Content uploaded by Michael Barden
Author content
All content in this area was uploaded by Michael Barden on Apr 15, 2020
Content may be subject to copyright.
WSRC-TR-2003-00331
December 31, 2004
Multiple Lines of Evidence Used to Evaluate
Natural Attenuation and Enhanced Remediation
of Chlorinated Solvents
Prepared by:
Todd H. Wiedemeier, P.G., Michael J. Barden, and W. Zachary Dickson
www.thwa.com
Contibuting Author:
Dave Major, P.E., GeoSyntec, Inc.
WSRC-TR-2003-00331
December 31, 2004
This report is a work prepared for Westinghouse Savannah River Company and the United States
Department of Energy by T.H. Wiedemeier & Associates, LLC. In no event shall either Westinghouse
Savannah River Company, the United States Government, or T.H. Wiedemeier & Associates, LLC have
any responsibility or liability for any consequences of any use, misuse, inability to use, or reliance upon
the information contained herein, nor do any of these parties warrant or otherwise represent in any way
the accuracy, adequacy, efficacy, or applicability of the contents hereof.
WSRC-TR-2003-00331
December 31, 2004
-i-
TABLE OF CONTENTS
Page
SECTION 1 INTRODUCTION ..................................................................................................1-1
1.1 Lines of Evidence Used to Evaluate Natural Attenuation ............................................... 1-1
1.2 Data Required to Evaluate Natural Attenuation .............................................................. 1-4
SECTION 2 PRIMARY LINES OF EVIDENCE ....................................................................... 2-1
2.1 Contaminant Concentration Data..................................................................................... 2-1
2.1.1 Graphical Evaluation of Plume Behavior .......................................................2-2
2.1.2 Statistical Methods for Evaluating Plume Behavior ....................................... 2-6
2.1.2.1 Nature of Groundwater Concentration Data and Appropriate
Statistical Methods....................................................................... 2-8
2.1.2.2 Tests for Trend.............................................................................2-9
2.1.2.3 Tests for Differences Between Groups of Data ......................... 2-17
2.1.2.4 Using Statistical Results ............................................................2-19
2.2 Contaminant/Daughter Product and Geochemical Data................................................ 2-24
2.2.1 Soil/Sediment Data ....................................................................................... 2-24
2.2.1.1 Total Organic Carbon ................................................................2-25
2.2.1.2 Field Observation of Iron Minerals............................................ 2-25
2.2.2 NAPL Data....................................................................................................2-27
2.2.3 Groundwater Analytical Data ....................................................................... 2-27
2.2.3.1 Daughter Products...................................................................... 2-27
2.2.3.2 Chloride...................................................................................... 2-27
2.2.3.3 Dissolved Oxygen......................................................................2-28
2.2.3.4 Nitrate ........................................................................................ 2-29
2.2.3.5 Fe(II) .......................................................................................... 2-29
2.2.3.6 Sulfate and Sulfide.....................................................................2-29
2.2.3.7 Methane...................................................................................... 2-30
2.2.3.8 Ethene/Ethane ............................................................................ 2-30
2.2.3.9 Acetylene ................................................................................... 2-30
2.2.3.10 Oxidation-Reduction Potential................................................... 2-30
SECTION 3 SUPPLEMENTAL LINES OF EVIDENCE .......................................................... 3-1
3.1 Supplemental Groundwater Geochemical Data............................................................... 3-1
3.1.1 Mn(II)..............................................................................................................3-1
3.1.2 Carbon Dioxide ............................................................................................... 3-4
3.1.3 Alkalinity ........................................................................................................ 3-4
3.1.4 Dissolved Inorganic Carbon............................................................................ 3-4
3.1.5 Major and Minor Ions .....................................................................................3-5
3.1.6 Dissolved Hydrogen........................................................................................ 3-5
3.1.7 Dissolved Organic Carbon .............................................................................. 3-7
3.1.8 Volatile Fatty Acids (VFAs) ........................................................................... 3-8
WSRC-TR-2003-00331
December 31, 2004
-ii-
3.1.9 Stable Isotopes ................................................................................................ 3-9
3.2 Mineralogical Analyses ................................................................................................... 3-9
3.2.1 Petrography ................................................................................................... 3-10
3.2.2 Wet Chemistry Techniques........................................................................... 3-12
3.2.3 X-Ray Diffraction ......................................................................................... 3-13
3.2.4 X-Ray Fluorescence...................................................................................... 3-14
3.2.5 Electron Microprobe Analysis ...................................................................... 3-14
3.2.6 Mass Magnetic Susceptibility Analysis ........................................................ 3-15
3.3 Microbiological Methods............................................................................................... 3-16
3.3.1 Microcosm Studies........................................................................................ 3-16
3.3.2 Phospholipid Fatty Acids (PLFA)................................................................. 3-16
3.3.2.1 Biomass...................................................................................... 3-17
3.3.2.2 Community Structure................................................................. 3-17
3.3.2.3 Diversity..................................................................................... 3-17
3.3.2.4 Physiological Status................................................................... 3-17
3.3.2.5 Example PLFA Analysis............................................................ 3-19
3.3.3 Denaturing Gradient Gel Electrophoresis ..................................................... 3-23
SECTION 4 DEDUCING GEOCHEMICAL ENVIRONMENTS AND DEGRADATION
PATHWAYS ............................................................................................................................... 4-1
4.1 Type 1 Environment: Systems that are Anaerobic Due to Anthropogenic Carbon......... 4-2
4.2 Type 2 Environment: Systems that are Anaerobic Due to Naturally-Occurring
Carbon........................................................................................................................... 4-10
4.3 Type 3 Environment: Aerobic Systems Due to Lack of Fermentation Substrates ....... 4-10
4.4 Mixed Environments...................................................................................................... 4-11
SECTION 5 REFERENCES........................................................................................................ 5-1
WSRC-TR-2003-00331
December 31, 2004
-iii-
LIST OF TABLES
Table 1.1 Summary of the Lines of Evidence Used to Evaluate Natural Attenuation and
Enhanced Remediation ...............................................................................................1-3
Table 1.2 Aquifer and Soil/Sediment Data Required to Evaluate Natural Attenuation
and Enhanced Remediation......................................................................................... 1-5
Table 1.3 Groundwater Data Required to Evaluate Natural Attenuation and Enhanced
Remediation ................................................................................................................ 1-6
Table 2.1 Example Calculation of the Mann-Kendall Statistic for TCE Concentrations in a
Monitoring Well with Ten Sampling Events............................................................ 2-10
Table 2.2 Table of Null Probabilities for the Mann-Kendall Statistic, n = 4 through 20 ......... 2-12
Table 2.3 Example Calculations for the Wilcoxon Signed-Rank Test Comparing Groups of
Paired Data for A) Quarterly Concentration Data in a Single Monitoring Well for
Two Years (µg/L), and B) Concentrations in Several Monitoring Wells for Two
Years (µg/L).............................................................................................................. 2-20
Table 2.4 Critical Test Statistic Values for the Signed-Rank Statistic W+,
n = 4 through 20........................................................................................................ 2-21
Table 3.1 Supplemental Groundwater Data for Evaluating Natural Attenuation and
Enhanced Remediation ...............................................................................................3-2
Table 3.2 Range of Hydrogen Concentrations for a Given Terminal Electron-Accepting
Process ....................................................................................................................... 3-6
Table 3.3 Supplemental Soil/Sediment Data for Evaluating Natural Attenuation and Enhanced
Remediation .............................................................................................................. 3-11
Table 3.4 Description of PLFA Structural Groups ................................................................... 3-18
Table 3.5 Viable Microbial Biomass Expressed as Picomoles of PLFA per mL of Sample
and as Cells per mL of Sample, Fatty Acid Structural Groups as Percent of Total
PFLA, and Physiological Status Biomarkers as Mole Ratio ................................... 3-22
Table 4.1 General Characteristics of the “Type” Geochemical Environments .......................... 4-3
Table 4.2 Matrix Showing Some of the Potential Geochemical Environments
Encountered in the Terrestrial Subsurface and the Impact of These Environments
on the Fate of Chlorinated Ethenes.............................................................................4-6
WSRC-TR-2003-00331
December 31, 2004
-iv-
LIST OF FIGURES
Figure 2.1 Isopleth Maps of Total VOC Concentrations in Groundwater at the Depth of
Highest Contaminant Concentration......................................................................... 2-3
Figure 2.2 Sampling Locations for the Plots of Contaminant Concentration versus Time
and Distance Downgradient Presented in Figure 2.3................................................ 2-4
Figure 2.3 Plots of Contaminant Concentration versus Time and Distance
Downgradient ........................................................................................................... 2-5
Figure 2.4 Solute Plume Behavior Illustrated by Concentration Trends over Time
for Monitoring Points in the Vicinity of the Source, Mid-Plume, and the
Distal Portion of the Plume..................................................................................... 2-23
Figure 2.5 Photograph of Reduced and Oxidized Iron Minerals .............................................. 2-26
Figure 3.1 Biomass Content Presented as Total PLFAs........................................................... 3-20
Figure 3.2 Relative Percentages of PLFA Structural Groups. Table 3.4 Describes the
Various Structural Groups...................................................................................... 3-21
Figure 3.3 Microbial Physiological Stress Markers.................................................................. 3-24
Figure 4.1 Conceptual Model of Type 1 Environment for Chlorinated Solvent Plumes Due
to a PCE and TCE Release........................................................................................ 4-8
Figure 4.2 Conceptual Model of Type 3 Environment for Chlorinated Solvent Plume Due
to a PCE and TCE Release...................................................................................... 4-12
Figure 4.3 Conceptual Model of Type 3 Environment for Chlorinated Solvent Plume with
VC and 1,2-DCA.................................................................................................... 4-13
Figure 4.4 Conceptual Model of Mixed Environments with Type 1 Environment in the
Source Zone and Type 3 Environment in the Downgradient Portion of the
Plume ......................................................................................................................4-14
Figure 4.5 Conceptual Site Model of Mixed Type 3/Type 1/Type 3 Environments ................ 4-15
Figure 4.6 Conceptual Model of a Plume Discharging to Surface Water ................................ 4-16
Figure 4.7 Example of a Plume that Grades from a Type 1 Environment with
Dehalorespiration into a Type 1 Environment with Abiotic Degradation into
a Type 3 Environment ............................................................................................ 4-17
WSRC-TR-2003-00331
December 31, 2004
1-1
SECTION 1
INTRODUCTION
Multiple distinct and converging lines of evidence, in various forms have been used in recent
years to evaluate natural attenuation and bioremediation (e.g., National Research Council [NRC],
1993; Wiedemeier et al. 1995, 1996a, and 1999; United States Environmental Protection Agency
[USEPA], 1998 and 1999; and American Society for Testing and Materials [ASTM], 1998). By
extension, many of these lines of evidence are useful for evaluating the effectiveness of
enhanced remediation. The focus of this document is on the meaning and use of the lines of
evidence for determining the efficacy of natural attenuation and enhanced remediation of
chlorinated solvents.
1.1 LINES OF EVIDENCE USED TO EVALUATE NATURAL ATTENUATION
The “lines of evidence” used to evaluate natural attenuation are the different types of data that
provide evidence for contaminant degradation and allow for quantification of the processes of
natural attenuation. The most common lines of evidence include historical trends in contaminant
data showing plume behavior and/or loss of contaminant mass over time, analytical data showing
that geochemical conditions are suitable for degradation of the constituents of concern, and data
that support the occurrence of biological or abiotic degradation. The use of these different types
of data provide a weight of evidence for the actual or potential efficacy of natural attenuation.
The three lines of evidence commonly used to evaluate natural attenuation are:
1) Contaminant concentration data demonstrating a reduction in concentration and/or loss
of mass at the field scale;
2) Geochemical data providing an indication of whether conditions are favorable for
contaminant degradation (biological or abiotic) and a signature of those process; and
3) Laboratory data providing evidence of degradation.
Each of these lines of evidence include a variety of different parameters that will vary
depending upon the specific contaminants involved and the natural attenuation processes that are
important for those contaminants. Additionally, the “weight of evidence” needed to evaluate
natural attenuation at any given site typically will depend upon the site-specific understanding of
the behavior of the contaminants involved.
In the case of chlorinated solvents, the geochemical environment is fundamentally important
for controlling the biological and abiotic transformation and degradation processes that can
occur. Therefore, in this document the lines of evidence are grouped into “primary” and
WSRC-TR-2003-00331
December 31, 2004
1-2
“supplemental” categories that are summarized in Table 1.1. The primary lines of evidence are
essential to the evaluation of natural attenuation of chlorinated solvents. These include the
contaminant concentration data and the critical geochemical parameters in the first and second
lines of evidence that are needed to evaluate solute plume behavior and geochemical conditions.
The supplemental lines of evidence include additional geochemical parameters and
microbiological data in the second and third lines of evidence that can better elucidate
transformation and degradation processes.
The primary lines of evidence are used to evaluate if natural attenuation is occurring at a rate
sufficient to meet remedial objectives and to help determine which mechanisms of natural
attenuation dominate. The data included in the supplemental lines of evidence can be useful for
refining the understanding of a system. If the evaluation of natural attenuation shows that
monitored natural attenuation (MNA) will not be protective, the data collected for the evaluation
also can be used to evaluate and design alternate remedial measures. The actual combination of
lines of evidence used to evaluate a site should be determined on a site-specific basis but, at a
minimum, the data for the primary lines of evidence should be collected and analyzed to help
estimate the natural attenuation capacity of the system. As a site becomes better characterized,
the specific parameters needed to evaluate the efficacy of continuing natural attenuation or
enhanced remediation will likely change. In many cases a subset of the data required to evaluate
the primary lines of evidence may be sufficient to ensure that the chosen remedial approach
remains effective.
The primary lines of evidence include the minimum data that must be collected to evaluate
MNA. Section 2 describes the primary lines of evidence in more detail. Evaluation of these
lines of evidence includes an assessment of plume behavior and an evaluation of plume
geochemistry. One of the primary lines of evidence involves using historical contaminant data to
evaluate if the contaminant plume is shrinking, stable, or growing. Although this line of
evidence can be used to show that a contaminant plume is being attenuated, it does not
necessarily show that contaminant mass is being destroyed. If the solute plume is adequately
defined in three dimensions and a statistically sufficient historical data set is available, this line
of evidence may be all that is required to show that natural attenuation is protective of human
health and the environment, regardless of the operant mechanism(s). However, because of their
complexity, and because almost every solute plume of chlorinated compounds behaves
differently, it is usually necessary to determine the dominant mechanisms working to effect
natural attenuation. This is the purpose of collecting daughter product and geochemical data
which is useful for determining the relative importance of attenuation mechanisms and to
determine whether or not contaminant mass is being destroyed. Based on what is learned from
the primary lines of evidence it may be necessary to collect some of the data listed under the
supplemental lines of evidence.
WSRC-TR-2003-00331
December 31, 2004
1-3
Table 1.1
Summary of the Lines of Evidence Used to Evaluate Natural Attenuation
and Enhanced Remediation
Category Line of Evidence Data Requirements Applicability/Comments
Contaminant
concentration data
Historical contaminant data
from groundwater
monitoring
Older sites with good historical data.
Should begin building historical
database for newer sites.
Used to evaluate solute plume
behavior and/or loss of contaminant
mass over time
Primary
Contaminant/daughter
product and geochemical
analytical data
Aquifer and groundwater
data
See Tables 1.2 and 1.3
Used to evaluate the geochemical
environment, determine if biological
and/or abiotc degradation is
occurring, and to estimate natural
attenuation capacity.
Use to evaluate the relative
importance of natural attenuation
mechanisms (e.g., sorption,
dispersion, degradation, etc).
Determine relative importance of
biological versus abiotic degradation
mechanisms.
Supplemental Geochemical and
microbiological data that
allow a refined
interpretation of the
distribution of biological
or abiotic degradation
mechanisms
Additional groundwater
geochemical parameters,
mineralogical analyses,
microcosm studies,
microbiological data, stable
isotope data, etc.
See Tables 3.1 and 3.3
Should be used only when the
additional data will provide useful
information, such as at sites where
the predominant degradation
mechanism(s) is (are) not readily
apparent.
WSRC-TR-2003-00331
December 31, 2004
1-4
At more complicated sites, select supplemental lines of evidence may be used to better
elucidate mechanisms of natural attenuation. At sites where enhanced remediation is required,
some combination of the primary lines of evidence and the data listed under the supplemental
lines of evidence may be useful for determining the most efficient pathway(s) to remediation.
Section 3 describes the supplemental lines of evidence in more detail.
1.2 DATA REQUIRED TO EVALUATE NATURAL ATTENUATION
Tables 1.2 and 1.3 summarize the data required to evaluate the primary lines of evidence used
to evaluate natural attenuation of chlorinated solvents. The interpretation of these data is
discussed in Section 2. Additional data for the supplemental lines of evidence that can be useful
for evaluating MNA, and the interpretation of these data, are discussed in Section 3.
One of the overarching considerations when evaluating MNA is how well the site is
characterized. Evaluation of the lines of evidence is entirely contingent upon the quality of the
available site characterization data and the understanding of the hydrogeologic and groundwater
chemistry relationships at the site. Also, the interpretation of data from the lines of evidence
requires an understanding of what is being sampled and the reason for the sampling. The degree
of characterization required will be site specific and will depend upon, among other things, the
velocity and direction of groundwater flow, the complexity of the hydrogeologic system, and the
distance to potential receptors.
Development of a good conceptual site model (CSM) is fundamental to the evaluation of
natural attenuation. Development of the CSM should include an evaluation of:
• The nature, extent, and magnitude of contamination, including:
¾ The nature and history of the contaminant release;
--Catastrophic or gradual release of non-aqueous phase liquid (NAPL)?
--More than one source area possible or present?
--Divergent or coalescing plumes?
¾ The three-dimensional distribution of mobile and residual NAPL and dissolved
contaminants. The distribution of mobile and residual NAPL is used to define the
dissolved plume source area;
¾ Groundwater and soil/sediment chemical data;
¾ Historical water quality data showing variations in contaminant concentrations;
¾ Chemical and physical characteristics of the contaminants; and
¾ Potential for degradation of the contaminants.
WSRC-TR-2003-00331
December 31, 2004
1-5
Table 1.2
Aquifer and Soil/Sediment Data Required to Evaluate Natural Attenuation and Enhanced Remediation
Analysis
Method
Data Use
Comments
Contaminants of
Concern/Daughter
Products
8260B Used to determine presence of parent
and daughter compounds and rates of
attenuation.
Decreasing source concentrations are important
for natural attenuation and daughter products are
a good indicator of contaminant degradation.
Total Organic
Carbon
SW9060 modified
for soil samples.
Sorption/solute retardation calculations. Procedure must be accurate over the range of
0.1–5 percent TOC.
Bulk Density Geotechnical
Laboratory
Procedure.
Sorption/solute retardation calculations. May be estimated from literature values.
Hydraulic Gradient Determined from
site potentiometric
surface maps.
Estimation of seepage velocity.
Required for groundwater flow and
solute transport models.
At least three measurement points required.
Hydraulic
Conductivity
Determined from
slug tests or
pumping tests.
Estimation of seepage velocity.
Required for groundwater flow and
solute transport models.
Critical parameter with the potential for the most
measurement error. Sensitivity analyses on this
parameter may be useful when estimating
seepage velocity.
Total and Effective
Porosity
Determined from
tracer tests or
estimates from
literature values.
Estimation of seepage velocity.
Required for groundwater flow and
solute transport models.
Literature values typically are used.
Field observation
of reduced iron
minerals
Visual observation. Estimation of the potential for abiotic
degradation through reaction with
reduced iron minerals.
Oxidized iron minerals typically impart some
shade of rust color to the soil ranging from rusty
brown, to yellow or orange, to red or maroon.
Reduced iron minerals typically cause the soil to
be gray to black in color.
WSRC-TR-2003-00331
December 31, 2004
1-6
Table 1.3
Groundwater Data Required to Evaluate Natural Attenuation and Enhanced Remediation*
Analysis Method/Reference Comments Data Use Sample Volume, Sample Container,
Sample Preservation
Field or
Laboratory
Chemicals of
Concern
SW8260B EPA Handbook
method.
Used to determine presence of parent and
daughter compounds and rates of attenuation.
Collect 3x40 mL VOA vials, preserve
with HCL and cool to 4oC.
Fixed-base
laboratory
Chloride IC method E300 a/ Method SW9050
may also be used.
Final product of chlorinated solvent reduction.
Can be used as a tracer.
Collect 1x 1 Liter poly container and
cool to = 4 oC.
Fixed-base
laboratory
Dissolved
Oxygen
E360.1- Dissolved
oxygen membrane
electrode.
Avoid exposure to
atmospheric oxygen.
Concentrations less than about 0.5 mg/L
generally indicate an anaerobic pathway – use
in conjunction with other geochemical data.
Measure dissolved oxygen onsite
using a flow-through cell.
Field
Nitrate IC method E300 a/ Method E300 is a
Handbook method.
Substrate for microbial respiration if oxygen is
depleted. Absence is required for Fe(III)
reduction to occur.
Collect 1 Liter poly container and cool
to 4 oC.
Fixed-base
laboratory
Iron (II) (Fe2+) Colorimetric
Hach Method
Filter with 0.45
micron inline filter.
Indicates an anaerobic degradation process due
to depletion of oxygen, nitrate, and
manganese. Required for abiotic reductive
dechlorination.
Collect 100 mL of water in a
headspace-free container to eliminate
introduction of oxygen and analyze as
soon as possible.
Field
Sulfate (SO42-) IC method E300 a/ Method E300 is a
Handbook method.
Substrate for anaerobic microbial respiration. Collect 1 Liter poly container and cool
to 4 oC.
Fixed-base
laboratory
Sulfide E376.1 Handbook method. Required for abiotic reductive dechlorination. Collect 500 mL in plastic or glass
container, preserve with NaOH to pH
< 9, cool to 4 oC, no headspace.
Fixed-base
laboratory
Oxidation-
Reduction
Potential (ORP)
Direct-Reading Probe Avoid introduction
of oxygen during
sampling.
The ORP of groundwater influences and is
influenced by the nature of biologically
mediated reactions.
Measure ORP onsite using a flow-
through cell.
Field
Methane, Ethane,
and Ethene
RSK-175 Method published by
researchers at the US
EPA.
Presence of methane suggests biodegradation
via methanogenesis. Ethane and ethene are
daughter products of complete dechlorination.
Collect 6x40 mL VOA vials, preserve
with HCL, and cool to cool to 4 oC.
Fixed-base
laboratory
pH E150.1 - Field probe
with direct reading
meter.
Field. Fundamental measurement which is critical for
interpretation of carbonate data. Used as a
well stabilization criterion.
Measure in flow-through cell during
well purging.
Field
Temperature 170.1 - Field probe with
direct reading meter.
Field only. Fundamental measurement required in all
thermodynamic calculations.
Measure in flow-through cell during
well purging.
Field
Conductivity E120.1/SW9050, direct
reading meter
Protocols/Handbook
methods.
General water quality parameter that is
proportional to dissolved ions in solution.
Measure in flow-through cell during
well purging.
Field
Acetylene Under
Development\Not
available
GC-FID method. Product of abiotic reductive dechlorination by
iron sulfide minerals.
Preservation techniques under
development.
Fixed-base
laboratory
*Not all analytes will be required for every site or every sampling event. a/ Ion Chromatography
WSRC-TR-2003-00331
December 31, 2004
1-7
• Geologic and hydrogeologic data (in three dimensions, if feasible):
¾ Lithologic and stratigraphic relationships (e.g., well boring logs, geologic cross-
sections, etc);
¾ Grain size distribution (sand vs. silt vs. clay);
¾ Aquifer hydraulic conductivity;
¾ Groundwater flow gradients and potentiometric or water table surface maps (over
several seasons, if possible);
¾ Preferential flow paths (geologic features, utility conduits, abandoned wells, etc);
and
¾ Interactions between groundwater and surface water and rates of
infiltration/recharge.
