f/EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington DC 20460
EPA/600/R-00/006
January 2000
Natural Attenuation of
MTBE in the
Subsurface under
Methanogenic Conditions
-------
EPA/600/R-00/006
January 2000
Natural Attenuation of MTBE in the
Subsurface under
Methanogenic Conditions
John T. Wilson
Jong Soo Cho
Barbara H. Wilson
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
Ada, Oklahoma 74820
James A. Vardy
Civil Engineering Branch
United States Coast Guard
Cleveland, Ohio 41199-2060
Project Officer
John T. Wilson
Subsurface Protection and Remediation 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
This work was carried out by staff of the U.S. Environmental Protection Agency (Office of Research and
Development, National Risk Management Research Laboratory) and by staff of the U.S. Coast Guard, in a
collaboration funded in part under Interagency Agreement # RW-69-937352. 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 or commercial products does not constitute endorsement or recommendation for use. Certain
samples were collected or analyzed by employees of ManTech Environmental Research Services Corp., an in-
house contractor to the U.S. Environmental Protection Agency.
Representations and interpretations of the behavior of methyl-tertiary-butyl-ether (MTBE) or tertiary-butyl
alcohol (TEA) apply only to the site of the case study presented in this report. The authors, the U.S.
Environmental Protection Agency, and the U.S. Coast Guard make no claim in this report concerning the
behavior of methyl-tertiary-butyl-ether (MTBE) or tertiary-butyl alcohol (TEA) at other sites.
All research projects making conclusions or recommendations based on environmentally related measure-
ments and funded by the Environmental Protection Agency are required to participate in the Agency Quality
Assurance Program. This project was conducted under an approved Quality Assurance Project Plan. The
procedures specified in this plan were used without exception. Information on the plan and documentation of the
quality assurance activities and results are available from John T. Wilson.
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress to protect 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 these mandates, EPA's research program is providing data and technical support for solving
environmental problems of today and building a science knowledge base necessary to manage our ecological
resources wisely, understand how pollutants affect our health, and prevent or reduce environmental risks in the
future.
The National Risk Management Research Laboratory is the Agency's center for investigation of technological
and management approaches for reducing risks from threats to human health and the environment. The focus of the
laboratory's research program is on methods for the prevention and control of pollution to air, land, water, and
subsurface resources; protection of water quality in public water systems; remediation of contaminated sites and
ground water; and prevention and control of indoor air pollution. The goal of this research effort is to catalyze
development and implementation of innovative, cost-effective environmental technologies; develop scientific and
engineering information needed by EPA to support regulatory and policy decisions; and provide technical support
and information transfer to ensure effective implementation of environmental regulations and strategies.
The U.S. Environmental Protection Agency's Office of Underground Storage Tanks uses a risk management
approach to protect ground water from contamination with the soluble components of fuels that are accidentally
spilled or released from underground storage tanks. Contamination of ground water with MTBE and TEA associated
with spills from underground storage tanks is an emerging problem in the United States. Little is known of the
prospects for biodegradation of MTBE and TEA in ground water. Consistent with the Agency's goal of sound science
as a basis for risk management, the Subsurface Protection and Remediation Division is developing information on
the rate and extent of natural attenuation of MTBE and TEA in ground water. This research effort emphasizes
natural biodegradation under various geochemical environments. This report describes natural attenuation of MTBE
under methanogenic conditions. It is the first in a series of reports; subsequent reports will examine natural
attenuation of MTBE under aerobic conditions, and under sulfate-reducing and iron-reducing conditions.
Clinton W. Hall, Director
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
-------
Abstract
At many fuel spill sites, the spread of contamination from benzene, toluene, ethylbenzene, and the xylenes
(BTEX compounds) is limited by natural biodegradation of the petroleum hydrocarbons in the ground water. At
present there is much uncertainty about whether MTBE from fuel spills will follow the same pattern as the petroleum-
derived hydrocarbons, or whether MTBE is biologically recalcitrant in ground water. If MTBE does not biodegrade
in ground water, then dilution and dispersion are the only mechanisms that are available to attenuate MTBE. As a
consequence, plumes of MTBE could expand farther than plumes of benzene or the BTEX compounds in the
absence of biodegradation.
This case study was conducted at the former Fuel Farm Site at the U.S. Coast Guard Support Center at
Elizabeth City, North Carolina. The geochemistry of the site is typical of sites where natural biodegradation limits the
spread of BTEX compounds. The plume is undergoing extensive anaerobic oxidation of petroleum hydrocarbons, as
well as fermentation of hydrocarbons to methane. The hydrocarbon metabolism through sulfate and iron oxidation
is approximately equivalent to the hydrocarbon metabolism through methanogenesis. The amount of hydrocarbon
metabolized through anaerobic pathways is about ten times the amount degraded with molecular oxygen.
There are two laboratory studies in the literature that report the biotransformation of MTBE in aquifer material
under methanogenic conditions. Neither study included an evaluation of the field-scale performance of natural
attenuation. This case study is intended to answer the following questions: Can MTBE be biodegraded under
methanogenic conditions in ground water that was contaminated by a fuel spill? Will biodegradation produce
concentrations of MTBE that are less than regulatory standards? Is the rate of degradation in the laboratory
adequate to explain the distribution of MTBE in the ground water at the field site? What is the relationship between
the degradation of MTBE and degradation of the BTEX compounds? What is the rate of natural attenuation of the
source area?
The apparent first order rate of removal of MTBE in the field was a sensitive function of ground-water seepage
velocity. The rate of removal was calculated for an upper boundary on velocity, an average velocity, and a lower
boundary on velocity. The rate was 5.0 per year at the upper boundary; 2.7 per year at the average velocity, and
2.2 per year at the lower boundary. Methane was considered to be a conservative tracer of ground-water flow at the
site. The apparent rate of removal of methane was taken as an estimate of attenuation along the flow path due to
dilution and dispersion. The apparent first order rate of removal of methane at the average estimate of seepage
velocity was 0.50 +/- 0.65 per year.
Biodegradation was evaluated in laboratory microcosms that were constructed with material from the contami-
nated portion of the aquifer. After 490 days of incubation, the average concentration of MTBE remaining in six
replicates of a treatment that was supplemented with BTEX compounds was 81 ug/l, compared to 5680 ug/l at the
beginning of incubation. The average concentration remaining in the control treatment after 490 days was 1470 ug/l,
compared to 3330 ug/l at the beginning of incubation. MTBE was also removed in microcosms that were not
supplemented with alkylbenzenes. After 490 days of incubation, the concentration of MTBE in all six of the replicate
microcosms that were sampled was below 40 ug/l, compared to 3110 ug/l at the beginning of incubation. Removal
of MTBE in the microcosms did not require the presence of BTEX compounds. The removal of MTBE did not begin
until the removal of the BTEX compounds was complete.
The first order rate of removal of MTBE in microcosms supplemented with alkylbenzenes was 3.02 per year +/
- 0.52 per year at 95% confidence. Removal in the corresponding controls was 0.39 +/- 0.19 per year at 95%
confidence. The removal in the microcosms without added alkylbenzenes was 3.5 per year+/-0.65 per year at 95%
confidence. Removal in the corresponding controls was 0.30 per year+/-0.14 per year at 95% confidence. The rate
of removal of MTBE in the laboratory studies can explain the apparent attenuation of MTBE at field scale.
The rate of natural attenuation of the source area was evaluated by comparing that flux to the total mass of
MTBE in the source area. The mass transfer of MTBE from the source LNAPL to the ground water moving
underneath was estimated by calculating the flux of MTBE moving away from the source, then dividing the flux into
the quantity of MTBE remaining. The flux of MTBE away from the source area in 1996 was 2.76 kg/year. The lower
boundary on the total quantity of MTBE in the source area was 46 kg. If the rate of transfer of MTBE to ground water
is proportional to the amount of MTBE in the source, the instantaneous rate of transfer is 0.06 per year. The average
concentration at the most contaminated location in the transect is 1200 ug/l. At this rate of attenuation of the source,
it would require at least sixty years for the concentration to reach 30 ug/l.
Tertiary Butyl Alcohol (TEA) has been documented as a transformation product of MTBE in a number of studies.
At the Old Fuel Farm Site, there is no evidence of accumulation of TEA in the ground-water plume as a whole. With
two exceptions, the concentration of TEA in ground water downgradient of the source area was less than 200 ug/l.
Ground water from a location immediately downgradient of the source area had a higher concentration of TEA, near
2000 ug/l. In this sample there was a corresponding reduction in the concentration of MTBE. At this location the TEA
was probably produced from transformation of MTBE.
iv
-------
Contents
Notice ii
Foreword iii
Abstract iv
Figures vi
Tables viii
Acknowledgments ix
Section 1 Introduction 1
Evidence for Biodegradation of MTBE under Anaerobic Conditions in Ground Water 1
Purpose of the Case Study 2
Section 2 Laboratory Studies 5
Construction, Sampling and Analysis of Microcosms 5
Removal of MTBE 6
Removal of Benzene, Toluene, and Ethylbenzene 7
Relationship between removal of BTEX compounds and removal of MTBE 10
Section 3 Site Characterization 11
Site Description and History 11
Core Sampling the Source Area 11
Estimation of Total Quantity of TPH and MTBE and the Area Impacted 13
Vertical Distribution of TPH and MTBE in Core Samples 13
Distribution of Total Petroleum Hydrocarbons and Hydraulic Conductivity with Depth 15
Distribution of MTBE and BTEX Compounds with Depth 16
Section 4Transport and Fate of MTBE in the Ground Water 23
Estimated Rate of Attenuation in Ground Water 23
Transfer from the Entire Source Area to the Plume 27
Transfer of MTBE to TBE 28
Section 5 Summary and Conclusions 31
Extent of Biodegradation of MTBE 31
Role of BTEX Compounds 31
Rate of Removal of MTBE 31
Expected Persistence of the Source of Ground-water Contamination 31
Production and Depletion of TEA 31
Geochemical Context of the Plume that Biodegraded MTBE 32
References 33
Appendix A: Temporal Variation in the Hydraulic Gradient and the Direction of Ground Water Flow 34
Appendix B: Geochemical Context of the MTBE Plume 45
-------
Figures
Figure 1.1 Site selected for the case study of natural attenuation of MTBE under methanogenic conditions. The shaded
concentric circles represent the residual LNAPL from a fuel spill. The arrow represents the distance traveled
by ground water in three years 4
Figure 2.1 Removal of MTBE in microcosms constructed with MTBE and BTEX compounds compared to removal in
control microcosms that were autoclaved to prevent biotransformation of MTBE 7
Figure 2.2 Removal of MTBE in microcosms constructed with MTBE but without supplemental concentrations of BTEX
compounds compared to removal in control microcosms that were autoclaved to prevent biotransformation of MTBE 8
Figure 2.3 Removal of Toluene in microcosms constructed with MTBE and BTEX compounds compared to removal in
control microcosms that were autoclaved to prevent biotransformation of toluene. The solid line is fit through the
removal in the controls 8
Figure 2.4 Removal of Benzene in microcosms constructed with MTBE and BTEX compounds compared to removal in
control microcosms that were autoclaved to prevent biotransformation of benzene. The solid line is fit through the
removal in the controls 9
Figure 2.5 Removal of Ethylbenzene in microcosms constructed with MTBE and BTEX compounds compared to removal in
control microcosms that were autoclaved to prevent biotransformation of ethylbenzene 9
Figure 2.6 Comparison of the time lags for removal of MTBE, and of benzene, toluene, and ethylbenzene in microcosms
constructed with all the compounds present together 10
Figure 3.1 Relationship between the sampling locations for characterization of the LNAPL source area (labeled CPT-1
through CPT-5), and the former location of storage tanks for fuels 12
Figure 3.2 Inferred location of the fuel release, based on vertical core samples and the location of the steel underground
storage tanks 14
Figure 3.3 Vertical distribution of MTBE and Total Petroleum Hydrocarbon (TPH) in core samples at location CPT-1
(See Figure 3.1 for map) 15
Figure 3.4 Vertical distribution of MTBE and Total Petroleum Hydrocarbon (TPH) in core samples from location CPT-2
(See Figure 3.1 for map) 15
Figure 3.5 Vertical distribution of MTBE and Total Petroleum Hydrocarbon (TPH) in core samples from location CPT-3
(See Figure 3.1 for map) 15
Figure 3.6 Relationship between the vertical extent of Hydraulic Conductivity and the vertical extent of Total Petroleum
Hydrocarbon at location CPT-1 16
Figure 3.7 Relationship between two transects of ground-water samples and the fuel release. The arrow represents
the average direction of ground-water flow 17
Figure 3.8 Location of vertical sampling points along the north-south transect, collected in August 1996. Distance
along the transect extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 19
Figure 3.9 Location of vertical sampling points along the east-west transect, collected in December 1997. Distance
along the transect extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow 19
Figure 3.10 Distribution of hydraulic conductivity along the north-south transect, collected in August 1996. Distance along
the transect extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 20
Figure 3.11 Distribution of hydraulic conductivity along the east-west transect, collected in December 1997. Distance
along the transect extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow 20
Figure 3.12 Distribution of MTBE along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 21
Figure 3.13 Distribution of MTBE along the east-west transect, collected in December 1997. Distance along the transect
extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow 21
Figure 3.14 Distribution of BTEX along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 22
Figure 3.15 Distribution of BTEX along the east-west transect, collected in December 1997. Distance along the transect
extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow 22
Figure 4.1 Variation in ground-water flow calculated from eighteen rounds of quarterly monitoring. The length of the
arrow is the distance that would be traveled by MTBE in one year at that hydraulic gradient 24
Figure 4.2 Variation in ground-water flow calculated from fourteen rounds of monthly monitoring. The length of the
arrow is the distance that would be traveled by MTBE in one year at that hydraulic gradient 24
Figure 4.3 Attenuation in concentrations of MTBE, methane, and iron (II) with travel time downgradient from the
location with the highest concentration of MTBE 26
Figure 4.4 Locations of ground-water samples included in the calculation of the rate of natural attenuation. The arrow
represents the average direction of ground-water flow. The dark shaded area is the area with LNAPL.
