EPA/600/A-95/142
Field-Derived Transformation Rates for
Modeling Natural Bioattenuation
of Trichloroethene and its Degradation Products
James W. Weaver* John T. Wilsont Don H. Kampbell-
Mary E. Randolphs
Abstract
Subsurface contamination by trichloroethene (TCE) was detected at the St. .Joseph,
Michigan site in 1982. The contamination resulted from disposal of TCE and other
chemicals at an industrial facility located near the eastern shore of Lake Michigan.
Investigation of the site revealed the presence of TCK degradation products, including
three dichloroethene isomers, vinyl chloride and ethene. The plume was found to be
depleted of oxygen and methanogemc at certain depths. Portions of the plume were
sampled by slotted auger borings in 1991 and 1992. In August of 1994, water samples
were taken from a barge .situated about 100 in off-shore in Lake Michigan, Each round of
samples were taken along transects that crossed the width of the plume; as determined
in the field by gas chromatography. From the data set, the average concentration of each
chemical and net apparent loss coefficients between appropriate pairs of transects were
calculated. The loss rates were calculated from the .solution of the one-dimensional
advective-dispersive-reactive transport equation. The net apparent rate coefficients
and a set of coupled reaction rate equations were used to extract the apparent loss
coefficients from the field data.
1 Background
Ground water contamination was detected at the St. Joseph Michigan Superfund site in
1982. The contamination was found as dissolved organic chemicals in the ground water,
including trichloroethene (TCE). TCE is sparingly soluble (1100 mg/L), but is immiscible
with water so the two liquids remain distinct. Because of the industry located at the site, the
dissolved TCE found in the aquifer likely originated as a solvent that had been discharged
to the subsurface. At St. Joseph, the mass of TCE and its distribution in the subsarface
are unknown. Further, the rate and timing of the release are unknown as these records
are simply unavailable. Together these characteristics of the release define the source of
contamination, and, when used in a model, define the boundary conditions. Thus the real
system that we wish to simulate was not completely characterized.
The site is located four miles south of St. Joseph and one-half mile east of Lake Michigan.
Since 1942, the site has been used for auto parts manufacturing. The aquifer is primarily
composed of medium, fine and very fine sands that are of glacial origin. The base of the
aquifer is defined by a clay layer that lies between 21 and 29 meters below the ground
'Research Ilydrologist. United Stales Environmental Protection Agency, Ada. OK.
1 Research Microbiologist, United Slates Environmental Protection Agency. Ada, OK.
'Research Chemist, United States Environmental Protection Agency, Ada. OK.
sMicrobiologist, United States Environmental Protection Agencv. Ada. OK.

