. .51.. .
L .313}

71. vx

.33]? 126.4.ch 3...,

 

.. .2.me $.74.

 

(3 " LIBRARY
51 (c 3: xi/ 519 Michigan State
v ' University

This is to certify that the
dissertation entitled

DEVELOPMENT AND APPLICATION OF A METHODOLOGY
TO EVALUATE NATURAL ATI'ENUATION OF
CHLORINATED SOLVENTS USING CONCEPTUAL AND
NUMERICAL MODELS

presented by
Jaime A. Graulau-Santiago

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Civil and Environmental
E:ngineen'nj

fl C fl/Wi wif/
Major Pidjlsor’ 5 Signature

December 8, 2003

 

 

Date

MSU is an Amnnative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 chIRCJDateDuepss-pjs

 

 

DEVELOPMENT AND APPLICATION OF A METHODOLOGY TO
EVALUATE NATURAL ATTENUATION OF CHLORINATED SOLVENTS
USING CONCEPTUAL AND NUMERICAL MODELS
By

Jaime A. Graulau-Santiago

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY
Civil and Environmental Engineering Department

2003

ABSTRACT
DEVELOPMENT AND APPLICATION OF A METHODOLOGY TO
EVALUATE NATURAL ATTENUATION FOR CHLORINATED SOLVENTS
USING CONCEPTUAL & NUMERICAL MODELS
By

Jaime A. Graulau-Santiago

Natural attenuation of chlorinated solvents has being increasingly used by
engineers and regulators as a remedial alternative to contaminated aquifers because the
numerous advantages it has over more traditional engineered solutions. However, natural
attenuation has been clearly demonstrated only for a selected class of organic
compounds, primarily fuel hydrocarbons. The scientific community has recognized that
the surge in use of natural attenuation has outpaced certain guidelines that have been
developed for its application. This study evaluates a novel approach to evaluate natural
attenuation of chlorinated solvents by using models incorporating important physical and
biogeochemical processes that can be critical for a successful evaluation of natural
attenuation.

A hydraulic characterization methodology was developed to estimate parameters
that influence the transport of solutes in groundwater environments. Biogeochemical
data coupled with historical development of the VOC plume were used to develop a
conceptual model for the contaminated site. This model was translated into a numerical
model incorporating the compiled information and two snapshots of the plume historical

development were successfully validated.

The successful application of the methodology developed in this dissertation
could become the foundation for the development of a series of guidelines to evaluate

natural attenuation of chlorinated solvents.

Copyright by
JAIME A. GRAULAU-SANTIAGO
2003

 

ACKNOWLEDGMENTS

I would like to express my gratitude to all of those that directly or indirectly
helped me in the completion of this work.

First I would like to thank God for giving me the strength, health, patience, and
the character to always work hard for the things I want.

To my parents, for teaching me essential values in life that I could not be taught in
any school or college, thanks for accepting me in our Home University.

My sisters and brother, thanks for the long phone conversations during the last
five years and for all the experiences we share together. You guys are my best friends.

To Parissa, for the love we share and for being with me during these tough times.
Finally I can say “I’m done”. I know you have heard this before but this time I think is
for real. I am looking forward to our bright future together.

To Natalia Zohreh, for telling me to hurry up with the dissertation, even when She
could not even talk.

To my advisor; Dr. David C. Wiggert, thanks for giving me the opportunity to
reach all my academic goals. Your support during the last five years was invaluable.

Thank you, Dr. Michel J. Dybas, for showing me all the aspects of purging and
sampling monitoring wells in the field. You are an inspiration for any one who wants to
be a good scientist.

To Dr. David W. Hyndman; for introducing me to the wonderful world of

hydrogeology and numerical modeling.

To Dr. Manta S. Phanikumar, for teaching me numerous tools that I can further
apply in my career as a modeler and scientist.
Finally, to all the friends and co-workers at MSU for making my years in East

Lansing more enjoyable.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... ix

LIST OF FIGURES ........................................................................................................... x

LIST OF SYMBOLS ..................................................................................................... xiv

CHAPTER 1

INTRODUCTION ............................................................................................................. 1
1.1 Environmental concern of chlorinated solvents .................................................. 1
1.2 Natural attenuation: overview of processes ........................................................ 4
1.3 Natural attenuation of chlorinated solvents ...................................................... 11
1.4 Hypothesis and research objective .................................................................... 13
1.5 Scope of work .................................................................................................... 13
1.6 Literature cited ................................................................................................... 14

CHAPTER 2

HYDROGEOLOGIC CHARACTERIZATION OF A VOC

CONTAMINATED AQUIFER ZONE ........................................................................... 17
2.1 Abstract ............................................................................................................. 17
2.2 Introduction ....................................................................................................... 18
2.3 Site description .................................................................................................. 19
2.4 Methods ............................................................................................................. 21
2.5 Results and discussion ...................................................................................... 32
2.6 Conclusions ....................................................................................................... 46
2.7 Acknowledgments ............................................................................................. 48
2.8 Literature cited .................................................................................................. 48

CHAPTER 3

EVALUATION OF NATURAL ATTENAUTION
IN A CONTROL VOLUME OF A VOC

CONTAMINATED AQUIFER ....................................................................................... 51
3.1 Abstract ............................................................................................................. 51
3.2 Introduction ....................................................................................................... 52
3.3 Site Description ................................................................................................. 55
3.4 Materials and methods ...................................................................................... 65
3.5 Results and discussion ...................................................................................... 67
3.6 Conclusions ....................................................................................................... 79
3.7 Acknowledgments ............................................................................................. 80
3.8 Literature cited .................................................................................................. 81

vii

CHAPTER 4

DEVELOPMENT AND APPLICATION OF A NATURAL
ATTENUATION MODEL FOR A VOC CONTAMINATED
AQUIFER UNDER LIMITED ELECTRON DONOR

CONDITIONS ................................................................................................................. 84
4.1 Abstract ............................................................................................................. 84
4.2 Introduction ....................................................................................................... 85
4.3 Model development and methodology .............................................................. 88
4.4 Results and discussions ................................................................................... 102
4.5 Summary and conclusions .............................................................................. 118
4.6 Acknowledgments ........................................................................................... 119
4.7 Literature cited ................................................................................................ 120
CHAPTER 5
SUMMARY AND CONCLUSIONS ............................................................................ 124
5.1 Introduction ..................................................................................................... 124
5.2 Hydrogeological characterization of the VOC
contaminated area ........................................................................................... 125
5.3 Geochemical analyses: insight into the natural attenuation
process ............................................................................................................. 128
5.4 Reductive dechlorination model for the VOC
contaminants plume ........................................................................................ 131
5.5 Summary ......................................................................................................... 131
5.6 Conclusions ..................................................................................................... 132
5.7 Literature cited ................................................................................................ 133

APPENDIX A - Boring logs for the wells installed for the pilot
scale study ........................................................................................... 136

APPENDIX B - MatLab script for the optimization of aquifer
parameters ........................................................................................... 147

APPENDIX C - RT3D user defined reaction code for the
evaluation of reductive dechlorination linked to
carbon source degradation .................................................................. 149

viii

Table 1.1.

Table 1.2.

Table 2.1.

Table 2.2.

Table 2.3.

Table 2.4.

Table 3.1.

Table 3.2.

Table 3.3.

Table 3.4.

Table 3.5.

Table 4.1.

Table 4.2.

Table 4.3.

Table 4.4.

LIST OF TABLES

Estimated annual production of the most important

hydrocarbons in the US. and its applications .................................................. 3
Gibbs free energy for reductive dechlorination of

chlorinated ethene compounds ....................................................................... l 1
Details of tracer experiments ......................................................................... 24
Logistics of tracer tests .................................................................................. 25
Bail test results ............................................................................................... 35
Depth-specific dispersivities and corresponding

velocities ........................................................................................................ 43

Chlorinated compounds and concentrations found in
sediment samples at ARCO facilities ............................................................ 57

Depth-specific summary of physical parameters .................... . ...................... 62

Description of the sampling events in the control
volume for the three year study ..................................................................... 66

Flow through cell parameters ......................................................................... 72

EPA (1998) screening process applied to a selected

aquifer interval (well no. MP-A3 at 20. 4m bgs) ............................................ 79
Gibbs free energy for reductive dechlorination with

toluene as the electron donor ......................................................................... 90
Parameters for the reductive dechlorination model ....................................... 96
Flow model parameters and initial conditions for the

natural attenuation simulation of the VOC contaminated
site ................................................................................................................ 101

Details of the numerical model domain ....................................................... 108

ix

LIST OF FIGURES

Images in this dissertation are presented in color (Figures 2.8, 3.10, 3.11, and 3.12).

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 2.1.

Figure 2.2.

Figure 2.3.

Figure 2.4.

Figure 2.5.

Figure 2.6.

Figure 2.7.

Figure 2.8.
Figure 2.9.

Figure 2.10.

Commonly chlorinated organic compounds found in
groundwater .................................................................................................. 2

Steps in the degradation of PCB by reductive
dechlorination ............................................................................................. l 0

Redox potential (Eh) in millivolts for various electron
acceptors in groundwater ............................................................................ 12

Location of study area ................................................................................. 20

Well network in study area (stimulation and
augmentation grids) .................................................................................... 22

Cross section A-A’ (Figure 2.2) through monitoring
wells MW-2, MW-5, and MW-8 ................................................................. 26

Main features that control regional groundwater flow
in the unconfined aquifer (from Lipinski 2002) ......................................... 27

Numerical model grid in (a) horizontal, and (b)
vertical directions. (No. of cells in the horizontal is
52,890, with 15 layers in the vertical direction) ......................................... 29

Scatter plots of (a) hydraulic conductivity and

(b) porosity .................................................................................................. 34
(a) log K frequency distribution, and (b) experimental

and model variograms of the data ............................................................... 36
Kriged Images of (a) Log K, and (b) total porosity .................................... 37
Layout of the tracer injection system .......................................................... 39

Tracer breakthrough curves at (a) flux control well,
and (b) delivery wells .................................................................................. 40

Figure 2.11.

Figure 2.12.

Figure 2.13.

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 3.4.

Figure 3.5.

Figure 3.6.

Figure 3.7.

Figure 3.8.

Simulated and observed tracer breakthrough curves in
downgradient monitoring points: (°) observed, (—)

optimum per-layer case, (— —) Single optimum
value case .................................................................................................... 42

Tracer concentration (C/Co) contours at (a) 4 hrs,
(b) 1 day, and (c) 15 days after tracer injection .......................................... 45

Root Mean Square Errors (RMSE) in velocity at each

depth interval for a 10% change in porosity, and 10%

change in dispersivity (deviation from the optimal

values) ......................................................................................................... 46

Schoolcraft Village showing the extent of the VOC
contaminant plume ....................................................................................... 56

Location of the major source areas of contaminants
identified during the 1986 investigation (modified
form Lipinsky, 2002) ................................................................................... 58

Regional hydrologic boundaries that control
groundwater flow in the Schoolcraft area (modified
from Lipinski, 2002) ................................................................................... 60

Control area with monitoring wells for the natural
attenuation study ......................................................................................... 61

Details of the multi-level wells installed for the
natural attenuation study ............................................................................. 63

Cross section A-A’ (Figure 3.4) showing the location
of two preferential flow pathways ............................................................... 64

Chlorinated ethene concentration in aquifer sediments

(Fall 2000 sampling event); TCE(o), cis-DCE(o), and

VC(o). Vertical axis represents the depth in meters

below the ground surface at which the sample was

collected. Horizontal axis is the concentration in ug/kg ............................ 68

Chlorinated ethene compounds in groundwater

samples (Spring 2001 sampling event); TCE(o), cis-

DCE(o), and VC(o). Vertical axis represents the

depth in meters below the ground surface at which the

sample was collected. Horizontal axis is the

concentration in pg/L .................................................................................. 69

xi

Figure 3.9.

Figure 3.10.

Figure 3.11.

Figure 3.12.

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.

Figure 4.6.

Figure 4.7.

Figure 4.8

Geochemical constituents concentration in

groundwater samples (Spring 2001 sampling event);

NO;'(-), 504210), and or (0). Vertical axis

represents the depth in meters below the ground

surface at which the sample was collected.

Horizontal axis is the concentration in mg/L .............................................. 71

Solid phase chlorinated ethene concentration (pg/kg)
in B-B’ cross section (spring 2002 sampling event).
(a) TCE, (b) ciS-DCE, and (c) VC .............................................................. 74

Liquid phase chlorinated ethene concentration (pg/L)
in B-B’ cross section (spring 2002 sampling event).
(a) TCE, (b) cis-DCE, and (c) VC .............................................................. 75

Geochemical parameter concentrations (mg/L) in
B-B’ cross section (spring 2002 sampling event).
(a) NO;’, (b) so}; (c) cr ........................................................................... 77

General pathway for the reductive dechlorination

process of PCB to ethene. Bold arrows indicate the

most likely pathway under the influence of microbial

processes (Garant & Lynd, I998) ............................................................... 89

Site location and model boundaries for reactive
transport simulation .................................................................................. 101

Toluene and biomass concentration in the
hypothetical batch reactor simulation ....................................................... 102

Liquid and solid phase PCE concentration for the
hypothetical batch reactor simulation ....................................................... 104

Liquid and solid phase TCE concentration in the
batch reactor .............................................................................................. 104

Liquid and solid phase DCE concentration in the
batch reactor .............................................................................................. 106

Liquid and solid phase VC concentration in the batch
reactor ....................................................................................................... 106

Liquid and solid phase ethene concentration in the
batch reactor .............................................................................................. 107

xii

Figure 4.9. Groundwater contour map in the VOC contaminated
region ........................................................................................................ 109

Figure 4.10. Comparison between observed and computed heads
for several wells in the region ................................................................... 110

Figure 4.11. Simulated (solid) and delineated (dashed) 5 ug/L
isoconcentration line for (a) PCE, (b) TCE, (c)cis-
DCE, and (d) VC at the end of the loading period (35
years) ......................................................................................................... 112

Figure 4.12. Simulated (solid) and delineated (dashed) Sug/L
isoconcentration line for (a) PCE, (b) TCE, (c)cis-
DCE, and (d) VC at 47 years .................................................................... 115

Figure 4.13. Simulated Sug/L isoconcentration line for (a)PCE, (b)
TCE, (c) cis-DCE, and ((1) VC for the year 2050 ..................................... 117

Figure 5.1. Cross section of the Plume G site with distinctive
stratigraphic tracer breakthrough curve .................................................... 126

Figure 5.2. Tracer breakthrough curves for the simulated
scenarios. The solid line represents the depth specific
case, the dashed line represents the entire model grid
case, and the solid circles are the observed data. ...................................... 128

Figure 5.3. PCB to ethene breakdown through reductive
dechlorination (Clement et al. 2000) ........................................................ 129

Figure 5.4. Illustration of three different scenarios that can be
found in co-contaminated environments. (a) Non-
interacting petroleum and chlorinated solvents
plumes, (b) partly interacting plumes, and (c)
completely interacting plumes (NRC 2000) ............................................. 130

xiii

csim

D,‘ Dij
D
[DCE ]

[DCElsoiI
[ethene ]
[ethene]

soil
§[DCI~:]
S[ethene]
S[tol]
S[PCE]
€[TCE]
€[vc]
foc

H

h

K

Kd

K d [DCE]
K d [ethene]
K d [PCB]
K d [TCE]
K d [tol]
Kd [vc]