The larger lightly shaded area is the area downgradient where the ground water contains high
concentrations of methane and iron (II). Only wells in the shaded area were included in the calculation of the
rate of natural attenuation 27
VI
-------
Figure 4.5 Relationship between the direction of ground-water flow and the ground-water sampling locations in the
transect sampled in December 1997. Ground-water flow vectors were calculated from the gradients in
water table elevation in eighteen different rounds of monitoring The length of arrow is the distance that
would be traveled by MTBE in one year of flow at that gradient 27
Figure 4.6 Concentrations of MTBE in a transect that extends across the plume in a direction that is roughly
perpendicular to ground-water flow. See Figure 4.5 for the positions of the sampling locations identified as
15 through 25 in both Figures. Depicted at each location are the flow-weighted average concentrations of
MTBE in ground-water samples from a vertical profile extending across the aquifer at each location 28
Figure 4.7 Relationship between the concentration of TBA in ground water and the concentration of MTBE, in water
samples collected in a transect across the plume in December 1997 29
Figure 4.8 Depth distribution of MTBE in three locations downgradient of the LNAPL source area. See Figure 4.5 for
position of the locations on a map. Compare location 19 to location 19 in Figure 4.9 30
Figure 4.9 Depth distribution of TBa in locations downgradient of the source area. See Figure 4.5 for position of the
locations on a map. Compare location 19 to location 19 in Figure 4.8 30
Figure A.1 Variation in elevation of water in the Pasquotank River over a time interval extending from
Septembers, 1996 to October 30, 1996 34
Figure A.2 Location of the permanent monitoring wells used to estimate the hydraulic gradient and direction during each
round of monitoring 35
Figure A.3 Variation in elevation of the water table at the fuel farm site over time. Consult Figure A.1 for the
location of the monitoring wells. Well ESM-10 is closest to the Pasquotank River, the point of ground-water
discharge. Wells ESM-14, ESM-6, and ESM-7 are farther inland 35
Figure A.4 Direction and gradient of ground-water flow on a sample date in September 1994 40
Figure A.5 Direction and gradient of ground-water flow on a sample date in December 1994 40
Figure A.6 Direction and gradient of ground-water flow on a sample date in March 1995 41
Figure A.7 Direction and gradient of ground-water flow on a sample date in May 1995 41
Figure A.8 Direction and gradient of ground-water flow on a sample date in August 1995 41
Figure A.9 Direction and gradient of ground-water flow on a sample date in December 1995 41
Figure A.10 Direction and gradient of ground-water flow on a sample date in March 1996 42
Figure A. 11 Direction and gradient of ground-water flow on a sample date in June 1996 42
Figure A.12 Direction and gradient of ground-water flow on a sample date in September 1996 42
Figure A.13 Direction and gradient of ground-water flow on a sample date in December 1996 42
Figure A.14 Direction and gradient of ground-water flow on a sample date in March 1997 43
Figure A.15 Direction and gradient of ground-water flow on a sample date in June 1997 43
Figure A.16 Direction and gradient of ground-water flow on a sample date in September 1997 43
Figure A.17 Direction and gradient of ground-water flow on a sample date in December 1997 43
Figure A.18 Direction and gradient of ground-water flow on a sample date in March 1998 44
Figure A.19 Direction and gradient of ground-water flow on a sample date in June 1998 44
Figure A.20 Direction and gradient of ground-water flow on a sample date in September 1998 44
Figure A.21 Direction and gradient of ground-water flow on a sample date in December 1998 44
Figure B.1 Distribution of methane along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 46
Figure B.2 Distribution of MTBE along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 46
Figure B.3 Distribution of methane along the east-west transect, collected in December 1997. Distance along the transect
extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow 47
Figure B.4 Distribution of MTBE along the east-west transect, collected in December 1997. Distance along the transect
extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow 47
Figure B.5 Distribution of oxygen along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 48
Figure B.6 Distribution of sulfate along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 48
Figure B.7 Distribution of iron (II) along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 49
Figure B.8 Distribution of alkalinity along the north-south transect, collected in August 1996. Distance along the transect
extends from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow 49
VII
-------
Tables
Table 1.1 Temporal variation in the concentrations of MTBE, Benzene, and Methane at the most contaminated permanent
sampling location that is downgradient of the LNAPL area 3
Table 2.1 The concentration of MTBE and alkylbenzenes in the most contaminated sample of ground water from the LNAPL
source area, in the permanent monitoring well at the location where the sediment used to construct the microcosms
was acquired, and the initial concentrations achieved in the microcosms 6
Table 3.1 Quantity of Total Petroleum Hydrocarbon and MTBE at seven sampling locations in or near the point of release of fuel 13
Table 3.2 Distribution of Hydraulic Conductivity (K) in the North-South transect sampled in August, 1996 (Figure 3.7) 18
Table 3.3 Distribution of Hydraulic Conductivity (K) in the East-West transect sampled in December, 1997 (Figure 3.7) 18
Table 4.1 Sensitivity analysis of the estimates of the seepage velocity of ground water at the site. These estimates were
used to calculate a first order rate of attenuation of MTBE in ground water downgradient of the source area 24
Table 4.2 Concentration of MTBE, methane, and iron (II) at monitoring locations used to calculate the rate of attenuation of
MTBE, methane, and iron (II) with time of travel downgradient of the location with the highest concentration 25
Table 4.3 The apparent first order rate of attenuation of MTBE, methane, and iron (II) with time of travel downgradient from the
location with the highest concentration of MTBE 26
Table A.1. Elevation of the water table in permanent monitoring wells during eighteen rounds of quarterly
monitoring extending from September 1994 through December 1998. The elevations are reported in
feet above mean sea level. Compare Figure A.1 for the location of the monitoring wells 37
Table A.2. Elevation of the water table in permanent monitoring wells during fourteen rounds of monthly
monitoring extending from September 1994 through December 1998. The elevations are reported in
feet above mean sea level. Compare Figure A.1 for the location of the monitoring wells 37
Table A.3. Equation of a linear plane that was fit using a least-squares regression through the elevation of the
water table in permanent monitoring wells during each of eighteen rounds of quarterly sampling. The
plane is in an x,y,z coordinate system where x increases toward the east, y increases toward the north,
and z increases with elevation above mean sea level. The equation is in the form Ax+By+C+z where
x and y are the grid location in UTM meters and z is the elevation of the water table in feet 38
Table A.4. Equation of a linear plane that was fit using a least-squares regression through the elevation of the
water table in permanent monitoring wells during each of fourteen rounds of monthly sampling. The
plane is in an x,y,z coordinate system where x increases toward the east, y increases toward the north,
and z increases with elevation above mean sea level. The equation is in the form Ax+By+C+z where
x and y are the grid location in UTM meters and z is the elevation of the water table in feet 39
VIM
-------
Acknowledgments
The authors appreciate the excellent technical support extended by Frank Beck and Cherri Adair of the SPRD,
by Kelly Hurt while he was a National Research Council associate, by staff of ManTech Environmental Research
Services Corp., and by staff of Dynamac, Inc.
IX
-------
SECTION 1
Introduction
MTBE is widely distributed in ground water. The U.S.
Geological Survey sampled shallow ambient ground water
from eight urban areas in 1993 and 1994. MTBE was
detected at concentrations at or above 0.2 u,g/l in 27% of
the 210 wells and springs that were sampled (Squillace et
al.,1996). The U.S. Environmental Protection Agency has
tentatively classified MTBE as a possible human carcino-
gen (U.S. EPA, 1996). There is currently much concern
about the occurrence and behavior of MTBE in ground
water that might be used for a drinking water supply.
Higher concentrations of MTBE in ground water are the
result of releases of gasoline containing oxygenates from
underground storage tanks (Landmeyer et al., 1998).
Fuels also contaminate ground water with benzene and
alkylbenzenes including toluene, ethylbenzene, and the
xylenes (BTEX compounds). Careful and detailed studies
of the transport and fate of the BTEX compounds demon-
strated that these compounds were biologically degraded
under natural conditions in ground water (summarized in
Wiedemeier et al. 1999).
The data supports the theory that the spread of BTEX
contamination at many sites was limited by natural bio-
degradation processes. As a result of our increased un-
derstanding of benzene plume behavior, natural attenua-
tion is now being formally recognized as a component of
many risk-based remedies at petroleum fuel spill sites.
There is little recognition in the literature that natural
biodegradation may control the spread of MTBE contami-
nation in ground water. In his review, Chapelle (1999)
noted that "Field studies of MTBE biodegradation relative
to BTEX compounds ... indicate that MTBE is biodegraded
in shallow aquifers, but that biodegradation is less than for
BTEX compounds." Mormile et al. (1994) conclude "the
common ether oxygenates resist both anaerobic and aero-
bic decay and must be considered recalcitrant chemicals."
As will be discussed in the next section, there is evidence
in the literature that anaerobic degradation of MTBE is
possible. There are also many experiments where degra-
dation was not detected.
Evidence for Biodegradation of MTBE
under Anaerobic Conditions in Ground
Water
There are two reports of MTBE biotransformation in
laboratory studies under methanogenic conditions. Yeh
and Novak (1994) constructed static soil and water micro-
cosms with material from three sites; a site at a wooded
area at the Virginia Polytechnic Institute (VPI) at
Blacksburg, Virginia, that is largely unsaturated clay; a
site at VPI in a low area that is mainly sandy loam that
receives runoff from a feedlot, and a site at Newport
News, Virginia, that is mainly silty loam. Ethanol and
starch were added as a source of molecular hydrogen.
Potassium phosphate and ammonium chloride were added
as nutrients. Cysterine and sodium sulfide were added to
encourage anaerobic processes. Sodium molybdate was
added to inhibit sulfate-reducing microorganisms in the
microcosms that were intended to simulate methanogenic
conditions. The initial concentration of MTBE in the
microcosms was 100 mg/l. The microcosms were incu-
bated at 20°C for times extending from 250 days to 300
days. Removals in microcosms were compared to remov-
als in autoclaved controls prepared for each site.
After 250 days of incubation, there was no removal of
MTBE in excess of removal in autoclaved controls in
microcosms constructed with material from the sandy
loam site downgradient of the feed lot, or the silty loam
site. However, in excess of 99 percent of MTBE was
removed in a microcosm constructed with clay material
collected at a depth of 1.5 meters below land surface, and
80 percent of MTBE was removed in a microcosm con-
structed from material collected at a depth of 3.0 meters
below land surface. The rate of removal in the microcosm
constructed with material from a depth of 1.5 meters
corresponded to a first-order rate of removal of 3.3 per
year, or a half life of eleven weeks. The removal in the
microcosm constructed with material from a depth of
3.0 meters corresponded to a first-order rate of removal of
2.0 per year or a half life of eighteen weeks.
In their studies, Yeh and Novak (1994) found no evi-
dence of MTBE biodegradation under anaerobic condi-
tions where nutrients and a hydrogen source were not
added, or under denitrifying conditions or sulfate-reducing
conditions when nutrients and a hydrogen source were
added. The removal of MTBE was only associated with
methanogenic conditions.
Mormile et al. (1994) examined material from a sandy
water-table aquifer near Empire, Michigan, that had been
contaminated with gasoline; sediment from the Ohio River
that had been impacted by oil storage and barge loading
facilities; and, sediment from Mill Creek in Cincinnati,
Ohio, that had been impacted with industrial and munici-
pal sewage sludge. Microcosms were constructed with
slurries of sediment and ground water. The slurries were
-------
amended with sodium sulfide, but no additional nutrients or
hydrogen sources were added. Some of the microcosms
received sodium sulfate or sodium nitrate to stimulate
sulfate-reducing and nitrate-reducing conditions. The ini-
tial concentration of MTBE was 50 mg/l as carbon.
There was removal of MTBE in one of three replicate
microcosms constructed with the sediment from the Ohio
River. After 152 days of incubation, the concentration of
MTBE was reduced to 22 mg/l carbon from an initial
concentration of 48 mg/l carbon. There was no removal of
MTBE in an autoclaved control. In the microcosm, the
removal of MTBE was associated with the production of
tertiary butyl alcohol (TEA).
In contrast to the removal of MTBE in the one micro-
cosm constructed with sediment from the Ohio River, there
was no removal in sediment that was contaminated with
gasoline after 230 days of incubation, or in the sediment
impacted with sewage sludge after 180 days of incubation.
In an earlier study, Suflita and Mormile (1993) had exam-
ined the degradtion of a variety of oxygenate in material
from a sandy water table aquifer at Norman, Oklahoma,
that had been contaminated with landfill leachate. After
249 days of incubation, there was no evidence of removal
of MTBE.
Consistent with the work of Yeh and Novak (1994),
Mormile et al. (1994) found no removal of MTBE under
sulfate-reducing conditions or nitrate-reducing conditions
in the three materials they examined.
There are two additional reports of MTBE degradation
under anaerobic conditions. Landmeyer et al. (1998)
examined MTBE degradation at a site on Port Royal
Island, South Carolina, in the Lower Coastal Plain of the
Atlantic Coastal Plain geophysical province. The aquifer
had been contaminated with a gasoline spill. Microcosms
were constructed with material from an area with high
concentrations of BTEX contamination and MTBE con-
tamination, and a second area with high concentrations of
MTBE but much lower concentrations of BTEX. Micro-
cosms were constructed with material from each site with a
high or a low concentration of MTBE. Transformation of
MTBE was assayed by collecting radio-labeled carbon
dioxide produced from the transformation of MTBE that
was uniformly labeled with carbon 14 (radio-labeled impu-
rity less than 0.2% of the total label).
The headspace of the microcosms was helium, which
resulted in iron-reducing conditions in the microcosms.
After 28 weeks of incubation, between 2.0% and 3.0% of
the label was transformed to carbon dioxide (mean of
triplicate microcosms for all four experimental treatments).
Attenuation under iron-reducing conditions was real, but
the rate of transformation was slow, corresponding to a
first-order rate of attenuation of 0.06 per year. Anaerobic
biodegradation did not attenuate the plume of MTBE in
ground water before it approached the receptor.
Church et al. (1997) examined the effluent from a col-
umn microcosm constructed with core material from a site
in Trenton, New Jersey. The influent concentration of
MTBE was near 100 u,g/l. After 35 days of operation, the
effluent concentration of MTBE was 160 u,g/l, and there
was no detectable concentration of TEA (interpretation of
their Figure 2). After 44 days of operation, the effluent
concentration of MTBE was 160 u,g/l and the concentration
of TEA was 20 u,g/liter. After 52 days of operation, the
effluent concentration of MTBE was reduced to 40 u,g/l,
and the concentration of TEA increased to 60 mg/l.
Purpose of the Case Study
This case study is intended to answer the following
questions. Can MTBE be biodegraded under methanogenic
conditions in ground water that was contaminated by a
fuel spill? Will biodegradation reach concentrations of
MTBE that are less than regulatory standards? Is the rate
of degradation in the laboratory adequate to explain the
distribution of MTBE in the ground water at the field site?
What is the relationship between the degradation of MTBE
and degradation of the BTEX compounds? How long can
the fuel release continue to contaminate ground water at
the site?
The case study was conducted at a former fuel farm that
had been operated by the United States Coast Guard at
their Support Center at Elizabeth City, North Carolina. Fuel
for aircraft was stored at the site until December 31,1991.
The fuel farm had been in use since 1942, and originally
consisted of a 50,000-gallon concrete underground stor-
age tank, and two steel underground storage tanks with a
volume of 12,000-gallons and 15,000-gallons, respectively
(adjacent to location CPT-1 in Figure 1.1). The steel tanks
were apparently removed in the mid-1980s.
