EPA/600/A-96/076
Extraction of Degradation Rate Constants from the
St. Joseph, Michigan Trichloroethene Site
James W. Weaver, John T. Wilson and Don 11. Kampbell
National Risk Management Research Laboratory
United States Environmental Protection Agency
Ada. Oklahoma 74820
Extended Abstract to Appear in
Symposium on Natural Attenuation of Chlorinated Organics
in Ground Water
September 11-13. 1996
Dallas, Texas
1 Background
Anaerobic biodegradation of TCE occurs through successive dechlorination from trichloroethene
lo dichloroet liene. vinyl chloride and ethene [2]. The process produces three isomers of DCE (1,1-
I)(,'E. cis-l,2-I)CE. and trans- 1.2-DCE). Although TCE was commonly used in industry, the DCEs
were not: and ethene would not be expected in most ground waters. Thus the presence of these
compounds are indicative of degradation when found in anaerobic ground waters. Implicit in the
work of [Ij and [3] 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 as
described in this proceedings [9j. The purpose of this paper is to present estimates of averaged
concentrations, mass flux and degradation rate constants.
2 Ground Water Flow
Ground water flows at the St. Joseph site from the contaminant source toward Lake Michigan. The
average hydraulic conductivity at. the site was estimated at 7.5 m/d from a calibrated ground water
flow model [0], The estimated travel time for TCE between the source and the lake is approximately
18 years (Table 1). If the contamination was released only in the aqueous phase, one would expect
that contaminants released 18 years or longer ago would by now have discharged into the lake. The
observed contaminant distribution suggests a continuing source, most likely a DNAPL.
3 Averaged Concentrations
Data were collected from the site from sets of borings that formed four on-shore and one off-
shore transects thai, crossed the plume (.see Figure 1 of [9]). These range from 130 in to >s."t m
from the suspected source of contamination. From the borings, a. three-dimensional view of the
contamination was developed. A field gas chromatograpli was used to determine the boundaries of
1
-------
Average Concentration (j/g/L)
Highc.Nl Concentration (fig/L)
Distance from source
Transport Time
Transect Width
Vinyl
(m)
(y)
(m)
TCE
cis DOE
Chloride
130
3.2
108
6.500
8.100
930
68,000
128,000
4,J,00
390
9.7
150
520
830
'150
8.700
9,800
1,660
550
12.5
192
15
18
106
56
870
205
855
17.9
395
< 1
< 1
< 1
/ J,
0.8
0.5
Table 1: Attenuation of the chlorinated etheoes along the length of the plume
the plume. Sampling continued until the entire width of the plume was crossed at each transect.
My following this procedure the transects are known to have contained the entire plume. This
approach allows calculation of total mass that crosses each transect, and thus gives an estimate of
flux of each contaminant as a function of distance from the lake.
Transect-averaged concentration estimates were developed by using the SITE-3I) graphics pack-
age. [7]. The data, were represented as sets of blocks that are centered around each boring. The
blocks were each 5 ft high, corresponding to the length of ( lie slotted auger. At each transect, the
average concentration was calculated by summing over the blocks and dividing by the area of the
transects.
In Table 2. concentration estimates are presented for the perpendicular transects ordered from
furthest up gradient (transect, 2) to furthest down gradient (transect 5). The concentration esti-
mates are based only upon blocks from the anaerobic portion of the aquifer (and thus differ from
the averages in Table 1). All of the chlorinated ethenes siiow decreasing concentration with dis-
tance 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. However, on a molar basis the increase in average chloride concentration is greater
than which would result from dechlorination alone.