LIST OF SYMBOLS

dispersivity tensor, [L]

cell endogenous decay coefficient, [1"]
contaminant concentration in soil air, [M L'3]
dissolved concentration of the kth specie, [M L'3]

concentration of the kth sorbed specie, [M M'I]

source-sink flux term concentration for the kth specie, [M L'3]

contaminant concentration in water, [M L'3]
tracer concentration in pore fluid, [M L'3]
tracer concentration in source or sink, [M L'3]

observed tracer concentration, [M L'3]

simulated tracer concentration, [M L'3]
hydrodynamic dispersion coefficient tensor, [L2 TI]

effective molecular diffusion coefficient, [L2 T1]
liquid phase DCE concentration, [M L'3]

sorbed DCE concentration [M M"]

liquid phase ethene concentration, [M L'3]

sorbed ethene concentration [M M"]

first-order DCE mass transfer rate coefficient [1"]
first-order ethene mass transfer rate coefficient [Tl]
first-order mass solid-liquid mass transfer coefficient, [1"]
first-order PCE mass transfer rate coefficient [1"]
first-order TCE mass transfer rate coefficient ["1"]
first-order VC mass transfer rate coefficient [Tl]

fraction of total organic carbon, [M M"]

dimensionless Henry’s constant
hydraulic head [L]
hydraulic conductivity tensor [L T‘]

solid-liquid partition coefficient, [L3 M"]
solid-liquid DCE partitioning coefficient, [L'3 M"]

solid-liquid ethene partitioning coefficient, [L'3 M"]
solid-liquid PCE partitioning coefficient, [L'3 M"]
solid-liquid TCE partitioning coefficient, [L'3 M"]
toluene partitioning coefficient, [L3 M"]

solid-liquid VC partitioning coefficient, [L'3 M"]

xiv

Ki

Koc

Ks [DCE]
Ks [PCE]
Ks [TCE]
Ks [tol]
Ks [VC]
k

“max
[PCE ]

[ma 1..
[PCE]
<1[DCE]
<1[PCE]

soil

q max
£1[TCE]
5l[vc]
¢

¢e

R

Rk

p, Pb
95

S

t[TCE ]
[TCElsoil
[tol]

[mllsoil

Y

Vl

IVCl
[VClsoil
W9 C’k a ‘15
X,y,Z

hydraulic conductivity tensor principal component, [L Tl]
organic carbon-water partition coefficient, [L3 M'l]

DCE half-velocity coefficient, [M L'3]

PCE half-velocity coefficient, [M L'3]

TCE half-velocity coefficient, [M L'3]

half-velocity coefficient for toluene consumption [M L‘3]
VC half-velocity coefficient, [M L'3]

chemical specie index

maximum specific growth rate of dehalogenators, [”1"]

liquid phase PCE concentration, [M L'3]
equivalent PCE concentration, [M L'3]

sorbed PCE concentration [M M"]

maximum specific utilization rate of DCE, [M M'l 1"]
maximum Specific utilization rate of PCE, [M M'l Tl]
maximum specific toluene utilization rate, [M M'1 Tl]
maximum specific utilization rate of TCE, [M M'I Tl]
maximum specific utilization rate of VC, [M M'l T'l]
total porosity of the soil, [L3 L'3]

effective porosity

retardation coefficient

reaction term for the kth specie

bulk mass density of the sediments, [M L'3]

particle mass density, [M L'3]

specific aquifer yield, [L3 L'3]

time, [T]

liquid phase TCE concentration, [M L'3]

sorbed TCE concentration [M M"]

concentration of toluene, [M L'3]

solid-phase toluene concentration, [M M"]

groundwater velocity vector, [L T']
contaminant retarded velocity, [L T']

longitudinal advective groundwater velocity, [L T']
liquid phase VC concentration, [M L'3]

sorbed VC concentration [M M"]

fluid source Sink term, [L T1]
Cartesian coordinates, [L]
distance along a Cartesian coordinate, [L]

XV

[X] biomass concentration, [M L'3]
Y[DCE] 41-05] stoichiometric TCE to DCE yield coefficient, [M M"]

Y[ethene]/[VC] stoichiometric VC to ethene yield coefficient, [M M"]
Y[TCE]/[pCE] stoichiometric PCE to TCE yield coefficient, [M M"]
Y[VC]/[DCE] stoichiometric DCE to VC yield coefficient, [M M"]
Y[x]/ [ml] yield coefficient for cell synthesis, [M M"]

V gradient operator (ii/ax, a/ay, 6/62)

xvi

CHAPTER 1

INTRODUCTION

1.1 Environmental concern of chlorinated solvents

Due to their widespread use as industrial solvents and degreasers, chlorinated
compounds are among the top five contaminants found in groundwater around the world
(Prakash and Gupta 2000). These compounds are substituted hydrocarbons in which
hydrogen atoms have been replaced with a chlorine atom. Figure 1.1 illustrates some of
the chlorinated compounds of environmental concern. Annual production levels in the
US. for some of these compounds are on the order of 105 to 106 tons per year (Table 1.1).

Halogenated hydrocarbons can be degraded under both, anaerobic and aerobic
conditions. The anaerobic degradation process is called reductive dechlorination. In this
process, halogen atoms are sequentially removed from the compound molecule and
replaced by hydrogen. The chlorinated compound in this microbially mediated reaction
is used as an electron acceptor; not as a carbon source (Holliger and Schumacher 1994;
McCarty 1997; Sims et al. 1991).

Reductive dechlorination depends on the redox state of the halogenated molecule,
which is determined primarily by the strength of the halogen-carbon bond. The higher
the bond strength, the less likely the halogen will be removed. In general, bromine and
iodine substitutions, which have lower bond strengths than chlorine, are easier to remove.

Fluorine, for example, forms stronger bonds with carbon than chlorine,

C1 C1
\ /
C = C
/ \
C1 C1

tetrachloroethene

(PCE)

H\ /H

/ \
C1 C1

cis-1,2 dichloroethene
(cis-DC E)

vinyl chloride
(V C)

fl

1
H— —$—H

z—o—Q

H

I, 2 dichloroethane
(1,2 DCA)

C1
C1— ('2— C1
C1

carbon tetrachloride
(CT)

H CI
\ /
C = C
/ \
C1 C1

trichloroethene

(T CE)

trans-1, 2 dichloroethene
(trans-DCE)

CI“?

0—p—c—H
|

a H

1,1,1 trichloroethane

(TC/1)

C1 C1

l |
0—p—f—41
H H

1,1,2 trichloroethane
(1,1,2 T CA)

Cl

H—CIZ—Cl
H

methylene chloride

Figure 1.1. Commonly chlorinated organic compounds found in groundwaters.

Table 1.1. Estimated annual production of the most important hydrocarbons in the U.S. and its

applications.8

Chlorinated
hydrocarbon

Tetrachloroethene (PCE)

Trichloroethene (TCE)

1,1 Dichloroethene (1,1 DCE)

Vinyl chloride

1,1,2 Trichloroethane

1,1,1 Trichloroethane

1,1 Dichloroethane

1,2 Dichloroethane

Monochloroethane

Carbon tetrachloride (CT)

Trichloromethane

Dichloromehtane

Monochloromethane

" from Fetzner (1998)
NA. = not available

Production
(* 1 03 tons)

165

110

NA.

6000

NA.

327

NA.

7200

70

150

230

162

390

Major use

Solvent for dry cleaning, metal degreasing,
textile finishing, dyeing, extraction
processes; intermediate for the production
of trichloroacetic acid and some
fluorocarbons.

Solvent for vapor degreasing in the metal
industry and for dry cleaning; extraction
solvent; solvent in formulations for rubbers,
elastomers, and industrial paints.

Basic material for polyvinylidene chloride and its
copolymers; production of 1,1 ,1-
trichloroethane.

Production of polyvinyl chloride (PVC);
production of chlorinated solvents
(primarily 1,1,1- trichloroethane).

Intermediate for production of 1,1,1-
trichloroethane and 1 , l -dichloroethane.

Dry cleaning; vapor degreasing; solvent for
adhesives and metal cutting fluids; textile
processing.

Feedstock for the production of 1,1,1-
trichloroethane

Production of vinyl chloride; production of
chlorinated solvents such as 1,1,1
trichloroethane and tri- and
tetrachloroethene: synthesis of
ethylenediamines.

Production of tetraethyllead; production of
ethylcellulose; ethylating agent for fine
chemical production; solvent for extraction
processes.

Production of trichloromonofluoromethane and
dichlorodifluoromethane; solvent.

Production of monochlorodifluoromethane (for
the production of tetrafluoroethene, which is
used for the manufacture of Hostaflon and
Teflon); extrath for pharmaceutical
products.

Degreasing agent; paint remover; pressure
mediator in aerosols; extraction technology

Production of silicones, tetramethyllead,
methylcellulose; other methylation
reactions.

consequently, the energy required to remove chlorine atoms is less than fluorine.
Bouwer et al. (1981) and Bouwer and McCarty (1983(3); 1983(b)) were the first
to demonstrate conclusively that biological transformation of these compounds not only
could occur, but occurred at a faster rate than abiotic transformations. They
demonstrated that a microbial consortium enriched under methanogenic conditions with
acetate as a source of carbon could transform C1 and C2 halocarbons into carbon dioxide
and methane. Of the compounds screened, only carbon tetrachloride was transformed

using enrichments under denitrifying conditions.

1.2 Natural attenuation: overview of processes

Due to the complexities in the subsurface and the inherent problems and costs
associated with conventional treatment technologies, interest in natural attenuation of
groundwater contaminants has increased over the last decade (Azadpour-Keeley et al.
2001). Numerous potential advantages of natural attenuation over more traditional

engineered approaches have been identified (Swett and Rapaport 1998):

0 since it is an in situ process, less volume of remediation wastes are generated

0 the site can be used with minimal disruption while remediation is occurring

0 few surface structures are required

- implementation of natural attenuation enables managers, remediators, and
regulators to differentiate between sites that are cleaning themselves from those
that are not, so that engineering resources can be allocated to sites where they will

provide the greatest benefit

0 associated costs are lower than any engineered remediation technology

A challenge to acceptance of natural attenuation can be public perception since it may
be viewed as a “do nothing approach” in which responsible parties were employing
natural attenuation to avoid remediation costs. As scientists developed better
understanding of the processes and disseminate this information, community and
regulatory perceptions have changed as natural attenuation has become defined (EPA,
1998) and accepted as a remedial approach.

However, real technical challenges exist with natural attenuation. Usually, long
time frames are required to achieve contaminant concentration levels that are protective
of the human health and the environment. Moreover, these times are not easily predicted
even with historical data. Depending on the complexity of the site hydrogeology,
characterization costs can be high and long term monitoring is necessary to Show the
effect of changing conditions on the overall remedial effectiveness of natural attenuation.

Natural attenuation has been defined as all naturally occurring physical, chemical,
and biological processes that can reduce water-phase concentration of contaminants
(NRC 2000). These processes can be divided in two categories: non-destructive or
destructive (EPA 1998). Non-destructive are those that reduce the contaminant’s
concentration but not the total contaminant mass. On the other hand, destructive
processes are those that reduce both contaminant mass and concentration. They include

biological and chemical transformation of the contaminants.

Non-destructive processes
Non-destructive processes are all physical mechanisms and include advection,

hydrodynamic dispersion, molecular diffusion, sorption, dilution, and volatilization.

Advection is the migration of solutes in the direction parallel to the groundwater flow.
Transport by advection alone results in strong, sharp concentration fronts. The advective

velocity of the groundwater is given by:

v=——4WI OJ)

C

where v is the groundwater velocity vector [L'T']; V is the gradient operator (6/6x, 6/6y,
6/62); x, y, z, are spatial coordinates [L]; K is the hydraulic conductivity tensor [L-T']; h

is the hydraulic head [L]; and (be is the effective porosity.

Hydrodynamic dispersion results in the spreading of contaminants in directions
longitudinal and transverse to the principal groundwater flow direction. This
phenomenon is attributed to two physical processes: mechanical dispersion and molecular
diffusion. Mechanical dispersion is the mixing of contaminants that result from small
scale variability in velocity around the average linear groundwater velocity. Molecular
diffusion occurs when the contaminant migrates due to the thermal-kinetic energy of the
solute molecules. A hydrodynamic dispersion coefficient which accounts for the two

processes can be obtained from (Freeze and Cherry 1979):

D=av+D‘ 02)

where D is the hydrodynamic dispersion coefficient tensor, [Lz-T'], a is the dispersivity

fi
tensor, [L]; and D is the effective molecular diffusion coefficient for the solute [Lz-T'].

At normal groundwater velocity, mechanical dispersion is often more dominant than the
Spreading that occurs due to molecular diffusion.

Sorption is the process whereby contaminants partition from the water and adhere
to the soil particles comprising the aquifer matrix. This mechanism results in retardation
or slowing of solute migration relative to the advective groundwater velocity. The ratio
of groundwater to contaminant velocity is a measure of the relative slowness of the

contaminant:

R=— (1.3)

where R is the retardation coefficient; v1 is the advective groundwater velocity in the

longitudinal direction, [L-T‘]; and vc is the contaminant retarded velocity [L-T'].

The retardation coefficient can be estimated from:

R=1+——PbKd (1.4)

where pb is the bulk mass density of the sediments, [M-L'3]; Kd is the partition

coefficient between soil particles and water, [L3- M'l]; and (l) is the soil porosity, [L3 L'3].

It has been found that partition coefficient values normalized to the total organic carbon
content eliminates the variations observed between different soil types (EPA 1998).

Therefore, the partition coefficient can be estimated by:

Kd = Koc foc (1.5)

where K0c is the organic carbon-water partition coefficient, [L3 - M'l]; and f0c is the

fraction of total organic carbon in the sediments [M° M'l].

Volatilization, while not considered a destructive mechanism, does reduce the
contaminant mass in the groundwater. Factors that affect the volatilization rate of
contaminants in groundwater include concentrations, depth dependant concentration
gradients, Henry’s Law and diffusion coefficients, mass transport coefficients, sorption,
and water and soil gas temperature. The partition coefficient between the contaminant

concentration in water and soil gas is given by Henry’s Law:

ca =HCl (1.6)

where C8 is the contaminant concentration in soil air, [M- L'3]; C] is the contaminant

concentration in water, [M- L3]; and H is Henry’s constant (dimensionless). The impact
of volatilization for chlorinated compounds can usually be neglected. It has been found

that volatilization could have a significant impact only for vinyl chloride removal.

Destructive processes

Destructive attenuation mechanisms consist of biological and abiotic processes
that result in transformation and reduction of the contaminant’s mass. Biological
mediated processes for chlorinated solvents in groundwater include reductive

dechlorination, cometabolism, and direct biological oxidation. Reductive dechlorination

and direct biological oxidation are the most common processes responsible for biological
destruction of chlorinated compounds in groundwater under natural conditions.
Hydrolysis is an abiotic process in which H20 or 0H substitutes for an electron-
withdrawing group such as chlorine. 1,1,1-TCA is the only chlorinated compound that
can be hydrolyzed within the one to two-decade time span under conditions likely to be
found in most groundwater (NRC 2000).

Reductive dechlorination of chlorinated compounds has been documented
elsewhere (Ferguson and Pietari 2000; Ndon et al. 2000; Maymo-Gatell et al. 1997;
Vogel et al. 1987; Vogel and McCarty 1987). In this process, the chlorinated compounds
serve as an electron acceptor in anaerobically mediated biological processes. An
appropriate source of carbon for microbial grth must be available in order for
reductive dechlorination to occur. Potential carbon sources include low molecular weight
organic compounds, fuel hydrocarbons, byproducts of fuel hydrocarbons, or naturally
occurring organic matter. The steps involved in the biological degradation of PCE by
reductive dechlorination are illustrated in Figure 1.2 (McCarty 1997).

Gibbs free energy for the reductive dechlorination of PCE to ethene with
hydrogen as the electron donor is shown in Table 1.2. The AG°' values in this table
indicate that all these reactions are feasible at standard temperature and pressure from the
thermodynamic standpoint. Also, redox potential ranges at which reductive
dechlorination reactions are feasible are given in Figure 1.3 (Nyer and Duffin 1997).

Due to the oxidized nature of chlorinated compounds they are unlikely to undergo direct
biological oxidation. However, it has been observed that vinyl chloride can be oxidized

to carbon dioxide and water via iron reduction.

Hydrolisis

Fermentation

Acetate and
Hydrogen
Formation

Oxidation

I Sulfate if

     

Electron
Donors

 

 

Mixed
Complex
Organic
Materials

 

 

 

 

Organic
Monomers

 

 

 

 

 

 

Alcohols
and
Acids

 

 

all Sulfide l

 

I Iron (III) lfi

Carbon

 

   

Electron
Acceptors

=1] Iron (11) l

 

 

 

Dioxide

 

=1] Methane I

 

[ PCE if TCE )L cis-DCE }[

Figure 1.2. Steps in the degradation of PCE by reductive dechlorination.
The chlorinated compounds must compete for electrons with

H Ethene l

sulfate, iron, and carbon dioxide (from McCarty, 1997).

