.G. 1982. Monitoring in the Vadose Zone: Part II.Ground Water Monitoring Review 2(2):31-42.
Wilson, L.G. 1983. Monitoring in the Vadose Zone: Part III.Ground Water Monitoring Review 3(l): 155-166.
Wilson, D.J., R.O., Mutch, Jr., and A.N. Clarke. 1989. Model-ing of Soil Vapor Stripping. Presented at Workshop onSoil Vacuum Extraction held at U.S. EPA Robert S. Kerr
28 (Dominic DiGiulio, Technical Coordinator).
241
I
Chapter 16Aquifer RestorationRonald C. Sims and Judith L. Sims
Currently, several remedial techniques are being used torestore contaminated ground water and aquifer material. Par-ticipants in the U.S. Environmental Protection Agency’s (EPA)SITE program that are testing technologies applicable tocontaminated ground water are listed in Tables 16-1 and 16-2(U.S. EPA, 1989a). Table 16-3 summarizes technologies ap-plicable to contaminated ground water currently being evalu-ated and demonstrated in the NATO/CCMS Pilot Study ondemonstration of remedial action technologies for contami-nated land and ground water (U.S. EPA, 1989b).
The pattern of contamination from a release of contamin-ants into the subsurface environment, such as would occurfrom an underground leaking storage tank containing non-aqueous phase liquids (NAPLs), is complex (Figure 16-1)(Palmer and Johnson, 1989; Wilson et al., 1989). As contami-nants move through the unsaturated zone, a portion is leftbehind, trapped by capillary forces. If the release containsvolatile contaminants, a plume of vapors forms in the soilatmosphere in the vadose zone. If the release contains NAPLsless dense than water (LNAPLs), they may flow by gravitydown to the water table and spread laterally. Ground watermoving through subsurface sediments contacts the release andthe more water-soluble components are dissolved into thewater phase. Therefore, three distinct regions of contaminantsare formed in the release: a plume of fumes in the soilatmosphere, a ground-water plume, and the region that con-tains the oily phase material that serves as the source area forboth plumes. This latter region may include both recoverablefree product (i.e., continuous phase material), and sorbed orcapillary-held material (i.e., residual saturation material). Ifthe release contains DNAPLs, these contaminants can pen-etrate to the bottom of an aquifer, forming pools in depres-sions.
This chapter discusses three techniques concerning aqui-fer restoration: (1) product removal, (2) pump-and-treat, and(3) biorestoration.
16.1 Product RemovalProduct removal generally consists of product character-
ization, product location, and product recovery. Product char-acterization refers to identifying the type of product (e.g.,petroleum, wood-preserving, or solvent) and associated indi-vidual chemicals (e.g., BXT, PAHs, TCE). Product locationsinclude characterizing the product mobility at the site (e.g.,
LNAPL following the water table, DNAPL following thebedrock). Knowledge of whether the product is an LNAPL ora DNAPL may help locate the product in the subsurface.
Physical rccovcry techniques to remove free productinclude (1) a single pump system producing a mixture ofhydrocarbon and water that must be separated, but requiringminimal equipment and drilling; (2) a two-pump, two-wellsystem utilizing one pump to produce a water table gradientand a second well to recover floating product; or (3) a singlewell with two pumps in which a lower pump produces agradient and an upper pump collects free product (Lee andWard 1986). Vacuum extraction of volatilizing contaminantsalso may be used to recover floating free product from aperched water table.
Pumping systems commonly used for recovery of LNAPLsarc shown in Figures 16-2 and 16-3. An aboveground oil/water separator generally is used to recover product for futureuse. Subsurface drains also have been used for recovery ofDNAPLs (Figure 16-4a). When only the oil recovery drainline(ORD) is used (Figure 16-4b), water truncates the flow ofproduct (DNAPL) due to the poor relative permeability of theproduct as described previously in the discussion of soilflushing. The water table depression drainline (WTDD) is anefficient method (see Figure 16-4c) to drag an oily productacross the subsurface by viscous forces and thereby create ahydraulic head of oil above the ORD; however, oil also entersthe WTDD, thereby creating the need for aboveground sepa-ration of product and water. When both ORD and WTDD areused (Figure 16-4d), subsurface separation of oil and water isachieved, thereby minimizing aboveground separation re-quirements. This system (Figure 16-4d) is also efficient sincethe permeability of oil is greatest in the oily contaminatedsubsurface, and the underground separation maintains waterflowing in the water compartment and oil flowing in the oilycompartment.