The United States Coast Guard has conducted exten-
sive free product recovery efforts at the site. Figure 1.1
depicts the location of residual LNAPL, the direction and
speed of ground-water flow, and the ground-water sam-
pling locations at the site.
A GeoProbe™ push point sampler was used to acquire
water samples at the location in Figure 1.1. At each
location the aquifer was sampled in a vertical profile that
extended from the water table, through a shallow silty clay
layer, into a fine sand unit, and then into a silty clay unit
beneath the sand. The GeoProbe™ push points were
screened over a vertical interval of 1.5 feet. The
GeoProbe™ samples extended from 5 feet below land
surface to 30 feet below land surface. The aquifer was
confined to an interval between 10 and 25 feet below land
surface. At a minimum, every other 1.5 foot vertical
interval was sampled.
The hydraulic conductivity of the material sampled was
estimated for each GeoProbe™ sample. The measured
concentration of MTBE, benzene, and methane were
weighted by the hydraulic conductivity of the interval being
sampled before the vertical samples were averaged. Be-
cause the GeoProbe™ samples are flow-weighted aver-
ages, there is no danger that the data could give a false
impression of natural attenuation due to hydraulic averag-
ing along the flow path in the aquifer.
The concentration of MTBE in ground water under the
spill was high, 1740 u,g/L at location CPT-1 in Figure 1.1.
However, the only monitoring locations downgradient of
-------
the source area that had concentrations of MTBE that
exceeded regulatory standards were monitoring locations
ESM-14 and ESM-3, and the concentrations were approxi-
mately 20% of the maximum concentration in the source
area.
The concentration of MTBE in a permanent monitoring
well at location ESM-14 (see Figure 1.1) showed good
agreement with the weighted-average concentration from
the vertical profile sampling (see Table 1.1). In order to
estimate the temporal variability of the plume, data from
the permanent monitoring well at location ESM-14 are
reported for sampling events extending from August 1996
through September 1999. The August 1996 sampling
event was almost five years after the site was no longer
used for fuel storage. In the interval from August 1996 to
September 1999 there are no recognizable trends in the
concentration of MTBE, benzene, or methane. There is no
evidence that the concentrations of MTBE leaving the
source are decreasing overtime.
The site was selected for study for two reasons. The
source of the MTBE plume was relatively stable, based on
trends in MTBE concentrations in the most contaminated
permanent monitoring well at the site (Table 1.1). This
made it possible to differentiate attenuation in concentra-
tion as water moved downgradient from attenuation of the
source area itself. The site was also selected because the
concentrations of MTBE were greatly attenuated at moni-
toring locations that were relatively close to the source
area in terms of time of travel of ground water. The time
required for ground water to travel from the source area to
the most distant sampling location is less than five years.
This suggested that the kinetics of natural attenuation at
this site should be rapid.
Table 1.1 Temporal variation in the concentrations of MTBE, Benzene, and Methane at the most contaminated
permanent sampling location that is downgradient of the LNAPL area.
IVTTBE
Benzene
Methane
Vertical
Profile
8/1996
Permanent Well ESM-14
8/1996
10/1997
10/1998
12/1998
7/1999
9/1999
(|ag/l)
383
139
7,780
353
631
11,500
194
389
15,400
154
1280
16,200
65
1300
10,900
259
2185
8,400
609
1070
962
-------
PASQUOTANK RIVER
ESM-10 ESM.9
100 150 200 100
Approximate Scale in Meters Approximate Scale in Feet
Figure 1.1 Site selected for the case study of natural attenuation of MTBE under methanogenic conditions. The shaded concentric circles
represent the residual LNAPL from a fuel spill. The concentration of MTBE at location CPT-1 was 1740 |ig/l. The arrow represents
the distance traveled by the ground water in three years. The only sampling location with concentrations of MTBE above 20 |ig/l
was ESM-14.
-------
SECTION 2
Laboratory Studies
Construction, Sampling and Analysis of
Microcosms
Microcosms were constructed with aquifer material from
location ESM-14. This location had the highest concen-
tration of MTBE in the permanent monitoring wells that
were available at the time the samples for the microcosm
study were collected. At this location, approximately 10
feet (3 meters) of silty clay overlies 15 feet of silty sand
and fine sand. The water table is near 10 feet below land
surface. A hollow stem auger was advanced into the
earth to a depth of approximately 15 feet (4.6 meters).
The auger was maintained at this depth, and was rotated
to elevate material on the auger flights. The initial material
that was elevated was silty clay; this material was dis-
carded. Approximately 0.3 cubic meters of fine sand was
elevated on the auger flights and discarded, then 8 liters of
sediment was collected for construction of microcosms.
The sediment was collected and stored in 1-quart glass
jars. To protect the anaerobic microorganisms that might
be present in the samples from oxygen in the atmosphere,
the head space above the sediment was replaced with
ground water from the borehole immediately after collec-
tion. The samples were cooled and shipped to the Robert
S. Kerr Environmental Research Center with water ice,
and stored at 4°C until used to construct microcosms.
To protect anaerobic microorganisms from oxygen in
the atmosphere, all manipulations to prepare the micro-
cosms were carried out in a glove box with a concentra-
tion of oxygen in the atmosphere that was less than
1 ppm (v/v). This corresponds to a concentration of oxy-
gen in water (at equilibrium) of 0.00004 mg/l. Microcosms
were prepared in glass serum bottles with a volume of 25
ml. Ground water from the bore hole was added to the
sediment to make a thick slurry. This slurry was trans-
ferred to the serum bottles with a scoop. Each microcosm
received 40 gm wet weight of slurry and 1.0 ml of a dosing
solution containing MTBE, or MTBE and alkybenzenes.
The remaining volume (3 to 4 ml) was filled with auto-
claved ground water from the bore hole. The microcosms
were sealed with a grey butyl rubber septum and a crip
cap. The microcosms were stored in the same glove box,
under an atmosphere that was 2% to 5% hydrogen and
contained less than 1 ppm oxygen. Ground-water tem-
peratures from permanent wells at the site varied from 19°
to 24°C between sampling dates in December and Sep-
tember. The microcosms were incubated at room tem-
perature (20° to 22°C).
To prepare abiotic controls, a portion of the sediment
was autoclaved overnight. Four treatments were pre-
pared; sediment amended with MTBE alone, autoclaved
sediment amended with MTBE alone, sediment amended
with MTBE and alkylbenzenes, and autoclaved sediment
amended with MTBE and alkylbenzenes. The initial con-
centrations of MTBE and alkylbenzenes that were achieved
in the microcosms are listed in Table 2.1.
The microcosms were sampled and analyzed as fol-
lows. The contents of the microcosms were vigorously
stirred with a vortex mixer. The microcosms were centri-
fuged to settle the solids. Then the crimp cap and septa
were removed, and 1.0 ml of water was transferred to
39 ml of dilution water and sealed in 40 ml VOA bottle with
a Teflon™-faced silicone septum and a screw cap. The
dilution water was distilled water that had been boiled,
each sample received one drop of sulfuric acid to pre-
serve the sample. The concentration of MTBE and
alkylbenzenes were determined by purge and trap analy-
sis using gas chromatography with a PID detector. The
limit of quantification for MTBE and the alkylbenzenes was
1 u,g/l, corresponding to a limit of quantification of 40 u,g/l
in the pore water of the microcosms.
The sediment selected to construct the microcosms
was intended to represent the region in the aquifer where
natural attenuation of MTBE was in progress. The micro-
cosms were constructed with material from a location
where the apparent natural attenuation of MTBE was
extensive, but was not complete. Table 2.1 compares the
concentration of MTBE in the source area of the plume to
the concentration in a permanent monitoring well at the
location that was used to acquire the sediment for the
microcosms. At the sample location, the concentration of
MTBE was reduced approximately tenfold from the con-
centration in the LNAPL source area. Another tenfold
reduction would approach concentrations that would meet
regulatory standards for MTBE.
The microcosms incubated over time were intended to
represent the travel of a representative volume of water
along the entire flow path, starting with the source and
extending to a potential receptor. The sediments in the
microcosms were amended with MTBE to simulate the
highest concentration in the source area. The travel time
of MTBE from the source to a potential receptor is on the
order of four years. The microcosm study was designed to
last for eighteen months to two years.
-------
Table 2.1 The concentration of MTBEand alkylbenzenes in the most contaminated sample of ground water from
the LNAPL source area, in the permanent monitoring well at the location where the sediment used to
construct the microcosms was acquired, and the initial concentrations achieved in the microcosms.
Most Sample
Impacted Location
Ground Water Ground Water
CPT-1 Well
Compound 3.1 m to ESM-14
3.6 m bis
Single Analysis
MTBE MTBE
alone alone
control
Microcosms
MTBE
plus BTEX
Mean (Sample Standard Deviation,
MTBE
plus BTEX
control
n = 3)
van
MTBE
Benzene
Toluene
Ethylbenzene
o-xylene
m-Xylene
p-Xylene
m+p-Xylene
1,2, 3-trimethylbenzene
1,2,4-trimethylbenzene
1, 3, 5-trimethylbenzene
3640
7830
383
396
23
1250
286
430
107
353
631
1.8
1.9
<1
2.9
1.2
<1
<1
3112
(188)
<40
<40
<40
<40
<40
<40
<40
<40
<40
2908
(538)
<40
<40
<40
<40
<40
<40
<40
<40
<40
5680
(138)
2079
(218)
2183
(233)
1256
(130)
1911
(170)
1592
(158)
1556
(151)
887
(54)
625
(40)
664
(45)
3330
(360)
1953
(174)
1996
(176)
1109
(108)
1697
(146)
1404
(134)
1361
(139)
747
(78)
521
(62)
551
(69)
To determine whether there was any interaction between
the presence of alkylbenzenes and the removal of MTBE,
one set of treatments was amended with benzene, tolu-
ene, ethylbenzene, the three xylenes, and the three
trimethylbenzenes, and one set of treatments was not
amended. The initial concentrations of individual
alkylbenzenes in the microcosms were higher than their
concentration in the source area, with the exception of
benzene, where the concentration in the microcosms was
approximately one-fourth the maximum concentration in
the source area (Table 2.1). Background concentrations
of alkylbenzenes were not detected in microcosms that
were not amended with alkylbenzenes.
Removal of MTBE
Removal of MTBE in material that was supplemented
with alkylbenzenes was extensive (Figure 2.1). There was
no evidence of MTBE removal over removal in the controls
in the first 175 days of incubation. The concentrations of
MTBE in replicate microcosms from both the living and
control treatments show relatively little scatter. After
385 days of incubation, there is evidence of removal in the
living treatment. The data after 385 days of incubation
show a great deal of scatter. The range in concentrations
in six replicates is over an order of magnitude wide. After
490 days of incubation, there is consistent removal of
MTBE compared to the controls.
-------
10000
1000 -
PQ
H
100
10
+ MTBE plus BTEX
D MTBE plus BTEX
Control
-100
100 200 300
Time (Days)
400
500
600
Figure 2.1 Removal of MTBE in microcosms constructed with MTBE and BTEX compounds compared to removal in control microcosms that
were autoclaved to prevent biotransformation of MTBE.
The average concentration remaining in six replicates of
the living treatment was 81 u,g/l, compared to 5680 u,g/l at
the beginning of incubation. The average concentration
remaining in the control treatment after 490 days was 1466
u.g/1, compared to 3330 u.g/1 at the beginning of incubation.
The removal in the controls was a little more than twofold,
while removal in the living microcosms was 70-fold.
Removal of MTBE in material that was not supplemented
with alkybenzenes was also extensive (Figure 2.2). There
was little evidence of removal in the first 175 days of
incubation. After 385 days, the removal of MTBE in the
living microcosms was extensive. After 490 days of incu-
bation, the concentration of MTBE in six replicate micro-
cosms was below 40 u,g/l, compared to 3112 u,g/l at the
beginning of incubation. After 490 days, the average
concentration of MTBE in the control microcosms was
1571 u,g/l, compared to an initial concentration in the
controls of 2908 u,g/l.
A first-order rate of removal was fitted to the data by a
linear regression of the natural logarithm of the concentra-
tion of MTBE on the time of incubation. The rate of
removal of MTBE in microcosms supplemented with
alkylbenzenes was 3.02 per year ±0.52 per year at 95%
confidence. Removal in the corresponding controls was
0.39 +/- 0.19 per year at 95% confidence. The removal in
the microcosms without added alkylbenzenes was 3.5 per
year +/- 0.65 per year at 95% confidence. Removal in the
corresponding controls was 0.30 per year, ±0.14 per year
at 95% confidence. The rate constants were fit to the
entire data; no correction was made for the apparent lag
period.
A container control was not done. There is no way to
determine if the removals in the controls are due to kineti-
cally slow sorption to the aquifer solids, or to diffusion out
of the microcosm through the septa.
Removal of Benzene, Toluene, and
Ethylbenzene
Toluene was removed rapidly and extensively in the
microcosms. After only 40 days or 47 days of incubation,
toluene removal was extensive in most of the microcosms
sampled (Figure 2.3). After 110 days of incubation, the
concentration of toluene was less than 40 u,g/l in all of the
microcosms sampled. Benzene was also rapidly removed
from the microcosms. There was no evidence of benzene
removal after 40 days or 47 days of incubation (Figure 2.4)
although removal of toluene was evident. After 110 days
of incubation, the concentration of benzene in all the
microcosms sampled was less than 40 u,g/l. There is no
evidence of a lag for removal of toluene. There may have
been a slight lag in the removal of benzene.
The behavior of ethylbenzene (Figure 2.5) is representa-
tive of the behavior of the xylenes and trimethylbenzenes
as well. Removal was extensive in both living microcosms
and controls. There was no difference in removal in living
and control microcosms through 110 days of incubation.
After 175 days of incubation, the concentration of
ethylbenzene in the living microcosms is below the limit of
quantification (40 u,g/l) but the average concentration in the
controls is only 113 u,g/l compared to an initial concentra-
tion of 1109 u,g/l. After 490 days of incubation, the concen-
tration in both living and control microcosms was less than
the quantification limit.
-------
PQ
H
10000
1000 -
100
10
+ MTBE Alone
D MTBE Alone Control
-100
0
100
200 300
Time (Days)
400
500 600
Figure 2.2 Removal of MTBE in microcosms constructed with MTBE but without supplemental concentrations of BTEX compounds compared
to removal in control microcosms that were autoclaved to prevent biotransformation of MTBE.
10000
1000
o
H
100
10
+ Toluene
° Toluene Control
-100 0 100 200 300 400 500 600
Time (Days)
Figure 2.3 Removal of Toluene in microcosms constructed with MTBE and BTEX compounds compared to removal in control microcosms
that were autoclaved to prevent biotransformation of toluene. The solid line is fit through the removal in the controls.