-------
2 Weaver et al.
surface. The elevation of the clay layer increases toward Lake Michigan. The suspected
source was situated over a ground water divide, so the contamination was divided into
eastern and western plumes. Both plumes were found to contain trichloroethene (TCE),
cis- and trans- 1,2-dichloroethene (c-DCE and t-DCE), 1,1-dichloroethene (1.1-DCE) and
vinyl chloride (VC). Previous investigation indicated that natural anaerobic degradation of
the TCE was occurring in the western plume, because of the presence of transformation
products and significant levels of ethene and methane [4], [6].
Anaerobic biodegradation of TCE occurs through successive dechlorination from
trichloroethene to dichloroethene, vinyl chloride and ethene [5]. The process produces
three isomers of DCE (1,1-DCE. cis-l,2-DCE, and trans-l,2-DCE). Although TCE was
commonly used as a degreaser. the DCEs were rarely used in industry and ethene would not
be expected in most ground waters. Thus the presence of these compounds are indicative
of degradation when found in oxygen-depleted ground water.
Modeling the degradation requires, first, knowledge that degradation is occurring and,
second, values of rate constants that describe the process. Implicit in the work of [4] and [6]
is the fact that degradation of TCE at the St. Joseph site was not predicted from theoretical
considerations; rather degradation of TCE was established from the field data. The purpose
of this paper is to show a portion of the data set and to estimate the rate constants.
2 Data Summary
Data collected in 1984 and 1985 were used for general delineation of the western plume
(Figure 1). Borings were made with a 10 foot long slotted auger along a transect which
runs from the industry toward the lake. This data set formed the basis for selecting later
sampling locations. Water samples were taken in October 1991 and March 1992 from a
5 foot-long slotted auger [11]. These borings form transects that crossed the contaminant
plume. Three transects (1, 2 and 3 on Figure 1) which include 17 borings were completed
near the source of the western plume [4]. Data from the first three transects has been
analyzed by [8]. In 1992, two additional transects (4 and 5 on Figure 1) consisting of 9
additional slotted auger borings were completed. These two transects were chosen to sample
the plume in the vicinity of Lake Michigan. In each boring, water samples were taken in 5
foot intervals from the water table to the base of the aquifer. On-site gas chromatography
was used to determine the width of the plume and find the point of highest concentration
in each transect. Three of the transects (2, 4, and 5) were roughly perpendicular to the
contaminant plume. Of the remaining transects, transect 1 crosses the plume at an angle
and transect 3 lies along the length of the plume. In August 1994, data were collected from
a transect located about 100 m off-shore. A transect was formed that was roughly parallel
to the shore line. Water samples were taken from lake sediments with a geoprobe mounted
on a barge [11]. Data from the lake transect showed the location of the plume by reduction
in dissolved oxygen and the redox potential. The perpendicular on-shore and lake transects
were used for the analysis presented below.
As expected for anaerobic processes, significant methane concentrations occur where
the dissolved oxygen concentration is below 50 micromoles/L (about 2 rng/L). Conversely,
most high oxygen concentrations occur where the methane concentration is low. Variation
in concentration occurring on a scale smaller than the length of the auger is not accurately
represented, as waters of differing chemistry may mix upon sampling. This may explain
why there are a few data points that have high methane and high oxygen concentrations.
The entire chlorinated ethene (TCE, the DCEs and VC) data set is plotted as a chlorine

-------
Field-Derived Transformation Rates 3
iNSECTS
TRANSECT 1
TRANSECT 4
TRANSECT 2
JRANSECT 3
Fig. 1. St. Josayh. Michigan Suvarfh
number, Nch that is defined as
, ,	HmCi
(1)	Nci - ^
E
-------
4 Weaver et al.
CD
_Q
CD
C
o
_c
O
3
2.5
-n	Her
~
~
~
~
~
~ ~ ~
Jn
~
~
~dp S ~
~
~
~
~
a
2
1.5
~
~
~ ~
[ED
~ ~
~
~
~
~ ~~~ED
~
0.5
aA A
S5a
aAa
~
~
a 02 < 2 mg/L
~ 02 > 2 mg/L
0
141.
0 100 200 300 400
Micromolar Oxygen Concentration
Fig. 2. Composited chlorinated eth.ene concentrations as chlorine number plotted, against the
dissolved, oxygen concentration
the transect by one-half the distance to the innermost subsequent data point. If the field
sampling had perfectly crossed the plume, then the concentrations on the outermost blocks
would be zero and the estimate of concentration would be independent of the assumption on
the outermost blocks. By following this procedure, the measured chemical concentrations
are not extrapolated below the clay layer underlying the site. At each transect, the average
concentration was calculated by summing over the blocks and dividing by the area of the
transects.
In Table 1, concentration estimates are presented for the perpendicular transects
-------
Field-Derived Transformation Rates 5
CD
_Q
E
Z5
0
C
o
JZ
O
3
"TT" _ _
K A
L/V*
A
if.
a A A a
A
a	A
A A	A
A
A
A
A
~ A .
A A
aA A
2^ 4aA4
A. .	AAa ^
~ A
A
A A
A A
A M
A.
A A
A A
~\A
Aa
A
A
A
A
AA
A
A
A
IP & A
A A
A A S
1 I D A
A
A
A A
A
A
A A
A
A A