Caution should be exercised during product recovery ofLNAPL when an extraction well is used to control localgradients and collect free product in a cone of depression. Dueto capillary forces in the subsurface aquifer material, trappedresidual will constitute a continuous source of contaminationto ground water that will persist after product removal fromthe water table is completed.
243
Table 16-1. SITE Demonstration Program Participants with Technologies Applicable to Remediation of Contaminated GroundWater
Applicable Waste
Developer Technology Inorganic Organic
AWD Technologies, Inc.Burbank, CA(004)
Biotrol, Inc.Chaska, NM(003)
Integrated Vapor Extractionand Steam VacuumStripping
NA Volatile Organic Compounds
Bioiogical AqueousTreatment Systems
Can be applied to Nitrates Chlorinated andNonchlorinatedHydrocarbons
Readily BiodegradableOrganic Compounds
DETOX, inc.Dayton, OH(003)
Submerged Aerobic Fixed-Film Reactor
Metals inhibit process
E.I. Du Pont de Nemoursand Co./Oberiln Filter Co.Newark, DE(003)
Membrane Microfiltration Heavy Metals, Cyanide,Uranium
Non-specific
Ecova CorporationRedmond, WA(003)
In Situ Biological Treatment NA Chlorinated Solvents,Nonchlorinated OrganicCompounds
Exxon Chemicals, Inc./Rio Linda Chemical Co.Long Beach, CA(004)
ChemicalOxidation/OrganicsDestruction
NA Non-specific
Freeze Technologies Corp.Raleigh, NC(003)
Freezing Separation Non-specific Non-specific
Ozonics Recycling Corp.Boca Raton, FL(004)
Soil Washing, Catalytic/Ozone Oxidation
Cyanide Semivolatiles, Pesticides,PCBs, PCP, Dioxin
Silicate Technology Corp.Scottsdale, AZ(003)
Solidification/Stabilizationwith Silicate Compounds
Metals, Cyanide, Ammonia High Molecular WeightOrganics
Ultrox International, Inc.Santa Ana, CA(003)
Ultraviolet Radiation andOzone Treatmen!
NA Halogenated Hydrocarbons,Volatile Organic Compounds,Pesticides, PCBs
Volatile and SemivolatileZimpro/Passavant,Inc., Rothschild, WI(002)
NA = non applicable
U.S. EPA, 1989a
PACT®/Wet Air Oxidation NAOrganic Compounds
16.2 Pump-and-Treat RemediationBoth hydrogeologic information and contaminant infor-
mation are required for pump-and-treat remediation. Hydro-geologic information about ground-water flow includesgeological and hydraulic factors (described in Chapters 3 and4) as well as ground-water use/withdrawal factors.
concerning extraction wells include whether to use continu-ous pumping, pulsed pumping, or pumping combined withcontainment. While continuous pumping maintains an inwardhydraulic gradient, pulsed pumping allows maximum concen-trations to be pumped and requires only minimum volumes ofpumping. Containment (physical or hydraulic) limits theamount of uncontaminated water that requires treatment. In-jected water can contain nutrients or electron acceptors wherebioremediation is used, or can contain enhanced oil recoverymaterials (EOR) for NAPL contaminants, or can be reinfectedtreated water without nutrient or EORs (U.S. EPA, 1990).
This chapter discusses pump-and-treat systems in twocategories: (1) pumping systems, and (2) treatment systems.Pumping systems may be used for plume containment and
When pump-and-treat remediation is selected, a decisionneeds to be made about the use of wells or drains (U.S. EPA,1990). If the hydraulic conductivity is sufficiently high toallow flow to wells, then wells are recommended. For low-permeability material, drains may be required. Wells can becategorized as extraction, injection, or a combination. Injec-tion wells reduce cleanup time required by flushing chemicalsto the extraction wells. Design and management decisions
244
Table 16-2. SITE Emerging Technology Program Participants with Technologies Applicabie to Remediation of ContaminatedGround Water
Developer
Atomic Energy of Canada,Ltd. Chalk River, Ontario(EO1)
Bio-Recovery Systems, Inc.Las Cruces, NM(EO1)
Eiectro-Pure Systems, Inc.Amherst, NY(E02)
Energy and EnvironmentalEngineering, Inc.East Cambridge, MA(EO1)
University of Washington,Dept. of Civil EngineeringSeattle, WA(E02)
Wastewater Tech. CentreBurlington, Ontario(E02)
Chemical TreatmentUltrafiltration
Bioiogical Sorption
A/C ElectrocoagulationPhase Separated andRemoval
Laser StimulatedPhotochemical Oxidation
Adsorptive Filtration
Cross-Flow PervaporationSystem
Applicable Waste
inorganic Organic
Specific for Heavy Metals NA
Specific for Heavy Metals NA
Heavy Metals Petroleum Byproducts, Coal-TarDerivatives
NA Non-specific
Metals NA
NA Volatile Organic Compounds
. , —NA = non applicableSource: U.S. EPA, 1989a
.