10

End
Products

Table 1.2. Gibbs free energy for reductive dechlorination of chlorinated
ethene compounds.

 

 

AG°'
Reductive dechlorination reaction (kJ/mol)a
C2CI4 + H2 H CzHC13 + I'I’+ + Cl. '17I.8
C2HCI3 + H2 H C2H2Cl2 + H+ + CF 466.1
CszClz + H; H C2H3Cl + I-F + Cl' -l44.8
C2H3Cl + H2 H C2H4 + I'I+ + Cl. '154.5

 

a from Dolfing (2000)

1.3 Natural attenuation of chlorinated solvents

Several investigations have demonstrated the biodegradability of chlorinated
solvents in natural environments (Clement et al. 2002; Davis et al. 2002; Rbling and van
Verseveld 2002; Witt et al. 2002). However, based on field evidence, the likelihood of
using natural attenuation as a stand alone approach to bring contaminant concentration to
levels that do not represent a risk to the human health and the environment are moderate
(Macdonald 2000). It has been estimated that only 20% of sites contaminated with
chlorinated organics may be amenable to using just natural attenuation (Swett and
Rapaport 1998).

Under anaerobic conditions, reductive dechlorination has been identified as the
major mechanisms for the biological destruction of chlorinated compounds (McCarty
1997). A carbon source capable of creating a reduced environment is required for this
process to occur naturally. This criterion imposes a limitation on the application of
natural attenuation to sites where presence of carbon source ensures a long term

biodegradation of chlorinated solvents.

ll

I000 ——

 

02 + 4H+ + 4e' —+ 2H,o (Eh = +820)

 

 

 

AEROBIC
2NO3' + 12H+ +10e-——> 2N2 + 6H20 (Eh = +740)
500 __ Mno,(s) + HCO3' + 314* + e' —> MnCO3(s) + 211,0
(Eh = +520)
ANAEROBIC

possible range for
reductive dechlorination
1

0—.—

 

FeOOH(s) + HCO3' + 211* + e' —. FeCO3 + 2H,o
(Eh = -50)

Decreasing amount of energy released during electron transfer

__ so}- + 9H+ + 8e' ——+ Hs- + 4H20 (Eh = -220)

. /v I _— €02 + 8H+ + 8e’ —9 CH4 + 2H20 (Eh = -240)

Optimal range
V for reductive
dechlorination

 

 

 

-500 __

Figure 1.3. Redox potential (Eh) in millivolts for various electron acceptors in

groundwater. Values for redox pairs at 25°C and pH = 7.0 (modified from
Nyer and Duffin, 1997).

Numerous protocols for evaluating natural attenuation of chlorinated solvents
have been developed (NRC 2000). These protocols have been applied to document
natural attenuation of chlorinated solvents at numerous sites (Clement et al. 2002;
Wiedemeier et al. 1997; Witt et al. 2002). However, they are used to reach conclusions

about whether a site is candidate for natural attenuation based on approaches that do not

12

consider the field experience or the literature to date.

1.4 Hypothesis and research objective

The research presented in this dissertation provides an insight into the current
discussion of natural attenuation for remediation and restoration of chlorinated solvent
contaminated aquifers. It has been emphasized that existing methodologies for
evaluating natural attenuation of chlorinated solvents should be replaced by approaches
considering the specific conditions of the site.

The hypothesis for this research is that at the Schoolcraft site reductive
dechlorination was the dominant mechanism responsible in the past for the reduction of
the contaminants; however it is no longer a major contributor to the natural attenuation
process occurring at this site. The specific objective of this research is to develop a
methodology for evaluating natural attenuation of chlorinated solvents considering the
hydrogeology, microbial processes, and geochemical processes linked to the historical
development of a contaminant plume. This will help to identify which of the natural
attenuation components is the current dominant mechanism and whether or not this

technology can be applied at this site

1.5 Scope of work
To accomplish the specific objective, the following approach was undertaken in
this research:

0 Characterization of hydraulic parameters influencing transport and distribution of
contaminants in the subsurface (Chapter 2)

13

0 Characterization of natural attenuation processes based on geochemical
parameters and contaminant concentration data (Chapter 3)

0 Development of a conceptual and numerical model incorporating the
hydrogeology, microbial processes, and geochemical processes linked to the
contamination history of the aquifer under consideration (Chapter 4)

The case study for this research is a VOC contaminated aquifer located in
Schloolcraft, MI. Past industrial and commercial activities have resulted in the
development of a plume of chlorinated organic contamination extending approximately

2km from the suspected source of contamination.

1.6 Literature cited

Azadpour-Keeley, A., Keeley, J. W., Russell, H. H., and Sewell, G. W. (2001).
"Monitored natural attenuation of contaminants in the subsurface: Processes."
Ground Water Monitoring and Remediation, 21 (2), 97-107.

Bouwer, E. J ., and McCarty, P. L. (l983(a)). "Transformations of 1- and 2-carbon
halogenated aliphatic organic compounds under methanogenic conditions."
Applied and Environmental Microbiology, 45, 1286-1294.

Bouwer, E. J ., and McCarty, P. L. (1983(b)). "Transformations of halogenated organic
compounds under denitrifying conditions." Applied and Environmental
Microbiology, 45, l 295- 1 299.

Bouwer, E. J ., Rittrnan, B. E., and McCarty, P. L. (1981). "Anaerobic degradation of
halogenated 1- and 2-carbon organic compounds." Environmental Science &
Technology, 15(5), 596-599.

Clement, T. P., Truex, M. J., and Lee, P. (2002). "A case study for demonstrating the
application of US EPA's monitored natural attenuation screening protocol at a
hazardous waste site." Journal of Contaminant Hydrology, 59(1-2), 133-162.

Davis, J. W., Odom, J. M., DeWeerd, K. A., Stahl, D. A., F ishbain, S. S., West, R. J.,
Klecka, G. M., and DeCaroliS, J. G. (2002). "Natural attenuation of chlorinated
solvents at Area 6, Dover Air Force Base: characterization of microbial
community structure." Journal of Contaminant Hydrology, 57(1-2), 41-59.

14

Dolfing, J. (2000). "Energetics of anaerobic degradation pathways of chlorinated
aliphatic compounds." Microbial Ecology, 40(1), 2-7.

U.S. Environmental Protection Agency (EPA). (1998). "Technical protocol for evaluating
natural attenuation of chlorinated solvents in groundwater." EPA/600/R-98/128,
USEPA, Cincinnati, OH.

Ferguson, J. F., and Pietari, J. M. H. (2000). "Anaerobic transformations and

bioremediation of chlorinated solvents." Environmental Pollution, 107(2), 209-
215.

Fetzner, S. (1998). "Bacterial dehalogenation." Applied Microbiology and Biotechnology,
50(6), 633-657.

Freeze, R. A., and Cherry, J. A. (1979). Goundwater, Prentice Hall, Inc.

Holliger, C., and Schumacher, W. (1994). "Reductive dehalogenation as a respiratory
process." Antonie van Leeuwenhoek, 66, 239-246.

Macdonald, J. A. (2000). "Evaluating natural attenuation for groundwater cleanup."
Environmental Science & Technology, 34(15), 346A-353A.

Maymo-Gatell, X., Chien, Y., Gossett, J., and Zinder, S. (1997). "Isolation of a bacterium
that reductively dechlorinates tetrachloroethene to ethene." Science, 276(6), 1568-
1571.

McCarty, P. L. (1997). "Breathing with chlorinated solvents." Science, 166, 1521-1522.

Ndon, U. J., Randall, A. A., and Khouri, T. Z. (2000). "Reductive dechlorination of
tetrachloroethylene by soil sulfate-reducing microbes under various electron
donor conditions." Environmental Monitoring and Assessment, 60(3), 329-336.

NRC. (2000). Natural attenuation for groundwater remediation, National Academy
Press, Washington, DC.

Nyer, E. K., and Duffin, M. E. (1997). "The state of the art bioremediation." Ground
Water Monitoring and Remediation, 2, 64-69.

Prakash, B., and Gupta, S. K. (2000). "Effect of carbon source on PCE dehalogenation."
Journal of Environmental Engineering, 126(7), 622-628.

Rbling, W. F. M., and van Verseveld, H. W. (2002). "Natural attenuation: What does the
subsurface have in store?" Biodegradation, 13(1), 53-64.

Sims, J. L., Suflita, J. M., and Russell, H. H. (1991). "Reductive dehalogenation of
organic contaminants in soils and ground water." EPA Ground Water Issue.

15

Swett, G. H., and Rapaport, D. (1998). "Natural Attenuation: Is the Fit Right?" Chemical I
Engineering Progress, 94, 37-43.

Vogel, T. M., Criddle, C., and McCarty, P. L. (1987). "Transformation of halogenated
aliphatic compounds." Environmental Science & Technology, 21, 722-736.

Vogel, T. M., and McCarty, P. L. (1987). "Abiotic and biotic transformation of 1,1,1-
trichloroethane under methanogenic conditions." Environmental Science &
Technology, 21 , 1208-1213.

Wiedemeier, T. H., Wilson, J. T., and Kampbell, D. H. "Natural Attenuation of
Chlorinated Aliphatic Hydrocarbons at Plattsburgh Air Force Base, New Yor ."
Proceedings of the Symposium on Natural Attenuation of Chlorinated Organics in
Ground Water, 76-84.

Witt, M. E., Klecka, G. M., Lutz, E. J., Ei, T. A., Grosso, N. R., and Chapelle, F. H.
(2002). "Natural attenuation of chlorinated solvents at Area 6, Dover Air Force
Base: groundwater biogeochemistry." Journal of Contaminant Hydrology, 57(1-
2), 61-80.

16

CHAPTER 2

HYDROGEOLOGIC CHARACTERIZATION OF A
VOC CONTAMINATED AQUIF ER ZONE

2.1 Abstract

A methodology that couples laboratory and field data with numerical optimization
methods was developed to estimate aquifer physical parameters that influence
distribution and migration of contaminants. Hydraulic conductivity, porosity, and tracer
test data were used to characterize an aquifer zone contaminated with VOC using an
optimization method for solving the generalized groundwater flow and transport
equations. Based on the results a strong correlation between laboratory and field
determined hydraulic conductivity was found. These tests revealed the presence of high
conductivity zones that act as preferential contaminant pathways. Hydraulic conductivity
values for these zones are in the range of 10'1 to 1cm/s. Based on the optimization
technique, simulated depth-specific average linear velocities agreed reasonably well with
observed tracer velocities. A simulated average linear velocity of 98.4 cm/day was found
for the fastest zone in this aquifer, a value two times higher than the velocity at the
slowest depth interval. Sensitivity analysis using the root mean square errors (RSME)
showed that a 10% change in porosity field yielded a 25% average linear velocity
deviation from optimal values. Although length scale dependency of longitudinal
dispersivity was not taken into account, a single value per depth interval described tracer
distribution reasonably well. This methodology can be used to characterize aquifer zones

intended to be use for engineered remediation purposes.

17

2.2 Introduction

The physical characteristics composing the aquifer matrix exert a profound
influence on the movement and distribution of chemical compounds in groundwater. It is
widely accepted that transport of contaminants migrating to the saturated zone of an
aquifer occur through high energy or less resistant stratigraphic units (Hyndman et al.
2000b; Lee et al. 2001). Therefore, study of the fate and transport of pollutants in a
contaminated site must incorporate a detailed hydraulic characterization. Moreover, if an
assessment of a contaminated site is being conducted, sufficient site-specific data is
necessary to develop reliable numerical models that incorporate the heterogeneous nature
of subsurface environments. This is critically important if numerical models are going to
be used for predicting the future extent of contaminant migration (Wiedemeier et al.
1998). Among the many physical properties of geologic units that influence subsurface
contaminant migration, the most important are hydraulic conductivity, porosity, and
dispersivity. These parameters are strongly correlated to the heterogeneous nature of the
aquifer matrix.

Field and laboratory methods have been developed to estimate these parameters.
Hyndman et al. (1994) developed a theoretical algorithm, which combines field data from
cross-well tomograms with tracer concentration tests to estimate transport parameters of
lithologic zones in two dimensions. That methodology was further expanded and applied
to a field site to estimate properties in three dimensions (Hyndman and Gorelick 1996;
Hyndman et al. 2000b). Cho et al. (2000) developed a field procedure to measure
vertical profiles of hydraulic conductivity using direct push methods. They identified

small-scale variations in hydraulic conductivity that would have not been detected by

18

laboratory or conventional field methods. A similar procedure was used at a BTEX
contaminated site to estimate the role of small-scale heterogeneities upon contaminant
distribution (Hurt et al. 2001). Concentrations of BTEX compounds were highly
underestimated when variations in parameters such as hydraulic conductivity and
porosity were not properly measured in the field.

In the laboratory, several assays have been developed to estimate hydraulic
conductivity, porosity, and dispersivity from core samples (Freeze and Cherry 1979).
Although it has been recognized that it is unlikely to reproduce field conditions, a good
indication of approximate values of transport parameters can be obtained with well
designed laboratory experiments.

The objective of this study is to develop and test a methodology to quantify
heterogeneities and physical properties of an aquifer region contaminated with volatile
organic compounds. The main goal is to identify small-scale heterogeneities that
influence the migration and distribution of contaminants in an unconfined aquifer. For
this purpose, a series of laboratory and field tests coupled with numerical groundwater
flow and transport models were developed to estimate hydraulic conductivity, porosity,

and optimal depth-specific dispersivity values within the region.

2.3 Site description

The Village of Schoolcraft is a small rural community located approximately 16
km south of Kalamazoo, MI, USA (Figure 2.1). The unconfined aquifer underneath the
village has been contaminated with organic and metal compounds as result of previous

industrial and commercial activities in the village. Regional and local hydrogeologic

l9

.88 .835 «0 5:83 .mm oBME

 

      

1 mMMHmE ZH mq<um

com com 0

6C ESE 00>

 

 

 

 

 

 

 

 

20

conditions have been described elsewhere (Lipinski 2002; Dybas et al. 1998; Mayotte et
al. 1996). The focus of this investigation is an area where a pilot scale study is being
conducted to evaluate the feasibility of remediating a volatile organic compound plume in
this aquifer (Figure 2.1).

Currently, bioremediation and natural attenuation studies are being conducted on-
site to evaluate the potential of using a combined treatment strategy to reduce
concentration of organic compounds to acceptable levels. For this purpose, two “side-by-
side” delivery and monitoring well networks have been drilled approximately in the
plume’s center of mass (Figure 2.2). The purpose of the study on the north well network,
i.e. stimulation grid, is to stimulate the native microbial flora to degrade VOC
contaminants. Bioaugmentation effects on contaminant degradation are being evaluated

on the south well network, i.e. the augmentation grid (Figure 2.2).

2.4 Methods
Hydraulic conductivity and porosity tests

To estimate depth-specific values of hydraulic conductivity and porosity, soil
cores from 9m to 24.3m below ground surface (bgs) (in 1.5m intervals) were collected
from monitoring well locations using the Waterloo cohesionless continuous sand sampler
method (Dybas et al. 1998). These cores were visually inspected in the laboratory using
the AST M Standard D2488-00 Practice for Description and Identification of Soils (Visual
Manual Procedure) to determine relative grain size distribution of the glaciofluvial
sediments composing the unconfined aquifer. Cores from pump wells (9 — 24.3 m bgs)

were inspected on-site since they were extracted in approximately 6 meter intervals and

21

 

.Amvtw confine—Ema can nozflsfiumv «one beam 5 €9.33 :03 .N.N DSwE

 

 

lll 20.732
i Wmm :W G
3.2.3 >mm>imo@ I am: a.
oEeomEOmQ I we a. 4% «as a. a.
3m? 3258 xbd® , n... 46 ® a my. .4 (A mxmhmEEBSw
we .
925A camotzoz e W .1. oo@ 9% week.” a. S m o
llll I wflflfl .. t... We coo... o
a, e a. a.
> ”.mm .ww my we :9 ® KW
< Q m. .V. «0 ® 9: 7mm
”.32 o o. no 9 732
a .
4 O
.V
a. we.
saw I
I
0 do 0
as e 6% one 3% W
G we ® ® a a.
0.32 c D we Nrmw
.p/s oQ0 e s». .% mm 0
2W0 #9 were am. he :22
a e e e .0 w
are 9% % EV ® Var <
. d x s
0/8 0.3:.Wd W0 RIP 0%,.
no 6 Evan; Sasséa
a. a.
a as.
Q
m.mm

 

 

22

undisturbed transport to the laboratory was not possible. The Rotosonic drilling method
(Lewis 2001) was used for those holes. Boring logs for all monitoring and pumping
wells are given in Appendix 1.