~
A
A
AAA
A A 4«
A
A
A
0
A AA
A
A
A A
A
a 02 < 2 mg/L
~ 02 > 2 mg/L
0 200 400 600 800 1000
Micromolar Methane Concentration
Fig. 3. Composited chlorinated ethane concentrations as chlorine number plotted, against, the
methane concentration
ordered from furthest up gradient (transect 2) to furthest down gradient (transect 5).
The concentration estimates are based only upon blocks from the anaerobic portion of the
aquifer. These were used below for extracting degradation rate constants from the field
data set. All of the chlorinated ethenes show decreasing concentration with distance down
gradient. Thus, all of the rate coefficients developed below reflect a net loss of the species.
The chloride concentrations increase down gradient as expected from the dechlorination of
the ethenes.
-------
6 Weaver et al.
Tabu-; 1
Transect-Averaged Concentrations (/.ig/Ljfrom the Anaerobic Zone
Chemical
Transect 2
Transect 4
Transect 5
Lake Transect
TCE
7411
864
30.1
1.4
c-DCE
9117
1453
281
(0.80)
t-DCE
716
34.4
5.39
1.1
1,1-DCE
339
24.3
2.99
nd
VC
998
473
97.7
(0.16)
Ethene
480
297
24.2
no data
Sum of the Ethenes
19100
3150
442
3.5
Chloride
65073
78505
92023
44418
4	Transient Solute Transport
The transport of each chemical is parametrized by the ground water flow velocity, the
retardation coefficient and the decay constant and the dispersivities. Estimates of the first
three were obtained from the field data as discussed below. The longitudinal dispersivity
was estimated to be equal to one-tenth the distance between the transects, based upon [3].
The transverse dispersivity was ignored because the width of the plume does not increase
down gradient at St. Joseph. Thus, a one-dimensional analysis was applied to the site.
One-dimensional solute transport with first order decay obeys
, ,	Or. d2<: Oc *
2	R— = D—TT-v	Ac.
v '	VI dx2 0.,:
where R is the retardation coefficient, c is the concentration, I, is time, D is the dispersion
coefficient, x is distance, v is the seepage velocity, and A* is the first order decay constant.
First order decay is assumed for this analysis, because it is the usual way to report
degradation rates of chlorinated hydrocarbons, [7]. By formally neglecting transport
phenomena in equation 2, the rate of change of concentration is given by
(3)	R — = -A*<: = -ARc
v ! 
-------
Field-Derived Transformation Rates
7
was given by [10] as
(6)
= cq exp
-/
i , 4A*D
1 +
2 D
Rearranging gives the value of the net apparent rate coefficient, A*, as
(7)
A* = In
l(f)
<0
In
C-O
D
7*
The ground water flow velocities were as determined from a calibrated ground water flow
model developed by [9]. By setting the upstream transect-averaged concentration equal
to c„ and the distance to the downstream transect to x, A* is the rate that causes the
concentration c(x) in equation 6 to equal the downstream transect-averaged concentration.
For small D/x , A* in equation 7 becomes
(8)
A ~ - In
^0
- = - In
c(x)
which has the same form as the solution of the first order decay equation 3. Table 2 shows
a comparison of the retardation coefficient and the ratios of the rate coefficients determined
from equations 4 and 7. Because of the small impact of dispersivity at the distances between
the transects, the two sets of coefficients differ almost exactly by The similarity of the
two sets of rate coefficients, suggests that rate equations could be used to separate the rate
of loss of each component from the net rates determined from the field data. As noted
below, the rates determined from equation 4 or 7 include both production and loss of the
constituents. Rate coefficients, A j are needed that include only the transformation of each
individual constituent j.
Table 2
Ctnaparison of Estimated Degradation Rate Constant*
Chemical
Retardation
Transect 2 to
Transect 4 to
Transect 5 to