• Locations of potential receptor exposure points:
¾ Groundwater production wells; and
¾ Downgradient and cross-gradient groundwater discharge points.
The CSM is a representation of release mechanisms, the NAPL source, groundwater flow and
solute transport, including transport pathways, exposure points, and receptors, and should be
based on available geological, biological, geochemical, hydrological, climatological, analytical
data, and current and future uses for the site. After development, the CSM can be used to help
determine optimal placement of additional monitoring points, as necessary, to aid in the natural
attenuation investigation and to develop a solute fate and transport model, if required.
Contracting and management controls must be flexible enough to allow for the potential for
revisions to the CSM and thus the data collection effort. Successful CSM development involves:
• Definition of the problem to be solved (generally the nature, magnitude, and extent of
existing and future contamination).
• Integration and presentation of available data, including:
¾ local geologic and topographic maps;
¾ geologic data;
¾ hydraulic data;
¾ geochemical data;
¾ current and future land use; and
¾ contaminant concentration and distribution data.
WSRC-TR-2003-00331
December 31, 2004
1-8
• Determination of additional data requirements, including:
¾ borehole locations and monitoring well spacing;
¾ quality assurance project plan (QAPP);
¾ sampling and analysis plan (SAP); and
¾ any data requirements that have not been adequately addressed.
In some cases, available site-specific data are limited. If this is the case, future site
characterization activities should include collecting the data necessary to evaluate natural
attenuation at the site. Regardless of whether natural attenuation is selected as a sole remedial
strategy or in conjunction with an engineered remediation system, the additional costs incurred
by such data collection will likely be outweighed by the cost savings that will be realized. Much
of the data collected to evaluate natural attenuation also can be used to design and evaluate other
remedial measures.
WSRC-TR-2003-00331
December 31, 2004
2-1
SECTION 2
PRIMARY LINES OF EVIDENCE
The primary lines of evidence used to evaluate MNA include: 1) an evaluation of plume
behavior using temporal and spatial contaminant data; and 2) field and laboratory data for
evaluating the relative importance of natural attenuation mechanisms. This section describes the
primary lines of evidence and their use.
2.1 CONTAMINANT CONCENTRATION DATA
A statistically significant historical database of contaminant concentration data showing
plume stabilization and/or loss of contaminant mass over time is used to demonstrate that natural
attenuation is occurring at a site. This is perhaps the fundamental line of evidence for evaluating
monitored natural attenuation as a remedial approach. In fact, historical data are important for
evaluating any remediation technique. It is important to note that plume stabilization can occur
with or without destructive attenuation mechanisms. In some cases, nondestructive mechanisms
of natural attenuation such as dispersion, sorption, and volatilization may be sufficient to cause
the solute plume to reach steady-state equilibrium, or even recede if the strength of the NAPL
source is decreasing due to natural weathering or engineered remediation. In addition, the data
to support this line of evidence must be collected to help evaluate the effectiveness of enhanced
remediation. The aquifer and soil/sediment data described in Section 2.2 are used to help
separate the components of natural attenuation and to help quantify the natural attenuation
capacity of the system.
Both graphical and statistical methods can be used to evaluate plume behavior, and ultimately
some combination of graphical representations and statistical techniques likely will be used to
evaluate plume behavior at most sites. It is important when evaluating the behavior of a
contaminant plume that the historical data demonstrate a clear and meaningful pattern in
contaminant mass and/or concentration over time at appropriate monitoring or sampling points.
Methods to accomplish this are discussed in USEPA (1992) and Helsel and Hirsch (2002).
Useful graphical approaches to evaluate solute plume behavior are briefly described in
Section 2.1.1. Statistical approaches to evaluate plume behavior are presented in Section 2.1.2.
WSRC-TR-2003-00331
December 31, 2004
2-2
2.1.1 Graphical Evaluation of Plume Behavior
There are several ways to present data showing changes in contaminant concentrations and
plume configuration over time. One method consists of preparing isopleth maps of contaminant
concentration over time. Figure 2.1 shows isopleth maps of volatile organic compound (VOC)
concentrations in groundwater where there is a NAPL source. The VOC concentrations shown
on this figure were mapped at the depth of highest concentration. Note that VOC data were
collected during the same season (the season with the highest VOC concentrations). This is
important because seasonal variations in recharge can cause significant changes in contaminant
concentrations and groundwater geochemistry and an apparent reduction in plume size and/or
contaminant concentrations could simply be the result of seasonal variation.
Another method that can be used to present data showing changes in contaminant
concentrations and plume configuration over time is to plot contaminant concentrations versus
time for individual monitoring wells, or to plot contaminant concentrations versus distance
downgradient for several wells along the groundwater flow path for several sampling events. It
is important when plotting data in this manner that at least one data point is located a short
distance downgradient of the solute plume in the groundwater flow path. This ensures that
contaminants are not migrating downgradient of the observation wells. To ensure that
contaminants are not migrating, it is important that downgradient wells are located in the path of
contaminated groundwater flow. Geochemical data can be used to confirm that downgradient
wells are sampling groundwater that was once contaminated with organic compounds, as
discussed in Wiedemeier and Haas (2002).
Visual inspection of plotted data for concentration versus time and/or distance can yield
qualitative information on the presence of trends that are indicative of solute plume behavior.
While such plots are recommended for most plume stability analyses, visually discerning trends
in the plotted data is a subjective process, particularly if the data does not display a relatively
obvious uniform trend and shows some variability over time. For data that shows considerable
variation in concentrations, such as is common for concentration versus distance plots, the use of
semi-logarithmic plots can be useful. Plotting the concentration data on a logarithmic scale
against time on an arithmetic (linear) scale can counter the relatively large changes in
concentration (e.g., a concentration reduction from 1 mg/L to 1 ug/L represents a 1,000-fold
reduction) and effectively “linearize” the data.
As an example, Figure 2.2 shows sampling locations for the plots of contaminant
concentration versus time and distance downgradient which are presented in Figure 2.3. Based
on the geochemical data presented in Figure 2.2 it can be concluded with reasonable certainty
that well H is in the plume’s flow path, therefore if the plume were migrating downgradient it
should be detected. Wells F and H are spaced 100 feet apart, and the groundwater seepage
velocity is 150 feet per year; with 8 years of sampling data from the same season it can probably
be concluded with reasonable certainty that the plume is not migrating downgradient.
D:\doe\report\lines of evidence\Figure 2.1.ai
Westinghouse Savannah River MNA/EA Project
Figure 2.1
Isopleth Maps of Total VOC Concentrations in
Groundwater at the Depth of Highest
Contaminant Concentration
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
600
600
300
300
FEET
FEET
0
September 1993
ND
ND
4,000
4,000
8,000
8,000
September 1994
ND
ND
4,000
4,000
8,000
8,000
20,000
20,000
September 1995
ND
ND
4,000
4,000
4,000
4,000
8,000
8,000
4,000 Line of Equal Total
VOC Concentration (µg/L)
Groundwater Flow
Velocity ~ 1,600 feet per year
Groundwater
Groundwater
Flow Direction
Flow Direction
September 1998
ND
ND
4,000
4,000
8,000
8,000
2-3
West
West East
East
D
ABCEFGH
ABCDEFGH
L
K
J
I
O2 = 8 mg/L
NO3- = 10 mg/L
SO42- = 100 mg/L
CH4 < 0.001 mg/L
O2 = 6.5 mg/L
NO3- = 13 mg/L
SO42- = 96 mg/L
CH4 < 0.001 mg/L
O2 = 8.5 mg/L
NO3- = 13 mg/L
SO42- = 89 mg/L
CH4 < 0.001 mg/L
100 feet
O2 = 0.3 mg/L
NO3- = 0.5 mg/L
SO42- = 8 mg/L
CH4 = 7 mg/L
O2 = 0.1 mg/L, NO3- = 0.2 mg/L
SO42- = 3 mg/L, CH4 = 10 mg/L
Groundwater Flow Direction
Velocity = 150 feet per year
Extent of NAPL
Extent of Dissolved
Contaminant Plume
Extent of Dissolved
Contaminant Plume
D:\doe\report\lines of evidence\Figure 2.2.ai
Westinghouse Savannah River MNA/EA Project
Figure 2.2
Sampling Locations for the Plots of
Contamination versus Time and Distance
Downgradient Presented in Figure 2.3
Date: 12/12/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
100 feet
2-4
6/89
6/93
6/91
6/95
BC
June, 1989
Well C
June, 1991
June, 1993
June, 1995
June, 1997
F
6/97
Time
Concentration
Concentration
D:\doe\report\lines of evidence\Figure 2.3.ai
Westinghouse Savannah River MNA/EA Project
Figure 2.3
Plots of Contaminant Concentration versus
Time and Distance Downgradient
Date: 12/17/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
Distance Downgradient
2-5
WSRC-TR-2003-00331
December 31, 2004
2-6
The combination of decreasing contaminant concentrations shown by the plots in Figure 2.3,
and the lack of contaminant migration provide reasonable evidence for natural attenuation and
contaminant mass destruction. The chemical and geochemical data discussed in Section 2.2 can
be used to show that this loss of contaminant mass is the result of destructive attenuation
mechanisms.
2.1.2 Statistical Methods for Evaluating Plume Behavior
Statistical methods are powerful tools for identifying significant changes and trends in
groundwater concentration data. They provide for an objective evaluation of the data and allow
statements to be made about the confidence in results. This provides a quantitative indication of
the likelihood that conclusions drawn from the data are correct. In evaluating natural
attenuation, statistical methods are used to assess groundwater monitoring data for the presence
of significant trends or changes in concentrations over time that can provide insight into solute
plume behavior. Once again, it is paramount to verify the monitoring events and data subject to
statistical analyses are comparable. If high water levels correlate with higher contaminant
concentrations, then data from high and low water table events may not be comparable. If
sampling was conducted during an extreme weather event (e.g. 100-year flood), then it may not
be comparable to previous events. A more detailed discussion of the concept of comparibility is
found in Gilbert (1987).
The application of statistics requires an understanding of the underlying assumptions of the
tests and nature of the data since these determine the selection of appropriate methods and
interpretation of the results. While a detailed review of the statistical analysis of concentration
data and its application are beyond the scope of this document, a brief discussion of the
significant factors and some methods that are applicable in the majority of situations is provided.
More detailed discussion is available in several statistics texts (e.g., Gilbert, 1987; Gibbons,
1994; USEPA, 2000; Helsel and Hirsch, 2002). This is not a theoretical discussion; rather, it
provides practical considerations where statistical results are used in decision-making.
Statistical tests are a form of hypothesis testing and their basis is the comparison of what
statisticians call the “null hypothesis” (H0) to an alternative hypothesis (H1). The null hypothesis
is the statistical hypothesis being tested; generally that the test results are merely a product of
chance factors. For example, to test for a trend in a concentration time series, H0 would be that
there is no change in concentration over time, and H1 would be that the concentration is either
increasing or decreasing with time. The two hypotheses are compared using a test statistic that is
calculated from the data series being tested.
Most statistical tests are intended to detect a significant difference between a group of
samples or from a predefined condition. This is determined by comparing the value for the test
statistic calculated from the data set to the probability of obtaining that value purely due to
chance. The probability values are determined from the “null” distribution for the test statistic
that is the distribution of values for the test statistic under the null hypothesis (H0). The
WSRC-TR-2003-00331
December 31, 2004
2-7
significance level is a means of determining whether the test statistic is “significantly” different
from values that would typically occur under H0. If the probability for the test statistic value
calculated from the data set is less than the level of significance, the null hypothesis (H0) can be
rejected in favor of the alternative hypothesis (H1).
There are two possible types of decision errors associated with statistical hypothesis testing.
A Type I error is when H0 is incorrectly rejected. A Type II error is when H0 is accepted when
H1 is true. Both types of decision errors have implications for the conclusions drawn from
results of statistical tests.
A Type I error is rejecting the null hypothesis (H0) when it is in fact true. This is essentially
equivalent to a “false positive” result, such as concluding that there is an increasing or decreasing
trend in concentration over time when no trend is actually present. The probability of incorrectly
rejecting H0 is the “significance level” (α) of the test. Type I errors are controlled by selecting an
appropriate α-value to reduce the likelihood of drawing an incorrect conclusion from the test.
The inability to reject the null hypothesis (failure to accept the alternative hypothesis) at some
level of significance does not imply that the null hypothesis is true. A Type II error is failing to
reject (accepting) the null hypothesis (H0) when it is false and the alternative hypothesis (H1) is
true. This is essentially equivalent to a “false negative” result, such as concluding that there is
no trend in concentration over time when an increasing or decreasing trend is actually present.
The probability of this occurring is β and the power of a statistical test to detect a significant
difference is 1-β. The statistical power of a test is related to both the α-value selected and the
sample size (n).
Ideally, we would like to minimize both Type I and Type II errors in using statistical tests, but
this is difficult in practice. The importance of either type of decision error should be evaluated in
terms of the ultimate use of the results of the statistical test. A pragmatic approach is to specify
an acceptable value for α and concurrently reduce β by 1) increasing the sample size and/or 2)
using a statistical test with the greatest power for the type of data being evaluated (Helsel and
Hirsch, 2002).
Statistical tests are described as one-sided or two-sided depending upon the specific
alternative hypothesis involved. A two-sided test is used when a difference in either direction
from H0 would cause H0 to be to be rejected, such as a test for detecting the presence of a
trend/change in concentration. For example, if there is no reason to assume that concentrations
are not stable or that departures from H0 in only one direction are of interest, a two-sided test is
appropriate. A one-sided test is used when a change in only one direction from H0 would cause
H0 to be to be rejected, such as a test for detecting an increase (or decrease) in concentration
over time. For example, if only evidence that concentration is increasing (or decreasing) over
time is considered important, H0 would be stated as “the change in concentration over time is
less (or greater) than or equal to zero (0),” and H1 would be “the change in concentration over
time is greater (less) than zero (0).”
WSRC-TR-2003-00331
December 31, 2004
2-8
The null distributions for most test statistics are symmetrical and the probability values for
only one “tail” of the distribution are given. For detecting an increase (or decrease), only the
difference in one direction is important and the critical test statistic value at α is used (one-sided
tail). For detecting the presence of a trend/change in concentration, both a positive or negative
difference is important and the critical test statistic value at α/2 is used (two-sided tail).
The issue of confidence levels, or significance levels, is important when applying statistics.
The practical implication of the confidence level is that there is error associated with the decision
to reject the null hypothesis. If the calculated value of the test statistic leads you to reject the null
hypothesis, it does not mean that the value for the test statistic you obtained could not have
occurred by chance. It means that the probability of obtaining that value by chance alone is
sufficiently small that it is reasonable to conclude that the result is not due to chance and that the
decision to reject the null hypothesis is correct. The confidence level simply quantifies the
likelihood that rejecting the null hypothesis is appropriate.
The confidence level for a statistical test is related to the significance level (α) and is simply
described by the value 1-α, typically expressed as a percentage. The significance level (α) is
specified in advance of the test and defines the “acceptable” level of Type 1 error that the user is
willing to tolerate in deciding to reject the null hypothesis. For example, if the desired
confidence level for a statistical test is 95% (0.95), the significance level would be specified as
0.05 and the null hypothesis would be rejected if the calculated test statistic value has a
probability ≤ 0.05. This means that the likelihood of making an incorrect decision to reject the
null hypothesis is 5 in 100 (1 in 20) and, conversely, the likelihood that the decision to reject the
null hypothesis is correct is 95 in 100 (19 in 20).
The confidence level simply quantifies the “confidence” associated with obtaining a
“significant” result for a statistical test, such as concluding that there is a trend in concentration
over time or a difference in concentrations. There is no magic to defining the appropriate
confidence level and adjusting the confidence level simply changes the tolerance for Type I error
in decision-making. In most scientific applications, a 95% confidence level is used since there is
general concurrence that the associated error (5%) is sufficiently small. Decreasing the
confidence level for a statistical test will increase the likelihood of obtaining a “significant”
result, but will also increase the chances that the null hypothesis will be incorrectly rejected. The
specified confidence level is simply a reflection of the user’s willingness to accept a mistaken
conclusion for a statistical test.
2.1.2.1 Nature of Groundwater Concentration Data and Appropriate Statistical Methods
Issues involved with the statistical analysis of groundwater concentration data are myriad, but
most commonly involve missing values, non-detect (censored) values, small number of data
points, and the lack of certain knowledge of the underlying distribution. All of these complicate
the application of statistical methods and either require significant data manipulation or the use
of methods that are little affected by these data characteristics. Trend analysis, in particular, is
WSRC-TR-2003-00331
December 31, 2004
2-9
sensitive to these issues, as well as to changes in sampling and analytical procedures, seasonal or
other cyclic variation in the data, and correlated data (Gilbert, 1987).
Statistical approaches can be separated into parametric and non-parametric methods. The
familiar parametric statistics, such as regression analysis, rely on data conforming to an
underlying distribution, such as normal (Gaussian) or log-normal. Parametric statistics are
sensitive to missing data points and outliers, how non-detect values are handled, and departures
from the assumed distribution. Non-parametric statistical methods do not depend on
assumptions regarding the underlying data distribution and are also known as “distribution-free”
methods. They can accommodate missing data points and non-detect values that are common in
groundwater concentration data sets. These methods rely on the ranks or relative magnitudes of
the data rather than the actual values and are fairly straightforward to use. In many situations,
particularly those involving small data sets, non-parametric methods perform as well or better
than parametric methods (Helsel and Hirsch, 2002).
The selection of statistical methods is frequently limited by the availability of sufficient data.
Aside from the issues mentioned previously, parametric methods are sensitive to sample size and
their power is reduced for small data sets. This is a common problem for groundwater
concentration data. Non-parametric methods typically are equally or more powerful for
discerning trends and changes for small data sets.
Because of the issues associated with most groundwater concentration data, the use of non-
parametric techniques are generally preferred for environmental concentration data (Gilbert,
1987; Gibbons, 1994) and some commonly used methods are described briefly below.
Additional information on these non-parametric methods is provided in Hollander and Wolfe
(1999), Conover (1999), and Helsel and Hirsch (2002).
2.1.2.2 Tests for Trend
The Mann-Kendall test for trend (Mann, 1945; Kendall, 1975) is used to determine the
presence or absence of a trend in concentration over time for individual monitoring points. It is a
test for zero slope of time-ordered data that is based on a non-parametric analog of linear
regression. The basic methodology and its variants (such as the Seasonal Mann-Kendall test) are
described in Gilbert (1987) and Helsel and Hirsch (2002) and four or more independent sampling
events are required. The results of the Mann-Kendall test indicate the presence or absence of a
statistically significant increasing or decreasing trend in concentrations over time at a monitoring
point. These results can be used to help evaluate whether the solute plume is receding,
expanding, or stable.
The Mann-Kendall test for 4-40 comparable data points is very straightforward to apply and
an example calculation is provided in Table 2.1. Concentration data are ordered sequentially
over time and a matrix is constructed comparing each data value to subsequent values.
WSRC-TR-2003-00331
December 31, 2004
2-10
Table 2.1
Example Calculation of the Mann-Kendall Statistic for TCE Concentrations in a Monitoring
Well with Ten Sampling Events.
Event 1 2 3 4 5 6 7 8 9 10
Concentration (µg/L) 56 78 63 43 45 36 38 40 46 42 Row Sums
1 1 -1 -1 -1 -1 -1 -1 -1 -5
-1 -1 -1 -1 -1 -1 -1 -1 -8
-1 -1 -1 -1 -1 -1 -1 -7
1 -1 -1 -1 1 -1 -2
-1 -1 -1 1 -1 -3
1 1 1 1 4
1 1 1 3
1 1 2
-1 -1
Mann-Kendall Statistic S = -17
WSRC-TR-2003-00331
December 31, 2004
2-11
Starting with the earliest data point, each subsequent data point is compared and a value entered
into the matrix: +1 if the later value is greater, -1 if the later value is less, and 0 if the later value
is equal to the earliest data point. The process is repeated for the next data point in the sequence,
comparing its value to subsequent ones, until all data points in the sequence have been compared
and appropriate values entered into the matrix. The values in each row in the matrix are then
summed and the row sums are then summed to generate the Mann-Kendall statistic (S).
Once the S-statistic has been calculated, it is compared to the table of null probability values
for S for the number of data points (n) in the series (Table 2.2). If the probability value for the
calculated S-statistic and the number of data points (n) is less than the specified significance
level for the test (α for one-sided; α/2 for two-sided), the result is significant at the 1-α
confidence level and a trend is present. The calculated S-statistic (-17) and n (10) for the
example calculation in Table 2.1 correspond to a probability of 0.078 in Table 2.2. For a one-
sided test, this result is less than the α for the 90% confidence level (α = 0.1) indicating a
significant result, but is greater than the α for the 95% confidence level (α = 0.05) indicating
that the result is not significant at this level of confidence. Because the S value is negative, it is
concluded that a decreasing trend in concentration is present at the 90% confidence level.
Whether this result is “significant” depends upon the significance level (α) specified for the test.
The Mann-Kendall test is robust to missing data points and non-detect values. Missing data
points are simply ignored since they don’t influence the test result. Non-detect values are
replaced with a common value less than the smallest concentration value in the data series. If
multiple detection limits are involved, the data must be further censored at the highest detection
limit (Helsel and Hirsch, 2002). This decreases the power of the test to detect trends due to the
increased number of tied values, but the impact in most situations involving small data sets is not
significant. If the number of tied values is a significant proportion of the data series, the tie
correction for the large-sample approximation described below can be used.