-------
10000
^ 1000
~
N
PQ
100
10
* +
+ Benzene
D Benzene Control
-100
100
200 300
Time (Days)
400
500
600
Figure 2.4 Removal of Benzene in microcosms constructed with MTBE and BTEX compounds compared to removal in control microcosms
that were autoclaved to prevent biotransformation of benzene. The solid line is fit through the removal in the controls.
10000
N
s
X
J3
w
1000
" 100
10
+ Ethylbenzene
D Ethylbenzene Control
D
D
-100
100 200 300 400
Time (Days)
500 600
Figure 2.5 Removal of Ethylbenzene in microcosms constructed with MTBE and BTEX compounds compared to removal in control micro-
cosms that were autoclaved to prevent biotransformation of ethylbenzene.
-------
Relationship between removal of BTEX
compounds and removal of MTBE
Removal of MTBE did not require the presence of BTEX
compounds. Figure 2.6 plots the removal of MTBE, ben-
zene, toluene, and ethylbenzene. Toluene was entirely
depleted within 40 to 47 days, benzene was entirely de-
pleted within 110 days, and ethylbenzene was entirely
depleted within 175 days. During this time period there
was no evidence of removal of MTBE. After 385 days
there was evidence of extensive removal of MTBE in one
microcosm, limited removal in three microcosms, and no
evidence of removal in two microcosms. The removal of
MTBE did not begin until the removal of the BTEX com-
pounds was complete.
There is a possibility that removal of MTBE may have
been inhibited by the presence of BTEX compounds. How-
ever, the lag for removal of MTBE in microcosms without
BTEX compounds was also long, at least 175 days (Fig-
ure 2.6). Removal of MTBE in microcosms that did not
contain detectable concentrations of alkylbenzenes was
not detected until 385 days of incubation. There is not
enough resolution in the sampling schedule to determine if
the presence of BTEX compounds inhibited MTBE re-
moval.
10000
e 1000
a 10°
10
*
+ MTBE
" Benzene
A Toluene
° Ethylbenzene
i i
O
i
-100 0 100 200 300
Time (Days)
400
500
600
Figure 2.6 Comparison of the time lags for removal of MTBE, and of benzene, toluene, and ethylbenzene in microcosms constructed with all
the compounds present together.
10
-------
SECTION 3
Site Characterization
Site Description and History
The case study was conducted at a former fuel farm
located at the U.S. Coast Guard Support Center at Eliza-
beth City, North Carolina. The following description is
excerpted from the Former Fuel Farm Work Plan, a part of
the Remediation Feasibility Assessment Work Plan pre-
pared for the U.S. Coast Guard Support Center, Elizabeth
City (SCES), North Carolina, by Parsons Engineering Sci-
ence, 1996.
The Support Center is located on the southern bank of
the Pasquotank River. The former fuel farm was located
south of a concrete ramp used to recover seaplanes from
the Pasquotank River (Figure 3.1). Currently a plume of
MTBE and fuel hydrocarbons in ground water emanates
from a source area in the location of the former fuel farm,
and flows under the concrete ramp toward the Pasquotank
River to the north, and toward a drainage canal along the
western side of the seaplane ramp. This source area
corresponds to the former location of fuel storage tanks on
the site (Figure 3.1).
Fuel was stored at the site until December 31,1991. The
fuel farm had been in use since 1942, and originally con-
sisted of a 50,000-gallon concrete underground storage
tank (TANK 23 in Figure 3.1), and two steel underground
storage tanks with a volume of 12,000-gallons and
15,000-gallons, respectively (adjacent to location CPT-1 in
Figure 3.1). The steel tanks were apparently removed in
the mid-1980s. In addition to the underground storage
tanks, two steel, above-ground storage tanks with a capac-
ity of 50,000 gallons were installed in the mid-1980s. There
was evidence of corrosion in the transfer lines from these
tanks. They were taken out of service and removed from
the site. No evidence of a release from the pipes was
discovered.
The U.S. Coast Guard began a free product recovery
effort at the site in September 1990. Eight recovery wells
were arranged around the source area in a circle. By March
1992, a total of 79,000 gallons of fuel was recovered.
Core Sampling the Source Area
In September 1996, a GeoProbe™ was used to acquire
core samples in continuous vertical profiles at seven loca-
tions in or near the source area (locations CPT-1 through
CPT-7 in Figure 3.1). The water table was 7.0 to 8.0 feet
below land surface (2.1 to 2.4 meters below land surface).
The cores extended from the surface to a depth 12 to 16
feet (3.7 to 4.9 meters).
The cores were cut into subcores that were 4 inches long
(10 cm long). A plug of material approximately 1.0 cm in
diameter and 5 cm long was acquired from each subcore
with a paste sampler. The plugs were immediately trans-
ferred in the field into 40-ml glass vials with 5 ml of
methylene chloride and 10 ml of distilled water. The vials
were sealed with Teflon™-faced septa and screw caps,
then they were shaken to extract organic components into
the methylene chloride. The contents were allowed to
settle, then the methylene chloride was taken for analysis
by gas chromotography using a mass spectrometer as a
detector. The limit of quantitation for MTBE and for BTEX
compounds was 0.01 mg/kg; the limit of quantitation for
Total Petroleum Hydrocarbons was 50 mg/kg.
The cores at each sampling location extended from clean
soil above the release, through the release to clean aquifer
material below the release. The quantity of Total Petro-
leum Hydrocarbon (TPH) or MTBE in individual cores was
summed to determine the total amount of TPH and MTBE
present at each location. The subcores were 10 cm long.
The concentration reported in mg/kg was considered repre-
sentative of a block of soil that was 1.0 meter square and
0.1 meter deep. The dry bulk density of the soil or
sediment was assumed to be 1,820 kg/ m3. Each block of
soil would have a weight of 182 kg/m2. The concentration
reported in mg/kg was multiplied by 182 kg/m2 to determine
the quantity in each block. The quantity in each block was
summed to determine the total quantity at each location.
Results are presented in Table 3.1.
The greatest quantity of TPH was found at locations
CPT-2 and CPT-1 (Table 3.1). These locations were near
the location of the original steel underground storage tanks
(Figure 3.1). These two locations also had the greatest
mass of MTBE, and the highest concentration of MTBE in
the residual fuel. The quantity of TPH at location CPT-3
was high, but the concentration of MTBE in the TPH was
lowerthan the concentration in locations CPT-1 and CPT-2.
MTBE was detected at location CPT-7, but the quantity
and concentration in the TPH was much lower than at
locations CPT-1 and CPT-2. The fuel release that contains
MTBE is centered around locations CPT-1 and CPT-2, and
11
-------
PASQUOTANK RIVER
Drainage Canal
STEEL AST'S STEEL UST'S
0
50
100 150 200 100
Q0 19° Q 2Q
200
Approximate Scale in Meters Approximate Scale in Feet
Figure 3.1 Relationship between the sampling locations for characterization of the LNAPL source area (labeled CPT-1 through CPT-5), and
the former location of storage tanks for fuels.
12
-------
Table 3.1. Quantity of Total Petroleum Hydrocarbon and MTBE at seven sampling locations in ornearthe point of
release of fuel.
Location Total Petroleum Hydrocarbons MTBE Mass Fraction MTBE
(kg/m2) (kg/m2) ppm MTBE per TPH
CPT-7
CPT-5
CPT-3
CPT-1
CPT-2
CPT-6
CPT-4
9.7
0.220
26
54
75
1.3
30
0.00017
<0.0001
0.0015
0.0152
0.0104
<0.0001
<0.0001
18
57.7
282
139
is roughly bound by CPT-7 to the west, by CPT-5 and CPT-3
to the north, and by CPT-6 and by CPT-4 to the east.
Estimation of Total Quantity of TPH and
MTBE and the Area Impacted
The former location of the steel underground storage
tanks was selected as the location of CPT-1. This location
had the highest quantity of MTBE (Table 3.1), and will be
taken as the center of the release. Location CPT-1 is
60 meters from CPT-7, 67 meters from CPT-5, 37 meters
from CPT-3, 88 meters from CPT-6, and 60 meters from
CPT-4 (Figure 3.2). If the source is a circle that fits within
the space bounded by CPT-7, CPT-5, and CPT-4, then its
radius is 60 meters, and its area is 11,000 m2. If the
concentration of TPH in the source is 46 kg/m2 (the average
of locations CPT-1, CPT-2, CPT-3, and CPT-4 in Table 3.1),
the total quantity of fuel hydrocarbons remaining in the
source is 500,000 kg. If the density of the fuel is 0.82, this
corresponds to 620,000 liters or 180,000 gallons of fuel.
Addendum 2 to the Corrective Action Plan, Former Fuel
Farm (SWMU No. 32), U.S. Coast Guard Support Center
Elizabeth City, Elizabeth City, North Carolina, Parsons
Engineering Science, 1997, provides an independent as-
sessment of the mass of fuel remaining. They estimate that
approximately 100,000 gallons (380,000 liters) remain in the
soil, and that the impacted area is approximately 150,000
square feet (14,000 m2). The agreement between their
estimate and our estimate is acceptable.
The average quantity of MTBE at locations CPT-1 and
CPT-2 was 12.8 g/m2 (Table 3.1). If the area of the source is
11,000 m2, this corresponds to a total quantity of 140 kg of
MTBE. This would be an upper boundary on the quantity of
MTBE in the source.
If the source "hot-spot" is restricted to the interval be-
tween CPT-1 and CPT-2 (see Figure 3.2), then the radius of
the "hot-spot" is 29 meters, and the area of the hot spot is
2,600 m2. The average concentration of MTBE at CPT-1
and CPT-2 is 12.8 g/m2, for a total in the "hot spot" of
33.2 kg of MTBE. If the remaining area of the source
(8,400 m2) has a quantity of MTBE equal to that of location
CPT-3 (1.5 g/m2), the addition quantity is 12.6 kg fora grand
total 46 kg in the source area. This would be a lower
boundary on the quantity of MTBE in the source.
Vertical Distribution of TPH and MTBE in
Core Samples
Figure 3.3 presents the vertical distribution of MTBE and
TPH in the continuous core samples from location CPT-1.
The majority of TPH was confined to a depth interval
between 1.5 and 3.0 meters. The relative proportions of
MTBE in the TPH were very consistent over this interval.
Below 3 meters TPH disappears, while the concentration of
MTBE declines gradually with increase in depth. The
MTBE in core samples below a depth of 3 meters can only
be dissolved in the ground water. In this interval there is no
TPH to partition into, and sorption to aquifer solids should
be negligible.
The vertical distribution of TPH and MTBE at location
CPT-2 (Figure 3.4) was very similar to the distribution at
location CPT-1 (Figure 3.3). The majority of the TPH was
confined to an interval between 1.5 and 3 meters below land
surface. The relative proportions of MTBE and TPH were
consistent across the vertical profile, and it was the same
proportion seen at location CPT-1. At the time the cores
were collected, the depth to the water table was 7.0 feet
(2.13 m). One core sample, located right at the water table,
had a TPH concentration of 139,000 mg/kg. This concen-
tration is high enough to represent free product floating on
the watertable. The concentration of MTBE in this sample
was 11.5 mg/kg. At location CPT-2, both TPH and MTBE
disappeared below a depth of 3 meters.
At location CPT-3, the TPH was also confined to an
interval between 1.5 and 3.0 meters (Figure 3.5). The
quantity of MTBE relative to TPH was less than the propor-
tion at locations CPT-1 and CPT-2. As was the case at the
other locations, the relative proportion of MTBE to TPH did
not change across the vertical profile.
13
-------
PASQUOTANK RIVER
50 100 150 2(?£
B
200
9
-»
Approximate Scale in Meters Approximate Scale in Feet
Figure 3.2 Inferred location of the fuel release, based on vertical core samples and the location of the steel underground storage tanks.
14
-------
Depth
(meters)
MTBE (mg/kg)
Figure 3.3 Vertical distribution of MTBE and Total Petroleum
Hydrocarbon (TPH) in core samples at location CPT-1
(see Figure 3.1 for map).
Total Petroleum Hydrocarbon (mg/kg)
20000 40000
60000
MTBE
TPH
Depth
(meters)
0 2 4 6 8 10
MTBE (mg/kg)
Figure 3.5 Vertical distribution of MTBE and Total Petroleum
Hydrocarbon (TPH) in core samples from location CPT-
3 (see Figure 3.1 for map).
Depth
(feet)
Total Petroleum Hydro carbon (mg/kg)
0 20000 40000
4 -
TPH
Depth
(meters)
4
MTBE (mg/kg)
Figure 3.4 Vertical distribution of MTBE and Total Petroleum
Hydrocarbon (TPH) in core samples from location CPT-
2 (see Figure 3.1 for map).
Distribution of Total Petroleum
Hydrocarbons and Hydraulic Conductivity
with Depth
Water samples were acquired using GeoProbe™ rods
with an outer diameter of 1.0 inch (2.54 cm). The leading
rod had 1.5 vertical feet (0.46 meter) of vertical mill slot
screens. In addition to collecting samples for analysis of
chemical parameters, the hydraulic conductivity was deter-
mined at each depth interval using an inverse specific
capacity test following the procedure of Wilson etal. (1997).
Figure 3.6 depicts the vertical relationship of Total
Petroleum Hydrocarbon (TPH) and Hydraulic Conductivity
at location CPT-1 (Figure 3.6). The TPH was confined to
an interval extending from 5 to 10 feet below land surface.
This may represent a "smear" zone around the watertable,
which was located 7 feet below land surface when the core
and water samples were collected in August 1996. The
core material containing significant concentrations of TPH
was silty clay to clay with very low hydraulic conductivity.
The depth interval with significant hydraulic conductivity
extended from 10 to 27 feet (3 to 8 meters) below land
surface (Figure 3.6). The interval containing TPH was just
above the interval allowing significant flow of ground water.
Figure 3.7 plots the locations of ground-water samples at
the site. At each location, sampling was attempted at
depths that extended across the first aquifer, starting in the
15
-------
Total Petroleum Hydrocarbon (mg/kg)
0 10000 20000 30000 40000
0 J ' ' ' H 0
Depth 15 _
(feet)
20 -
25 -
Water Table
30
— 4
Hydraulic Conductivity
Depth
5 (meters)
— 9
0 0.01 0.02 0.03 0.04
Hydraulic Conductivity (cm/sec)
Figure 3.6 Relationship between the vertical extent of Hydraulic
Conductivity and the vertical extent of Total Petroleum
Hydrocarbon at location CPT-1.
low conductivity material near the surface and extending
across the aquifer to the confining zone below the aquifer.
The vertical distribution of hydraulic conductivity in Fig-
ure 3.6 is representative of the distribution of hydraulic
conductivity across the transects depicted in Figure 3.7. At
a minimum, the hydraulic conductivity was measured in
every other 1.5 foot interval at each location. At many
locations, the hydraulic conductivity was measured in ev-
ery successive 1.5 foot interval. For each location in
Figure 3.7, Tables 3.2 and 3.3 compare the average hy-
draulic conductivity across the aquifer, the highest hydrau-
lic conductivity measured, and the conductivity in the
confining layers above and below the aquifer.