4 Mass Flux
The concentration results (Table 2) show that by the time the contaminants reach the lake their
concentrations are reduced to very low levels. It is equally important, to determine the mass of
chemicals released to the la,ke per year. Given the approximate ground water velocities and the
contaminant concentrations in the transects, an estimate of the mass flux of chemicals can also be
estimated. Advective mass fluxes of each chemical were estimated per transect by multiplying the
seepage velocity by concent ration in each block formed by STTE-3I). The results are given in Table 3
which shows a decline in mass flux of each chlorinated ethene. The reduction in flux ranged from
2
-------
Chemical
Traaisect. 2
Transect 4
Transect 5
Lake Transc
TCK
74 11
864
30.1
(1.4)
c-I)CE
9117
1453
281
(0.80)
t-DCE
716
31.4
5.39
(1.1)
1.,1-DC I'-
339
21.3
2.99
blq
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
Table 2: Transect-averaged concentrations (//g/L) from the anaerobic zone. Values in parenthesis
were based upon one or more estimated values and blq indicates no detection above the limit of
quantitat ion
Mass flux kg/y
Total
Transect TCE c-DCE VC Ethene Ethenes Methane Chloride
2 (Any-Sept, 1991)
117
133
16.8
7.60
283
65.7
1456
1 (Any-Sept, 1991)
50.0
45.2
16.8
7.95
125
49.2
54 5
J, (Mar. 1992)
30.9
4 1.7
3.87
10.8
88.4
101
4610
5 (Apr, 1992)
0.95
10.0
1.68
0.164
13.1
16.7
5290
Reduction Ratio
123
13
10
46
22
Table 3: Mux Estimates for Transects 1. 2, I and 5. The reduction ratio is the ratio of mass (lux
at transect 2 to that at transect 5,
a factor of 10 to 123. The flux of methane showed no consistent pattern. Chloride flux increased
beyond Transect 1.
5 Degradation Rates
The transport of each chemical is parametrized by the ground water flow velocity, the retardation
coefficient, the dispersivities. and the decay constant. Specifically, two-dimensional solute transport
with first order decay obeys
oi ~ xxd? + Dyydf v~th- " ^ ( J
s ^ s—v—' Decay
DiaprTsion Adurctton
where 11 is the retardation coefficient, c is the concentration, t is time, f)rj. and Dyy are the
longitudinal and transverse dispersion coefficients, respectively, ar is longitudinal distance, y is the
distance transverse to the plume centerline in the horizontal plane, v is the seepage velocity, and
3
-------
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 [4]. This form of t.he transport
equation assumes that the ground water flow is uniform and aligned with the axis of the plume as
observed for the plume. This assumption also allows application of analytic solutions as described
in the appendix.
The concentration of dissolved chemicals can change because of the effect of the terms on the
right hand side of equation 1. Dispersion is used to characterize apparent, physical dilution in
aquifers. Dispersion is currently understood to result primarily from ground water (low through
heterogeneous materials. In multi-dimensional flow, advection can cause concentrations to de-
crease because of divergence of flow lines. Advection does not directly change concentrations in
one-dimensional flow, but influences the contribution of dispersion. Decay changes concentration
through removal of mass from the aquifer.
The significance of these observations is that when presented with a set of contaminant con-
centrations. the distribution of contamination may depend upon physiochemical and biological
processes. Observed concentrations in themselves do not indicate the contribution of each process
to the plume shape. Extraction of apparent rates from the field data needs to account for the
multiple processes. In Table 5 estimated rate constants are given for St. Joseph. These constants
were determined from the solution of the transport equation presented in the appendix. The so-
lution included advection. retardation, longitudinal and transverse dispersion, and first order loss.
Inclusion of transverse dispersion is important because this characterizes down-gradient spreading
of the plume. The observed widths of the plume at St. Joseph are given in Table 1 and were used
to estimate the transverse dispersivity according to the proceedure given in the appendix. The
effect of transverse dispersivity on the estimated rate constants, however, decreases as the plume
widens and the centerline concentrations decrease. Longitudinal dispersivity lias been shown to
have a minor impact on the estimated rate constants at distances between transects on the order
of 100 meters [S],
The rates given in Table 4 are called net rates, because, for the daughter products, the observed
concentrations are a result of production of the daughter from decay of the parent ami decay of
the daughter itself. 'J lie gross rate of decay of the daughter (Table 5) that does not include its
production was determined by the procedure given in the appendix. The two rates are the same
for TCE. since no production of TCE occurred. The gross rates are. as expected, higher than the
net rates, because production of a compound must: be balanced by high gross rates to attain the
observed net rate.
