United States
Environmental Protection
Agency
Monitored Natural
Attenuation of MTBE as a
Risk Management Option at
Leaking Underground
Storage Tank Sites
oooooooooooo
o»oo»oo»oo»o
oooooooooooo
oooooooooooo
o«oo»oo«oo«o
oooooooooooo
oooooooooooo
o»oo»oo»oo«o
QQOO
1°
b
oooooooooooo
o»oo»oo»oo»o
oooooooooooo
oooooooooooo
o»oo»oo«oo«o
oooooooooooo
oooooooooooo
o»oo»oo»oo«o
oooooooooooo
oooooooooooo
o»oo«oo«oo«o
oooooooooooo
oooooooooooo
o«oo»oo«oo»o
oo»oooooo»oo
000*0000»ooo
o«oo»oo«oo»o
oooooooooooo
oooooooooooo
o«oo«oo«oo»o
ooo«oooo»ooo
oo»o00000*00
o»oo»oo»oo«o
oooooooooooo
•oooo«ooooo«
0«00«00«00«0
oo»oooooo«oo
ooo»oooo«oo
o«oo«oo«oo»o
•ooooo»oooo»
•00*0*
o«oo»o
•o»ooo
ooo»o«
o«oo»o
•ooo«o
-------
EPA/600/R-04/179
January 2005
Monitored Natural Attenuation of MTBE
as a Risk Management Option at Leaking
Underground Storage Tank Sites
John T. Wilson, Philip M. Kaiser and Cherri Adair -
U.S. Environmental Protection Agency
Office of Research and Development
National Risk Management Research Laboratory -
Ada, Oklahoma 74820 -
Project Officer -
John T. Wilson -
Ground Water and Ecosystems Restoration Division
National Risk Management Research Laboratory -
Ada, Oklahoma 74820 -
National Risk Management Research Laboratory -
Office of Research and Development -
U.S. Environmental Protection Agency -
Cincinnati, OH 45268 -
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and
Development funded the research described here. It has been subjected to the
Agency's peer and administrative review and has been approved for publication
as an EPA document. Mention of trade names and commercial products does not
constitute endorsement or recommendation for use.
All research projects making conclusions and recommendations based on environ-
mentally related measurements and funded by the U.S. Environmental Protection
Agency are required to participate in the Agency Quality Assurance Program. This
project was conducted under two Quality Assurance Project Plans. Work performed
by U.S. EPA employees or by the U.S. EPA on-site analytical contractor followed
procedures specified in these plans without exception. Information on the plans
and documentation of the quality assurance activities and results is available from
John Wilson or Cherri Adair.
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's land, air, and water
resources. Under a mandate of national environmental laws, the Agency strives to formulate and implement actions
leading to a compatible balance between human activities and the ability of natural systems to support and nurture
life. To meet this mandate, EPA's research program is providing data and technical support for solving environmental
problems today and building a science knowledge base necessary to manage our ecological resources wisely, under-
stand how pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for investigation of techno-
logical and management approaches for preventing and reducing risks from pollution that threatens human health
and the environment. The focus of the Laboratory's research program is on methods and their cost-effectiveness for
prevention and control of pollution to air, land, water, and subsurface resources; protection of water quality in public
water systems; remediation of contaminated sites, sediments and ground water; prevention and control of indoor
air pollution; and restoration of ecosystems. NRMRL collaborates with both public and private sector partners to
foster technologies that reduce the cost of compliance and to anticipate emerging problems. NRMRL's research
provides solutions to environmental problems by: developing and promoting technologies that protect and improve
the environment; advancing scientific and engineering information to support regulatory and policy decisions; and
providing the technical support and information transfer to ensure implementation of environmental regulations and
strategies at the national, state, and community levels.
In the United States of America, the responsibility for managing spills of gasoline from underground storage tanks
falls to the individual states. Where it has been appropriate, many states have selected monitored natural attenuation
as a remedy for organic contaminants in ground water. Many states also use a formal process of risk management
to select the most appropriate remedy at gasoline spill sites. Both monitored natural attenuation (MNA) and risk
management require an understanding of the environmental processes that control the behavior of a contaminant in
ground water. This report is intended for technical staff in the state agencies with responsibility for administering the
underground storage tank program as mandated by RCRA. The information is intended to allow the state regulators
to determine whether they have adequate information to evaluate MNA of fuel oxygenates at a site, and to allow the
regulators to separate sites where MNA of fuel oxygenates may be an appropriate risk management alternative from
sites where MNA is not appropriate.
Stephen G. Schmelling, Director
Ground Water and Ecosystems Restoration Division
National Risk Management Researchr Laboratory
-------
Abstract
This report reviews the current state of knowledge on the transport and fate of
MTBE in ground water, with emphasis on the natural processes that can be used
to manage the risk associated with MTBE in ground water or that contribute to
natural attenuation of MTBE as a remedy. It provides recommendations on the site
characterization data that are necessary to manage risk or to evaluate monitored
natural attenuation (MNA) of MTBE, and it illustrates procedures that can be used
to work up data to evaluate risk or assess MNA at a specific site.
-------
Contents
Notice ii •
Foreword iii •
Abstract iv •
Acknowledgments xi •
Section 1 Relationship Between Risk Management and MNA 1 •
1.1 Monitored Natural Attenuation and Risk Management 1 •
1.2 Suggestions to Improve Plume Management and Risk Evaluation 2
1.3 Suggestions to Improve MNA to Meet a Cleanup Goal 2
Section 2 Typical Behavior of MTBE Plumes 3
2.1 Concentration of MTBE Expected in Gasoline Spills 3
2.2 Definitions and Expressions Used for Rate Constants 5
2.3 Role of Attenuation in the Lifecycle of MTBE Plumes 6
Section 3 Recommendations for Monitoring 11
3.1 Biogeochemical Footprints to Evaluate Coverage of Monitoring Wells 12
3.2 Monitoring the Direction of Ground Water Flow 13
3.3 Monitoring to Predict Biological Processes 16
3.4 Concerns with Analytical Issues 16
Section 4 Biological Degradation 19
4.1 Microbiology of Aerobic MTBE Biodegradation 19
4.2 Biochemistry of Aerobic MTBE Biodegradation 21
4.3 Aerobic MTBE Biodegradation in Ground Water 22
4.4 Anaerobic Biodegradation of MTBE 23
4.5 Acclimation to Anaerobic Biodegradation of MTBE 26
4.6 Zero Order Biodegradation of MTBE at High Concentrations 29
Section 5 Monitoring MTBE Biodegradation with Stable Isotope Ratios 31
5.1 Monitoring MTBE Biodegradation with Stable Isotope Ratios 31
5.2 Predicting Biodegradation of 513C in MTBE in Gasoline 34 •
5.3 Sources of Uncertainty in Estimates of Biodegradation 35 •
5.4 A Conservative Estimate of the Extent of Biodegradation 36 •
Section 6 Application of Stable Isotope Ratios to Interpret Plume Behavior 37 •
6.1 Application of Stable Isotope Ratios to Interpret Plume Behavior 37 •
6.2 Using Stable Carbon Isotope Ratios to Recognize Natural Biodegradation 38 •
6.3 Using Stable Carbon Isotope Ratios to Estimate the Projected Rate of Natural
Biodegradation 40
6.4 Using the Projected Rate of Biodegradation to Estimate the Length of Plumes 41
6.5 Using 513C to Distinguish the Source of TEA in Ground Water 42 •
6.6 Caveats and Limitations Concerning the Use of 513C of MTBE to Estimate
Biodegradation 44
-------
Section 7 Statistical Evaluation of Rates of Attenuation of Sources 47
7.1 Risk Management and U.S. EPA Expectations for MNA 47
Section 8 Typical Rates of Attenuation in Source Areas 55
8.1 Typical Rates of Attenuation over Time in Source Areas 55
8.2 Number of Sampling Dates Needed to Calculate Rates of Attenuation 56
8.3 Effect of Number of Sampling Dates on the Detectable Rate of Attenuation 58
8.4 Effect of Seasonal Variations 58
Section 9 Quality Assurance Statement 61
9.1 Analysis of Concentrations in Water 61
9.2 Stable Carbon Isotope Analyses 62
Section 10 References 69
-------
Figures
Figure 2.1 Comparison of the distribution of the maximum concentration of MTBE at gasoline spill
sites in regions of the United States that used MTBE to meet the federal oxygenate
standard (Texas and Southern California) to a region that did not (Kansas) 3 -
Figure 2.2 Relationship between a first order rate constant and a half life 6 -
Figure 2.3 Evolution of a plume of MTBE at Parsippany, New Jersey, that is receding back on itself. ...8 -
Figure 2.4 Evolution of a plume of MTBE in Laurel Bay, South Carolina, where the "hot spot" has
detached from the source and moved down gradient 9 -
Figure 3.1 Variation in the direction and magnitude of ground water flow at an MTBE site in
Elizabeth City, North Carolina 14 -
Figure 3.2 Variation in the direction of ground water flow at gasoline spill sites in Texas 15 -
Figure 3.3 Association of sulfate and methane with natural anaerobic biodegradation of MTBE
in ground water collected from 61 monitoring wells at thirteen gasoline spill sites in
Orange County, California 17 -
Figure 3.4 A change in the relative concentration of TEA and MTBE in a monitoring well as evidence
for natural biodegradation of MTBE 18 -
Figure 4.1 Significant products of the aerobic biodegradation of MTBE 21 -
Figure 4.2 Distribution of TEA and MTBE in the most contaminated wells at gasoline spill sites in
Orange County, California, in 2002 26 -
Figure 4.3 Transition from MTBE to TEA in monitoring wells at gasoline spill sites in Orange
County, California 27 -
Figure 4.4 Anaerobic biodegradation of MTBE and production of TEA in microcosms constructed
with sediment from a gasoline spill site 28 -
Figure 4.5 Anaerobic biodegradation of MTBE at high concentration in an enrichment culture
constructed with sediment from the microcosms used to produce the data in Figure 4.4 30 -
Figure 5.1 An illustration of the kinetic isotope effect 31 -
Figure 5.2 Typical changes in the value of 513C as MTBE is degraded under aerobic and anaerobic
conditions 33 -
Figure 5.3 The data from Figure 5.2 have been plotted in units commonly used in the literature on
stable isotopes 34 -
Figure 5.4 MTBE biodegradation under anaerobic conditions predicted from the 513C of MTBE in
ground water. 36 -
Figure 6.1 Concentration of MTBE (|ig/L) in selected monitoring wells at a gasoline spill site in
Dana Point, California, in August 2004 39-
Figure 6.2 Location of monitoring wells and water table elevations at a gasoline spill site in Newark,
Delaware, with high concentrations of TEA in the ground water. 43 -
-------
Figure 6.3 -Distribution of 513C of MTBE in ground water and the fraction of MTBE remaining from
biodegradation as calculated from the concentrations of MTBE and TEA in ground water
and the assumption that TEA was produced by biodegradation of MTBE 45
Figure 7.1 -The variance in monitoring data is often proportional to the concentration 48
Figure 7.2 -Location of monitoring wells in a plume of MTBE at Parsippany, New Jersey 49
Figure 8.1 -Distribution of the rates of attenuation of MTBE over time in source areas of plumes
from gasoline spills 56
Figure 8.2. Variation in the number of sampling dates in a data set that are required to extract a
rate of natural attenuation that is statistically significant 57
Figure 8.3. Potential for error when a short data set is used to estimate the rate of attenuation of
concentrations over time 57
Figure 8.4 -Effect of the number of samples used to calculate a rate of attenuation on the minimum
rate of attenuation that is statistically different from zero at 90% confidence 59
Figure 8.5 -A monitoring record from a well in Maryland with seasonal maximum concentrations of
MTBE in certain years 59 •
-------
Tables
Table 2.1. Rates of Attenuation of MTBE in Source Areas Over Time Contrasted to Rates of
Attenuation Along Flow Paths in Ground Water 7 -
Table 3.1 - Variation in the Standard Deviation of the Direction of Ground Water Flow at
132 Gasoline Stations in Texas 15 -
Table 4.1. - Comparison of the Growth Rate of Aerobic Bacteria During Growth on MTBE as the
Primary Substrate to the Growth Rate of Bacteria that Grow on Pentane and Fortuitously
Metabolizes MTBE and to the Growth Rate of Bacteria that Grow on BTX Compounds 20 -
Table 4.2 Distribution (in percent) of Biodegradation Products of MTBE under Nitrate-reducing,
Sulfate-reducing, Iron-reducing, and Methanogenic Conditions 23 -
Table 4.3. Anaerobic Biodegradation of MTBE, TEA, Benzene, and Ethanol in Microcosms
Constructed with Aquifer Sediment 25 -
Table 4.4 - Biodegradation of MTBE in the Most Contaminated Well at 13 Gasoline Spill Sites in
Orange County, California, as Predicted by Fractionation of 13C in MTBE 29 -
Table 5.1 - Fractionation of Black Dots and White Dots in the Visual Example in Figure 5.1 32 -
Table 5.2 - Typical Changes in the Ratio of 13C to 12C in MTBE During Biodegradation of MTBE
under Anaerobic Conditions 33 -
Table 6.1 - A Comparison Between the Distribution of MTBE and TEA in Ground Water
Contaminated by a Fuel Spill in Dana Point, California, and the Extent of MTBE
Biodegradation Predicted from the Stable Carbon Isotope Ratio (513C) of the
Residual MTBE 39-
Table 6.2 - Rates of Natural Biodegradation of MTBE Projected along a Flow Path in Ground
Water to Monitoring Wells 41 -
Table 6.3 - Relationship Between the Extent of Contamination and Biogeochemical Parameters at a
Site in Newark, Delaware 43 -
Table 6.4 - Concentrations of TEA Predicted from Biodegradation of MTBE to TEA at a Site in
Newark, Delaware 43 -
Table 7.1 - Relationship Between the Rate of Attenuation Necessary for Risk Management or for
Monitored Natural Attenuation, and the Achieved Rates of Attenuation During Long-term
Monitoring 49 -
Table 7.2 - Long-term Monitoring Data at a Gasoline Spill Site at Parsippany, New Jersey 50 -
Table 7.3 - Progress of Natural Attenuation of MTBE at a Gasoline Spill Site at Parsippany,
New Jersey 54 -
Table 9.1 - Typical Quality Performance Data for Continuing Calibration Check Standards for
MTBE in Water 63-
Table 9.2 Typical Quality Performance Data for Analysis of MTBE in Water, Including Blanks,
Laboratory Duplicates, and Matrix Spikes 64 -
-------
Table 9.3 - Typical Quality Performance Data for Continuing Calibration Check Standards for
TEA in Water. 65'
Table 9.4 - Typical Quality Performance Data for Analysis of TEA in Water, Including Blanks,
Laboratory Duplicates, and Matrix Spikes 66
Table 9.5 - Typical Quality Performance Data for Continuing Calibration Check Standards for
Benzene in Water 66
Table 9.6 - Typical Quality Performance Data for Analysis of Benzene in Water, Including Blanks,
Laboratory Duplicates, and Matrix Spikes 67
Table 9.7 Typical Quality Performance Data for Continuing Calibration Check Standards for
Ethanol in Water 68
Table 9.8 - Typical Quality Performance Data for Analysis of Ethanol in Water, Including Blanks,
Laboratory Duplicates, and Matrix Spikes 68
-------
Acknowledgments
Douglas Mackay at the University of California at Davis, Michael Hyman at North
Carolina State University, Theresa Evanson and Aristeo Pelayo with the Wiscon-
sin Department of Natural Resources, and Harold White with U.S. EPA Office of
Underground Storage Tanks provided formal peer reviews. Tomasz Kuder at the
University of Oklahoma, Ravi Kolhatkar with Atlantic Richfield Company, Patricia
Ellis with the State of Delaware Department of Natural Resources and Environmental
Control, and Matthew Small with U.S. EPA Region 9 provided technical reviews.
Seth Daugherty with the Local Oversight Program within the Environmental Health
Division of the Health Care Agency of Orange County, California, provided technical
reviews and valuable suggestions for the technical approach taken in the document
and the format and organization of the document.
-------
Section 1
Relationship Between Risk Management and MNA -
Monitored natural attenuation (or MNA) is defined by U.S. EPA in the OSWER Directive (U.S. EPA, 1999) as one
alternative means of achieving remediation objectives that may be appropriate for specific, well-documented site cir-
cumstances where its use meets the applicable statutory and regulatory requirements. The remedial objective may
be chemical-specific cleanup levels. The remedial objective may also include preventing exposure to contaminants,
preventing further migration of contaminants from source areas, preventing further migration of the groundwater con-
taminant plume, reducing contamination in soil or groundwater to specified cleanup levels appropriate for current or
potential future uses, or other objectives.
Natural attenuation processes, such as biodegradation and dispersion along a flow path, can bring the concentration of
contaminants to a chemical-specific cleanup level. This is particularly true when the source of contamination has been
controlled. The same natural attenuation processes can also prevent exposure to contaminants, prevent further migration
of contaminants from source areas, or prevent further migration of the ground water contaminant plume.
This section discusses the relationship between risk management and MNA, describes the common remedial objectives
for monitored natural attenuation, identifies the behavior of ground water plumes that is crucial for success in monitored
natural attenuation, and makes suggestions to improve the current state of practice for monitored natural attenuation of
methyl tertiary butyl ether (MTBE).
1.1 Monitored Natural Attenuation and Risk Management
In the United States of America, the responsibility for managing spills of gasoline from underground storage tanks falls
to the individual states. Where it has been appropriate, many states have selected monitored natural attenuation as a
remedy for organic contaminants in ground water (U.S. EPA 1999; New England Interstate Water Pollution Control
Commission; 2000, 2003).
Many states use a formal process of risk management to select the most appropriate remedy at gasoline spill sites. The
potential receptors of contamination are identified, and the behavior of the plume is characterized to determine the
potential for contamination to migrate along a flow path and impact the receptors. Many states estimate that risk with
mathematical formulas or mathematical models that describe the rate of transport of contaminants in ground water and
the rate of attenuation of the contaminant along the flow path through dilution and dispersion, sorption, and biodegra-
dation.
If the remediation objective for MNA is to prevent exposure or to prevent further migration of MTBE, the critical issue
is the distance MTBE can move in ground water. The size and long-term behavior of a plume of MTBE are controlled
by the rate of dissolution of MTBE from the residual gasoline in the source area and the rate of attenuation along the
flow path in the aquifer through biodegradation, dilution, and dispersion (Small and Weaver, 1999). This interaction is
described and illustrated in detail in Section 2 Typical Behavior of MTBE Plumes.
If the remedial objective is a specific cleanup level, the critical issue is the time required to reach the cleanup goal. In
most gasoline spills, the long-term source of MTBE in ground water is MTBE in residual gasoline trapped in the aqui-
fer. The time required to reach the clean up goal is controlled by the rate at which MTBE dissolves from the residual
gasoline into ground water as the ground water flows past the residual gasoline. If the MTBE dissolves rapidly, the
residual gasoline will be depleted of MTBE. As ground water from up gradient comes into contact with the depleted
gasoline, the concentration of MTBE that dissolves into the ground water will be less. The concentrations of MTBE
in water will drop rapidly. If the MTBE dissolves slowly, residual gasoline will be depleted slowly, and MTBE will
persist in the source area for long periods of time.
The OSWER Directive (U.S. EPA, 1999) requires that MNA will meet site remediation objectives within a time frame
that is reasonable compared to that offered by other methods. The progress toward achieving the remedial objective
is monitored until the remedial objective is obtained. When the remedial objective is to prevent exposure and prevent
-------
further migration of contaminants, many states will monitor the concentrations of contaminants until a statistical analysis
of the data reveals that the concentrations are declining over time at some predetermined level of confidence. When the
remedial objective is a specific cleanup level, many states will monitor a site until the cleanup level is met.
1.2 Suggestions to Improve Plume Management and Risk Evaluation
There is room for improvement in the current practice for risk evaluation of MTBE plumes in ground water. Most MTBE
plumes are anaerobic. Until recently, it was generally believed that MTBE would not degrade in anaerobic ground
water, and most risk assessments for MTBE have ignored the possibility that the MTBE might biodegrade. Recent work
has documented anaerobic MTBE biodegradation in laboratory microcosm studies. Section 3 Recommendations for
Monitoring discusses the monitoring needed to recognize and properly describe anaerobic MTBE biodegradation at field
scale. Section 4 Biological Degradation of MTBE and Other Fuel Oxygenates reviews the current state of knowledge
concerning the microbiology of MTBE biodegradation.
In the past, the primary evidence for biodegradation at field scale was the accumulation of degradation products. The
primary degradation product of MTBE is tertiary butyl alcohol (TEA). Because TEA has intentionally been added
to gasoline as a fuel oxygenate, and because it occurs as a trace component of commercial MTBE in gasoline, TEA
accumulation by itself is not convincing evidence of MTBE biodegradation. This makes it particularly difficult to use
conventional monitoring data to document biodegradation of MTBE at field scale, or to extract rate constants for at-
tenuation that can be used in predictions of the future behavior of plumes.
Recent work has shown the stable carbon isotopes in MTBE are fractionated when MTBE is biologically degraded
(Hunkeler et al., 2001; Gray et al., 2002; Kolhatkar et al., 2002; Kuder et al., 2005). As biodegradation proceeds, the
MTBE that has not been degraded has a progressively greater proportion of the heavy carbon isotope 13C, compared to
the more common isotope 12C. Advances in compound-specific stable isotope analyses make it possible to accurately
measure the shift in the ratio of the isotopes in MTBE in water at low concentrations. The fractionation of the MTBE
that has not degraded becomes the equivalent to a "metabolic product" that is used to document biodegradation. This
makes it possible for the first time to unequivocally identify and measure anaerobic biodegradation of MTBE at field
scale. Section 5 Monitoring MTBE Biodegradation with Stable Isotope Ratios explains the units used to measure carbon
isotope fractionation and discusses the simple formulas used to estimate the extent of biodegradation from the extent of
fractionation. Section 6 Application of Stable Isotope Ratios to Interpret Plume Behavior illustrates the use of stable
carbon isotope analyses to recognize anaerobic biodegradation of MTBE at field scale, to extract a rate constant of
biodegradation of MTBE at field scale, and to evaluate the contribution of MTBE biodegradation to the concentrations
of TEA measured at a gasoline spill site.
1.3 Suggestions to Improve MNA to Meet a Cleanup Goal
There is also room for improvement in the current practice to evaluate the rate of natural attenuation over time. Often
the monitoring data are presented to state regulators as a simple chart or table without any statistical evaluation of the
data. If the data are examined, the evaluation is often cursory and incomplete. The rate of attenuation over time is
conventionally estimated as the slope of a linear regression of the monitoring data on the date of sampling. The report
may provide the regulator with a chart showing the regression line. It may also provide the correlation coefficient
(r2) of the regression as an indication of the variability of the data. The value of r2 in itself is not a test for statistical
significance. The monitoring data should be evaluated to determine if the concentrations are actually declining. More
specifically, the data should be evaluated to determine if the slope of the regression line is statistically significant from
zero at some predetermined level of confidence.
Section 7 Statistical Evaluation of Rates of Attenuation of Sources provides detailed step-by-step instructions to extract
a rate of attenuation from field data and to evaluate the data to determine whether the rate is statistically significant
from zero. The approach that is illustrated relies on conventional parametric statistics. It is the approach that is most
likely to be familiar and accessible to a ground water scientist or engineer that does not have extensive experience with
environmental statistics.
Section 8 Typical Rates of Attenuation in Source Areas presents data on typical rates of attenuation in the source area
of selected MTBE plumes. The typical rates provide a benchmark for the relative rate of attenuation at a particular site
of interest. They also provide a realistic view of the prospects for rapid natural attenuation of MTBE in the source area
of plumes in general. This section also provides information on the number of samples that are typically necessary to
determine a rate of attenuation that is statistically significant.
-------
Section 2 -
Typical Behavior of MTBE Plumes -
This section describes the maximum concentrations of MTBE in the source areas of plumes from gasoline spills. The
concentration of MTBE in the source area has a strong influence on the length of a plume at steady state and on the
prospects that the concentrations will decline to meet a goal for cleanup. This section also describes the typical behavior
of the plumes in terms of the rate of natural decline in concentrations of MTBE over time in the source area and the
rate of natural decline in concentration of MTBE in ground water along the flow path. The decline in concentration
along the flow path is the critical behavior of a plume that determines its ability to impact a receptor. The decline in
concentration over time is the critical behavior of a plume that determines its ability to reach a cleanup goal.
2.1 Concentration of MTBE Expected in Gasoline Spills
The concentrations of MTBE in ground water at gasoline spill sites are much lower than would be expected from typical
concentrations of MTBE in gasoline. Figure 2.1 compares the distribution of the maximum concentration of MTBE in
monitoring wells in southern California, Texas, and Kansas. Oxygenated gasoline is not required in Kansas; however,
it is required in California and in the Dallas/Fort Worth and Houston markets in Texas. The data from Texas are the
maximum concentrations of MTBE at 609 gasoline spill sites that had at least one analysis for MTBE in monitoring
wells at the site (Mace and Choi, 1998). The data from California are the maximum MTBE concentrations at gasoline
spill sites in Orange County, California, in 2002 (data courtesy Seth Daugherty, Orange County Local Oversight Pro-
gram, compare Odencrantz, 1998), and in Los Angeles County, California, in 2002 (Shih et al., 2004). The data from
Kansas are the maximum concentration reported at sites in Kansas UST trust fund 2003 (Hattan et al., 2003). The data
presented in Figure 2.1 are the maximum concentrations of MTBE in any well at the site within a particular year of
monitoring. They are not the maximum concentrations that have ever been recorded at the sites.
10000
jfi
= 1000 -
en
S 100
10
1 -
o n 1
O
E 0.01
'8
0.001
Los Angeles
County, California
Texas
Kansas
Orange County,
California
20
40
60
80
100
Figure 2.1 Comparison of the distribution of the maximum concentration of MTBE at gasoline spill sites in regions
of the United States that used MTBE to meet the federal oxygenate standard (Texas and Southern Califor-
nia) to a region that did not (Kansas). The dashed line represents the concentration of MTBE that would
be expected in ground water in contact with residual gasoline at 1,000 mg/kg when the gasoline contained
11% MTBE. The concentrations of MTBE are similar in regions that used MTBE as the fuel oxygenate.
In a region that does not require a fuel oxygenate, the concentrations are lower. The concentrations are
less than would be expected based on dissolution of MTBE from reformulated gasoline.
-------
The frequency distribution of MTBE in ground water in the three regions of the United States that require an oxygenate
in gasoline was very similar. Approximately 25% of sites had concentrations less than 0.1 mg/L, the median concentra-
tion was near 1 mg/L, approximately 25% of sites had concentrations above 10 mg/L, and approximately 5% of sites
had concentrations above 100 mg/L. The frequency distribution of concentrations of MTBE in ground water in Kansas
was similar to the distribution in Texas and southern California, but the concentrations of MTBE in ground water in
Kansas are on the order of three-fold less than concentrations in these areas where reformulated gasoline is required
(Figure 2.1).
The equilibrium partitioning of MTBE between gasoline and water was used to provide a basis for comparison between
the actual concentration of MTBE in ground water and the concentration that would be expected from a spill of
reformulated gasoline.
w v , 9W Equation 2.1
NAPL
Equation 2.1 calculates the concentration of MTBE in the ground water Cw, from the concentration of MTBE in the
gasoline that was spilled C0 NAPL, the porosity filled with gasoline 9NAPL, the water-filled porosity 9W , and the distribu-
tion coefficient between gasoline and water KNAPL. The specific gravity of gasoline is near 0.78, which is equivalent to
780,000 mg/L. If the gasoline contains 11 % MTBE by volume, then the concentration of MTBE in the gasoline C0 NAPL
is 85,800 mg/L. Assume the concentration of residual gasoline in the aquifer sediment is 1,000 mg/kg, and the bulk
density of the sediment is 1.7 kg/L. The concentration of gasoline in a volume of aquifer material would be 1,700 mg/L.