If more than 40 data points are available, a modification of the Mann-Kendall test based on
the normal approximation can be used. This version of the Mann-Kendall test uses “Z” as the
test statistic. The test is performed by calculating the S-statistic for the data set as described
previously. The variance of the S-statistic is then calculated as:
() ()( )
()( )
1
1
VAR 1 2 5 1 2 5
18 =
=−+− −+
∑
q
pp p
p
Snnn tt t
where n is the number of data points in the data set, q is the number of groups of tied values, and
tp is the number of data points in pth group of tied values. If the calculated S is 0, the Z-statistic
is also 0. Otherwise, the Z-statistic is then calculated as follows:
()
()
1 if 0
VAR
1 if 0
VAR
−
=
>
+
=
<
S
ZS
S
SS
S
WSRC-TR-2003-00331
December 31, 2004
2-12
Table 2.2
Table of Null Probabilities for the Mann-Kendall Statistic, n = 4 through 20
Number of data points (n)
S
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0 0.625 0.592 0.548 0.540 0.527 0.524 0.518 0.516 0.513
±1 0.500 0.500 0.500 0.500 0.500 0.500 0.500 0.500
±2 0.375 0.408 0.452 0.460 0.473 0.476 0.482 0.484 0.487
±3 0.360 0.386 0.431 0.440 0.457 0.461 0.470 0.473
±4 0.167 0.242 0.360 0.381 0.420 0.429 0.447 0.452 0.462
±5 0.235 0.281 0.364 0.381 0.415 0.423 0.441 0.445
±6 0.042 0.117 0.274 0.306 0.369 0.383 0.412 0.420 0.436
±7 0.136 0.191 0.300 0.324 0.374 0.385 0.411 0.418
±8 0.042 0.199 0.238 0.319 0.338 0.378 0.388 0.411
±9 0.068 0.119 0.242 0.271 0.334 0.349 0.383 0.391
±10 0.008 0.138 0.179 0.273 0.295 0.345 0.358 0.387
±11 0.028 0.068 0.190 0.223 0.295 0.313 0.354 0.365
±12 0.089 0.130 0.230 0.255 0.313 0.328 0.362
±13 0.008 0.035 0.146 0.179 0.259 0.279 0.327 0.339
±14 0.054 0.090 0.190 0.218 0.282 0.299 0.339
±15 0.001 0.015 0.108 0.141 0.225 0.248 0.300 0.314
±16 0.031 0.060 0.155 0.184 0.253 0.271 0.315
±17 0.005 0.078 0.109 0.194 0.218 0.275 0.290
±18 0.016 0.038 0.125 0.153 0.225 0.245 0.293
±19 0.001 0.054 0.082 0.165 0.190 0.250 0.267
±20 0.007 0.022 0.098 0.126 0.199 0.220 0.271
±21 0.000 0.036 0.060 0.140 0.164 0.227 0.245
±22 0.002 0.012 0.076 0.102 0.175 0.196 0.250
±23 0.023 0.043 0.117 0.141 0.205 0.223
±24 0.001 0.006 0.058 0.082 0.153 0.174 0.230
±25 0.014 0.030 0.096 0.120 0.184 0.203
±26 0.000 0.003 0.043 0.064 0.133 0.154 0.211
±27 0.008 0.020 0.079 0.101 0.165 0.184
±28 0.001 0.031 0.050 0.114 0.135 0.193
WSRC-TR-2003-00331
December 31, 2004
2-13
Table 2.2 - Continued
Table of Null Probabilities for the Mann-Kendall Statistic, n = 4 through 20
Number of Data Points (n)
S
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
±29 0.005 0.013 0.063 0.084 0.147 0.166
±30 0.000 0.022 0.038 0.097 0.118 0.176
±31 0.002 0.008 0.050 0.070 0.130 0.149
±32 0.016 0.029 0.083 0.102 0.159
±33 0.001 0.005 0.040 0.057 0.115 0.133
±34 0.010 0.021 0.070 0.088 0.144
±35 0.000 0.003 0.031 0.046 0.100 0.119
±36 0.007 0.015 0.058 0.076 0.130
±37 0.002 0.024 0.037 0.088 0.105
±38 0.004 0.011 0.048 0.064 0.117
±39 0.001 0.018 0.029 0.076 0.093
±40 0.003 0.007 0.039 0.054 0.104
±41 0.000 0.013 0.023 0.066 0.082
±42 0.002 0.005 0.032 0.046 0.093
±43 0.010 0.018 0.056 0.072
±44 0.001 0.003 0.026 0.038 0.082
±45 0.007 0.014 0.048 0.062
±46 0.000 0.002 0.021 0.032 0.073
±47 0.005 0.010 0.041 0.054
±48 0.001 0.016 0.026 0.064
±49 0.003 0.008 0.034 0.047
±50 0.001 0.013 0.021 0.056
±51 0.002 0.006 0.029 0.040
±52 0.000 0.010 0.017 0.049
±53 0.002 0.004 0.024 0.034
±54 0.008 0.014 0.043
±55 0.001 0.003 0.020 0.029
±56 0.006 0.011 0.037
WSRC-TR-2003-00331
December 31, 2004
2-14
Table 2.2 - Continued
Table of Null Probabilities for the Mann-Kendall Statistic, n = 4 through 20
Number of Data Points (n)
S
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
±57 0.001 0.002 0.016 0.025
±58 0.004 0.009 0.032
±59 0.000 0.001 0.013 0.021
±60 0.003 0.007 0.027
±61 0.001 0.011 0.017
±62 0.002 0.005 0.023
±63 0.001 0.009 0.014
±64 0.002 0.004 0.020
±65 0.000 0.007 0.012
±66 0.001 0.003 0.017
±67 0.005 0.010
±68 0.001 0.002 0.014
±69 0.004 0.008
±70 0.001 0.002 0.012
±71 0.003 0.006
±72 0.000 0.001 0.010
±73 0.003 0.005
±74 0.001 0.008
±75 0.002 0.004
±76 0.001 0.007
±77 0.001 0.003
±78 0.000 0.006
±79 0.001 0.003
±80 0.005
±81 0.001 0.002
±82 0.004
±83 0.001 0.002
WSRC-TR-2003-00331
December 31, 2004
2-15
Table 2.2 - Concluded
Table of Null Probabilities for the Mann-Kendall Statistic, n = 4 through 20
Number of Data Points (n)
S
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
±84 0.003
±85 0.000 0.001
±86 0.002
±87 0.001
±88 0.002
±89 0.001
±90 0.002
±91 0.001
±92 0.001
±93 0.000
±94 0.001
±95
±96 0.001
±97
±98 0.001
±99
±100 0.000
WSRC-TR-2003-00331
December 31, 2004
2-16
The sign of the calculated Z indicates whether a trend is increasing (positive) or decreasing
(negative). Once the Z-statistic has been calculated, it is compared to the table of null probability
values for Z that can be found in most statistics texts. Critical values for the Z-statistic at
probabilities for the commonly used significance levels for one-sided (p=α) and two-sided
(p=α/2) tests are 1.29 (p=0.1), 1.64 (p=0.05), and 1.96 (p=0.025).
A general consideration for using the Mann-Kendall test is that a non-significant result does
not demonstrate stability since the result could be due to concentrations at the monitoring point
actually being at steady state (stable) or to the data set being inadequate to provide a statistically
significant result (Barden, 2003). Failing to reject H0 does not mean that it was "proven" that
there is no trend. Rather, it is a statement that the evidence available is not sufficient to conclude
that there is a trend at the specified confidence level.
A suggested approach to dealing with the issue of a non-significant result for the Mann-
Kendall test is to use the coefficient of variation as an indication, or “test,” of stability
(Wiedemeier et al., 1999; GSI, 1998; Ling et al., 2003). The coefficient of variation (CV)
measures the spread of a set of data as a proportion of its mean and the suggested approach
concludes that a Mann-Kendall test that is not significant at the 90% confidence level where
CV < 1 indicates stability. However, the coefficient of variation is a relative measure of variation
described by the ratio of the sample standard deviation to the sample mean. Thus, it depends
upon both values and has no implicit meaning. If the mean value is large, even a small CV can
include significant variation. Data series with “low” values for CV certainly show less scatter in
the data, but there is no objective basis for using a particular value of CV to determine
“stability.”
A useful variation on the Mann-Kendall test is a test for “homogeneity of stations” (Gilbert,
1987; Helsel and Hirsch, 2002). This test essentially pools the results for Mann-Kendall tests at
individual monitoring points and allows statements to be made about consistency of trends
throughout the plume or portions of the plume (e.g., whether the trends at all monitoring points
are in the same direction - all increasing or all decreasing). Such a general statement about the
presence or absence of monotonic trends is useful for making interpretations of the overall
behavior of the entire plume or specific portions of the plume. For chlorinated solvent solute
plumes, these results can be used in combination with geochemical data to discern different types
of environments.
The presence of seasonal variability in groundwater concentration time series data can make
discerning trends difficult brcause it contributes short-term variation, caused by water level
fluctuations and other seasonal effects, that appear as background noise in a Mann-Kendall test
for the whole time series. If the source of the seasonal effect can be identified, one way to
“remove” the effect is to normalize the concentration data to the source variable. For example, if
groundwater concentrations are shown to be correlated with water levels in monitoring wells,
they could be “normalized” by dividing concentrations by water levels. This is a simplistic
WSRC-TR-2003-00331
December 31, 2004
2-17
approach and more sophisticated data normalization techniques can be used (Helsel and Hirsch,
2002).
The “Seasonal Kendall test” (Hirsch et al.,1982; Hirsch and Slack, 1984) is a modification of
the Mann-Kendall test that addresses this short-term variability due to seasonality and allows
evaluation of overall trends in the time series. In a seasonal Kendall test, the Mann-Kendall test
is applied to each season (e.g., quarter) separately and then the results are combined for an
overall test (Hirsch et al., 1982). Each season by itself may show a positive trend, none of which
is significant, but the overall seasonal Kendall statistic can be quite significant. The test has all
the advantages of the Mann-Kendall test, but is more robust because it removes short-term
variability caused by seasonality. When successive seasons are correlated, a correction must be
used based on the covariance among seasons (Hirsch and Slack, 1984).
The seasonal Kendall test consists of calculating the Mann-Kendall statistic, S, and its
variance, VAR(S), for the data from each season collected over a period of years. These
“seasonal” statistics are then summed and the test statistic Z is calculated as previously described
using the summed values. As with the normal approximation, the sign of the calculated Z
indicates whether a trend is increasing (positive) or decreasing (negative). The calculated Z-
statistic then is compared to the table of null probability values for Z that can be found in most
statistics texts. There is some question regarding the direct application of the standard Z table
values for a small number of “seasons” and few years of sampling data (Gilbert, 1987).
However, the exact distribution for the test statistic can be determined using the technique
described in Hirsch et al. (1982).
A practical limitation on the use of the seasonal Kendall test for evaluating groundwater data
in long-term monitoring of natural attenuation is that seasonal (e.g., quarterly) data must be
available. If the monitoring frequency is changed to annual or semi-annual basis, this seasonal
data may be lost. If seasonal effects are identified during site characterization, or in the early
stages of the long-term monitoring program, continued quarterly monitoring may be warranted to
adequately define the impact of seasonal effects on trend results and to determine the appropriate
frequency for later monitoring. Additionally, the number of data points for each season and the
number of seasons considered can impact the results of the seasonal Kendall test. Generally, at
least three years of monitoring data should be included in the analysis.
2.1.2.3 Tests for Differences Between Groups of Data
Another type of statistical test that is commonly suggested for evaluating groundwater
concentration data for natural attenuation is a test for significant differences between groups of
data. Several non-parametric methods are available for performing such comparisons and the
appropriate method depends upon the number of groups to be compared and whether the data is
paired (Gilbert, 1987; Helsel and Hirsch, 2002). All of these methods are non-parametric
analogs of the Student's t-test. These methods test whether measurements from one data set are
WSRC-TR-2003-00331
December 31, 2004
2-18
consistently larger or smaller than those from another data set, either using relative ranks of the
data or the differences.
Two-sample tests are typically used for comparing earlier data sets to those from later time
periods. These can include comparing concentrations in several monitoring points at two time
points or comparing concentrations from an individual monitoring point for one time period to
those for another time period (e.g., quarterly monitoring results for one year to those for another
year). Such a comparison can essentially identify the presence or absence of a step trend in
concentrations over time. Two-sample procedures should only be used when the data set(s)
being analyzed can be naturally broken into two distinct time periods or when a known event has
occurred that is likely to have resulted in a significant change in concentrations (Helsel and
Hirsch, 2002). In general, the monotonic trend methods discussed previously are more
appropriate.
The Mann-Whitney U test (Mann and Whitney, 1947), also called the Wilcoxon rank sum
test, is commonly suggested for the purpose of identifying step trends and has been specified in
some States’ regulations (e.g., New Jersey; Wisconsin). The typical application of this test is to
compare concentrations from individual monitoring points for one time period to those for
another time period (e.g., quarterly monitoring results for one year to those for another year).
The Mann-Whitney U test is based on the assumption that the two data sets are independent,
meaning that there is no natural way to pair the data. However, in the typical use of this test for
evaluating natural attenuation, the data for the two groups can be considered paired by "seasons"
and are not really independent. Use of the Mann-Whitney U test should be limited to the
situations noted previously and where data set independence can be assured.
Data is considered paired when there is a natural way to spatially or temporally associate data
values in each group. In many cases, the data involved in evaluating natural attenuation will be
paired by location or by season (e.g., quarterly data). In such situations, a paired-sample test,
such as the "sign test" or the "Wilcoxon signed rank test" (not to be confused with the Wilcoxon
rank sum test), is more appropriate (Gilbert, 1987).
The sign test is more versatile than the Wilcoxon signed rank test since it has no distributional
assumptions and can accommodate a few non-detect values. However, it has less ability to
detect differences between populations. The test statistic is the number of data pairs where x1i <
x2i; the number of positive differences. At small sample sizes the sign test has limited utility.
The Wilcoxon signed rank test is a more powerful alternative to the sign test that is more likely
to detect significant differences between data sets. However, it does require that the underlying
distribution is symmetrical. In some cases where the differences are not symmetric in the
original units, but a logarithmic transformation of the two data sets produces symmetric
differences, the Wilcoxon signed rank test is also appropriate (Helsel and Hirsch, 2002).
The Wilcoxon signed rank test involves calculating and ranking the differences (Di) of the
data pairs. The H0 for the test is the median of the differences is zero (0). Example calculations
WSRC-TR-2003-00331
December 31, 2004
2-19
are shown in Table 2.3a for quarterly concentration data in a monitoring well from two years,
and in Table 2.3b for concentration data from multiple monitoring wells for two years. The
difference between each pair of values (xi - yi) in the two data sets are calculated and the absolute
value of the differences (|Di|) are then ranked from smallest to largest. The test uses only non-
zero differences, so tied values (xi - yi = 0) are deleted and the sample size is reduced by the
number of tied values. When two non-zero differences are tied, the average of the ranks
involved is assigned to the tied values.
The signed rank (Ri) for each pair is determined by the sign of the difference for each pair (xi -
yi); “+” for a positive difference and “–“ for a negative difference. The test statistic W+ is then
calculated as the sum of the positive ranks. The W+ statistic is compared to a table of critical
values for W+ quantiles (Table 2.4). For the appropriate sample size in Table 2.4, the critical
values (x and x′) are obtained for the significance level of the test. For a two-sided test (p = α/2),
the null hypothesis is rejected if W+ ≥ x or W+ ≤ x′ (x tends to be larger or smaller than y). For
a one-sided test (p = α), the null hypothesis is rejected if either W+ ≥ x (x tends to be larger than
y; concentrations decrease) or W+ ≤ x′ (x tends to be smaller than y; concentrations increase).
The calculated W+ statistic (9) and n (4) for the example shown in Table 2.3a correspond to a
probability of 0.125 in Table 2.4. For a one-sided test, the W+ statistic (9) is not greater than the
critical value for a significant decrease (x = 10) or less than the critical value for a significant
increase (x′ = 0) at the 90% confidence level (α = 0.1) indicating a non-significant result so the
null hypothesis of no increase, or decrease, of concentration in this monitoring well is accepted.
For the sample size in this example, the 95% confidence level for a one-sided test cannot be
resolved and neither the 90% or 95% confidence levels can be resolved for a two-sided test. This
illustrates the limitation of small sample sizes for such tests.
In the example shown in Table 2.3b, the symmetry of the differences for the data pairs is
questionable. Recalculating the differences using the logarithms of the data values, log(xi) –
log(yi), gives a distribution of differences that is more symmetrical. These differences are then
ranked as described previously and the W+ statistic is calculated. The calculated W+ statistic (24)
for the example shown in Table 2.3b is less than the critical values for x and greater than the
critical values for x′ at the 90% (α = 0.1) and 95% (α = 0.05) confidence levels for the sample
size, n (8). This indicates a non-significant result for either a one-sided or two-sided test at these
confidence levels so the null hypotheses would be accepted and no significant change in overall
concentrations for these monitoring well is indicated.
2.1.2.4 Using Statistical Results
The use of results from statistical tests in evaluating the first line of evidence for natural
attenuation allows quantifiable patterns in contaminant concentrations over time to be
determined. These can provide insight into solute plume behavior and changes over time in
different parts of
WSRC-TR-2003-00331
December 31, 2004
2-20
Table 2.3
Example Calculations for the Wilcoxon Signed-Rank Test Comparing Groups of Paired Data
for A) Quarterly Concentration Data in a Single Monitoring Well for Two Years (µg/L), and B)
Concentrations in Several Monitoring Wells for Two Years (µg/L).
A
Quarter Year 1 Year 2 Difference Rank
1st 32 27 5 4
2nd 46 42 4 2.5
3rd 28 30 -2 -1
4th 30 26 4 2.5
W+ = 9
B
Raw Values Log of Values
Well Year 1
(x)
Year 2
(y) Difference Rank Difference Rank
MW-1 1045 890 155 8 0.070 5
MW-2 352 241 111 7 0.165 8
MW-3 256 287 -31 -6 -0.050 -3
MW-4 132 128 4 2.5 0.013 1
MW-5 46 40 6 5 0.061 4
MW-6 28 30 -2 -1 -0.030 -2
MW-7 30 25 5 4 0.079 6
MW-8 10 14 -4 -2.5 -0.146 -7
W+ = 26.5 W+ = 24
WSRC-TR-2003-00331
December 31, 2004
2-21
Table 2.4
Critical Test Statistic Values for the Signed-Rank Statistic W+, n = 4 through 20
[ reject H0: at one-sided α when [ reject H0: at one-sided α when
W+ ≤ x (table entry) (small W) ] W+ ≥ x′ (table entry) (large W) ]
α-level α-level
n 0.025 0.05 0.1 n 0.025 0.05 0.1
4 0 4 10
5 0 2 5 15 13
6 0 2 3 6 21 19 18
7 2 3 5 7 26 25 23
8 3 5 8 8 33 31 28
9 5 8 10 9 40 37 35
10 8 10 14 10 47 45 41
11 10 13 17 11 56 53 49
12 13 17 21 12 65 61 57
13 17 21 26 13 74 70 65
14 21 25 31 14 84 80 74
15 25 30 36 15 95 90 84
16 29 35 42 16 107 101 94
17 34 41 48 17 119 112 105
18 40 47 55 18 131 124 116
19 46 53 62 19 144 137 128
20 52 60 69 20 158 150 141
WSRC-TR-2003-00331
December 31, 2004
2-22
the solute plume that reflect the efficacy of natural attenuation processes. An important note is
that none of the statistical tests described previously are tests for solute plume stability; none
presently exist. In evaluating solute plume stability, it is important to combine statistical results
with observations of the solute plume boundaries. The presence or absence of statistically
significant trends in concentration over time at monitoring points does not necessarily translate
into spatial changes in solute plume configuration. The lack of statistically significant trends in
concentration over time can generally be taken to represent a steady-state condition at a given
monitoring point, but this implies nothing about solute plume behavior. Consideration of results
at all the monitoring points is necessary.
In evaluating statistical results for concentration data, it is necessary to consider all of the
monitoring points in the solute plume. Depending on the dynamics of mass transfer from the
source and the specific natural attenuation processes involved, different portions of a solute
plume may exhibit different types of behaviors (Figure 2.4). No single monitoring point can
provide statistical results that are definitive since different monitoring points will be located in
different geochemical environments which impact the ambient degradation and transformation
processes.
A general consideration for the use of statistical methods in identifying trends and evaluating
solute plume behavior is that statistical significance does not necessarily imply real-world
significance and statistical test results can provide a false sense of assurance regarding
conclusions (Barden, 2003). It is important to always relate statistical results and evaluation
back to the physical problem in the field to ensure that the results are meaningful. Changes in
concentration and trends in concentration time series should be evaluated in the context of the
scientific understanding of the relevant natural attenuation processes. The point is to be able to
explain why the observed patterns (Figure 2.4) indicated by the statistical results are occurring.
The reason for a “statistically significant” change in concentrations is not provided by the
statistics themselves.
As an example, consider the results from tests for step trends. Comparison of concentration
data for two successive years does not imply that the result is meaningful. The fact that
concentrations in the second year are lower (or higher) than those from the previous year only
demonstrates a “statistically significant” difference. This does not imply that data from
subsequent years would produce the same result. A fundamental flaw in this sort of analysis is
that two years of data in most hydrogeologic settings is not a very large amount and the resulting
evaluation may not be substantive in the real world.
A consideration with the seasonal Mann-Kendall test is that trends of opposite sign in
different seasons may offset each other, giving the impression that no trends are present. This is
typically not a substantive concern since the point of the test is to determine overall trends in the
data series that may help to describe solute plume behavior. However, the individual seasonal
trends may be of importance for helping to unravel relationships between parameters, in which
case they could be examined individually in more detail.
Increasing
trend
Decreasing
trend No trend
Groundwater flow
D:\doe\report\lines of evidence\figures\Figure 2.4.ai
Westinghouse Savannah River MNA/EA Project
Figure 2.4
Solute Plume Behavior Illustrated by
Concentration Trends Over Time for Monitoring
Points in the Vicinity of the Source, Mid-Plume
and the Distal Portion of the Plume
Date: 12/17/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
2-23
●●●●●●●●●
●●●●●●
●●●
●●●●●●
●●●
A
DEF
GH I
BC
WSRC-TR-2003-00331
December 31, 2004
2-24
Similarly, it is common for the test for “homogeneity of stations” for monitoring points in the
entire solute plume to show no significant overall trend, even though trends are significant within
contiguous portions of the solute plume. Careful consideration of how monitoring points should
be grouped is necessary to evaluate portions of the solute plume. Graphical evaluation of the
data combined with a scientific understanding of the problem should be a good guide on how to
group contiguous monitoring points for statistical analysis.
2.2 CONTAMINANT/DAUGHTER PRODUCT AND GEOCHEMICAL DATA
Groundwater analytical data used to evaluate degradation products and geochemical
parameters are essential parameters for demonstrating that geochemical conditions in the aquifer
are appropriate for degradation. When the USEPA published the Technical Protocol for
Evaluating Natural Attenuation of Chlorinated Solvents (USEPA, 1998), the prevailing thought
was that biodegradation was the dominant mechanism working to remove tetrachloroethene
(PCE), trichloroethene (TCE), and carbon tetrachloride (CT) from groundwater. In recent years
it has become apparent that there are also several abiotic reductive dechlorination mechanisms
that work to remove chlorinated solvents from groundwater. All of these abiotic mechanisms of
reductive dechlorination rely on some form of reduced metal, typically iron. The iron can be in
the form of FeS (troilite), marcasite and pyrite (FeS2), magnetite, green rust (Fe2+-Fe3+ hydroxy
compounds), or other iron-bearing minerals. The formation of these materials in sedimentary
environments requires iron- and sulfate-reducing conditions; conditions which are brought about
by naturally occurring microorganisms degrading natural or anthropogenic carbon and utilizing
Fe(III) and sulfate as electron acceptors. Iron(II) and S2- are very reactive and iron-reducing and
sulfate-reducing bacteria are ubiquitous in the terrestrial subsurface, so if organic carbon, Fe(III),
and sulfate are present in sufficient quantities FeS will be formed.
In many cases abiotic reductive dechlorination reactions are biologically predicated (i.e., the
abiotic reaction will not occur without contemporaneous biologic activity). In these cases, the
degradation of PCE and TCE, whether abiotic or biological, and the abiotic degradation of
dichloroethene (DCE) appears to require oxidation-reduction potential (ORP) conditions that are
at least sulfate reducing. Thus, this line of evidence involves collecting data that helps determine
if the system is sufficiently reducing to allow degradation of the chlorinated VOCs listed
previously. Visual observations and mineralogical analyses can also be valuable for evaluating
the potential for abiotic degradation. In addition, this line of evidence should allow the
investigator to determine if the degradation mechanism is biological, abiotic, or a combination of
the two. This line of evidence includes the collection of both soil/sediment and groundwater
data.
2.2.1 Soil/Sediment Data
In addition to the data typically collected to evaluate the hydrogeology of a site which include
hydraulic conductivity, hydraulic gradient, and total and effective porosity, two types of
WSRC-TR-2003-00331
December 31, 2004
2-25
soil/sediment data are required to help determine the relative importance of natural attenuation
mechanisms. These include TOC and field analysis of reduced iron minerals.
2.2.1.1 Total Organic Carbon
Total organic carbon analyses are required to determine the sorptive capacity of an aquifer.
Sorption of organic compounds has two ramifications; 1) sorption causes slowing (retardation) of
an organic solute plume, and 2) sorption generally is a reversible processes which can cause low
concentrations of organic compounds to exist over a long period of time. Sorption is an
important component of the natural attenuation capacity of aquifers with high TOC
concentrations. Sorption generally is considered important when the TOC concentration is
greater that 0.1 percent by weight of the aquifer matrix.
2.2.1.2 Field Observation of Iron Minerals
Qualitative information on the occurrence of reduced Fe(II) and iron sulfide minerals can be
obtained by visual observation of samples collected during borehole drilling. Much of the color
associated with sediments is provided by the presence of iron-based minerals and their oxidation
state. In general there is a correlation between decreasing sediment size and increasing iron
concentration (Kennedy et al., 2000). For example, the presence of oxidized iron (Fe[III])
minerals can impart the following colors to sediments (Kennedy et al., 2000):
Mineral Color
Ferrihydrite - FeOH3 reddish brown
Göethite - FeO(OH) yellow to orange
Hematite - Fe2O3 red to maroon
Magnetite – Fe3O4 black
Conversely, sediments containing reduced iron such as iron sulfide minerals are typically gray to
black in color. Figure 2.5 is a photograph of reduced and oxidized iron minerals. Because FeS
and FeS2 cannot be differentiated visually in the field a simple field test to differentiate these
materials is as follows (from Kennedy et al., 2000):
Place a few grams of sediment into a small mouth bottle such as a 160 milliliter
serum bottle and add several milliliters of 6 N HCl. Cover the mouth of the bottle
for 30 seconds or so. If FeS is present, then the distinctive “rotten egg” odor
associated with H2S will be detected. Because the gas being produced by this
reaction is toxic, care should be taken to avoid inhaling too much of the gasses
produced by this test. Use of such a test must be addressed through a site specific
health and safety plan.