After visually classifying the cores, approximately 250 sediment samples were
collected for hydraulic conductivity and porosity tests. A constant head permeameter
apparatus was used to estimate the hydraulic conductivity of repacked soil samples
(Hoard 2002). Bail tests were also performed on selected wells in the field screened at
different depths to evaluate the accuracy of laboratory methods. The methods of
Hvorslev (1951) and Bower and Rice (1976) were used.

The following equation was used to estimate the total porosity of sediment

samples:

¢=1—9—b— (2.1)
95

where cl) is the porosity; pb is the bulk mass density of the sample [M-L'3]; and pS is the

particle mass density [M-L'3] (Freeze and Cherry 1979).

Field tracer tests

Field tracer tests were conducted on the augmentation and stimulation well
networks to estimate flow and transport parameters that influence distribution of solutes
within this aquifer section. The main objective of these experiments was to study the
impacts that the stratigraphy exert upon contaminant migration and distribution in this

region.

23

Table 2.1. Details of tracer experiments.

 

 

Solute Parameter value Description.

Bromide Tracer injected into DW ’s by
injected concentration (t =0) [00 ppm pumping water fi'om F C W ’5,
injection into DWA I -D WAZ adding the salute to the
pumping rate 1_ 3 x [0'3 m3/s (each well) pumped water and injecting it
pumping time 4 hrs back into DW ’s.

Fluorescein Same as Bromide injection.
injected concentration (t=0) 100 ppb
injected into DWAl-DWAZ, DWSI-
pumping rate DWSZ
pumping time 1.3 x I 0'3 m3/s (each well)

4 hrs

 

‘
see Figure 2.2 for well notation

F luorescein and bromide tracers were chosen because experiments have shown
they are relatively conservative tracers, do not sorb to soil particles, do not affect
bacterial activity, and are very easy to detect (Hyndman et al. 20003). An “injection-
extraction” strategy was used to deliver tracer solution to the saturated zone of the
aquifer. To achieve uniform tracer delivery in the geologic formation, a 4hr injection
time was chosen since preliminary numerical models predicted an approximate 80%
tracer breakthrough in the extraction well at the end of the pumping period. The screened
interval for the extraction and injection wells in the network is between 18.3 to 25.0m
bgs. Details of the field tracer experiments are provided in Table 2.1.

Samples for tracer measurement were taken at downgradient multi-level
piezometers at time and depth intervals specified in Table 2.2. A bromide electrode
(Cole-Panner) and a digital field fluorometer (Model 10-005-CE Turner Designs Inc.)
were used to measure bromide and fluorescein concentrations in collected samples,

respectively.

24

Table 2.2. Logistics of tracer tests.

Sampling time

 

Interval Sampled. (days after tracer
Sampled wells (m bgs) injection)
Augmentation well network
Bromide Tracer Test
MPA4—MPA6 16.8to 24.] 7
MPA7 16.8 to 24.1 8, I4, 15
MPA8—MPA9 15.8t024.l 8, 12, I4, 15
MFA/0 16.8t024.l 8
MFA]! 16.8to 24.! 12,14, 15
MPA12 16.8t024.l 12, 14
F luorescein Tracer Test
MPA4-MPA5 16.8t024.l 1-4, 6-8, 10, I4
MPA6 16.8to 24.1 3, ,6-8, 10, I4
MPA7—MPA10 15.8t024.l 8.10.14.17.22.
24, 27, 30
MPAIl—MPAIZ 16.8to 24.1 8,10,14,17,22.
24, 27, 30
Stimulation well network
F luorescein Tracer Test
MPA7—MPA10 17.1 to 24.4 7,9,12,15
MPAll—MPA15 18.0to 24.1 7,9,12,15

 

t
All multi-level wells consist of multiple 15.3cm screens spaced at 0.9 m intervals

Conceptual and numerical models

A conceptual model for this site was constructed to develop an understanding of

the important features that affect solute transport within this region. A northwest-

southeast cross section through the site (Figure 2.3) shows the main geologic features of

the region. Groundwater flow in the vicinity is from northwest to southeast at an average

velocity of lScm/day (Dybas et al. 2002; Mayotte et al. 1996) and recharge for this area

is approximately 23.7cm/yr. Pertinent boundary conditions were assigned based on a

regional-scale model constructed by Lipinsky (2002). Figure 2.4 shows the boundaries

for the regional aquifer which control groundwater flow in the study area.

25

 

 

 

MW-8

 

 

 

MW-S

MW-2

 

 

 

 

 

 

0m

 

{IT
9
D
Z
<
C0
C3
l-I-l
3
C?
._I
.4
LL]
3
ll

E
5,1,.
:8
”3c:
Ea
<2

wé
OLD
E...)
341—1
8:

...I
9.
>.
<1
._1
U

 

 

 

 

 

26

Figure 2.3. Cross section A-A’ (Figure 2.2) through monitoring wells MW -2, MW-5, and MW-8.

.Amoom .3835 :85 Stave voaaqoons 05 E 26: 833.5:on BEES 75:8 :2: 858m :32 .vd onE

Tammi

new 2.3..

, buucaon two: 3.3%
w

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

canon 26: o:

 

 

 

L.-_I..,lil Fax/Mi

 

 

 

 

 

 

 

+2

27

Information compiled from the conceptual model was incorporated into a three
dimensional numerical groundwater flow and transport model. The objective for this
experiment was to understand the influence of local heterogeneities on solute distribution
in this region. By analyzing vertical profiles of tracer concentration a quantitative
description of those heterogeneities can be obtained.

A three dimensional finite difference grid was constructed by telescopic mesh
refinement (Anderson and Woessner 1992) from the numerical regional groundwater
flow model of Lipinski (2002). The horizontal resolution of the grid was 15cm in both,
longitudinal and transverse direction. Vertical resolution of the grid varied from 90cm at
the bottom of the aquifer to 620cm at the top resulting in 15 model layers. This grid
discretization was adequate to represent the physical resolution of the monitoring well
networks on the site while minimizing artificial dispersion effects. Local boundary
conditions were determined based on the regional groundwater flow model. Figure 2.5
Shows the numerical model grid in horizontal and vertical directions.

Groundwater flow was simulated using the classical mathematical expression

from Bear (1979):

Vo(K-Vh)+W=Sy-%IE- (2.2)

where V is the gradient operator (6/6x, B/éy, 53/62); x, y, z, are spatial coordinates [L]; K
is the hydraulic conductivity tensor [L-T']; h is the hydraulic head [L]; W is the fluid

source sink term [L'T']; Sy is the specific aquifer yield; and t is the time [T].

28

 

 
 
 

3 lllllllllllllllllllllllllllllllllllllllllllllll|||l|ll||ll|||l|llll||lll|l||l||ll|lll|||l|lll||l|||llllllll

‘° [ljllwlllllullllfl Illllmlllllll"lllillllllllllllllllllll'lll lllllll “mIIllllll ll!
“”15 2 'lIIIlllIIl'llIIhl “.[Il] '[1 ll“! [. 'l' l Illlllllllll'll [gull “lilllll' ,

 

(b)

Figure 2 5 Numerical model grid in (a) horizontal, and (b) vertical directions. (No. of
cells in the horizontal is 52,890, with 15 layers in the vertical direction.

29

This equation was coupled to the advection-dispersion equations for simulating

tracer transport:

%=VO(DVc)—V(voc)+c W

 

(2.3a)

e

v = -—Vh (2.3b)

C

where c is the tracer concentration in pore fluid [M°L'3]; v is the groundwater velocity

vector [L'T’]; c' is the tracer concentration in source or sink [M-L'3]; be is the effective

porosity; and D is the hydrodynamic dispersion coefficient tensor [LZ-T'] with
coefficients a-v, where a is the dispersivity vector [L].

Equations 2.2 and 2.3 were solved numerically using MODF LOW (McDonald
and Harbaugh 1988) and MT3DMS (Zheng and Wang 1999), respectively. For the flow
model, hydraulic conductivity and porosity data obtained from field and laboratory
analysis were kriged to the 3-D finite difference grid. Two stress periods were Simulated:
1) 4 hr of tracer injection followed by; 2) 29.83 days of tracer transport under natural
gradient influence, for a total simulation time of 30 days.

A constrained optimization method was developed to solve the objective function

(Equation 2.4) and find a set of optimal dispersivity values for the geologic formation.

to a)= Z3:— :1“ )) (2.4)

30

In the previous equation i is the index for the dispersivity at a particular depth; N
are depths at which tracer data was collected in monitoring wells; and the A operator is

given by:

c bs" ——c—d5im|t
Mai )= (3.1—d o (25)
Z] cobs .

where the integral represents normalized absolute differences between observed (cobs)

and simulated (csim) tracer concentrations over the entire simulation time t, for depth

interval i; and the summation is over all multi-level piezometers M, located at depth

interval i. Simulated tracer concentrations in Equation 2.5 are given by Equation 2.3.
The denominator (710) in Equation 2.4 is the operator evaluated at an initial set of

dispersivity values.

The objective function (Equation 2.4) was minimized by solving a set of
generalized Kuhn-Tucker equations (Phanikumar et al. 2002). Methods for solving these
equations are available in many programming packages. The sequential quadratic
methods in MATLAB Optimization Toolbox (Coleman et al. 1999) were used in this
study. The MATLAB script for this optimization is provided in Appendix 11.

Depth intervals included in the objective function (Equation 2.4) were 18.6, 19.5,
20.4, 21.3, 22.3, and 23.2m bgs. Multi-level piezometers at 15.8, 16.8, and 17.7 were not
included because tracer was not detected at any of these depths. This was expected since

the screen top of injection wells is at 18.3m bgs. Also, piezometers with insignificant

31

tracer concentration were excluded in the formulation of Equation 2.5.

2.5 Results and discussion
Hydraulic conductivity and porosity

Figure 2.3 Shows the main features found during the visual classification analysis.
Material from shallower depths (0 ~ 10m bgs) of this aquifer is mostly fine sand
changing to medium sand as one moves deeper into the formation. From 10m bgs to
approximately 18m bgs the aquifer is composed of medium to coarse sands. A layer of
poorly sorted sands with gravel and cobbles was identified at an average depth of 19m in
most of the cores. The thickness of this layer varies from 0.5m to about 1.0m. The
underlain sediments changed again to a mixture of coarse and medium sands with an
approximate thickness of 3m. A second layer of coarse material at an approximate depth
of 23m bgs was apparent from the analysis. Well graded gravels with a significant
amount of coarse gravel (40mm average diameter) were the principal materials of this
second layer. Thickness of this layer is approximately 1.2m on the average. Below this
material, a medium to coarse sand layer was deposited. Cobbles of significant size (~
100mm) were found interbedded within this layer. Hard, gray clay was found
approximately at 24.5m bgs in soil cores from MWI to MW3. Apparently, this clayey
layer gradually slopes downward as evidenced by the location (25.9m bgs) at which clay
was found in cores from wells located at the southeast boundary of the network.

Based on the visual classification, it is evident that distinct stratigraphic units
(may act as preferential flow paths for contaminant migration) exist at average depths of

20 and 23m bgs. The extent of these layers; however, cannot be predicted since cores

32

were taken in a relatively small area compared to the entire extension of the VOC plume
(Figure 2.1). In a study by Dybas et al. (2002), cores from a location approximately
1.8km north of this site showed the underlying unconfined aquifer to be composed mostly
of medium sands.

Figure 2.6 and Table 2.3 show hydraulic conductivity results from laboratory
permeameter and field bail tests, respectively. Laboratory determined hydraulic
conductivity ranged from 1x10'3 to 4x10'lcm/s with an average of 6.6x10'2cm/S and a
standard deviation of 0.066cm/s. The scatter plot in Figure 2.6(a) shows that between 5
to about 15m bgs hydraulic conductivity is in the range of 10'3 to 10"5cm/s. Data is
clustered around 10"'5 to lcm/s below the 15m bgs interval.

Bail test results (Table 2.3) showed that values obtained by the Hvorslev method
were, in general, an order of magnitude higher than hydraulic conductivity obtained by
the Bower and Rice method. However, these values are within the expected range of
hydraulic conductivity for glacial outwash sediments (Fetter 2001). Laboratory measured
hydraulic conductivity correlated well to field values obtained by the Hvorslev (1951)
method.

A total porosity scatter plot from laboratory repacked samples is shown in Figure
2.6(b). Average total porosity was 0.331 with a standard deviation around the mean of
0.050. Maximum and minimum values were 0.504 and 0.208, respectively. Total
porosity of shallower sediments seems to be around 10% lower than deepest ones. From
Figure 2.6 an inverse correlation between hydraulic conductivity and total porosity can be

observed. A similar pattern was also reported in the study conducted by Hoard (2002).

33

 

 

 

 

[ 1
L + t
10 i 4' "'“
>-+ —1
+ + ++
l +N++ +++
r +41»,
b #— fl”, + + +“H’+ fit} .4
++ + +4..
6‘ K + #4” ++++
or 15 - +3 +1 3
so )- I'H' ##1##
t it
> +
g ”‘4‘ 4’s“ 4
* + 43* ¢++ + :+ 1
5 . 4* ¢+e+ + {++ ++ + ,
a 20 +++ +
0 T * H++¢T gab-3nd T
“O r + + ++ + 1
. + + + ++ X? +
0*” ++ +0 T
> + +itfi++¢+ fi' + + 4
+
25 - +++ +++Wr + —
F + 4
30 A LA A 1A A A A A L A L 1 A A I ILLAAAAAJ

-3 -2.5 -2 -1.5 -1 -0.5 0
logK(cm/s)

(a)

 

 

 

 

5 rjrrfij rTTTI I 1 VII YYT‘VVY'rTrirITTfifi
I- -4
. t .
10 + m... +
t + +++1t+++
++ ++ +
_ +$ +*.++++ T!
A > +++ v... #*+ + -I
U) +
O) 15 ’- ++++++++ t; + j
.0 r +++ *¢ +
+++++ +
E ++ 11+“: +"TI-‘1' 1
5 . + ++ ‘1 +++ *+ +++I++ + -I
320—.+.......,.+++ . -
“O ’h» +4++f+ + ++ + ‘
++~++++1I++* i
I— + + vi
l + ”an”: 4: + J
25]- :+T+ ++<I3 ++ -l
t
>- + '4
4
h J
30 4.1.1....I....I....III..I....1.II.

0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55

porosity, n

(b)

Figure 2.6. Scatter plots of (a) hydraulic conductivity and
(b) porosity.

34

Table 2.3. Bail test results.