If the density of gasoline is 780,000 mg/L, the porosity filled with gasoline 9NAPL is 0.00218 L/L. At a bulk density of
1.7 kg/L, the total porosity is 30% of the volume of the aquifer material. Because the porosity filled with gasoline is so
small, the water-filled porosity 9W will be near 0.30. The gasoline to water partition coefficient of MTBE is assumed
to be 16 (Cline et al., 1991; Rixey and Joshi, 2000).
_ 85,800mg/L
w ~ ., 0.30 Equation 2.1 solved
16 -I
0.00218
Under these assumptions, the predicted concentration of MTBE in ground water in contact with the residual gasoline
is 560 mg/L. This prediction is the horizontal dashed line in Figure 2.1.
The assumed concentration of residual gasoline in the aquifer was 1,000 mg/kg TPH. This is a relatively low value for
residual gasoline at spill sites. Most gasoline spills would have higher concentrations of gasoline at residual saturation
which would produce higher concentrations of MTBE in water. The measured concentrations of MTBE in the most
contaminated monitoring wells in Texas and southern California are from two to three orders of magnitude less than
the concentrations that would be expected from the content of MTBE in oxygenated reformulated gasoline.
There are a number of explanations why the measured concentrations of MTBE are lower than the expected concen-
trations. Not every gasoline spill in these data sets contained 11% MTBE. Many sites have been subjected to active
remediation. In some sites, the rate of dissolution of MTBE from the residual gasoline to ground water is limited by
mass transfer phenomena. As a consequence, the MTBE in the residual gasoline is not in equilibrium with the MTBE
in water.
Because the measured concentrations of MTBE at gasoline spill sites are so much lower than the expected concentra-
tions, the prospects of reaching clean up goals are more attainable. The length of a plume at equilibrium is related to the
concentration at the source. Because the concentrations are lower, the possibility of impacting a receptor is less. The
lower range of the U.S. EPA health advisory is 20 |ig/L. If half the MTBE plumes in those areas of the United States
that used MTBE as a fuel oxygenate have a maximum concentration of 1,000 |ig/L or less (see Figure 2.1), then half
the plumes in their source areas are within a factor of fifty or less of the lower range of the U.S. EPA health advisory.
If 25% have a maximum concentration of 100 |ig/L or less (Figure 2.1), then 25% are within a factor of five or less of
the lower range.
Kansas is probably representative of the regions of the United States where MTBE is not intentionally added to gaso-
line to meet the federal oxygenate standard. However, it should not be surprising to find MTBE in gasoline spills in
-------
Kansas. Gasoline may contain MTBE for a variety of reasons. The gasoline may have been refined for another market
where an oxygenate is required, or MTBE may have been added to meet octane requirements for the fuel, or MTBE
may have entered the gasoline through incidental blending with other products or feed stocks that contained MTBE
during refining and distribution.
To determine the amount of MTBE in the fuel supply in Kansas, the Kansas DHE analyzed 1,380 fuel samples for
the content of MTBE. Only 5.3% of samples had MTBE concentrations in a range from 11% to 15.4% as would be
expected for oxygenated reformulated gasoline. Samples of gasoline that contained MTBE for purposes of meeting
octane requirements were more common; 21.3% of samples had concentrations of MTBE between 11% and 6%, and
37.6% of the gasoline samples had concentrations of MTBE between 6% and 1.5%. Gasoline that had MTBE from
incidental activities were also common; 35.8% of the samples of gasoline had less than 1.5% MTBE (Hattan et al.,
2003). The average concentration of MTBE over all the samples of gasoline from Kansas was near 4.5%, between a
half and a third of the concentration expected in oxygenated reformulated gasoline. The concentration of MTBE in
ground water at gasoline spill sites in Kansas was between one-half and one-third of the concentrations in Texas and
southern California where oxygenated reformulated gasoline was required (Figure 2.1).
2.2 Definitions and Expressions Used for Rate Constants
At field scale, the concentration of MTBE in a well can change through the combined influence of dilution and dispersion,
biodegradation, sorption, and mixing of the contaminant plume with cleaner water in a monitoring well. It is usually
impossible to separate the individual contributions of each process. It is important to acknowledge this uncertainty. In
this document, rate constants calculated from changes in concentrations in wells will be identified as rates of attenua-
tion. In laboratory studies conducted with batch microcosms, there is no opportunity for dilution and dispersion, and
the effects of sorption are evaluated in sterilized controls. In this document, rate constants that are calculated from
controlled laboratory studies will be termed rates of biodegradation.
Dispersion and dilution of MTBE with the flow of water is independent of the concentration of MTBE. As a conse-
quence, their effect on concentrations of MTBE will be proportional to the initial concentration of MTBE. A constant
quantity of MTBE does not partition between water and residual gasoline. Rather, a constant proportion of the MTBE
will partition, regardless of the absolute concentration of MTBE.
At the concentrations of MTBE most commonly seen in ground water, the rate of biodegradation is not a fixed number,
but is proportional to the concentration of MTBE present in the ground water. However, an exception to this general
rule will be discussed in Section 4 Biological Degradation.
Chemists and engineers describe a process where the rate of the process is directly proportional to the amount of material
subject to the process as being a first order process. A first order process for biodegradation is quantitatively defined
by Equation 2.2, where F is the fraction of the original material remaining at some time t, and k is the first order rate of
removal through biodegradation or attenuation. F is conventionally calculated as the concentration remaining C divided
by the original concentration Co.
-F ~~ ^ ^O — 6 Equation 2.2
Equation 2.2 describes changes in concentration over time in a particular monitoring well. Equation 2.3 uses the same
relationship to describe changes in concentration with distance along a flow path in the aquifer, where d is the distance
along the flow path between the up gradient well producing water with the contaminant at concentration Co and the
down gradient well producing water with the contaminant at concentration C.
Equation 2.3
The most familiar example of a first order process is the decay of radioactive elements. A familiar unit for the rate of
radioactive decay is the half life of the element, the time required for half the material originally present to decay. From
Equation 2.2, this would be the time required, at a particular first order rate of removal, for C/Co to equal Vi A first
order rate constant is not intuitively obvious to most people. Most people find a half life easier to understand. How-
ever, a first order rate constant is a direct expression of the rate and is the best way to compare rates with each other.
As the rate goes up or down, the value of the constant goes up or down proportionately. First order rate constants are
also more convenient to use in equations. As a consequence, the first order rate constant will be used throughout this
manuscript. If a reader prefers to think in terms of half lives, Figure 2.2 provides a convenient means to translate a first
-------
order rate constant into a half life. A half life is easily calculated from a first order rate constant using the relationship
in Equation 2.4, where t is the time required to reach half the initial concentration, and k is the first order rate constant
for attenuation. The unit of the half life is the reciprocal of the unit for the first order rate constant.
t1/2 = 0.693/k
Equation 2.4
TO
1000000
100000
10000.
1000-
100-
10-
1 -
0.1-
nm.
— «-
^
^^
*
>
>N
::=*=^
>
^
^
*
>
S^
*
^^
"" fc
fcfc
^
*
-
X
.
Ns
^
-^
I
^^
*
s
41"
•
s^
If Life
-: —
«
=1^
in
*
fc^
^^
=Y
D
*
ft
•
S
2tsi =
iw
jh
0.001 0.01 0.1 1 10
First Order Rate Constant (per year)
100
Figure 2.2 Relationship between a first order rate constant and a half life.
2.3 Role of Attenuation in the Lifecycle of MTBE Plumes
The size of a plume reflects a balance between the rate of release of the contaminant from the source area into the
aquifer, the rate of transport of the contaminant away from the source area, and the rates of degradation and dispersion
in ground water which remove mass or reduce plume concentrations. Depending on the relationship between the rate
of dissolution of the contaminant from the fuel spill and the rate of attenuation of the contaminant in ground water,
plumes may grow, or plumes may shrink and eventually disappear. When source dissolution and advection dominate
over dispersion and biodegradation, plumes expand. When biodegradation and dispersion dominate, plumes contract
or attenuate back toward their source. When the plume begins to attenuate, the actual distribution of MTBE in ground
water will depend on the relationship between two rates of attenuation: the rate of attenuation of the source and the
rate of attenuation along the flow path in the plume. Methods to calculate these rates from field data are presented in
Newell et al., (2002).
The long-term source of MTBE at gasoline spill sites is MTBE retained in residual gasoline trapped in the aquifer.
The longevity of the plume will depend on the amount of MTBE in the residual gasoline, on the rate that MTBE is
transferred from the residual gasoline to ground water, and the rate that the flow of ground water carries the MTBE
away from the residual gasoline. Transfer of MTBE will weather the residual gasoline and, over time, will reduce the
concentrations in the ground water that are in contact with the residual gasoline. The rate of weathering is the rate of
natural attenuation of the source which determines how long a plume will persist over time. Biodegradation, dilution,
dispersion and mixing will attenuate concentrations of MTBE as water moves away from the source. They determine
the rate of attenuation along the flow path which determines how far a plume will extend away from the source.
When a plume has come to a steady state, the rate of attenuation over time in all the wells should be zero. As the
source starts to weather away and the concentrations of contaminant at the source start to attenuate, there will be a
-------
corresponding attenuation in all the monitoring wells down gradient of the source. It is important to not confuse at-
tenuation of concentrations over time in monitoring wells down gradient of the source with attenuation along the flow
path in the aquifer.
Table 2.1 presents data from six plumes that contrast the roles of these two distinct rates of natural attenuation. The
data are from plumes described by Landmeyer et al., (2001) and Wilson and Kolhatkar (2002). All of the plumes in
Table 2.1 are old releases that had reached a steady state at the time of the study. The method to calculate the rates
of attenuation is described in detail in Newell et al., (2002). The rate of attenuation of the source is calculated from
Equation 2.2 as the first order rate of attenuation over time of concentrations of MTBE in the most contaminated well.
The rate of attenuation of the plume was calculated from Equation 2.3 as the first order rate of attenuation of concen-
trations of MTBE with distance along the flow path in the plume, multiplied by an estimate of the plume's seepage
velocity. The method to estimate the confidence intervals is illustrated in Section 7 Statistical Evaluation of Rates of
Attenuation of Sources.
Table 2.1 Rates of Attenuation of MTBE in Source Areas Over Time Contrasted to Rates of Attenuation Along Flow
Paths in Ground Water
Location
Brandon, FL
Elizabeth City, NC
Long Island, NY
Parsippany, NJ
Port Hueneme, CA
Laurel Bay, SC
Attenuation of Plume
Rate
Slower 90%
Confidence Interval
Attenuation of Source
Rate
Slower 90%
Confidence Interval
per year
2.02
1.80
0.79
1.17
0.56
<0.04
0.71
1.20
0.53
0.61
0.47
not significant at 90%
confidence
0.27
0.15
0.75
0.19
0.23
0.70
0.14
0.04
0.29
0.15
0.09
0.60
In the plumes at Brandon, Florida; Elizabeth City, North Carolina; Parsippany, New Jersey; and Port Hueneme, Cali-
fornia, the rate of attenuation along the flow path in the plume is faster than the rate of attenuation over time in the
source area. As they age, these plumes tend to recede back on themselves. The tendency to recede back to the source
is illustrated in Figure 2.3 with data from the plume at Parsippany, New Jersey.
In the plume on Long Island, New York, the source area was remediated. After remediation, the residual "hot spot of
MTBE" moved down gradient over time. In the plume at Laurel Bay, South Carolina, the rate of attenuation over time
in the source area was faster than attenuation in the ground water. Over time, the "hot spot of MTBE" in the plume
detached from the source area and moved down gradient. The tendency for the "hot spot" to detach and move down
gradient is illustrated in Figure 2.4 with data from the plume at Laurel Bay, South Carolina.
There was no significant natural biodegradation of MTBE in the anaerobic portion of the plume at Laurel Bay, South
Carolina. Because there was little variation in the direction of ground water flow, there was little contribution of disper-
sion to attenuation. The plume continued to grow until it approached its point of discharge to surface water drainage
in a concrete-lined ditch.
As the plume approached the ditch, it was oxygenated as it mixed with ground water in the bed sediments beneath the
concrete liner (Landmeyer et al., 2001). On a sampling date in the winter, all of the MTBE in the plume was biode-
graded before the plume discharged. On another date in the summer, more than 96% of the MTBE in the plume was
biologically degraded before the plume discharged to surface water. The bed sediments of many surface water streams
have a considerable capacity for aerobic biodegradation of MTBE (Bradley et al., 200Ic).
In general, when the rate of attenuation of the source over time is faster than the rate of attenuation along the flow
path, the "hot spot" will detach from the source and move down gradient. This is true whether the attenuation of the
source is purely natural, or is a result of the efforts to control or remediate the source area. In general, when the rate
of attenuation of the source over time is slower than the rate of attenuation along the flow path, the plume will appear
to recede back to the source area over time.
-------
1993/1994
>20
1998
30 meters
>20 ng/L
>100 MS/L
2003
>20|
>ioo n-9/L
Figure 2.3 Evolution of a plume ofMTBE at Parsippany, New Jersey, that is receding back on itself.
-------
^
I
Former Underground Storage Tank Area
\
30 meters
>20u,g/L
> 10,OOOn,g/L
January 1998
^^^H \ .n \
April 1993
\
Drainage Ditch
Figure 2.4 Evolution of a plume ofMTBE in Laurel Bay, South Carolina, where the "hot spot" has detached from the
source and moved down gradient.
-------
Preceeding Page Blank
Section 3 -
Recommendations for Monitoring -
This section discusses a number of interrelated issues concerned with monitoring the behavior of MTBE plumes. Data
that are collected to understand the behavior of MTBE at gasoline spill sites are generally collected for three purposes.
The primary purpose is to document the distribution of MTBE contamination at the site in time and space. A second
purpose is to determine if the coverage of the monitoring wells is adequate to properly determine the distribution of
contamination. A third purpose is to predict the physical, chemical, and biological processes that control the distribu-
tion of the fuel oxygenates
This section discusses biogeochemical parameters that can be used to recognize the footprint of a plume of contamination
from a gasoline spill, even when the concentrations of MTBE have declined below detection limits. This information is
valuable to evaluate the coverage of monitoring wells. If biogeochemical data are collected in each round of sampling,
they can also be used to determine if a plume has shifted its location.
The distribution of MTBE in ground water is controlled by the direction of ground water flow and by changes in the
direction of ground water flow over time. The direction of ground water flow is usually inferred from the elevation of
the water table in monitoring wells. Information on the depth to water is usually collected with each round of sampling,
but reports to regulators often present ground water contours for only one "typical" round of sampling, or a few rounds
of sampling. The flow direction in the "typical" round of sampling becomes ingrained in the site conceptual model;
and the variation in the direction of flow is ignored. The variation in flow direction should be considered to determine
whether a particular well should sample the plume or sample ambient ground water in each round of sampling. This
section illustrates the wide variation in flow direction at a representative field site and introduces two simple computer
applications that can be used to extract the direction of ground water flow from data on water elevations in monitoring
wells.
The prospects for natural attenuation of chlorinated solvents, or the BTEX compounds, have been correlated with the
geochemistry of the ground water. Preliminary studies suggested that anaerobic biodegradation of MTBE is favored
under methanogenic conditions and sulfate reducing conditions (Kolhatkar et al., 2000). This section examines the
potential correlation in more detail and finds that there is no correlation between the extent of anaerobic biodegradation
of MTBE in ground water and the concentration of methane or sulfate in water.
Traditionally, microcosm studies have been used to demonstrate that microorganisms at a site can degrade a contami-
nant. Microcosm studies of MTBE biodegradation are expensive, time consuming, and often yield equivocal results.
As a consequence, they are rarely done as part of the risk evaluation at gasoline spill sites. Either the possibility of
natural MTBE biodegradation is ignored altogether, or rate constants published in the literature are extrapolated to a
site without any site specific evidence that they are appropriate.
An analysis of the change in the stable carbon isotope ratios in MTBE can provide unequivocal evidence for biodegra-
dation of MTBE at field scale. Unfortunately, at this writing, these analyses are not currently offered by commercial
analytical laboratories. They are only commercially available from a few university laboratories. At many sites, the
onset of anaerobic biodegradation of MTBE can be recognized by a change in the ratio of TEA to MTBE in the moni-
toring record. This section illustrates the use of data on the relative concentration of MTBE and TEA as a practical
alternative to microcosm studies or stable carbon isotope analyses.
Conventional monitoring wells can provide an incomplete picture of the true distribution of MTBE in ground water.
If the screen of a monitoring well is long compared to the thickness of the plume of contamination, it can sample the
plume of contaminated ground water and cleaner ground water above or below the plume, giving a false impression of
natural attenuation from one well to another. Long plumes of MTBE may dive below the screens of monitoring wells
altogether. Any evaluation of natural attenuation between monitoring wells should consider the screened intervals of
the wells, the depth interval contaminated with gasoline (if that information is available), and the lithological features
sampled by the wells. This report does not further discuss the vertical spacing of monitoring wells in the assessment
11
-------
of natural attenuation and the evaluation of risk. These considerations are discussed in detail in Performance Monitor-
ing ofMNA Remedies for VOCs in Ground Water (Pope et al., 2004), which is available on the Kerr Center web page.
Search for "Kerr" on the U.S. EPA web page.
3.1 Biogeochemical Footprints to Evaluate Coverage of Monitoring Wells
If natural attenuation is reducing the concentrations of contaminants in ground water, the concentrations should be
higher in the source area and lower in the down gradient wells. The concentration in monitoring wells down gradient
of a source area can be lower because the contaminants were attenuated, or concentrations may be lower because the
monitoring well missed the plume (Wilson, 2003a). To distinguish between the two, the National Research Council
recommended the use of biogeochemical footprints to distinguish attenuation of contaminants in ground water that has
been impacted by a fuel spill from clean water that was never impacted. (NRC, 2000; Rittmann, 2003).
Many organic materials can be metabolized in ground water through an oxidation / reduction reaction where the organic
material is oxidized to carbon dioxide while an electron acceptor is reduced. Oxygen, nitrate, and sulfate are often
referred to as soluble electron acceptors. During metabolism they are reduced to water, nitrite or molecular nitrogen,
and sulfide. Metabolism of the organic matter can be recognized by the depletion of oxygen, nitrate, and sulfate and
the accumulation of nitrite or sulfide in water. Iron (III) minerals in sediments can also serve as an electron acceptor.
In the process, iron (III) is reduced to iron (II). Although iron (III) minerals are not very soluble in ground water, the
iron (II) that is produced from microbial metabolism is more soluble and can accumulate in ground water.
Not all metabolism requires an external electron acceptor. Some organic materials, including the BTEX compounds,
can be fermented. One end product of the fermentation of BTEX compounds is methane.
In summary, the metabolism of the BTEX compounds in gasoline can consume oxygen, nitrate, and sulfate or produce
methane and iron (II) (Wiedemeier et al., 1995). If changes in these biogeochemical parameters can be associated
with a particular spill of gasoline, as distinct from organic materials already present in the aquifer, they can be used as
footprints for the plume of contamination produced from the spill.
If the biogeochemical parameters are consistent with the ambient conditions in the aquifer and contaminants are absent,
this situation indicates that the well has not been impacted by the gasoline spill. The well is outside the footprint of
the plume. If biogeochemical parameters show the depletion of oxygen, nitrate and sulfate, and the accumulation of
iron (II) and methane, and the contaminants are absent, this situation indicates that the water was contaminated at one
time, but natural attenuation processes have removed the contaminants. The well is inside the footprint of the plume.
To be useful as footprints of the plume, these biogeochemical indicator parameters should be measured in each round
of sampling where possible.
Biogeochemical footprints are footprints of gasoline contamination as a whole, and not necessarily of MTBE contamina-
tion alone. The biogeochemical indicators cannot distinguish a plume of gasoline with MTBE from a plume of gasoline
without MTBE. The biogeochemical indicators work best if the MTBE entered ground water from direct contact of
gasoline with ground water. If the MTBE entered ground water through a vapor pathway, the readily degradable hy-
drocarbon components of gasoline may have been removed before the MTBE entered the ground water. They may fail
to provide any indication of contamination of ground water by MTBE vapors.
Some of the biogeochemical parameters are more useful than others. Bacterial communities acclimate readily to de-
grade BTEX compounds using oxygen and nitrate as electron acceptors. Depletion of oxygen and nitrate should be
expected at almost every gasoline spill. Bacterial communities also acclimate readily to degrade sulfate. Depletion of
sulfate should be expected at most sites as well. Bacterial communities require from months to years to acclimate to
ferment BTEX compounds to methane.
The depletion of dissolved oxygen is the most sensitive indication of contamination with gasoline. It is also the most
problematic. It is difficult to prevent reoxygenation of ground water samples. It is almost impossible to prevent re-
oxygenation of the ground water sample if the well is purged and sampled with bailers. Oxygen meters may provide
reliable data if they are properly maintained and are recalibrated in the field; but if the appropriate quality assurance
procedures are not implemented, they can produce data that are misleading. Simple field test kits can also provide us-
able data on the concentration of dissolved oxygen in ground water (Wilkin et al., 2001).
Nitrate is the next most sensitive indicator of gasoline contamination. Sampling of nitrate is not problematic, as is the
case with oxygen. Analysis of water samples for nitrate is straightforward and inexpensive. Unfortunately, nitrate is
often absent under natural conditions in many ground waters. If nitrate is present under ambient conditions, it should
be considered as an alternative to dissolved oxygen if the monitoring wells are sampled with bailers.
12
-------
Sulfate is the most important soluble electron acceptor in ground water at most fuel spill sites (Wiedemeier et al., 1995).
Sulfate samples do not require preservation, and analysis for sulfate is straightforward and inexpensive. At most fuel
spills, the depletion of sulfate is the best single indicator of impact from a gasoline spill. If it is only possible to moni-
tor one biogeochemical parameter, sulfate should be the parameter of choice.
The accumulation of methane as a tracer works best in old spills. Often the microbial communities do not acclimate
to produce methane. Perhaps one third of sites will fail to accumulate methane to concentrations above 0.5 mg/L
(Kolhatkar et al., 2000). Methane samples must be collected in a manner that avoids losses due to volatilization, and
the samples must be preserved.
The BTEX compounds in gasoline will support iron reduction in almost every site. However, iron (II) often fails to
accumulate in the ground water. In many contaminated aquifers, iron reduction and sulfate reduction occur at the
same time. The iron (II) will react with any sulfide produced by sulfate reduction and precipitate. If the rate of sulfide
production exceeds the rate of iron (II) production, iron (II) may never accumulate in the ground water. Iron is useful
as a tracer at less than one-fourth of sites (personal experience of authors). Iron (II) is best determined immediately
after water samples are collected.
3.2 Monitoring the Direction of Ground Water Flow
The second approach used to identify the possible footprint of a plume is to measure the hydrological properties of the
aquifer receiving the fuel spill, characterize the distribution of the gasoline spill, and then calibrate a computer model
such as BIOSCREEN (Newell et al., 1996). For model results to be applicable, the model assumptions must be valid
for the site.
Most computer models assume that ground water flows in a uniform direction with a uniform velocity. Any spreading
of the plume is attributed to dispersion, and when the models are calibrated to field data, the values for longitudinal and
horizontal dispersion are adjusted to match the field data. In some aquifers, the direction and speed of ground water
flow are stable, and in these aquifers plumes are usually long and narrow. Often the width of the plume down gradi-
ent of the source is no wider than the width of the source area, indicating that transverse and vertical dispersion make
a minimal contribution to the distribution of MTBE. If there is little spreading of MTBE to the sides of the plume, a
conventional ground water model may provide an accurate forecast of the distribution of contamination.
Other plumes appear to spread laterally as well as longitudinally. This apparent lateral dispersion may be the direct result
of changes in the direction of ground water flow. When there is significant variation in the direction and magnitude of
ground water flow, conventional models can be misleading about the expected footprint of the plume. What appears to
be lateral dispersion is really longitudinal dispersion occurring in different directions. In plumes where there is a wide
variation in the direction of ground water flow, simplistic ground water models that assume the ground water moves
only one direction at one velocity can be misleading.
Figure 3.1 presents data on the direction and magnitude of ground water flow at an MTBE site at Elizabeth City, North
Carolina (Wilson, 2003a). The site is near the Pasquotank River, and the average direction of ground water flow is
toward the river; however, the flow at any particular time is sensitive to the stage of the river. The plume was moni-
tored monthly for one year. Figure 3.1 presents predictions to the direction and velocity of ground water flow from
the monitoring data in each month.
Regression analysis was used to fit a plane through the elevation of the water table in the monitoring wells (Wilson et
al., 2000, Srinivasan, 2004). An arrow is used in the figure to represent the direction and velocity of ground water. The
arrows are given different shades to allow them to be distinguished in the figure.
It is apparent that the direction and magnitude of flow vary widely from one month to the next at this site. The standard
deviation of the direction of ground water flow over 12 months of sampling, as depicted in Figure 3.1, was 23 degrees.
One round of sampling, or even a few rounds of sampling at this site, would not be adequate to define the direction and
magnitude of ground water flow. At this site, the contaminant plume occupies the area encompassed by the variation
in the direction of ground water flow. This is probably true for all sites where the plume has come to a steady state.
Mace et al., (1997) used a similar approach to calculate the standard deviation of the direction of ground water flow at
132 gasoline stations in Texas (Table 3.1). At each site, the direction of flow was estimated from water table elevations
on at least ten separate occasions. The direction of ground water flow at most of the sites in Texas (Figure 3.2) was
more variable than the site in North Carolina illustrated in Figure 3.1. The median of the standard deviation of the
direction of ground water flow in Texas was 36 degrees.
13
-------
50
100 150
Approximate Scale in Meters
Sampling Point
Figure 3.1. Variation in the direction and magnitude of ground water flow at an MTBE site in Elizabeth City, North
Carolina. The arrows represent the distance that water would move in one year, based on the direction
and hydraulic gradient present in a particular round of sampling. Some of the arrows are shaded grey to
allow them to be distinguished from the other arrows. The origin of the arrows is the center of the LNAPL
source area. The black dots are locations of monitoring wells. The shaded area includes all the monitor-
ing wells with concentrations of MTBE above 20 ug/l.
To facilitate the visualization of the variability of the flow direction of the sites in Texas, Figure 3.2 summarizes the
data in Table 3.1 by comparing the arcs subtended by one standard deviation in flow direction on either side of the mean
direction of flow. A comparison is made between the sites with low variation (± 20 degrees), moderate variation (± 30
degrees), high variation (± 50 degrees), very high variation (± 70 degrees), and extremely high variation (± 120 degrees).
The variability in the direction of ground water flow in Texas is probably typical of many regions of the United States.
For roughly one-third of the sites in Texas, the direction in ground water flow is highly variable, and the concept of a
single flow direction is not the best representation of the behavior of the plume. A conventional ground water model
could be misleading at these sites.
Site investigation reports may include maps showing the contour of the water table, particularly if there are significant
variations in the direction of flow from one sampling event to the next. However, the variation in the direction of ground
water flow is rarely evaluated in any formal way. As a consequence, the monitoring wells on the perimeter of a site
may not be in the best position to detect a plume of contamination.
Data on water table elevations are frequently collected at gasoline spill sites. The U.S. EPA has created a decision
support tool that evaluates data on water table elevations in monitoring wells to make predictions of the most likely
footprint of a plume of contamination. The Optimal Well Locator (OWL) is a screening tool to evaluate the locations
of existing monitoring wells, to identify the best locations for new wells, and to identify the existing wells that are least
likely to detect the plume (Srinivasan, 2004). OWL was used to generate the predictions on the direction and magnitude
of ground water flow that are presented in Figure 3.1. A simple calculator for the direction of ground water flow is also
available on the Athens Laboratory web page, (http://www.epa.gov/athens/learn2model/part-two/onsite/index.html)
14
-------
Table 3.1 - Variation in the Standard Deviation of the Direction of Ground Water Flow at 132 Gasoline Stations in
Texas (Data from Figure 14 in Mace et al., 1997)
Standard Deviation
(degrees)
Oto 10
10 to 20
20 to 30
30 to 40
40 to 50
50 to 60
60 to 70
70 to 80
80 to 90
90 to 100
100 to 110
110 to 120
Occurrence
6
22
22
23
10
14
7
9
13
3
1
1
Frequency %
4.6
16.8
16.8
17.6
7.6
10.7
5.3
6.9
9.9
2.3
0.8
0.8
Cumulative Frequency %
4.6
21.4
38.2
55.7
63.4
74.0
79.4
86.3
96.2
98.5
99.2
100.0
Figure 3.2. - Variation in the direction of ground water flow at gasoline spill sites in Texas. The solid arrow represents
the direction of ground water flow. Presented are the arcs subtended by one standard deviation of the di-
rection of ground water flow. Numbers on the left side of the figures are the percentage of 132 sites where
one standard deviation is contained within the arc to the left of the number. Numbers on the right hand
side of the figure are the percentage of 132 sites where the arc subtended is greater than the arc to the left
of the number but less than the arc to the right of the number in the figure.