Tabie 16-3. NATO/CCMS Projects for Remediation of Contaminated Ground Water
245
Figure 16-1. Regions of contamination in a typical release from an underground storage tank (from Wilson et al., 1989).
plume recovery for aboveground treatment (Figures 16-5 and16-6). Ground-water pumping systems utilize the principlethat ground water flows in response to a hydraulic gradient,i.e., a drop in hydraulic pressure created by the combinedeffects of elevation, fluid density, and gravity.
The migration of a plume away from its source area,which is related to hydraulic containment, often can be pre-vented by capturing the plume with a purge well. The wellmust pump hard enough to overcome regional flow in theaquifer. Hydrodynamic control of a contaminated ground-water plume is accomplished by manipulating the hydraulicgradient. Passive hydrodynamic controls, or interceptor sys-tems, function by gravity. Active hydrodynamic controls relyon injection and production wells to control the hydraulicgradient (Canter and Knox, 1985).
Physical containment techniques include installing barri-ers to ground-water flow (e.g., slurry walls (see Figure 16-7),grout curtains, sheet pilings, block displacement, and clayliners) or diverting divert uncontaminated surface water awayfrom waste sites or contaminated water away from clean areas(Ehrenfield and Bass, 1984). These containment systems alsomay provide for temporary containment while ground water isremoved and treated and aquifer material is decontaminated.
Contaminated ground water that is withdrawn from anaquifer can be treated by various methods, depending on thetype(s) of contamination. Treatment methods may include oneor more of the following: (1) physical processes, such asadsorption onto activated carbon or resins, ion exchange,reverse osmosis, filtration, or transfer to the gaseous phase byair stripping; (2) chemical processes, such as neutralization,coagulation, precipitation, oxidation, or reduction reactions,which involve inactivating or immobilizing contaminants withchemical agents; or (3) biological processes, using conven-tional wastewater treatment methods such as suspended growth(e.g., activated sludge, lagoons, waste stabilization ponds, andfluidized bed reactors) and freed film (e.g., trickling filtersand rotating biological contractors) processes (Thomas et al.,1987).
With pumping systems that are used to bring contami-nated ground water to the surface for treatment contaminantsare transported by advection (velocity) and dispersion. Watervelocity for pumping systems can be calculated using Darcy’sLaw; however, spatial variability in hydraulic conductivityresults in a corresponding distribution of flow velocities;therefore, contaminant removal and transport rates are distrib-uted. Chemical contaminants in ground water also may notmove at the same rate as the water due to subsurface pro-cesses, including sorption or retardation, ion-exchange, and
246
Figure 162. Product recovery using two pumps in one well—a probe scavenger pump and a water table depression pump (fromNyer, 1985).
chemical precipitation (described under soil immobilizationtechniques), and bioremediation (described under soil biore-mediation).
Monitoring water and soil cores within the plume whilepumping is occurring allows a determination of area of reme-diation and remediation rate. These results allow rationalmanagement of the remediation wellfield. Keely (1989) ex-plains that, using this approach, the flow rates of extractionwells that pump from relatively clean zones would be de-creased, while flow rates from extraction wells that pumpfrom highly contaminated zones should be increased. Also,Keely (1989) points out that the exclusive use of monitoringpoints downgradient from a plume does not assist in anunderstanding of plume dynamics during remediation, exceptto indicate “out-of-control” conditions when contaminants aredetected.
During the continuous operation of an extraction well-field, the level of contamination in water flowing through thesubsurface usually is decreased in a relatively short period of
time, after which a low-level residual concentration is presentin the extracted water (Figure 16-8). After the residual con-centration in water is attained, a pump-and-treat system isusually characterized by treatment of large volumes of slightlycontaminated water over a long period of time. In addition, ifremediation is terminated before removal of residual contami-nation at a site, the concentration of contaminant(s) in theaquifer water may increase due to slow release of contaminantresiduals relative to pumpage-induced water movement (Fig-ure 16-9) (Keely, 1989). Transport processes that cause thiscontaminant behavior in the subsurface include (1) diffusionof contaminants in low-permeability sediments; (2) hydrody-namic isolation, or dead spots, within the wellfield (3) des-orption of contaminants from sediment surfaces; and (4)liquid-liquid partitioning of immiscible contaminants (Keely,1989).