In the area that was sampled, hydraulic conductivity in
this aquifer was remarkably uniform. The aquifer abruptly
pinched out to the southeast (compare GP-21 and GP-22
in Table 3.3), and pinched out to a more limited extent to
the northwest. There was no systematic change in hy-
draulic conductivity moving north toward the Pasquotank
River. In general, the highest hydraulic conductivity was
twice the average conductivity, as would be expected if the
shape of the distribution of hydraulic conductivity with
depth was triangular.
Figures 3.8 and 3.9 depict the vertical distribution of the
actual sampling locations on transects that are depicted in
Figure 3.7, and summarized in Tables 3.2 and 3.3. These
figures are offered to provide the reader an indication of the
density of the data that are contoured in Figures 3.10
through 3.15.
The depth to the water table was approximately 7 feet in
the source area and 6 feet below land surface farther
toward the Pasquotank River. Figures 3.10 through 3.15
depict contours as depth below the water table.
Figures 3.10 and 3.11 contour the distribution of hydrau-
lic conductivity. In both transects, the conductive interval
starts approximately 5 to 6 feet below the water table and
extends to 18 to 20 feet below the water table. In the
horizontal plane, there is little indication that flow of ground
water is confined to preferential flow channels with mark-
edly higher hydraulic conductivity than surrounding aquifer
material.
Distribution of MTBE and BTEX Compounds
with Depth
The distribution of MTBE in the north-south transect is
depicted in Figure 3.12. The highest concentrations of
MTBE are in the shallow ground water underneath the
LNAPL at the south end of the transect. As the ground
water moves north (to the right in Figure 3.12), the highest
concentrations of MTBE are found in the depth intervals
with the highest hydraulic conductivity. There is a threefold
reduction in the concentration of MTBE at the most con-
taminated depth interval with each 200 feet north of the
source area. All the ground water from the location that
was closest to the Pasquotank River had less than 1 u,g/l
MTBE.
The transect of samples collected in December 1997,
runs approximately northwest to southeast. The transect
is oriented 50 degrees west of north. As discussed in
Appendix A, the average direction of ground-water flow is
8 degrees west of north. The angle between the transect
and the direction of flow is 42 degrees. The distribution of
MTBE in the northwest to southeast transect reveals a
plume that is approximately 350 feet wide. The distance
containing MTBE that is perpendicular to ground-water
flow is calculated by multiplying the sine of 42 degrees by
the contaminated length along the transect. The width of
the plume perpendicular to ground-water flow (230 feet, or
90 meters) is slightly less than the diameter of the source
area (compare Figure 3.7). The MTBE plume attenuates
abruptly on its northwest side, and attenuates more gradu-
ally on its southeastern side.
The distribution of BTEX compounds along the north-
south transect is depicted in Figure 3.14. As was the case
with MTBE, the highest concentrations are in the shallow
ground water underneath the LNAPL. As ground water
moves away from the source area, the highest concentra-
tions are found in the most conductive depth intervals.
Unlike the pattern seen at other sites, there is no evidence
that BTEX compounds are attenuating while MTBE is
persistent. There is little practical difference in the pattern
of attenuation for MTBE and BTEX compounds.
16
-------
PASQUOTANK RIVER
Average Direction of
Ground Water Flow
50 100 150 2001000
200
Approximate Scale in Meters Approximate Scale in Feet
Figure 3.7 Relationship between two transects of ground-water samples and the fuel release. The arrow represents the average direction of
ground-water flow.
17
-------
Table 3.2 Distribution of Hydraulic Conductivity (K) in the north-south transect sampled in August 1996
(Figure 3.7).
South to Number of
North measurements
CPT-2
CPT-1
CPT-3
CPT-5
ESM-14
ESM-10
GP-1
Table 3.3 Distribution
(Figure 3.7).
11
9
10
10
9
10
7
of Hydraulic
West to Number of
East measurements
GP-25
GP-24
GP-23
GP-1 9
GP-1 8
GP-1 7
GP-1 6
GP-1 5
GP-20
GP-21
GP-22
4
6
8
7
8
8
7
8
6
7
6
Average K
10 to 26. 5
feet
cm/sec
0.015
0.020
0.026
0.022
0.025
0.031
0.024
Conductivity (K)
Average K
10 to 26.5
feet
cm/sec
0.013
0.019
0.025
0.032
0.027
0.024
0.017
0.024
0.024
0.022
0.0001
Highest K
cm/sec
0.029
0.033
0.041
0.042
0.046
0.060
0.041
in the east-west
Highest K
cm/sec
0.038
0.047
0.057
0.084
0.063
0.052
0.031
0.050
0.051
0.052
0.020
Lowest K in
Higher Interval
cm/sec
0.0004
0.0036
0.0014
0.00001
0.0003
0.0077
0.0015
transect sampled in
Lowest K in
Higher Interval
cm/sec
0.0007
0.00015
0.00004
0.00001
0.0016
0.0003
0.00024
0.0034
0.00013
0.00003
Lowest K in
Lower Interval
cm/sec
0.00035
0.016
0.0001
0.012
0.0005
0.0033
0.0015
December 1997
Lowest K in
Lower Interval
cm/sec
0.0064
0.013
0.00021
0.012
0.00004
0.00004
0.00023
0.010
0.00004
0.00002
It is important not to interpret the apparent attenuation
along this transect as sufficient evidence for natural attenu-
ation along the flow path. The transect may be askew of
the flow path, and may sample water at the distal locations
that was never in the plume. The north-south transect is
oriented 16 degrees east of north (Figure 3.2). This is only
24 degrees from the average direction of ground-water
flow, but this difference is large enough to move out of a
plume that is 230 feet wide at a distance of 600 feet from
the edge of the source area.
To determine whether a particular location along the
transects sampled the plume or missed the plume, the
concentration of MTBE at each location was compared to a
number of geochemical indicators that correlate with bio-
logical activity in ground water. The concentrations of
methane, iron (II), oxygen, sulfate, and alkalinity at the
transect locations were compared to their concentrations
in ground water that was not impacted by the plume (see
Figures B.1 through B.6 in Appendix B). The geochemistry
of the site is discussed in detail in Appendix B. All of the
sampling locations on the north-south transect that were
downgradient of the source area were depleted of oxygen
and sulfate, and had elevated concentrations of methane,
iron (II), and alkalinity, indicating that these locations in the
north-south transect sampled the plume.
Figure 3.15 reveals that the plume, as sampled by the
east-west transect, is heterogeneous with respect to the
distribution of MTBE and BTEX compounds. The higher
concentrations of MTBE extend from 50 to 350 feet along
the transect (Figure 3.13). The higher concentrations of
BTEX extend from 200 feet along the transect to 400 feet
along the transect (Figure 3.15). The northwestern reach
of the transect has MTBE but no BTEX, the central reach
has both MTBE and BTEX, and the southeastern reach
has BTEX but no MTBE.
18
-------
Q
(i ft-"
30.0
Location of Sampling Points on North-South Transect
CS
3
03
H
ta
|
13
PQ
-4—*
6.0-
12.0-
<
18.0 ,
4
24.0 ,
»
i
»
»
•
i
*
•
*
•
•
ft 1 fill 2011
400 500 6 DO
Location on Transect (ft)
'lift gift 9ftO
Figure 3.8 Location of vertical sampling points along the north-south transect, collected in August 1996. Distance along the transect extends
from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow.
•M 6
H 9
= 12
| 15
|l8
f.21
3 2-1 -
,4'tcahon ot'Ssmpiitig Points on Fast-We^t Transect
15f t 200 250 3f«) .£50 -1(9) -150
Location on Transect eft?
Figure 3.9 Location of vertical sampling points along the east-west transect, collected in December 1997. Distance along the transect
extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow.
19
-------
Hydraulic Conductivity (cm/sec) at North-South Transect
s
1
«
PQ
P
0.0-
3.0-
6.0-
9.0-
12.0-
15.0-
18.0-
21.0-
24.0-
27.0-
30.
0.01-
•0.001
°-—o.ooooi-
•0.0001
0 100 200 300
400 500
600 700 800 900
Location on Transect (ft)
Figure 3.10 Distribution of hydraulic conductivity along the north-south transect, collected in August 1996. Distance along the transect extends
from south to north (bottom to top in Figure 3.7), in the direction of ground-water flow.
Figure 3.11
Hydraulic Conductivity (cm/sec) at East-West Transect
50 100 150 200 250 300
Location on Transect (ft)
350 400
450
Distribution of hydraulic conductivity along the east-west transect, collected in December 1997. Distance along the transect
extends from west to east (left to right in Figure 3.7), opposite the direction of ground-water flow.
20
-------
0.0-
3.0-
6.0-
,0
03
9.0
- 12.0
OJ
I 15-°
* 18.OH
J 21.0-
24.0-
OH
OJ
Q 27.OH
30.0
0
MTBE ((jg/1) at North-South Transect
00
200
300 400
500
600
700 800 900
Location on Transect (ft)
Figure 3.12 Distribution of MTBE along the north-south transect, collected in August 1996. Distance along the transect extends from south to
north (bottom to top in Figure 3.7), in the direction of ground-water flow.
0
3
6™
9~
|12^
|15H
gl8H
^ 21-
27-
MTBE (jug/1) at East-West Transect
0 50 100 150 200 250 300 350 400 450
Location on Transect (ft)
Figure 3.13 Distribution of MTBE along the east-west transect, collected in December 1997. Distance along the transect extends from west to
east (left to right in Figure 3.7), opposite the direction of ground-water flow.
21
-------
BTEX (jig/1) at North-South Transect
o.o.
_ 3.0^
? 6'°
I 9-°-l
s 12-°'
^ 18.0-1
•§ 21.0-
I 24-0-
&27.0-
° 30.0-
0 100 200 300 400 500 600 700
Location on Transect (ft)
800 900
Figure 3.14 Distribution of BTEX along the north-south transect, collected in August 1996. Distance along the transect extends from south to
north (bottom to top in Figure 3.7), in the direction of ground-water flow.
0™
^ijmmmj ^%
S^^^rf* •% ™"
^^
x> 6~
M 9
S18'
t21"
Q24»
27™
0
BTEX ((ig/1) at East-West Transect
50
100 150 200
250 300
350 400 450
Location on Transect (ft)
Figure 3.15 Distribution of BTEX along the east-west transect, collected in December 1997. Distance along the transect extends from west to
east (left to right in Figure 3.7), opposite the direction of ground-water flow.
22
-------
SECTION 4
Transport and Fate of MTBE in the Ground Water
Estimated Rate of Attenuation in Ground
Water
Ground-water flow carrying the plume of contamination
is contained within a semi-confined aquifer. Most of the
plume occurs under the concrete of an operational apron
at the U.S. Coast Guard Support Center. The bottom of
the drainage ditch on the northwest side of the site does
not penetrate the upper layer of silty clay, and does not
communicate with the sandy layer that carries the plume
of contamination. As a result, there is little opportunity for
the ground-water flow field to be influenced by local re-
charge or local discharge. As a simplification, we assume
that the shape of the water table is a plane during any
particular round of sampling.
The direction of ground-water flow was determined by
using a least squares regression technique to fit a plane
through the elevation of the watertable in eight permanent
monitoring wells at the site. A separate regression was
performed for each of eighteen rounds of quarterly moni-
toring starting in September 1994 and extending through
December 1998, and each of fourteen rounds of monthly
sampling starting in February 1998 and extending to
March 1999. Appendix A contains a map showing the
location of the permanent monitoring wells used to esti-
mate flow direction, tables showing the watertable eleva-
tion in the monitoring wells at each round of sampling, and
tables showing results of the regression analyses, with
estimates of plume direction, hydraulic gradient, and good-
ness of fit. Appendix A also contains maps that compare
the contours of the water table as estimated by the
regressions to the measured elevations of the water for
the eighteen rounds of quarterly monitoring.
The regional ground-water flow direction is north, di-
rectly toward the Pasquotank River; however, the hydrau-
lic gradient at any one time in the study area is strongly
influenced by the stage of the Pasquotank River. The
direction and magnitude in the flow of ground water as
predicted in each round of quarterly sampling is depicted
in Figure 4.1. For each round of sampling, an arrow in the
figure represents the distance ground water would move
in one year under the conditions of gradient and direction
that were encountered in that round of sampling. The
variation in direction and magnitude in the fourteen rounds
of monthly sampling is depicted in Figure 4.2.
The average direction of ground-water flow was calcu-
lated by weighting the direction of flow on any particular
round of sampling by the hydraulic gradient at that particu-
lar round, then taking an average of the weighted flow
directions.
Forthe eighteen rounds of quarterly sampling, the aver-
age flow direction was 8.6 degrees west of north, with a
standard deviation of 9.1 degrees. Forthe fourteen rounds
of monthly sampling, the average flow direction was 8.7
degrees west of north, with a standard deviation of 23
degrees.
The seepage velocity of the plume was estimated from
Darcy's Law. To calculate seepage velocity, the hydraulic
conductivity was multiplied by the hydraulic gradient, and
divided by the porosity. Table 4.1 summarizes the statis-
tical properties of the parameters used to calculate seep-
age velocity. There is relatively little variation in hydraulic
conductivity; the range of eight samples is only 56% of the
mean. There is greater variation in hydraulic gradient
from one round of sampling to another. For eighteen
rounds of quarterly sampling, the hydraulic gradient varied
over an order of magnitude. Nothing is known directly
about the range of porosity at the site. The total pore
space in core samples was calculated by comparing the
wet and dry weight of core samples from five locations.
The total porosity was very close to 0.34 to 0.36. A value
of 0.35 was taken as an upper boundary on effective
porosity. A survey of the literature and professional
judgment was used to assign an average effective poros-
ity of 0.3, and lower boundary on effective porosity of 0.25.
The average seepage velocity was calculated by multi-
plying the average hydraulic conductivity by the average
hydraulic gradient, and dividing by an assumed effective
porosity of 0.3. The upper boundary on seepage velocity
was calculated by multiplying the upper 95% confidence
intervals for hydraulic conductivity and hydraulic gradient,
then dividing by an assumed effective porosity of 0.25.
The lower boundary was calculated by multiplying the
lower 95% confidence intervals for hydraulic conductivity
and hydraulic gradient, then dividing by an assumed
porosity of 0.35. The average calculated seepage velocity
at the site was 82 meters per year. The upper boundary
was 150 meters per year, and the lower boundary was
67 meters per year. These estimates of seepage velocity
23
-------
PASQUOTANK RIVER
Approximate Scale in Meters Approximate Scale in Feet
PASQUOTANK RIVER
Approximate Scale in Meters Approximate Scale in Feet
Figure 4.1. Variation in ground-water flow calculated from eighteen
rounds of quarterly monitoring. The length of the arrow
is the distance that would be traveled by MTBE in one
year at that hydraulic gradient.