15
-------
3.3 Monitoring to Predict Biological Processes
If the biogeochemical parameters are used to determine the chemical environment of the organisms that degrade MTBE
in the aquifer, it is not necessary to determine oxygen, nitrate, sulfate, methane, and iron (II) at every round of sampling.
Oxygen should be monitored as a routine parameter only when the monitoring wells are sampled using pumps. If
the wells are sampled with bailers, it is best not to attempt to measure dissolved oxygen in the ground water. If the
concentration of oxygen in the well water is less than 0.5 mg/L, there is not an adequate supply of oxygen to support
extensive metabolism of MTBE.
At gasoline spill sites, the vertical gradients of biogeochemical parameters may be strong. It is not unusual for a moni-
toring well that is screened across the water table to produce oxygenated uncontaminated water from the depth interval
right at the water table and water that is contaminated and strongly anaerobic from an interval less than a meter deeper.
The best indication of anaerobic conditions in the contaminated portion of the aquifer is the presence of iron (II) or
methane in water from the monitoring well. The blended water produced by the well may contain oxygen even when
oxygen is totally absent in the contaminated portion of the aquifer.
Analysis of methane in samples can be expensive and is not recommended for routine monitoring. The measured
concentrations of iron (II) in ground water have no straightforward relationship to the amount of biological activity
in the water. The presence of iron (II) in water at detectable concentrations indicates that biological iron reduction is
occurring in the aquifer. The actual concentration of iron (II) has no direct interpretation. Routine monitoring for iron
(II) is not recommended.
Early work on natural anaerobic biodegradation of MTBE indicated an association of MTBE biodegradation with
methanogenic conditions in ground water (Wilson et al., 2000; Kolhatkar et al., 2000; Kolhatkar et al., 2001; Kolhatkar
et al., 2002). More recent work at thirteen gasoline spill sites in Orange County, California, used the stable carbon
isotope ratio of MTBE in ground water to estimate the extent of natural biodegradation (Kuder et al., 2005; Wilson et
al., 2005b, discussed in detail in Sections 5 and 6).
Figure 3.3 compares the extent of biodegradation of MTBE in ground water at the thirteen sites in Orange County,
California, to the concentration of methane or sulfate in the water. There was no clear association of natural MTBE
biodegradation with high concentrations of methane or with low concentrations of sulfate in the ground water (Figure 3.3).
Although these biogeochemical parameters may be useful to recognize the footprint of the plume from a spill of gasoline,
these parameters have little value to predict anaerobic biodegradation of MTBE.
The first biodegradation product of MTBE is TEA. At many sites, the TEA produced from MTBE biodegradation accu-
mulates in the ground water. If the long-term monitoring record includes analysis of TEA, many sites show a transition
in the relative proportions of TEA and MTBE in the ground water (see Figure 3.4 for an example). The best indication
for MTBE biodegradation that is available from conventional monitoring parameters is an abrupt and persistent increase
in the ratio of TEA to MTBE in ground water.
Depending on its concentration, the TEA produced by natural biodegradation of MTBE in the ground water may also
be a concern.
3.4 Concerns with Analytical Issues
Good environmental monitoring requires good chemistry. The conventional analytical approach for analysis of gasoline
components in ground water is preparation with a purge-and-trap unit, separation on a gas chromatograph, and deter-
mination with a flame ionization detector (EPA method 8015). Crumbling and Lesnik (2000) noted that alcohols, such
as TEA, are not efficiently recovered from water by purge-and-trap, and as a result, the analytical detection limits for
TEA and other fuel alcohols are high. Often the reporting limit is above the Provisional Action Goal set by California
for TEA of 12 |ig/L. To improve the recovery efficiency of TEA and other fuel alcohols, Lin et al., (2003) prepared
water samples using a heated headspace sampler. With a heated headspace sampler and determination with a mass
spectrometer (EPA method 8260), they could achieve a method detection limit of 0.8 |ig/L for TEA.
MTBE is manufactured from a mixture of isobutylene and methanol, with the aid of an acid (H+) catalyst. Acid can
also catalyze the hydrolysis of MTBE to produce TEA and methanol. If ground water is preserved with acid to pH<2,
MTBE can be hydrolyzed to TEA (Lin et al., 2003; McLoughlin et al., 2004; O'Reilly et al., 2001; White et al., 2002).
In samples that are refrigerated, the rate of hydrolysis is slow. When samples are stored at 10° C, less than 5% of the
MTBE is hydrolyzed within 30 days. There is little hydrolysis during sample preparation by purge-and-trap at room
16
-------
temperature. However, when the water sample is heated to improve the recovery of TEA, a major fraction of the MTBE
is hydrolyzed during analysis. Hydrolysis of MTBE to TEA can be avoided by preserving the sample with 0.1% triso-
dium phosphate instead of acid (Lin et al., 2003; McLoughlin et al., 2004; White et al., 2002).
As will be discussed in detail in Sections 5 and 6, the most unequivocal indicator of natural MTBE biodegradation is
the ratio of stable carbon isotopes in the residual MTBE. Samples collected for analysis of stable carbon isotope ratios
should be preserved with trisodium phosphate whenever possible. The effective holding time for samples preserved
with trisodium phosphate is more than three months. If the analytical laboratory is backed up and the samples cannot
be analyzed within a few weeks, the samples will not be compromised by the longer holding times.
0.0001
°-001
O)
c
c c
'co .9
I | 0.01
QL 2
o
LL
s
0.1 -
0.001
0.01 0.1 1
Methane (mg/L)
10
0.0001
I
O)
c
'c c
ro o
0.001
°i Ł 0.01
iu S?
CQ «>
o
1
CQ
0.1
******* * i *
* *••**• « >,
10 100
Sulfate (mg/L)
1000
10000
Figure 3.3. Association of sulfate and methane with natural anaerobic biodegradation of MTBE in ground water col-
lected from 61 monitoring wells at thirteen gasoline spill sites in Orange County, California.
17
-------
10 n
8-
6-
1
§
2 -
Jan-01
Jan-02
Jan-03
Jan-04
Date of Sample
Figure 3.4. A change in the relative concentration of TEA and MTBE in a monitoring well as evidence for natural
biodegradation of MTBE. Data are from a gasoline spill site in Tustin, California.
18
-------
Section 4 -
Biological Degradation -
Many ground water scientists and engineers think that MTBE does not biodegrade in ground water. The idea may come,
in part, from the fact that it is difficult to culture MTBE degrading bacteria. The idea may also come, in part, from
the fact that many MTBE plumes are much longer than the associated plumes of benzene and BTEX compounds that
come from the same spill of gasoline. Finally, the idea that MTBE does not degrade is consistent with the behavior of
MTBE in a controlled release experiment that was done by the University of Waterloo at Canadian Forces Base Borden
in Ontario, Canada. In an aquifer that had not experienced MTBE contamination, MTBE did not degrade in the first
years of the experiment. In subsequent years, MTBE did biodegrade in the field-scale plume, but by then, the idea that
MTBE does not biodegrade was fixed in the minds of many ground water professionals.
This section briefly reviews the current state of knowledge concerning microbiology of MTBE biodegradation. The
information provided in this section draws from information provided in several recent reviews of the biodegradation of
MTBE (Deeb et al., 2000a; Cozzarelli and Baehr, 2003; Fiorenza and Rifai, 2003; Finneran and Lovley, 2003; Rittmann,
2003; Wilson, 2003a; Wilson, 2003b, and Schmidt, 2004). The review by Schmidt et al., (2004) is particularly detailed
and comprehensive.
There are three general approaches to document biodegradation; the loss of the parent substrate, the accumulation
of an intermediate biodegradation product, and mineralization of the organic carbon originally present in the parent
substrate. Mineralization studies are conventionally done by labeling the parent compound with 14C, and measuring
the accumulation of radio-label in carbon dioxide or methane. All three approaches have been applied to understand
biodegradation of MTBE.
4.1 Microbiology of Aerobic MTBE Biodegradation
Organisms that grow aerobically on MTBE are difficult to isolate and culture in the laboratory. In laboratory micro-
cosms, Yeh and Novak (1994) found no evidence for aerobic biodegradation of MTBE after 100 days of incubation.
After 60 days of incubation, Jensen and Arvin (1990) found no degradation of MTBE in samples of activated sludge,
topsoil, or aquifer material. In 1994, Salanitro et al. published the first report of the aerobic biodegradation of MTBE
by a mixed culture of microorganisms; in 1997, Mo et al. published the first report of aerobic biodegradation of MTBE
by pure cultures of bacteria; and in 1997, Borden et al. published the first report of aerobic MTBE biodegradation in
microcosms. In the study of Borden et al. (1997), the aerobic biodegradation of MTBE in the microcosms was incom-
plete; the biodegradation of MTBE stopped after 100 days of incubation, leaving approximately 1.0 mg/L of MTBE
in the pore water.
There are several explanations of why it is difficult to isolate MTBE degrading micro-organisms. All of the carbon-
to-carbon bonds in MTBE involve bonds with the central tertiary carbon. It is difficult for microorganisms to break
a carbon bond with a tertiary carbon atom (one more carbon-to-carbon bond would result in the bonding structure of
diamond). It is also difficult for microorganisms to degrade ethers, and MTBE is an ether. If the ether bond is broken
by enzymatic hydrolysis, the products are TEA and formaldehyde, and formaldehyde is toxic. Finally, it may be dif-
ficult to isolate pure cultures of MTBE degrading organisms because MTBE is degraded in nature by a consortium of
different types of microorganisms acting together, and not by single organisms.
Because it was difficult to isolate and culture microorganisms that aerobically degraded MTBE, there was a limited
appreciation of the capacity of aerobic bacteria to degrade MTBE in ground water. In their comprehensive review of
the behavior of MTBE in the environment, Squillace et al. (1997) concluded, "In general, most studies to date have
indicated that MTBE is difficult to biodegrade, and some have classified MTBE as recalcitrant."
This initial difficulty in isolating and culturing MTBE degrading bacteria was caused by the impatience of the scientists
as much as the metabolic capability of the microbes. Traditionally, microbiologists working in the laboratory base their
expectations, and design their experimental protocols, on the behavior of organisms that grow rapidly. They rarely
incubated their enrichment studies for more than a few months.
19
-------
The difficulty in isolating microorganisms that can degrade MTBE is easily understood if we compare the growth rate
of microorganisms that grow on MTBE to the growth rate of microorganisms that degrade ordinary petroleum hydro-
carbons such as benzene, toluene, and xylenes (Table 4.1). Typical strains of bacteria growing aerobically on petroleum
hydrocarbons can divide and double their numbers every two to five hours at room temperature. As a consequence,
laboratory enrichment cultures will grow up and remove the hydrocarbons in a few days. On the other hand, cultures
of bacteria using MTBE as a growth substrate require several days to several weeks to double their numbers. Their
growth rate is from one-tenth to one-hundredth of the growth rate of bacteria that degrade conventional petroleum hy-
drocarbons. Their very slow growth rate has an important effect on the time required for a culture to grow to densities
that will entirely consume MTBE.
Table 4.1. Comparison of the Growth Rate of Aerobic Bacteria During Growth on MTBE as the Primary Substrate
to the Growth Rate of Bacteria that Grow on Pentane and Fortuitously Metabolizes MTBE and to the
Growth Rate of Bacteria that Grow on BTX Compounds
Growth
Substrate
Source of Data
Doubling Time (hours)
Reference
Organisms that can degrade MTBE as the primary substrate
MTBE
MTBE
MTBE
MTBE
BC-1 culture
Enrichment from Refinery
Activated Sludge
Enrichment from Biofilter
ENV735 Hydrogenophaga
flava
>340
58
670
41
Salanitro et al. (1998)
Park and Cowan (1997)
Fortin and Deshusses (1999)
Steffan et al. (2000)
Organisms that co-metabolize MTBE
Pentane
Pseudomonas aeruginosa
3.6
Garnier et al. (1999)
Organisms that cannot degrade MTBE
Benzene
Toluene
Xylenes
Median of 10 studies
Median of 15 studies
Median of 8 studies
2.1
2.9
5.0
Suarez and Rifai (1999)
Suarez and Rifai (1999)
Suarez and Rifai (1999)
The BC-1 culture acquired by Salanitro et al., (1998) requires at least 340 hours to double its density (see Table 4.1).
The complete metabolism of an initial concentration of MTBE of 1.0 mg/L will produce a final density of bacteria of
approximately 106 per milliliter. If the initial density of the BC-1 culture were one active cell per milliliter, it would
require 20 cycles of cell division over 283 days for the culture to grow up and degrade MTBE. Strain ENV735 is an-
other bacterium that can degrade MTBE as the sole carbon source. Its doubling time is near 41 hours; it would require
34 days for the culture to grow up and degrade MTBE.
Fortin and Deshusses (1999) note that most of the cultures of MTBE degrading bacteria that were available to them at
that time were acquired from bioreactors or biofilters that had already acclimated to degrade MTBE. It is likely that
early attempts to enrich for the organisms that degrade MTBE failed because the enrichment cultures were not incubated
for an adequate period of time.
Some organisms can biodegrade MTBE, but they cannot grow on MTBE alone; these organisms require another substrate
for growth. Biodegradation of MTBE under these circumstances is termed co-metabolism or co-oxidation. The organisms
that grow on other substrates and co-metabolize MTBE can grow rapidly. Compare the growth rate of Pseudomonas
aeruginosa when growing on pentane to the growth rate of microorganisms growing on MTBE (Table 4.1). With a
doubling time of 3.6 hours, this organism could grow from an initial density of one cell per ml up to densities that can
degrade 1.0 mg/L of MTBE in only three days.
20
-------
Many of the natural hydrocarbons in gasoline can support the growth of organisms that will degrade MTBE. This is
particularly true of the straight-chained alkanes and iso-alkanes (Hyman et al., 2000). Because they grow more rapidly,
adding oxygen to environmental samples that contain a mixture of petroleum hydrocarbons and MTBE will most likely
enrich for organisms that co-metabolize MTBE.
4.2 Biochemistry of Aerobic MTBE Biodegradation
A simplified and generalized pathway for complete aerobic metabolism of MTBE is presented in Figure 4.1. The figure
combines features in the pathways published by Fiorenza and Rifai (2003), Steffan et al., (1997) and Deeb et al., (2000a).
In every aerobic organism studied to date, the first transformation is believed to be carried out by a mono-oxygenase
enzyme. These enzymes insert one oxygen atom from molecular oxygen into the organic compound being metabolized.
The other oxygen atom is reduced to form water. The first stable products are TEA and either formaldehyde or formic
acid. Formaldehyde and formic acid are very readily degraded.
CH,
CH3-C-O-CH3
l
CH,
O
C
i
OH
CH3-C-OH
I
CH,
H
l
CH3-C-OH
CH3
l
CH3-C-OH
l
CH,
CH2-OH
CH,-C-OH
CH,
O
CH3-C
CH,
Figure 4.1. Significant products of the aerobic biodegradation of MTBE. (After Wilson 2003b). MHP is methyl-2-hy-
droxy-1-propanol. HIBA is 2- hydroxyisobutyric acid.
Often the resulting TEA will accumulate in ground water. Kane et al., (2001) showed the transitory accumulation of
TEA during aerobic biodegradation of MTBE by naturally occurring bacteria from a gasoline spill site in Palo Alto,
California. Hunkeler et al. (2001) showed transitory accumulation of TEA in laboratory cultures during aerobic growth
on MTBE and during aerobic co-metabolism of MTBE supported by 3-methylpentane. Apparently, the native microbial
population at the spill site included organisms that could degrade MTBE and TEA. The MTBE was degraded first, and
then the TEA was degraded if the supply of oxygen was sufficient.
The TEA can be further transformed through a second attack by a mono-oxygenase to form 2-methyl-2-hydroxy-l-
propanol (MHP). The MHP is further oxidized to 2-hydroxyisobutyric acid (HIBA). HIBA has been detected in ground
water at a gasoline spill site (Personal Communication, Pat McLoughlin, Microseeps Inc., Pittsburg, PA). Elimination
of the carboxylic acid group from HIBA produces 2-propanol (isopropyl alcohol), which in turn can be oxidized to
acetone.
Acetone and 2-propanol are rapidly degraded in aerobic ground waters to carbon dioxide, water, and biomass. These
compounds should be more persistent in anaerobic ground water than in aerobic ground water, but they should eventually
degrade. Acetone is occasionally reported in ground water from gasoline spills. It has conventionally been attributed
21
-------
to contamination of the field sample by acetone in the laboratory. There is a strong possibility that the acetone reported
in ground water samples from gasoline spills was a biodegradation product of MTBE or TEA.
4.3 Aerobic MTBE Biodegradation in Ground Water
In 1988, the University of Waterloo conducted a large controlled-release study of MTBE degradation in a sandy gla-
cial aquifer at Canadian Forces Base Borden in Ontario, Canada, (Hubbard et al., 1994; Schirmer and Barker 1998;
Schirmer et al., 1999). They injected ground water containing 19 mg/L BTEX and 269 mg/L MTBE into the aquifer
and monitored the degradation of BTEX and MTBE in the plume. The BTEX compounds were completely removed
within 476 days. However, there was no statistically significant evidence for biodegradation of MTBE (Hubbard et al.,
1994). Many of the readers of Hubbard et al. (1994) interpreted the lack of evidence for biodegradation of MTBE as
evidence that MTBE would not biologically degrade in ground water. This report supported and reinforced the con-
ventional wisdom at the time that MTBE did not biodegrade in aquifers.
Researchers at the University of Waterloo sampled the MTBE plume again in 1995. The concentrations of MTBE were
much lower than expected based on dilution and dispersion alone. In 1996, they sampled the plume using a fine grid to
give themselves greater confidence in their estimate of the mass of MTBE remaining in the aquifer. After 3,000 days of
residence time, only 3% of the MTBE originally injected into the aquifer remained in the aquifer (Schirmer and Barker,
1998; Schirmer et al., 1999). When they used sediment from the aquifer to construct laboratory microcosms, acclima-
tion to degrade MTBE was a rare event; only 3 of 40 microcosms acclimated after 20 months of incubation. However,
once the acclimation event occurred in laboratory microcosms, biodegradation was rapid and extensive. Based on this
finding, they attributed the disappearance of MTBE in the field scale plume to aerobic biodegradation.
There seems to be a wide variation from one gasoline spill to another in the distribution and activity of native microor-
ganisms that can degrade MTBE. Salinitro et al., (1998) surveyed sites for the presence of MTBE degrading bacteria.
They examined ground water and soil from ten sites: two retail sites in California, refineries in Louisiana and Illinois,
distribution terminals in Nevada and Ohio, a pipeline in Texas, and retail sites in Michigan, Texas, and New Jersey.
They were able to isolate MTBE degrading organisms from two of the ten sites and demonstrate MTBE degradation
in microcosms constructed with material from two sites. Similarly, Kane et al., (2003) [see also Kane et al., (2001)]
constructed microcosms with material from seven MTBE spills in California; MTBE was degraded in sediment from
only three of the sites. In sediment from the other four sites, MTBE was not degraded within the period of incubation
(170 days to 350 days depending on the site).
Hanson et al., (1999) reported the isolation of strain PM-1, a pure culture that can degrade MTBE as the primary sub-
strate. The strain was isolated from a biofilter that was used to treat the off-gases from a sewage treatment plant that
received discharges from a local refinery. Recent developments in molecular genetics make it possible to identify the
genes of particular bacteria in samples of aquifer material. Using denaturing gradient gel electrophoresis, Kane et al.,
(2003) showed that the sediment from all three of the gasoline spill sites in California that degraded MTBE harbored
bacteria with DNA similar to strain PM-1. Sediment that did not degrade MTBE did not harbor bacteria with DNA
similar to PM-1. Hristova et al., (2003) isolated bacteria containing DNA very similar to PM-1 from an aerobic bio-
logical treatment system for MTBE at a gasoline spill site on Vandenberg AFB, California. Many of the bacteria that
have been studied to date that degrade MTBE under aerobic conditions in ground water seem to be closely related to
strain PM-1.
Strain PM-1, which can grow on MTBE as a sole carbon source, also grows readily on benzene. While PM-1 is growing
on benzene, it does not degrade MTBE. Once the benzene is exhausted, it produces the enzymes necessary to degrade
MTBE and starts to grow on MTBE (Deeb et al., 2000a; Deeb et al., 2000b; Deeb et al., 2001). Benzene present in a
gasoline spill could enrich PM-1 to high density and prepare the site for rapid MTBE biodegradation once the BTEX
compounds were exhausted. Hyman (2000) noted that organisms that can degrade alkanes and isoalkanes and can co-
metabolize MTBE are common in ground water at gasoline spill sites. Alkanes, isoalkanes, and the BTEX compounds
are the major components of gasoline. These components of gasoline may enrich for organisms that can degrade MTBE
once the gasoline has been exhausted.
The prospects for natural aerobic biodegradation of MTBE by native microorganisms may be related to the age of the
spill, the time that has been available for acclimation of the native microorganisms to MTBE, and perhaps to the seep-
age velocity of ground water. Because the organisms that can degrade MTBE grow so slowly, acclimation to degrade
MTBE may require several years, as was the case at Canadian Forces Base Borden in Ontario, Canada. If a release
starts as a slow pinhole release, and only later grows large enough to be noticed, there may be time for acclimation
22
-------
before the major portion of the release occurs. When oxygen is added to support aerobic biodegradation of MTBE in
old anaerobic plumes, they often acclimate in weeks to months (Salanitro et al., 2000; Wilson et al., 2002).
Can the leading edge of an MTBE plume outrun the bacteria and escape biodegradation? Not in the long term. Al-
though most bacteria in aquifers are associated with surfaces, many of them are planktonic. The planktonic bacteria
are already in the ground water and move with the ground water. Any transport process that will advance the MTBE
will advance the bacteria. Their motion is with respect to the plume itself. They move within the plume even as the
plume advances through the aquifer. In general, flagellated planktonic bacteria can move no more than 6 cm a day
through flowing ground water. If the seepage velocity of a plume is high, a plume may get to be very large before the
acclimation event occurs. Once acclimation has occurred at a particular point, it may take a long period of time for the
bacteria to spread throughout the rest of the plume.
4.4 Anaerobic Biodegradation of MTBE
There are reports in the literature of MTBE biodegradation under nitrate-reducing conditions (Bradley et al., 200la),
sulfate-reducing conditions (Bradley et al., 2001b; Somsamak et al., 2001), iron-reducing conditions (Landmeyer et al.,
1998; Bradley et al., 2001b; Finneran, and Lovley, 2003), and methanogenic conditions (Mormile et al., 1994; Wilson
et al., 2000; Bradley et al., 2001b; Kolhatkar et al., 2002; Somsamak et al., 2005; Wilson et al., 2005a).
Bradley and co-workers added radio-labeled MTBE to stream bed sediments and compared the distribution of biodeg-
radation products under nitrate-reducing, sulfate-reducing, iron-reducing, and methanogenic conditions (Bradley et al.,
2001a and 2001b). Table 4.2 summarizes some of their results. Under nitrate-reducing conditions, the MTBE that was
degraded was completely metabolized to carbon dioxide. Under sulfate-reducing conditions, iron-reducing conditions,
and methanogenic conditions, TEA accumulated to a greater or lesser extent in the different sediments.
In these experiments, the MTBE was uniformly labeled with 14C. Only 20% of the radio-label in the MTBE added to
the sediment was associated with the methoxyl-carbon of MTBE. If the label recovered as carbon dioxide exceeded
25% of the label recovered as TEA, then some portion of the TEA produced from MTBE biodegradation must have been
further metabolized. A sulfate reducing culture described by Somsamak et al., (2001) did not degrade TEA. However,
in the microcosm study of Bradley et al. (2001b), some portion of the TEA produced during MTBE biodegradation
was further oxidized to carbon dioxide under sulfate-reducing conditions in all three sediments tested. In one of the
sediments tested by Bradley et al. (2001a, 2001b), most of the TEA produced during biodegradation of MTBE under
iron-reducing conditions was further oxidized to carbon dioxide (Table 4.2).
Table 4.2 Distribution (in percent) of Biodegradation Products of MTBE under Nitrate-reducing, Sulfate-reducing,
Iron-reducing, and Methanogenic Conditions (Bradley et al., 2001a and 2001b)
Amendment
Nitrate
Sulfate
Iron (III)
None, all
sediments are
methanogenic
Location
FL
SC
NJ
SC
FL
FL
SC
NJ
FL
SC
NJ
FL
SC
NJ
SC
FL
Duration
days
166
166
166
77
77
166
166
166
166
166
166
166
166
166
77
77
MTBE
TEA
CO2 Methane Total
Percent of original radio-label in MTBE
29 ±2
72 ± 1
81 ±2
70 ± 1
71 ±4
82 ±3
81 ±9
82 ±3
88 ±3
92 ± 12
81 ± 10
82 ±3
81 ±9
82 ±3
85 ±5
78 ±4
nd
nd
nd
nd
1 ± 1
1 ± 1
9±7
3±0
9± 1
8±4
4± 1
1±1
9±7
3±0
10 ±2
9±2
75 ± 1
33 ±8
23 ±5
26 ±10
23 ±5
20 ±4
9±3
12 ±3
nd
nd
14 ±4
20 ±4
9±3
12 ±3
3±3
5±2
nd
nd
nd
nd
nd
nd
nd
nd
3±2
3±2
nd
9±3
9±3
9±3
1±1
5±1
104 ± 12
105 ±7
104 ±4
96 ±10
95 ±5
103 ±4
104 ± 12
97 ±2
100 ±3
102 ± 10
99 ±10
103 ±4
104 ±2
97 ±2
99 ±5
97 ±4
nd - means not detected.
23
-------
These studies were done with mixed microbial communities in sediments. The redox potential and the exposure to
MTBE varied from sediment to sediment. An observation that MTBE biodegradation occurred under methanogenic
conditions in the sediment does not mean that the MTBE was degraded by methanogenic organisms. Similarly, an
observation that MTBE degradation occurred under iron-reducing conditions does not mean that the MTBE was de-
graded by iron-reducing organisms. Several electron-accepting processes can occur concomitantly in aquifer material.
In a recent review, Schmidt et al., (2004) noted that most of the laboratory studies conducted to date have failed to
associate anaerobic MTBE biodegradation with a specific electron accepting process. However, these studies do show
that MTBE can be degraded under oxidation-reduction conditions that are common in contaminated ground water at
gasoline spill sites.
Somsamak et al., (2005) reported degradation of MTBE to TEA in enrichments from a microcosm that was originally
methanogenic. Oxygen, nitrate, sulfate, and iron (III) were not available. When methanogenesis was inhibited with
20 mM 2-bromomethanesulfonic acid (BBS), MTBE continued to degrade. Somsamak et al., (2005) speculated that
anaerobic MTBE biodegradation in their culture may have been carried out by homoactogenic bacteria. These bacteria
can metabolize a methyl-ether using molecular hydrogen and bicarbonate ion to produce acetate and the corresponding
alcohol. Wilson et al., (2005b) compared the Gibbs free energy for the metabolism of molecular hydrogen by methano-
gens to metabolism by homoactogens using MTBE. At concentrations of molecular hydrogen that would be expected
in ground water, there is more energy available to the homoactogens, and they would be expected to have a competitive
advantage over the methanogens that use hydrogen.
Although these studies prove that anaerobic biodegradation of MTBE in sediments is possible, they do not indicate that
anaerobic biodegradation in aquifer sediments is common or pervasive. Amerson and Johnson (2002) added MTBE
labeled with 13C to a large MTBE plume at Port Hueneme, California. They found no evidence for loss of the MTBE
labeled with 13C over the course of one year. Landmeyer et al. (1998) documented degradation of MTBE under iron-
reducing conditions in microcosms constructed with sediment impacted by a gasoline spill site in South Carolina.