One promising innovation in the use of pump-and-treatremediation is pulsed pumping. Pulsed pumping of hydraulicsystems is the cycling of extraction or injection wells on andoff in active and resting phases (Figure 16-10) (Keely, 1989).
247
Figure 16-3. Product recovery using a watertable depression pump and a floating oil/water filter (from Nyer, 1985).
The resting phase allows chemical contaminants to movefrom low-permeability sediments, dead spots, sediment sur-faces, and immiscible fluids in the subsurface into the waterphase. The pumping phase removes the minimum volume ofwater at the maximum contaminant concentration. By periodi-cally cycling selected wells, stagnation zones may be broughtinto active flow paths and remediated.
When pulsed pumping systems are used, peripheral gra-dient control must be ensured to prevent offsite migration ofcontamination. If migration is slow, water would be rapidlyrecovered by the high flow velocities back toward extractionwell(s) during the pumping phase. If migration is rapid, thenadditional containment controls are necessary to prevent offsitemigration during the resting phase of pulsed pumping.
16.3 BiorestorationIn addition to the overviews presented by Thomas and
Ward (1989) and Lee et al. (1988), there are several milestone
publications on biological restoration of contaminated groundwaters. These publications include those that address (1)hydrogen peroxide as a supplemental source of oxygen (Hulinget al., 1990); (2) new approaches for site characterization,project design, and performance evaluation (Wilson et al.,1989); (3) methanotrophic destruction of chlorinated aliphaticchemicals (Roberts et al., 1989); and (4) modeling aspects(Rifai et al., 1988 and 1989).
Biological in situ treatment of subsurface contaminants inaquifers is usually accomplished by stimulating indigenoussubsurface microorganisms to degrade organic waste con-stituents (Thomas and Ward, 1989). The activity of microor-ganisms is stimulated by injection of inorganic nutrients and,if required, an appropriate electron acceptor, into aquifermaterials. Most biological in situ treatment techniques cur-rently used are variations of techniques developed by re-searchers at Suntech to remediate gasoline-contaminatedaquifers. The Suntech process received a patent titled Recla-mation of Hydrocarbon Contaminated Ground Waters
248
Figure 16-4. Recovery of dense nonaqueous phase liquid(DNAPL) from subsurface bedrock using drains toaccomplish underground separation of DNAPL(oil) and water (from Sale and Piontek, 1989).
(Raymond, 1974). The process, as described by Lee et al.(1988), involves circulating oxygen and nutrients through acontaminated aquifer using injection and production wells.Well placement depends on the area of contamination and theporosity of the formation, but they are usually no more than100 ft apart. The nutrient amendment consists of nitrogen,phosphorus, and other inorganic salts, as required, at concen-trations ranging from 0.005 to 0.02 percent by weight. Oxy-gen for use as an electron acceptor in microbial metabolism issupplied by sparging air into the ground water. If the growthrate of microorganisms is 0.02 g/L per day, the process isestimated to require approximately 6 months to achieve 90percent degradation of the hydrocarbons present. Cleanup isexpected to be most efficient for ground waters contaminatedwith less than 40 ppm of gasoline. After termination of theprocess, the numbers of microbial cells are expected to returnto background levels.
Figure 16-5. Cross-sectional view of pump-and-treat system(from U.S. EPA, 1985).
Figure 16-8. Plan view of pump-and-treat system (from U.S.EPA, 1985).
Another technique, which has not yet been fully demon-strated, is the addition of microorganisms with specific meta-bolic capabilities to a contaminated aquifer (Lee et al. 1988).Populations that are specialized in degrading specific com-pounds are selected by enrichment culturing, which involvesexposure of microorganisms to increasing concentrations of acontaminant or mixture of contaminants. The type of organ-ism (or group of organisms) that is selected, or acclimates, tothe contaminant, depends on the source of the inoculum, theconditions used for the enrichment, and the substrate. Ex-amples of changes that may occur during an acclimationperiod include (1) an increase in population of contaminantdegraders; (2) a mutation that codes for new metabolic capa-bilities; and/or (3) induction or derepression of enzymes re-
249
Figure 16-7. Schematic of the preparation of a slurry wall forphysical containment of contaminated groundwater or for diversions of clean water around acontaminated subsurface (from U.S. EPA, 1985).
Figure 16-8. Decrease in aquifer water concentration causedby pump-and-treat system where contaminantconcentration in pumped water reaches anirreducible level that is frequently above theregulated limit (from Keely, 1989).
sponsible for degradation of specific contaminants (Aelion etal., 1987).