Coefficient
Transect 4
Transect 5
Lake Transect


a/ = 2.6 m
ai = 1.6 m
o / = 3.6rr;


x = 260m
x — 158 in
x — 360m


A*/A
A*/A
A*/A
TCE
1.78
1.81
1.84
1.84
c-DCE
1.20
1.22
1.22
1.27
t-DCE
1.37
1.41
1.40
1.39
1,1-DCE
1.40
1.44
1.43
1.31
VC
1.05
1.06
1.07
1.12
6 Extraction of Rate Coefficients via Simulation using Rate Equations
The first order decay rate equation is written for TCE as
d(-'-TCE ;
(9)	—-— = -atceKtckCtce
at
-------
8 Weaver et al.
The similar rate equations for the DCE isomers are
(10)	—jr = fi^TCER-TCECTCE ~ AiRiCi
at
where subscript i represents cDCE, tDCE or 1,1-DCE and /,; is the fraction of DCE that
is in the form of isomer i. The equation for vinyl chloride degradation includes production
from each DCE isomer and is written as
^cDCeIIcDCeC'cDCE
^tDCE RtDCEC tDCE
^1,1 -DCeR\,1~DCeC\,\-DCE
^vcR-vcCvc
An additional equation can be added for the chloride produced from the degradation of the
chlorinated ethenes.
/ \	ciO f */ n	v—^ '	*
(12)	~~dt~ = ^tce^tceCtce + ^2 KRiCi + ^vcJivcCvc
where the subscript i represents each of the DCE isomers. All of the constants in the chloride
equation are positive, thus the chloride concentration should increase with time. Further,
when each mole of TCE is completely degraded three moles of chloride are produced.
The reaction rate model (equations 9 to 12) is in the form of a coupled system of
nonlinear equations for the concentrations over time. These equations were solved with
a variable time step Runge-Kutta solver [2]. For simulation of each pair of transects, the
field-calculated rate coefficients were used as parameters of the model. The concentrations
at the up gradient transect were used as the initial conditions. Figure 4 shows the predicted
percent loss of TCE and each daughter product. Because the data were fitted to the model,
each curve exactly matches the field data at the ends of the transects. The breaks in
the curves, that are indicated in Figure 4 by the arrows, correspond to locations of the
transects. At these points the rates change abruptly, because the net rate coefficients were
calculated from the pairs of transects.
Since the chloride data were not used in fitting any of the degradation rate constants, the
predicted chloride concentrations are an independent check of the model. From transect 2 to
transect 5 there is a decrease in TCE concentration of 7380 //g/L which corresponds to 56.8
micromoles/L. The increase in chloride concentration of 26,950 //g/L (770 micromoles/L)
is 13.6 times greater than the molar loss of TCE. This value contrasts with the factor of
three which would be expected if degradation of TCE was the only source of chloride.
The effective rate constants were determined from the reaction rate equation solution
and are shown in Figure 5. The sharp breaks in the curves shown on this figure occur as
the transects are crossed. For any given pair of transects, the rates vary because of the
DCE isomer fractions appearing in equation 10. The range of values of the apparent loss
coefficients, A j, for each chemical and each pair of transects is shown in Table 3. The values
were converted from the rate equation time basis to that used by the transport equation 2
and thus are compatible with the A* rate constants. There is remarkably low variation in
these constants when compared to the four order-of-magnitude variation in some of the
transect-averaged concentrations.
dC vc _
dt
+
+
(11)
-------
Field-Derived Transformation Rates
9
c
o
o
D
~o
CD
cc
c
o
H—t
(S3
c
0
a
c
o
O
c
CD
y
0
CL
100
40

	
*
^ *

»
0

* k
* 's /
¦ f

/ i
•'//
/1/'
i
:>//
i »