Although the removal was statistically significant, the rate was very slow (2% ± 0.6% oxidized to CO2in four months).
The overall rate of biodegradation in the plume in South Carolina was less than 0.04 per year (Landmeyer et al., 2001).
Biodegradation would have had minimal influence on distribution of contamination in the plume. Anaerobic biodegrada-
tion of MTBE may have little effect on the distribution of the plume of MTBE at many gasoline spill sites.
Wilson and coworkers constructed microcosms with contaminated sediment from aquifers that had been impacted by
spills of gasoline or neat MTBE (Wilson et al., 2000; Kolhatkar et al., 2002; Kuder et al., 2002; Kuder et al., 2003;
Wilson et al., 2005a; Adair et al., unpublished). Sediment from gasoline spill sites at Boca Raton, Florida; Parsippany,
New Jersey; Deer Park, New York; Petaluma, California; and Vandenberg AFB, California; were amended with either
2 mg/L MTBE, or 2 mg/L TEA, or 2 mg/L benzene, or 2,000 mg/L ethanol. Sediment from a gasoline spill site at
Port Hueneme, California, was amended with 10 mg/L MTBE or 10 mg/L TEA, but not with benzene. Sediment from
a JP-4 jet spill at Elizabeth City, North Carolina, was amended with MTBE or TEA or benzene, but not with ethanol.
Sediment from a spill of neat MTBE at a tank farm in Nederland, Texas, was mixed together until the concentration
of MTBE and TEA already present in the sediment was uniform, but the sediment was not amended with additional
MTBE, TEA, benzene, or ethanol. The sediments were incubated for up to 24 months in an anaerobic glove box. If
more than 90% of the material was removed compared to a sterilized control, the material was considered to have
biologically degraded.
Benzene degraded under anaerobic conditions in all six of the sediments tested (Table 4.3). The sediments were depleted
of oxygen, nitrate, and sulfate before the microcosms were constructed. Biodegradation of benzene in the microcosms
occurred under methanogenic conditions or possibly under iron-reducing conditions. Biodegradation of benzene was
not tested in sediment from the Port Hueneme, California, site. At this site, benzene degrades readily under sulfate
reducing conditions in the field scale plume.
There was no evidence that MTBE degraded under anaerobic conditions at the Port Hueneme, California, site (Amerson
and Johnson; 2002), and the Vandenberg AFB, California, site (Wilson et al., 2002). Because MTBE did not degrade
at field scale, degradation was not expected in the microcosms. As expected, MTBE did not degrade in sediment from
these sites (Table 4.3). There was evidence at field scale that MTBE was degrading under anaerobic conditions at the
Parsippany, New Jersey, site (Kolhatkar et al., 2002), at the Elizabeth City, North Carolina site (Wilson et al., 2000),
at the Deer Park, New York, site (Kolhatkar et al., 2002), at the Boca Raton, Florida, site (case files), the Petaluma,
California, site (case files), and at the Nederland, Texas, site (case files). Because MTBE appeared to degrade at field
scale, MTBE degradation was expected in the microcosms. In contrast to the behavior of benzene, MTBE degraded in
24
-------
Table 4.3 Anaerobic Biodegradation of MTBE, TEA, Benzene, and Ethanol in Microcosms Constructed with Aqui-
fer Sediment
Location
Boca Raton, FL
Parsippany, NJ
Deer Park, NY
Petaluma, CA
Port Hueneme,
CA
Vandenberg AFB,
CA
Elizabeth City,
NC
Nederland,TX
MTBE
in plume
TEA
in
plume
Benzene
in plume
Hg/L
106
790
1,240
9,900
4,700
49,000
154
1,500,000
11,600
600
96
<2,500
3,300
3,400
46
68,000
132
4,900
8
30,000
<5
124
1,280
65
MTBE
degraded
No
Yes
No
No
No
No
Yes
Yes
TEA
degraded
No
No
No
Yes
No
No
No
No
Benzene
degraded
Yes
Yes
Yes
Yes
Not Tested
Yes
Yes
Not Tested
Ethanol
degraded
No
Yes
Yes, but Slow
Yes
Yes
Yes
Not Tested
Not Tested
only three of the six sediments tested where there was evidence that MTBE degraded at field scale. In the microcosms
where MTBE degraded, there was a near stoichiometric production in TEA (Wilson et al., 2005a).
Novak and his co-workers surveyed anaerobic biodegradation of TEA in sediments from uncontaminated subsurface
material from Blacksburg, Virginia; Newport News, Virginia; and Williamsport, Pennsylvania (Novak et al., 1985;
Hickman et al., 1989; Hickman and Novak, 1989; Yeh and Novak, 1994). In these studies, TEA degraded readily un-
der nitrate-reducing and sulfate-reducing conditions. Finneran and Lovley (2001) reported degradation of TEA under
iron-reducing conditions in sediment from a gasoline spill site in South Carolina and in sediment from the Potomac
River. Finneran and Lovely (2003) demonstrated TEA biodegradation under methanogenic conditions in sediment from
a refinery spill site in Oklahoma. Day and Gulliver (2003) used long-term monitoring data and analysis of stable carbon
isotopes to document natural anaerobic biodegradation of TEA in ground water at a refinery site in Texas.
In contrast to this precedent for TEA biodegradation in the literature, TEA did not degrade in aquifer sediment under
anaerobic conditions in studies reported by Suflita and Mormile (1993) and Mormile et al. (1994), and TEA degraded
in only one of eight sediments from gasoline spill sites (Table 4.3). At this writing, anaerobic biodegradation of TEA
at gasoline spill sites has not been well documented.
The degradation of ethanol was rapid in sediments from the gasoline spill sites at Petaluma, California; Parsippany, New
Jersey; and Vandenberg AFB, California; the first order rate of attenuation was greater than 0.1 per day. In sediments
from the sites at Deer Park, New York, and Port Hueneme, California, the rate was slower, but still environmentally
significant. The rates were greater than 2 per year.
Ethanol did not degrade in sediment from Boca Raton, Florida. The first order rate of degradation was less than 0.05 per
year. Acidic conditions can inhibit anaerobic biodegradation of ethanol; however, low pH does not explain the absence
of ethanol degradation in the sediment from Boca Raton, Florida. The pH of the sediment from Boca Raton was 6.5.
The sediment used to prepare the microcosms had high concentrations of BTEX compounds and trimethylbenzenes.
The demand for fixed nitrogen or for sulfate to metabolize the aromatic hydrocarbons may have depleted the supply of
the nutrients. The pore water in the microcosms contained approximately 1.0 mg/L of ammonia N, and less 0.3 mg/L
of sulfate.
25
-------
4.5 Acclimation to Anaerobic Biodegradation of MTBE
Prior to 1999, the Local Oversight Program of the Environmental Health Division of the Health Care Agency in Orange
County (California) used EPA method 8020 or 8021 for routine monitoring of fuel-derived contaminants. Concentra-
tions of TEA were not reported. In 2000 and 2001, they transitioned their monitoring to EPA method 8260 or 8260B
(purge and trap with gas chromatography with a mass spectrometer detector), and concentrations of TEA were routinely
reported. The concentrations of TEA were higher than they had expected (Figure 4.2). On average, the concentrations
of TEA were higher than the concentrations of MTBE. A similar relationship for TEA and MTBE was reported for
neighboring Los Angeles County (Shih et al., 2004).
The Local Oversight Program selected thirteen gasoline spill sites for detailed evaluation. All of the sites had high
concentrations of TEA. Some had low concentrations of MTBE, and some had higher concentrations of MTBE
(Figure 4.2). In 35 of 59 wells at the sites selected for detailed study, the concentrations of TEA and MTBE followed
a pattern illustrated in Figure 4.3. Initially, each monitoring well produced ground water with higher concentrations
of MTBE and lower concentrations of TEA. After a period of time, the well started to produce ground water with
high concentrations of TEA and much lower concentrations of MTBE. The transition from water that was dominated
by MTBE to water that was dominated by TEA was usually rapid, taking only a few months to a year. The decline
in concentrations of MTBE and increase in concentrations of TEA occurred in water that was anaerobic. This sharp
transition was probably a result of the acclimation of anaerobic microorganisms in the aquifer to degrade MTBE.
The field data in Figure 4.3 are typical of field data from the 35 wells that exhibited the transition from MTBE to TEA.
Sometimes more TEA is produced than would be expected from the biodegradation of concentrations of MTBE that
were present at an earlier time; sometimes less. This most likely reflects changes in the concentration of MTBE in
the ground water before biodegradation, caused by changes in ground water elevation, or changes in flow direction, or
some other influence.
This behavior of MTBE at field scale is mirrored in the behavior of MTBE in anaerobic microcosms. Figure 4.4 pres-
ents data from a microcosm study conducted with sediment from Parsippany, New Jersey (Table 4.3; Kolhatkar et al.,
2002; Wilson et al., 2005a). After a lag, the transition from MTBE to TEA was rapid. The first order rate of attenuation
of concentrations of MTBE (including the lag) in the microcosms was 11.7 ± 2.4 per year at 90% confidence. In the
microcosm data presented in Figure 4.4, the MTBE that was removed was replaced with an equivalent concentration
of TEA, confirming that MTBE was being transformed to TEA.
1,000,000
100,000
10,000
o
52 1,000
CO
fa
g 100
o
o
to
CD
g> 10
±
,0
'••*.'.I- ''•*.(••:••«"Ł"• . •
• . v •.•.•.•"•;.*•-.••• ••• •
*.•*!** • *.i • „ . •**•• * *•
• • • V| •/*•••#• . • •
... t. •.•;. .yrf .••
D
Sites Selected for
Detailed Study
100 1,000 10,000 100,000 1,000,000
Highest MTBE at Station
10
Figure 4.2 Distribution of TEA and MTBE in the most contaminated wells at gasoline spill sites in Orange County,
California, in 2002.
26
-------
200
3/15/2000 10/1/2000 4/19/2001 11/5/2001 5/24/2002 12/10/2002
Date of Sample
100
80 -
60 H
o
•j=
I
§ 40 H
<§
20 -
0
OOUT038 MW-08
10/1/2000 4/19/2001 11/5/2001 5/24/2002 12/10/2002 6/28/2003
Date of Sample
Figure 4.3 Transition from MTBE to TEA in monitoring wells at gasoline spill sites in Orange County, California.
The extent of biodegradation of MTBE in the microcosms and enrichment cultures from the microcosms could be pre-
dicted from the fractionation of the stable carbon isotopes of MTBE remaining in the microcosm after biodegradation.
(Kolhatkar et al., 2002; Kuder et al., 2004). The process is discussed in detail in Section 5. The relationship between
the fractionation of the stable carbon isotopes in MTBE and the extent of biodegradation of MTBE in microcosms was
used to predict the extent of biodegradation of MTBE in the ground water in Orange County, California (Kuder et al.
2004; Wilson et al., 2005b). The results are presented in Table 4.4. At 12 of the 13 sites, most of the MTBE in the
most contaminated well had been biologically degraded. At seven of the thirteen sites, more than 90% of the MTBE
had degraded.
27
-------
50 100 150
Time of Incubation (days)
200
.g
-b 0.1 -
I
o
O
0.01 -
0.001
50 100 150
Time of Incubation (days)
200
Figure 4.4 -Anaerobic biodegradation ofMTBE and production of TEA in microcosms constructed with sediment from
a gasoline spill site. The error bars are the 95% confidence intervals on the geometric mean concentra-
tion. The same data are plotted on an arithmetic scale in the upper panel and a logarithmic scale in the
lower panel.
28
-------
Table 4.4 Biodegradation of MTBE in the Most Contaminated Well at 13 Gasoline Spill Sites in Orange County,
California, as Predicted by Fractionation of 13C in MTBE. 513C is the Unit Used to Measure the Isotopic
Fractionation (See Section 5 for a Detailed Explanation.)
Location
99UT015
99UT032
96UT028
87UT211
91UT086
OOUT038
86UT175
89UT007
86UT062
88UT138
88UT198
85UT114
MTBE
(MS/Q
280,000
2,500
2,650
268
890
5.49
100
820
20
16.9
100
100
TEA
(M8/L)
300,000
24,000
190,000
136,000
81,000
51,600
80,000
29,000
13,000
180,000
41,000
110,000
513C MTBE
(°/ )
v oo7
-30.32
-18.51
-13.29
-12.37
-0.71
5.29
6.04
6.84
15.95
24.03
27.06
56.78
Fraction
MTBE remaining
Biodegradation not expected
0.478
0.312
0.289
0.111
0.068
0.064
0.060
0.028
0.015
0.011
0.001
4.6 Zero Order Biodegradation of MTBE at High Concentrations -
All the equations used in this report assume that the anaerobic biodegradation of MTBE is a first order process. If the
removal is a first order process, the data will plot along a straight line when time or distance is plotted on an arithmetic
scale and concentrations of MTBE are plotted on a logarithmic scale. The lower panel in Figure 4.4 plots the data with
concentration of MTBE on a logarithmic scale. The concentrations of MTBE fall along a straight line from day 61 of
the incubation to day 179 of the incubation, indicating that the anaerobic biodegradation of MTBE in the microcosms
was a first order process. The initial concentration of MTBE in the microcosms constructed with sediment from Par-
sippany, New Jersey, was near 2 mg/L. The anaerobic degradation of MTBE in microcosms constructed with material
from a JP-4 spill in Elizabeth City, North Carolina, was also first order (Wilson et al., 2000). The initial concentration
of MTBE in the sediment from Elizabeth City, North Carolina was near 3 mg/L. At these concentrations, and at lower
concentrations, the anaerobic biodegradation of MTBE can be expected to be a first order process. First order rate
constants have a unit of reciprocal time, such as, per day or per year.
At higher concentrations, the enzymes that metabolize MTBE may become saturated. As a consequence, the bacteria
degrade MTBE at some fixed maximum rate, regardless of the concentration of MTBE. Under these conditions, the
rate of degradation is described as a zero order process. Degradation follows Equation 4.1, where Co is the initial
concentration, C is the final concentration, t is the elapsed time, and K is the zero order rate constant. Typical units for
K would be mg/L per day.
C = Co-Kt Equation 4.1
The microcosms that produced the data in Figure 4.4 were respiked with MTBE at an initial concentration near
100 mg/L. The biodegradation at this higher concentration is presented in Figure 4.5. If biodegradation is zero order,
the data should plot along a straight line when the concentrations are plotted on an arithmetic scale. The upper panel
in Figure 4.5 indicates that biodegradation proceeded without a lag, and that the degradation was a reasonably good fit
to a zero order process.
The lower panel of Figure 4.5 plots the same data on a logarithmic scale. The logarithmic plot suggests that the biodeg-
radation of MTBE behaved like a zero order process for the first four sampling dates, and like a first order process for
the remainder of the incubation. In the first four sampling dates, the zero order rate of biodegradation was 0.20 mg/L
29
-------
per day. In the remainder of the incubation, the first order rate of biodegradation was 10 ± 3.4 per year, which is not
statistically different from the first order rate achieved by the organisms in the sediment in the experiments described
in Figure 4.4.
The transition from zero order to first order biodegradation occurred at concentrations of MTBE between 65 and 40 mg/L.
Using concentrations of MTBE above 40 mg/L in Equations 2.2 and 2.3 will likely produce errors, but the errors are
conservative. The MTBE will degrade more rapidly than would be predicted by the equations.
50
100 150 200
Time of Incubation (days)
250
300
100
LU
P
o 10
o
U
50 100 150 200
Time of Incubation (days)
250
300
Figure 4.5 -Anaerobic biodegradation of MTBE at high concentration in an enrichment culture constructed with
sediment from the microcosms used to produce the data in Figure 4.4. The same data are plotted on an
arithmetic scale in the upper panel and a logarithmic scale in the lower panel.
30
-------
Section 5 -
Monitoring MTBE Biodegradation with Stable Isotope Ratios -
A new technique has been developed to evaluate the extent of MTBE biodegradation at field scale. The technique is
based on the fractionation of the stable carbon isotopes in the remaining MTBE during the course of degradation. The
fractionation of the stable carbon isotopes of MTBE can provide an unequivocal indication of MTBE biodegradation.
At this writing, these analyses are only commercially available from a few university laboratories. Their costs are on
the order of $250 per analysis. In the future, reports provided to regulators to evaluate the risk from MTBE plumes are
likely to contain data on the ratio of stable carbon isotopes in the residual MTBE in the ground water. This is particularly
true when the possibility of natural biodegradation of MTBE is crucial to the risk evaluation of a gasoline spill.
These analyses are not necessary at gasoline spill sites where the possibility of MTBE biodegradation does not change the
site conceptual model or the strategy for risk management at the site. It is not necessary to evaluate MTBE fractionation
in every well at a site. The analyses should be reserved for water from wells that are critical to the risk analysis.
This section provides a visual illustration of the process of isotope fractionation during biodegradation. It also provides
numerical examples of the fractionation that can be expected during anaerobic biodegradation. This section explains the
units used to express the ratio of stable carbon isotopes and presents simple formulas and graphs to predict the extent
of biodegradation from the measured stable carbon isotope ratio.
5.1 Monitoring MTBE Biodegradation with Stable Isotope Ratios
There are two stable isotopes of carbon: carbon twelve (12C) and carbon thirteen (13C). Unlike carbon fourteen (14C), the
stable isotopes are not subject to radioactive decay. The most prevalent stable isotope of carbon is 12C. Approximately
1% of the carbon on the Earth is 13C. During MTBE biodegradation, MTBE molecules with 12C at the methoxy group
are metabolized more rapidly than MTBE molecules with 13C at the methoxyl group (Hunkeler et al., 2001; Gray et al.,
2002; Kolhatkar et al., 2002; Kuder et al., 2004, 2005). This discrimination against the heavier isotope is called the
kinetic isotope effect. As biodegradation of MTBE proceeds, the remaining MTBE contains a progressively greater
proportion of the 13C isotope. As a consequence, the extent of biodegradation can be determined from the change in
the ratio of stable isotopes in the MTBE. Figure 5.1 is a pictorial illustration of this fractionation. The figure greatly
exaggerates the relative fractionation compared to 12C and 13C to make it easier to see the effect.
oooooooooooo
o«oo«oo«oo«o
oooooooooooo
oooooooooooo
o»oo«oo«oo«o
oooooooooooo
oooooooooooo
O»OO»OO»OO«O
oooooooooooo
oooooooooooo
O»OO«OO»OO«O
oooooooooooo
oooooooooooo
o«oo«oo«oo«o
oooooooooooo
oooooooooooo
o»oo«oo»oo«o
oooooooooooo
oooooooooooo
O»OO»OO»OO»O
oooooooooooo
oooooooooooo
O»OO»OO»OO«O
oooooooooooo
•oooo»ooooo«
o»oo»oo»oo»o
oo«oooooo«oo
ooo»oooo»ooo
o»oo»oo»oo«o
•ooooo«oooo«
oooooooooooo
o»oo«oo»oo«o
oo«oooooo»oo
ooo«oooo»ooo
o«oo«oo»oo«o
oooooooooooo
oooooooooooo
ooo« 0000*000
oo»o 00000*00
o»oo»oo»oo»o
oooooooooooo
•oo«o«
o»oo»o
•o«ooo
ooo«o«
o»oo«o
•ooo«o
Figure 5.1 An illustration of the kinetic isotope effect. Shown is the enrichment in black dots, or fractionation of the
proportion of black dots to white dots, when the rate of removal of white dots is faster than the removal of
black dots.
31
-------
The first order rate of biodegradation (k) can be calculated from the initial concentration of MTBE (C ), the final
concentration of MTBE (C), and the time elapsed (t) following Equation 5.1. Equation 5.1 rearranges the terms in
Equation 2.2 in Chapter 2.
k = -ln(C/C0)/t Equation 5.1
The numbers of black dots and white dots at each time step in Figure 5.1 are presented in Table 5.1. Half of the dots
are degraded in one time step. Each time step is one half-life for the dots. From Table 5.1, the number of black dots at
time zero was 32, and the number after time step three was 13. If 13 is substituted for C, 32 is substituted for C0, and
3 is substituted for t in Equation 5.1, the first order rate of attenuation of black dots is 0.300 per time step. Similarly,
the rate of attenuation of white dots is 0.782 per time step, and the rate of attenuation of all the dots is 0.693 per time
step. The rate of removal of white dots was 2.6 times faster than the rate of removal of black dots, and the black dots
became progressively more abundant relative to the white dots with each time step. This visual analogy will be applied
to the fractionation of stable carbon isotopes in MTBE during the course of biodegradation.
Table 5.1 Fractionation of Black Dots and White Dots in the Visual Example in Figure 5.1
Time Step
(half-lives)
zero
one
two
three
Number of All
Dots Remaining
288
144
72
36
White Dots Re-
maining
240
120
54
23
Black Dots Re-
maining
32
24
18
13
Ratio Black Dots to White Dots
0.13
0.20
0.33
0.57
Recent advances in analytical chemistry make it possible to determine the ratio of stable isotopes in MTBE dissolved in
a water sample at concentrations that are near regulatory standards (Hunkeler et al., 2001; Kolhatkar et al., 2002). The
MTBE is separated from water by purge and trap (Kolhatkar et al., 2002) or by solid phase microextraction (Hunkeler
et al., 2001), and then further separated by gas chromatography, and finally the ratio determined with an isotope ratio
mass spectrometer. The effective minimum concentration of MTBE for analysis of the stable carbon isotope ratio is
near 10 |ig/L.
The isotope ratio mass spectrometer does not measure the ratio of the stable carbon isotopes directly to each other.
Rather, it measures the deviation of the ratio in the sample from the ratio in a standard substance that is used to calibrate
the instrument. The substance used as the international standard for stable carbon isotopes has a ratio of 13C to 12C of
0.0112372.
The conventional notation for the ratio of 13C to 12C in a sample (513C) reports the ratio in terms of its deviation from
the ratio in the standard.
cl3x-<
o C =
(13c/12c) -(13c/12c)
V / sample \ ' standard
(13C/12C)
V 'standard
xlOOO
Equation (5.2)
The units for 513C are parts per thousand, often represented as %o, or per mil, or per mill. Table 5.2 presents calcula-
tions that illustrate the changes in 5 13C when carbon in MTBE is fractionated during biological degradation. In the
example calculations, the abundance of 13C starts out at 1.092% of the total carbon. This abundance is in the range
typically encountered in MTBE in gasoline. The 513C calculated following Equation 5.2 is -28.2 %o. The first order
rate of biodegradation of MTBE containing only 12C was 0.6927 per time step, while the rate of biodegradation of
MTBE containing one atom of 13C was 0.6852 per time step. The rate of degradation of MTBE containing only 12C is
approximately 1.09% faster than the rate of degradation of MTBE containing one atom of 13C. This ratio of the rates
of biodegradation is typical for anaerobic biodegradation of MTBE.
After ten time steps, the concentration of MTBE is reduced nearly one thousand fold, and the ratio of 13C to 12C in-
creased from 0.01092 to 0.01182. The value of 513C shifts from -28.4 %0 to +52.2 %0. This is the range of values of
513C in MTBE typically seen at field sites to date. The value of 513C can usually be determined with a reproducibility
32
-------
Table 5.2 Typical Changes in the Ratio of 13C to 12C in MTBE During Biodegradation of MTBE under Anaerobic
Conditions
Step
0
1
2
3
4
5
6
7
8
9
10
Fraction Remaining
MTBE
1.000
0.500
0.250
0.125
0.063
0.031
0.016
0.008
0.004
0.002
0.001
Fraction Remaining
12C
0.989
0.495
0.247
0.124
0.0618
0.0309
0.0155
0.00773
0.00386
0.00193
0.00097
Fraction Remaining
!3C
0.0108
0.00544
0.00274
0.00138
0.000697
0.000351
0.000177
8.9E-05
4.4E-05
2.2E-05
1.1E-05
13Q12C
0.01092
0.01101
0.01109
0.01118
0.01127
0.01136
0.01145
0.01154
0.01164
0.01173
0.01182
5 13C (%o)
-28.2
-20.2
-13.1
-5.1
2.9
10.9
18.9
26.9
35.8
43.9
51.9
of better than ± 0.5%o. The value of 513C before biodegradation was negative because MTBE in gasoline has relatively
less 13C than does the international standard used to calibrate the mass spectrometer. As biodegradation proceeded, the
value of 513C became more positive.
Figure 5.2 plots the value of 513C in Table 5.2 against the fraction of MTBE remaining after each step. Notice that
the extent of biodegradation increases and the fraction of MTBE remaining decreases toward the right-hand side of
the x-axis. Table 5.2 contains data that are typical for anaerobic biodegradation of MTBE; Figure 5.2 also plots data
typical of aerobic biodegradation. In the calculations for aerobic biodegradation, the first order rate of degradation of
MTBE composed entirely of 12C was 0.6927 per time step (the same as for calculations for anaerobic biodegradation).
The rate of aerobic biodegradation of MTBE containing one atom of 13C was 0.6915 per time step. The rate of aerobic
degradation of MTBE composed entirely of 12C is approximately 0.17% faster than the rate of degradation of MTBE
containing one atom of 13C. This ratio of the rates of biodegradation is typical for aerobic biodegradation of MTBE.
0)
a.
-"- anaerobic biodegradation
•*- aerobic biodegradation
-40
0.1 0.01
Fraction MTBE remaining (C/Co)
0.001
Figure 5.2 Typical changes in the value of?>BC as MTBE is degraded under aerobic and anaerobic conditions.
33
-------
The rate of anaerobic biodegradation of MTBE containing one atom of 13C in each time step is slower than the rate
under aerobic conditions. There will be a greater tendency for 13C MTBE to accumulate under anaerobic conditions,
and MTBE will fractionate more rapidly under anaerobic conditions.
These assumptions are consistent with what is known of the physiology of MTBE biodegradation. The initial reaction
under anaerobic conditions involves hydrolysis of the ether bond (Kuder et al., 2005, Somsamak et al., 2005). Aerobic
metabolism of MTBE is initiated by mono-oxygenase enzymes which extract a proton from the methoxyl group. The
difference between the strength of the 13C-O bond and 12C-O bond is more pronounced than the difference between the
13C-H bond and the 12C-H bond (Huskey, 1991), resulting in greater fractionation of carbon during anaerobic metabo-
lism of MTBE.
5.2 Predicting Biodegradation from 513C in MTBE in Gasoline
Smallwood et al., (2001) reported that the normal range of 513C for MTBE in gasoline is from -28.3%o to -31.6%c; more
recent surveys indicate that the normal range extends between -27.5 %o and -33 %o(O'Sullivan et al., 2003). This is
the range of 513C that would be expected for MTBE in ground water in the absence of biodegradation. The variation in
513C during anaerobic biodegradation is much larger than the variation in 513C in MTBE from one sample of gasoline
to another (Figure 5.2).
Notice the straight-line relationship when values of 513C on an arithmetic scale are plotted against the fraction of MTBE
remaining on a logarithmic scale in Figures 5.2 and 5.3. The simplified version of the Rayleigh equation, originally
developed by Mariotti et al. (1981), is commonly used in the literature to relate the extent of biodegradation of MTBE
(and other organic compounds) to the 513C of the material remaining after biodegradation.
61JC
MTBE in ground water
Sl3^»
= o C
MTBE in gasoline
Equation 5.3
In Equation 5.3, s is the isotopic enrichment factor and is an expression of the extent of isotopic fractionation dur-
ing biodegradation, and F is the fraction of MTBE remaining after biodegradation. The value of F is simply C/Co in
Figure 5.2. The value of z is usually calculated as the slope of a linear regression of 513C on the natural logarithm of
F (Figure 5.3). Notice that the data in Figure 5.3 are plotted in an unconventional fashion. The natural logarithm of F
decreases toward the right-hand side of the x-axis.
-•- anaerobic biodegradation
•*- aerobic biodegradation
-1 -2 -3 -4 -5 -6
natural logarithm of fraction MTBE remaining
Figure 5.3 The data from Figure 5.2 have been plotted in units commonly used in the literature on stable isotopes.