However, inoculation of a specialized microbial popula-tion into the environment may not produce the desired degreeof degradation for a number of reasons (Goldstein et al., 1985;Lee et al., 1988; Suflita 1989). Possible causes that may limitthe success of inoculants include both abiotic and bioticfactors. Environmental factors, such as pH, temperature, sa-linity, and osmotic or hydrostatic pressure may act alone orcollectively to inhibit the survival of the microorganisms. Theconcentration of the specific organic constituent of concernmay be too low to support growth and activity. The environ-ment may contain substances or other organisms that are toxic
Figure 18-9. Following temporary termination of pumping,aquifer water concentration increases, orrebounds, due to the presence of contaminantresiduals (from Keely, 1989).
or inhibitory to the growth and activity of the inoculatedorganism(s). The inoculated organism(s) may utilize someother organic compound than the one it was selected tometabolize. In addition, adequate mixing and transport toensure contact of the organism with the specific organicconstituent of concern may be difficult to achieve in groundwater. Successful inoculation of organisms into simpler, morecontrollable environments (e.g., bioreactors such as wastewa-ter treatment plants) to accomplish degradation has beendemonstrated. However, effectiveness of inoculation into un-controlled and poorly accessible environments (e.g., the sub-surface) is much more difficult to achieve, demonstrate, andassess (Thomas and Ward, 1989).
In a contaminated aquifer, some regions will clean upfaster than others, and the most contaminated flow path willbe the last to be cleaned. If this flow path can be identified,then its properties can be used to determine how much effortand time are required to remediate the entire area. The timerequired to clean the most contaminated flow path can bedetermined using a modification of the relationship given byU.S. EPA (1989c), correcting for units:
Time required to clean most contaminated flow path=
where
Massc/Massoxygen represents the stoichiometric amount ofoxygen required to biodegrade (mineralize) contami-nant (hydrocarbon) present;
Massoxygen/Volumewater represents the concentration of oxy-gen in the ground water; and
Volumewater/Time represents the seepage velocity alongthe contaminant flow path.
250
Figure 16-10. Pulsed pumping removal of residual contaminant minimizes volume of water required for pumping and maximlzescontaminant concentration in pumped water (from Keely, 1989).
Generally, if the supply of mineral nutrients is adequate,the rate of bioremediation is directly related to the rate ofsupply of electron acceptor. As a result, the rate of remedia-tion is directly proportional to the concentration of electronacceptor in the injected water and the flow velocity of waterthrough the contaminated area.
When in situ bioremediation of a contaminant ground-water plume involves using methods to enhance the processdiscussed above, such as the addition of nutrients, additionaloxygen sources, or other electron acceptors, hydraulic con-trols might be required to minimize (i.e., contain) migration ofthe plume during the in situ treatment process (Thomas et al.1987; U.S. EPA 1989c). In general, hydraulic control systemsare less costly and time consuming to install than physicalcontainment structures such as slurry walls. Well systems alsoare more flexible, because pumping rates and well locationscan be altered as the system is operated over a period of time.Wells should be installed under the direction of a hydrogeolo-gist to ensure proper placement and operation.
With respect to biorestoration of aquifers, pumping-in-jection systems can be used to (1) create stagnation (no flow)zones at precise locations in a flow field, (2) create gradientbarriers to pollution migration, (3) control the trajectory of acontaminant plume, and (4) intercept the trajectory of a con-taminant plume (Schafer, 1984). The choice of a hydrauliccontrol method depends on geological characteristics, vari-ability of aquifer hydraulic conductivities, background veloci-ties, and sustainable pumping rates (Lee et al., 1988). Typicalpatterns of wells that are used to provide hydraulic controlsinclude (1) a pair of injection-production wells, (2) a line ofdowngradient pumping wells, (3) a pattern of injection-pro-duction wells around the boundary of a plume, and (4) the“double-cell” hydraulic containment system. The “double-cell” system utilizes an inner cell and an outer recirculationcell, with four cells along a line bisecting the plume in thedirection of flow (Wilson, 1984).
Well systems also serve as injection points to add materi-als used to enhance microbial activity into the aquifer and forcontrol of circulation through the contaminated portion. Thesystem usually includes injection and production wells andequipment for the addition and mixing of the nutrients (Lee etal., 1988). Figure 16-11 illustrates a typical system in whichmicrobial nutrients are mixed with ground water and circu-lated through the contaminated portion of the aquifer througha series of injection and recovery wells (Raymond et al., 1978;Thomas and Ward, 1989). Wells should be screened to ac-commodate seasonal fluctuations in the level of the watertable and air can be supplied through a system of diffusers.Some operational designs are closed loop in which the water
Figure 16-11. Typical sc