Figue 4.2. Variation in ground-water flow calculated from fourteen
rounds of monthly monitoring. The length of the arrow
is the distance that would be traveled by MTBE in one
year at that hydraulic gradient.
Table 4.1 Sensitivity analysis of the estimates of the seepage velocity of ground water at the site. These estimates were used to calculate a
first-order rate of attenuation of MTBE in ground water downgradient of the source area.
Parameter
Unit
Basis of
Boundary
Number of
Samples
Mean
Maximum
Minimum
Upper Boundary
Lower Boundary
Hydraulic
Conductivity
cm per second
95% Confidence
Interval
8
0.027
0.036
0.020
0.031
0.023
Hydraulic
Gradient
meter per meter
95% Confidence
Interval
18
0.0029
0.0067
0.00059
0.0037
0.0028
Porosity
Fraction pore
space
Range of
Literature
0.30
0.25
0.35
Seepage Velocity
meter per year
Calculated
Calculated
82
150
67
24
-------
Table 4.2. Concentration of MTBE, methane, and iron (II) at monitoring locations used to calculate the rate of attenuation of MTBE, methane,
and iron (II) with time of travel downgradient of the location with the highest concentration.
Location
Source Area
CPT-1
CPT-3
Downgradient of
Source
CPT-5
ESM-14
ESM-3
ESM-9
ESM-10
ESM-1 1
GP-1
Distance from
Source (CPT-1)
meters
Date
Sampled
Month/Year
MTBE
Mg/l
Methane
mg/l
Iron (II)
mg/l
0
40
8/1996
8/1996
1,740
823
3.5
13.5
34
96
70
104
134
180
195
238
250
8/1996
8/1996
6/1999
8/1996
8/1996
6/1999
8/1996
672
383
319
<1
9.7
13.5
<1
5.9
7.8
1.3
3.24
4.6
0.12
1.0
56
84
86
42
20
33
59
were used to calculate time of travel of ground water from
the most contaminated location (CPT-1) to the downgradient
locations.
For various sampling locations presented in Figure 4.1,
Table 4.2 compares the distance downgradient to the
concentration of MTBE, and the concentrations of meth-
ane and iron (II). Methane is expected to be a conserva-
tive tracer in ground water once it forms, and iron (II)
appears to be conservative at this site as well. Methane
and iron (II) will be used as tracers for the plume of
contaminated ground water. Monitoring locations CPT-1,
CPT-3, CPT-5, ESM-3, ESM-10, ESM-14, and GP-1 have
high concentrations of methane and iron (II) indicating that
these locations sample the plume. These locations were
included in the calculation of the natural biodegradation
rate constant.
Although location ESM-11 is directly downgradient of the
source area (Figure 4.1), it had low concentrations of
methane (Table 4.2). The low concentration of MTBE at
location ESM-11 may have resulted from simple dilution;
as a result, location ESM-11 was not included in the
calculation of the biodegradation rate constant. In con-
trast, location ESM-9 is not directly downgradient from the
source area. However, the geochemical parameters indi-
cate that ground water sampled in this location was in the
plume of contamination. As depicted in Figure 4.1, the
plume of methane and iron (II) moves more to the east
than would be expected from the hydraulic gradient alone.
This may reflect anisotrophic flow in the aquifer, rather
than error in estimating the direction of the hydraulic gradi-
ent. Location ESM-9 was included in the calculation of the
biodegradation rate constant.
Figure 4.3 plots the logarithm of concentration of iron (II),
methane, and MTBE against the calculated travel time of
ground water downgradient of the source. In three years'
travel time, there is no attenuation of iron (II), the concen-
tration of methane is attenuated by an order of magnitude,
and the concentration of MTBE is attenuated by three
orders of magnitude. The attenuation of methane is the
best estimate of the effects of attenuation due to dilution
and dispersion. The attenuation of MTBE in excess of the
attenuation of methane must be due to natural biodegrada-
tion.
The first-order rate of attenuation was calculated by a
linear regression of the natural logarithm of concentration
on time of travel along the flow path. Table 4.3 compares
the apparent rate of attenuation, and the 95% confidence
interval on that rate, for MTBE, methane, and iron (II). The
average total rate of attenuation of MTBE was near 2.7 per
year. The attenuation of methane, which may be taken as
25
-------
O)
(U
o
c
o
o
100000
10000
1000
100
10
1
0.1
-0.5
+ Iron
• Methane
AMTBE
0.5 1.5 2.5
Travel Time (years)
3.5
Figure 4.3 Attenuation in concentrations of MTBE, methane, and iron (II) with travel time downgradient from the location with the highest
concentration of MTBE.
Table 4.3 The apparent first-order rate of attenuation of MTBE, methane, and iron (II) with time of travel downgradient from the location with
the highest concentration of MTBE.
Analyte
MTBE
MTBE
MTBE
Methane
Iron (II)
Estimate of
Plume Velocity
Upper Boundary
Average
Lower Boundary
Average
Average
Apparent Rate of
Attenuation
Upper 95%
Confidence
Interval
Lower 95%
Confidence
I nterval
per year
5.0
2.7
2.2
0.5
0.12
7.21
3.89
3.26
1.15
0.619
2.69
1.45
1.22
-0.160
-0.384
26
-------
a surrogate for attenuation of MTBE due to dilution and
dispersion, was 0.5 per year. The lower 95% confidence
interval for attenuation of MTBE does not overlap the
upper 95% confidence interval for methane attenuation.
The rates are different at 95% confidence. There was no
appreciable attenuation in concentrations of iron (II).
The rate of attenuation in the field compares well with the
rate of attenuation in the laboratory microcosm study. The
average rate of attenuation in the field was 2.7 per year.
The average rate of attenuation in the laboratory was
3.02 ± 0.52 per year at 95% confidence.
Transfer from the Entire Source Area to the
Plume
The mass transfer of MTBE from the source LNAPL to
the ground water moving underneath was estimated by
calculating the flux of MTBE moving away from the source
across the east-west transect, then comparing that flux to
the total mass of MTBE in the source area.
Figure 4.5 depicts the relationship between the locations
on the east-west transect, the direction of ground-water
flow, and the source area.
The flow-weighted average concentration of MTBE in
ground water at each location along the east-west transect
was calculated by multiplying the concentration of each
vertical sample by the hydraulic conductivity at that point,
dividing each product by the average hydraulic conductiv-
ity to produce a weighted concentration, then taking the
simple arithmetic average of the weighted concentrations.
The results are presented in Figure 4.6. The highest
concentrations of MTBE were directly downgradient of the
"hot-spot" at CPT-1.
Each location was considered to represent a length
along the transect equal to the distance to the mid points
between the neighboring locations. The linear distance
between the transect locations was 50 feet. For locations
18,17,16,15,20,21, and 22, the length that was perpen-
dicular to ground-water flow was 34 feet. For locations 25,
PASQUOTANK RIVER
Approximate Scale in Meters Approximate Scale in Feet
Figure 4.4 Locations of ground-water samples included in the
calculation of the rate of natural attenuation. The arrow
represents the average direction of ground-water flow.
The dark shaded area is the area with LNAPL. The
larger lightly shaded area is the area downgradient
where the ground water contains high concentrations of
methane and iron (II). Only wells in the shaded area
were included in the calculation of the rate of natural
attenuation.
PASQUOTANK RIVER
Direction of Ground
Water Flow
Approximate Scale in Meters Approximate Scale in Feet
Figure 4.5 Relationship between the direction of ground-water flow
and the ground-water sampling locations in the transect
sampled in December 1997. Ground-water flow vectors
were calculated from the gradients in water table
elevation in eighteen different rounds of monitoring The
length of arrow is the distance that would be traveled by
MTBE in one year of flow at that gradient.
27
-------
o
-4— '
O
O
O
00
1400
1200
B 1000
100
200
300
400
500
Distance Along Transect (feet)
Figure 4.6 Concentrations of MTBE in a transect that extends across the plume in a direction that is roughly perpendicular to ground-water
flow. See Figure 4.5 for the positions of the sampling locations identified as 15 through 25 in both Figures. Depicted at each
location are the flow-weighted average concentrations of MTBE in ground-water samples from a vertical profile extending across
the aquifer at each location.
24, and 23 the distance perpendicular to ground-water flow
was 17.1 feet. For location 19, the length was 26 feet.
For each location, the vertical interval that was averaged
to get the average concentration (25.5 feet) was multiplied
by the length perpendiculartoflowtogetthe cross section,
then by the Darcy velocity to get the flux of water. The flux
at all the locations was summed. Approximately 5,300
cubic meters of water crosses the transect per year. The
flux of water at each location was multiplied by the average
concentration at each location, then summed. The flux of
MTBE from the source area across the transect was
2.8 kg/year.
The lower boundary on the total quantity of MTBE in the
source area was 46 kg (Section 3). If this flux did not
change over time, it would take seventeen years to remove
the MTBE from the source.
When the MTBE diffuses out of the LNAPL to be swept
away by flow in the aquifer, the less soluble petroleum
hydrocarbons are left behind. As a result, the concentra-
tion of MTBE in the fuel decreases over time. It is not likely
that the mass transfer rate of MTBE from the LNAPL
remains the same, regardless of the concentration of
MTBE remaining in the LNAPL. Diffusion is the mecha-
nism that drives the mass transfer process. The diffusive
flux is proportional to concentration gradients, and the
gradients of MTBE are proportional to the remaining con-
centration of MTBE in the LNAPL. The mass transfer of
MTBE would be directly proportional to the concentration
of MTBE remaining in the LNAPL. If the rate of transfer of
MTBE to ground water is proportional to the amount of
MTBE in the source, the instantaneous rate of transfer is
0.06 per year. The average concentration at the most
contaminated location in the transect is 1200 u,g/year. At
this rate of attenuation of the source, it would require
approximately sixty years for the concentration to reach
30 u,g/liter.
Transformation of MTBE to TBA
Tertiary Butyl Alcohol (TBA) has been documented as a
transformation product of MTBE in a number of studies
(Mormile et al., 1994; Squillace et al., 1996; Church et al.,
1997). In the water samples collected in the transect
across the plume in December 1997, the concentration of
TBA was measured using solid phase micro-extraction.
The limit of quantitation was 1 u,g/l. Figure 4.7 compares
the concentration of TBA to the concentration of MTBE in
all the water samples from the transect. With two excep-
tions, the concentration of TBA was less than 200 u,g/l.
There is no evidence of accumulation of TBA in the transect
as a whole.
In general, it is difficult to determine whether TBA in
ground water was a component of the original spill, or if it
was produced from biological transformation of MTBE
(Church etal., 1997; Landmeyeretal., 1998). Two ground-
water samples had higher concentrations of TBA. In one
28
-------
H
2000
1800
1600
1400
1200
1000
800
600
400
200
0
0
500
1000
MTBE
1500
2000
2500
Figure 4.7 Relationship between the concentration of TEA in ground water and the concentration of MTBE in water samples collected in a
transect across the plume in December 1997.
of these samples there was a corresponding reduction in
the concentration of MTBE. At these locations the TEA
was probably produced from transformation of MTBE.
Figure 4.8 depicts the distribution of MTBE with depth at
three locations immediately downgradient of the source
area. Location 19 is between locations 18 and 23. Loca-
tions 18 and 23 have a peak in MTBE concentration near
2000 u,g/l at a depth of 6 meters. At location 19, at a depth
of 6 meters there is a decline in the concentration of MTBE,
with higher concentrations at depths of 5 and 7 meters.
Figure 4.9 depicts the distribution of TEA with depth at the
same locations. There is little accumulation of TEA at
locations 18 and 23. However, there is a large accumula-
tion of TEA at a depth of 6 meters at location 19, at the
same location where MTBE was depleted.
29
-------
MTBE (|ig/l)
500 1000 1500 2000 2500
TEA (|lg/l)
500 1000 1500 2000 2500
Depth
(meters)
0
Depth
(meters)
1
2 -
3
4
5 -
6
7 -
23
18
19
Figure 4.8 Depth distribution of MTBE in three locations
downgradient of the LNAPL source area. See Figure 4.5
for position of the locations on a map. Compare location
19 to location 19 in Figure 4.9.
Figure 4.9 Depth distribution of TEA in locations downgradient of
the source area. See Figure 4.5 for position of the
locations on a map. Compare location 19 to location 19
in Figure 4.8.
30
-------
SECTION 5
Summary and Conclusions
Extent of Biodegradation of MTBE
Removal of MTBE in microcosms that were supple-
mented with alkylbenzenes was extensive. There was no
evidence of MTBE removal over removal in the controls in
the first 175 days of incubation. After 385 days of incuba-
tion, there is strong evidence of removal in the living
treatment. After 490 days of incubation, there was very
extensive removal of MTBE compared to the controls.
The average concentration remaining in six replicates of
the living treatment was 81 ug/l, compared to 5680 ug/l at
the beginning of incubation. The average concentration
remaining in the control treatment after 490 days was
1470 ug/l, compared to 3330 ug/l at the beginning of
incubation. The removal in the controls was a little more
than twofold, while removal in the living microcosms was
70-fold.
Removal of MTBE in microcosms that were not supple-
mented with alkybenzenes was also extensive. There
was little evidence of removal in the first 175 days of
incubation. After 385 days, the removal of MTBE in the
living microcosms was extensive. After 490 days of
incubation, the concentration of MTBE in all six of the
replicate microcosms that were sampled was below 40 ug/l,
compared to 3110 ug/l at the beginning of incubation.
Role of BTEX Compounds
Removal of MTBE in the microcosms did not require the
presence of BTEX compounds. Toluene was entirely
depleted within 40 to 47 days, benzene was entirely
depleted within 110 days, and ethylbenzene was entirely
depleted within 175 days. During this time period there
was no evidence of removal of MTBE. The removal of
MTBE did not begin until the removal of the BTEX com-
pounds was complete.
Rate of Removal of MTBE
The first order rate of removal of MTBE in microcosms
supplemented with alkylbenzenes was 3.02 per year ±0.52
per year at 95% confidence. Removal in the correspond-
ing controls was 0.39 per year ±0.19 per year at 95%
confidence. The removal in the microcosms without
added alkylbenzenes was 3.5 per year ±0.65 per year at
95% confidence. Removal in the corresponding controls
was 0.30 per year ±0.14 per year at 95% confidence.
The apparent first order rate of removal of MTBE in the
field was a sensitive function of ground-water seepage
velocity. The rate of removal was calculated for an upper
boundary on velocity, an average velocity, and a lower
boundary on velocity. The rate was 5.0 per year at the
upper boundary, 2.7 per year at the average velocity, and
2.2 per year at the lower boundary. Methane was consid-
ered to be a conservative tracer of ground-water flow at
the site. The apparent rate of removal of methane was
taken as an estimate of attenuation along the flow path
due to dilution and dispersion. The apparent first order
rate of removal of methane at the average estimate of
seepage velocity was 0.50 ±0.65 per year.