34
-------
Hunkeler et al., (2001) determined the enrichment factor for four aerobic cultures that degraded MTBE as the sole carbon
source and one culture that co-metabolized MTBE when grown on 3-methylpentane. All the cultures were enriched
from Borden aquifer sediments from Ontario, Canada. The enrichment factors varied from -1.52 ± 0.06 to -1.97 ±
0.05. Gray et al., (2002) determined the enrichment factors during aerobic growth of pure culture PM-1 and a mixed
culture from Vandenberg AFB, California. The enrichment factors varied from -1.4 ± 0.1 to -2.4 ± 0.3.
Under anaerobic conditions, the enrichment factors are more negative. Kolhatkar et al., (2002) extracted an enrichment
factor of-8.1 ± 0.85 for anaerobic biodegradation along a flow path at a gasoline spill in New Jersey and -9.1 ± 5.0
for degradation in anaerobic microcosms constructed with core material from the site. More recent work indicates
that the isotopic enrichment factor in a mixed culture of the organisms from New Jersey is -12.5 ±1.4 (Kuder et al.,
2005). Somsamak et al., (2005) determined the enrichment factor for MTBE biodegradation by an enrichment culture
isolated from sediment from the Arthur Kill inlet between Staten Island, New York, and New Jersey. In a methanogenic
culture, the value of z was -15.6 ± 4.1. These experimental values are in good agreement with a theoretical value of
z (-12.2) that would be expected from the cleavage of a C-O bond in a molecule with five carbon atoms (Kuder et al.,
2005 following Huskey, 1991).
The values of z in the hypothetical data in table 5.2 and Figure 5.3 would be -11.6 for anaerobic biodegradation and
-2.3 for aerobic biodegradation. The values of z in the examples were based on typical values derived from laboratory
experiments and field studies.
5.3 Sources of Uncertainty in Estimates of Biodegradation
The relationship between the 513C of MTBE in a ground water sample and the true extent of biodegradation is often
complex. It is influenced by the starting 513C of the MTBE in the gasoline spill, and this value is rarely known with
certainty. There may be different releases of gasoline with different values of 513C contributing to the same spill. The
MTBE in a sample of gasoline floating in a monitoring well or in gasoline extracted from a core sample may already be
biologically weathered, and the 513C may not be representative of the 513C of the original release. Because a trustworthy
value of 513C for the MTBE in the original gasoline in the spill is rarely available, most evaluations compare the 513C
in the MTBE in the ground water to the published range of 513C in samples of gasoline that have not been released to
the environment.
The relationship between the value of 513C for MTBE and the extent of biodegradation is also influenced by the relative
contribution of aerobic and anaerobic biodegradation to the attenuation of MTBE. Because the value of z is less nega-
tive for aerobic biodegradation, there is less fractionation for a given amount of biodegradation. When 513C is used to
predict the fraction of MTBE remaining, the predicted extent of biodegradation is much greater for aerobic conditions
compared to anaerobic conditions. As an example, if the MTBE in the original gasoline spill had a value of 513C of
-27.5 %c, and the MTBE in the ground water was -10.0 %c, the fraction remaining calculated using Equation 5.3 under
aerobic conditions would be 0.0006, and the fraction remaining under anaerobic conditions would be 0.20.
The value of 513C is also influenced by site specific interactions between the ground water and the source of contamina-
tion. If the ground water is in contact with residual gasoline, unfractionated MTBE can dissolve from the gasoline into
ground water while biodegradation is in progress. Imagine ground water flowing through and under a region containing
residual gasoline. At the leading edge, MTBE dissolves into the ground water. As this MTBE moves with ground water,
it is biodegraded and fractionated. As the ground water flows through and underneath the area with residual gasoline,
additional MTBE from the center and down gradient edge of area with residual gasoline will dissolve into the ground
water. This MTBE will not be fractionated.
The fractionated MTBE remaining in ground water after biodegradation will be diluted by additional MTBE that has
not been fractionated. This dilution will shift the value of 513C to an extent that is directly proportional to the relative
concentrations of fractionated and unfractionated MTBE in the water sample. However, the shift in the value of 513C
will affect the estimate of biodegradation to an extent that is related to the natural logarithm of the fraction remaining
after biodegradation. The overall effect of dilution is to produce a value of 513C which will underestimate the true
extent of MTBE biodegradation.
A related interaction affects the estimate of biodegradation in a plume that is heterogeneous with depth across the well
screen of the monitoring well. If MTBE is almost entirely degraded at one depth interval but not at another, then the
MTBE produced from the well will be dominated by water from the interval where MTBE did not degrade as extensively.
Fractionated and unfractionated MTBE will be mixed in the monitoring well, and the value of 513C would underestimate
the true extent of biodegradation in the entire plume.
35
-------
If the TEA that is produced by biodegradation of MTBE is not further degraded, the next effect of these interactions
will be higher concentrations of TEA than would be expected from the extent of MTBE biodegradation as predicted
from the value of 513C in MTBE in the water sample.
5.4 A Conservative Estimate of the Extent of Biodegradation
These sources of uncertainty make it difficult to predict an exact value for the extent of biodegradation of MTBE in
ground water. However, if certain assumptions are made, it is possible to calculate an upper boundary on the fraction
of MTBE remaining from the 513C. In other words, for a certain value of 513C, we can be certain that the fraction of
MTBE remaining is this small or smaller.
This conservative boundary on the extent of biodegradation is presented as the lower line in Figure 5.4. The lower line in
the figure is calculated assuming a value of s of -12, which is the most negative value that is theoretically possible. The
value of 513C in the gasoline was assumed to be -27.5 %o, which is the highest value that has been reported for MTBE
in gasoline. Under these assumptions, a value of 513C of 0 %o would correspond to C/Co of 0.1 or 90% removal.
The upper line in Figure 5.4 is the boundary calculated at a value of 513C in the gasoline of -33 %o, which is the lowest
value that has been reported for gasoline. This line is added to show the effect of uncertainty in the value of 513C in
the gasoline originally spilled on the predicted extent of biodegradation. A value of 513C of 0 %o would correspond to
C/Co of 0.06 or 94% removal. At a value of 513C for MTBE in ground water of 0 %o, the published range of 513C in
gasoline would produce a two-fold range in the fraction of MTBE remaining that was predicted from Equation 5.3 or
Figure 5.4. The effect is of minor importance at values of 513C above 0 %o.
Anaerobic Biodegradation, Isotopic Enrichment Factor Ł=-12
0.001
o
O
O,
0)
•Ł 0.01
10
I
g
1
LL
LU
m
-40 -30 -20 -10
0 10 20
513C(°/00)
Figure 5.4 MTBE biodegradation under anaerobic conditions predicted from the &13C of MTBE in ground water. The
lower line is a conservative prediction of the extent of biodegradation of MTBE.
36
-------
Section 6 -
Applications of Stable Carbon Istotope Ratios -
As discussed in Section 5, the stable carbon isotope ratio in MTBE can be used to predict the extent of biodegradation
of MTBE in ground water. This section illustrates three practical applications that regulators are likely to see in reports.
In the first application, the analysis of stable carbon isotope ratios is used as an alternative to a microcosm study to
document that anaerobic biodegradation was occurring at a field site. In the second application, the analyses are used
along with data on the hydraulic properties of the aquifer to estimate a first order rate constant for anaerobic biodegra-
dation. In the third application, the analyses are used to show that TEA in ground water came from the biodegradation
of MTBE, and not from TEA that was originally present in gasoline.
6.1 Applications of Stable Isotope Ratios to Interpret Plume Behavior
The U.S. EPA recognizes a three-tiered approach to evaluate site specific data in support of monitored natural attenua-
tion (U.S. EPA, 1999), specifically: (1) historical groundwater and/or soil chemistry data that demonstrate a clear and
meaningful trend of decreasing contaminant mass and/or concentration over time at appropriate monitoring or sam-
pling points, (2) hydrogeologic and geochemical data that can be used to demonstrate indirectly the type(s) of natural
attenuation processes active at the site, and the rate at which such processes will reduce contaminant concentrations
to required levels, 3) data from field or microcosm studies (conducted in or with actual contaminated site media) which
directly demonstrate the occurrence of a particular natural attenuation process at the site and its ability to degrade the
contaminants of concern. Unless EPA or the overseeing regulatory authority determines that historical data (Number 1
above) are of sufficient quality and duration to support a decision to use MNA, data characterizing the nature and rates
of natural attenuation processes at the site (Number 2 above) should be provided. Where the latter are also inadequate
or inconclusive, data from microcosm studies (Number 3 above) may also be necessary.
As a practical matter, it is difficult to provide the second line of reasoning for MTBE in ground water. For any con-
taminant, it is challenging to obtain the convincing hydrogeologic and biogeochemical data that demonstrate the type
of processes and the rate at which the processes operate. The second line of reasoning has been provided in a few
instances by correcting the apparent attenuation of the contaminant along a flow path by the attenuation of a tracer.
Wiedemeier et al., (1996) were able to use trimethylbenzenes in a plume of ground water contaminated by JP-4 jet
fuel as a conservative tracer to correct the apparent attenuation of BTEX compounds along the flow path in the plume.
The trimethylbenzenes and BTEX compounds occurred together in the fuel spill in the same relative proportions. Any
reduction in concentrations of BTEX compounds in excess of the dilution of the tracer was assumed to be biological
degradation. Varadhan et al., (1998) were able to use chloride in a plume of landfill leachate to correct the apparent
attenuation of benzene, 1,1-dichloroethane, and 1,2-dichloroethane. Chloride was abundant in the leachate and not in
the ambient ground water, and the concentration of chloride was orders of magnitude higher than the concentration of
organic chlorine in 1,1-dichloroethane, and of 1,2-dichloroethane.
Unfortunately, at gasoline spill sites, a good tracer for MTBE is usually not available. The biogeochemical parameters,
such as depletion of oxygen or production of methane, are associated with gasoline components in general and not
MTBE in particular. Any trimethylbenzenes in ground water may have been part of an earlier spill of gasoline that did
not contain MTBE. Chloride is not an important component of unleaded gasoline.
Although microcosm studies (the third line of reasoning) are easy to interpret, they are expensive and tend to be time-
consuming. Microcosm studies often take several months to over a year to complete, and frequently the results are
equivocal. Microcosm studies can only show that the aquifer harbors microorganisms that are capable of degrading
the contaminants under the conditions that pertained at the time the aquifer material was sampled. They do not provide
direct evidence that the contaminant in the aquifer was actually biologically degraded.
As a consequence, most evaluations of MNA rely heavily on the first line of reasoning, using the long term monitoring
data. The rate of decrease in concentration over time in monitoring wells reflects the rate of attenuation of the source
of contamination, not the rate of transformation of contaminants along the flow path in ground water (Newell et al.,
37
-------
2002; Wilson, 2003a; Wilson and Kolhatkar, 2002). Most studies in support of MNA do not provide a solid estimate of
the rate of degradation of the contaminant in ground water along the flow path. Without a solid estimate of the rate of
degradation in the ground water, a conservative evaluation of the risk to a receptor is restricted to the assumption that
the contaminant does not degrade at all.
One conventional approach to evaluate biodegradation of organic contaminants in ground water is to demonstrate an
increase in the concentration of transformation products. This approach is problematic for MTBE from gasoline spills
because the primary transformation product (TEA) can also be a component of gasoline (compare Landmeyer et al.,
1997). Kramer and Douthit (2000) extracted gasoline from six service stations in New Jersey using a fuel-to-water
ratio of one to four. TEA was detected in extracts of the gasoline from five of the six stations. The concentration in the
water extracts varied from 1,120 to 1,690 mg/1. These are concentrations that would be expected if the MTBE added
to the gasoline contained 11% by volume TEA (equivalent to 1.5% by volume in the gasoline). Using TEA to estimate
MTBE biodegradation is further complicated by the possibility that TEA may be biologically degraded in ground water
under both aerobic and anaerobic conditions.
The Committee on Intrinsic Remediation of the National Academy of Sciences (NRC, 2000) determined that biologi-
cal transformation was the dominant process responsible for attenuation of MTBE in ground water. Because of these
uncertainties, they further determined that the current level of understanding of biological transformation of MTBE
is moderate, and as a result, the likelihood of success for using monitored natural attenuation as a remedy for MTBE
contamination at a particular site is low (NRC, 2000, page 8).
Recent work shows a strong discrimination during anaerobic biodegradation between molecules of MTBE containing the
stable isotope 12C and molecules containing the stable isotope 13C (Kolhatkar et al., 2002; Kuder et al., 2005; Somsamak
et al., 2005, see discussion in Section 5). The 12C isotope is preferred, and the molecules containing the 13C isotope
accumulate in the residual MTBE. The stable isotopes are said to be fractionated during biodegradation.
6.2 Using Stable Carbon Isotope Ratios to Recognize Natural Biodegradation.
As discussed in Section 5, the extent of degradation of MTBE can be estimated from the change in the ratio of 12C to
13C in MTBE in the plume compared to the ratio in the MTBE in the gasoline that was originally spilled. The stable
isotope approach provides direct information on the extent to which MTBE has been biologically degraded in the
ground water. Compound-specific stable isotope analysis provides a useful extension to the conventional practice for
interpreting the behavior of MTBE plumes.
The approach will be illustrated with data from a plume of MTBE from a gasoline spill site in Dana Point, California
(Figure 6.1). The plume of MTBE is contained in a layer of silty fine sand and clean sands that lies beneath a layer of
clay and silt. The water table is in the layer of clay and silt. Based on aquifer testing, the average hydraulic conductivity
of the layer of silty fine sand and clean sand is 11 meters per day.
The direction of ground water flow at the site was estimated by using linear regression to fit a plane to the elevation of
the water in 14 monitoring wells at the site. The regression was fit using the Optimal Well Locator (OWL) application
(Srinivasan et al., 2004). Data were available for 14 rounds of sampling between April 1999 and July 2002. For six of the
regressions, the value of r2 was low (0.17 or less), and data from these dates were ignored. For the remaining eight dates,
the value of r2 for the regression ranged from 0.62 to 0.73. For these eight dates, the direction of ground water flow is
presented by the "flow rose" in Figure 6.1. The length of each arrow was calculated by multiplying the hydraulic gradient
by the average hydraulic conductivity (11 meters per day), then dividing by an estimate of porosity (0.25).
There were two sources of MTBE contamination in the ground water at this site. The major source was associated with
leaking under ground storage tanks (Figure 6.1). The tanks and the surrounding fill material were excavated. Residual
gasoline in the aquifer acts as a continuing source of MTBE in ground water. The highest concentrations of MTBE are
immediately down gradient of the underground storage tanks. A second source is associated with the distribution lines
to the eastern dispenser island. In the wells that are side gradient and far down gradient of the underground storage
tanks and the dispenser island, the concentrations of MTBE are lower.
To estimate the fraction of MTBE remaining after biodegradation from the 513C of MTBE in water from the wells,
Equation 5.3 was solved for the fraction remaining to produce Equation 6.1.
/ (Ys13r S13r \/A Equation 6.1
j-j f I f \\ 'MTBE in ground water ° cMTBE in gasoline j /*=•] n
38
-------
MW-11
318
10 meters
TPHg>1,OOOmg/kg
TPHg>100mg/kg
•
MW-12
3.6
MW-9 Underground
Storage Tanks
<0 5 Dispenser
Islands
Figure 6.1 Concentration ofMTBE (fig/L) in selected monitoring wells at a gasoline spill site in Dana Point, Califor-
nia, in August 2004. The cluster of arrows is a "flow rose" indicating the direction and distance ground
water would move in one year based on the elevation of the water table in monitoring wells on particular
sampling dates. TPHg is total petroleum hydrocarbons within the range of molecular weights expected
for gasoline.
Table 6.1 compares the concentrations of MTBE and TEA in the monitoring wells to the fraction of MTBE remaining as
predicted from the 513C of MTBE using Equation 6.1. A conservative value for s of -12 was used in the calculation.
Table 6.1 A Comparison Between the Distribution of MTBE and TEA in Ground Water Contaminated by a Fuel
Spill in Dana Point, California, and the Extent ofMTBE Biodegradation Predicted from the Stable Car-
bon Isotope Ratio (513C) of the Residual MTBE
Well
MW-14
MW-3
MW-8
MW-6
MW-7
MW-11
Date
5/20/03
8/18/04
5/20/03
8/18/04
5/20/03
8/18/04
5/20/03
8/18/04
8/18/04
5/20/03
8/18/04
TEA
Measured
(M8/L)
13,000
107,000
20,000
32,000
10,000
32,000
3,600
19,200
1,220
<10
135
MTBE
Measured
(M8/Q
11,000
26,000
870
164
19
25
47
490
106
1
318
813C MTBE
(%«)
-23.88
-21.58
6.84
8.53
18.11
37.99
9.83
-1.58
-27.33
-31.5*
-28.92
MTBE
Fraction Remaining
(C/Co)
0.75
0.62
0.058
0.050
0.023
0.0043
0.045
0.116
0.994
1.41
1.14
mate o
39
-------
Biodegradation makes the value of 513C larger. The highest value of 513C that has been measured for MTBE in gasoline
is -27.4 %0. This conservative value was used for 513CMTBE fa gasoline to calculate the fraction of MTBE remaining of the
MTBE originally spilled in the aquifer.
The most contaminated well at the site (MW-14 in Figure 6.1) is located in an area that had 9,000 mg/kg of gasoline
range Total Petroleum Hydrocarbons (TPHg). When sampled in May 2003, MTBE in water from MW-14 had a low
value of 513C. The concentrations of MTBE and TEA were essentially equivalent, and there was little evidence of
biodegradation (Table 6.1). When sampled again in August 2004, the value of 513C was slightly higher, and the concen-
tration of TEA was now four fold higher than the concentration of MTBE. Biodegradation was beginning to influence
the distribution of MTBE and TEA in well MW-14.
Wells MW-3 and MW-8 are further down gradient of the source of MTBE that was associated with the underground
storage tanks. The sum of the concentrations of MTBE and TEA in wells MW-3 and MW-8 are roughly equivalent to
the sum of MTBE and TEA in well MW-14; however, the concentration of MTBE is much lower than the concentration
of TEA in wells MW-3 and MW-8, indicating that MTBE may have been degraded to TEA. The 513C of MTBE in wells
MW-3 and MW-8 is much heavier than MTBE in gasoline (Table 6.1). The calculated fraction of MTBE remaining
corresponds to 94% to 99.6% biodegradation of MTBE. The attenuation in concentration of MTBE in wells MW-3
and MW-8 compared to well MW-14 can safely be attributed to biodegradation.
Well MW-6 appears to be side gradient to the source of MTBE associated with the underground storage tanks (Figure
6.1). However, well MW-6 is directly down gradient of the secondary source associated with the dispenser islands. The
behavior of MTBE in well MW-6 is very similar to wells MW-3 and MW-8. Concentrations of MTBE are low, and
concentrations of TEA are high. The 513C of MTBE is high compared to MTBE in gasoline, and the predicted fraction
remaining corresponds to 88% to 96% biodegradation of MTBE.
Wells MW-7 and MW-11 are even further down gradient of the source of MTBE. The concentrations of MTBE are
low, and it would be tempting to attribute the low concentrations to biodegradation. However, the 513C of MTBE in
these wells is even lower than the 513C in MW-14, the most contaminated well. As discussed in Section 5, the expected
range of 513C for MTBE in gasoline is -27.5%0 and -33%0(O'Sullivan et al., 2003). In fact, the 513C of MTBE in these
wells falls near or within the range of 513C expected for gasoline. There is no evidence from the 513C of MTBE that
biodegradation contributed to attenuation of MTBE in these wells.
6.3 Using Stable Carbon Isotope Ratios to Estimate the Projected Rate of Natural Biodegradation.
Because the 513C of MTBE in ground water provides a direct estimate of the fraction of MTBE remaining after bio-
degradation, it can be used to extract an estimate of the rate of natural biodegradation of MTBE along the flow path.
Earlier approaches to extract rate constants from field data used conservative tracers to correct for dilution (Wiedemeier
et al., 1996; Varadhan et al., 1998) or made an estimate of the attenuation due to dilution from dispersion (Buscheck
and Alcantar, 1995). Because the 513C of MTBE in a sample of ground water is not changed by dilution, the ratio C/Co
estimated from the 513C of MTBE is not changed by dilution, and there is no need to correct for dilution in the estimate
of biodegradation. The rate constant can be calculated directly from the fraction of MTBE remaining as estimated from
the 513C of MTBE, the distance between wells, and an estimate of the interstitial seepage velocity. The projected rates
of biodegradation are not equivalent to overall rates of natural attenuation because they do not include the contribution
from dilution and dispersion, or sorption.
The projected rate of biodegradation can be expressed directly as a first order rate of removal with distance, or the rate
of removal with distance can be multiplied by an estimate of the seepage velocity of ground water to calculate a rate
of removal with time of travel (Newell et al., 2002). The projected rate of biodegradation with distance is calculated
following Equation 6.2. Biodegradation with time follows Equation 6.3.
kwithdistance = ~ m (F)/d Equation 6.2
kwithttme = - In (F) * V/d Equation 6.3
In Equations 6.2 and 6.3, k is the projected rate of natural biodegradation, F is the fraction of MTBE remaining as
estimated from 513Cfleld using Equation 6.1, and d is the distance along the flow path between the up gradient well and
the down gradient well, and v is the ground water seepage velocity.
The average hydraulic conductivity at the site in Dana Point, California, is 11 meters per day. The average hydraulic
gradient over eight rounds of sampling was 0.0023 meter per meter. Assuming the effective porosity is 0.25, the aver-
40
-------
Table 6.2 - Rates of Natural Biodegradation of MTBE Projected along a Flow Path in Ground Water to Monitoring
Wells. Projected Rates Were Calculated from the Estimated Seepage Velocity of Ground Water and the
Fraction of MTBE Remaining After Biodegradation. Projected Rates of Biodegradation are not Equiva-
lent to Overall Rates of Natural Attenuation, Because They do not Include Contributions from Dilution
and Dispersion, or Sorption -
Well
MW-3
MW-3
MW-8
MW-8
MW-7
MW11
MW-6
MW-6
Date Sampled
May, 2003
August, 2004
May, 2003
August, 2004
August, 2004
August, 2004
May, 2003
August, 2004
Fraction
MTBE
Remaining
(C/C0)
0.058
0.050
0.023
0.0043
0.994
1.0
0.045
0.116
Distance from
MW-14
(meters)
9.6
9.6
11.7
11.7
23.0
44.1
Distance from
Dispenser Island
(meters)
31.1
31.1
Projected Rate of
Biodegradation with
Distance
(per meter)
0.30
0.31
0.32
0.46
0.00025
0
0.10
0.069
Projected Rate of
Biodegradation with Time
(per year)
10.9
11.5
11.9
17.1
0.0093
0
3.7
2.6
age ground water seepage velocity should be near 37 meters per year. The projected rates of biodegradation of MTBE
along flow paths between the most contaminated well (MW-14), and down gradient wells MW-3, MW-7, MW-8, and
MW-11 are presented in Table 6.2.
In wells MW-3 and MW-8, the projected first order rate of biodegradation is rapid, on the order of 0.3 per meter of
travel, or 10 per year of residence time. In well MW-7, the projected rate of biodegradation was one thousand fold
slower, and in well MW-11 biodegradation was not detected at all.
Projected rates along the flow path are most useful to predict the possible extent of plumes, as will be discussed in the
following paragraph. The projected rate of biodegradation with time is more convenient to compare the behavior of
the plume to other plumes, or to rates published in the literature. The rate of anaerobic biodegradation of MTBE in a
microcosm study constructed with material from a gasoline spill in Parsippany, New Jersey, varied from 11 ± 2.3 per
year to 12 ± 2.9 per year (Wilson et al., 2005a). The rate of anaerobic MTBE biodegradation in a microcosms study
constructed with core material from a JP-4 jet fuel spill in Elizabeth City, North Carolina, was 3.02 ± 0.52 per year and
3.5 ± 0.65 per year (Wilson et al., 2000). These laboratory rates are in reasonable agreement with the rates projected
by Equation 6.3 for the flow path to wells MW-3, MW-6 and MW-8 at the Dana Point, California, site (Table 6.2).
If the aquifer carrying the plume is heterogeneous (and most are), it is best to use the highest value for the hydraulic
conductivity measured at the site to estimate the seepage velocity used to calculate the projected first order rate with
respect to time. It is most likely that the plume is spreading the fastest through the most conductive material. Using
the highest value for hydraulic conductivity will provide a conservative estimate of the projected rate constant.
6.4 Using the Projected Rate of Biodegradation to Estimate the Length of Plumes
The distance traveled before the concentration reaches a particular goal (d goal) can be calculated by rearranging Equation
6.2 to produce Equation 6.4, where F is the ratio of the goal to the existing concentration in the monitoring well.
Equation 6.4
ith distance
41
-------
If the maximum concentration of MTBE in monitoring wells MW-3 and MW-8 is 1,000 |ag/L (compare Table 6.1 for
real monitoring data), and the goal for MTBE is the U.S. EPA advisory limit of 20 |ag/L, and the projected first order
rate of biodegradation with distance is 0.3 per meter; then the plume would be expected to move only 13 meters further
before it reaches the goal.
In well MW-7, the projected first order rate of biodegradation is much slower. At a rate of 0.00025 per meter, start-
ing at a concentration of 106 |ag/L, the MTBE plume would be expected to move 6,700 meters further down gradient
before it reaches the advisory limit. In well MW-11, biodegradation of MTBE could not be established based on the
513C for MTBE in the ground water. The only processes that can be reasonably expected to attenuate MTBE further
down gradient of wells MW-7 and MW-11 are dilution and dispersion.
This pattern has been seen by the authors in three other MTBE plumes. The biodegradation of the MTBE in the core
of the plume was rapid and extensive, but MTBE in the periphery of the plume was not degraded. As a consequence,
the extent of the plume was underestimated when a single rate constant for biodegradation was applied to the maximum
concentration of MTBE in the source area. On the other hand, the maximum extent of the plume was seriously overes-
timated if biodegradation was ignored. At this point in the evolution of risk evaluation, a conservative course of action
is to recognize that plumes are heterogeneous. An independent estimate of the extent of MTBE contamination further
down gradient should be made for each well used in the risk evaluation, based on the concentration of MTBE in each
well, and the projected rate of biodegradation in the flow path leading to each well. For many flow paths, the extent of
MTBE contamination will be determined by dilution and dispersion, not by anaerobic biodegradation.
6.5 Using 513C to Distinguish the Source of TEA in Ground Water
There are two plausible sources of TEA in ground water (compare Landmeyer et al., 1997). Commercial MTBE may
have contained as much as 1% to 10% TEA in the past (Kramer and Douthit, 2000). One process for the chemical
synthesis of MTBE produces MTBE by reacting isobutylene with methanol. Any water that is present in the methanol
feed stock will react with isobutylene to produce TEA, which is carried over into the commercial grade MTBE. In addition,
biodegradation of MTBE may produce TEA as a transformation product.
Regulators often need to know the source of TEA in ground water. Tank owners may be reluctant to accept responsi-
bility for a plume that contains high concentrations of TEA if there were low concentrations of TEA in their gasoline.
Pump-and-treat can effectively remove TEA from ground water in the source area at a gasoline spill. However, if
MTBE continues to partition to ground water from residual gasoline in the aquifer, and the MTBE is degraded to TEA,
the concentrations of TEA can rebound. Unfortunately, it is usually impossible to identify the source of TEA present
in ground water using conventional chemical analyses.
Figure 6.2 depicts the location of monitoring wells at a gasoline spill site in Delaware that had high concentrations of
TEA in the ground water. Well MW-1 is near and slightly up gradient of the underground storage tanks. Well MW-
2 is side gradient of the underground storage tanks and down gradient of the dispenser islands. Well MW-3 is down
gradient of the underground storage tanks, and MW-4 is further down gradient of MW-3.
Table 6.3 presents the concentrations of contaminants and biogeochemical parameters. A background well (not shown
on Figure 6.2) was devoid of contaminants and had low concentrations of methane and moderate concentrations of
oxygen and sulfate. The two wells closest to the underground storage tanks (MW-1 and MW-3) had high concentra-
tions of benzene, MTBE, and TEA. The ratio of TEA to MTBE was high. The concentrations of methane were high,
and the concentrations of sulfate were low in both wells, and oxygen was depleted in one of the wells. The chemistry
of the ground water in these two wells indicates that the water was anaerobic and that natural biodegradation would
proceed through the anaerobic pathway.