///

,///
:> A


	 TCE
;}/
	gDCE
M *
r
'm 0
.T 0
Jf '
Jr *
3 #
m 0
§ *
V #
. 1 /
0
0
0
, , , . 1
	 tDCE
		 1,1-DCE
	 vc
500 1000 1500 2000 2500 3000
Time in Days
Fig, 1, Percent, lass of each chlorinated ethene as a function of travel time between transects
7 Conclusions
Investigation of the TCE contamination at the St. Joseph, Michigan Superfund site
indicated that TCE is degrading under ambient conditions. This statement is supported
by the field data which shows the presence of degradation products (three isomers of DCE
and vinyl chloride), depletion of oxygen and the abundance of methane. In general, the
degradation products are found in oxygen depleted portions of the aquifer, where methane
is abundant. There are a fraction of sample locations, however, where degradation products
are found in conjunction with high oxygen concentrations.
Degradation rate constants were calculated from the field data set by averaging
-------
10 Weaver et al.
Q 0.08
C
_cd
tn
c
o
Q
0
1	""*
CO
~c
L_
CD
"O
o
-i—>
U)
iZ
0.06
0.04
0.02
cDCE
1,1 DCE
500 1000 1500 2000 2500 3000
Time in Days
Fig. 5. Degradation rate. constants determined from, the reaction rate model.
concentrations over the transects and then applying a onc-dirnensional analysis to the
resulting concentration data. The application of a steady state solute transport model
gives values of the coefficients based on each individual chemical. Because the production
of the DCE isomers depends upon the degradation rate of TCE, and the production of vinyl
chloride depends on the degradation rate of TCE and the DCE isomers, the rate coefficients
estimated from the field data are net rate coefficients that include the production of each
chemical and its subsequent degradation. Thus, the net rate coefficients must be modified
to separate the rates of production and loss. The form of the net rate coefficient differs
from that of simple first order decay only by the inclusion of a dispersion-dependent term
-------
Field-Derived Transformation Rates
11
Tablk 3
Rate. Constants Estimated from the Degradation Rate. Model.
Chemical
Transect 2 to
Transect 4
m = 2.6m
Transect 4 to
Transect 5
ol = 1.6in
Transect 5 to
Lake Transect
= 3.6 m
A(l/ci)
X(l/d.)
A(l/rf)
TCE
c-DCE
t-DCE
1,1-DCE
vc
0.00160
0.00218 - 0.00201
0.00305 - 0.00288
0.00276 - 0.00259
0.0147 - 0.00474
0.00611
0.00518 - 0.00334
0.00556 - 0.00372
0.00598 - 0.00414
0.0135 - 0.00921
0.00557
0.0107- 0.0121
0.00324 - 0.00458
0.00221 - 0.00414
0.0319 - 0.0845
and a constant. For the St, Joseph parameter values, the rate coefficients estimated by
the advective-dispersive equation were almost the same as those determined from the first
order decay equation. This fact suggests that a coupled reaction rate model that includes
production and decay could be used to estimate the apparent loss coefficients for each
chemical.
Analysis of contamination at St. Joseph and similar sites is limited by lack of knowledge
of the mass of contamination in the aquifer, and inability to predict qualitatively if
degradation of TCE can occur. As a result, the rate coefficients extracted from the field
data are not intended to be predictive of the phenomena, but rather representative of the
behavior at this field site.
8 Disclaimer
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency. It has been subjected to Agency review and approved
for publication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
References
[1]	Engineering Science Inc.. Remedial Investigation and Feasibility Study, St. Joseph, Michigan,
Phase I Technical Memorandum., Liverpool. NY, (1990),
[2]	E. Fehlbcrg, Low-Order Classical. Runge-Kntta Formulas with Step Size Control and Their
Application to Some. Heat Transfer Problems, NASA, Washington, DC, TR R-315 (1969).
[3]	L. W, Gclhar. C. Welty and K. R. Rehfeldt, A critical review of data on fie Id-scale, dispersion
in aquifers. Water Resources Research, 28 (1992) pp. 1955-1974.
[4]	F. K. Kitanidis, L. Seinprini, D. II. Kampbell and J. T. Wilson, Natural anaerobic hioreme-
diation of TCE at. the St. Joseph, Michigan, superfund site, Symposium on Bioremediation