D:\doe\report\lines of evidence\figures\Figure 2.5.ai
Westinghouse Savannah River MNA/EA Project
Figure 2.5
Photograph of Reduced
and Oxidized Iron Minerals
Date: 12/17/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
Black FeS enriched
sand mixed with
gravel (originally
red/orange colored)
Black FeS enriched
sand mixed with
gravel (originally
red/orange colored)
Base confining clay
(no change)
Base confining clay
(no change)
Source: Kennedy and Everett (2004)
2-26
WSRC-TR-2003-00331
December 31, 2004
2-27
In addition to these tests, a magnet can be used in the field to determine if magnetic minerals
such as magnetite which facilitate abiotic degradation are present.
2.2.2 NAPL Data
NAPL analyses can be helpful in establishing which compounds impact a site and the strength
of the source. At sites where it can be demonstrated that compounds appearing in groundwater
were not present or used at a site these compounds are likely daughter products. It is possible
that vinyl chloride (VC) and some of the DCE isomers can be primary contaminants in some
groundwater systems. However, VC is not normally present as a primary contaminant in solvent
spills. The reasons for this are; 1) VC was not used as a solvent, and 2) VC is a gas at
temperatures as low as 15oC. Thus, the presence of VC in groundwater associated with a
chlorinated ethene spill is strong evidence of reductive dechlorination. Also, cis-DCE (rather
than trans-DCE or 1,1-DCE) is usually produced from the reductive dechlorination of TCE (both
abiotically and biologically). As a rule of thumb, if the ratio of cis-1,2-DCE to trans-1,2-DCE
plus 1,1-DCE is greater than about 5:1, then the observed DCE is likely the result of degradation
of TCE and/or PCE. When 1,1,1-TCA is present or is suspected to have been present, 1,1-DCE
may be present as the result of the abiotic dehydrohalogenation of this compound. This must be
taken into account when evaluating the ratio of DCE isomers. Based on these concepts, VC
and/or cis-DCE are usually reliable indicators of reductive dechlorination. In addition, the
presence of ethene and ethane typically are indicative of reductive dechlorination. However,
these products are extremely transitory and low detection limits typically are required to identify
and quantify these compounds.
2.2.3 Groundwater Analytical Data
There are several groundwater analyses that can be used to evaluate and quantify natural
attenuation and enhanced remediation. These analyses are presented in Table 1.3 and discussed
in the sections that follow.
2.2.3.1 Daughter Products
Concentrations of chlorinated solvents and their degradation (daughter) products give a direct
indication of the presence or absence of degradation (both abiotic and biological) processes. In
many cases the production of cis-DCE, VC, and chloride ions along aquifer flowpaths is direct
evidence of degradation. For example, if TCE was the only contaminant disposed of at a site,
then any cis-DCE or VC present at the site must have come from the degradation of the parent
TCE.
2.2.3.2 Chloride
During degradation of chlorinated hydrocarbons dissolved in groundwater, chloride is
released into the groundwater, resulting in the accumulation of chloride. This results in chloride
WSRC-TR-2003-00331
December 31, 2004
2-28
concentrations in groundwater in the contaminant plume that are elevated relative to background
concentrations. In aquifers with low background concentrations of chloride, the concentration of
this material in the solute plume can be seen to increase as chlorinated solvents are degraded.
Elemental chlorine is the most abundant of the halogens. Although chlorine can occur in
oxidation states ranging from Cl- to Cl+7, the chloride form (Cl-) is the only form of major
significance in natural waters (Hem, 1985). Chloride forms ion pairs or complex ions with some
of the cations present in natural waters, but these complexes are not strong enough to be of
significance in the chemistry of fresh water (Hem, 1985). The chemical behavior of chloride is
neutral. Chloride ions generally do not enter into oxidation-reduction reactions, form no
important soluble complexes with other ions (unless the chloride concentration is extremely
high), do not form salts of low solubility, are not significantly adsorbed on mineral surfaces, and
play few vital biochemical roles (Hem, 1985). Thus, physical processes control the migration of
chloride ions in the subsurface. Kaufman and Orlob (1956) conducted tracer experiments in
groundwater, and found that chloride moved through most of the soils/sediments tested more
conservatively (i.e., with less retardation and loss) than any of the other tracers tested. Because
of the neutral chemical behavior of chloride, it can be used as a conservative tracer to estimate
degradation rates.
2.2.3.3 Dissolved Oxygen
The concentration of dissolved oxygen in an aquifer is a very important parameter for
determining if the system is capable of supporting the degradation of chlorinated solvents in the
terrestrial subsurface. If dissolved oxygen is present at concentrations greater than about 0.5
mg/L (assuming accurate measurements) then reductive dechlorination (biological or abiotic)
will not occur to any great extent.
Dissolved oxygen is the favored electron acceptor used by microbes for the biodegradation of
most forms of organic carbon. Strictly anaerobic bacteria generally cannot function at dissolved
oxygen concentrations greater than about 0.5 mg/L and hence Fe(III) reduction, sulfate
reduction, methanogenesis, and reductive dechlorination (biological or abiotic) cannot occur.
This is why it is important to have a source of carbon in the aquifer that can be used by aerobic
microorganisms as a primary substrate. During aerobic respiration, dissolved oxygen
concentrations decrease. The absence of dissolved oxygen is a prerequisite for iron- and sulfate-
reduction.
Dissolved oxygen measurements should be taken during well purging and immediately before
sample acquisition using a direct-reading meter. Each of these measurements should be recorded
during well purging. Because many well purging techniques can allow aeration of groundwater
samples, it is important to minimize the potential for aeration during well purging. Because of
the difficulty in obtaining accurate dissolved oxygen measurements, especially when the
concentration falls below about 1 milligram per liter, these measurements should be used in a
qualitative manner. One use of dissolved oxygen measurements is during well purging.
Stabilization of dissolved oxygen concentrations, in conjunction with pH, temperature,
WSRC-TR-2003-00331
December 31, 2004
2-29
conductivity, and ORP can be useful during well purging to determine when the well has been
purged sufficiently to give representative samples.
If the aquifer is aerobic, then the measurement of reduced dissolved gasses should not be
undertaken. The reason for this is that the presence of dissolved oxygen precludes the formation
of these gasses which include hydrogen sulfide, methane, ethene, ethane, and hydrogen.
2.2.3.4 Nitrate
After dissolved oxygen has been depleted in the microbiological treatment zone, nitrate is
used as an electron acceptor for anaerobic biodegradation of organic carbon via denitrification.
During denitrification, nitrate concentrations measured in groundwater decrease. Thus, nitrate
concentrations below background in areas with dissolved contamination provide evidence for
denitrification. Nitrate is an electron acceptor that competes with dehalorespiration. The
absence of nitrate is a prerequisite for iron- and sulfate-reduction so it is important that this
compound is absent in groundwater for biological and abiotic reactions to proceed.
2.2.3.5 Fe(II)
When Fe(III) is used as an electron acceptor during anaerobic biodegradation of organic
carbon, it is reduced to Fe(II), which is somewhat soluble in water. Fe(II) concentrations are
used as an indication that anaerobic degradation of organic carbon has occurred via Fe(III)
reduction. The presence of Fe(II) (and sulfide) is required in order for many of the abiotic
reactions described elsewhere in this document to occur. In addition, Bradley and Chapelle
(1996; 1997) have shown that VC and DCE can be biologically oxidized under iron-reducing
conditions. Fe(III) is an electron acceptor that competes with dehalorespiration.
2.2.3.6 Sulfate and Sulfide
Sulfate is used as an electron acceptor for anaerobic biodegradation during sulfate reduction
wherein sulfate (SO42-) is reduced to sulfide (HS- or H2S). During this process, sulfate
concentrations measured in groundwater decrease and sulfide is produced. The sulfide produced
during sulfate reduction is very reactive and in most cases is quickly complexed with Fe(II).
From the standpoint of chlorinated solvent degradation, sulfate reduction is important for two
reasons; 1) reductive dechlorination caused by biological processes does not become efficient
until the dominant terminal-electron accepting process is sulfate reduction or methanogenesis,
and 2) sulfate reduction is important for abiotic mechanisms of reductive dechlorination because
it results in the production of sulfide. High sulfate concentrations will likely have the following
two ramifications:
1) They will reduce the efficiency of biological reductive dechlorination because sulfate is a
competing electron acceptor; and
WSRC-TR-2003-00331
December 31, 2004
2-30
2) They will increase the efficiency of abiotic reductive dechlorination, especially if
appreciable amounts of Fe(II) are present.
2.2.3.7 Methane
As implied by the name, methanogenesis results in the production of methane during the
biodegradation of organic carbon. The presence of methane in groundwater is indicative of
strongly reducing conditions and biologically mediated reductive dechlorination is typically very
efficient under these conditions. Analysis of methane concentrations in groundwater should be
conducted by a qualified laboratory. It is important that the detection limit for methane be on the
order of 1 µg/L, especially when evaluating the natural attenuation or enhanced remediation of
chlorinated solvents.
The presence of methane generally is indicative of a Type I/Type II Environment where
reductive dechlorination to VC and then to ethene/ethane is likely. If no VC is present then
abiotic reactions should be evaluated. Methane can also be transported by advective
groundwater flow. Because of this, its presence in a groundwater does not ensure that the
immediate environment is methanogenic; only that methanogenic conditions exist in the vicinity.
Evaluating the presence of methane in concert with the other geochemical indicators (e.g., Fe[II]
and SO42-) is essential.
2.2.3.8 Ethene/Ethane
Ethene and ethane are the end products of both biological and abiotic reductive
dechlorination. Because these compounds are extremely transitory their concentrations typically
remain low with concentrations at sites with active reductive dechlorination on the order of
hundreds of micrograms per liter.
2.2.3.9 Acetylene
Acetylene is a product of the abiotic dechlorination of chlorinated aliphatic hydrocarbons
(e.g., PCE and TCE) by iron sulfides. Although the exact pathway has not been fully
determined, it is thought that the pathway for TCE oxidation is via the cis-dichlorovinyl radical
directly to acetylene (Butler and Hayes, 1999). Because acetylene is unstable and very transient,
it is hard to detect. Researchers are currently working on preservation methods that should allow
the detection of acetylene at low concentrations.
2.2.3.10 Oxidation-Reduction Potential
The oxidation-reduction potential (ORP), or “redox” potential, of groundwater is a measure of
the electrochemical potential associated with transfer of electrons from one substance to another
that indicates the redox condition of the system. Oxidation-reduction reactions involving organic
compounds (natural or anthropogenic) in groundwater are usually biologically mediated, and the
ORP of a groundwater system depends upon and influences rates of degradation (both biological
WSRC-TR-2003-00331
December 31, 2004
2-31
and abiotic). The ORP of groundwater as measured with field meters generally ranges from -400
millivolts (mV) to 800 mV. In general, the lower the ORP of groundwater the more reducing the
system is, and the more likely that reductive dechlorination will be efficient.
The measured ORP value depends upon the potential of the reference electrode used in the
measurement. The “Eh” is a measure of ORP relative to the “standard hydrogen electrode” that
is assigned a potential of 0 mV and is used in theoretical redox calculations. Measured ORP
values can be converted to equivalent Eh values by:
measured ref .electrode
Eh
=
+EE.
The silver-silver chloride (Ag-AgCl) reference electrode used in most field ORP meters has E =
222.3 mV at 25ºC. Another commonly used measure of the redox condition is “pE” which is
defined as the negative common logarithm of the “free electron” activity (pE = -log10[e-]). The
pE is a purely theoretical construct that is used to visualize relationships among redox sensitive
species. The pE is related to the Eh by the Nerst factor:
[
]
Eh Eh(mV)
pE @ 25º C
2.303 59.157
==
F
RT
where F is the Faraday constant (9.64846x104 C/mol.), R is the Universal Gas Constant (8.3144
J/mol-ºK), and T is temperature (ºK).
In general, ORP readings should only be used on a qualitative basis in comparing the relative
ORP values. ORP measurements assume that the system is at equilibrium, and this is rarely the
case in most groundwater systems (Deutsch, 1997; Stumm and Morgan, 1996). Complicating
factors include problems with the measurement electrodes, kinetic constraints on redox processes
resulting in redox disequilibria, and mixing of waters with different potentials. If the accuracy of
ORP measurements are of interest, the measured values should be verified with measured
relative concentrations of the members of one or more redox couples (Stumm and Morgan,
1996).
ORP measurements can be used to provide real-time data on the location of the contaminant
plume, especially in areas undergoing anaerobic biodegradation. Mapping the ORP of
goundwater while in the field helps the field scientist to determine the approximate location of
the contaminant plume. To map the ORP of the groundwater while in the field, it is important to
have at least one ORP measurement (preferably more) from a well located upgradient from, or
peripheral to, the plume. ORP measurements should be taken during well purging and
immediately before sample acquisition using a direct-reading meter. Because most well purging
techniques can allow aeration of collected groundwater samples (which can affect ORP
measurements), it is important to minimize potential aeration by using a flow-through cell as
outlined previously.
WSRC-TR-2003-00331
December 31, 2004
3-1
SECTION 3
SUPPLEMENTAL LINES OF EVIDENCE
In many cases, evaluation of the primary lines of evidence discussed in the preceeding section
will be adequate to evaluate the natural attenuation of a chlorinated solvent solute plume.
However, at some sites additional evidence may be useful for deducing the natural attenuation
capacity of an aquifer and the effects of enhanced remediation. These lines of evidence can be
particularly useful for those portions of the plume where the fate of VOCs is not apparent. For
example, in some cases degradation can be enhanced through bioaugmentation or the addition of
materials that stimulate the production of iron sulfide minerals to aid abiotic degradation. In
many cases recent studies suggest that biologically predicated abiotic degradation may be more
efficient than dehalorespiration. According to one researcher, abiotic degradation can occur with
half-lives for PCE/TCE as low as 19 days (Dr. Lonnie Kennedy, personal communication).
The supplemental lines of evidence discussed in this section can, in some cases, provide
critical information to support monitored natural attenuation and enhanced attenuation
approaches. However, evaluation of these supplemental lines of evidence will result in varying
amounts of additional effort and cost, both in sampling and analysis and in data interpretation.
Therefore, it is important that the investigator consider how the information provided by the
supplemental data will contribute to the necessary understanding of natural attenuation processes
at a site.
3.1 SUPPLEMENTAL GROUNDWATER GEOCHEMICAL DATA
In some cases additional geochemical data can be useful for evaluating the predominant
geochemical environment in groundwater. Table 3.1 summarizes some of the supplemental
groundwater parameters that may be useful for evaluating natural attenuation and enhanced
remediation.
3.1.1 Mn(II)
When Mn(IV) is used as an electron acceptor during anaerobic biodegradation of organic
carbon, it is reduced to Mn(II). Mn(II) concentrations can thus be used as an indicator that
anaerobic degradation of organic carbon has occurred via Mn(IV) reduction. Changes in Mn(II)
concentrations inside the contaminant plume versus background concentrations can be used to
estimate the mass of contaminant that has been biodegraded by Mn(IV) reduction. Mn(IV) is an
electron acceptor that competes with iron reduction and dehalorespiration.
WSRC-TR-2003-00331
December 31, 2004
3-2
Table 3.1
Supplemental Groundwater Data for Evaluating Natural Attenuation and Enhanced Remediation
Analysis
Method/Reference
Comments
Data Use
Sample Volume, Sample Container,
Sample Preservation
Field or
Fixed-Base
Laboratory
Geochemical analyses
Manganese Colorimetric Hach
Method
Filter with 0.45
micron inline
filter.
May indicate an anaerobic degradation
process due to depletion of oxygen, and
nitrate. Interferences can occur if hydrogen
sulfide and high concentrations of calcium
are present.
Collect 100 mL of water in a
headspace-free container to eliminate
introduction of oxygen and analyze as
soon as possible.
Field
Alkalinity Hach digital titrate Field filter with
0.45 micron inline
filter. Carbonate
alkalinity only
significant at pH
>8.5.
General water quality parameter used (1) to
measure the buffering capacity of
groundwater, and (2) as a marker to verify
that all site samples are obtained from the
same groundwater system.
Collect 100 mL of water in glass
container. Analyze as soon as
possible.
Field
Dissolved
Inorganic
Carbon
(DIC)
E415.1 Filter in the field
with 0.45 micron
inline filter.
Minimize aeration
of sample and fill
sample container
completely to
avoid loss of CO2.
An increase of DIC above background
concentrations provides a footprint in
groundwater that has been remediated by
biological processes. Carbon dioxide is the
most universal end product of chlorinated
hydrocarbon biodegradation. DIC is the
sum of dissolved carbon dioxide, carbonic
acid, bicarbonate and carbonate.
Collect 1x250 mL glass amber
container, preserve with H2SO4 and
cool to = 4 oC.
Fixed-base
laboratory
Anions - Cl,
Fl, and SO4,
NO3, HCO3,
CO3, Br
Method E300 (Cl,
Fl, SO4, NO3, and
Br). Method
E310.1 (HCO3 and
CO3)
Filter in the field
with 0.45 micron
inline filter.
Can be used graphically (e.g., Piper and
Stiff diagrams) with cations to identify
different hydrogeologic units and identify
areas impacted by contamination.
Collect 1x1 Liter poly container and
cool to = 4 oC.
Fixed-base
laboratory
Cations –
Ca, Mg, K,
and Na, Mn,
Fe
SW6010 Filter in the field
with 0.45 micron
inline filter.
Can be used graphically with anions to
identify different hydrogeologic units and
identify areas impacted by contamination.
Collect 1x500 mL poly container,
preserve with HNO3 and cool to = 4
oC.
Fixed-base
laboratory
WSRC-TR-2003-00331
December 31, 2004
3-3
Table 3.1 (Concluded)
Supplemental Groundwater Data for Evaluating Natural Attenuation and Enhanced Remediation
Analysis
Method/Reference
Comments
Data Use
Sample Volume, Sample
Container, Sample Preservation
Field or Fixed-Base
Laboratory
Hydrogen Equilibration with
gas in the field
analyzed with a
reducing gas detector
in the lab.
Supplemental
specialized analysis
to be completed on
select wells.
Determine current terminal electron
accepting process and if sufficient
hydrogen is available for reductive
dechlorination.
Sampled at well head. Requires
the production of 100mL per
minute of water for 30 minutes.
Fixed-base laboratory
Dissolved Organic
Carbon
E415.1 Field filter with 0.45
micron inline filter.
Minimize aeration
and fill sample
container completely.
Used to classify plume and to
evaluate the potential for biologic and
biologically predicated abiotic
degradation.
Collect 1x250 mL glass amber
container, preserve with H2SO4
and cool to 4 oC.
Fixed-base laboratory
VFAs Ion chromatography. Can be a useful
indicator of microbial
metabolism of added
substrate.
Biomarkers of anaerobic metabolism.
Anaerobic bacteria produce these
compounds by fermentation.
Contact Laboratory. Fixed-base laboratory
Stable Isotopes Sherwood Lollar et
al. Org Geochem.
1999, 30, 813-820
May need more than
one quarter of data
and interpretation
may become
complicated if there
is more than one
source.
Helps elucidate biotic vs. abiotic
dechlorination pathways.
Collect 1 x44mL VOA vial no
headspace and cool to 4 oC.
Fixed-base laboratory
Microbiological Analyses
PLFAs White et al. (1997) PLFA data can be
readily correlated
with contaminant and
geochemical trends.
Provides microbial biomass,
community structure and
physiological status data.
Collect 1 to 2 Liters of
groundwater in a sterile
widemouth poly bottle and cool to
= 4 oC.
Fixed-base laboratory
DGGE Muyzer et al. AEM
1993, 59:695-700.
Mainly for use in
forensic or failure
analyses.
Identifies most dominant
microorganisms in the groundwater.
Collect 500 to 1000 mL of
groundwater in a sterile
widemouth poly bottle and cool to
4 oC.
Fixed-base laboratory
WSRC-TR-2003-00331
December 31, 2004
3-4
3.1.2 Carbon Dioxide
Metabolic processes operating during biodegradation of organic compounds lead to the
production of carbon dioxide (CO2). However, CO2 released into groundwater rapidly reacts to
form carbonic acid (H2CO3). Accurate measurement of the amount of carbon dioxide produced
during biodegradation is difficult because carbonate in groundwater (measured as alkalinity)
serves as both a source and sink for free carbon dioxide. If the carbon dioxide produced during
metabolism is not completely removed by the natural carbonate-buffering system of the aquifer,
carbon dioxide concentrations higher than background may be observed. Comparison of
empirical data to stoichiometric calculations can provide estimates of the degree of
microbiological activity and the occurrence of in situ mineralization of contaminants provided
the carbonate buffering is accounted for. However CO2 measurements alone typically are
uninformative.
3.1.3 Alkalinity
Biologically active portions of a solute plume typically can be identified by an increase in
alkalinity. This increase in alkalinity is brought about by the production of carbon dioxide
during the biodegradation of organic carbon. Alkalinity results from the presence of hydroxides,
carbonates, and bicarbonates of cations such as calcium, magnesium, sodium, and potassium.
These species result from the dissolution of rock (especially carbonate rocks), the transfer of
carbon dioxide from the atmosphere, and respiration of microorganisms. Alkalinity is important
in the maintenance of groundwater pH because it buffers the groundwater system against acids
generated during both aerobic and anaerobic biodegradation. In general, areas with organic
carbon that has been reduced exhibit a total alkalinity that is higher than that seen in those areas
with low organic carbon concentrations. This is expected because the microbially mediated
reactions causing biodegradation of organic carbon cause an increase in the total alkalinity in the
system. Changes in alkalinity are most pronounced during aerobic respiration, denitrification,
Fe(III) reduction, and sulfate reduction, and less pronounced during methanogenesis (Morel and
Hering, 1993).
3.1.4 Dissolved Inorganic Carbon
Dissolved inorganic carbon (DIC) is a measure of the inorganic carbon species in solution,
including carbon dioxide (CO2), carbonic acid (H2CO3), bicarbonate (HCO3-), and carbonate
(CO32-). Similar to alkalinity, biologically active portions of a solute plume typically exhibit an
increase in DIC due to the production of CO2 by microbial metabolism of organic carbon.
In general, alkalinity (as CaCO3) measurements are predominantly due to carbonate species in
solution. Therefore, the measurement of DIC is typically redundant if alkalinity is measured and
if carbonate speciation data is needed, it can be calculated from the alkalinity measurement.
WSRC-TR-2003-00331
December 31, 2004
3-5
3.1.5 Major and Minor Ions
In general, data on major and minor ion concentrations in groundwater does not contribute
substantively to the evaluation of natural attenuation for chlorinated solvents, with the exception
of the geochemical parameters previously discussed (e.g., Cl-, NO3-, Fe2+, Mn2+, SO42-, HCO3-
/CO32-). However, since chlorinated hydrocarbon solute plumes are superposed on, and interact
with, the ambient geochemical system, in some cases knowledge of the major and minor ion
chemistry in groundwater can provide useful information for evaluating natural attenuation.
Major ions include constituents that typically are present in groundwater at concentrations
greater than 1 mg/L. The major cations include sodium (Na+), potassium (K+), calcium (Ca2+),
and magnesium (Mg2+) and the major anions include chloride (Cl-), suflate (SO42-), the carbonate
species (HCO3- and CO32-), and sometimes nitrate (NO3-). A major nonionic constituent is
silica, that generally is present as the neutral dissolved species H4SiO4°. The minor ions include
constituents that typically are present in groundwater at concentrations between 1 and 0.01 mg/L.