42
-------
MW-1
90.5
UST
Dispenser Island
•MW-2
10 meters
Figure 6.2 Location of monitoring wells and water table elevations at a gasoline spill site in Newark, Delaware, with
high concentrations of TEA in the ground water.
Table 6.3 Relationship Between the Extent of Contamination and Biogeochemical Parameters at a Site in Newark,
Delaware
Location
Background
MW-1
MW-2
MW-3
MW-4
Benzene
MTBE
TEA
(M8/Q
<0.5
1,300
<0.2
1,440
<0.5
<1.0
475
45.2
18,000
15.9
<10
245,000
49.4
306,000
406
Oxygen
Sulfate
Methane
(mg/L)
2.7
0.5
0.2
1.6
3.2
10.3
<0.1
4.4
0.4
24.4
0.003
5.4
7.1
2.7
0.06
Water from well MW-2 is also decidedly anaerobic, the concentrations of the contaminants are much lower, and the
concentration of TEA is almost equal to the concentration of MTBE. Water from well MW-4 is aerobic, the concentra-
tion of contaminants is low, but the concentration of TEA is higher than the concentration of MTBE.
Table 6.4 compares the 513C of MTBE in the four wells. The 513C of MTBE in well MW-2 is very near the center of
the range of values expected for gasoline. There is no evidence of MTBE biodegradation at this location. The value
of 513C in MTBE in the other three wells is higher than of 513C in MTBE in well MW-2.
Table 6.4 Concentrations of TEA Predicted from Biodegradation of MTBE to TEA at a Site in Newark, Delaware
Location
MW-1
MW-2
MW-3
MW-4
MTBE
Measured
Qjg/L)
475
45.2
18,000
15.9
813CofMTBE
(%«)
43.05
-30.17
7.94
10.22
MTBE
C/Co
Figure 5.4
0.002
1
0.05
0.04
MTBE
C/Co
Equation 6.1
0.0021
1.0
0.042
0.035
Prediction of
TEA Produced
Qjg/L)
190,000
0
350,000
370
TEA
Measured
Qig/L)
245,000
49.4
306,000
406
43
-------
The value of 513C for MTBE was used to predict the fraction of MTBE remaining in two ways. The estimate of C/Co
was estimated graphically from the lower line of Figure 5.4. The fraction remaining (C/Co or F) was also calculated
from Equation 6.1.
The value of 513C in MTBE measured in MW-2 was assumed to be the value of 513C in the gasoline that was originally
spilled. Because natural biodegradation of MTBE occurred under anaerobic conditions, the value of z was assumed
to be -12. Results of the calculations are presented in Table 6.4. In the three wells with high values of 513C, 95% to
99.8% of the MTBE was degraded to TEA.
The amount of TEA that was expected from the biodegradation of MTBE was calculated following Equation 6.5.
CTBAproduced =(CMTBE/F)(l-F)(74/88) Equation 6.5
The measured concentration of MTBE is the concentration after biodegradation. The concentration of MTBE before
biodegradation is Co in the expression C/Co. The concentration of MTBE before biodegradation was calculated by
dividing the measured concentration by the fraction remaining (C/[C/Co] =Co). The fraction of the original concentra-
tion of MTBE before biodegradation that was transformed to TEA is one minus the fraction of MTBE remaining after
biodegradation or (1-F). The concentration of MTBE that was transformed to TEA was calculated by multiplying the
concentration of MTBE before biodegradation by the fraction of MTBE transformed to TEA. One molecule of MTBE
produced one molecule of TEA. The concentration of TEA that was produced was calculated by multiplying the
concentration of MTBE that was transformed to TEA by the molecular weight of TEA and dividing by the molecular
weight of MTBE ([TEA] = [MTBE] • (74/88)).
Table 6.4 compares the estimated concentration of TEA produced from biodegradation of MTBE to the actual con-
centration of TEA. The good agreement between the measured concentration of TEA and the expected concentration
of TEA indicates that biodegradation of MTBE was responsible for the major portion of the TEA that was present in
wells MW-1 and MW-3. Although the concentration of TEA in well MW-4 is lower than the concentrations in MW-1
and MW-3, the TEA that was present in well MW-4 was also produced by biodegradation of MTBE.
The distribution of MTBE at this site is counter intuitive. The MTBE in well MW-2 has not been degraded, but the
concentration of MTBE is very low. More than 99% of MTBE in well MW-3 has been degraded, but MW-3 has the
highest concentration of MTBE in any well at the site. This disparity can be attributed to heterogeneity in anaerobic
biodegradation. The concentrations of MTBE at MW-2 may have been too low to allow acclimation of an MTBE de-
grading microbial community. As was discussed earlier for the site in Dana Point, California, the core of the MTBE
plume participated in anaerobic biodegradation, while flow paths at the dilute margins of the plume did not.
6.6 Caveats and Limitations Concerning the Use of 513C of MTBE to Estimate Biodegradation
As discussed in Section 5, a variety of processes operates at field scale to confound the simple relationship in Equa-
tion 6.1 between the 513C of MTBE in ground water and the extent of biodegradation of MTBE. Some portion of the
MTBE may be degraded through an aerobic pathway. Because the value of z for aerobic metabolism is on the order
of -2.5, compared to -12 for anaerobic pathway, a given amount of biodegradation produces a smaller shift in the value
of513CofMTBE.
If the MTBE is degraded in ground water that is in proximity to residual gasoline, then fresh MTBE can partition from
the gasoline to ground water and dilute the fractionated MTBE with MTBE that has not been fractionated. This effect
will be most important near the source areas of plumes, particularly in wells that contained free product at some time
in the past.
As illustrated with the data from the site at Dana Point, California, biodegradation in MTBE plumes can be heteroge-
neous. If one portion of a plume has degraded and a second portion has not, and the two portions are mixed when the
water is sampled from a monitoring well, the MTBE in the water from the well will be dominated by MTBE from the
portion that did not degrade. The blended value of 513C of MTBE in the well water will not accurately reflect the extent
of biodegradation of all the MTBE originally present in the ground water.
All of these processes act to underestimate the extent of biodegradation. As a consequence, the fraction remaining
calculated by Equation 6.1 is a conservative upper boundary on the fraction remaining. The fraction remaining may be
much lower, and the extent of biodegradation of MTBE to TEA may be much larger.
To illustrate this point, Kuder et al., (2004) estimated the fraction of MTBE degraded from the measured concentra-
tions of MTBE and TEA, and then compared the fraction remaining to the 513C of MTBE in the ground water. Their
44
-------
results are presented in Figure 6.3. If one molecule of MTBE is degraded to one molecule of TEA, and the TEA is not
further degraded, the sum of the molar concentrations of MTBE and TEA after biodegradation should equal the molar
concentration of MTBE before biodegradation. The fraction remaining is simply the molar concentration of MTBE
after biodegradation divided by the sum of the molar concentrations of MTBE and TEA after biodegradation.
The data presented in Figure 6.3 were collected from 99 wells at 19 sites. The two solid lines are the relationship that
would be expected between the fraction of MTBE remaining and the 513C of MTBE in ground water if the value of e
is -12, and the value of 513C before biodegradation is -33%o and -27.5%o. Almost all the estimates of the fraction of
MTBE remaining based on accumulation of TEA are smaller than the fraction remaining estimated from the 513C of
MTBE (above the lines in Figure 6.3). The 513C underestimated the extent of biodegradation.
The absence of evidence for a process is not evidence for the absence of a process. The approach outlined in Section
5 and illustrated in this section contains a number of conservative assumptions. These include the assumption that all
biodegradation of MTBE goes through the anaerobic pathway, and that the 513C of MTBE in gasoline that was spilled
at any site is as heavy as the heaviest value of 513C that has been measured for MTBE in gasoline at any time anywhere
in the world. Equation 6.1 may fail to detect natural biodegradation of MTBE when it is really occurring. The situa-
tion is directly analogous to a "not detected" in analytical chemistry. A "not detected" does not mean the analyte was
not present. If the stable carbon isotope data fail to predict natural biodegradation of MTBE, they should not be further
interpreted. In particular, the stable isotope data should not be interpreted to prove that natural biodegradation is not
occurring.
0.00001 T
-40
-20
0
513C in MTBE
20
40
Figure 6.3 Distribution of513C of MTBE in ground water and the fraction of MTBE remaining from biodegradation
as calculated from the concentrations of MTBE and TBA in ground water and the assumption that TBA
was produced by biodegradation of MTBE (from Kuder et al, 2005). The solid lines bound the fraction
remaining as calculated from the 5"C of MTBE in ground water.
45
-------
Preceeding Page Blank
Section 7 -
Statistical Evaluation of Rates of Attenuation of Sources -
7.1 Risk Management and U.S. EPA Expectations for MNA
The expectations of the U.S. EPA for natural attenuation in ground water are specified in the OSWER Directive 9200.4-
17P (U.S. EPA, 1999). The Directive identifies the first line of evidence for MNA as historical groundwater and/or soil
chemistry data that demonstrate a clear and meaningful trend of decreasing contaminant mass and/or concentration
over time at appropriate monitoring or sampling points. Further, the OSWER Directive 9200.4-17P (U.S. EPA, 1999)
notes that EPA expects that MNA will be an appropriate remediation method only where ... it will be capable of achiev-
ing site-specific remediation objectives within a timeframe that is reasonable compared to other alternatives.
Most states agencies choose to manage gasoline spill sites with a combination of risk management, active clean up,
and monitored natural attenuation. The state agencies start with a risk evaluation. If a supply of drinking water is at
risk, most state agencies will require active efforts to control the source of MTBE contamination. The risk is reduced
through a variety of techniques to remove the gasoline (e.g. free product recovery, excavation of the residual gasoline,
and surfactant flushing) and a variety of techniques to treat the gasoline in situ (e.g. air sparging and vacuum extrac-
tion, or in situ bioremediation, or electrical heating). Most state agencies focus their effort on source reduction versus
remediation of the plume in ground water.
Most state agencies monitor concentrations of MTBE and other fuel components at gasoline spill sites on a fixed sched-
ule. This monitoring has two purposes. It provides documentation that the concentrations in the plume are actually
declining over time. It also monitors the plume for a radical change in its behavior that would require a new evaluation
of risk. There may be a new release of gasoline at the site. The direction of ground water flow from the spill may have
changed due to changes in pumping of ground water from the aquifer or development of land in the recharge zone of
the aquifer.
If the trend in contaminant concentrations is down, and the state agency is satisfied that the risk of exposure is properly
managed at a site, they may not require an active remedial technology for the site. A clear and meaningful trend can
be documented with conventional parametric statistics such as the slope of a regression line or non-parametric statistics
such as the Mann-Kendall test.
If the goal is simply to establish that the concentrations are declining over time, either the parametric or non-parametric
statistics are useful and appropriate. If the goal is to determine how rapidly the concentrations are declining, or to project
how soon the concentration will reach a particular goal, then it is necessary to use parametric statistics.
Most practitioners assume a first order rate law to describe the rate of attenuation in concentration over time. If the rate
of attenuation of MTBE in ground water in a monitoring well is controlled by the rate of physical weathering of MTBE
from residual gasoline in the source area of a plume, the rate of weathering is constrained by mass transfer limitations
on dissolution of the MTBE from gasoline such as diffusion from regions of low hydraulic conductivity to regions of
higher conductivity. As a consequence, the instantaneous rate of weathering should be proportional to the amount of
MTBE in the residual gasoline, and the rate of attenuation over time should be a first order process.
If the time allowed to reach a specific cleanup level for MTBE has been determined, and if attenuation follows a first
order law, the rate of attenuation necessary to meet the goal (k^,^^) can be calculated from Equation (7.1), where Cg
is the cleanup goal, C0 is the current concentration, and t is the time allowed for meeting the goal.
k = - In (Cg/ 1 It Equation (7.1)
^•necessary m I /Ł \ll
In Equation (7.1), k^^^is defined as a rate of attenuation. It has a positive value when concentrations are decreasing
over time.
If risk is being managed at the site, but the time to reach a cleanup goal has not been determined, then Cg is any value
less than C0, and k necessaryis any value greater than zero.
47
-------
The achieved rate of attenuation over a time interval being evaluated (kadlieved) will be defined as the best estimate of the
rate of attenuation that is extracted by statistical analysis of the monitoring data. The achieved rate is best determined as
the slope of a regression of the natural logarithm of the concentration on time. In the following material in this section,
a spreadsheet will be used to extract kadlieved from monitoring data. The spreadsheet will also be used to extract statisti-
cal confidence intervals on kadlieved. The slower confidence interval (k^-confidence) can be compared to zero to determine
if the achieved rate of attenuation is statistically significant. A spreadsheet will also be used to extract the slowest rate
of attenuation (kdetect) that is statistically different from zero at the predetermined level of confidence.
As will be discussed later, the slower confidence interval is not the "lower" confidence interval identified in the spread-
sheet. The spreadsheet calculates a rate of change, not a rate of attenuation. If concentrations are attenuating, the rate
of change is negative. If concentrations are attenuating more rapidly, the rate constant is more negative. A "lower"
confidence interval as identified by the spreadsheet will be the most negative confidence interval. The "lower" con-
fidence interval will actually be the faster confidence interval, and the "higher" confidence interval will be the slower
confidence interval.
The spreadsheet uses the t statistic to calculate confidence intervals on the rate of attenuation. Use of the t statistic
requires an assumption that the variance of the data is independent of the values of the data. Variance is a statistical
definition of the variation in sample data about a calculated summary statistic. It is a numerical measure of the scatter
in the data.
The concentrations of contaminants in ground water generally do not meet the assumption that the variance is indepen-
dent of the values of the data. The variance in ground water data tends to be proportional to the concentration. Large
concentrations have high variance, and small concentrations have a smaller variance. However, the variance of the
logarithms of the concentrations is much less dependent on the concentration. To adjust the variance between the higher
concentrations in the early samples and the lower concentrations in the later samples, the statistical comparisons will
be made between the natural logarithms of the concentrations. This relationship is illustrated in Figure 7.1.
Jan-92 Jan-94 Jan-96 Jan-98 Jan-00 Jan-02
Date of Sampling
1000
100
10
Jan-92 Jan-94 Jan-96 Jan-98 Jan-00 Jan-02
Date of Sampling
Figure 7.1 The variance in monitoring data is often proportional to the concentration. The upper panel plots con-
centration on an arithmetic scale, while the lower panel plots concentration on a logarithmic scale. The
variance of the logarithm of the concentration is less dependent on the concentration.
48
-------
Table 7.1 compares the relationship between the rates of attenuation necessary to achieve a goal for cleanup, and the
achieved rates of attenuation in particular wells. The process of extracting the rate constants from monitoring data
will be illustrated with monitoring data from a gasoline spill site in Parsippany, New Jersey. Natural attenuation of
MTBE in ground water in this plume is dominated by natural anaerobic biodegradation (Kolhatkar et al., 2002). The
locations of the monitoring wells are depicted in Figure 7.2. A portion of the long-term monitoring record is presented
in Table 7.2.
Table 7.1 Relationship Between the Rate of Attenuation Necessary for Risk Management or for Monitored Natural
Attenuation, and the Achieved Rates of Attenuation During Long-term Monitoring
Remediation Objective a Shrinking
Plume
*Svith-confldence ^* "
kadiieved > 0 > J^vith-confldence
Remediation Objective a Cleanup Goal.
^Svith-confidence ^ ^Siecessary
kadiieved > ^necessary DUt
k <• k
with-confidence Tiecessary
kadiieved > ^detect DUt
k <• k
achieved Tiecesary
k <• k
Tiecessary detect
kadiieved < kdetect
Interpretation
At the predetermined level of confidence, the concentrations are at-
tenuating over time.
Concentrations might be attenuating over time, but there is no statistical
confidence that concentrations are attenuating.
At the predetermined level of confidence, the concentration goal should
be achieved in the specified time.
Attenuation might achieve the goal in the specified time, but there is no
statistical confidence that the rate is adequate.
Attenuation is happening, but it may not be rapid enough to reach the
goal in the specified time.
Data are too variable or too few to determine if attenuation is proceeding
at a rate necessary to meet the goal in the specified time.
The data are too variable or too few to determine if attenuation is oc-
curring over time.
MW-12
•
Ground Water Flow
Figure 7.2 Location of monitoring wells in a plume of MTBE at Parsippany, New Jersey.
49
-------
Table 7.2 Long-term Monitoring Data at a Gasoline Spill Site at Parsippany, New Jersey
Date
12-Mar-1993
11 -May- 1993
17-Sep-1993
23-Sep-1994
17-May-1996
10-Aug-1996
7-Nov-1996
8-Dec-1997
27-Mar-1998
23-Jul-1998
18-Sep-1998
16-Dec-1998
1 -Mar- 1999
21-Jun-1999
7-Sep-1999
30-Dec-1999
20-Mar-2000
22-Jun-2000
MW-5
MW-11
MW-6
MW-10
MW-7
Ug/L
1,500
—
1,900
1,800
1,300
980
620
500
635
470
1,210
379
700
574
1,050
525
501
420
—
—
—
2,200
880
1.1*
660
339
426
419
—
144
123
464
155
220
173
146
140
—
270
200
120
120
66
—
71.2
—
44
—
42.2
—
43.2
—
36
140
—
290
180
5.3
5.3
23
13
—
3.4
—
ND*
—
4.41
—
16
—
6.4
5.2
19*
—
30*
150
100
20
17
—
—
—
—
—
—
—
—
—
—
3.7
: Not included in the regression.
The following illustrates the process to extract the rate constants from monitoring data. Microsoft EXCEL will be used
in the illustration because it is widely available to regulators in state agencies. The data are from MW-5 in Table 7.2.
Despite the best efforts at quality control, any large data set contains spurious data. Professional judgment was used
to exclude selected data in Table 7.2 from the regression.
Enter the dates the well was sampled (Column G in the example), and the concentrations of MTBE (Column H in the
example). Enter the formula for taking the natural logarithm of the contents of Cell HI into Cell II [ =LN(H2) ].
Use the mouse to click any cell other than Cell II, and Excel will accept the formula. After the formula is accepted,
drag it through the other cells in Column I to calculate the natural logarithm of all the data.
V EH. Ed« Vw* Inset Formal Iool> Dolo Window Help
12
A
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
d
B
=l i li,H2
c
*
D
— ror
E
IIUIU 111 I
F
, en LŁ
G
Date
12-Mar-1993
17-Sep-1993
23-Sep-1994
17-May-1996
10-Aug-1996
7-Nov-1996
B-Dec-1997
27-Mar-1998
23-JuM998
18-Sep-1998
16-Dee-1998
1-Mar-1999
21-Jun-1999
7-Sep-1999
30-Dec-1999
H
MTBE (ug/L)
1500
1900
1800
1300
980
620
500
635
470
1210
379
700
574
1050
525
I
LN MTBE
7.31322
7.549609
7.495542
7.17012
6.887553
6.429719
6.214608
6.453625
6.152733
7.098376
5.937536
6.55108
6352629
6.956545
6.263398
A
50
-------
Then open the Tools menu from the menu bar and select Data Analysis.
">le Łd* yiew insert Formal loots Qato
1
3
4
5
6
7
8
9
10
11
12
13
J21 »
A
14
15
16
17
18
•
B
I I
Wriord >
^vwwanm^^BB
^^^^T^^T^^H
I
E
F
G H
Date MTBE(ug/L)
12-Mar-1993 1500
17-Sep-1993
23-Sep-1994
17-May-1996
1900
1800
1300
10-Aug-1996 980
7-Nov-1996 620
8- Dec- 199 7
27-Mar-1998
23-JuH998
18-Sep-1998
500
635
470
1210
16- Dec- 1998' 379
1-Mar-1999| 700
21-Jun-1999| 574
7-Sep-1999 1050
30-Dec-1999 525
20-Mar-2000
22-Jun-2000
501
420
I
LN MTBE
7,31322
7.549609
7.495542
7 17012
6 887553
6.429719
6.214608
6.453625
6.152733
7.098376
5.937536
655108
6.352629
6.956545
6.263398
6.216606
6.040255
*
Open the Data Analysis menu and select Regression. -
1
2
3
•'.
5
'
8
9
10
11
1?
A B
I
Data Analysis
Analysis Toote
C D E
Covariarce
Descriptive StkUtct
JExpcnenoal Smoothug
•f-Twt Two-Sample fcr Variances
•Fourier Analyse
Histogram
Moving Avwag*
P.3Tdom Hunter Generation
|p 7* ^rrj P^rcenoie
3
i
F
.Uxl
OK |
. .1 , .-
Uelp |
G
Date
12-Mar-1993
17-Sep-1993
23-Sep-1994
17-May-1996
10-Aug-1996
T-Nov-1996
S-Dec-1997
27-Mar-1998
23-JUH998
18-Sep-1998
16-Dec-1998
H
MTBE (ug/L)
1500
1900
1800
1300
980
620
500
635
470
1210
379
L
LN MTBE
7.31322
7 54961
7.49554
7.17012
6.88755
6.42972
6.21461
6.45362
6.15273
7.09838
593754
—
Perform a linear regression of the natural logarithm of the concentrations of MTBE on the date the water samples were
collected from monitoring well #5.
Cell locations of the calculated natural logarithms of the concentrations are entered into the Input Y Range window, and
the dates are entered in the Input X Range window. The cell names can be typed into the windows, or click with the
mouse on the window, erase any names already entered, then select the data to be entered with the mouse. If you want
a level of statistical confidence that is different than 95% confidence, enter the desired level in the appropriate window.
In the example, a confidence level of 80% was selected. Click OK to perform the regression.
51
-------
>' File Edit View Insert Formal loots DOT a
G2 J
A B D E F G
-5— L Regression _U.xJ
« I
4 Cancel
5 Input i Range: i I ^j
6
7 r Latwfc r Constant is fcro
P CoriEkfrnc* L«v«l pi" %
q
*•» ^ j
t1 rftj%>ut Range: | _3
12 ^NswwerfcshwtBy: !~~
("Nnwlttrttxx*
15 PRjUKijal* r Rtsi^jal Pbts
. g r Standardized Restiwls T Lpe Fit Ptots
17 Normal Probability
18 r normal Robabttny Ptott
19
->r\ . ....
«*> |
[[Date
H
MTBE (ug/L)
l'2-Mar-1993] 1500
17-Sep-1993J 1900
23-Sep-1994j 1800
17-May-1996! 1300
10-Aug-1996
7-Nov-1996
980
620
8-Dec-1997! 500
27-Mar-1998
23-Jul-1998
18-Sep-1998
635
470
1210
16-Dec-1998J 379
1-Mar-l999
21-Jun-1999
700
574
7-Sep-1999! 1050
30-Dec-1999
j 20-Mar-20QO
• 22-Jurv2000
525
501
420
I
LN MTBE
7.31322
7.54961
7.49554
7.17012
6.88755
642972
6.21461
6.45362
6.15273
7.09838
5.93754
6.55108
6.35263
6.95655
6.2634
6.21661
6.04025
Excel returns a SUMMARY OUTPUT of the regression as a new tab in the workbook. The first order rate of change
of concentration with time is located in cell B18. This is the negative of the rate of attenuation. Because Excel was
supplied a date as the X variable, the rate constant is reported in units of day1.
Cells F18 and G18 contain the 95% confidence interval on the rate, and cells HI8 and 118 have the 80% confidence
interval on the rate. The rate of change and the confidence intervals were converted from units of per day to units of
per year by inserting a formula into Cell B19 that multiplied the rate by the number of days in a year, accepting the
formula, and dragging the formula onto cells F19 through 119.
r*% file Edit
Insert Fg/mai loole
dow yelp
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
l?n
A
.. _| '[,/IA.:-
Regressn
Mulbple R
R Square
Adjusted R
Standard E
'. 'I •-.••.••'•-.!'
ANOVA
Regressioi
Residual
Total
Intercept
X. Vanable
per year
B I
' OUTPUT
w Statistics
077612007
060236237
057585319
034248874
'
at ss
1 2 665349815
15 1 759478077
16 4424827892
Coefficients Standard Error
244325875 3730990758
-0.0004977 0000104402
-0 1817725|-^^^~
D
MS
266535
0 117299
tStat
6548552
-4.76684
E
F
2272279
P-value
922E-06
000025
F
Significance f-
0 00024962
_:.<>•,-• '-.-'.' •
16 480164
-0 0007202
-0 2630504
G
Upper 95%
3238501093
-000027514
-0.10049455
H
Lower 800%
19 430801
-000063763
-023289335
I
Upper 800%
2943437392
-0000357705
-01306516
52
-------
The confidence intervals on the rate of change are calculated in Excel using the Student's t distribution as defined in
Equation 7.2, where -k is the statistic of interest (the slope of the regression line), and sk is the standard deviation or
standard error of -k.
t =
-y\
At I
Equation 7.2
Think of the t statistic as the ratio of signal to noise. If the data used to calculate the t statistic have a normal distribu-
tion, the values of t are known for any level of confidence.
Excel calculates the confidence interval on -k following Equation 7.3:
confidence interval on - k = -k -1 * s,
Equation 7.3
The probability of error (a) is the probability that a calculated rate constant will be accepted as a statistically significant
rate, even though the calculated rate was a result of random variation and was not truly different from zero. The confi-
dence level is one minus the probability of error. At a confidence level of 90%, the probability of error is 10%. Excel
calculates the t statistic using a two-tailed distribution of errors in the estimate of the rate of change. Half of the error
in a is associated with rates where concentrations are increasing over time, and half is associated with rates of attenu-
ation. We are only interested in the half of the error that is associated with rates of attenuation. The 90% confidence
intervals calculated by Excel on the rate of change are 95% confidence intervals on the rate of attenuation. The 95%
confidence intervals on the rate of change are 97.5% confidence intervals on the rate of attenuation.
insert Ffirmo) Tools Qeta Window Help
•-•
17
18
Coefficients
Standard Error t Stat P-value Lower 9S% Upper 95%
jo •"•••••
Upper 800%
Intercept
24 4325875
3 730990758
6548552
9 22E-06
16480164
3238501093
19430801
2943437392
X Variable
-0 0004977
0000104402
-4.76684
t 0002
-00007202
-0.00027514
-0 00063763
-0,000357705
19 per year
-0.181772
-0.2630504
-0.10049455
-023289335 -0.1306516
\
-k
achieved
-k
\\ith-eonfidence
In the screen shot above from output of the linear regression, the rate of change in concentration with time is
-0.181772 per year (Cell B19), corresponding to a rate of attenuation kadlieved of 0.18 per year. The Upper 80.0% Con-
fidence interval on the rate of change is -0.1306516 per year (Cell G19), which corresponds to a k with. confldence on the
rate of attenuation of 0.13 per year at the 90% confidence level.
The difference between kadlieved and kwith.confldence is the value that a calculated kadlieved must exceed to be statistically differ-
ent from zero (Equation 7.4). This difference is the minimum rate of attenuation that can be detected at the accepted
level of confidence (kdetect) with the existing variability in the data. In this case, the value of kdetectis 0.181772 per year
minus 0.1306516 per year or 0.051 per year.
^achieved ~ ^with-confidence = ^detect Equation 7.4
Table 7.3 compares the concentrations of MTBE and the rates of attenuation of MTBE in all five wells in the plume
at Parsippany, New Jersey (see Table 7.2 and Figure 7.2). Table 7.3 summarizes an analysis of monitoring data from
March 1993 through June 2000. The current goal for MTBE in New Jersey is 70 |ig/L. To illustrate comparisons of
the rate constants, we will assume that a hypothetical goal of 20 |ig/L should be reached in a "reasonable" interval of
five years. The rate of attenuation necessary to meet the goal (k necessary) was calculated using Equation 7.1.
53
-------
Table 7.3 Progress of Natural Attenuation of MTBE at a Gasoline Spill Site at Parsippany, New Jersey
Well
MW-5
MW-11
MW-6
MW-10
MW-7
20 |ig/L is the hypothetical goal to close the site chosen for this illustration.
5.0 years is the hypothetical "reasonable" time period to reach the goal for this site, chosen for this illustration.