The rate of removal of MTBE in the laboratory studies
can explain the apparent attenuation of MTBE at field
scale.
Expected Persistence of the Source of
Ground-water Contamination
The mass transfer of MTBE from the source LNAPL to
the ground water moving underneath was estimated by
calculating the flux of MTBE moving away from the source,
then comparing that flux to the total mass of MTBE in the
source area. The flux of MTBE away from the source area
in 1996 was 2.76 kg/year. The lower boundary on the total
quantity of MTBE in the source area was 46 kg. If this flux
did not change overtime, it would take seventeen years to
remove the MTBE from the source. If the rate of transfer
of MTBE to ground water is proportional to the amount of
MTBE in the source, the instantaneous rate of transfer is
0.06 per year. The average concentration at the most
contaminated location in the transect is 1200 ug/l. At this
rate of attenuation of the source, it would require approxi-
mately sixty years for the concentration to reach 30 ug/l.
Production and Depletion of TBA
Tertiary Butyl Alcohol (TBA) has been documented as a
transformation product of MTBE in a number of studies.
At the Old Fuel Farm Site, there is no evidence of accu-
mulation of TBA in the ground-water plume as a whole.
With two exceptions, the concentration of TBA in ground
water downgradient of the source area was less than 200
ug/l. Ground water from a location immediately
downgradient of the source area had a higher concentra-
tion of TBA, near 2000 ug/l. In this sample there was a
corresponding reduction in the concentration of MTBE. At
this location the TBA was probably produced from trans-
formation of MTBE.
31
-------
Geochemical Context of the Plume that
Biodegraded MTBE
The entire MTBE plume is contained within a plume of
methane. Methane concentrations generally exceed
3.0 mg/l, and often exceed 10 mg/l . Concentrations of
methane average 7 mg/l, which corresponds to 9 mg/l of
hydrocarbon originally metabolized.
Ground water in the region of the aquifer that contains
MTBE and BTEX compounds is also depleted of sulfate.
Sulfate concentrations are reduced from a background of
near 28 mg/l to less than 4 mg/l; many regions have less
than 1 mg/l. A depletion of 24 mg/l of sulfate would oxidize
5 mg/l of fuel hydrocarbons. The same regions that are
depleted in molecular oxygen and sulfate have significant
accumulations of iron (II) . Background concentrations of
iron (II) are less than 0.1 mg/l. Many regions of the aquifer
with MTBE and BTEX compounds have iron (II) concentra-
tions greater than 50 mg/l. This accumulation of iron (II)
would be capable of oxidizing 3 mg/l of hydrocarbons.
The plume is also undergoing extensive anaerobic oxi-
dation of petroleum hydrocarbons, as well as fermentation
of hydrocarbons to methane. The hydrocarbon metabo-
lism through sulfate and iron oxidation is approximately
equivalent to the hydrocarbon metabolism through
methanogenesis. The amount of hydrocarbon metabo-
lized through anaerobic pathways is about ten times the
amount degraded with molecular oxygen.
32
-------
References
Chapelle, F.H., 1999, Bioremediation of petroleum
hyrocarbon-contaminated ground water: the
perspectives of history and hydrology, Ground Water
37(1):122-132.
Church, C.D., Isabelle, L.M., Pankow, J.M., Rose, D.L.,
and Tratnyek, P.G., 1997, Method for determination of
methyl terf-butyl ether and its degradation products in
water, Environ. Sci. Technol. 31(12):3723-3726.
Landmeyer, J.E., Chapelle, F.H., Bradley, P.M.., Pankow,
J.F., Church, C.D., and Tratnyek, P.G., 1998, Fate of
MTBE relative to benzene in a gasoline-contaminated
aquifer (1993-98), Ground Water Monitoring and
Remediation, Fall, 1998, pp. 93-102.
Mormile, M.R., Liu, S, and Suflita, J.M., 1994, Anaerobic
biodegradation of gasoline oxygenates:extrapolation
of information to multiple sites and redox conditions,
Environ. Sci. Technol. 28(9):1727-1732.
Squillace, P.J., Zogorski,J.S., Wilber,W.G., and Price, C.V.,
1996, Preliminary assessment of the occurrence and
possible sources of MTBE in ground water in the
United States, 1993-1994, Environ. Sci. Technol.
30:1721-1730.
Suflita, J.M., and Mormile, M.R., 1993, Anaerobic
biodegradation of known and potential gasoline
oxygenates in the terrestrial subsurface, Environ. Sci.
Technol. 27(5):976-978.
U.S. Environmental Protection Agency. 1996. Drinking
Water Regulations and Health Advisories. Washington,
D.C.
Wiedemeier, T.H., Rifai, H.S., Newell, C.J., and Wilson,
J.T., 1999, Natural Attenuation of Fuels and Chlorinated
Solvents in the Subsurface, John Wiley & Sons, New
York, ISBN 0-471-19749-1.
Wilson, J.T., Cho,J.S., Beck, F.P., and Vardy,J.A., 1997,
Field Estimation of Hydraulic Conductivity for
Assessments of Natural Attenuation. In Proceedings
of the Fourth International In-Situ and On-site
Bioremediation Symposium, New Orleans, LA, pp.
309-314.
Yeh, C.K., and Novak, J.T., 1994, Anaerobic biodegradation
of gasoline oxygenates in soils, Water Environ. Res.,
66(5):744-752.
33
-------
Appendix A. Temporal Variation in the Hydraulic Gradient and the Direction
of Ground-water Flow
Ground-water flow at the site is strongly influenced by
the elevation of the Pasquotank River. At Elizabeth City,
North Carolina, the Pasquotank River makes a transition
from a conventional river to a large estuary. At Elizabeth
City, the Pasquotank River is a few hundred feet wide. At
the U.S. Coast Guard Air Station, only a few miles farther
down river, the Pasquotank is more than two miles wide.
The average elevation of the Pasquotank River at the fuel
farm site is near sea level. The elevation of the river is not
controlled by recent precipitation and runoff, as is usually
expected for a river. The strongest influence on the
elevation of the river is the recent direction of the wind,
producing a phenomenon known as a wind seiche. The
trend of the valley is from northwest to southeast. Friction
from strong winds coming from the north will drive water
out of the river valley, and lower the elevation of the river.
Winds from the south drive water into the valley, and raise
the water table.
This effect is illustrated in the data in Figure A.1. A
pressure transducer was used to record the elevation of
the Pasquotank River every fifteen minutes over a time
interval extending from September 5, 1996 to October 30,
1996. To minimize the confounding effects of wave action
on the measurement of water table elevation, the trans-
ducerwas located in a drainage ditch in close communica-
tion with the river. Over this two-month interval, the eleva-
tion of the Pasquotank River varied from 1.5 feet above
sea level to 1.5 feet below sea level. The most rapid
changes in elevation were drops in elevation on the order
of 1.5 feet within one or two days that were associated with
cold fronts that came through in the first week of Septem-
ber, and the first week of October. There was also a
diurnal cycle that varied from 0.1 to 0.5 feet.
These changes in elevation of the receptor of the
ground-water plume are large, compared to the change in
elevation of ground water across the fuel farm site. As an
illustration, examine Figure A.2 and identify Well ESM-14
located in the center of the study area. The average
elevation of ground water in Well ESM-14 over 27 rounds
of quarterly or monthly sampling was 1.25 feet above sea
level. As a consequence of the temporal variation in the
elevation of the Pasquotank River, the hydraulic gradient
and the ground-water flow velocity across the site vary
widely from one sampling event to another.
Figure A.3 plots the elevation of the water table against
time for four representative monitoring wells extending
across the site (see Figure A.2 forthe location of the wells).
Over the period from March 1994 to December 1998, the
elevation of the water table in individual wells changed
from 1.5 to 2.0 feet. The elevation of the water table in all
the wells tended to track each other over time. In some
time intervals the elevations of ground water in all the wells
are nearly the same. At other intervals, the hydraulic
gradient is more strongly expressed. There is no obvious
correlation of the hydraulic gradient to seasons of the year,
or to the average elevation of ground water across the site.
-0.8
-5
15 25 35 45 55 65
Time (days)
Figure A.1 Variation in elevation of water in the Pasquotank River over a time interval extending from September 5, 1996 to October 30, 1996.
34
-------
283200-
283150-
283100-
283050-
283000-
PASQUOTANK
RIVER
862700 862750 862800 862850 862900 862950 863000 863050
0 50 100
SCALE IN METERS
Figure A.2 Location of the permanent monitoring wells used to estimate the hydraulic gradient and direction during each round of monitoring.
3.5
a 2.5 ^
.a
co
* 2
"co
1 H
CO
CD
Lu 0.5 -
0
Dec-93 Dec-94 Dec-95 Dec-96 Dec-97 Dec-98 Dec-99
Date of Sampling
Figure A.3 Variation in elevation of the water table at the fuel farm site overtime. Consult Figure A.1 for the location of the monitoring wells.
Well ESM-10 is closest to the Pasquotank River, the point of ground-water discharge. Wells ESM-14, ESM-6, and ESM-7 are
farther inland.
35
-------
Water table elevations were available from eighteen
rounds of quarterly monitoring and for fourteen rounds of
monthly monitoring. The elevations of ground water in
wells depicted in Figure A.2 are presented in Table A.1 for
eighteen rounds of sequential quarterly sampling, and
Table A.2 for fourteen rounds of sequential monthly sam-
pling. There are four sampling dates common to the
quarterly monitoring data and the monthly monitoring data.
Rather than take one or a few rounds of sampling and use
professional judgment to construct ground-water elevation
contours that would be representative of the site, a simple
statistical approach was used to give equal weight to each
round of sampling.
The aquifer containing the plume of contamination is
semi-confined across the entire site. The superficial layer
of silty clay, and the concrete operational apron prevent
recharge of precipitation in the study area. Recharge
occurs far inland from the Pasquotank River. The superfi-
cial confining layer extends out into the bed of the river.
Discharge occurs some distance into the river, not at its
bank. As a result, ground-water flow at the site is con-
trolled by regional flow in the aquifer. As an approximation,
the ground-water elevation at any round of sampling will be
considered to be a linear plane in three-dimensional space.
The slope of the plane is the hydraulic gradient and flow
direction of ground water.
Table A.3 provides a summary of the regression on the
eighteen rounds of quarterly monitoring. The table pro-
vides the equation of the regression, the number of wells
involved in the regression at each date, the coefficient of
correlation r2, the variance of the estimate of elevations,
and the fitted hydraulic gradient and direction for each
round of sampling. The seepage velocity of the ground
water was calculated from Darcy's Law using an average
estimate of hydraulic conductivity of 0.27 cm/sec and an
assumed effective porosity of 0.3. In general, the assump-
tion that the water table was a plane was reasonable. The
coefficient of correlation exceeded 0.8 for fifteen of the
eighteen rounds of sampling. Table A.4 provides the sum-
mary of the regression of the fourteen rounds of monthly
monitoring. The coefficient of correlation exceeded 0.8 for
nine of the fourteen rounds of sampling.
The results of the individual regressions are presented in
graphical form in Figures 4.1 and 4.2.
The average direction of ground-water flow was calcu-
lated by weighting the direction of flow on any particular
round of sampling by the hydraulic gradient at that particu-
lar round, then taking an average of the weighted flow
directions. For the eighteen rounds of quarterly sampling,
the average flow direction was 8.6 degrees west of north,
with a standard deviation of 9.1 degrees. The mean of the
hydraulic gradient was 0.00252 with a 95% confidence
interval of 0.0010. For the fourteen rounds of monthly
sampling, the average flow direction was 8.7 degrees west
of north, with a standard deviation of 23 degrees. The
mean of the hydraulic gradient was 0.00252 with a 95%
confidence interval of 0.00070. The results of the regres-
sions for each of the eighteen rounds of quarterly monitor-
ing are presented in Figures A.4 through A.18. The
measured elevations submitted to the regressions are
listed in the figures to allow a well-by-well evaluation of the
fit of the regression.
36
-------
Table A.1 Elevation of the water table in permanent monitoring wells during eighteen rounds of quarterly
monitoring extending from September 1994 through December 1998. The elevations are reported in
feet above mean sea level. Compare Figure A.1 for the location of the monitoring wells.
Well
ESM-3
ESM-4
ESM-6
ESM-7
ESM-9
ESM-10
ESM-11
ESM-14
Table A.2.
Well
ESM-3
ESM-4
ESM-6
ESM-7
ESM-9
ESM-10
ESM-11
ESM-14
9/94 12/94
0.74 0.86
0.44 0.89
1.08 1.03
1.63 1.55
0.73 0.92
0.68 0.87
0.4 0.9
0.92 1.03
3/95 5/95
1.08 1.23
0.89 1.11
2.18 1.38
3.16 1.65
0.95 1.23
0.83 1.19
0.8 1.09
1.34 1.33
8/95
1.35
1.36
1.2
1.17
1.33
1.3
1.38
1.27
12/95 3/96 6/96
0.49 0.97 1.32
0.17 0.65 1.05
0.97 1.71 1.85
1.67 2.56
0.42 0.89 1.25
0.37 0.8 1.16
0.09 0.61 1.02
0.68 1.24 1.5
9/96
1.5
1.45
1.89
2.45
1.45
1.39
1.45
1.64
12/96 3/97
1.4 1.16
1.18 0.84
1.81 1.73
2.3
1.34 1.09
1.32 1.11
1.13 0.78
1.53 1.39
6/97
1.31
1.53
1.28
1.66
1.23
1.25
1.34
1.24
9//97 12/97
1.29 0.5
1.37 0.29
1.14 0.71
1 1.05
1.31 0.52
1.32 0.44
1.34 0.26
1.23 0.59
3/98 6/98
1.56 1.27
1.18 1.19
2.35 1.56
2.82 1.95
1.47 1.25
1.4 1.34
1.06 1.15
1.79 1.37
9/98
1.03
1.21
1.06
1.76
1.08
1.13
1.17
1.07
12/98
0.58
0.83
0.74
1.38
0.67
0.6
0.59
0.59
Elevation of the water table in permanent monitoring wells during fourteen rounds of monthly
monitoring extending from September 1994 through December 1998. The elevations are reported in
feet above mean sea level. Compare Figure A.1 for the location of the monitoring wells.