 

 

 

 

Screen depth K (cm/s)b
Well ida (m bgs) Bower and Rice (1976) Hvorslev (1951)
MP-AZ 23.8 — 24.4 2.9 x 10’2 d 1.58 x 10'1d
MP-AlS 23.8 - 24.4 4.02x10’2 4 0.001 2,211.10] :1: 0.008
MP-A6° 22.3 —22.9 4.30x10‘2 + 0.017 2.35x10'l a: 0.091
MP-83 24.4 —25.0 328.102 a 0.004 1.82x10" 4 0.021
MP-SlS 24.4 —25.0 3.15:.10'2 :I 0.004 1.72x10'l a 0.023
w-55° 22.9-23.5 8.32x10‘2 + 0.027 4.54x10'1 4 0.146
MSU-1 19.2 — 19.8 4.11x10’2 .8 0.016 2.25x10" 3: 0.089
a see Figure 2.2 for well location

b . .
average value :I: one standard devratron

measurements taken on a nearby well not identified in Figure 2.2

d . . .
average of two bail tests; standard devratrons not calculated

To incorporate hydraulic conductivity to the flow and transport model a statistical
analysis of the data was performed. A log K histogram analysis showed this data to be
approximately normally distributed (Figure 2.7(a)). Vertical anisotropy in this geologic
formation was evident from the variogram analysis shown in Figure 2.7(b). From this
analysis, horizontal and vertical correlation lengths were 18.2 m and 4.4 m, respectively.
A variance at zero separation distance of 0.03cm2/S2 was found, with a log K variance of
0.13cm2/52. These results were similar to the findings of Hyndman et al. (2000), and
Hoard (2002).

An exponential variogram model with parameters from the experimental
variograms was used to krig hydraulic conductivity data to the numerical model grid. A
similar procedure was used for porosity data. Cross section images of log K and porosity
are shown in Figure 2.8. High K areas in the image at approximate depths of 20m bgs
and 24m bgs were found to correlate well with the two layers of coarse material
identified in the conceptual model. These high K zones might act as preferential

pathways for contaminant migration in this aquifer.

35

 

 

 

 

 

 

 

 

 

 

0.2
0.16--
6‘
g 0.12 ~- 4;:
3 9
9 a
f, 0.08-- ‘5
9 E
8
0.04 l»
0 -
I?) 3 F: 8’; 8 i3 8
‘1‘ “I‘ "I' "I' ‘I' C? o
logK(cm/s)
(a)
0.30J o horizontal /
0 vertical 0/
0.20: ,o-r‘”
3 °\ #0”
v P—
* 0.10-l v” "r
/
0.0 l , . . . .

 

 

 

lag distance, h(m)
(b)

Figure 2.7. (a) log K frequency distribution, and (b) experimental and
model variograms of the data.

36

.baea is E as. .M as 3.6 89.5 Bag .3 new;
05 a

9N0
8N0
98.0
000.0
09.0

09.0

a $38.80 30

 

e5 .5

0
Ir.
0
V
O
M
O
N
0
F1
0

00d.

on 8.”-
2 a .- m
05 N 8. F- _
S-OI %

8.0

anew M was 3

 

o

37

Tracer tests and Numerical models

A series of tracer tests were conducted to test the hypothesis that local
heterogeneities influence solute distribution in the study region. Observed depth-specific
tracer breakthrough curves were coupled to an optimization algorithm to estimate
dispersivity values across the formation. Groundwater was pumped from the flux control
well to a mixing tank where tracer was added using an in-tank recirculation system
(Figure 2.9). Solution from the mixing tank was injected into the ground in delivery
wells 1 and 2. The screened interval for these wells is between 18.9 and 25.0m bgs.

Preliminary numerical transport models using limited hydraulic conductivity from
wells in the outer boundary of the grid predicted an average tracer travel time of 3.5
days/m in the high conductivity zones. Tracer samples were taken from downgradient
wells at an interval of 7 day for 21 days. Due to the limited hydraulic conductivity data
used in this model, the raising limb of the observed tracer breakthrough was not properly
described. In this first tracer test, a l40ppm bromide solution was injected into DW-Al
and DW-AZ.

A second tracer test using fluorescein was designed with sampling intervals of l
and 2 days. Figure 2.10(a) and 2.lO(b) show breakthrough curves during tracer injection
at the flux control and delivery wells, respectively. Hydraulic conductivity results from
sediment samples of monitoring and delivery wells were incorporated into the flow and
transport model and simulation results validated field observations reasonably well.
Based on these results, about 60% tracer breakthrough occurred in the flux control after
l70min of pumping time. Tracer breakthrough at delivery wells reached a maximum

relative concentration (C/Co) of 1.8 at 4hr of pumping. Relative concentration in delivery

38

.EBmmm 838? Bush 05 .«o 893.— .oN 8&5

 

 

 

       

3ng
I
N c
6 av
25.: 3&2
6 33:2 v.25
3&2 @ 02:22 «5.2
.32 0mm 1 I G
V 223..
g 1
25:2
x25 6
«85:
O
:9. 95:25
_:l l ,ll- E: l : -l E
D _ Sm? >mm>5mo @w

3m? 405.28 x3: mow

9253

 

__
.
w 5:: 0228.202 G
_

 

 

 

39

 

O
(D
fiTfi T r

0.6 ~

0.4 l

relative concentration (C/Co)

 

 

 

 

 

 

 

 

0.2 ~
* . ' obs
Obfl’.’nxiltgLJ...1...1...<
0 40 80 120 160 200 240
time (min)
(a)
2 l i l l I
1.8 b

1.6
1.4 >

1.2 ’

 

relative concentration (C/Co)

model

 
 

 

 

 

 

 

1 l 0 obs ,
08 0. 4 J . . r 1 a . . l . . . 1 . . , 1 , , , l
0 40 80 120 160 200 240
time (min)
(b)

Figure 2.10. Tracer breakthrough curves at (a) flux control well,
and (b) delivery wells.

40

wells are greater than 1 as soon as tracer breaks through in the flux control well. When
this condition occurs, the injected concentration will increase by the relative
concentration coming from the flux control well. To model this recirculation system in
MT3DMS (Zheng and Wang 1999) a series of additions and modifications were made to
the main code and subroutines. Breakthrough results indicate that at least 2.7hrs of pump
time at a rate of 2.6x10'3m3/s is required in order to create a uniform solute distribution
region across the well screens.

Heterogeneity effects upon tracer distribution could not be evaluated during tracer
injection due to strong hydraulic gradients induced by pump operation. Dispersion
effects were more evident during transport under natural gradient conditions.

Two scenarios were simulated with the optimization approach described by
Equations 2.4 and 2.5. In the first scenario, depth-specific dispersivity values were
estimated. During the second scenario, only one value of dispersivity was used by
assuming that a single value was enough to describe the differences seen in tracer
breakthrough curves.

Tracer breakthrough curves for the optimized per-layer dispersivity and the grid
optimal dispersivity scenarios along with observed data are shown in Figure 2.11.
Breakthrough curves obtained in the optimum per-layer scenario closely matched
observed tracer concentration for some depth intervals in monitoring points A4 and A5.
Dispersivity values assigned on a per-layer basis seems a better representation than a
single dispersivity value for the entire model domain. Although dispersivity is a length
dependent parameter (Kim et al. 2002), one value per model layer was adequate for this

simulation. Peak arrival times for the per-layer simulation scenario and observed data

41

Well No.

MBA-4 MPA-S MFA-6 MFA-7 MFA-8 MFA-9

2 _— _ ._.¥ - -._‘_ _._-__—_ ._---a

 

,___-_. _.___. _____‘ __ _ __

 

relative concentration (C/Co)

i 22.3 m

 
   

 

 

 

 

A
,- 7 . 7. .3
01020010200102001020010200102030
time (days)
Figure 2.1 1. Simulated and observed tracer breakthrough curves in downgradient

monitoring points: (') observed, (—) optimum per-layer case, (— —)
single optimum value case.

42

Table 2.4. Depth-specific d'gpersivities and cormpondinggelocities.

 

 

 

Velocity (cm/day).
optimal dispersivity, a
depth (m bgs) (cm) observed simulated
19.5 15.25 57.3 a: 10.7 47.7 :1: 2.0
20.4 15.24 44.9 d: 5.5 40.1 d: 1.5
21.3 33.43 39.0 i 10.5 33.2 :t 3.00
22.3 15.24 49.8 i 17.3 50.1 3: 7.8
23.2 12.19 98.4 at 29.0 99.5 :h 9.8

 

0
Average of multi-level wells at the same depth at uncertainty in the mean
based on the 95% confidence interval as given by the l-distribution.

were used to estimate simulated and observed depth-specific average linear velocity,
respectively. Table 2.4 shows optimal dispersivity values and corresponding velocity
calculated from observed and simulated tracer breakthrough curves.

Based on these results, the average linear velocity at 23.2m is around two times
higher than velocities for other intervals. Similarly, the depth interval with slowest
velocity is 21 .3m bgs. Results from Table 2.4 indicate that the bottom of this aquifer will
serve as a preferential flow path for contaminant migration. This fact can be seen in
Figure 2.12, where the concentration contours in a cross through the augmentation grid
illustrates the position of the tracer plume at 4hrs, 1 day, and 15 days after tracer
injection. From figure 2.l2(c), it is evident that the tracer solution is more distributed in
the high conductivity zones.

Sensitivity with respect to porosity and dispersivity was evaluated by perturbing
these parameters and examining the relative change in the root mean square error
between simulated and observed average linear velocity (Anderson and Woessner 1992).
This analysis revealed that average linear velocity is more sensitive to changes in
porosity than dispersivity. A 10% change in porosity caused approximately a 25%

deviation in average linear velocity when compared to optimal values (Figure 2.13). A

43

3% overall perturbation in the simulated velocity occurred when dispersivity values

changed by 10% from the optimal values.

44

10 “r . . . .
mj cotton/extraction

 

 

 

 

 

 

 

  

 

 

A zone
5° 15 a_ 1 T
5
'5
g 20 2_
flow —’
25 1 1 1 1
l l l 7
0 10 20 30 40
horizontal distance (111)
(a)
10 ‘—
50 15 ——
E
'5
£- 20 _...
25 1 1 4
l l T
0 10 30 40
horizontal distance (111)
(b)
10 1
E? 15 4—
5
5
g- 20 -L
25 % I / 0.2\% 0.1\0.05\%
0 10 20 3O 40
horizontal distance (m)
(C)

Figure 2.12. Tracer concentration (C/Co) contours at (a) 4 hrs, (b) 1 day, and (c) 15 days
after tracer injection. Cross section through augmentation grid.

45

35

 

 

 

 

 

 

30 - u

25 I I
3
2 20 b I
CE

15 - -

10 I- J

5 I l I l l
19 20 21 22 23 24 25
depth (m)
|+opt --G--10%disp c— 10%nl

 

Figure 2.13. Root Mean Square Errors (RMSE) in velocity at each
depth interval for a 10% change in porosity, and 10%
change in dispersivity (deviation from the optimal
values).

2.6 Conclusions

A methodology to investigate the influence of local heterogeneity on the
distribution of solutes in an unconfined aquifer was evaluated. This methodology
coupled laboratory and field tests with optimization techniques for solving the
generalized groundwater flow and transport equations. Hydraulic conductivity results
revealed the existence of high-energy zones that could act as preferential flow pathways

for contaminant migration. This was further confirmed by observations made during

46

tracer experiments.

Optimal depth-specific dispersivity values for these zones were found by solving
the advection-dispersion equation using an inverse modeling approach. This approach
yielded dispersivity values that better represent the distribution of tracer as compared to a
solution scenario where a single dispersivity value was used for the entire model domain.
The values obtained during these simulations were used to estimate depth-specific
average linear velocities. The magnitude of the average linear velocity at the 23.2m bgs
interval was, in general, two times greater than the rest of the zones.

Simulated tracer concentrations were more sensitive to changes in dispersivity
than porosity. One limitation of the methodology developed here is that it does not
consider the length scale dependency of the dispersivity parameter. However, since a
small study area was selected, a single value per depth interval was adequate to describe
the longitudinal dispersivity value in the region. This approach might be adequate in
small aquifer areas where a treatment technology is being evaluated for in-situ
groundwater remediation. Since laboratory experiments for hydraulic conductivities are
tedious, a similar approach can be developed to estimate optimal values for hydraulic
conductivity and dispersivity based solely on tracer concentration histories.

This characterization provides the foundation for understanding the impacts that
the hydrogeologic environment exert upon the distribution of contaminants and
geochemical constituents. This could reveal important information regarding the natural
processes that are responsible for changes in concentrations observed in chemically

polluted environments.

47

2.7 Acknowledgments

Financial support for this study was provided by the Michigan Department of
Environmental Quality under contract Y40386. Partial support was also provided by a
Graduate Assistance in Areas of National Needs (GAANN) fellowship. Special thanks to

Brian Lipinski for providing the regional groundwater flow model.

2.8 Literature cited

Anderson, M. P., and Woessner, W. W. (1992). Applied Groundwater Modeling:
Simulation of flow and advective transport, Academic Press, INC., San Diego,
Ca.

Bear, J. (1979). Hydraulics of Groundwater, McGraw-Hill.

Bower, H., and Rice, R. C. (1976). "A slug test for determining hydraulic conductivity of
unconfined aquifers with completely or partially penetrating wells." Water
Resources Research, 12(3), 423-428.

Cho, .1. S., Wilson, J. T., and Beck, F. P. (2000). "Measuring vertical profiles of hydraulic
conductivity with in situ direct-push methods." Journal of Environmental
Engineering-ASCE, 126(8), 775-777.

Coleman, T., Branch, M. A., and Grace, A. (1999). "Optimization Toolbox for use with
MATLAB: User's Guide." The Math Works, Inc., Natick, Mass.

Dybas, M. J., Barcelona, M., Bezborodnikov, 8., Davies, S., Fomey, L., Heuer, H.,
Kawka, O., Mayotte, T., Sepulveda-Torres, L., Smalla, K., Sneathen, M., Tiedje,
J., Voice, T., Wiggert, D. C., Witt, M. E., and Criddle, C. S. (1998). "Pilot-scale
evaluation of bioaugmentation for in-situ remediation of a carbon tetrachloride
contaminated aquifer." Environmental Science & Technology, 32(22), 3598-3611.

Dybas, M. J., Hyndman, D., Heine, R., Tiedje, J ., Linning, K., Wiggert, D., Voice, T.,
Zhao, X., Dybas, L., and Criddle, C. (2002). "Development, operation, and long-
term performance of a full-scale biocurtain utilizing bioaugmentation."
Environmental Science & Technology, 36(16), 3635-3644.

Fetter, C. W. (2001). Applied Hydrogeolog, Prentice Hall, Upper Saddle River, NJ.

Freeze, R. A., and Cherry, J. A. (1979). Goundwater, Prentice Hall, Inc.

48

Hoard, C. J. (2002). "The influence of detailed aquifer characterization of groundwater
flow and transport models at Schoolcrafi, Michigan," M.S. thesis, Michigan State
University, East Lansing, MI.

Hurt, K. L., Beck, F. P., and Wilson, J. T. (2001). "Implications of subsurface
heterogeneity at a potential monitored natural attenuation site." Ground Water
Monitoring and Remediation, 21(3), 59-63.

Hvorslev, M. J. (1951). "Time lag and soil permeability in groundwater observations."
Bull. 36, U.S. Army Corps of Engineers Waterways Experimental Station,
Vicksburg, Miss.

Hyndman, D., Dybas, M. J ., Fomey, L., Heine, R., Mayotte, T., Phanikumar, M. S.,
Tatara, G., Tiedje, J., Voice, T., Wallace, R., Wiggert, D., Zhao, X., and Criddle,
C. S. (2000a). "Hydraulic characterization and design of a full-scale biocurtain."
Ground Water, 38(3), 462-474.

Hyndman, D. W., and Gorelick, S. M. (1996). "Estimating lithologic and transport
properties in three dimensions using seismic and tracer data." Water Resources
Research, 32(9), 2659-2670.

Hyndman, D. W., Harris, J. M., and Gorelick, S. M. (1994). "Coupled seismic and tracer

test inversion for aquifer property characterization." Water Resources Research,
30(7), 1965-1977.

Hyndman, D. W., Harris, J. M., and Gorelick, S. M. (2000b). "Inferring the relation
between seismic slowness and hydraulic conductivity in heterogeneous aquifers."
Water Resources Research, 36(8), 2121-2132.

Kim, D. J., Kim, J. S., Yun, S. T., and Lee, S. H. (2002). "Determination of longitudinal

dispersivity in an unconfined sandy aquifer." Hydrological Processes, 16, 1955-
1964.