The minor cations include iron (Fe2+ and Fe3+), manganese (Mn2+), aluminum (Al3+), etc., and
the minor anions include fluoride (F-), bromide (Br-), phosphate (PO42-), etc.
The primary application of data for major and minor ions in the evaluation of natural
attenuation of chlorinated solvents is in cases where the bulk groundwater geochemistry is
important for characterizing and understanding geochemical processes. Such situations may
include evaluation of the formation and stability of mineral phases, detailed examination of the
behavior of inorganic constituents, or the design of enhanced remediation systems. Major ions
can be plotted on Stiff and Piper diagrams and used to graphically depict different hydrogeologic
units and areas impacted by contamination. Geochemical models such as MINTEQ and
PHREEQ that require concentrations of major ions as input parameters can be useful for
evaluating groundwater geochemistry. In addition, minor ion data may be required for these
models if they are important to the reactions in the system. A detailed discussion of the
application of major and minor ion chemistry data is complex and beyond the scope of this
document, and the interested reader is referred to Deutsch (1997) for a practical overview.
3.1.6 Dissolved Hydrogen
Concentrations of dissolved hydrogen can be used to evaluate terminal electron-accepting
processes in groundwater systems (Lovley and Goodwin, 1988; Lovley et al., 1994; Chapelle et
al., 1995). Because each terminal electron-accepting process (TEAP) has an associated
characteristic hydrogen concentration range, hydrogen concentrations can be an indicator of the
predominant TEAP. These characteristic ranges are given in Table 3.2.
Oxidation-reduction potential (ORP) measurements are based on the concept of
thermodynamic equilibrium and, within the constraints of that assumption, can be used to
evaluate TEAPs in groundwater systems. The use of dissolved hydrogen to classify the system is
WSRC-TR-2003-00331
December 31, 2004
3-6
Table 3.2
Range of Hydrogen Concentrations for a Given Terminal Electron-Accepting Process
Dissolved Hydrogen (H2) Concentration
Terminal Electron
Accepting Process nmol/L 1 atm
2 µg/L
Denitrification < 0.1 < 1.3 x 10-7 < 0.2 x 10-3
Iron (III) Reduction 0.2 to 0.8 0.26 - 1.0 x 10-6 0.4 - 1.6 x 10-3
Sulfate Reduction 1 to 4 1.3 - 5.0 x 10-6 2.0 - 8.0 x 10-3
Methanogenesis 5 to 20 63 - 250 x 10-6 10 - 40 x 10-3
1 nanomole per liter (nanomolar)
2 In gas phase in equilibrium with water containing dissolved hydrogen
Source: Adapted from Lovley et al. (1994) and Chapelle et al. (1995)
WSRC-TR-2003-00331
December 31, 2004
3-7
based on the ecological concept of interspecies hydrogen transfer by microorganisms and, within
the constraints of that assumption, can also be used to evaluate TEAPs. These methods,
therefore, are fundamentally different. A direct comparison of these methods (Chapelle, 1997)
has shown that while oxidation-reduction potential measurements were effective in delineating
oxic from anoxic groundwater, they could not reliably distinguish between nitrate-reducing,
Fe(III)-reducing, sulfate-reducing, or methanogenic zones in an aquifer. In contrast, the
measurement of dissolved hydrogen could readily distinguish between different anaerobic zones.
At those sites where distinguishing between different anaerobic processes is important (such as
at sites contaminated with chlorinated solvents), hydrogen measurements can be useful for
delineating the distribution of terminal electron-accepting processes.
In practice, it is preferable to interpret hydrogen concentrations in the context of electron
acceptor (dissolved oxygen, nitrate, Mn(IV), Fe(III), sulfate) availability and the presence of the
final products (Mn(II), Fe(II), hydrogen sulfide, methane) of microbial metabolism (Chapelle et
al., 1995). For example, if sulfate concentrations in groundwater are less than 0.5 mg/L,
methane concentrations are greater than 0.5 mg/L, and hydrogen concentrations are in the 5-20
nM range, it can be concluded with a high degree of certainty that methanogenesis is the
predominant terminal electron-accepting process in the aquifer. Similar logic can be applied to
identifying denitrification (presence of nitrate, hydrogen <0.1 nM), Fe(III) reduction (production
of Fe(II), hydrogen 0.2 to 0.8 nM), and sulfate reduction (presence of sulfate, production of
sulfide, hydrogen 1-4 nM).
Chapelle et al. (1997) compare three methods for measuring hydrogen concentrations in
groundwater; a downhole sampler, a gas stripping method, and a diffusion sampler. The
downhole sampler and gas stripping methods gave similar results. The diffusion sampler
appeared to overestimate hydrogen concentrations. Of these methods, the gas stripping method
is better suited to field conditions because it is faster (approximately 30 minutes for a single
analysis as opposed to two hours for the downhole sampler and eight hours for the diffusion
sampler), the analysis is easier (less sample manipulation is required), and the data computations
are more straightforward (hydrogen concentrations need not be corrected for water sample
volume) (Chapelle et al. 1997). At least one commercial laboratory uses the gas stripping
method (called the “bubble strip” method) for hydrogen sampling and analysis.
3.1.7 Dissolved Organic Carbon
Dissolved organic carbon (DOC) is an operationally defined parameter that measures the
organic carbon that passes through a 0.45 µm filter. In general, organic carbon dissolved in
groundwater is more readily available to microbes and DOC can be used as an indicator of the
availability of organic carbon that can maintain reducing conditions and provide fermentable
substrates needed to support biological reductive dechlorination.
Naturally occurring DOC consists of humic and fulvic acids, sugars, fatty acids, low
molecular weight organic acids, and alkanes. Anthropogenic organic carbon can include organic
acids in landfill leachate, fuel hydrocarbons, and other reduced organic compounds.
WSRC-TR-2003-00331
December 31, 2004
3-8
Measurements of DOC concentrations are more relevant for solute plumes where the reduced
organic substrates are not “contaminants” that are otherwise sampled and analyzed. However,
even in these cases DOC measurements will include organic compounds that are not chemicals
of concern and provide a more representative indication of substrate availability.
Changes in DOC concentrations may be correlated to changes in organic substrate
availability. A statistically significant decline over time or space may indicate that readily
available organic substrates have become depleated. This could signal a change to conditions
that are less conducive to biological or abiotic reductive dechlorination.
3.1.8 Volatile Fatty Acids (VFAs)
Short-chain organic acids, or “volatile fatty acids” (VFAs), are end products of the
fermentation of organic matter and are important substrates for anaerobic respiration, such as for
sulfate-reducing bacteria, in natural settings. Also, the biodegradation of simple hydrocarbons
almost always produces organic acid intermediates, particularly under anaerobic conditions. The
major volitle fatty acids are acetate, propionate, formate, butyrate, lactate and pyruvate. These
are biomarkers of anaerobic bacterial metabolism and accumulation of VFAs is indicative of the
availability of fermentable organic matter that also yields hydrogen (H2) which is essential for
dehalorespiration.
The VFAs are analyzed by ion chromatography methods; generally by ion exclusion
chromatography (Bradley et al., 1993; Hansen, 1998). Analysis of VFAs can provide an
indication of the availability of fermentation substrates that can support reductive dechlorination
(dehalorespiration) and, potentially, information on the stability of redox conditions in a plume
(Christensen et al., 2000). The use of VFAs as redox indicators is equivocal, but shifts in the
predominant TEAP can yield transient concentration peaks in VFAs that could be used as an
indicator of a change in redox conditions (Christensen et al., 2000).
Enhanced attenuation approaches for chlorinated solvents in groundwater that use substrate
additions (e.g., vegetable oil or HRC®) are intended to augment the substrates available for
fermentation to yield hydrogen (H2). The presence, or change in concentrations, of various
VFAs characteristic of the substrate after substrate addition provides an indication that
fermentation in proceeding and the VFAs can also serve as a tracer to monitor where the treated
water flows.
Similar to VOCs, VFAs are subject to evaporation losses as well as degradation under both
aerobic and anaerobic conditions. Groundwater samples collected for analysis of VFAs must be
preserved immediately and handled similar to VOC samples. Preservation of VFA samples by
the addition of 0.2% chloroform and immediate freezing in polypropylene vials has been found
to be an appropriate method (Hansen, 1998; Albrechtsen et al., 1999) and also prevents
generation of additional VFAs in samples containing dissolved organic matter (e.g., humic and
fulvic acids) (Christensen et al., 2000).
WSRC-TR-2003-00331
December 31, 2004
3-9
3.1.9 Stable Isotopes
Analysis of stable isotope ratios between parent and daughter compounds can be useful for
identifying the biodegradation of chlorinated compounds because isotopic fractionation
commonly occurs during biodegradation. This fractionation results in a characteristic pattern of
isotope ratios between parent compounds and daughter products. For the chlorinated ethenes,
non-degradative subsurface processes such as dissolution, sorption, and volatilization do not
involve isotopic fractionation greater than 0.5‰, and this is the typical accuracy and
reproducibility of continuous flow isotope analysis techniques (Slater et al., 2001).
Hunkeler et al., (1999) studied the occurrence of stable carbon isotope (13C/12C) fractionation
during the reductive dechlorination of PCE to ethene in the field and in the laboratory using
aquifer material from the same site located in Toronto, Ontario, Canada. According to these
researchers, all dechlorination steps in the microcosm were accompanied by stable carbon
isotope fractionation with similar results for the field study. In the microcosm study the largest
fractionation occurred during dechlorination of cis 1,2-DCE and VC, resulting in a large
enrichment of 13C in the remaining cis-1,2-DCE and VC. Stable carbon isotope ratios (δ13C) of
cis-1,2-DCE and VC increased from -25.7 to 1.5‰ and -37 to -2.5‰, respectively. The δ13C of
ethene was initially -60.2‰ and approached the δ13C of the added PCE (-27.3‰) as
dechlorination came to completion. Based on their work, these researchers conclude that strong
enrichment of 13C in cis-1,2-DCE and VC during microbial dechlorination may serve as a
powerful tool to monitor the last two steps of dechlorination. These steps frequently determine
the rate of dechlorination of chlorinated ethenes at field sites where degradation is occurring.
3.2 MINERALOGICAL ANALYSES
The purpose of mineralogical analysis as a line of evidence for natural attenuation of
chlorinated solvents is to evaluate the presence of reduced iron minerals that are known to
facilitate abiotic degradation of chlorinated compounds. These include iron sulfides, such as
troilite and mackinawite (FeS), marcasite and pyrite (FeS2), iron oxides and hydroxides, such as
magnetite (Fe3O4) and green rust (Fe2+-Fe3+ hydroxy compounds), and other iron-bearing
minerals. Mineralogical analysis can also provide information on the amount of Fe(III) available
for reduction to Fe(II). These analyses can be used to supplement and refine the macroscopic
indicators of iron mineralogy described in Section 2.2.1.2.
Characterization of aquifer mineralogy can involve the identification of individual minerals,
identification and quantification of the bulk content of minerals or species in aquifer sediment,
and identification and quantitation of “reactive” fractions for minerals and species that are
available for a given reaction. There are numerous approaches that can be used to determine the
mineral composition of soil, sediment, and rock (Amonette and Zelazny, 1994; Amonette, 2002)
and approaches specific to iron minerals are discussed in Bigham et al. (2002) and Ribbe (1974).
These approaches to mineralogical analysis generally involve standard petrographic methods and
bulk elemental and mineral analysis techniques. Some of the most readily applied and useful
techniques relavent to evaluation of natural attenuation are summarized in Table 3.3 and briefly
discussed below. Other methods, such as 57Fe Mössbauer spectroscopy, may have application in
WSRC-TR-2003-00331
December 31, 2004
3-10
the evaluation of iron-bearing minerals, but are not as readily available as the techniques
discussed here. It is best to contact a qualified laboratory for these analyses.
Some iron sulfides, such as troilite and mackinawite (FeS), pyrrhotite (FeS0.9) and greigite
(Fe3S4), and the green rusts are highly unstable under oxidizing conditions and care should be
exercised in sampling to prevent contact with atmospheric oxygen. Surface oxidation can occur
within 30 minutes or less of exposure to aerobic conditions (Fanning et al., 2002). Anaerobic
sampling techniques for aquifer sediments are described in Kennedy et al. (1998, 2000).
3.2.1 Petrography
Petrographic evaluation of aquifer solids using optical methods is a technique that can provide
useful information on the mineral composition of the aquifer matrix and the distribution of
minerals in the aquifer at the grain scale. Optical petrography is a fairly simple, observational
technique that allows the user to see features of the aquifer matrix directly. It can involve the use
of thin sections and/or grain separates, and can involve the use of either a petrographic
microscope using transmitted and reflected light or with a simple binocular microscope using
reflected light. Minerals are identified by their optical characteristics. Good overviews of the
methods and application of optical petrography can be found in Cady et al. (1986) and Drees and
Ransom (1994), as well as texts on optical mineralogy (e.g., Kerr, 1977).
Transmitted light microscopy with a petrographic microscope is used to examine non-opaque
minerals, microcrystalline aggregates, and “amorphous” materials with plane-polarized and
cross-polarized light. Reflected-light microscopy is used to examine opaque minerals. The
methodology is similar to that used in ore petrography and identification is based on reflectance
properties under plane-polarized light (Spry and Gedlinske, 1987). Often, the same specimen
that is viewed using the light microscope can be analyzed using the advanced x-ray and ion
microprobe techniques discussed below.
Petrographic evaluation of thin sections is useful for determining both the identity and
distribution of iron minerals in the aquifer matrix. This is particularly useful in providing a
visual representation that retains the original spatial relationships of the minerals. Thin-section
petrography also is applicable to both poorly crystallized and “amorphous” components, such as
some iron oxides, hydroxides and sulfides, and can be complemented with selective extraction
techniques. Quantification can be achieved through standard point-counting techniques or
through image analysis (Drees and Ransom, 1994). The construction of thin-sections from
unconsolidated aquifer material requires that an intact sample be impregnated with an epoxy or
acrylic resin, such as described by Buol and Faddness (1961), Innes and Pluth (1970), and
Ashley (1973), to fill the pore space and stabilize the specimen.
WSRC-TR-2003-00331
December 31, 2004
3-11
Table 3.3
Mineralogical Analysis Methods for Evaluating Aquifer Solids for Natural Attenuation and
Enhanced Remediation
Analysis
Comments
Data Use
Field or
Fixed-Base
Laboratory
Petrography Either a petrographic
microscope or a simple
reflecting microscope with a
polarized light source can be
used.
Mineral identification and distribution. Field or Fixed Base
Wet Chemistry Useful for determining the
amount of “reactive” species
and extimating the amount of
other iron & sulfur minerals.
Determining amount of Fe(III) available for
reduction to Fe(II).
Characterization of iron and sulfur mineral
species.
Field or Fixed Base
X-Ray Diffraction Useful for identifying
crystalline minerals; cannot be
used to identify “amorphous”
materials.
Mineral Identification and concentration Fixed Base
X-Ray Fluorescence Elemental analysis method
useful for determining bulk
element composition.
Bulk element composition used to characterize
potential minerals present.
Fixed Base
Electron Microprobe Elemental analysis method
useful for identifying
microcrystalline and
amorphous iron-bearing
mineral phases.
Identification of iron-bearing mineral phases
from cemical composition in textural context.
Fixed Base
Mass Magnetic
Susceptibility Analysis
Method under development Useful for determining the presence and
concentration of magnetic minerals (e.g.,
magnetite) and changes in iron mineral
compostition.
Fixed Base
WSRC-TR-2003-00331
December 31, 2004
3-12
Information on the distribution of iron minerals in the aquifer matrix is of particular interest
since the abiotic dechlorination reactions with iron minerals are surface controlled, making an
understanding of the distribution and potential contact areas of these minerals useful. This is an
advantage of thin-section petrography, since bulk composition analysis methods, such as x-ray
diffraction (XRD), do not yield this information.
While petrography can provide definitive identification of individual mineral components, the
processes involved in preparing a sample for examination are time consuming and cumbersome
and a number of samples will be needed. Specialized equipment (i.e., petrographic and/or
binocular microscopes) is required and individuals with specific training and experience in
petrographic examination and mineral identification techniques should perform the examination,
particularly for iron minerals. Due to this complexity, professional petrographic examination can
be expensive and, therefore, only samples deemed to adequately represent aquifer materials
should be considered for such evaluation.
3.2.2 Wet Chemistry Techniques
Wet chemistry methods for evaluation of iron minerals rely on selective extraction approaches
that are based on differential reactivity and dissolution of the various iron-bearing constituents.
These approaches have a long history of use in soil science studies and, over time, numerous
methods have been developed or refined, particularly for the iron oxides and hydroxides
(Loeppert and Inskeep, 1996). Several wet chemical methods have been developed and refined
in recent years for speciation of iron and sulphur minerals (e.g., Heron et al., 1994a; 1994b;
Kennedy et al., 1998; 2000). These chemical extraction techniques involve measuring
operationally defined quantities of extracted or reacted species and are simple and easy to use.
They yield a bulk composition analysis of the minerals in the aquifer matrix based on the
composition of the extracted phase.
The wet chemistry methods involve the mixing of the sediment with an extractant followed by
determination of the extractant composition. Analysis of extracts and evolved gases is used for
determining the composition of the reacted mineral phases. Due to the variable composition,
degree of crystallinity, and reactivity of the various iron-bearing minerals, specific identification
by selective extraction methods is equivocal. However, the wet chemistry methods provide a
useful indication of the “reactive fraction” of iron-bering minerals in the aquifer matrix. A
standard reaction time is selected for each extraction method, but a given method does not
necessarily extract the same “reactive fraction” from different sediments due to differences in
specific mineral composition.
Several speciation techniques for iron and sulfur species and a protocol for their use is
presented in Kennedy et al. (2000). However, this protocol focuses on capacity determination
for fuel hydrocarbon biodegradation; for chlorinated solvents a more detailed speciation may be
needed to address the constituents important for abiotic dechlorination. Heron et al. (1994a)
describes several iron speciation techniques and determined that the “traditional” soil science
extractions for Fe(III) (dithionate and ammonium oxalate) were inappropriate in reduced aquifer
sediments due to interference from Fe(II). Techniques for sulfur species that separate acid-
WSRC-TR-2003-00331
December 31, 2004
3-13
volatile sulphides (AVS), pyrite, chromium-reducible sulphur, elemental sulphur, and solid
sulphates are described in Heron et al. (1994b). All these methods involve HCl extractions that
are inappropriate in sediments with high background content of carbonates due to neutralization
of the acid during extraction (Amirbahman et al., 1998). In systems supersaturated with respect
to ferrous carbonate, precipitation of siderite and ferroan calcite can occur (Tuccillo et al., 1999).
This can result in interferences for Fe(II) quantification and a milder extraction using ascorbic
acid may be appropriate (Amirbahman et al., 1998).
The techniques using “mild” extractants (e.g., weak HCl solutions) preferentially dissolve
surface-related, very reactive, or “amorphous” minerals, such as ferrihydrite [Fe(OH)3] and
ferrous monosulphides (FeS), while leaving most of the bulk iron and sulfur species behind and
intact. Hydrochloric acid and ascorbic acid extractions are the most promising approaches for
reactive fraction measurements (Christensen et al., 2000).
3.2.3 X-Ray Diffraction
X-ray diffraction (XRD) is a technique that allows the determination of minerals based on
measuring the diffraction pattern of x-rays reflected off of mineral crystal surfaces. It is one of
the primary techniques used to examine the physico-chemical composition of unknown solids
and can provide unambiguous determination of mineralogical composition of a sample.
However, the major limitation on the application of XRD is that it typically will not resolve
poorly crystallized or “amorphous” (noncrystalline) materials.
X-Ray diffraction can be a useful method for identifying crystalline iron-bearing minerals,
such as magnetite and pyrite, and can therefore be useful for determining the potential for abiotic
reductive dechlorination. However, poorly crystallized or “amorphous” monosulfides (FeS),
perhaps one of the most important minerals contributing to abiotic reductive dechlorination, will
not be detected with this technique.
XRD uses a powdered sample of the material placed in a holder and illuminated with x-rays
of a fixed wavelength. The reflection angles and intensity of the reflected radiation are recorded
using a goniometer. This data is then analyzed for the reflection angle to calculate the inter-
atomic spacing (D). The intensity (I) is measured to discriminate the various D spacings using
relative intensities. The results are compared to single-phase XRD patterns of known minerals to
identify possible matches.
When minerals can be obtained in adequate quantity and purity, their identification by XRD
powder methods is likely the simplest and most reliable approach. The technique can also be
applied to bulk (mixed mineral) samples, but the number of intensity peaks from the various
minerals that are present complicates the evaluation of results. Various sample pre-treatment
techniques, such as discussed in Laird and Dowdy (1994) Jones and Malik (1994), may be
useful, depending upon the information needed from the XRD results. Differential XRD
(DXRD) is widely used for determination of iron-oxide minerals. DXRD involves comparing
XRD patterns of samples before and after selective dissolution treatment for one or more
WSRC-TR-2003-00331
December 31, 2004
3-14
constituents. This allows discerning the components of a bulk analysis. An overview of DXRD
approaches is provided in Schulze (1994).
3.2.4 X-Ray Fluorescence
X-Ray fluorescence (XRF) is a non-destructive, bulk elemental analysis technique that can be
useful for characterizing both crystalline and amorphous iron-bearing solids, such as magnetite,
pyrite, and FeS. Therefore, it can be useful for determining the potential for abiotic reductive
dechlorination.
XRF spectrometry is based on measuring the emitted radiation (fluorescence) of element
atoms generated by exposure to an x-ray source. Primary x-rays from the source that are
absorbed by atoms in the target sample displace electrons from the inner shell of the atoms.
Electrons from the outer shell of the atoms transfer to the inner shell to fill the vacancies and the
energy change yields secondary x-rays with energies equivalent to the difference in the binding
energies of the two electron shells. Each element produces characteristic secondary x-ray
energies and measurement of these allows quantitation of the elemental composition of the
sample.
XRF is capable of detecting elements with atomic number ≥ 8 (oxygen) and can commonly
resolve concentrations of a few parts per million. It is commonly used in the laboratory and in
the field to determine concentrations of metals in soils. Reviews of the technique are presented
in Amonette and Sanders (1994) and Karathanasis and Hajek (1996).
Bulk elemental analysis by XRF yields information on the total elemental content of the
sample; it does not distinguish individual mineral phases. While information on the amounts of
each element present in a sample can be useful in constraining the mineral phases that may be
present, in itself this information on a bulk sample of aquifer material does not necessarily aid in
mineralogical characterization. Selection of sample material that represents the specific minerals
of interest can, to some extent, alleviate this issue. XRF cannot distinguish between oxidation
states [e.g., Fe(II) and Fe(III)], among mineral polymorphs (minerals with different structures,
but the same elemental stoichiometry), or between species (e.g., S2- and SO42-). Thus, it does not
lead to the quantitation of minerals directly and must be supplemented by other types of analysis.
However, it can aid in determining what additional analyses are appropriate.
3.2.5 Electron Microprobe Analysis
Electron microprobe analysis is a non-destructive technique that provides for an in situ
determination of chemical composition on specimen surfaces. It is applicable to all iron-bearing
mineral phases, regardless of their degree of crystallinity. While the technique is typically
limited to concentrations of 50-100 parts per million, the ability to resolve very small volumes
(e.g., 1 µm3) compensates for this limitation (Sawhney, 1986). Electron microprobe analysis is a
combination of the techniques of scanning electron microscopy and x-ray spectroscopy. An
overview of aspects of the basic methodology is provided in Sawhney (1986) and Reed (1996).