Chemical Transect 2 to I Transect 4 to 5 Transect 5 to Lake
TOE 0.30 1.7 1.7
cDCE 0.26 0.58 3.3
YO 0.15 0.78 2.6
Table 4: Net apparen* degradation rate constants { I/y) from the two-dimensional model (equa-
tion :j)
4
-------
Chemical Transect 2 to-I Transect 4 to 5 Transect 5 to Lake
TCK
cDCE
VC
0.30
0.54
•2.6
1.7
1.1
3.1
1.7
1.0
20
Table 5: Apparent degradation rate constants (1/y) from the two-dimensional model equation 3
and the gross rate correction given by equation 7
6 Conclusions
The western TCE plume at St. Joseph. Michigan showed a decrease of maximum TCE concentration
by a factor of 50,000 from the furthest up-gradient transect to the lake transect. Concentrations
of each contaminant declined to values below the respective MCLs when sampled from the lake
sediments. Mass fluxes decrease by factors of 10 to 123 from the source to the last on-shore tran-
sect (number 5). Thus, not only do the concentrations decline, so does the loading in the ground
water. The reduction in loading is attributed to degradation, because of the geocliemical evidence
presented by [9], Further, when site-specific estimates of the transport, parameters are used in
solutions of the transport equations the apparent reduction in concentration is only accounted for
by loss of mass. These apparent degradation rate constants were calculated from the St. Joseph.
Michigan data set through application of a two-dimensional analytical solution of the transport
equation. Since transverse spreading of the plume reduces the contaminant concentrations, the
effect of transverse dispersivity was included in the analysis.
7 Disclaimer
This is an abstract of a proposed presentation and does not necessarily reflect EPA policy.
[1] P. K. Kitanidis, L. Semprini. 1). H. Kampbell. and J. T. Wilson. Natural anaerobic bioremedi-
ation of TCE at the St. Joseph, Michigan, superfund site. In Symposium on Biorcmediation of
.Hazardous Wastes: Research. Development, and Field Evaluations, EPA/fiOO/R-93/054. pages
57-00. United States Environmental Protection Agency. 1993.
[2] P. L. McCarty and L. Semprini. Ground-water treatment for chlorinated solvents. In Norris
et al.. editors. Handbook of Hioremediation. pages 87-11(>. Lewis Publishers. 1994.
[3] P. L. McCarty and J. 1'. Wilson. Natural anaerobic treatment of a TCI" plume St. Joseph,
Michigan NPL site. In Biorcnu dial ion of Hazardous U-'V/sfe*. EPA/600/11-92/126, pages -17-50,
References
1992.
;>
-------
[ t] II. S. Rifai, It. C. Borden, J. T. Wilson, and ('. II. Ward. Intrinsic bioattenuation for subsurface
restoration. In II. E. Hhiehee. J. T. Wilson, and D. C. Downey, editors. Intrinsic Bioremedation.
volume .*5(1). pages 1-29, Columbus Ohio. 1995. Hat,telle Press.
[5] V. J, Smith and Randall J. Charbeneau, Probabilistic soil contamination exposure assess-
ment procedures. American Society of Civil Engineers, Journal of Kncironmental Engineering,
116(6):1143-1163, 1990.
[0] 0. Tiedeman and S. Gorelick. Analysis of uncertainty in optimal groundwater contaminant
capture design. Water Resources Research. "29:2139 2153. 1993.
[7] J. W. Weaver, Animated three-dimensional display of field data with SITE-3D: User's guide for
version 1.00. Technical Report EPA/G00/R-96/00-1. United States Environmental Protection
Agency, 1996.
[8] .1. W. Weaver, J. T. Wilson, I).II. Kampbell. and M.E. Randolph. Field-derived transformation
rates for modeling natural bioattenuation of trichloroethene and its degradation products. In
(leiieration of Computational Models Computational Methods, August 17-19, Bay Citij,
Michigan. Society of Industrial and Applied Mat hematics. 1995.
[9] .lames W. Weaver. John T. Wilson, and Don H. Kampbell. Case study of natural attenuation
of tricldoroethene at St. Joseph, Michigan. In Natural Attcnaulion of Chlorinated Solvents in
the Subsurface. United States Environmental Protection Agency. 1996.
A Appendix: Extraction of Rate Constants via Two-Dimensioiial,
Steady-State Transport Analysis
The two-dimensional transport equation, subject, to the boundary conditions
Vertically averaged concentrations and the distances between each borehole were used to develop the
boundary condition (c(0, ij.t)\w equation 2) for application of equation 3. The unknown parameters
c(x.y,0) = 0
(2)
c(oc,yJ,) = e(x.-oc,t) = c(x.oo.t) = 0
has the approximate steady state solution [5]
6
-------
arc the iip-gradient peak concentration, c.„ and the standard deviation, rr. of the distribution. Since
the width of the plume, IT", was established via the field sampling program, the stand a,rd deviation
of the distribution can be estimated as 11" = Oct. A mass balance can then be solved for the peak
concentration of the gaussian distribution. c0. from
nc.dy = tic,, exp dy = nc.,a\/2TT ('1)
when; n is the porosity, c is the vertically averaged concentration, and the y coordinate runs parallel
to the transect.