Lee, J. Y., Cheon, J. Y., Lee, K. K., Lee, S. Y., and Lee, M. H. (2001). "Factors affecting
the distribution of hydrocarbon contaminants and hydrogeochemical parameters
in a shallow sand aquifer." Journal of Contaminant Hydrology, 50(1-2), 139-158.

Lewis, M. (2001). "Installing continuous multi-chamber tubing using sonic drilling."
Water Well Journal, 16-17.

Lipinski, B. A. (2002). "Estimating natural attenuation rates for a chlorinated
hydrocarbon plume in a glacio-fiuvial aquifer, Schoolcraft, Michigan," M.S.
thesis, Michigan State University, East Lansing, MI.

49

Mayotte, T. J ., Dybas, M. J ., and Criddle, C. S. (1996). "Bench-scale evaluation of
bioaugmentation to remediate carbon tetrachloride-contaminated aquifer
materials." Ground Water, 34(2), 358-367.

McDonald, M. G., and Harbaugh, A. W. (1988). "A modular three-dimensional finite-
difference groundwater flow model." Book 6, Chapter A], USGS Open-file
Report.

Phanikumar, M. S., Hyndman, D. W., Wiggert, D. C., Dybas, M. J., Witt, M. E., and
Criddle, C. S. (2002). "Simulation of microbial transport and carbon tetrachloride

biodegradation in intennittently-fed aquifer columns." Water Resources Research,
38(4), 1-13.

Wiedemeier, T. H., Swanson, M. A., Moutoux, D. E., Gordon, B. K., Wilson, J. T.,
Wilson, B. H., Kampbell, D. H., Haas, P. E., Miller, R. N., Hansen, J. E., and
Chapelle, F. H. (1998). "Technical protocol for evaluating natural attenuation of
chlorinated solvents in groundwater." EPA/600/R-98/128, USEPA, Cincinnati Oh.

Zheng, C., and Wang, P. P. (1999). "MT3DMS: A modular three-dimensional
multispecies transport model for simulation of advection, dispersion, and
chemical reactions of contaminants in groundwater systems." U.S. Army Corps of
Engineers, Washington, DC.

50

CHAPTER 3

EVALUATION OF NATURAL ATTENUATION IN A CONTROL
VOLUME OF A VOC CONTAMINATED AQUIFER

3.1 Abstract

Natural attenuation was evaluated in a study region of an aquifer located
approximately in the center of mass of VOC contaminant plume. The approach
undertaken in this chapter provides a methodology for evaluating natural attenuation of
chlorinated solvents when both parent and daughter products of dechlorination are
identified in the contaminant source. This study includes solid and liquid phase data
collected over a three year period. Generally, studies of natural attenuation of chlorinated
solvents ofien emphasize the liquid phase concentration, which downplays the role of
sediment associated constants. In this study, TCE concentrations in the solid phase up to
1,200ug/kg were detected. This value is twice the concentration found in the liquid
phase for the same sampling event. Levels of cis-DCE were similar for solid and liquid
phase concentrations (on a per Kg basis).

The co-existence of high levels of TCE and cis-DCE, with only low
concentrations of the parent PCE, indicates that reductive dechlorination has occurred to
some extent at some point in time. However, VC concentrations were very low
compared to its parent products, and coupled with the accumulation of cis-DCE, indicates
incomplete reductive dechlorination of chlorinated parent products.

Low nitrate concentrations in the deepest zones of this aquifer along with elevated

chloride concentration suggest that dechlorination of VOC has occurred in the anaerobic

51

zones of this aquifer. However this is not strong evidence to support the hypothesis that

dechlorination processes are still taking place within the study region.

3.2 Introduction

Over the past few years there has been an increased reliance on natural processes
to clean-up contaminated sites (NRC 2000). Natural attenuation include physical,
chemical, and biological processes that, under favorable circumstances, act upon
contaminants to reduce its mass, toxicity, mobility, volume, or concentration (EPA 1998).
Although natural attenuation has been implemented at a significant number of sites
contaminated with a broad range of organic compounds, concern has been raised in the
scientific community because the behavior of most of these compounds in the
environment is not fully understood.

Natural attenuation has been successfully demonstrated at sites contaminated with
BTEX compounds. Suarez and Rifai (2002) analyzed historical data for a benzene plume
at an industrial facility and found a constant decrease in extent and concentration of
contaminants with time. These results were further confirmed by a geochemical
characterization of the site. F razmann et al. (2002) demonstrated toluene degradation
under sulfate reducing conditions in a coastal plain in Western Australia. However,
degradation of benzene within the plume was not significant. Extensive natural
attenuation of BTEX compounds was observed at a Coast Guard base facility in Traverse
City, MI (NRC 2000). In this aquifer, the presence of breakdown products of BTEX
degradation along with oxygen depleted zones followed by methane and rich Fe(II) zones

where indicative footprints of aerobic and anaerobic decomposition of BTEX

52

compounds.

Due to their wide spread use during the last few decades perchloroethene (PCE),
trichloroethene (TCE), and trichloroethane (1,1,1 TCA) are among the most common
contaminants found in groundwater. Attenuation of these compounds presents a major
challenge since they are highly recalcitrant in the environment. Although the
biodegradability of these compounds have been demonstrated in laboratory (Ferguson
and Pietari 2000; Kao and Wang 2001; Maymo-Gatell et al. 1997; Ndon et al. 2000;
Vogel et al. 1987; Vogel and McCarty 1987; Witt et al. 1999; Yang and McCarty 1998),
attenuation of these compounds in natural environments is mostly attributed to physical
mechanisms such as dilution, dispersion, and immobilization. However, sequential
biodegradation of these compounds has been observed at some contaminated sites.

Weaver et al. (1996) observed sequential dechlorination of TCE at the St. Joseph,
MI, Superfund site. Presence of methanogenic zones evidenced the strongly reducing
conditions suitable for sequential dechlorination of TCE to ethene. Data from individual
boreholes at the site confirmed that high cis-DCE concentrations correlated with declines
in oxygen and sulfate concentrations. Also, high concentrations of ethene and vinyl
chloride where found in the most methanogenic zones of this aquifer.

At Area 6, Dover Air Force Base, metabolic byproducts of the degradation of
PCE/TCE, and oxygen depleted zones with elevated methane and hydrogen concentration
evidenced that natural attenuation processes are acting to reduce the contaminants at this
site (Davis et al. 2002; Witt et al. 2002). Similar observations were made at a Superfund
hazardous waste site in Louisiana (Clement et al. 2002) where the “lines of evidence”

approach suggested by the United States Environmental Protection Agency (EPA 1998)

53

was employed to demonstrate that natural attenuation processes were responsible for the
degradation of PCB and TCE.

In all these cases, the presence of electron donors necessary for the reductive
dechlorination process is readily available either as co-contaminants to the chlorinated
organics or as naturally occurring organic matter. However, partial dechlorination is
frequently encountered due to deficiency or depletion of electron donors (McCarty et al.
1998; Richmond et al. 2001). This deficiency translates into accumulation of metabolic
by-products such as cis-DCE or VC that might be more toxic or are regulated at lower
concentrations than their parent compound.

To evaluate the extent to which natural attenuation is occurring at a particular site,
several protocols have been developed (NRC 2000). A “lines of evidence” approach is
the basis of most of these protocols where evidence of natural attenuation can be
provided by: (1) documented loss of contaminants from the site; (2) evidence that the
biodegradation potential is actually realized in the field; and (3) laboratory assays
showing that microorganisms from site have the potential to transform the contaminants
(Roling and van Verseveld 2002). A “scoring system” assigns a numeric value based on
the evidence collected and that number is used to reach conclusions about the natural
attenuation potential for a particular site. This approach has been criticized for assigning
a numerical value to a series of qualitative assessments; an assignment which implies
more confidence thanjustified by field evidence (NRC 2000).

The objective of this study is to analyze geochemical and VOC concentration data
in a control volume located within the contaminant plume to determine if natural

attenuation is occurring in this particular area. This might be a useful approach to

54

evaluate natural attenuation of chlorinated organic compounds, especially when both,
parent and daughter products are found in the contaminant source area. Clement et al.
(2002) indicate that the scoring system proposed by the EPA (1998) might result in
underestimation of overall attenuation because metabolic by-products from the
degradation of parent chlorinated compounds have to be excluded from the evaluation if
they are present in the source area. Analysis of data in a control volume might be useful
in resolving this conflict.

Also, a detailed hydraulic characterization of the control volume will be used to
investigate the influence that variations in geophysical parameters may have upon natural
attenuation of these compounds. A control volume approach will be compared to the
scoring system used by the EPA to derive conclusions regarding the natural attenuation

processes in this area.

3.3 Site description
Site history

The village of Schoolcraft is a small rural community located in Kalamazoo, MI
(Figure 3.1). Several plumes of organic and metal contaminants were discovered in the
unconfined aquifer beneath the village approximately two decades ago by the Michigan
Department of Environmental Quality (MDEQ). The most extensive of them is a VOC
plume, designated by MDEQ as Plume G. It has been estimated that this plume has
impacted about 1.3x107m3 of aquifer material. ARCO Industries, a former manufacturer

of automobile plastic parts, was identified as the source of this contamination. A detailed

55

6:53 EmEESEo 00> 05 mo 25:8 05 9:30am owfl=> 080335 ._.m “:de

 

  

$952120. 35m
cool man . .0

6v 0:5: 00>

     

 

 

 

 

 

 

 

 

56

investigation conducted in 1986 on the ARCO premises identified five major areas where
usage and disposal of chlorinated compounds occurred throughout the history of the
facility (Figure 3.2). Table 3.1 shows the major chlorinated compounds and the
maximum concentration found in soil samples collected at different locations over the
property during the 1986 site investigation. In addition to the compounds listed in Table
3.1, toluene was found in several sediment samples ranging from non-detectable levels to
a maximum concentration of 115,000ug/kg. Although there were no historical record of
toluene usage in the industrial operations at ARCO, one of the chemical providers
admitted that one supplied product was, at least, contaminated with toluene.

Water and bottom sediment samples from the wastewater disposal pond (Figure
3.2) showed low level concentrations of xylene, ethylbenzene, and chlorobenzene in
addition to VOC. However, VOC concentration from these sediments and from the
pond’s water was several orders of magnitude lower than concentrations found at other
locations.

The combined effects of the five areas shown in Figure 3.2 resulted in a VOC
contaminant plume extending about 2km southeast from the ARCO facility (Figure 3.1).

Near the center of mass, the plume is approximately 400m wide.

Table 3.1. Chlorinated compounds and concentrations found in
sediment samples at ARCO facilities.

 

 

Compound concentration, ug/kg
PCE 300,000
TCE 280,000
C15 '1' trans-DCE 520 0008
VC N.A.b

 

a . .
estimated based on percent recovery from a pilot vapor
extraction system.

not available in historical reports

57

 

 

 

 

 

 

 

H DUNCAN ST. H

 

 

 

 

 
 
 

wastewater
disposal pond

LEE ST.

ELIZA ST.

1.

O 50

'meters

 

Legend

1r areas with VOC
contamination

 

 

 

 

Figure 3.2. Location of the major source areas of contaminants identified during the

58

1986 investigation (modified form Lipinsky, 2002).

 

Hydrogeological characterization

The unconfined aquifer beneath the Schoolcraft Village is mostly composed of
highly stratified glacial outwash deposits (Lipinski 2002; Mayotte et al. 1996). The
surficial layer overlaying this outwash sand is a dark brown, silty sandy clay that extends
from the ground surface to about 0.6 to 1.2m below ground surface (bgs). Underlying the
outwash sediments, a light gray clay deposit was found at depths ranging from 21.3 to
30.5m bgs. It is believed that this layer prevents further contaminant migration to the
underlying confined sandy aquifer.

The water table in the unconfined aquifer is, on the average, at 4.6m bgs. General
groundwater flow direction is towards the southeast at an average linear velocity of
15cm/day (Dybas et al. 2002; Mayotte et al. 1996). Hydrogeological features that control
the regional groundwater flow are shown in Figure 3.3. Recharge for the entire area is
approximately 23.7cm/yr (Lipinski 2002). It is believed that upon further migration, this
plume will eventually discharge at the surface water bodies located southeast of the

Schoolcraft region.

Characterization of the control volume for evaluating natural attenuation

An area located approximately in the center of mass of a VOC contaminated
plume was selected to study the natural attenuation processes using a control volume
approach. Figure 3.4 shows the control area with all the monitoring wells installed for
this study. The approximate dimensions for this area are 30.5m x 30.5m x 24.4m
(LxWxD). Groundwater flow direction is approximately perpendicular to the northwest

and southeast boundaries and parallel to the other two sides.

59

 

 

i A
N

 

 

 

 

Y

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

J? l Sunset Lake
_ l - A
Shug 0% E [D]
. E. , l -- _
Schoolcraft | Portage Creek
_ g I l i
- I - __ }
Flowerfield Creek l l
\ Eli- 1
1 7 '
, - l,— P
m7? PVW— ”7 . ——-—---—- _ - ‘5-
. ~—1;—-
- Lakes V\ Rivers/Streams l 1
. . , o 2
m Wetlands w Dram Spring Creek T
Rggg Moraine a

 

 

 

Figure 3.3. Regional hydrologic boundaries that control groundwater flow in the
Schoolcraft area (modified from Lipinski, 2002).

60

 

 

 

\
\\\SB-3
9} \ ~1
5 G \\\ 59
Q 4"
\
5 ® ‘5 ‘3 83-6 Q ~
A 6* 9‘ r3" 9
\\:w\z 3'! $9? 5}\Q Q
513 2 \\ 39 G‘\ $9 raw-9
\\\ 5? a? éflfl 6 5:9
\\ é § 6‘8 §\§\
\\ we 9} 9‘ \\
B \ \. § 5 \ Cc
‘ 3- \SB-S
\\ $7 \\
‘Q .9 \
\ 3 \
\\ 9 Si" \\
W" G \ MW-l
SB-l \ ® a p S \‘ .
V" ‘ °‘ 5 ‘9 3 ‘\ 53.:
9 a! f‘ G 5 9 o “s .
o ‘ \G G 3‘,» $3. 49' s A
.v ‘6 ,5,
3 W4 \ G 5: x‘
0 5 10 53.4 3 ¢ m 3 ,-
E 9 9' 9‘ m 5‘69 9 w l
SCALE IN METERS 4°" 5” 09‘ \Q 95‘ G MONITORING POINT l
O
,9" ‘K @FLUX CONTROL WELL
0 5’ W"‘\
W.,,” S“ ‘3. C SOIL BORING :
DELIVERY WELL l
—CONTROL VOLUME BOUNDARY l ‘

 

 

 

 

 

Figure 3.4. Control area with monitoring wells for the natural attenuation study.

All monitoring wells consists Of 5cm internal diameter PVC pipes with 8 or 9
multi-level ports attached at different locations for depth specific sampling. Details of
the monitoring wells with the multi-level ports are illustrated in Figure 3.5.

A hydraulic characterization of the control volume using tracer tests coupled to
numerical models and Optimization techniques were conducted to estimate relevant flow
and transport parameters. A summary of the depth-specific physical parameters found in
this study is given in Table 3.2. Total porosity based on the bulk density of the sediments

in this region ranged from 0.21 to 0.50 With a mean value of 0.33 i 0.006 (95%

61

confidence interval). A cross-section through the site (A-A’ in Figure 3.4) is shown in
Figure 3.6. Two preferential flow path regions were identified from the soil cores

extracted from the site. These zones are located on the average at 20m, and 23m bgs.

Table 3.2. Depth-specific summary of ghfiical parameters.

 

 

hydraulic average linear optimal solute
depth conductivity, velocity, v dispersivity, retention
(m bgs) K (*10'2) (cm/day)a a time
(cm/”a (cm) (l/day)
19.5 8.1 at. 3.4 57.3 d: 10.7 15.3 53.2
20.4 11.6 :1: 4.4 44.9 d: 5.5 15.2 67.9
21.3 10.7 :t 4.0 39.0 :1: 10.5 33.4 ' 78.1
22.3 9.0 :t 2.9 49.8 :1: 17.3 l5.2 61.2
23.2 16.6 :t 4.9 98.4 :t 29.0 12.2 31.0

 

8 . .
mean of samples at the same depth :1: uncertainty in the mean at the 95%
confidence interval given by a t-distribution.