WSRC-TR-2003-00331
December 31, 2004
3-15
An electron microprobe uses a high-energy focused beam of electrons to generate x-rays
characteristic of the elements within a sample. It can typically resolve volumes as small as 3 µm
across. Analysis locations are selected using a transmitted-light optical microscope, which
allows positioning accurate to about 1 µm. The resulting data yield quantitative chemical
information in a textural context and variations in chemical composition within a material, such
as a mineral grain, can be readily determined. Chemical composition is determined by
comparing the intensity of x-rays from standards with those from unknown materials and
correcting for the effects of absorption and fluorescence in the sample. The electron microprobe
allows quantitative determination of elements with atomic numbers >3 (beryllium) to >8
(oxygen), depending on the specific instrument, at routine levels as low as 100 ppm.
Electron microprobe analysis is a specialized technique that is not inexpensive, making it
inappropriate for routine use in evaluating mineral composition. Samples prepared for routine
optical petrographic evaluation can typically also be used for electron microprobe analysis and
this technique can help resolve ambiguities from optical petrographic evaluations.
3.2.6 Mass Magnetic Susceptibility Analysis
Magnetic susceptibility is essentially a measure of how liable a mineral is to magnetization
when subjected to a magnetic field. The mass specific magnetic susceptibility, χ, is the
magnetization induced by an applied field with units of m3/kg. Whole cores, or individual
sediment samples, are exposed to an external magnetic field that causes the sediments to become
magnetized according to the amount of various magnetizable, predominantly iron-bearing,
minerals present in the samples. Mc Bride (1986) provides a discussion of this type of analysis.
Magnetizable minerals include the ferri- and ferro-magnetic minerals (strongly magnetizable)
and any of the paramagnetic (moderately magnetizable) minerals and other substances. The
former include magnetite, hematite, iron titanium oxides, pyrrhotite, maghemite, greigite and
goethite. The paramagnetic minerals include a broad array of substances all of which contain
Fe2+, Fe3+, or Mn2+ ions, such as clay minerals (chlorite, smectite and glauconite), iron and
manganese carbonates (siderite, rhodochrosite), as well as a variety of ferric-oxyhydroxide
mineraloids. The ferri-magnetic iron oxides (e.g., magnetite, maghemite) typically dominate the
magnetic signature of soil, sediment and rock and have strong, positive mass magnetic
susceptibility values that are two to three orders of magnitude greater than those of the other iron
oxides (Bigham et al., 2002). For a mixed-mineral sample, the magnetic susceptibility is
primarily determined by the concentration of the ferri-magnetic iron oxides and related samples
will commonly show a positive, linear relationship with the magnetite and maghemite content
(Bigham et al., 2002). The magnetic susceptibility of a sample is ultimately related to the
concentration and composition (size, shape and mineralogy) of magnetizable material present
within the sample.
If the composition and structure of iron-bearing minerals changes as ferrous iron and sulfide
are incorporated into the aquifer matrix, there should be a variation in the magnetic
susceptibility. Evaluation of changes in the magnetic susceptibility signature may be useful for
determining the presence of these iron-bearing minerals. Although still under development for
WSRC-TR-2003-00331
December 31, 2004
3-16
application to natural attenuation evaluation, this analysis is rapid, inexpensive, and easy to use
in both the field and laboratory. It could prove useful in the future to help differentiate between
areas where abiotic dehalogenation is active or not.
3.3 MICROBIOLOGICAL METHODS
Several microbiological methods can yield useful information that provides more detailed
information on natural degradation mechanisms. These methods include microcosm studies and
molecular biology tools, such as phospholipid fatty acids (PLFA) and denaturing gradient gel
electrophoresis (DGGE). These methods are discussed briefly below. For more detailed
discussion and description of these, and other, molecular biology methods, see White and
Ringelberg (1998), Burlage (1998), Liu and Stahl (2002), and Pinkart et al. (2002).
3.3.1 Microcosm Studies
The most common type of data collected for evaluating the degradation of organic
contaminants in aquifer material is the laboratory microcosm study. If properly designed,
implemented, and interpreted, microcosm studies can provide very convincing documentation of
the potential for intrinsic bioremediation and/or abiotic reductive dechlorination. Microcosm
studies are the only “line of evidence” that allows an unequivocal mass balance on the
biodegradation of environmental contaminants. If the microcosm study is properly designed, it
will be easy for decision-makers with non-technical backgrounds to interpret. The results of a
microcosm study are strongly influenced by the nature of the geological material submitted for
study, by the physical properties of the microcosm, by the sampling strategy, and by the duration
of the study. Therefore, relating laboratory microcosm results back to in situ field conditions can
be difficult. Additionally, microcosm studies are time consuming and expensive to conduct. For
these reasons, microcosm studies should be used very selectively in assessing the efficiency of
natural attenuation and enhanced remediation.
There are some circumstances, however, when laboratory studies are useful. When specific
questions are raised concerning conditions under which degradation processes occur or do not
occur, controlled laboratory studies can be helpful. For example, if concentrations of a particular
compound are observed to decrease in the field, it is often not clear whether this decrease is due
to sorption, dilution, or biological or abiotic degradation. Laboratory studies in which the effects
of each process can be isolated and controlled (they usually cannot be controlled in the field) are
the only available method of answering these questions.
3.3.2 Phospholipid Fatty Acids (PLFA)
Examining the phospholipid fatty acids (PLFA) in environmental samples is an effective tool
for monitoring microbial responses to their environment. They are essential components of the
membranes of all cells (except for the Archea), so their sum includes most of the important
actors in microbial communities. Methanogens are members of the Archea and are not included
in this analysis. There are four different types of information in PLFA profiles – biomass,
community structure, diversity, and physiological status.
WSRC-TR-2003-00331
December 31, 2004
3-17
3.3.2.1 Biomass
PLFA analysis is the most reliable and accurate method available for the determination of
viable microbial biomass. Since phospholipids break down rapidly upon cell death (White et al.,
1979; White and Ringelberg, 1995), the PLFA biomass does not contain ‘fossil’ lipids of dead
cells. The sum of the PLFA, expressed as picomoles (1 picomole = 1 × 10-12 mole), is
proportional to the number of cells. The proportions used typically are taken from cells grown in
laboratory media, and vary somewhat with the type of organism and environmental conditions.
Starving bacterial cells have the lowest cells/pmol, and healthy eukaryotic cells have the highest.
3.3.2.2 Community Structure
The PLFA in an environmental sample is the sum of the microbial community’s PLFA, and
reflects the proportions of different organisms in the sample. PLFA profiles are routinely used to
classify bacteria and fungi (Tighe et al., 2000) and are one of the characteristics used to describe
new bacterial species (Vandamme et al., 1996). Broad phylogenic groups of microbes have
different fatty acid profiles, making it possible to distinguish among them (Dowling et al., 1986;
Edlund et al., 1985; White et al., 1997; White et al., 1996). Table 3.4 describes the six major
structural groups employed in a typical analysis of groundwater.
3.3.2.3 Diversity
The diversity of a microbial community is a measure of the number of different organisms
and the evenness of their distribution. Natural communities in an undisturbed environment tend
to have high diversity. Contamination with toxic compounds will reduce the diversity by killing
all but the resistant organisms. The addition of a large amount of a food source will initially
reduce the diversity as the opportunists (usually Proteobacteria) over-grow organisms less able to
reproduce rapidly. The formulas used to calculate microbial community diversity from PLFA
profiles have been adapted from those applied to communities of macro-organisms (Hedrick et
al., 2000).
3.3.2.4 Physiological Status
The membrane of a microbe must adapt to the changing conditions of its environment, and
these changes are reflected in the PLFA. Toxic compounds or environmental conditions that
disrupt the membrane cause some bacteria to make trans fatty acids from the usual cis fatty acids
(Guckert et al., 1986). Many Proteobacteria and others respond to starvation or highly toxic
conditions by making cyclopropyl (Guckert et al., 1986) or mid-chain branched fatty acids
(Tsitko et al., 1999). The physiological status biomarkers for toxic stress and starvation/toxicity
are formed by dividing the amount of the stress-induced fatty acid by the amount of its
biosynthetic precursor.
WSRC-TR-2003-00331
December 31, 2004
3-18
Table 3.4
Description of PLFA Structural Groups
PLFA Structural Group General classification
Monoenoic (Monos)
Abundant in Proteobacteria (Gram negative bacteria), typically fast
growing, utilize many carbon sources, and adapt quickly to a variety of
environments.
Terminally Branched Saturated (TerBrSats)
Characteristic of Firmicutes (Low G+C Gram-positive bacteria), and
also found in Bacteriodes, and some Gram-negative bacteria
(especially anaerobes).
Branched Monoenoic (BrMonos) Found in the cell membranes of micro-aerophiles and anaerobes, such
as sulfate- or iron-reducing bacteria .
Mid-Chain Branched Saturated (MidBrSats) Common in Actinobacteria (High G+C Gram-
p
ositive bacteria), and
some metal-reducing bacteria.
Normal Saturated (Nsats) Found in all organisms.
Polyenoic Found in eukaryotes such as fungi, protozoa, algae, higher plants, and
animals.
WSRC-TR-2003-00331
December 31, 2004
3-19
3.3.2.5 Example PLFA Analysis
Two groundwater samples were collected from downgradient of a landfill for analysis of
PLFAs. Over the course of operation the landfill accepted municipal solid waste and spent
chlorinated solvents. PLFAs were analyzed by extraction of the total lipid (White et al., 1979)
and then separation of the polar lipids by column chromatography (Guckert et al., 1985). The
polar lipid fatty acids were derivatized to fatty acid methyl esters, which were quantified using
gas chromatography (Ringelberg et al., 1994). Fatty acid structures were verified by
chromatography/mass spectrometry and equivalent chain length analysis. Results from the
analysis of these two samples revealed the following information.
Biomass content (as determined by the total PLFA concentration) was fairly similar in both
samples (Figure 3.1). The total biomass calculated based upon PLFA attributed to bacterial and
eukaryotic biomass is ~105 cells/mL using the proportion 20,000 cells/pmole taken from cells
grown in laboratory media.
Both samples had relatively diverse community structures that differed between the two
sampling points based on the relative proportions of PLFA structural groups (Figure 3.2).
Structural groups are assigned according to PLFA chemical structure, which is related to fatty
acid biosynthesis, as described in Table 3.4. The results for the two samples indicate (Table 3.5):
• Sample MNA-10 was primarily composed of Gram-negative proteobacteria (as indicated
by the percentage of monoenoic PLFA). High proportions of proteobacteria are of
particular interest at contaminated sites due to their ability to utilize a wide range of
carbon sources and adapt quickly to environmental conditions.
• Sample MNA-05 differed from MNA-10 in that proportions of Proteobacteria were lower
whereas proportions of biomarkers associated with Firmicutes (as indicated by terminally
branched PLFA) and anaerobic metal reducing bacteria (branched monoenoic and mid
chain branched PLFA) were noticeably higher. Terminally branched PLFA are common
in Firmicutes (Clostridia-like Gram-positive bacteria). An increase in terminally
branched PLFA is often seen in environmental transects from more aerobic to more
anaerobic conditions.
Physiological status ratios for starvation were highest in sample MNA-05, whereas both
samples were indicating a response to environmentally induced stress (Figure 3.3). The
starvation biomarker for the gram-negative bacterial community is assessed by the ratios of
cyclopropyl fatty acids to their metabolic precursors. An adaptation of the Gram-negative
community to toxic stress is determined by the ratio of trans to cis omega-7 fatty acids
(ω7t/ω7c). Gram-negative bacteria generate trans fatty acids to minimize the permeability of
their cellular membranes as an adaptation to a less favorable environment. Ratios greater than
0.1 have been shown to indicate an adaptation to a toxic or stressful environment, resulting in
decreased membrane permeability. Of course, one must be cautious when interpreting bacterial
response to an environment using data from only one sampling event.
0
2
4
6
8
10
12
14
MNA-10 MNA-05
pmoles of PLFA/mL water
D:\doe\report\lines of evidence\Figure 3.1.ai
Westinghouse Savannah River MNA/EA Project
Figure 3.1
Biomass Content Presented as Total PLFAs
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
3-20
D:\doe\report\lines of evidence\Figure 3.2.ai
Westinghouse Savannah River MNA/EA Project
Figure 3.2
Relative Percentages of PLFA Structual
Groups. Table 3.4 Describes the
Various Structural Groups.
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
0%
20%
40%
60%
80%
100%
MNA-10 MNA-05
% of total PLFA
Firmicutes (TerBrSats) Proteobacteria (Monos)
Anaerobic metal reducers (BrMonos) Actinomycetes (MidBrSats)
General (Nsats) Eukaryotes (polyenoics)
3-21
WSRC-TR-2003-00331
December 31, 2004
3-22
Table 3.5
Viable Microbial Biomass Expressed as picomoles of PLFA per mL of Sample and as Cells per mL of Sample, Fatty Acid
Structural Groups as Percent of Total PLFA, and Physiological Status Biomarkers as Mole Ratio
Samples Biomass Community Structure (Percent of Total PLFA) Physiological Status
Sample
Name
Sample
Date pmol/mL cells/mL
Anaerobic
Gram Neg./
Firmicutes
(TerBrSats)
Proteobacteria
(Monos)
Anaerobic
metal
reducers
(BrMonos)
Actinomycetes/SRB
(MidBrSats)
General
(Nsats)
Eukaryotes
(polyenoics) Starved cy/cis
Membrane
Stress,
trans/cis
MNA-10 6/5/2003 9 1.89E+05 13.7 56.5 2.8 4.7 21.3 1.0 0.39 0.17
MNA-05 6/5/2003 13 2.58E+05 35.2 31.8 4.1 8.6 17.7 2.6 0.65 0.15
WSRC-TR-2003-00331
December 31, 2004
3-23
3.3.3 Denaturing Gradient Gel Electrophoresis
The recovery of DNA and RNA and its subsequent analysis after amplification by polymerase
chain reaction (PCR) provides a powerful tool for characterizing microbial community structure
that complements the PLFA analysis. As with PLFA analysis, numerous studies have used PCR
amplification of ribosomal RNA genes (rDNA) to characterize microbial populations in a
number of different environments and have demonstrated that the dominant microorganisms
isolated by culture frequently do not match those identified by molecular techniques (Amann, et
al., 1995). Given that often only 0.1 - 10% of visually countable bacteria in samples are cultured
and previous studies have demonstrated that organisms obtained from culturing are not
necessarily the numerically dominant organisms in situ, it is apparent that the results from
culture-based community structure assessments can be noticeably incomplete.
Denaturing gradient gel electrophoresis (DGGE) analysis can be used to detect and identify
organisms from a whole community of organisms. The DGGE approach directly determines the
species composition of complex microbial assemblages based on the amplification of conserved
gene sequences (16S rDNA fragments for prokaryotes, 18S or 28S rDNA for eukaryotes). In
DGGE analysis, differences in gene sequences among organisms allow DNA from various
organisms to be physically separated in a denaturing gradient gel, thereby allowing one to
generate profiles of numerically dominant bacterial community members for a sample. The
profiles are visible as bands (or lines) in a gel. The banding patterns and relative intensities of
the bands provide a measure of difference among the communities. Gel bands from dominant
species, which constitute at least 1% of the total bacterial community, can be excised and
sequenced. Sequence analysis of individual bands is used to infer the identity of the source
organism based on database searches and phylogenetic methods. Phylogenetic affiliations are
determined by comparing the rDNA sequences retrieved from samples to rDNA sequences of
known bacterial sequences in national databases, such as the Ribosomal Database Project (RDP)
or GenBank).
D:\doe\report\lines of evidence\Figure 3.3.ai
Westinghouse Savannah River MNA/EA Project
Figure 3.3
Microbial Physiological Stress Markers.
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
MNA-10 MNA-05
Physiological Status Ratio
Starvation (cy/cis)
Stress (trans/cis)
3-24
WSRC-TR-2003-00331
December 31, 2004
4-1
SECTION 4
DEDUCING GEOCHEMICAL ENVIRONMENTS AND DEGRADATION PATHWAYS
Solute plumes formed by petroleum hydrocarbons and most other sources of reduced organic
carbon behave in a very predictable manner in the terrestrial subsurface. This consistent
behavior results from these compounds being used as electron donors that are oxidized by
bacteria utilizing the naturally-occurring electron acceptors that are ubiquitous in groundwater.
Conversely, the reductive dechlorination of the common chlorinated solvents (i.e., PCE, TCE,
CT, and TCA) is much more dependent upon the prevailing groundwater geochemistry because
these compounds are highly oxidized and are used as electron acceptors. In addition, the lesser
chlorinated compounds (e.g., DCE and VC) can be used as electron donors or electron acceptors;
that is they can either be oxidized or reduced depending upon the prevailing terminal electron-
accepting process. For these reasons an accurate assessment of groundwater geochemistry is
required to determine if biological or abiotic degradation is occurring, to determine degradation
pathways, and to determine the sustainability of the reactions. As stated by Dr. John T. Wilson
of the USEPA (per comm.) chlorinated solvent plumes are like children, each one is different.
The main factor affecting the degradation of chlorinated compounds is the prevailing
geochemical environment. Thus, an understanding of the groundwater geochemistry, principally
the distribution of aerobic versus anaerobic environs, is essential for deducing degradation
pathways, and interpreting chlorinated solvent behavior. Once degradation pathways are
elucidated it is possible to determine if natural attenuation will be sufficient as a remedial action.
In addition, if natural attenuation is shown not to be protective, an understanding of the ambient
groundwater geochemistry and degradation pathways is essential for determining the most
appropriate remedial enhancement. In many cases, portions of the plume with differing
geochemistry will require different remediation approaches.
To help understand the degradation pathways of chlorinated compounds in different
environments, Wiedemeier et al. (1998) proposed a classification system for chlorinated solvent
plumes which was adopted by the USEPA (1998). This classification system presented “Type”
geochemical environments that result in specific chlorinated solvent behavior. In essence, this
classification system groups the groundwater system into either anaerobic and strongly reducing
(Type 1 and Type 2) or aerobic and oxidizing (Type 3) end-member environments. In most
solvent plumes where reductive dechlorination is occurring, some mixture of environments will
be observed, and thus degradation pathways will be different in different portions of the plume.
WSRC-TR-2003-00331
December 31, 2004
4-2
Once the geochemistry of the groundwater has been characterized it is possible to deduce
degradation pathways and interpret solute plume behavior. This is essential and required for
evaluating the efficacy of natural attenuation and selecting enhanced remedial alternatives should
they be necessary.
As mentioned previously, Wiedemeier et al. (1998) proposed a classification system for
chlorinated solvent plumes which was adopted by the USEPA (1998). Because the predominant
thinking at the time was that the reductive dechlorination of chlorinated solvents was strictly a
biological process, this classification system was based on the amount and origin of fermentation
substrates that produce the hydrogen that drives dehalorespiration. Three types of groundwater
environments and associated plume behavior (Type 1, Type 2, and Type 3) are described in
USEPA (1998). Type 1 and Type 2 environments were those environments where the
groundwater was very strongly reducing (i.e., sulfate reducing or methanogenic) because of the
presence of anthropogenic or naturally-occurring organic carbon, respectively. In the Type 1 and
Type 2 environments chlorinated solvents can be completely biodegraded to ethane and ethane.
The Type 3 environment was characterized by oxidizing or only very slightly reducing
groundwater. In this environment the more highly chlorinated compounds (e.g., PCE, TCE, CT,
and TCA) are not biodegraded but VC and DCE can be biologically oxidized. Recent advances
in the understanding of chlorinated solvent degradation shows that abiotic reductive
dechlorination reactions can be important in some systems. Because this type of reaction
typically requires the availability of Fe(II) and sulfide to generate iron sulfides, the system must
be at least sulfate-reducing. Thus, when a site has a Type 1 or Type 2 environment, an
assessment of the relative importance of biological versus abiotic degradation mechanisms must
be made.
Table 4.1 summarizes the characteristics of the different “Type” geochemical environments
with subsets for biological (possibly combined with abiotic) versus exclusively abiotic reactions.
Table 4.2 is a matrix showing some of the potential geochemical environments encountered in
the terrestrial subsurface and the impact of these environments on the fate of chlorinated ethenes.
The potential variability in chlorinated solvent plume behavior is almost infinite and most
plumes exhibit mixed behavior in different portions of the plume. Thus, the attenuation capacity
will be different in different portions of the plume. This has a direct and profound impact on the
remedial approach selected for different portions of the plume.
4.1 TYPE 1 ENVIRONMENT: SYSTEMS THAT ARE ANAEROBIC DUE TO
ANTHROPOGENIC CARBON
For highly chlorinated solvents to be reductively dechlorinated via dehalorespiration or
biologically predicated abiotic processes, anaerobic, strongly reducing conditions must prevail
within the contaminant plume. Anaerobic conditions are typical at sites contaminated with fuel
hydrocarbons, landfill leachate, or other anthropogenic carbon because these organics exert a
tremendous electron-acceptor demand on the system, driving it to a strongly reducing condition.
WSRC-TR-2003-00331
December 31, 2004
4-3
Table 4.1
General Characteristics of the “Type” Geochemical Environments
Environment Reductive
Dechlorination?
General Characteristics Comments Natural Attenuation Capacity
Type 1 Yes; biological
(dehalorespiration)
and possibly
abiotic
Depleted:
Dissolved Oxygen
Nitrate
Sulfate
Elevated:
Fe(II) [if biologically available Fe(III) is present]
H2S (if sulfate is available)
Methane/ethene/ethane
Hydrogen
Daughter Product Concentrations
(If chlorinated ethenes are present then VC
will likely be present)
Acetylene (+/-)
(results from abiotic reductive dechlorination)
Microbiology:
Dehalococcoides or other dehalorespirators present
Strongly reducing environment
due to the presence of
anthropogenic organic carbon
where biologically-mediated
dehalorespiration is an important
reaction.
Abiotic reactions may or may not
occur. Degradation of chlorinated
ethenes to VC and ethene/ethane
common.
All of the chlorinated ethenes,
ethanes, and methanes will
degrade in this environment.
PCE, TCE = High
DCE = Moderate
VC = Low (may accumulate)
TCA = High (biodegradation/hydrolysis)
DCA = Moderate
Chloroethane = Very High (Hydrolysis)
Cabon Tetrachloride (CT) = High
Chloroform (TCM) = High
Methylene Chloride (DCM) = Moderate
Chloromethane = Very High
(Hydrolysis)
Yes; Abiotic Only Depleted:
Dissolved Oxygen
Nitrate
Sulfate
Elevated:
Fe(II)
H2S
FeS & other iron sulfides (critical)
Methane (may or may not be elevated)
Acetylene likely present (from abiotic degradation)
Microbiology:
Dehalococcoides or other dehalorespirators absent
The presence of acetylene and
cis-1,2-DCE and absence of VC
are indicative of an environment
where abiotic degradation
dominates.
Requires the presence of
biologically available Fe(III) and
sulfate and Fe(III)- and sulfate-
reducing bacteria.
FeS appears to be the most
efficient reductant.
PCE, TCE = High
DCE = Moderate to High
VC = typically not present
TCA = High
DCA = Moderate
Chloroethane = Very High (Hydrolysis)
Cabon Tetrachloride (CT) = High
Chloroform (TCM) = Moderate
Methylene Chloride (DCM) = Moderate
Chloromethane = Very High
(Hydrolysis)
WSRC-TR-2003-00331
December 31, 2004
4-4
Table 4.1 - Continued
General Characteristics of the “Type” Geochemical Environments
Environment Reductive
Dechlorination?
General Characteristics Comments Natural Attenuation Capacity
Type 2 Yes; biological
(dehalorespiration)
and possibly
abiotic
Depleted:
Dissolved Oxygen
Nitrate
Sulfate
Elevated:
Fe(II) [if biologically available Fe(III) is present]
H2S
Methane/ethane/ethene
Hydrogen
Daughter Product Concentrations
(If chlorinated ethenes are present then VC
will likely be present)
Acetylene (+/-)
(results from abiotic reductive dechlorination)
Microbiology:
Dehalococcoides or other dehalorespirators present
Strongly reducing environment due
to the presence of natural organic
carbon where biologically-mediated
dehalorespiration is an important
reaction.
Abiotic reactions may or may not
occur. Degradation of chlorinated
ethenes to VC and ethene/ethane
common.
All of the chlorinated ethenes,
ethanes, and methanes will degrade
in this environment.