Current Concentration
(H8/L)
Maximum Concentration
(H8/L)
k neceSSary (per year)
k achieved (per year)
k with-confldence (P6r yCST)
95% confidence
k detect (per year)
number of sampling dates
420
1,900
0.61
0.18
0.13
0.05
17
146
2,200
0.39
0.45
0.36
0.09
13
51
270
0.19
0.27
0.22
0.05
11
5.2
290
Already met
goal
0.41
0.21
0.20
11
3.7
150
Already met
goal
0.64
0.36
0.28
5
Initially, five wells had concentrations of MTBE that were higher than 20 |ig/L. By 2000, two of the five wells had
reached the cleanup goal. These wells (MW-10 and MW-7) were at the lateral margins of the plume (compare Figure
7.2). Two wells near the source area (MW-5 and MW-6) and one distant well (MW-11) still maintained significant
concentrations of MTBE in 2000. Values for kadlieved in the wells distant from the source area (MW-7, MW-10, MW-11)
were greater than values in the wells near the source (MW-5 and MW-6) by a factor of two or three. Concentrations
of MTBE in the plume appeared to be retreating back toward the source area.
In one of the wells with concentrations of MTBE above the goal (MW-6), the kwith.confldence was greater than k^^^
Natural attenuation was on track to meet the goal in a reasonable time period. In a second well (MW-11), kadlieved was
greater than
but kwith.confldence was less than
Natural attenuation may have been on track to meet the goal,
but the data were too variable or too few to support the projection at a 90% confidence level. In the well with the highest
concentration of MTBE (MW-5), kachieved was much less than k^.^. In order to meet the goal in five years, it would
be necessary to actively remediate the source area near MW-5.
54
-------
Section 8 -
Typical Rates of Attenuation in Source Areas -
Regulators are often asked to determine the number of rounds of sampling that is necessary for them to evaluate the
behavior of an MTBE plume. To plan a monitoring effort, it is necessary to have some idea of the rate of attenuation
MTBE in the source area of a plume and the variation in that rate over time. This section illustrates the range of rates
that might be expected at a typical gasoline spill site. At a major proportion of MTBE sites, the long-term monitoring
data will fail to show that MTBE is attenuating in the most contaminated wells. It also illustrates the number of rounds
of sampling that are necessary to document the rate of natural attenuation with statistical significance. In general, short
data sets with less than twelve samples may fail to detect rates of attenuation that have environmental significance.
8.1 Typical Rates of Attenuation Over Time in Source Areas
Wilson and Kolhatkar (2002) compared the rate of attenuation of MTBE over time in the source area of five plumes to
the rate of attenuation of MTBE in ground water along the flow path. The rates of attenuation in ground water were
from two-fold to more than ten-fold faster. Thus, in the source area, the persistence of the plumes is controlled by the
rate of attenuation of concentrations in the source areas. Durrant et al., (1999) explained the long-term persistence of a
plume of MTBE in California by modeling the diffusion of high concentrations of MTBE into regions of low hydraulic
conductivity early in the spill. Over time, the MTBE in the residual gasoline dissolved into ground water and was car-
ried away by the flow of ground water. After MTBE in the residual gasoline was depleted, the plume was sustained by
the slow diffusion of MTBE back out of the regions of low hydraulic permeability.
Other situations can produce the same behavior. If residual gasoline remains in the aquifer, MTBE can slowly parti-
tion from the residual gasoline to the ground water. Peargin (2000, 2001) compared the relative rate of attenuation of
MTBE, benzene, and xylenes in wells in the smear zones of gasoline spills. The rate of attenuation was independent of
the water solubility of the contaminant and could not be explained by the expected rate of dissolution from gasoline into
water. Peargin (2000, 2001) concluded that mass transfer limitations slowed the transfer of MTBE from the gasoline
to the active flow paths in the aquifer.
Wilson and Kolhatkar (2002) extracted the rate of attenuation of MTBE in the most contaminated well at gasoline spills
in California, Florida, North Carolina, New York, and the site at Parsippany, New Jersey, that was discussed extensively
in the previous section. The rate of attenuation at the five sites varied from 0.15 per year to 0.75 per year. The rate
that was statistically significant at 90% confidence varied from 0.04 per year to 0.29 per year. Peargin (2000, 2001)
extracted the rate of natural attenuation of MTBE in 23 wells in the smear zone of 15 gasoline stations in the eastern
U.S. (primarily Maryland). The fastest rate of attenuation of the source was 0.7 per year, equivalent to a half-life of
one year. The mean rate of attenuation was 0.04 per year, equivalent to a half-life of 17 years. Shorr and Rifai (2002)
calculated the rate of change in the concentration of MTBE over time for 694 monitoring wells at gasoline spill sites in
Texas. In two thirds of the wells, the concentrations of MTBE declined over time. In the wells where the concentra-
tions of MTBE declined over time, the median rate of attenuation was 0.043 per year, corresponding to a half-life of
16 years, and 25% had a rate equal to, or greater than, 0.37 per year, equivalent to a half-life near two years. Robb and
Moyer (2003) provided monitoring data on a site in the Midwestern U.S. The rate of attenuation of MTBE in the most
contaminated well was 0.62 per year (0.29 per year at 90% confidence) equal to a half-life of 1.1 years.
Figure 8.1 collates the rates of attenuation of MTBE over time in the most contaminated well at thirteen of the sites
described by Peargin (2000, 2001) that had not been subjected to remediation, at five sites discussed in Wilson and
Kolhatkar (2002), at the site in the Midwestern U.S. discussed by Robb and Moyer (2003), and at the site in South
Carolina described in Landmeyer et al., (1998). Landmeyer (personal communication J. Landmeyer, USGS, Columbia,
SC) provided monitoring data on attenuation of MTBE in the most contaminated well at the site in South Carolina.
Twenty sites are not a statistically representative sample of the hundreds of thousands of MTBE sites in the United States;
however, the distribution of the rates of attenuation can at least illustrate the possible behavior of MTBE in gasoline spill
sites. In six of twenty sites, the concentration of MTBE increased over time in the most contaminated well, instead of
55
-------
Jo 0.8
CD
•
0.4
0.2
0
-0.2
O
iZ -0.4
-0.6
-*- rate of attenuation (per year)
^- 90% one-tailed confidence interval
6 8 10 12 14 16
Sites Ranked by the Rate of Attenuation
18
20
Figure 8.1 -Distribution of the rates of attenuation ofMTBE over time in source areas of plumes from gasoline spills.
Negative rates indicate increasing concentrations with time.
attenuating. At 14 sites, concentrations of MTBE attenuated over time. At 11 of the sites, the 90% one-tailed confi-
dence interval on the rate was greater than zero. At a major proportion of MTBE sites, long-term monitoring data will
fail to show that MTBE is attenuating in the most contaminated wells, or will fail to show that MTBE is attenuating at
90% confidence. If the rate of attenuation is truly greater than zero, but the rate of attenuation is slow and variable, the
chance that a calculated rate of attenuation will be found to be statistically different from zero is strongly influenced
by the number of samples used to calculate the rate. As the number of samples goes up, the proportion of calculated
rates that are found to be statistically greater than zero will also increase. To illustrate this effect, Figure 8.2 compares
the minimum rate of attenuation that was detectable at 90% confidence to the number of samples used to extract the
rates presented in Figure 8.1.
8.2 Number of Sampling Dates Needed to Calculate Rates of Attenuation
A regression analysis of long-term trends in monitoring data is subject to two kinds of error. The analysis may fail to
detect attenuation when the attenuation is really happening. This happens most often when the rate of attenuation is
slow and variable, and there are simply not enough data to distinguish the true trend above the natural variation in the
data. If this is the case, the chance that a calculated rate of attenuation will be found to be statistically different from
zero is strongly influenced by the number of sampling dates used to calculate the rate. As the number of sampling dates
increases, the proportion of calculated rates that are found to be statistically greater than zero will also increase.
Often state agencies only have sampling data for a limited number of dates. To determine the minimum number of
sample dates that are needed to extract a rate that is statistically significant, a regression analysis was performed on
small portions of the long-term monitoring data used to extract the rates in Figure 8.1. Portions of each long-term
record were selected that contained four, five, six, or more dates. To avoid bias, the dates in the portions of the record
were selected to distribute the dates in the portion equally about the central date of the parent record. Then the portions
were analyzed as described in Section 7 to determine whether the rate of attenuation was statistically different from
zero. The number of dates in the portion were expanded until regression analysis indicated that rate of attenuation was
statistically significant with a one-tailed confidence level of 90%. Then the number of dates was expanded again until
the rate was significant at 95% confidence.
The number of sampling dates required for the rate to be statistically significant is presented in Figure 8.2. For many of
the data sets, as few as four samples were adequate to extract a rate statistically greater than zero. Several state agencies
will evaluate a site for natural attenuation after two years of quarterly monitoring. Eight quarters of monitoring would
have failed to recognize natural attenuation at 6 of the 14 sites. However, eight sampling dates are an efficient size for
the minimum data set to evaluate natural attenuation. Eight sampling dates were sufficient to recognize roughly one-
half of the sites where natural attenuation was occurring.
56
-------
It -
CO
o 12 -
1
"5 ~ 10 -
co co
w 'F a -
-1-1 — °
C D)
1 co
| 1 6 -
CO .9
co -K
S "co 4 -
M_ CO
o
Number
3 N5
rn
-
-
.-,
• Significant at 90%
D Significant at 95%
i-i r-i
r
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Sites Ranked by Rate of Attenuation
Figure 8.2. Variation in the number of sampling dates in a data set required to extract a rate of natural attenuation
that is statistically significant.
9 n
8 -
7 -
6 -
T 5 -
§
.E 4 -
LU
CO
i- 3 H
All the data from the site
ranked 12 in Figure 8.1
\
\
\
\
Data from the site ranked
12 in Figure 8.2 |
0
Jan-93 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00
Date
Figure 8.3. Potential for error when a short data set is used to estimate the rate of attenuation of concentrations over
time.
57
-------
The level of statistical confidence in the regression analysis had little effect on the number of sampling dates required.
There was little difference between the number of sample dates needed to extract a rate that was significant at 90%
confidence and a rate that was significant at 95% confidence.
In the second kind of error, a trend that is recognized as significant over a few dates is not borne out over the longer
interval, because the shorter interval used to evaluate the rate of attenuation was not representative of the longer interval.
This effect is illustrated in Figure 8.3, using data from the site ranked number 12 in Figures 8.1 and 8.2. The rate of
attenuation calculated for the entire data set as presented in Figure 8.1 was 0.06 per year, which was not significantly
different from zero at 90% confidence. The rate calculated for the five central dates as presented in Figure 8.3 is 0.49
per year. These five dates were used to calculate the minimum number of dates required in Figure 8.2. Over the smaller
data set, the rate of attenuation was much faster, and the rate was significant at 95% confidence. The only protection
from this second kind of error is more sampling dates that extend over a longer period of time.
8.3 Effect of Number of Sampling Dates on the Detectable Rate of Attenuation
The number of sampling dates in the data set also has an effect on the detection limit for the rate of natural attenuation.
As discussed in Section 7, the detection limit is the minimum rate of attenuation statistically different from zero at
some level of confidence (kdetect). Figure 8.4 presents the relationship between the number of dates in the data set and
kdetect at 90% confidence for the 14 sites presented in Figure 8.1 where MTBE was attenuating over time in the most
contaminated well.
Most of the data sets from the 14 sites are more extensive than is usually available. To estimate minimum detectable
rates that would be extracted with fewer data, Figure 8.4 also presents the minimum rates of attenuation that would be
detected when the rates were calculated with half of the available data. The rates were calculated from half of the data
that occupied the middle portion of the monitoring record. The first portion (approximately 25%) and final portion
(approximately 25%) of the monitoring data were excluded.
In their survey of the rate of attenuation of MTBE at sites in Texas, Shorr and Rifai (2002) found the rate of attenuation
met or exceeded 0.001 per day, or 0.37 per year in only 17% of wells. As discussed in Section 2, the median concen-
tration of MTBE in the most contaminated well at gasoline spill sites in Texas is near 1,000 |ag/L. If the rate of natural
attenuation is 0.37 per year, a site with a maximum concentration of 1,000 |ig/L of MTBE would require 12 years to
reach the EPA advisory concentration of 20 |ig/L.
A data set to evaluate natural attenuation of MTBE should have a detection limit (kdetect) lower than 0.37 per year.
When the number of sampling dates in Figure 8.4 was greater than 12, the detection limit for attenuation was less than
0.37 per year. When the number of sampling dates was less than 12, the detection limit for most of the data sets was
greater than 0.37 per year. Short data sets with less than 12 samples may fail to detect rates of attenuation that have
environmental significance.
8.4 Effect of Seasonal Variations
The statistical considerations discussed so far ignore any seasonal effects on the concentration of contaminants. If
there are strong seasonal effects on the recharge of precipitation to ground water, these effects may be reflected in the
measured concentrations of MTBE in monitoring wells. As the water table moves up and down in response to recharge,
the water may wet more or less of the gasoline in the smear zone, resulting in higher or lower concentrations of MTBE
in the well. If a plume is vertically heterogeneous, the screened interval of a monitoring well may sample different
vertical regions in the same plume as the water table shifts.
Seasonal variations can add to the variability in short-term data sets. When there are strong seasonal effects, it may be
useful to extract the rate of attenuation of the seasonal maximum concentrations, and compare that rate to the overall
rate of attenuation. Figure 8.5 illustrates a data set from a well with a strong seasonal component. The concentra-
tion maximums are associated with the summer months in 1993, 1994, 1996, and 1998. However, not every summer
shows a maximum. There was no maximum in the summer of 1995, and the well was not sampled in the summer of
1997. In this case, there was no difference in the rate of attenuation of the seasonal maximum concentrations and the
concentrations throughout the year. When the rate of attenuation is extracted from the complete data set (connected by
the solid line in Figure 8.5), the rate of attenuation is 0.48 per year, or 0.30 per year at 90% confidence. If the rate is
extracted from the four summer maxima (dashed line in Figure 8.3), the rate is 0.47 per year, or 0.34 per year at 90%
confidence.
58
-------
«
.a
=g I
n
E
c
z -
1.8
1.6
1.4 -
1.2 -
1
0.8
0.6 -
0.4 -
0.2 -
n
* complete
n halfofdal
n
D
n Half life of 2 years
n /
/
an /
a » /
/
n ~
n a
D 4
g« » » » »
* * *
data
a series
^
10 20 30
Number of Sampling Dates
40
Figure 8.4 Effect of the number of samples used to calculate a rate of attenuation on the minimum rate of attenuation
that is statistically different from zero at 90% confidence.
80000
70000-
60000-
50000-
40000-
30000-
20000-
10000-
Summer Maximum
Concentrations
"TV
Jul-92 Jul-93 Jul-94 Jul-95 Jul-96 Jul-97 Jul-98 Jul-99
Date of Sampling
Figure 8.5 A monitoring record from a well in Maryland with seasonal maximum concentrations ofMTBE in certain
years.
59
-------
Preceeding Page Blank
Section 9 -
Quality Assurance Statement -
9.1 Analysis of Concentrations in Water
Laboratory analyses for data presented in Table 4.3, Figure 4.4, and Figure 4.5 were conducted at the Robert S. Ken-
Environmental Research Center in accordance with a Quality Assurance Project Plan prepared for in-house task 10013
(Fate of Fuel Oxygenates in Aquifer Materials, approved January 2001). Concentrations of MTBE, TEA, and benzene
were determined following in-house SOPs very similar to Lin et al., 2003. Water samples were prepared with a heated
static headspace sampler, and determined by gas chromatography and mass spectrometry. Ethanol in water was deter-
mined by direct aqueous injection onto a gas chromatograph equipped with a flame ionization detector.
The stated data quality objectives for analysis of MTBE, TEA, benzene, and ethanol, were as follows: The reported
concentration of continuing calibration check standards and matrix spikes will agree with the expected concentration
plus or minus 15% of the known concentration. Analytical duplicates will agree with each other plus or minus 15%.
The microcosms constructed with sediment from the gasoline spill site at Parsippany, New Jersey, were the only mi-
crocosms that acclimated and degraded MTBE under anaerobic conditions. Tables 9.1 and 9.2 summarize typical data
quality for MTBE in the microcosms constructed with material from Parsippany, New Jersey. The first five sampling
dates correspond to the data presented in Figure 4.4. Five out of 22 of the calibration check standards did not meet the
goal of ± 15% of the nominal value (Table 9.1). However, none of the check standards varied by more 25% from the
nominal value. Two out of nine laboratory duplicates did not agree within 15%, but all the duplicates agreed within a
relative percent difference of 25% (Table 9.2). Two out of six of the matrix spikes exceed the goal of ± 15%, but all
the matrix spikes were within ±25% (Table 9.2).
The method blanks were always less than 0.5 |ig/L with the exception of samples collected on 11/16/01 (Table 9.2).
On this date, the analysts diluted all the samples provided to him by two fold, to have enough material for a laboratory
duplicate.
One of the treatments included in the experiment was a container control, containing the analytes of interest in sterile
water. At most sampling intervals, triplicate container controls were analyzed. The variation between the triplicate
analyses represents variability in the construction of the microcosms as well as any error in the analysis of the concen-
trations of the analyte. The sample standard deviation was never more than 5% of the mean of the triplicate analyses
(Table 9.2). The variation between the mean of triplicate samples in the "living" microcosms was always much higher
(See Figure 4.4), reflecting the influence of sorption and biodegradation on concentrations of the analytes.
Due to unavoidable problems with laboratory infrastructure, some of the samples were held for up to 116 days before
analysis. The samples were preserved in trisodium phosphate. There was no indication from the container controls that
the long holding time caused any loss of MTBE (Table 9.2). There was no statistically significant difference between
the mean concentration of MTBE at the beginning of the experiment and the concentration after 358 days of incubation.
The mean concentration of MTBE in the container controls varied as much as 9% from one sampling time to the next
(6/28/01 to 7/26/01). Variation in mean concentrations in the container controls most likely reflected normal variations
in the calibration of the analytical instrument, and not loss or gain of MTBE in the microcosms.
All the data for MTBE were determined to be of acceptable quality, and the data were used in the report.
Tables 9.3 and 9.4 summarize typical data quality for TEA in the microcosms constructed with material from Parsip-
pany, New Jersey. Analyses of TEA were not as accurate or precise as the analyses of MTBE. Six out of 22 of the
calibration check standards did not meet the goal of ± 15% of the nominal value (Table 9.3). Two of the six did not
meet a goal of ± 25% of the nominal value. One check sample was reported as 155% of the nominal value. Two of
seven matrix spike samples did not meet the goal of ± 15% of the spiked value (Table 9.4). One of the matrix spike
samples reported 126% of the spiked concentration.
61
-------
The method blanks were always less than 10 |ag/L with the exception of samples collected on 11/16/01. On this date,
the analysts diluted all the samples provided to him by two fold, to have enough material for a laboratory duplicate.
The standard deviation of the samples in the container controls was 10% or less of the mean. The mean concentration
of MTBE in the container controls varied as much as 36% from one sampling time to the next (5/2/01 to 10/25/01).
All the data for TEA were determined to be of acceptable quality, and the data were used in the report.
Tables 9.5 and 9.6 summarize typical data quality for benzene in the microcosms constructed with material from Par-
sippany, New Jersey. Analyses of benzene were not as accurate and precise as the analyses of MTBE. Seven out of
22 of the calibration check standards did not meet the goal of ± 15% of the nominal value (Table 9.5). However, all of
them did meet a goal of ± 25% of the nominal value. One check sample was reported as 123% of the nominal value.
Two of seven matrix spike samples did not meet the goal of ±15% of the spiked value (Table 9.6). However, all of the
matrix spike samples met a goal of ± 25% of the spiked concentration.
The method blanks were always less than 0.5 |ig/L with the exception of samples collected on 11/16/01. On this date,
the analysts diluted all the samples provided to him by two fold, to have enough material for a laboratory duplicate.
The standard deviation of the samples in the container controls was 10% or less of the mean. The mean concentration
of benzene in the container controls varied as much as 53% from one sampling time to the next (5/2/01 to 10/25/01)
All the data for benzene were determined to be of acceptable quality, and the data were used in the report.
Tables 9.7 and 9.8 summarize typical data quality for ethanol in the microcosms constructed with material from Parsip-
pany, New Jersey. The continuing calibration standards, duplicates, and matrix spike samples were all within the goal
of ± 15%. Ethanol was not included in the container controls. All the blanks were less than 1 mg/L. All the data for
ethanol were determined to be of acceptable quality, and the data were used in the report.
Several Tables and Figures in this report reference data provided by state agencies. To our knowledge, these analyses
were conducted following EPA 8260 or 8260B (purge and trap with gas chromatography with a mass spectrometer
detector). The results should be comparable to results obtained at the Kerr Center.
9.2 Stable Carbon Isotope Analyses
The analyses were performed by the University of Oklahoma, acting as a private contractor. The quality assurance
data provided with the samples indicated that the sample standard deviation of determination of 513C varied from ± 0.1
%e to ± 0.18 %o. The nature of the analysis makes it impossible to do a matrix spike. Each analysis is referenced to a
calibration standard.
62
-------
Table 9.1 Typical Quality Performance Data for Continuing Calibration Check Standards for MTBE in Water. All Values are |ig/L Unless
Otherwise Indicated
Date Collected
Date Analyzed
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
Check Standard Nominal
Check Standard Value
Percent of Check Standard
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
5/2/01
5/31/01
20
21.8
109.0%
200
220
110.0%
200
204
102.0%
5/31/01
9/24/01
200
209
104.5%
20
21.3
106.5%
200
193
96.5%
6/28/01
9/3/01
200
204
102.0%
200
237
118.5%
20
23.7
118.5%
7/26/01
9/21/01
200
239
119.5%
20
21.3
106.5%
200
233
116.5%
10/25/01
11/30/01
200
204
102.0%
20.0
20.4
102.0%
11/16/01
11/26/01
200
228
114.0%
200
236
118.0%
4/25/02
4/25/02
200
235
117.5%
20
18.1
90.5%
200
184
92.0%
5/15/03
5/16/03
20
19.6
98.0%
100
100
100.0%
200
210
105.0%
-------
Table 9.2 Typical Quality Performance Data for Analysis of MTBE in Water, Including Blanks, Laboratory Duplicates, and Matrix Spikes. All
Values are |ig/L Unless Otherwise Indicated
Date Collected
Date Analyzed
Blank 1
Blank 2
Sample Analysis 1
Laboratory Duplicate 1
Relative Percent Difference
Sample Analysis 2
Laboratory Duplicate 2
Relative Percent Difference
Spike Concentration
Sample Concentration
Spike Recovery (Percent)
Time of Incubation (Days)
Holding Time (Days)
Sterile Water Control 1
Sterile Water Control 2
Sterile Water Control 3
Mean
Standard Deviation
5/2/01
5/31/01
no report
no report
910
886
2.7%
159
164
3.1%
200
134
108%
0
29
39.2
39.6
38.2
39.0
0.72
5/31/01
9/24/01
<0.5
<0.5
20.2
20.6
2.0%
200
25.6
99%
29
116
37.4
37.0
37.6
37.3
0.31
6/28/01
9/3/01
<0.5
<0.5
13.3
15
12.0%
200
17.7
113%
57
67
35.8
35.4
36.4
35.9
0.50
7/26/01
9/21/01
<0.5
<0.5
15.9
15.6
1.9%
200
14.9
119%
85
57
37.4
39.0
41.2
39.2
1.91
10/25/01
11/30/01
<0.5
3.3
2.8
16.4%
6.2
5.6
10.2%
200
7.4
114%
176
36
36.6
38.5
38.5
37.9
1.10
11/16/01
11/26/01
<1
<1
<1
200
<1
116%
198
10
4/25/02
4/25/02
<0.3
<0.3
10.3
8.45
19.7%
358
0
39.9
39.0
36.1
38.3
1.99
5/15/03
5/16/03
<0.1
<0.1
65
61.9
4.9%
-------
Table 9.3 Typical Quality Performance Data for Continuing Calibration Check Standards for TEA in Water. All Values are |ig/L Unless
Otherwise Indicated
Date Collected
Date Analyzed
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
Check Standard Nominal
Check Standard Value
Percent of Check Standard
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
5/2/01
5/31/01
20
17.5
87.5%
200
168
84.0%
200
191
95.5%
5/31/01
9/24/01
200
207
103.5%
20
31.1
155.5%
200
186
93.0%
6/28/01
9/3/01
200
198
99.0%
200
192
96.0%
20
21.1
105.5%
7/26/01
9/21/01
200
216
108.0%
20
19.5
97.5%
200
185
92.5%
10/25/01
11/30/01
200
237
118.5%
20.0
24.3
121.5%
11/16/01
11/26/01
200
252
126.0%
200
241
120.5%
4/25/02
4/25/02
200
223
111.5%
20
22.5
112.5%
200
239
119.5%
5/15/03
5/16/03
20
22.5
112.5%
100
98.6
98.6
200
200
100.0%
-------
Table 9.4 Typical Quality Performance Data for Analysis of TEA in Water, Including Blanks, Laboratory Duplicates, and Matrix Spikes. All
Values are |ig/L Unless Otherwise Indicated
Date Collected
Date Analyzed
Blank 1
Blank 2
Sample Analysis 1
Laboratory Duplicate 1
Relative Percent Difference
Sample Analysis 2
Laboratory Duplicate 2
Relative Percent Difference
Spike Concentration
Sample Concentration
Spike Recovery (Percent)
Time of Incubation (Days)
5/2/01
5/31/01
no report
no report
<10
<10
41
42.8
4.3%
200
37.2
85%
0
5/31/01
9/24/01
<10
<10
84
88.2
4.9%
200
65.8
89%
29
6/28/01
9/3/01
<10
<10
61.8
67
8.1%
200
68
101%
57
7/26/01
9/21/01
<10
<10
87.2
86.6
0.7%
200
68
93%
85
10/25/01
11/30/01
<10
83.6
87.4
4.4%
55.2
53.6
2.9%
200
42.4
126%
176
11/16/01
11/26/01
<20
113
109.6
3.1%
200
71.8
121%
4/25/02
4/25/02
<2.6
<2.6
118
105
11.7%
603
667
10.1%
358
5/15/03
5/16/03
<2.8
<2.8
95.3
90.4
5.3%
Table 9.5 Typical Quality Performance Data for Continuing Calibration Check Standards for Benzene in Water. All Values are |ig/L Unless Other-
wise Indicated
Date Collected
Date Analyzed
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
Check Standard Nominal
Check Standard Value
Percent of Check Standard
Check Standard Nominal
Check Standard Measured
5/2/01
5/31/01
20
22.6
113.0%
200
246
123.0%
200
216
5/31/01
9/24/01
200
209
104.5%
20
21.3
106.5%
200
193
6/28/01
9/3/01
200
204
102.0%
200
237
118.5%
20
23.7
7/26/01
9/21/01
200
239
119.5%
20
21.3
106.5%
200
233
10/25/01
11/30/01
200
204
102.0%
20.0
20.4
102.0%
11/16/01
11/26/01
200
228
114.0%
200
236
118.0%
4/25/02
4/25/02
200
235
117.5%
20
18.1
90.5%
200
184
5/15/03
5/16/03
10
10.1
101.0%
50
50.6
101.2%
100
102
-------
Table 9.6 Typical Quality Performance Data for Analysis of Benzene in Water, Including Blanks, Laboratory Duplicates, and Matrix Spikes. All
Values are |ig/L Unless Otherwise Indicated
Date Collected
Date Analyzed
Blank 1
Blank 2
Sample Analysis 1
Laboratory Duplicate 1
Relative Percent Difference
Sample Analysis 2
Laboratory Duplicate 2
Relative Percent Difference
Spike Concentration
Sample Concentration
Spike Recovery (Percent)
Time of Incubation (Days)
Holding Time (Days)
Sterile Water Control 1
Sterile Water Control 2
Sterile Water Control 3
Mean
Standard Deviation
5/2/01
5/31/01
no report
no report
604
562
7.2%
48.2
48.6
0.8%
200
36
121%
0
29
24.8
25.8
24.8
25.1
0.58
5/31/01
9/24/01
<0.5
<0.5
20.2
20.6
2.0%
200
25.6
99%
29
116
20.4
20.2
20.2
20.3
0.12
6/28/01
9/3/01
<0.5
<0.5
13.3
15
12.0%
200
17.7
113%
57
67
21.8
21.2
21.0
21.3
0.42
7/26/01
9/21/01
<0.5
<0.5
15.9
15.6
1.9%
200
14.9
119%
85
57
21.6
21.2
20.6
21.1
0.50
10/25/01
11/30/01
<0.5
3.3
2.8
16.4%
6.2
5.6
10.2%
200
7.4
114%
176
36
16.0
14.8
13.1
14.6
1.46
11/16/01
11/26/01
<1
<1
<1
200
<1
116%
4/25/02
4/25/02
<0.3
<0.3
10.3
8.45
19.7%
358
0
16.1
14.8
13.1
14.7
1.50
5/15/2003
5/16/2003
<0.18
<0.18
15.6
13.6
13.7%
-------
Table 9.7 Typical Quality Performance Data for Continuing Calibration Check Standards for Ethanol in Water. All Values are mg/L Unless
Otherwise Indicated
Date Collected
Date Analyzed
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
Check Standard Nominal
Check Standard Value
Percent of Check Standard
Check Standard Nominal
Check Standard Measured
Percent of Check Standard
5/2/01
5/31/01
85
85.8
100.9%
15
15.5
103.3%
85
87.2
102.6%
5/31/01
9/24/01
15
14.1
94.0%
75
75.6
100.8%
6/28/01
9/3/01
15
15
100.0%
15
15.1
100.7%
7/26/01
9/21/01
15
15.8
105.3%
75
78.3
104.4%
10/25/01
11/30/01
15
15
100.0%
75.0
71.7
95.6%
15
15.5
103.3%
11/16/01
11/26/01
25
28.7
114.8%
25
27.4
109.6%
4/25/02
4/25/02
10
10.3
103.0%
50
49.9
99.8%
Table 9.8 Typical Quality Performance Data for Analysis of Ethanol in Water, Including Blanks, Laboratory Duplicates, and Matrix Spikes.