of Hazardous Wastes: Research. Development, and Field Evaluations, US EPA, EPA/600/R-
93/054 (1993) pp. 57-60.
[5]	P. L. McCarty and L. Seniprini, Ground-water treatment for chlorinated solvents, Handbook
of Bioremediation, Norris et al. (1994) pp. 5-1 to 5-30.
[6]	P, L, McCarty and .J. T, Wilson, Natural anaerobic, treatment of a TCE plume St.. Joseph,
Michigan NPL site,Bioremediation of Hazardous Wastes, US EPA, EPA/600/R-92/126 (1992)
pp. 47-50.
[7]	H, S. Rifai, R. C. Borden, .1. T. Wilson and C. II. Ward, Intrinsic Bioattenuation for Subsurface
Restoration, in Intrinsic Bioremedation, R. E. Hinchee, ,1. T. Wilson and D. C. Downey eds.,
Battelle Press, Columbus Ohio, 3(1). 1995, pp. 1-29.
-------
12
Weaver et al.
[8]	L. Semprini, P. K. Kitanidis, D. Kampbell and .J. T. Wilson, Anaerobic, transformation of
chlorinated aliphatic hydrocarbons in a sand, aquifer based on spatial, chemical distributions.
Water Resources Research 31 (1994) pp. 1051-1062.
[9]	C. Tiedeman and S. Gorelick, Analysis of uncertainty in optimal groundwater contaminant,
capture design. Water Resources Research. 29 (1993) pp. 2139-2153.
[10]	M. Tli. van Genucliten and W. .J. Alvcs, Analytical. Solutions of the One-Dimensional
Connective-Dispersive Solute Transport, Equation, U.S. Department of Agriculture. Technical
Bulletin No. 1661 (1982) 151 pp.
[11]	.]. W. Weaver, J. T. Wilson, and D. H. Kampbell, Natural Bioat.tenuat.ion of Trieh.loroet.hene
al the. St. Joseph. Michigan Superfund. Site. US. EPA, EPA/600/SV-95/001 (1995).
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing
1. REPORT NO. 2.
EPA/600/A-95/142
3. RE
4. TITLE AND SUBTITLE
Field-Derived Transformation Rates for Modeling
Natural Bioattenuation of Trichloroethene and its
Degradation Products
5. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7.AUTHORisi James W. Weaver, John T. Wilson,
Don H. Kampbell, and Mary E. Randolph
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME ANO AODRESS
USEPA
Subsurface Protection & Remediation Div-Ada
National Risk Management Research Laboratory
P.O. Box 1198
AH a 7AR9D
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
In-House RPDK4
In-House RSJW5
12. SPONSORING AGENCY NAME ANO ADDRESS
USEPA
Subsurface Protection and Remediation Div-Ada
National Risk Management Research Laboratory
P.O.Box 1198, Ada, OK 74820
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/15
15. supplementary notes publ. in the Proceedings of the Next Generation Computational Models
Computational Methods Conf., 8/7 - 9/95
16. ABSTRACT
Subsurface contamination by trichloroethene (TCE) was detected at the St.
Joseph, Michigan site in 1982. The contamination resulted from disposal of
TCE and other chemicals at an industrial facility located near the eastern
shore of Lake Michigan. Investigation of the site revealed the presence of
TCE degradation products, including three dichloroethene isomers, vinyl
chloride and ethene. The plume was found to be depleted of oxygen and
methanogenic at certain depths. Portions of the plume were sampled by slotted
auger borings in 1991 and 1992. In August of 1994, water samples were taken
from a barge situated about 100 m off-shore in Lake Michigan. Each round of
samples were taken along transects that crossed the width of the plume; as
determined in the field by gas chromatography. From the data set, the average
concentration of each chemical and net apparent loss coefficients between
appropriate pairs of transects were calculated. The loss rates were
calculated from the solution of the one-dimensional
advective-dispersive-reactive transport equation. The net apparent rate
coefficients and a set of coupled reaction rate equations were used to extract
the apparent loss coefficients from the field data.
17 KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIF IERS/OPEN ENDED TERMS
c. COSATi Field, Croup
Groundwater
Contamination
Bioremediation
Chlorinated Solvents


18. DISTRIBUTION STATEMENT
RELEASE TO THE PUBLIC
19. SECURITY CLASS iThts Report)
Unclassified
21. NO. OP PAGES
15
20. SECURITY CLASS /This po%c •
Unclassified
22. PRICE
EPA Form 2220-1 (R«». 4-77) devious coition is obsolete
-------