PCE, TCE = High
DCE = Moderate
VC = Low (may accumulate)
TCA = High (biodegradation and
hydrolysis)
DCA = Moderate
Chloroethane = Very High (Hydrolysis)
Cabon Tetrachloride (CT) = High
Chloroform (TCM) = High
Methylene Chloride (DCM) = Moderate
Chloromethane = Very High
(Hydrolysis)
Yes; Abiotic Depleted:
Dissolved Oxygen
Nitrate
Sulfate
Elevated:
Fe(II)
H2S
FeS & other iron sulfides (critical)
Methane (may or may not be elevated)
Acetylene likely present (from abiotic degradation)
Microbiology:
Dehalococcoides or other dehalorespirators absent
The presence of acetylene and cis-
1,2-DCE and absence of VC are
indicative of an environment where
abiotic degradation dominates.
Requires the presence of
biologically available Fe(III) and
sulfate and Fe(III)- and sulfate-
reducing bacteria.
FeS appears to be the most efficient
reductant.
PCE, TCE = High
DCE = Moderate to High
VC = typically not present
TCA = High
DCA = Moderate
Chloroethane = Very High (Hydrolysis)
Cabon Tetrachloride (CT) = High
Chloroform (TCM) = Moderate
Methylene Chloride (DCM) = Moderate
Chloromethane = Very High
(Hydrolysis)
WSRC-TR-2003-00331
December 31, 2004
4-5
Table 4.1 - Concluded
General Characteristics of the “Type” Biogeochemical Environments
Environment Reductive
Dechlorination?
General Characteristics Comments Natural Attenuation Capacity
Type 3 Insignificant to
Nonexistant
Dissolved Oxygen Elevated
Nitrate +/-
Fe(II) absent or +/-
Sulfate +/-
H2S absent
FeS absent
Methane absent
Acetylene absent
Essentially an aerobic or only
slightly reducing environment.
PCE, TCE, and CT will not
degrade in this environment.
VC is biologically oxidized
very rapidly.
TCA hydrolyzes and
dehydrochlorinates in this
environment.
PCE, TCE = Very Low
DCE = Moderate
VC = High to Very High
TCA = High (Hydrolysis)
DCA = Moderate
Chloroethane = Very High (Hydrolysis)
Cabon Tetrachloride (CT) = Very Low
Chloroform (TCM) = Very Low
Methylene Chloride (DCM) = Moderate
Chloromethane = Very High (Hydrolysis)
Table 4.2
Matrix Showing Some of the Potential Geochemical Environments Encountered in the Terrestrial Subsurface and Impact on the Fate of Chlorinated Ethenes
Denitrifying
High X X
Low X X X
Absent XXXXXXXXX
Present X
+/- X X
Absent XXXXXXXX
High XX XX XX
Moderate XX XX XX
Low- Absent X X X X X
High XXXXX
Low XXXXX
+/- XX X
Absent X X X X X
Hi XXXXXXX
Low X
+/-XXXX X
Present XXXX
+/-XXXX
Absent XXX
Present XXXXX
Absent XXXX
+/- XX X
Present X XXX
Absent X XXXXXX
+/-
Hi XXXXXXX
Moderate XXXXX
LowXXXX
Present XXXXXX
Absent XXXX
+/- X
present XXXXX
absent XXXX X X
Degrades XXXXXXX
Does Not Degraded/ XXXX
+/-
Degrades X XXXXXXX
TCE Does Not Degraded/ XXX
+/-
Degrades X X ? XXXXXXXX
Does Not Degrade
+/-
Degradese/ XX ? XXXX XX
Does Not Form X X
Does Not Degraded/
+/-
Direct Oxidizing Environment Cometabolic Environment Anaerobic Oxidizing Anerobic Oxidizing
Environment
A
naerobic Oxidizing with Iron
Sulfides
Dehalorespiration only Dehalorespiration and abiotic
reductive dechlorination
Abiotic reductive dechlorination
only Dehalorespiration only Dehalorespiration and abiotic
reductive dechlorination
Abiotic reductive dechlorination
only
Dehalorespirers convert PCE/TCE
to cDCE and VC, but competition
for e-donor by sulfate-reducers ma
y
significantly reduce conversion of
cDCE to VC. VC may be reduced
to ethene if VC-competent
Dehalococcoides is present.
Dehalorespirers convert PCE/TCE
to cDCE and VC, but competition
for e-donor by sulfate-reducers ma
y
significantly reduce conversion of
cDCE to VC. VC may be reduced
to ethene if VC-competent
Dehalococcoides is present.
Potential for formation of
significant amounts of metalic
sulfides, and abiotic dechlorination
of PCE/TCE to cDCE and
acetylene.
Potential for formation of
significant amounts of metalic
sulfides, and abiotic dechlorination
of PCE/TCE to cDCE and
acetylene.
Dehalorespirers convert PCE/TCE
to cDCE and VC. VC reduced to
ethene if VC-competent
Dehalococcoides is present.
Dehalorespirers convert PCE/TCE
to cDCE and VC. VC reduced to
ethene if VC-competent
Dehalococcoides is present.
Potential for formation of
significant amounts of metalic
sulfides, and abiotic dechlorination
of PCE/TCE to cDCE and
acetylene.
Potential for formation of
significant amounts of metalic
sulfides, and abiotic dechlorination
of PCE/TCE to cDCE and
acetylene.
c/ Parameters associated with methanotrophic cometabolism are used for illustration, other cometabolic reactions are possible, such as propane-oxidizers and ammonia-oxidizers, necessary combination of factors is uncommon in natural settings and typically must be engineered, methanotrophic cometabolism may occur in the vicinity of landfills.
d/ Typically does not degrade in any appriciable amount
b/Although being produced , concentrations may be too low to detect because compound is unstable.
Controlling Factors
Disclaimer: This table is intended to present some representative combinations of analytical parameters that are reflective of the various environments for example purposes. NO representation is made that these encompass all possible combinations of parameters and application to chlorinated compounds other than chlorinated ethenes (PCE/TCE) likely will be different.
Sulfate (Plume)
Environment Type
Comments
Dissolved Sulfide
b/
Ethene, ethane
b/
Type 3
PCE/TCE will not undergo
degradation. cDCE and VC likely
will be oxidized.
Compounds less chlorinated than
PCE will be cometabolically
oxidized in small transition zones
where O2 and CH4 are present.c/
Reductive dechlorination not
expected
Dehalorespirers may convert
PCE/TCE to cDCE but competitio
n
for e-donor by iron reducers
typically precludes this. Similarly,
conversion of cDCE to VC and
ethene is possible, but at
significantly reduced rates.
However cDCE can be
anaerobically oxidized by iron-
reducers and may be removed.
Migration of sulfide can produce
metalic sulfides marginal to sulfate-
reducing zone. Abiotic reductive
dechlorination of PCE/TCE to
cDCE and acetylene. cDCE can
be anaerobically oxidized by iron-
reducers and may be removed.
Type 1 or 2
Anaerobic, Strongly Reducing Environment
Aerobic
Environmental Redox Conditions
Iron Reducing Sulfate Reducing Methanogenic
Chemical Indicators
Compound Behavior
Oxygena/
Iron Sulfides (Solid)
Sulfate (Background)
Nitrate
Dissolved Fe(II)
Methane
Acetylene
b/
a/ Although dissolved oxygen may be detected in groundwater samples from iron reducing, sulfate reducing, and methanogenic zones this is inevitably due to poor sampling technique that allows contamination of a sample by atmospheric oxygen.
Organic Carbon
e/ Biodegradatin rate slows appreciably under sulfate reducing and methanogenic conditions and compound accumulates.
cis 1,2-DCE
VC
PCE
Biogeochemical Environment
Type 1 or 2
Anaerobic, Strongly Reducing Environment
Type 3 -- Transitional
C:\DOE\Report\Lines of Evidence\Draft-R1\Tables\Table 4-2.xls 4-6
WSRC-TR-2003-00331
December 31, 2004
4-7
This condition is referred to as a Type 1 environment. In a Type 1 environment,
anthropogenic carbon is fermented to produce hydrogen that drives dehalorespiration. In many
cases this environment also produces FeS or other reduced iron minerals which can cause
biologically predicated abiotic reductive dechlorination.
The geochemistry of groundwater in a Type 1 environment is typified by strongly reducing
conditions. This environment is characterized by the absence of dissolved oxygen and nitrate,
elevated Fe(II) and sulfide concentrations, sulfate concentrations that typically are lower than
those found in uncontaminated groundwater in the vicinity of the solute plume, and the presence
of methane. In addition, minerals containing reduced iron (e.g., FeS) also can be produced.
Unless background sulfate concentrations are very high, methane is almost always observed and
confirms that fermentation has occurred or is occurring at the site, generating hydrogen. If
measured, hydrogen concentrations are typically greater than 1 nanomolar.
A Type 1 environment results in the rapid and extensive degradation of the more highly-
chlorinated solvents such as PCE, TCE, CT, and TCA. A conceptual model of the Type 1
environment for a chlorinated ethene plume is shown in Figure 4.1, where map views and
centerline concentration profiles of PCE, TCE, cis-1,2-DCE, VC, inorganic electron acceptors
(dissolved oxygen, nitrate, sulfate, and carbon dioxide), metabolic by-products (methane and
dissolved iron), fermentation substrates (biochemical oxygen demand or BOD), and fermentation
products (acetate) are shown. If the reductive dechlorination is due to dehalorespiration alone,
the following sequence of reactions occurs:
PCE → TCE → DCE → VC → Ethene → Ethane
In this type of plume where degradation results from dehalorespiration only, cis-1,2-DCE and
VC degrade more slowly than TCE; thus, they tend to accumulate and form longer plumes
(Figure 4.1a). In Figure 4.1b, the PCE declines to zero and is replaced, in sequence, by a peak in
TCE concentrations, followed by a peak in cis-1,2-DCE, VC, and ethene. The oxidation-
reduction potential is depressed in the source zone by the anthropogenic carbon source and stays
depressed throughout the entire plume. Fermentation constituents (BOD and acetate) and
inorganics are shown in Figure 4.1c and 4.1d. Figure 4.1d illustrates how the fermentation
substrate (represented by BOD) extends beyond the source before being consumed. Both panels
show long chloride and methane plumes extending far downgradient from the plume area,
because chloride is conservative and methane is not biodegraded in an anaerobic environment.
The acetate curve indicates where active primary fermentation is occurring; declining acetate
concentrations are due to consumption by methanogens in the plume area.
a
b
c
d
Extent of Type 1 Environment
VC
cis-1,2-DCE
Ethene
PCE
TCE
Redox
cis-1,2-DCE
PCE
Ethene
TCE
CH4
Iron
BOD Chloride
Acetate
DO
NO3
Dissolved Iron
Chloride
Acetate
NO3
DO
CH4
BOD (e.g.,
methanol)
SO4
VC
SO4
Dissolved
Concentration or Mass Concentration or Mass
Back-
ground
Source
Area Downgradient
Distance Downgradient and
Direction of Groundwater Flow
D:\doe\report\lines of evidence\Figure 4.1.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.1
Conceptual Model of Type 1 Environment
for Chlorinated Solvent Plume Due to a PCE
and TCE Release
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
4-8
(After RTDF, 1997)
WSRC-TR-2003-00331
December 31, 2004
4-9
In addition to biologically-mediated reductive dechlorination, abiotic reductive dechlorination
also may be important in a Type 1 environment. The Fe(II) and sulfide generated by microbial
Fe(III)-reduction and sulfate-reduction will react readily to produce iron sulfide minerals that can
reductively dechlorinate the chlorinated aliphatic hydrocarbons. If chlorinated ethenes are
present in such a system, the degradation may appear to stall at DCE because no VC is produced.
In fact, in many cases the reaction has not stalled; instead the DCE is being abiotically degraded
to acetylene by iron sulfides. It appears that the acetylene then degrades to ethene and ethane.
Some important questions regarding the long-term behavior of plumes in a Type 1
environment are:
1) Is the supply of the anthropogenic carbon adequate to allow complete reductive
dechlorination?
2) Will the production of hydrogen from fermentation be sufficient to drive dechlorination to
completion if dehalorespiration is the dominant degradation mechanism?
3) What is the role of alternate electron acceptors (e.g., dissolved oxygen, nitrate, Fe(III) and
sulfate) in the competition for the hydrogen? In a Type I environment, there is usually
sufficient anthropogenic carbon delivered to the source zone (typically from dissolution of
residual NAPLs or from landfill leachate) to remove all the competing electron acceptors
and make dehalorespiration more efficient. If the supply of anthropogenic carbon declines
(such as through the dissolution of NAPL, removal of a floating product layer, or removal
of the landfill leachate), then the competing electron acceptors may no longer be depleted.
This would result in cessation of biologic, and possibly abiotic, reductive dechlorination.
In addition, the chlorinated solvent plume may migrate out of the area influenced by the
biodegradation of reduced organic carbon into a Type 3 environment where, again, there
would be a cessation of biologic and abiotic reductive dechlorination.
4) If produced, is vinyl chloride attenuated at rates sufficient to be protective of human health
and the environment? Vinyl chloride biodegrades the slowest of the chlorinated ethene
compounds in anaerobic environments and may accumulate, forming a longer vinyl
chloride plume relative to other chlorinated ethenes (PCE, TCE, or DCE). Based on
carcinogen slope factors, vinyl chloride poses a higher potential health risk than the other
chlorinated ethenes.
5) Are sufficent ferric iron (Fe3+) and sulfate available to yield the production of significant
FeS that can result in abiotic reductive dechlorination. The amount of FeS generated from
biogenically produced Fe(II) and sulfide will determine the potential for abiotic reductive
dechlorination and its importance relative to dehalorespiration.
WSRC-TR-2003-00331
December 31, 2004
4-10
4.2 TYPE 2 ENVIRONMENT: SYSTEMS THAT ARE ANAEROBIC DUE TO
NATURALLY-OCCURRING CARBON
The classification system of Wiedemeier et al. (1996a) recognized that anaerobic conditions
may also result from the biodegradation of naturally-occurring organic material in the
groundwater that flows through chlorinated solvent source zones. The Type 2 environment
occurs in hydrogeologic settings that have inherently high organic carbon concentrations, such as
coastal or stream/river deposits with high concentrations of organics, shallow aquifers associated
with organic-rich environments (such as swamps), or zones impacted by natural oil seeps. When
evaluating natural attenuation of a chlorinated solvent plume in a Type 2 environment, the same
questions as for a Type 1 environment apply. In addition, the same general conceptual model
applies (See Figure 4.1). A Type 2 environment typically will not occur in crystalline igneous
and metamorphic rock (see discussion of likely hydrogeologic settings for Type 3 environments).
As with the Type 1 environment, abiotic reductive dechlorination processes may be
important. The importance of abiotic reductive dechlorination relative to dehalorespiration is
dependant upon the amount of biologically available Fe(III) and sulfate in the system. A
potentially important consideration for Type 2 environments is that the ambient redox conditions
are well established and iron sulfide minerals are already precipitated prior to a chlorinated
solvent release occurring. The iron sulfides are aged over time to produce different mineral
species with varying oxidation states. Recent work suggests that aged iron sulfides may show
reduced reactivity to some chlorinated organic compounds (e.g., TCE) and that the oxidation
state of iron sulfide minerals likely will strongly influence the potential for abiotic reductive
dechlorination (Butler and Hayes, 2001).
4.3 TYPE 3 ENVIRONMENT: AEROBIC SYSTEMS DUE TO LACK OF
FERMENTATION SUBSTRATES
A Type 3 environment is characterized by a well-oxygenated groundwater system with little
or no organic matter. Concentrations of dissolved oxygen typically are greater than 1.0 mg/L. In
such an environment, dehalorespiration will not occur and chlorinated solvents such as PCE,
TCE, TCA, and CT will not be reductively dechlorinated. In this environment, very long
aqueous-phase solvent plumes can form. The most significant natural attenuation mechanisms
for PCE and TCE will likely be dispersion and sorption. However, VC and DCE can be rapidly
oxidized under these conditions. A Type 3 environment is often found in crystalline igneous and
metamorphic rock (fractured or unfractured) such as basalt, granite, schist, phyllite and also in
glacial outwash deposits, eolian deposits, thick deposits of well-sorted, clean, beach sand with no
associated peat or other organic carbon deposits, or any other type of deposit with inherently low
organic carbon content if no anthropogenic carbon has been released.
WSRC-TR-2003-00331
December 31, 2004
4-11
Two conceptual models are provided for environments in which Type 3 behavior occurs. For
sources with PCE and TCE, the major natural attenuation processes are dilution and dispersion
alone (no biodegradation). As shown in Figure 4.2, the PCE and TCE plumes extend from the
source zone and concentrations are slowly reduced by dilution, dispersion, and sorption
processes. Chloride concentrations and oxidation-reduction potential will not change as
groundwater passes through the source zone and forms the chlorinated ethene plume. If TCA is
the solvent of interest, significant abiotic hydrolysis and dehydrohalogenation may occur,
resulting in a more rapid decrease in TCA concentrations and an increase in chloride
concentrations with production of 1,1-DCE.
In Figure 4.3, a source releases VC and 1,2-DCA into the groundwater at a Type 3 site (an
unlikely occurrence as more highly chlorinated solvents are typically released at sites). Because
the VC and 1,2-DCA can be degraded aerobically, these constituents decline in concentration at
a significant rate. Chloride is produced, and a depression in dissolved oxygen concentration
similar to that occurring at fuel sites, is observed.
4.4 MIXED ENVIRONMENTS
Most chlorinated solvent plumes exhibit different types of behavior in different portions of the
plume. This generally is the result of the distribution of different geochemical environments
along the flow path. Figures 4.4 through 4.7 show several conceptual models for mixed
environments.
Mixed geochemical environments can be beneficial for natural biodegradation of chlorinated
solvent plumes. For example, Wiedemeier et al. (1996b) describe a plume at Plattsburgh AFB,
New York that exhibits Type 1 behavior in the source area and Type 3 behavior downgradient
from the source. For natural attenuation, this may be an efficient remediation scenario. PCE,
TCE, and DCE are reductively dechlorinated with accumulation of VC near the source area
(Type 1); then, VC is oxidized (Type 3) to carbon dioxide, either aerobically or via Fe(III)
reduction further downgradient and does not accumulate. Vinyl chloride is removed from the
system much faster under these conditions than under reducing conditions.
A less ideal variation of the mixed Type 1 and Type 3 environments is shown in the
conceptual model in Figure 4.4. An extended TCE and cis-1,2-DCE plume results because
insufficient fermentable carbon results in an anaerobic zone which is too short for complete
biodegradation. Therefore, TCE extends well into the aerobic zone where no biodegradation
occurs. A long DCE plume also extends into the aerobic zone, indicating insignificant direct
aerobic oxidation has occurred. While a long chloride plume will be observed, the short
anaerobic zone means much less methane is produced, allowing dilution/dispersion to limit the
extent of the methane plume.
PCE
TCE
TOC
Chloride
Redox
a
b
c
d
PCE
TCE
TOC
Chloride
Concentration or Mass Concentration or Mass
Back-
ground
Source
Area Downgradient
Distance Downgradient and
Direction of Groundwater Flow
D:\doe\report\lines of evidence\Figure 4.2.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.2
Conceptual Model of Type 3 Environment
for Chlorinated Solvent Plume due to
a PCE and TCE Release
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
4-12
(After RTDF, 1997)
Back-
ground
Source
Area Downgradient
Distance Downgradient and
Direction of Groundwater Flow
NO3-
NO3-
Chloride
a
b
c
d
TOC
DO
VC
1,2-DCA
DO TOC
Redox
Chloride
Concentration or Mass
Concentration or Mass
VC 1,2-
DCA
D:\doe\report\lines of evidence\Figure 4.3.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.3
Conceptual Model of Type 3 Environment
for Chlorinated Solvent Plume with
VC and 1,2-DCA
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
(After RTDF, 1997)
4-13
Back-
ground
Source
Area Downgradient
Type 3 Type 1 Type 3
Distance Downgradient and
Direction of Groundwater Flow
Redox
VC
cis-1,2-DCE
PCE
TCE
Redox
VC
PCE
TCE
Ethene
NO3-
DO
CH4
SO
DO
SO42-
NO3-
NO3-
Chloride
a
b
c
d
cis-1,2-DCE
CH4
4
Ethene
Chloride
BOD
Acetate
Acetate
BOD (e.g.,
methanol)
SO42-
DO
Concentration or Mass Concentration or Mass
D:\doe\report\lines of evidence\Figure 4.4.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.4
Conceptual Model of Mixed Environments with
Type 1 Environment in the Source Zone and
Type 3 Environment in the Downgradient
Portion of the Plume
Date: 12/17/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
4-14
(After RTDF, 1997)
FUEL
OIL
UNLEADED
GASOLINE DIESEL
Chips R Us
GROUNDWATER
FLOW DIRECTION
Not To Scale
3 Miles
RESIDUAL SOIL
CONTAMINATION LNAPL
TYPE III ENVIRONMENT
TYPE I ENVIRONMENT
PCE Only (No Daughter Products)
Dissolved Oxygen = 7 mg/L
Nitrate = 8 mg/L
Iron (II) < 0.01 mg/L
Sulfate = 75 mg/L
Methane < 0.001 mg/L
ORP = +225 mV
Fuel Hydrocarbons, PCE, TCE,
DCE, VC, and Ethene
Dissolved Oxygen = < 0.1 mg/L
Nitrate < 0.05 mg/L
Iron (II) = 25 mg/L
Sulfate < 0.01 mg/L
Methane = 15 mg/L
ORP = -200 mV
TYPE I
BEHAVIOR
BEGINS
Computer Chip
Manufacturing Plant
Bulk Fuel
Storage Facility
(System Reverts to Type 3
Environment Downgradient
From Storage Facility)
D:\doe\report\lines of evidence\Figure 4.5.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.5
Conceptual Site Model of Mixed Type 3/Type 1/
Type 3 Environments
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
4-15
DNAPL
Source
Groundwater
Flow
Dissolved PCE Only,
Aerobic Conditions
No Biodegradation
TCE, cis-1,2-DCE
and Vinyl Chloride Appear
Volatilization and
Photooxidation
DNAPL
Stringers
Aquifer
Aquitard
Not to Scale
DNAPL Pool Wetlands Sediment =
High Organic Carbon Content =
Biodegradation + Sorption = Excellent Bioreactor
Contaminant
Plume
Type II Behavior
(Redox = Methanogenic)
D:\doe\report\lines of evidence\Figure 4.6.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.6
Conceptual Model of a Plume
Discharging to Surface Water
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
4-16
Type 1 Environment
w/dehalorespiration
Dehalococcoides, etc. Present
DO
NO3
-
Fe2+
SO4
2-
H2S
CH4
Acetylene +/-
FeS
Dehalococcoides,
etc. Absent
Type 3 Environment
DO
NO3
- +/-
Fe2+ absent
SO4
2- +/-
H2S absent
CH4 absent
Acetylene absent
FeS absent
Type 1 Environment
w/abiotic degradation
DO
NO3
-
Fe2+
SO4
2-
H2S
CH4 +/-
Acetylene
FeS
SOURCE AREA - Comingled Solvents and
Anthropogenic Carbon (Petroleum
Hydrocarbons, Landfill Leachate, etc.)
cis-DCE
VC
cis-DCE
VC absent
cis-DCE +/-
VC absent
Groundwater Flow Direction
Transition Zone
+/- Constituent may or may not
be present
Elevated Concentration
Depleted Concentration
D:\doe\report\lines of evidence\Figure 4.7.ai
Westinghouse Savannah River MNA/EA Project
Figure 4.7
Example of a Plume that Grades from a
Type 1 Environment with Dehalorespiration
into a Type 1 Environment with Abiotic
Degradation Into a Type 3 Environment
Date: 12/27/04
Client: Westinghouse Savannah
River Company
Project Number: 031007
Revision Number: 0
4-17
WSRC-TR-2003-00331
December 31, 2004
4-18
A less common type of mixed environment occurs where a chlorinated solvent source zone in
a Type 3 environment produces a plume that extends downgradient to a zone in the aquifer
where fermentation substrates are supplied (Figure 4.5). For example, a downgradient leaking
underground storage tank could introduce benzene, toluene, ethylbenzene, and the xylenes
(BTEX) compounds into the chlorinated solvent plume, changing the environment to Type 1.