62

5 cm id PVC well

non-reactive
peristaltic pump
tubing

   

 

 

__ \l
1... ~———— —- — -8 or 9 @ 0.9 m each- -9 ~ 9 2,1

 

 

0.61 m screen length

 

 

:1

Figure 3.5. Details of the multi-level wells installed for the natural attenuation
study.

63

 

 

 

 

MW—8

 

 

 

MW—S

2

 

 

 

 

 

 

 

      
  

l

 

.K‘
II

 

- 2..“ - ‘m-z-‘U _

. nun-1 -mn“nn—-.-:-—- - -.n- ”4" gr- 1-1. -JI-T.’ on 04‘ ':m u n.-

  

I-‘-.jl'I-Ifl-‘I-III-llll-

t-..-1--j-

 

- ‘.I- .- ul- -‘x‘t‘ m-tnn - mmunv n.- -1- u - _ L - - -2- - o n n -
‘ - .mzn .11». I‘I'j.- I I

I.

   
  

 

 

I urtnLt n I. n I‘n-.:_ I I

; um- cum-

I-ul

 

 

i5,
5,.
3%
0d
53
Eu
‘0 a
Ea
<1
wé
Do
”A
52.4
52$

3
9
E
<
m
Q
m
3
L?
._1
.4
m
3
E

9
LL]
(D
m
u.)

 

§_CLAY(CL)

 

 

 

 

64

Figure 3.6. Cross section A-A’ (Figure 3.4) showing the location of two preferential flow pathways.

3.4 Materials and methods
Samples for geochemical constituents and organic compounds were collected in
four sampling events over a three year period. The logistics for the sampling event and

the matrix sampled in each one is given in Table 3.3.

Aquifer sediment sampling

Aquifer sediment samples were obtained during drilling of the monitoring wells.
Cores were extracted from each borehole using the Waterloo cohesionless continuous
sand sampler (Dybas et a1. 1998) from 15.5m to 25.0m bgs in 1.5m intervals or the
Rotosonic method, where cores where extracted in 6m intervals. Solid samples from
both core types were collected by inserting a syringe-type sampler to extract
approximately 2 to 4cm3 of sediment. These samples were place in 40mL headspace
vials containing 10mL of 2% (w/v) NaHSO4 solution and were sealed with Teflon-lined

septa. Samples were transported in ice to the laboratory for volatile organic analyses.

Groundwater sampling

Groundwater samples for organic and inorganic analyses were collected from
each port attached to the multi-level wells afier purging three well casing volumes.
Groundwater was extracted using peristaltic pumps at a rate of 200mL/min. For VOC
analyses, samples were collected in 40mL VOA vials containing lmL of 40% (w/v)
NaI-ISO4 solution and were sealed with Teflon-lined septa. Samples for inorganic
analyses were collected in 15mL tubes (0.45 pm filtration) with nitric acid (1% final

concentration) preservative in the case of soluble metals.

65

Table 3.3. Description of the sampling events in the control volume for the three year study.

 

Matrix sampled

Description

 

Sampling
Event Date Wells Sampled a
Fall 2000 MW-l — MW-9
Spring 2001 MW-l — MW-9
some MDEQ wells
Spring 2002 MW-l — MW-9
MP-Al — MP-A l 5
MP-Sl — MP-S15
FCW’s and DW’s
Summer 2002 MP-Al — MP—AlS

MP-Sl —MP-Sl$

a see Figure 3.4 for well location

aquifer sediments

Groundwater
groundwater

Groundwater
aquifer sediments
& groundwater
aquifer sediments
& groundwater
aquifer sediments

groundwater
groundwater

Volatile organic and inorganic analyses

Solid phase samples taken during
installation of monitoring wells MW-l
through MW—9

Liquid phase samples from the wells
installed during Fall 2000 and some
wells outside the control volume
installed by the Michigan Department of
Environmental Quality in 1986.

Liquid phase samples from the installed
MW wells and from the recently
installed multi-level wells (MP-A’s and
MP-S’s). Solid phase samples from the
cores extracted during MP-A’s, MP-S’s,
FCW’s and DW installation.

Liquid phase samples from the multi-
level wells installed during the Winter
2002.

Volatile organic compounds were analyzed using a Tekmar Precept II headspace

auto sampler interfaced with a sampler and concentrator (Teckmar-Dohrmann 3100) and

an Agilent 6890 gas chromatograph system.

A Dionex model 2000i/SP ion chromatograph with suppressed conductivity

detection, equipped with a Dionex model AS4-A column was used for anion analyses.

The mobile phase was a solution of 1.8mM bicarbonate and l.7mM carbonate at

3mL/min. Dissolved oxygen, temperature, pH, and oxidation-reduction potential were

measured in the field with a Purge-Saver Model F C2000 flow-through cell. Details for

these analyses can be found in Dybas et al. (1998).

66

Hydrogen Analysis

Hydrogen gas was measured using the bubble strip method (Chapelle et al. 1997;
Lovley et al. 1994). Groundwater was continuously pumped through a gas sampling bulb
containing a nitrogen or air “bubble” so that hydrogen can partition between the gas and
liquid phases. When an equilibrium between the dissolved and air concentration of
hydrogen was reached, a sample of the air bubble was collected and analyzed for H2 with
RCA-3, a reduced gas analyzer (Trace Analytical, Inc.). The dissolved H2 concentration

was found using the Ideal Gas Law and Henry’s Law.

3.5 Results and discussion

Electron acceptor profiles: correlation between geochemical and reductive
dechlorination indicators

Depth specific concentration of chlorinated ethene compounds in aquifer
sediments during the course of this study can be seen in Figures 3.7 and 3.8. Solid phase
samples were collected during drilling activities of MW wells surrounding the multi-level
samplers network (Figure 3.4). VOC concentration profiles in the solid phase showed
highest contamination approximately between 17.4 to 24.4m bgs. TCE concentrations as
high as 1,200ug/kg (MW-6) were found in sediments extracted from this region. Also, a
maximum cis-DCE concentration of 800pg/kg was found at 24.1m bgs in sediments from
MW-8. Vinyl chloride concentrations in the sediments were low compared to
concentrations of TCE and cis-DCE. VC concentrations were, on the average, two orders
of magnitude lower than its parent products.

The general patterns of chlorinated ethene concentrations in the liquid phase

67

 

 

 

 

 

 

 

 

r ' Y
‘ d
l l

:4»
20 '"'~ 4 d
o ..............
0% ------ .. .......... . . ..
. ‘9"‘ \‘~~;::;££‘G: ‘. ‘ -
”U . - 1: J a 0

 

 

25 a a
150 200 250 0 100 200 300 400 500

 

 

 

 

 

 

 

 

 

 

 

MW-S MW-6
d
.. ‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘‘ . . 'f.’.'.'.'.'.':.'_'::::::.'.,:,—.-.--. ---------- O
.0 -------- O -------- 0 ......... 01 . ------------------ .0 l
y,“
200 300 400 "0‘ ‘ ' ‘ l l
400 800 1200
MW-8 MW-9
I r ' r I ' ' l1
l a
* "r ------------ .
3:19.322. C o """ O 1 .. ‘
O
. L ""2 0". """ "°

 

 

 

 

 

 

Figure 3.7.

 

 

25 . . '
600 800 1000 o 50 100 150 200 250 300

Chlorinated ethene concentration in aquifer sediments (Fall 2000 sampling
event); TCE(o), cis-DCE(o), and VC(o). Vertical axis represents the depth
in meters below the ground surface at which the sample was collected.
Horizontal axis is the concentration in pg/kg.

68

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MW-l MW-Z MW-3
5 7 # I V ‘ ' 5 1 U V 1 r 5 1 U V I I U V
10 ‘ ......... '1 10 ’ ‘ 10 f] 0. o 4
Zn “““ .20 t;- q u [,0
"‘5 ,3. ll 93"
15 .0 “.9. -l 15:” _________ .111: d 15 F ‘1‘)“ q
‘T? ‘1‘ l, """""""" 9" """" o in “0T0.
o o ‘ ti o 0‘ L to
20 '1 Q ‘a‘ . 20 . b ,b . 20 :1 ‘b, .
a a l1 ‘199’ : 60
"go 0," : ,0 ' “fro 53 ..--‘ ‘0“
25 lo, I. l l ‘p 25 [-7] L 41 ,’ Lo"4l 25 ' l l _l S“‘4 1 J
0 100 200 300 400 500 600 0 50 100 ISO 200 250 300 0 100 200 300 400
MW-4 MW-S MW-6
5 I Y fi 1 5 I U ' I I I 5 I f I 1 r
10 F; O .‘ -I 10 5 ‘ G._ -I 10 is] O. T
l o "fa I ‘q‘ ‘0\ I 0'. 9'
¢ cycj' h ,0 [D b .9.
15-1 6? 3o - '5; _¢ 93' ' 15'» “9.? l
m O‘O'~~ o... o ............ ‘9‘
'. ‘o. “o F "‘0, """"""" 9 d ‘9
20 3,: ‘9 j. . 20 o o, as. 20 ogj .
l C! l "h 0 l ‘9‘
m _00 F" 9 . °‘~~.
25 ' 1 a L L. L 25 l 1 .IL 1 4 a a “l 25 ‘ a a J. a ‘ a _a 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
MW-7 MW-8 MW-9
5 I I ‘7 I' I 5 U I V I 5 T I r V I ‘ 1
10 . . 10 ~ - 10 - 9 9 q
E1 _.o lo {a a. o
I“ . .0" I". '5: O .2 . 15 1? .h‘o’D q
ll- 02' ....... .~ Ill 9 l‘ a!"
15 i. """"" o‘- :0 " “(5‘0
m 9 0.’ 20m. 1» . 20 “a l
l ’9‘ ‘.‘ p o -..:
20 :7 d. p. O at\ 0' o .......... '0
l 0 ’. 25m; “o l 25 ‘o "o l
m 0.}
25 J 4 A a a 0 30 a a a a a 30 L a a a a a n
0 50 100 150 200 250 300 0 50 100 150 200 250 300 0 100 200 300 400
Figure 3.8. Chlorinated ethene compounds in groundwater samples (Spring 2001

sampling event); TCE(o), cis-DCE(o), and VC(o). Vertical axis represents
the depth in meters below the ground surface at which the sample was
collected. Horizontal axis is the concentration in pg/L.

69

(Figure 3.8) were similar to the solid-phase concentration profiles on Figure 3.7.
Generally, concentrations of TCE ranged from 100 to 300pg/L in most of the wells
except for MW-l where liquid phase concentrations as high as 500pg/L were found. Cis-
DCE concentrations ranged from 50 to 350ug/L. A positive correlation between cis-DCE
and TCE concentrations is apparent in some of the profiles of Figure 3.8. Vinyl chloride
concentrations in the area were low compared to TCE and cis-DCE.

Depth specific concentration of geochemical parameters (Figure 3.9) shows an
apparent decrease in nitrate concentrations with depth for some of the wells. The most
noticeable concentration change occurs in MW-l where nitrate drops from lOOmg/L at
10m bgs to a concentration of 20mg/L at 23m bgs. However, no trend in nitrate
concentration data was observed in the profiles of MW-4, MW-6, MW-8, and MW-9.
Sulfate concentrations were between 40 and 80mg/L and although a slight increase in
concentration with respect to depth was observed in the wells, generally, the data shows
no significant depth related trend for this parameter.

Even though chloride is both a metabolic byproduct of dechlorination processes
and a geochemical parameter, it is included in these plots because increases in its
concentration could be indicative of microbially mediated dechlorination processes.
Chloride concentration ranged from 20 to 85mg/L with no appreciable concentration
trends. For most of the wells, chloride concentration was in the range of 60 to 80 mg/L.

Other groundwater parameters measured, such as dissolved oxygen, specific
conductivity, pH, and temperature are given in Table 3.4. Dissolved oxygen
concentration varies from 1.6mg/L at 9.1m bgs to 0.78mg/L at 24.4m bgs. This suggests

that aerobic zones overlay anaerobic zones in this aquifer. The source of oxygen

70

 

 

 

 

 

 

 

 

 

 

 

 

MW-1
5 ' ' j v '
10 b q
n\\o“ H.
:33: .
FC’ 0;- 5,
l5 . \ng \ J. ‘
t -‘fi
1' 9' \\
20' '1 dz. .
a" e [I
’." ..Q' :1
25 l. L A O L 4 \
0 20 40 60 80 100 120
MW_4
5 v ' ' ‘
10 _ II
C 0‘ ’0
' ,3”
O q,.[1/
15 . \p \9 .
.‘ d‘ l
'1 xv '
20 P 'I. ‘Q.’ .
4.. _- ,5
o... l‘
25 a O L i) ...... p
0 20 40 60 80 100
MW_7
5 ' I ' I
10 ‘
9° '0
O ’0‘ ’
91‘. BI
15 l . flea ) d
'9’ o, d\
‘0“ o \\ \
20 h :3 ’0 ‘7: II
"“ 0:: ‘‘‘‘‘ <\
,Io _‘ “has
25 .‘4 4 A 0'. L I
0 20 40 60 30 100

 

 

 

 

 

 

 

 

 

 

 

 

MW_2
5 ‘ , v '
10 _ 1
.° 2“ ’
‘5 x’, .a’
15 . 85.1»: fl
Q'” .‘o‘ \‘a
P 0‘ \\
20 . 9' ‘Q‘ I? ‘
.‘X
. .xo
25 . b . .04} .
0 20 40 6O 80 [00
MW-5
5 I I I ‘ ' ' '
10 _ .
° ,0 .
o I --------- .
,_.si’.-...-..-.’.,‘3
15 . w: I: q
.‘G:::--._.:.:’ \B“
I. ---------------- ::.-o‘::
20 .9, G ...... a, I 1
.‘ :0 \)
l! a Q]
25 6. . . . ‘3 I
0 10 20 30 40 50 60 70 30
MW-8
5 . . ' ' ' 1 I
10 . .
2°, 0
15 I x \\ .
r . ,3
20 - a. C(\ ‘
v i 5):)
25 l l l 1 L' a a" ‘l
0 10 20 30 40 50 60 70 80

 

 

 

 

 

 

 

 

 

 

 

 

10,, q
‘-° 9
.5 ,
‘49. r‘,
15. .3" ‘o \\ 4
H" \Q‘ \a
1". ‘0‘:
20.. 1., k
’ .9
f if“?
25 g . . 9‘
0 20 4O 60 80 100
MW_6
5 v I t , t '
10.. J
0‘ Q 0‘
:0 ['0 m
.‘l d, d
15 r x q
20. . -u‘ x 4
o o [m
0'." a
25 . . . ‘ Hg I
o 10 20 30 40 50 60 70
MW-9
5 . fl, I . ' t
10, ‘
.0 '0 ’0
oi. oz: K:
in 220 ”E
15 .. (1 o "I ‘
\
o ,2:
.36 ’,/’
20’ .5: 0'“. 5’.“ .
.-' H .20 “\‘\
0.3-"‘3 D
25 ‘ ,'; 1"‘03. a 1“ a
010 20 30 40 50 60 7o 80

Figure 3.9. Geochemical constituents concentration in groundwater samples (Spring
2001 sampling event); NO3'(o), SO42'(o), and Cl'(o). Vertical axis
represents the depth in meters below the ground surface at which the sample
was collected. Horizontal axis is the concentration in mg/L.

71

Table 3.4. Flow through cell parameters.

 

 

specific dissolved

depth conductivity oxygen temperature
(m bgs) (mS/cm) (mg/L). (°C) pH
9.1 0.989 t 0.032 1.62 :t 0.40 11.4 7.2
12.2 0.991 :h 0.028 1.59 :1: 0.40 11.4 7.2
15.2 0.993 :t 0.037 1.52 t 0.42 11.4 7.2
18.3 0.999 at 0.042 1.33 :1: 0.42 11.3 7.2
21.3 1.01] i 0.045 1.08 :1: 0.38 11.3 7.2
24.4 1.031 :t 0.052 0.78 :h 0.51 11.3 7.2

 

0
average 3: uncertainty in the average at the 95% confidence interval

in the shallower portion of this aquifer is believed to be recharge from precipitation or
snow melt. The pH for this water was 7.4 on the average and the temperature was
11.4°C.