The transverse dispersivity can also be estimated from the measured widths of the transects.
The width of a conta.minant distribution is related to the transverse dispersivity through
1 da2
(5)
where ayy is the transverse dispersivity. By applying equation 5 in a discrete form and substituting
At = Axll/v, an expression for uyy is obtained in terms of the seepage velocity, retardation
coefficient, distance between transects (A.t), and change in variance of the gaussian distributions
for the transect concentrations (Aer2);
1 Aa2
(<>)
2R Ax
The only remaining unknown in equation is the decay constant A*, which is determined through
a bisection search. Table 4 gives the rale constants from the two dimensional model.
A.l Net and Gioss Decay Rates
The rate constants derived from the solution (equation 3 and Table 4) are net rates which include
the production and decay of a given daughter product, ft. is necessary to separate production
of the compound from its decay to estimate the gross apparent decay rates for c-DCE. t-DCE.
IJ-DCE and VC. Previous work [8] used a reaction rate model that solved simultaneous ordinary
differential equations for this purpose. Here, simplified expressions for the rates were used to
estimate the apparent decay rates.
•\)(r0 — /Mj+i(«)5 + ^j(u)
where Aj(„j is the net decay rate determined by equation 3, fj is the fraction of an isomer (j)
produced from the degradation of the parent (j+1), A;-+1 („) is the apparent decay rate of the parent
defined from equation .'5. 5 is the ratio of molar concentration of parent (j+1) to daughter j, and
A;(7) is the gross apparent decay rate of daughter j. For the DCE isomers, f} is approximated by
the average ratio of an isomer j to the sum of the DCEs over the pairs of transects. For Y(\ f}
is equal to 1.0. The gross apparent decay rates for c-DCE. t~I)(*E. I.l-DCE and VC appear in
Table 5. Although equation 7 is concentration dependent because of H which was assumed to be
the average of the up and down gradient ratios, the results presented in Table 5 are essentially the
same as determined from the reaction rate model [8].
-------
TECHNICAL REPORT DATA
1 REWa7600/A-96/076
2.
3 . RE<
4. TITLE AND SUBTITLE
5. REPORT DATE
Extraction of Degradation Rate
Michigan Trichloroethene Site
constants from the St. Joseph,
6. PERFORMING ORGANIZATION CODE
7. AUTHORSS)
James W. Weaver, John T. Wilson, and Don H. Kampbell
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO,
USEPA/ORD
National Risk Management Research Laboratory
Subsurface Protection and Remediation Division
P. 0. Box 1198
Ada, OK 74820
11. CONTRACT/GRANT NO.
In-House RPDK4
In-House RSJW5
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
USEPA/ORD
National Risk Management Research Laboratory
P.O. Box 1198
Ada, OK 74820
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES Will be
published in proceedings and EPA Report
is. abstract Anaerobic biodegradation of TCE occurs through successive dechlorination from
Trichloroethene to dechloroethene, vinyl chloride and ethene [2]. The process produces
three isomers of DCE (1,1-DCE, cis-1, 2-DCE and trans-1»2-DCE). Although TCE was
commonly used in industry, the DCEs were not; and ethene would not be expected in most
ground waters. Thus the presence of these compounds are indicative of degradation when
found in anaerobic ground waters. Implicit in the work of [1] and [3] 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 as
described in this proceedings [9].
17 .
KEY WORDS AND DOCUMENT ANALYSIS
A. DESCRIPTORS
B. IDENTIFIERS/OPEN ENDED TERMS
C. COSATI FIELD, GROUP
Contamination
TCE
Degradation Rates
Subsurface
18. DISTRIBUTION STATEMENT
Release to the Public
19, SECURITY CLASS(THIS REPORT)
Unclassified
21. NO. OF PAGES
7
20. SECURITY CLASS(THIS PAGE)
Unclassified
22. PRICE
SPA FORM 2220-1 (REV.4-77) PREVIOUS EDITION IS OBSOLETE
------- |