2/11/98 3/10/98 4/7/98
2.34 2.22
2.15 1.98
3.14 2.75
3.6 3.13
2.17 2.11
2.12 2.09
2.1 1.94
2.5 2.37
1.53
1.29
1.98
1.54
1.37
1.32
2.07
1.61
5/13/98 6/16/98 7/9/98
1.06
0.45
1.59
2.16
0.91
0.85
0.35
1.21
1.41 1.16
1.43 1.4
1.49 1.17
1.73
1.39 1.06
1.4 0.95
1.44 0.91
1.45 1.19
8/6/98
0.52
0.32
0.82
1.38
0.47
0.47
0.29
0.65
9/2/98 10/1/98
1.36 1
1.37 1
1.78 1
2.26 1
1.46 1
1.33 1
1.36 1
1.39 1
.03
.21
.06
.76
.08
.13
.17
.07
11/14/98 12/7/98 1/6/99
0.85 0.58
1.07 0.83
0.88 0.74
1.63 1.38
0.91 0.67
0.99 0.6
1.02 0.59
0.9 0.59
0.78
1.15
1.62
0.86
0.77
0.79
2/1/99
0.94
0.69
1.42
1.79
0.66
0.68
0.62
0.98
3/5/99
1.05
0.73
1.36
1.64
0.94A
0.9
0.69
1.06
37
-------
Table A.3 Equation of a linear plane that was fit using a least-squares regression through the elevation of the water table in permanent
monitoring wells during each of eighteen rounds of quarterly sampling. The plane is in an x,y,z coordinate system where x increases
toward the east, y increases toward the north, and z increases with elevation above mean sea level. The equation is in the form
Ax+By+C+z where x and y are the grid location in UTM meters and z is the elevation of the water table in feet.
CO
oo
Date
Sept. 1994
Dec. 1994
Mar. 1995
May 1995
Aug. 1995
Dec. 1995
Mar 1996
June 1996
Sept. 1996
Dec. 1996
Mar. 1997
June 1997
Sept. 1997
Dec 1997
Mar. 1998
June 1998
Sept. 1998
Dec. 1998
n
number
of wells
8
8
8
8
8
8
8
7
8
8
7
8
8
8
8
8
8
8
A
x coefficient
1.03334E-03
7.0471 9E-04
-3.91751E-04
4.801 94E-04
-1.66687E-05
8.39634E-04
5.81895E-04
1.50708E-04
1.07050E-04
4.21094E-04
5.70420E-04
-2.55409E-04
-9.2901 3E-05
7.32342E-04
1.01025E-04
3.93228E-04
7.92146E-04
4.001 85E-04
B
y coefficient
-3.25785E-03
-1.76193E-03
-6.66909E-03
-1.49026E-03
5.84579E-04
-4.21060E-03
-5.39052E-03
-3.49088E-03
-2.82168E-03
-3.16140E-03
-4.02255E-03
-6.23482E-04
9.90993E-04
-2.09248E-03
-4.85332E-03
-2.09999E-03
-1.41290E-03
-1.85079E-03
C
constant
31.950
-108.013
2228.290
9.037
-149.892
468.710
1025.850
860.107
708.465
533.568
648.421
398.315
-199.269
-38.689
1289.130
256.873
-282.118
179.655
r2
0.96785
0.83375
0.95519
0.96893
0.86551
0.97192
0.96560
0.91667
0.90858
0.97066
0.89316
0.27942
0.93309
0.97696
0.96975
0.92957
0.50305
0.66928
z
Variance
0.00700
0.01227
0.04445
0.00141
0.00110
0.01028
0.02110
0.01053
0.01660
0.00613
0.01688
0.02470
0.00144
0.00205
0.01529
0.00672
0.03959
0.03389
Hydraulic
Gradient
0.0034
0.0019
0.0067
0.0016
0.0006
0.0043
0.0054
0.0035
0.0028
0.0032
0.0041
0.0007
0.0010
0.0022
0.0049
0.0021
0.0016
0.0019
Direction
(degrees
east from north)
-18
-22
3
-18
-2
-11
-6
-2
-2
-8
-8
22
-5
-19
-1
-11
-29
-12
Velocity
meters
/year
97
54
190
44
17
122
154
99
80
91
115
19
28
63
138
61
46
54
Velocity
feet/year
318
177
622
146
54
400
505
325
263
297
378
63
93
206
452
199
151
176
-------
Table A.4 Equation of a linear plane that was fit using a least-squares regression through the elevation of the water table in permanent
monitoring wells during each of fourteen rounds of monthly sampling. The plane is in an x,y,z coordinate system where x increases
toward the east, y increases toward the north, and z increases with elevation above mean sea level. The equation is in the form
Ax+By+C+z where x and y are the grid location in UTM meters and z is the elevation of the water table in feet.
CO
CD
Date
11-Feb-98
10-Mar-98
7-Apr-98
13-May-98
16-Jun-98
9-Jul-98
6-Aug-98
2-Sep-98
1-Oct-98
14-Nov-98
7-Dec-98
6-Jan-99
1-Feb-99
5-Mar-99
n
number
of wells
8
8
8
8
7
8
8
8
8
8
8
6
8
8
A
x coefficient
-8.05822E-04
-7.31494E-05
-1.75445E-03
9.33425E-04
-2.64965E-04
-1.28825E-04
7.07976E-04
2.28660E-04
7.92146E-04
8.51723E-04
4.001 85E-04
3.50042E-04
-7.16128E-04
2.01839E-04
B
y coefficient
-4.33641 E-03
-3.331 36E-03
9.29881 E-05
-4.79741 E-03
-1.24188E-04
-1.67917E-03
-2.891 56E-03
-2.54803E-03
-1.41290E-03
-1.43532E-03
-1.85079E-03
-2.31329E-03
-3.26365E-03
-2.5671 9E-03
C
constant
1925.990
1008.970
1489.030
554.451
265.225
587.936
208.726
525.917
-282.118
-327.332
179.655
354.144
1543.220
553.994
r2
0.96171
0.96560
0.19250
0.94123
0.65632
0.61699
0.95649
0.89449
0.50305
0.46232
0.89316
0.93028
0.95728
0.95110
z
Variance
0.01669
0.00833
0.09695
0.02886
0.00006
0.03750
0.00759
0.01567
0.03959
0.04834
0.03389
0.01330
0.01074
0.00695
Hydraulic C
Gradient (<
east
0.00441
0.00333
0.00176
0.00489
0.00029
0.00168
0.00298
0.00256
0.00162
0.00167
0.00189
0.00234
0.00334
0.00258
lirectii
degrei
from i
11
1
-87
-11
65
4
-14
-5
-29
-31
-12
-9
12
-4
Velocity
meters
125
95
50
139
8
48
84
73
46
47
54
66
95
73
Velocity
feet/year
411
310
164
455
27
157
277
238
151
155
176
218
311
240
-------
SCALE IN METERS
SCALE IN METERS
SCALE IN METERS
Figure A.4 Direction and gradient of ground-water flow on a sample Fi9ure A'5 Dir.ectio" and ****£?*ground-water flow on a sample
date in September 1994. date ln December 1994-
40
-------
50 100
SCALE IN METERS
SCALE IN METERS
Figure A.6 Direction and gradient of ground-water flow on a sample Figure A.7 Direction and gradient of ground-water flow on a sample
date in March 1995. date in May 1995.
50 100
SCALE IN METERS
SCALE IN METERS
Figure A.8 Direction and gradient of ground-water flow on a
sample date in August 1995.
Figure A.9 Direction and gradient of ground-water flow on a
sample date in December 1995.
41
-------
SCALE IN METERS
0 50 100
SCALE IN METERS
Figure A.10 Direction and gradient of ground-water flow on a
sample date in March 1996.
Figure A.11 Direction and gradient of ground-water flow on a
sample date in June 1996.
SCALE IN METERS
JOO
SCALE IN METERS
Figure A.12 Direction and gradient of ground-water flow on a
sample date in September 1996.
Figure A.13 Direction and gradient of ground-water flow on a
sample date in December 1996.
42
-------
SCALE IN METERS
SCALE IN METERS
Figure A.14 Direction and gradient of ground-water flow on a
sample date in March 1997.
Figure A.15 Direction and gradient of ground-water flow on a
sample date in June 1997.
joo
SCALE IN METERS
SCALE IN METERS
Figure A.16 Direction and gradient of ground-water flow on a
sample date in September 1997.
Figure A.17 Direction and gradient of ground-water flow on a
sample date in December 1997.
43
-------
JOO
SCALE IN METERS
Figure A.18 Direction and gradient of ground-water flow on a
sample date in March 1998.
SCALE IN METERS
Figure A.19 Direction and gradient of ground-water flow on a
sample date in June 1998.
0 50 100
SCALE IN METERS
SCALE IN METERS
Figure A.20 Direction and gradient of ground-water flow on a
sample date in September 1998.
Figure A.21 Direction and gradient of ground-water flow on a
sample date in December 1998.
44
-------
Appendix B: Geochemical Context of the MTBE Plume
Figures B.1 and B.2 compare the distribution of methane
and MTBE in the north-south transect (see Figure 3.7).
Figures B.3 and B.4 compare the distribution of methane
and MTBE in the east-west transect. The entire MTBE
plume is contained within a plume of methane. Methane
concentrations generally exceeded 3.0 mg/l, and often
exceeded 10 mg/l. In general, the distribution of MTBE
was contained within the distribution of methane. How-
ever, along the east-west transect, at the sampling loca-
tions 50, 100, and 150 feet along the transect, the highest
concentrations of MTBE extended about three feet deeper
into the aquifer than the higher concentrations of methane.
In general, this aquifer is strongly methanogenic. Concen-
trations of methane averaged 7 mg/l, which corresponds to
9 mg/l of hydrocarbon originally metabolized. (Wiedemeier
et al., 1999, table 5.3, page 214).
Figures B.5 through B.8 compare the other geochemical
parameters in the ground water along the north-south
transect. Figure B.5 shows that the MTBE plume is
contained within a region of the aquifer that is depleted of
molecular oxygen. Many regions of the aquifer have less
than 0.1 mg/l oxygen. Background concentrations of
oxygen in regions of the aquifer that are not impacted by
the fuel spill are near 3.6 mg/l. The depletion in oxygen
would account for 1 mg/l of petroleum hydrocarbon.
Ground water in the region of the aquifer that contains
MTBE and BTEX compounds is also depleted of sulfate
(Figure B.6). Sulfate concentrations are reduced from a
background of near 28 mg/l to less than 4 mg/l. Many
regions have less than 1 mg/l. A depletion of 24 mg/l of
sulfate would oxidize 5 mg/l of fuel hydrocarbons.
The same regions that are depleted in molecular oxygen
and sulfate have significant accumulations of iron (II)
(Figure B.7). Background concentrations of iron (II) are
less than 0.1 mg/l. Many regions of the aquifer with MTBE
and BTEX compounds have iron (II) concentrations greater
than 50 mg/l. This accumulation of iron (II) would be
capable of oxidizing 3 mg/l of hydocarbons.
The plume is undergoing extensive anaerobic oxidation
of petroleum hydrocarbons, as well as fermentation of
hydrocarbons to methane. The hydrocarbon metabolized
through sulfate and iron reduction is approximately equiva-
lent to the hydrocarbon metabolized through
methanogenesis. The amount of hydrocarbon metabo-
lized through anaerobic pathways is about seventeen times
the amount degraded with molecular oxygen.
The pH of the plume is generally near 6.5 and is below
6.0 only in the ground waterthat is in direct contact with the
LNAPL. Under these conditions, carbon dioxide produced
through oxidation of petroleum hydrocarbons will react with
carbonate minerals in the aquifer matrix to produce bicar-
bonate alkalinity in the ground water. Figure B.8 shows
that as much as 200 mg/l of alkalinity was produced by
oxidation of petroleum hydrocarbons. This corresponds to
88 mg/l of carbon dioxide produced or 28 mg/l of TPH
consumed. There is more than enough carbon dioxide
production to account for the depletion of oxygen and
sulfate, and production of iron (II) and methane.
45
-------
Methane (mg/1) at North-South Transect
^ ao^
xti' 3.U
| 6-°"
j2 9.0-
£ 12.0-
1 15.0-
^ 18.0-
•§21.0-1
PP24.0
^27.0-1
n 30.0-
0 100 200 300 400 500 600 700 800 900
Location on Transect (ft)
Figure B.1 Distribution of methane along the north-south transect, collected in August 1996. Distance along the transect extends from south
to north (bottom to top in Figure 3.7), in the direction of ground-water flow.
MTBE (jig/1) at North-South Transect
^ 0.0-
? 3.0-
1 6-°
t 9.0-
| 12.0-
t 15.0-
PQ
18.0-
21.0-
24.0-
27.0-
30.0-
30
0
100 200 300 400 500 600 700 800 900
Location on Transect (ft)
Figure B.2 Distribution of MTBE along the north-south transect, collected in August 1996. Distance along the transect extends from south to
north (bottom to top in Figure 3.7), in the direction of ground-water flow.
46
-------
£ 0-
CD
03
3-
6-
9-
t21'
o24-
27-
M ethane (mg/L) at East - West Transect
50 100 150 200 250 300 350
Location on Transect (ft)
i
400
450
Figure B.3 Distribution of methane along the east-west transect, collected in December 1997. Distance along the transect extends from west
to east (left to right in Figure 3.7), opposite the direction of ground-water flow.
OH
3
6
9
12H
15
£ 21-1
I 24-
i
0
MTBE (jig/1) at East-West Transect
50 100 150 200 250 300 350
Location on Transect (ft)
400 450
Figure B.4 Distribution of MTBE along the east-west transect, collected in December 1997. Distance along the transect extends from west to
east (left to right in Figure 3.7), opposite the direction of ground-water flow.
47
-------
Spatial Distribution of Oxygen (mg/1)
Sl2<
f
5i»
24-
30-
0 100 200 300 400 500 600
Location on Transect (ft)
700 800 900
Figure B.5 Distribution of oxygen along the north-south transect, collected in August 1996. Distance along the transect extends from south to
north (bottom to top in Figure 3.7), in the direction of ground-water flow.
Spatial Distribution of Sulfate (mg/1)
0
100 200 300 400 500 600
700 800 900
Location on Transect (ft)
Figure B.6 Distribution of sulfate along the north-south transect, collected in August 1996. Distance along the transect extends from south to
north (bottom to top in Figure 3.7), in the direction of ground-water flow.
48
-------
Spatial Distribution of Iron (II) (mg/1)
f
O.O
6.0-
12.0
18.0
24.0
30.0
26—
i
0
100 200
300 400
500 600 700 800 900
Location on Transect (ft)
Figure B.7 Distribution of iron (II) along the north-south transect, collected in August 1996. Distance along the transect extends from south to
north (bottom to top in Figure 3.7), in the direction of ground-water flow.
Spatial Distribution of Alkalinity (mg/1)
30
0
100 200 300 400 500 600
700 800 900
Location on Transect (ft)
Figure B.8 Distribution of alkalinity along the north-south transect, collected in August 1996. Distance along the transect extends from south
to north (bottom to top in Figure 3.7), in the direction of ground-water flow.
49
------- |