All Values are mg/L Unless Otherwise Indicated
Date Collected
Date Analyzed
Blank 1
Blank 2
Sample Analysis 1
Laboratory Duplicate 1
Relative Percent Difference
Spike Concentration
Sample Concentration
Spike Recovery (Percent)
5/2/01
5/31/01
<1
<1
1980
1950
1.5%
40
83.7
104%
5/31/01
9/24/01
<1
<1
40
83.2
102%
6/28/01
9/3/01
<1
<1
40
<1
104%
7/26/01
9/21/01
<1
<1
40
<1
101%
10/25/01
11/30/01
<1
<1
40
37.4
94%
11/16/01
11/26/01
<1
<1
17.5
18.3
4.5%
100
52
93.6%
4/25/02
4/25/02
<1
<1
37.5
37.2
0.8%
100
37.4
95.8%
-------
Section 10 -
References -
Amerson, I. and R. L. Johnson. A natural gradient tracer test to evaluate natural attenuation of MTBE under anaerobic conditions.
Ground Water Monitoring & Remediation 23 (1): 54-91 (2002).
Bradley, P. M., F. H. Chapelle, and J. E. Landmeyer. Methyl t-butyl ether mineralization in surface-water sediment microcosms
under denitrifying conditions. Applied and Environmental Microbiology 67 (4): 1975-1978 (2001a).
Bradley, P. M., F. H. Chapelle, and J. E. Landmeyer. Effect of redox conditions on MTBE biodegradation in surface water sedi-
ments. Environmental Science & Technology 35 (23): 4643-4647 (2001b).
Bradley, P. M., J. E. Landmeyer, and F. H. Chapelle. Widespread potential for microbial MTBE degradation in surface-water
sediments. Environmental Science & Technology 35 (4): 658-662 (200Ic).
Borden, R. C., R. A. Daniel, L. E. LeBrun IV, and C. W. Davis. Intrinsic biodegradation of MTBE and BTEX in a gasoline-con-
taminated aquifer. Water Resources Research 33 (5): 1105-1115 (1997).
Buscheck, T. E., and C. M. Alcantar. "Regression techniques and analytical solutions to demonstrate intrinsic bioremediation." In
Intrinsic Bioremediation, Proceedings of the Third International In Situ and On-Site Bioremediation Symposium, San Diego,
CA, (1995).
Cline, P. V., J. J. Delfino, and P. S. C. Rao. Partitioning of aromatic constituents into water from gasoline and other complex solvent
mixtures. Environmental Science & Technology 25 (5): 914-920 (1991).
Cozzarelli, I. M., and Baehr, A. L. "Volatile fuel hydrocarbons and MTBE in the environment treatise on geochemistry." In Treatise
on Geochemistry: Vol. 9, Environmental Geochemistry, Amsterdam: Elsevier/Pergamon, 2003, 433-474.
Crumbling, D. M., and B. Lesnik. Analytical issues with MTBE and related oxygenate compounds. L.U.S.T.LINE: A report on
federal and state programs to control leaking underground storage tanks. New England Interstate Water Pollution Control
Commission, Bulletin 36, 2000.
Day, M. J. and Gulliver, T. "Natural attenuation of tert butyl alcohol at a Texas chemical Plant." In MTBE Remediation Hand-
book, Amherst, MA: Amherst Scientific Publishers, 2003, 541-560.
Deeb, R. A., K. M. Scow, and L. Alvarez-Cohen. Aerobic MTBE biodegradation: An examination of past studies, current chal-
lenges, and future research directions. Biodegradation 11: 171-186 (2000a).
Deeb, R. A., S. Nishino, J. Spain, H. Hu, K. Scow, and L. Alvarez-Cohen. "MTBE and benzene biodegradation by a bacterial
isolate via two independent monooxygenase-initiated pathways." Paper presented at the Division of Environmental Chemistry,
American Chemistry Society Symposium, San Francisco, CA, March 26-30, 2000, (2000b).
Deeb, R. A., H. Y. Hu, J. R. Hanson, K. M. Scow, and L. Alvarez-Cohen. Substrate interactions in BTEX and MTBE mixtures by
an MTBE-degrading isolate. Environmental Science & Technology 35 (2): 312-317 (2001).
Durrant, G. C., M. Schirmer, M. D. Einarson, R. D. Wilson, and D. M. Mackay. "Assessment of the dissolution of gasoline con-
taining MTBE at LUST Site 60, Vandenberg Air Force Base, California. In Proceedings of the Petroleum Hydrocarbons and
Organic Chemicals in Groundwater: Prevention, Detection, and Restoration. National Ground Water Association, Houston,
TX, November 17-19, 1999.
Finneran, K. T. and D. R. Lovley. Anaerobic degradation of methyl tert-butyl ether (MTBE) and tert-butyl alcohol (TEA). Envi-
ronmental Science and Technology 35 (9): 1785-1790 (2001).
Finneran, K. T. and D. R. Lovley. "Anaerobic in situ bioremediation." In MTBE Remediation Handbook, Amherst, MA: Amherst
Scientific Publishers, 2003, 265-277.
Fiorenza, S. and H. S. Rifai. Review of MTBE biodegradation and bioaugmentation. Bioremediation Journal 1 (1): 1-36 (2003).
Fortin, N.Y. andM. A. Deshusses. Treatment of methyl tert-butyl ether vapors in biotrickling filters. 1. Reactor startup, steady state
performance, and culture characteristics. Environmental Science and Technology 33 (17): 2980-2986 (1999).
Francois, A. L., H. Gamier, F. Mathis, F. Fayolle, and F. Monot. Roles of tert-butyl formate, tert-butyl alcohol and acetone in
the regulation of methyl tert-butyl ether degradation by Mycobacterium austroafricanum IFP 2012. Applied Microbiology
Biotechnology 62: 256-262 (2003).
69
-------
Gamier, P., R. Auria, M. Magana, and S. Revah, S. "Cometabolic biodegradation of Methyl f-butyl Ether by a soil consortium."
In In Situ Bioremediation of Petroleum Hydrocarbon and Other Organic Compounds. The Fifth International In Situ and On-
Site Bioremediation Symposium, San Diego, CA, 1999.
Gray, J. R., G. Lacrampe-Couloume, D. Gandhi, K. M. Scow, R. D. Wilson, D. M. Mackay, and B. S. Lollar. Carbon and hydrogen
isotopic fractionation during biodegradation of methyl tert-butyl ether. Environmental Science & Technology 36 (9): 1931-1938
(2002).
Hanson, J. R., C. E. Ackerman, and K. M. Scow. Biodegradation of methyl tert-butyl ether by a bacterial pure culture. Applied and
Environmental Microbiology 65: 4788-4792 (1999).
Hattan, G., B. Wilson, and J. T. Wilson. Performance of conventional remedial technology for treatment of MTBE and benzene
at UST sites in Kansas. Remediation 14 (1): 85-94 (2003).
Hickman, G. T. and J. T. Novak. Relationship between subsurface biodegradation rates and microbial density. Environmental
Science and Technology 23 (5): 525-532 (1989).
Hickman, G.T., J.T.Novak, M.S.Morris, and M. Rebhun. Effects of site variations on subsurface biodegradation potential.
Research Journal of the Water Pollution Control Federation 61 (9): 1564-1575 (1989).
Hristova, K., B. Gebreyesus, D. Mackay, and K. M. Scow. Naturally occurring bacteria similar to the methyl tert-butyl ether
(MTBE) -degrading strain PM1 are present in MTBE-contaminated groundwater. Applied and Environmental Microbiology
69 (5): 2616-2623 (2003).
Hubbard, C. E., J. F. Barker, S. F. O'Hannesin, M. Vandegriendt, and R. W Gillham. Transport and fate of dissolved methanol,methyl-
tertiary-butyl ether, and monoaromatic hydrocarbons in a shallow sand aquifer. Health & Environmental Sciences Department,
American Petroleum Institute 4601. Washington, DC, 1994
Hunkeler, D., B. J. Butler, R. Aravena, and J. F. Barker. Monitoring biodegradation of methyl tert-butyl ether (MTBE) using com-
pound-specific carbon isotope analysis. Environmental Science & Technology 35 (4): 676-681 (2001).
Huskey, W.P. "Origins and interpretations of heavy atom isotope effects." In Enzyme Mechanism from Isotope Effects, Boca
Raton, FL: CRC Press 1991, 37-72.
Hyman, M., C. Taylor, and K. O'Reilly. "Cometabolic degradation of MTBE by wo-alkane-utilizing bacteria from gasoline-im-
pacted soils. Bioremediation and phytoremediation of chlorinated and recalcitrant compounds." In: The Second International
Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, 2000.
Jensen, H. M. and E. Arvin. "Solubility and degradability of the gasoline additive MTBE, methyl tert-butyl ether and gasoline
compounds in water." In Contaminated Soil '90, Dordrecht, the Netherlands, 1990.
Kane, S. R., H. R. Beller, T. C. Legler, C. J. Koester, H. C. Pinkart, R. U. Halden, and A. M. Happel. Aerobic biodegradation of
methyl tert-butyl ether by aquifer bacteria from leaking underground storage tank sites. Applied and Environmental Micro-
biology. 67 (12): 5824-5829 (2001).
Kane, S. R., T.C. Legler, and L. M. Balser. "Aerobic biodegradation of MTBE by aquifer bacteria from LUFT sites." In Proceed-
ings of the Seventh International In Situ and On-Site Bioremediation Symposium, Orlando, FL, 2003.
Kolhatkar, R., J. T. Wilson, and L.E. Dunlap. "Biodegradation of MTBE at multiple UST sites." In Proceedings of the Petroleum
and Organic Chemicals in Ground Water: Prevention, Detection and Remediation Conference & Exposition, Anaheim, CA,
2000.
Kolhatkar, R, J. T. Wilson, and G. Hinshalwood. "Natural biodegradation of MTBE at a site on Long Island, NY." In Proceedings
of the Sixth International In Situ and On-Site Bioremediation Symposium, San Diego, CA, 2001.
Kolhatkar, R., T. Kuder, P. Philp, J. Allen, and J. T. Wilson. Use of compound-specific stable carbon isotope analyses to demon-
strate anaerobic biodegradation of MTBE in groundwater at gasoline release site. Environmental Science & Technology 36
(23): 5139-5146 (2002).
Kramer, W. H. and T. L. Douthit. "Water soluble phase oxygenates in gasoline from five New Jersey service stations." In: Pro-
ceedings of the Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation
Conference and Exposition, Anaheim, CA, 2000.
Kuder, T., R. P. Philp, R. Kolhatkar, J. T. Wilson, and J. Allen. "Application of stable carbon and hydrogen isotopic techniques for
monitoring biodegradation of MTBE in the field. In Proceedings of the 2002 Petroleum Hydrocarbons and Organic Chemicals
in Ground Water: Prevention, Assessment, and Remediation, Conference and Exposition, Atlanta, GA, 2002.
Kuder, T., R. Kolhatkar, J. Wilson, K. O'Reilly, P. Philp, and J. Allen. "Compound-specific carbon and hydrogen isotope analysis-
field evidence of MTBE bioremediation." In Proceedings of the 2003 National Ground Water Association Focus Conference
on MTBE: Assessment, Remediation, and Public Policy, Baltimore, MD, 2003.
Kuder, T., J. T. Wilson, P. Kaiser, R. Kolhatkar, P. Philp, and J. Allen. Enrichment of stable carbon and hydrogen isotopes during
anaerobic biodegradation of MTBE-microcosm and field evidence. Environmental Science & Technology 39 (1): 213-220
(2005).
70
-------
Kuder, T., R. Kolhatkar, J. Wilson, P. Philp, and J. Allen. "Compound-specific isotope analysis of MTBE and TEA for bioreme-
diation studies. In Proceedings of the 2003 Natioinal Ground Water Association Focus Conference on MTBE: Assessment,
Remediation, and Public Policy, Baltimore, MD, 2004.
Landmeyer, J. E., J. E Pankow, and C. D. Church . "Occurrence of MTBE and tert-butyl alcohol in a gasoline-contaminated
aquifer." Paper presented at the Division of Environmental Chemistry, American Chemical Society, San Francisco, CA, April
13-17, 1997.
Landmeyer, J. E., E H. Chapelle, P. M. Bradley, J. E Pankow, C. D. Church, and P. G. Tratnyek. Fate of MTBE relative to benzene
in a gasoline-contaminated aquifer (1993-98), Ground Water Monitoring and Remediation 18 (4): 93-102 (1998).
Landmeyer, J. E., E H. Chapelle, H. H. Herlong, and P. M. Bradley. Methyl tert-butyl ether biodegradation by indigenous aqui-
fer microorganisms under natural and artificial oxic conditions. Environmental Science and Technology 35 (6): 1118-1126
(2001).
Lin, Z., J. T. Wilson, and D. D. Fine. Avoiding hydrolysis of fuel ether oxygenates during static headspace analysis. Environmental
Science & Technology 37 (21): 4994-5000 (2003).
McLoughlin, P. W, R. J. Pirkle, D. Fine, and J. T. Wilson. Production by acid hydrolysis of MTBE during heated headspace
analysis and evaluation of a base as a preservative. Ground Water Monitoring and Remediation 24 (4): 57-66 (2004).
Mace, R. E., R. S. Fisher, D. M. Welch, and S.P. Parra. Extent, mass, and duration of hydrocarbonplutnesfrotnleaking petroleum
storage tank sites in Texas. Geological Circular 97-1, Bureau of Economic Geology, University of Texas, Austin, TX, 1997.
Mace, R. E. and W. Choi. "The size and behavior of MTBE plumes in Texas." In Proceedings of the Petroleum Hydrocarbons
and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation Conference, API and NGWA Conference
and Exposition, Houston, TX, 1998.
Mariotti, A., J. C. Germon, P. Hubert, P. Kaiser, R. Letolle, A. Tardieux, P. Tardieux. Experimental determination of nitrogen
kinetic isotope fractionation: some principles; illustration for the denitrification and nitrification processes. Plant and Soil 62
(3): 413-430 (1981).
Mo, K., C. O. Lora, A. E. Wanken, M. Javanmardian, X. Yang, and C. E Kulpa. Biodegradation of methyl t-butyl ether by
pure bacterial cultures. Applied Microbiology and Biotechnology 47: 69-72 (1997).
Mormille, M. R., S. Liu, and J. M. Suflita. Anaerobic biodegradation of gasoline oxygenates: Extrapolation of information to
multiple sites and redox conditions. Environmental Science and Technology 28(9): 1727-1732(1994).
National Research Council. Natural Attenuation for Groundwater Remediation. Washington, DC: National Academy Press,
2000.
New England Interstate Water Pollution Control Commission. Summary report on a survey of state experiences with MTBE and
other oxygenate contamination at LUST sites. August 2000, 2003.
Newell, C. J., R. K. McLeod, and J. R. Gonzales. BIOSCREEN natural attenuation decision support system, EPA/600/R-96/087,
Cincinnati, OH: U.S. Environmental Protection Agency, 1996.
Newell, C. J., H. S. Rifai, J. T. Wilson, J. A. Connor, J. A. Aziz, and M. P. Suarez. Calculation and use of first-order rate con-
stants for monitored natural attenuation studies, EPA/540/S-02/500, Cincinnati, OH: U.S. Environmental Protection Agency,
2002.
Novak, J. T., C. D. Goldsmith, R. E. Benoit, and J. H. O'Brien. Biodegradation of methanol and tertiary butyl alcohol in subsurface
systems. Water Science Technology 17 (9): 71-85 (1985).
Odencrantz, J. E. "Implication of MTBE for intrinsic remediation of underground fuel tank sites." In: Proceedings of the Petroleum
Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation Conference. API, NGWA,
STEP Conference and Exposition, Houston, TX, 1998.
O' Sullivan, G., G. Boshoff, A. Downey, and R. M. Kalin. "Carbon isotope effect during the abiotic oxidation of methyl-tert-butyl ether
(MTBE). In Proceedings of the Seventh International In Situ and On-Site Bioremediation Symposium, Orlando, FL, 2003.
O'Reilly, K. T., M. E. Moir, C. D. Taylor, C. A. Smith, and M. R. Hyman. Hydrolysis of tert-butyl methyl ether (MTBE) in dilute
aqueous acid. Environmental Science & Technology 35 (19): 3954-3961 (2001).
Park, K. and R. M. Cowan, R.M. "Effects of oxygen and temperature on the biodegradation of MTBE." In Proceedings of the
213th ACS National Meeting, Division of Environmental Chemistry, San Francisco, CA, 1997.
Peargin, T. R. "Relative depletion rates of MTBE, benzene, and xylene from smear zone NAPL. In Proceedings of the 2000 Petro-
leum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation. Special Focus: Natural
Attenuation and Gasoline Oxygenates. API, NGWA, STEP Conference and Exposition, Anaheim, CA, 2000.
Peargin, T. R. "Relative depletion rates of MTBE, benzene, and xylene from smear zone non-aqueous phase liquid. In Proceedings
of the Sixth International In Situ and On-Site Bioremediation Symposium, San Diego, CA, 2001.
71
-------
Pope, D. E, S. D. Acree, H. Levine, S. Mangion, J. van Ee, K. Hurt, and B. Wilson. Performance monitoring ofMNA remedies for
VOCs in ground water, EPA/600/R-04/027. Cincinnati, OH: U.S. Environmental Protection Agency, 2004.
Rittmann, B. E. "Monitored natural attenuation of MTBE. In MTBE Remediation Handbook. Amherst. MA: Amherst Scientific
Publishers, 2003, 329-348.
Rixey, W. G. and S. Joshi. Dissolution of MTBE from a residually trapped gasoline source, API Soil and Groundwater Research
Bulletin No. 13, American Petroleum Institute, 2000.
Robb, Joseph and Ellen E. Moyer. "Natural attenuation of benzene and MTBE at four midwestern U.S. sites." In MTBE Reme-
diation Handbook, Amherst, MA: Amherst Scientific Publishers, 2003, 561-578.
Salinitro, J. P., L. A. Diaz, M. P. Williams, and H. L. Wisniewski. Isolation of a bacterial culture that degrades methyl t-butyl ether.
Applied and Environmental Microbiology 60: 2593-2596 (1994).
Salinitro, J. P., C. S. Chou, H. L. Wisniewski, and T. E. Vipond. "Perspectives on MTBE biodegradation and the potential for in
situ aquifer bioremediation." In Proceesings of the Southwest Ground Water Conference—Discussing the Issue of MTBE and
Percholorate in Ground Water, Anaheim, CA, 1998.
Salanitro, J. P., P. C. Johnson, G. E. Spinnler, P. M. Maner, H. L. Wisniewski, and C. Bruce. Field-scale demonstration of en-
hanced MTBE bioremediation through aquifer bioaugmentation and oxygenation. Environmental Science and Technology 34
(19): 4152-4162(2000).
Schirmer, M. and J. E Barker. A study of long-term MTBE attenuation in the Borden Aquifer, Ontario, Canada. Ground Water
Monitoring and Remediation 18(2): 113-122(1998).
Schirmer, M., B. J. Butler, J. E Barker, C. D. Church, and K. Schirmer, K. Evaluation of biodegradation and dispersion as natural
attenuation processes of MTBE and benzene at the Borden field site. Physics and Chemistry of the Earth, Part B: Hydrology,
Oceans and Atmosphere 24 (6): 557-560 (1999).
Schmidt, T., M. Schirmer, H. Weiss, and S. B. Haderlein. Microbial degradation of methyl tert-butyl ether and tert-butyl alcohol
in the subsurface. Journal of Contaminant Hydrology 70: 173-203 (2004).
Shih, T., Y. Rong, T. Harmon, and M. Suffet. Evaluation of the impact of fuel hydrocarbons and oxygenates on groundwater re-
sources. Environmental Science and Technology 38 (1): 42-48 (2004).
Shorr, G. L. and H. S. Rifai. "Characterizing the intrinsic remediation of MTBE at field sites in Texas." In Proceedings of the Third
International Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, 2002.
Small, M. C. and J. Weaver. "An updated conceptual model for subsurface fate and transport of MTBE and benzene." In Pro-
ceedings of the Petroleum Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and Remediation
Conference. API, NGWA, STEP Conference and Exposition, Houston, TX, 1999.
Smallwood, B. J., R. P. Philp, T. W. Burgoyne, and J. D. Allen. The use of stable isotopes to differentiate specific source markers
for MTBE. Environmental Forensics 2 (3): 215-221 (2001).
Smith, C. A., K. T. O'Reilley, and M. R. Hyman. Characterization of the initial reactions during the cometabolic oxidation of
methyl tert-butyl ether by propane-grown Mycobacterium vaccae JOBS. Applied and Environmental Microbiology 69 (2):
796-804 (2003).
Somsamak, P., R. M. Cowan, and M. M. Haggblom. Anaerobic biotransformation of fuel oxygenates under sulfate-reducing condi-
tions. FEMS Microbiology Ecology 37: 259-264 (2001).
Somsamak, P., H. H. Richnow, and M. M. Haggblom. Carbon isotope fractionation during anaerobic biotransformation of methyl
tert-butyl ether and tert-amyl methyl ether. Environmental Science and Technology 39 (1): 103-109 (2005).
Squillace, P. J., J. E Pankow, N. E. Korte, and J. S. Zorgorski. Review of the environmental behavior and fate of methyl tert-buty
Ether. Environmental Toxicology and Chemistry 16 (9): 1836-1844 (1997).
Srinivasan, P., D. E Pope, and E. Striz. Optimal Well Locator (OWL): A screening tool to evaluating locations of monitoring wells,
EPA 600/C-04/017. Cincinnati, OH: U.S. Environmental Protection Agency, 2004.
Steffan, R. J., K. McClay, S. Vainberg, C. W. Condee, and D. Zhang. Biodegradation of the gasoline oxygenates methyl tert-butyl
ether, ethyl tert-butyl ether, and tert-amyl ether by propane-oxidizing bacteria. Applied and Environmental Microbiology 63
(11): 4216-4222 (1997).
Steffan, R. J., S. Vainberg, C. Condee, K. McClay, and P. Hatzinger. "Biotreatment of MTBE with a new bacterial isolate. Bio-
remediation and phytoremediation of chlorinated and recalcitrant compounds." In Proceedings of the Second International
Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, 2000.
Suarez, M. P. and H. S. Rifai. Biodegradation rates for fuel hydrocarbons and chlorinated solvents in groundwater. Bioremedia-
tion Journal 3 (4): 337-362 (1999).
Suflita, J. M. and M. R. Mormile. Anaerobic biodegradation of known and potential gasoline oxygenates in the terrestrial subsur-
face. Environmental Science and Technology 27 (5): 976-978 (1993).
72
-------
United States Environmental Protection Agency. Use of monitored natural attenuation at superfund, RCRA corrective action, and
underground storage tank sites. Office of Solid Waste and Emergency Response, Directive 9200.4-17P (1999).
Varadhan, R., Jin-Song Chen, J. T. Wilson, J. A. Johnson, W. Gierke, andL. Murdie. Evaluation of natural attenuation of benzene
and dichloroethanes at the KL Landfill. Bioremediation Journal 2 (3-4): 239-258 (1998).
White, H., B. Lesnik, and J. Wilson. Analytical methods for fuel oxygenates. L.U.S.T.LINE: A Report on Federal and State Pro-
grams to Control Leaking Underground Storage Tanks. New England Interstate Water Pollution Control Commission Bulletin
42: 1-8(2002).
Wiedemeier, T. H., R. N. Miller, J. T. Wilson, and D. H. Kampbell. "Significance of anaerobic processes for the intrinsic biore-
mediation of fuel hydrocarbons. In Proceedings of the Petroleum Hydrocarbons and Organic Chemicals in Ground Water:
Prevention, Detection, and Remediation Conference: NWWA/API, Houston, TX, 1995.
Wiedemeier, T. H., M. A. Swanson, J.T.Wilson, D. H. Kampbell, R. N. Miller, and J. E. Hansen. Approximation of biodeg-
radation rates constant for monoaromatic hydrocarbons (BTEX) in ground water. Ground Water Monitoring & Remediation
16 (3): 186-194 (1996).
Wilkin, R. T., J. T. Wilson, M. S. McNeil, and C. J. Adair. Field measurement of dissolved oxygen: A comparison of methods..
Ground Water Monitoring and Remediation 21 (4): 124-132(2001).
Wilson, J. T., J. A. Vardy, J. S. Cho, and B. H. Wilson. Natural attenuation ofMTBE in the subsurface under methanogenic
conditions, EPA/600/R-00/006. Cincinnati, OH: U.S. Environmental Protection Agency, 2000.
Wilson, J. T., and R. Kolhatkar. Role of natural attenuation in the life cycle ofMTBE plumes. Journal of Environmental Engineer-
ing 128 (9): 876-882 (2002).
Wilson, J. T. "Fate and transport of MTBE and other gasoline components." In: MTBE Remediation Handbook, Amherst, MA:
Amherst Scientific Publishers, 2003a, 19-61.
Wilson, J. T. "Aerobic in situ bioremediation." In: MTBE Remediation Handbook. Amherst, MA: Amherst Scientific Publishers,
2003b, 243-264.
Wilson, J. T., C. Adair, P. M. Kaiser, and R. Kolhatkar. Role of anaerobic biodegradation in the natural attenuation ofMTBE at a
gasoline spill site. Ground Water Monitoring & Remediation In Press (2005a).
Wilson, J. T, R. Kolhatkar, T. Kuder, P. Philp, and S. J. Daugherty. Anaerobic degradation ofMTBE to TEA in ground water at
gasoline spill sites in Orange County, California. Ground Water Monitoring & Remediation (In review) (2005b).
Wilson, R. D., D. M. Mackay, and K. M. Scow. In situ MTBE biodegradation supported by diffusive oxygen release. Environmental
Science & Technology 36 (2): 190-199 (2002).
Yeh, C. K. and J. T. Novak. Anaerobic biodegradation of gasoline oxygenates in soils. Water Environment Research 66 (5): 744-
752 (1994).
73
-------
SEPA
United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
EPA/600/R-04/179
January 2005
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERED;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Recycled/Recyclable
Printed with vegetable based in on
paper that contains a minimum of
5O% post-consumer fiber content
processed chlorine free
------- |