Because petroleum hydrocarbons tend not to migrate a significant distance, the environment
would likely revert to a Type 3 environment and the solvent plume will likely continue
migrating.
Another possibility is an upland plume discharging into a wetland. In this case a Type 3
environment would be converted to a Type 2 environment if a downgradient recharge zone
introduced naturally-occurring fermentation substrates to the subsurface (Figure 4.6).
Figure 4.7 shows a conceptual site model where the plume grades from a Type 1 environment
with both biological and abiotic reductive dechlorination into a Type 1 environment where
abiotic reductive dechlorination dominates into a Type 3 environment.
WSRC-TR-2003-00331
December 31, 2004
5-1
SECTION 5
REFERENCES
Albrechtsen, H.-J., Bjerg, P.L., Ludvigsen, L., Rugge, K., Christensen, T.H., 1999. An anaerobic
field injection experiment in a landfill leachate plume Grindsted, Denmark: 2. Deduction
of anaerobic methanogenic, sulfate- and Fe(III)-reducing redox conditions. Water Resour.
Res. 35:1247–1256.
Amann, R. I., W. Ludwig, and K.-H. Schleifer, 1995, Phylogenetic identification and in situ
detection of individual microbial cells without cultivation. Microbiological Reviews
59:143-169.
Amirbahman, A., Schoenenberger, R., Johnson, C.A., Sigg, L., 1998. Aqueous- and solid-phase
biogeochemistry of a calcareous aquifer system downgradient from a municipal solid
waste landfill Winterthur, Switzerland . Environ. Sci. Technol. 32:1933–1940.
Amonette, J.E., 2002, Methods for determination of mineralogy and environmental availability,
p. 153-197: In J.B. Dixon and D.G. Schulze (eds.), Soil Mineralogy with Environmental
Applications. SSSA Book Series No. 7. Soil Science Society of America, Madison, WI.
Amonette, J.E., and Sanders, R.W., 1994, Nondestructive techniques for bulk elemental analysis,
p. 1-48: In J.E. Amonette and L.W. Zelazny (eds). Quantitative Methods in Soil
Mineralogy, SSSA Miscellaneous Publication. Soil Science Society of America, Madison,
WI.
Amonette, J.E., and Zelazny, L.W. (eds), 1994. Quantitative Methods in Soil Mineralogy, SSSA
Miscellaneous Publication. Soil Science Society of America, Madison, WI.
Ashley, G.M., 1973, Impregnation of fine-grained sediments with a polyester resin: A
modification of Altemuller’s method: Journal of Sedimentary Petrology 43:298-301.
ASTM, 1998, E1943-98 Standard and Guide for Remediation by Natural Attenuation at
Petroleum Release Sites: American Society for Testing and Materials, Philadelphia, PA.
Barden, M.J., 2003, Practical Use of Statistics for Natural Attenuation Trend Analysis.
Environmental Institute for Continuing Education Online Seminar EST-0101. June,
2003.
Bigham, J.M., Fitzpatrick, R.W., and Schulze, D.G., 2002, Iron oxides, p. 323-366: In J.B. Dixon
and D.G. Schulze (eds.), Soil Mineralogy with Environmental Applications. SSSA Book
Series No. 7. Soil Science Society of America, Madison, WI.
WSRC-TR-2003-00331
December 31, 2004
5-2
Bradley, P.M., and Chapelle, F.H., 1996, Anaerobic mineralization of vinyl chloride in Fe(III)-
reducing aquifer sediments: Environmental Science and Technology, v. 30, p. 2084 -
2086.
Bradley, P.M., and Chapelle, F.H., 1997, Kinetics of DCE and VC mineralization under
methanogenic and Fe(III)-reducing conditions: Environmental Science and Technology,
v. 31, p. 2692 - 2696.
Bradley, P.M., Chapelle, F.H., and Vroblesky, D.A., 1993. Does lead affect microbial-
metabolism in aquifer sediments under different terminal electron-accepting conditions.
Geomicrobiol. J. 11: 85–94.
Buol, S.W., and Faddness, D.M., 1961, A new method of impregnating fragile material for thin
sectioning. Soil Sci. Soc. Amer. Proc., v. 25, p. 253.
Burlage, R.S., 1998, Molecular techniques, p. 289-334: In R.S. Burlage, R. Atlas, D. Stahl, G.
Geesey and G. Sayler (eds.), Techniques in Microbial Ecology. Oxford University Press,
New York.
Butler, E. C., and K. F. Hayes, 1999, Kinetics of the transformation of trichloroethylene and
tetrachloroethylene by iron sulfide: Environmental Science and Technology, 33:2021-
2027.
Butler, E. C., and K. F. Hayes, 2001, Factors influencing rates and products in the transformation
of trichloroethylene by iron sulfide and iron metal: Environmental Science and
Technology, 35:3884-3891.
Cady, J.G., Wilding, L.P., and Drees, L.R., 1986, Petrographic microscope techniques, p. 185-
218: In A. Klute (ed.), Methods of Soil Analysis: Part 1, Physical and Mineralogical
Methods (Second Edition). SSSA Book Series No. 5 (Part 1). American Societry of
Agronomy and Soil Science Society of America, Madison, WI.
Chapelle, F.H., McMahon, P.B., Dubrovsky, N.M., Fujii, R.F., Oaksford, E.T., and Vroblesky,
D.A., 1995, Deducing the distribution of terminal electron-accepting processes in
hydrologically diverse groundwater systems: Water Resources Research, v. 31, p. 359-
371.
Chapelle, F.H., 1997, Identifying redox conditions that favor the natural attenuation of
chlorinated ethenes in contaminated groundwater systems, In Proceedings of the
Symposium on Natural Attenuation of Chlorinated Organics in Groundwater, Dallas, TX:
EPA/540/R-97/504, May 1997.
Chapelle, F.H., Vroblesky, D.A., Woodward, J.C., and Lovley, D.R., 1997, Practical
considerations for measuring hydrogen concentrations in groundwater: Environmental
Science and Technology, v. 31:2873-2877.
Christensen, T.H., Bjerg, P.L., Banwart, S.A., Jakobsen, R., Heron, G., and Albrechtsen, H-.J.,
2000. Characterization of redox conditions in groundwater contaminant plumes. Journal
of Contaminant Hydrology 45: 165–241.
WSRC-TR-2003-00331
December 31, 2004
5-3
Conover, W.J., 1999, Practical Nonparametric Statistics (3rd Edition): John Wiley & Sons, Inc.,
New York, 584 p.
Dawson K.J. and Istok, J.D., 1991, Aquifer Testing - Design and analysis of pumping and slug
tests: Lewis Publishers, Chelsea, Michigan, 344 p.
Deutsch, W.J., 1997, Groundwater Geochemistry: Fundamentals and Applications to
Contamination: Lewis Publishers, Boca Raton, FL. 221 p.
Dowling, N. J. E., F. Widdel, and D. C. White, 1986, Phospholipid ester-linked fatty acid
biomarkers of acetate-oxidizing sulfate reducers and other sulfide forming bacteria.
Journal of General Microbiology 132:1815-1825.
Drees, L.R., and Ransom, M.D., 1994, Light microscopic techniques in quantitative soil
mineralogy, p. 137-176: In J.E. Amonette and L.W. Zelazny (eds). Quantitative Methods
in Soil Mineralogy, SSSA Miscellaneous Publication. Soil Science Society of America,
Madison, WI.
Edlund, A., P. D. Nichols, R. Roffey, and D. C. White, 1985, Extractable and lipopolysaccharide
fatty acid and hydroxy acid profiles from Desulfovibrio species: Journal of Lipid Research
26:982-988.
Fanning, D.S., Rabenhorst, M.C., Burch, S.N., Islam, K.R., and Tangren, S.A., 2002, Sulfides
and sulfates, p. 229-260: In J.B. Dixon and D.G. Schulze (eds.), Soil Mineralogy with
Environmental Applications. SSSA Book Series No. 7. Soil Science Society of America,
Madison, WI.
Gibbons, R.D., 1994, Statistical Methods for Groundwater Monitoring: John Wiley & Sons, Inc.,
New York, 286 p.
Gilbert, R.O., 1987, “Statistical Methods for Environmental Pollution Monitoring,” Van
Nostrand Reinhold, New York.
GSI, 1998, Remediation by Natural Attenuation (RNA) ToolKit Users Manual: Groundwater
Services, Inc., Houston, Texas.
Guckert, J. B., C. P. Antworth, P. D. Nichols, and D. C. White, 1985, Phospholipid ester-linked
fatty acid profiles as reproducible assays for changes in prokaryotic community structure
of estuarine sediments: FEMS Microbiol. Ecol. 31:147-158.
Guckert, J. B., M. A. Hood, and D. C. White, 1986, Phospholipid ester-linked fatty acid profile
changes during nutrient deprivation of Vibrio cholerae: increases in the trans/cis ratio and
proportions of cyclopropyl fatty acids: Appl. Environ. Microbiol. 52:794–801.
Hansen, L.K., 1998. Biogeochemistry of Methane In A Shallow Sandy Aquifer. PhD
Dissertation, Technical University of Denmark, Lyngby, Denmark.
Hedrick, D.B., A Peacock, J.R. Stephen, S.J. Macnaughton, Julia Brüggemann, and David C.
White, 2000, Measuring soil microbial community diversity using polar lipid fatty acid
and denatured gradient gel electrophoresis data: J. Microbiol. Methods, 41, 235-248.
WSRC-TR-2003-00331
December 31, 2004
5-4
Helsel, D.R., and Hirsch, R.M., 2002, Statistical methods in water resources: U.S. Geological
Survey - Techniques of Water-Resources Investigations, Book 4, Ch. A3, 510 p.
[Available online only at http://water.usgs.gov/twri/twri4a3/]
Hem, J.D., 1985, Study and Interpretation of the Chemical Characteristics of Natural Water:
United States Geological Survey Water-Supply Paper 2254, 264 p.
Heron, G., Christensen, T.H., Tjell, J.C., 1994a. Oxidation capacity of aquifer sediments.
Environ. Sci. Technol. 28:153–158.
Heron, G., Crouzet, C., Bourg, A.C.M., Christensen, T.H., 1994b. Speciation of Fe(II) and
Fe(III) in contaminated aquifer sediments using chemical extraction techniques. Environ.
Sci. Technol. 28:1698–1705.
Hirsch, R.M., and Slack, J.R., 1984, A nonparameteric trend test for seasonal data with serial
dependence. Water Resources Research, vol. 20, p. 727-732.
Hirsch, R.M., Slack, J.R., and Smith, R.A., 1982, Techniques of trend analysis for monthly
quality data: Water Resources Research, vol. 18, p. 107-121.
Hollander, M., and Wolfe, D.A., 1999, Nonparametric Statistical Methods (2nd Edition): John
Wiley & Sons, Inc., New York, 787 p.
Hunkeler, D., Aravena, R., and Butler, B.J., 1999, Monitoring microbial dechlorination of
tetrachloroethene (PCE) in groundwater using compound-specific stable carbon isotope
ratios - microcosm and field studies: Environmental Science and Technology, v. 33, p.
2733-2738.
Innes, R.P., and Pluth, D.J., 1970, Thin section preparation using epoxy impregnation for
petrographic and electron microprobe analysis: Soil Sci. Soc. Amer. Proc. 34:483-485.
Johnson, A.I., 1967, Specific Yield – Compilation of Specific Yields for Various Materials.
United States Geological Survey Water-Supply Paper 1662-D, p. D1-D74.
Jones, R.C., and Malik, H.U., 1994, Analysis of minerals in oxide-rich soils by x-ray diffraction,
p. 296-329: In J.E. Amonette and L.W. Zelazny (eds). Quantitative Methods in Soil
Mineralogy, SSSA Miscellaneous Publication. Soil Science Society of America, Madison,
WI.
Karathanasis, A.D., and Hajek, B.F., 1996, Elemental analysis by x-ray fluorescence
spectroscopy, p. 161-223: In D.L. Sparks (ed.), Methods of Soil Analysis: Part 3,
Chemical Methods. SSSA Book Series No. 5 (Part 3). American Societry of Agronomy
and Soil Science Society of America, Madison, WI.
Kaufman, W.J. and Orlob, G.T., 1956, Measuring groundwater movement with radioactive and
chemical tracers: American Water Works Association Journal, v. 48, p. 559-572.
Kendall, M.G., 1975, Rank Correlation Methods, 4th ed. Charles Griffith, London.
WSRC-TR-2003-00331
December 31, 2004
5-5
Kennedy, L., and Everett, J, 2004, Draft Final Report, Field Test of Biogeochemical Reductive
Dechlorination at Dover Air Force Base, Dover Delaware: Air Force Center for
Environmental Excellence, San Antonio, Texas, December 2004.
Kennedy, L., Everett, J., and J. Gonzales, 2000, Aqueous and Mineral Intrinsic Bioremediation
Assessment (AMIBA) Protocol. U.S. Air Force Center for Environmental Excellence,
San Antonio, Texas.
Kennedy, L.G., Everett, J.W., Ware, K.J., Parsons, R., Green, V., 1998. Iron and sulfur mineral
analysis methods for natural attenuation assessments. Biorem. Jour. 2:259–276.
Kerr, P.F., 1977. Optical Mineralogy (4th edition). McGraw Hill, New York. 492 p.
Kruseman, G.P. and de Ridder N.A., 1991, Analysis and Evaluation of Pumping Test Data:
International Institute for Land Reclamation and Improvement, The Nederlands, 377 p.
Laird, D.A., and Dowdy, R.H., 1994, Preconcentration techniques in soil mineralogical analysis,
p. 236-266: In J.E. Amonette and L.W. Zelazny (eds). Quantitative Methods in Soil
Mineralogy, SSSA Miscellaneous Publication. Soil Science Society of America, Madison,
WI.
Ling, M., Rifai, H.S., Newell, C.J., Aziz, J.J., and Gonzales, J.R., 2003, Groundwater monitoring
plans at small-scale sites: An innovative spatial and temporal methodology: Journal of
Environmental Monitoring, vol. 5, p. 126-134.
Liu, W-.T., and Stahl, D., 2002, Molecular approaches for the measurement of density, diversity,
and phylogeny, p. 114-134: In C. J. Hurst, R.L. Crawford, G. R. Knudsen, M. J.
McInerney, and L. D. Stetzenbach (eds.), Manual of Environmental Microbiology, Second
Edition. ASM Press, Washington, DC.
Loeppert, R.H., and Inskeep, W.P., 1996, Iron, p. 639-664: In D.L. Sparks (ed.), Methods of Soil
Analysis: Part 3, Chemical Methods. SSSA Book Series No. 5 (Part 3). American Societry
of Agronomy and Soil Science Society of America, Madison, WI.
Lovley, D.R., and Goodwin, S., 1988, Hydrogen concentrations as an indicator of the
predominant terminal electron-accepting reaction in aquatic sediments: Geochimica et
Cosmochimica Acta, v. 52, p. 2993-3003.
Lovley, D.R., F.H. Chapelle, and J.C. Woodward, 1994, Use of dissolved H2 concentrations to
determine distribution of microbially catalyzed redox reactions in anoxic groundwater.
Environmental Science and Technology, v. 28, no. 7, p. 1205-1210.
Mann, H.B., 1945, Nonparametric tests against trend: Econometrica, vol. 13, p. 245-259.
Mann, H.B. and Whitney, D.R., 1947, “On a test of whether one or more random variables is
stochastically larger than in the other,” Ann. Math. Sciences, 18, pp. 52-54.
Mc Bride, M.B., 1986, Magnetic methods, p. 219-270: In A. Klute (ed.), Methods of Soil
Analysis: Part 1, Physical and Mineralogical Methods (Second Edition). SSSA Book
Series No. 5 (Part 1). American Societry of Agronomy and Soil Science Society of
America, Madison, WI.
WSRC-TR-2003-00331
December 31, 2004
5-6
Morel, F.M.M. and Hering, J.G., 1993, Principles and Applications of Aquatic Chemistry: John
Wiley & Sons, Inc., New York
National Research Council, 1993, In Situ Bioremediation, When Does it Work?: National
Academy Press, Washington, D.C., 207 p.
Pinkart, H.C., Ringelberg, D.B., Piceno, Y.M., MacNaughton, S.J., and White, D.C., 2002,
Biochemical approaches to biomass measurements and community structure analysis, p.
101-113: In C. J. Hurst, R.L. Crawford, G. R. Knudsen, M. J. McInerney, and L. D.
Stetzenbach (eds.), Manual of Environmental Microbiology, Second Edition. ASM Press,
Washington, DC.
Reed, S. J. B., 1996. Electron Microprobe Analysis and Scanning Electron Microscopy in
Geology. Cambridge University Press. 215 p.
Ribbe, P.H. (ed.), 1974, Sulfide Mineralogy. Mineralogical Society of America Short Course
Notes, v. 1. Southern Printing Company, Blacksburg, VA.
Ringelberg, D. B., G. T. Townsend, K. A. DeWeerd, J. M. Sulita, and D. C. White, 1994,
Detection of the anaerobic dechlorinating microorganism Desulfomonile tiedjei in
environmental matrices by its signature lipopolysaccharide branch-long-chain hydroxy
fatty acids. FEMS Microbiol. Ecol. 14:9-18.
RTDF, 1997, Natural Attenuaton of Chlorinated Solvents in Groundwater Seminar, Class Notes,
Remediation Technologies Development Forum
Sawhney, B.L., 1986, Electron Microprobe analysis, p. 271-290: In A. Klute (ed.), Methods of
Soil Analysis: Part 1, Physical and Mineralogical Methods (Second Edition). SSSA Book
Series No. 5 (Part 1). American Societry of Agronomy and Soil Science Society of
America, Madison, WI.
Schulze, D.G., 1994, Differential x-ray diffraction analysis of soil minerals, p. 412-429: In J.E.
Amonette and L.W. Zelazny (eds). Quantitative Methods in Soil Mineralogy, SSSA
Miscellaneous Publication. Soil Science Society of America, Madison, WI.
Slater, G.F., Lollar, B.S., Sleep, B.E., Edwards, E.A., 2001, Variability in carbon isotopic
fractionation during biodegradation of chlorinated ethenes – implications for field
applications: Environmental Science and Technology, v. 35, p. 901-907.
Spry, P.G. and Gedlinske, B.L., 1987. Tables for the determination of common opaque minerals:
Economic Geology, 52 p.
Stumm, W. and Morgan, J.J., 1996, Aquatic Chemistry, Chemical Equilibria and Rates in Natural
Waters: Wiley-Interscience, New York, New York, 1022 p.
Tighe, S.W., de Lajudie, P., Dipietro, K., Lindström, K., Nick, G. & Jarvis, B.D.W., 2000,
Analysis of cellular fatty acids and phenotypic relationships of Agrobacterium,
Bradyrhizobium, Mesorhizobium, Rhizobium and Sinorhizobium species using the
Sherlock Microbial Identification System: Int J Syst Evol Microbiol 50, 787-801.
WSRC-TR-2003-00331
December 31, 2004
5-7
Tuccillo, M.E., Cozzarelli, I.M., and Herman, J.S., 1999. Iron reduction in the sediments of a
hydrocarbon-contaminated aquifer. Appl. Geochem. 14:655–667.
Tsitko, I.V. Zaitsev, G.M., Lobanok, A.G., and Salkinoja-Salonen, M.S., 1999, Applied and
Environmental Microbiology 65(2):853-855.
USEPA, 1992, Methods for evaluating attainment of cleanup standards, volume 2, Groundwater:
EPA/230-R-92-014.
USEPA, 1998, Technical Protocol for Evaluating Natural of Chlorinated Solvents in
Groundwater: EPA/600/R-98/128, September, 1998.
US Environmental Protection Agency (USEPA), 1999, Use of Monitored Natural Attenuation at
Superfund, RCRA Corrective Action, and Underground Storage Tank Sites, Final: Office
of Solid Waste and Emergency Response. April 21. Directive Number 9200.
USEPA, 2000. Guidance for Data Quality Assessment, Practical Methods for Data Analysis.
U.S. Environmental Protection Agency, Office of Environmental Information,
Washington, DC. EPA QA/G-9 (QA00 Update).
Vandamme P, Pot B, Gillis M, de Vos P, Kersters K, Swings J., 1996, Polyphasic taxonomy, a
consensus approach to bacterial systematics: Microbiol Rev 1996 Jun;60(2):407-38.
White, D. C., W. M. Davis, J. S. Nickels, J. D. King, and R. J. Bobbie, 1979, Determination of
the sedimentary microbial biomass by extractable lipid phosphate: Oecologia 40:51-62.
White, D. C., H. C. Pinkart, and D. B. Ringelberg, 1997, Biomass measurements: Biochemical
approaches, p. 91-101: In C. J. Hurst, G. R. Knudsen, M. J. McInerney, L. D. Stetzenbach,
and M. V. Walter (eds.), Manual of Environmental Microbiology. ASM Press,
Washington, DC.
White, D. C., and D. B. Ringelberg, 1995, Utility of signature lipid biomarker analysis in
determining in situ viable biomass, community structure, and nutritional / physiological
status of the deep subsurface microbiota: In P. S. Amy and D. L. Halderman (ed.), The
microbiology of the terrestrial subsurface. CRC Press, Boca Raton.
White, D. C., and Ringelberg, D. B., 1998, Signature lipid biomarker analysis, p. 255-272: In
R.S. Burlage, R. Atlas, D. Stahl, G. Geesey and G. Sayler (eds.), Techniques in Microbial
Ecology. Oxford University Press, New York.
White, D. C., J. O. Stair, and D. B. Ringelberg, 1996, Quantitative comparisons of in situ
microbial biodiversity by signature biomarker analysis: Journal of Industrial Microbiology
17:185-196.
Wiedemeier, T.H., Wilson, J.T., Kampbell, D.H., Miller, R.N., and Hansen, J.E., 1995,
Technical protocol for implementing intrinsic remediation with long-term monitoring for
natural attenuation of fuel contamination dissolved in groundwater: US Air Force Center
for Environmental Excellence, San Antonio, Texas.
Wiedemeier, T.H., Swanson, M.A., Moutoux, D.E., Wilson, J.T., Kampbell, D.H., Hansen, J.E.,
and Haas, P., 1996a, Overview of the Technical Protocol for Natural Attenuation of
WSRC-TR-2003-00331
December 31, 2004
5-8
Chlorinated Aliphatic Hydrocarbons under Development for the U.S. Air Force Center for
Environmental Excellence: EPA/540/R-96/509, p. 35-59.
Wiedemeier, T.H., Swanson, M.A., Wilson, J.T., Kampbell, D.H., Miller, R.N., and Hansen,
J.E., 1996b, Approximation of biodegradation rate constants for monoaromatic
hydrocarbons (BTEX) in groundwater: Groundwater Monitoring and Remediation, v. 16,
no. 3, Summer 1996, p. 186-194.
Wiedemeier, T.H., Swanson, M.A., Moutoux, D.E., Gordon, E.K., Wilson, J.T., Wilson, B.H.,
Kampbell, D.H., Hansen, J.E., Haas, P., and Chapelle, F.H., 1998, Technical Protocol for
Evaluating the Natural Attenuation of Chlorinated Solvents Dissolved in Groundwater,
Air Force Center for Environmental Excellence, San Antonio, Texas.
Wiedemeier, T.H., Rifai, H.S., Newell, C.J., and Wilson, J.T., 1999, Natural Attenuation of Fuels
and Chlorinated Solvents in the Subsurface: John Wiley & Sons, New York, New York,
617 p.
Wiedemeier, T.H., and Haas, P.E., 2002, Designing Monitoring Programs to Evaluate the
Performance of Natural Attenuation: Ground Water Monitoring and Remediation,
Summer 2002.