Comparison of solid and liquid phase concentration data (Figure 3.7 and 3.8)
show a similar pattern in the profiles for the VOC concentrations. Generally,
concentrations are lower in the shallower portion of the aquifer and increase with depth.
Solid phase concentrations in MW-S show that cis-DCE is higher than TCE
concentrations at 25m bgs. Comparing this panel with the liquid phase concentration for
the same well in Figure 3.8 shows that cis-DCE is also higher that TCE. However, this
relation was not observed in all wells. For example, the solid-phase TCE profile for
MW-8 shows lower TCE concentrations at 25m bgs than cis-DCE. On the other hand,
liquid phase TCE is higher that cis-DCE concentration for the same well at the same
depth interval. The reason for this inconsistency is unknown at this moment. This type
of comparison could be useful in identifying inconsistencies in the analytical data,
collection procedures, and reformulation of conceptual views.

Even though solid and liquid phase samples were not collected at exactly the

same time, this data set can give an insight as to the concentration in the liquid and solid

72

phases since this data was collected within a 4 month time span and the migration of the
contaminants in the zone are not expected to be significant within this time period.

A comparison between Figures 3.8 and 3.9 was made to find patterns indicative of
microbial reductive dechlorination processes. The geochemical concentration profile of
MW-l show that nitrate concentration decrease with depth while chloride concentration
increases. However, the liquid-phase TCE data for the same well is higher in the deepest
portion of the aquifer than cis-DCE. Thus, the increase in chloride concentration cannot
be attributed to reductive dechlorination of TCE to cis-DCE. The geochemical profiles in
Figure 3.9 and the VOC concentration profiles (Figure 3.8) do not confirm that reductive
dechlorination is a significant component of the natural attenuation process at this

location.

Effects of hydrogeology on contaminant distribution and transport

To evaluate hydrogeological effects on transport and distribution of constituents,
data was collected at several depth specific wells located in cross section lB-B’ (Figure
3.4). Color-coded images of the interpolated data along with observed values are shown
in Figures 3.10 through 3.12

The interpolated image for solid-phase VOC concentrations is shown in Figure
3.10. Also, the location where data was collected is indicated by color-coded filled
circles. TCE concentrations in the cross section ranged from undetectable levels to
600ug/kg. Soil concentration of cis-DCE was as high as TCE. Maximum areas of cis-
DCE concentration are around 300pg/kg (Figure 3.10 b). However, sorbed VC

concentration was low compared to TCE and cis-DCE. VC concentration in the solid-

73

TCE (ug/Kg)
,H 600

 

ED ‘ 450
5 .
5 l 300
8'
'0 i 150
L. ‘ V‘ ‘__> 0
5 10 15 20 25 30
distance (m)
(a)
BONUS/Kg)
n 300
go 41- 225
5 l
.5 1 150
8'
'u

 

5 10 15 20 25 30

distance (in)
(b)
-1 6 YC(:15g/Kg)
3:; n 12
g -20 9
-§' 6

    

5 10 15 20 25 30
distance (m)

(C)

Figure 3.10. Solid phase chlorinated ethene concentration (pg/kg) in B-B’ cross section
(spring 2002 sampling event). (a) TCE, (b) cis-DCE, and (c) VC.

74

TCE (ug/L)

 

1000
2"?)
.n 750
5
33 500
'c 250

 

5 10 15 20 25 30
distance (m)

(a)

 

’32

.D

E

a

3

'U

5 10 15 20 25 30
distance (m)
(b)

VC(us/L)
Fl 3°

En Prim 22

5 : .

g . 15

% 7

 

5 10 1 5 20 25 30
distance (m)

(C)

Figure 3.11. Liquid phase chlorinated ethene concentration (pg/L) in B-B’ cross section
(spring 2002 sampling event). (a) TCE, (b) cis-DCE, and (0) VC.

75

phase ranged from non-detected levels to 15 ug/kg.

Liquid phase concentration images of VOC compounds show a more uniform
distribution than the solid phase concentration (Figure 3.11). Maximum concentrations
for TCE were approximately 700ug/L while for cis-DCE, the maximum concentrations
were around lOOOug/L. Vinyl chloride concentrations were within 0 and 20ug/L.

The geochemical parameters where also interpolated in the B-B’ cross section.
Figure 3.12 show interpolated images for nitrate, sulfate, and chloride. Nitrate
concentration seems to be distributed between 10 and 30mg/L. Sulfate and chloride
images show maximum concentrations in this cross section of 130 and 100mg/L,
respectively.

Comparing TCE and cis-DCE solid-phase concentrations in cross section B-B’
(Figure 3.10 (a) and (b)) with the geologic cross section of the site (Figure 3.6) show that
higher concentration areas are located in areas that were identified as preferential flow
pathways. VC concentrations, however, do not show a clear trend.

Liquid phase chlorinated ethene concentration (Figure 3.11) shows that for TCE,
the maximum concentration contours apparently do not fall within the preferential flow
regions identified in Figure 3.6. However, higher vinyl chloride areas are within this
region. It could be possible that in these high conductivity zones, equilibrium between
the solid and liquid phase concentration have not been reached. Therefore,
concentrations of compounds that have a tendency for sorption to the solid phase will be
higher in these high conductivity zones than in zones where quiescent water provides
more retention time to reach that equilibrium point. A sorption experiment using a

layered aquifer system could be helpful in the interpretation of these observations.

76

N03 (mg/L)
” so

  

depth (m bgs)

 

5 10 15 20 25 30

distance (m)
(a)

S04(Ing/L)
H 150

’8’? . 120

'0 .

é ‘ 90

a r' .0

'U

 

fl”
0

5 10 15 20 25 30
distance (m)
(b)
Cl(mg/L)
-16 H 100
75
. 50

depth (m bgs)
“a:

 

u 25
—24 1 , \ 0
5 10 15 20 25 30
distance (m)
(C)

Figure 3.12. Geochemical parameter concentrations (mg/L) in B-B’ cross section (spring
2002 sampling event). (a) NOg', (b) 8042', (c) Cl'.

77

The geochemical parameter images in Figure 3.12 do not show a significant trend
that could be correlated with the geologic information of Figure 3.6. Also, comparison of
the geochemical parameters with liquid phase chlorinated ethene concentrations do not
show a significant relation indicative of microbial reductive dechlorination in this cross
section. These observations confirm the results obtained for the depth specific data
sampling.

Methane concentrations were measured in selected wells in the area however, no
quantifiable levels of methane were found. Also, hydrogen concentration measurements
were all below detection limits indicating that major microbial processes such as iron
reduction, sulfate reduction, or manganese reduction, if occurring, are at low rates
balanced with H2 consumption. This supports the hypothesis that natural attenuation of
these compounds in this region can be attributed only to physical processes such as

dilution and dispersion.

Evaluation of natural attenuation by EPA (I 998) guidelines

An analysis of geochemical parameters and VOC compounds based on the EPA
(1998) protocol show that some of these wells will score between 6 to 14 points in the
scoring system proposed by the guidelines. These values indicate, “limited evidence for
anaerobic biodegradation of chlorinated organics” in the control volume area. However,
based on the geochemical and volatile organic analysis, there is no evidence that
reductive dechlorination is occurring in this region. Table 3.5 shows the score obtained

by analyzing geochemical and VOC data collected in a selected well at 20.4m bgs.

78

Table 3.5. EPA (1998) screening process applied to a
selected aquifer interval (MP-A3 at 20.4»:

 

 

 

b ).
Analyte concentration points awarded
dissolved oxygen 1.08 mg/L -2
Nitrate 11.8 mg/L 0
Iron (11) ND” 0
Sulfate 80.8 mg/L 0
ORP > 100 mvb o
Chloride 121.3 ug/L 2
TCE 286.5 ug/L 0
cis-DCE 163.3 ug/L 2
VC 2.4 ug/L 2
1,1,1TCA 100.2 ug/L 0
1,1 DCA 24.5 ug/L 2
Total 6

 

a . . .
not detected in prevrous sampling events

3.6 Conclusions

Data on geochemical and chlorinated organic compounds were collected on a
control volume in an area of a VOC contaminant plume to evaluate the extent to which
natural attenuation processes are occurring. Solid and liquid phase data show that TCE
concentrations in the solids are higher than concentrations found in the liquid phase. In
most natural attenuation studies in the literature, solid phase data is not considered or is
not evaluated. This can lead to underestimation of the contaminants in the area since
sorption increases with degree of chlorination of the compound.

The co-existence of elevated TCE and cis-DCE contaminated zones, both in the
liquid and solid phase, indicate that reductive dechlorination took place at certain times
within the contaminated area. The lower concentrations of VC and the accumulation of
cis-DCE provide evidence that reductive dechlorination is not happening at a significant
rate in this area. The absence of an electron donor that can promote biological usage of

chlorinated ethenes as electron acceptors could be the main reason for the apparent

79

accumulation of cis-DCE.

High concentration of solid-phase chlorinated ethene compounds in zones of high
conductivity indicates that contaminants in these zones have the potential to migrate
faster to downgradient receptors. However, the horizontal extent of these zones is
uncertain. Also, retardation with respect to the groundwater flow velocity is expected
since all of these compounds sorb to soil particles.

Although the EPA (1998) scoring system indicates that there is limited evidence
of reductive dechlorination in this area, the analysis performed in this study indicates that
at some point in time, reductive dechlorination was an important process for the
degradation of chlorinated compounds in the area, however, there is no evidence that
supports the hypothesis that these processes are still occurring in this region.

A conceptual analysis of the historical events for this site could help in identifying
all the physical, chemical, and biological processes that lead to the development of this

chlorinated solvents plume.

3.7 Acknowledgments

Financial support for this study was provided by the State of Michigan
Department of Environmental Quality. Partial support was also provided by a Graduate
Assistance in Areas of National Needs Fellowship and a GE Faculty for the Future
Fellowship. Special thanks to Dr. Xianda Zhao for his assistance during sample
collection. Also, the work of Leslie Dybas and Yan Pan with the laboratory analytical

chemistry is greatly appreciated.

80

3.8 Literature cited

Chapelle, F. H., Vroblesky, D. A., Woodward, J. C., and Lovley, D. R. (1997). "Practical
considerations for measuring hydrogen concentrations in groundwater."
Environmental Science & Technology, 31(10), 2873-2877.

Clement, T. P., Truex, M. J ., and Lee, P. (2002). "A case study for demonstrating the
application of US EPA's monitored natural attenuation screening protocol at a
hazardous waste site." Journal of Contaminant Hydrology, 59(1-2), 133-162.

Davis, J. W., Odom, J. M., DeWeerd, K. A., Stahl, D. A., Fishbain, S. S., West, R. J.,
Klecka, G. M., and DeCarolis, J. G. (2002). "Natural attenuation of chlorinated
solvents at Area 6, Dover Air Force Base: characterization of microbial
community structure." Journal of Contaminant Hydrology, 57(1-2), 41-59.

Dybas, M. J ., Barcelona, M., Bezborodnikov, 8., Davies, S., Fomey, L., Heuer, H.,
Kawka, 0., Mayotte, T., Sepulveda-Torres, L., Smalla, K., Sneathen, M., Tiedje,
J., Voice, T., Wiggert, D. C., Witt, M. E., and Criddle, C. S. (1998). "Pilot-scale
evaluation of bioaugmentation for in-situ remediation of a carbon tetrachloride
contaminated aquifer." Environmental Science & Technology, 32(22), 3598-3611.

Dybas, M. J., Hyndman, D., Heine, R., Tiedje, J ., Linning, K., Wiggert, D., Voice, T.,
Zhao, X., Dybas, L., and Criddle, C. (2002). "Development, Operation, and long-
term performance of a full-scale biocurtain utilizing bioaugmentation."
Environmental Science & Technology, 36(16), 3635-3644.

U.S. Environmental Protection Agency (EPA). (1998). "Technical protocol for evaluating
natural attenuation of chlorinated solvents in groundwater." EPA/600/R-98/128,
USEPA, Cincinnati, OH.

Ferguson, J. F., and Pietari, J. M. H. (2000). "Anaerobic transformations and

bioremediation of chlorinated solvents." Environmental Pollution, 107(2), 209-
215.

F ranzmann, P. D., Robertson, W. J., Zappia, L. R., and Davis, G. B. (2002). "The role of
microbial populations in the containment of aromatic hydrocarbons in the
subsurface." Biodegradation, 13(1), 65-78.

Kao, C. M., and Wang, Y. S. (2001). "Field investigation of the natural attenuation and

intrinsic biodegradation rates at an underground storage tank site." Environmental
Geology, 40(4-5), 622-631.

Lipinski, B. A. (2002). "Estimating natural attenuation rates for a chlorinated

hydrocarbon plume in a glacio-fluvial aquifer, Schoolcraft, Michigan," M.S.
thesis, Michigan State University, East Lansing, MI.

81

Lovley, D. R., Chapelle, F. H., and Woodward, J. C. (1994). "Use of dissolved H2
concentrations to determine distribution of microbially catalyzed redox reactions
in anoxic groundwater." Environmental Science & Technology, 28(7), 1205-1210.

Maymo-Gatell, X., Chien, Y., Gossett, J ., and Zinder, S. (1997). "Isolation of a bacterium
that reductively dechlorinates tetrachloroethene to ethene." Science, 276(6), 1568-
1571.

Mayotte, T. J ., Dybas, M. J ., and Criddle, C. S. (1996). "Bench-scale evaluation of
bioaugmentation to remediate carbon tetrachloride-contaminated aquifer
materials." Ground Water, 34(2), 358-367.

McCarty, P. L., Goltz, M. N., Hopkins, G. D., Dolan, M. E., Allan, J. P., Kawakami, B.
T., and Carrothers, T. J. (1998). "Full scale evaluation of in situ cometabolic
degradation of trichloroethylene in groundwater through toluene injection."
Environmental Science & Technology, 32(1), 88-100.

Ndon, U. J ., Randall, A. A., and Khouri, T. Z. (2000). "Reductive dechlorination of
tetrachloroethylene by soil sulfate-reducing microbes under various electron
donor conditions." Environmental Monitoring and Assessment, 60(3), 329-336.

NRC. (2000). Natural attenuation for groundwater remediation, National Academy
Press, Washington, DC.

Richmond, S. A., Lindstrom, J. E., and Braddock, J. F. (2001). "Assessment of natural
attenuation of chlorinated aliphatics and BTEX in subarctic groundwater."
Environmental Science & Technology, 35(20), 4038-4045.

Roling, W. F. M., and van Verseveld, H. W. (2002). "Natural attenuation: What does the
subsurface have in store?" Biodegradation, 13(1), 53-64.

Suarez, M. P., and Rifai, H. S. (2002). "Evaluation of BTEX remediation by natural
attenuation at a coastal facility." Ground Water Monitoring and Remediation,
22(1), 62-77.

Vogel, T. M., Criddle, C., and McCarty, P. L. (1987). "Transformation of halogenated
aliphatic compounds." Environmental Science & Technology, 21, 722-736.

Vogel, T. M., and McCarty, P. L. (1987). "Abiotic and biotic transformation of 1,1,1-
trichloroethane under methanogenic conditions." Environmental Science &
Technology, 21, 1208-1213.

Weaver, J. W., Wilson, J. T., and Kampbell, D. H. "Case study of natural attenuation of
trichloroethene at St. Joseph, Michigan." Proceedings of the symposium on
natural attenuation of chlorinated organics in ground water, Dallas, TX, 67-70.

82

Witt, M. E., Dybas, M. J ., Worden, R. M., and Criddle, C. S. (1999). "Motility-enhanced
bioremediation of carbon tetrachloride- contaminated aquifer sediments."
Environmental Science & Technology, 33(17), 2958-2964.

Witt, M. E., Klecka, G. M., Lutz, E. J., Ei, T. A., Grosso, N. R., and Chapelle, F. H.
(2002). "Natural attenuation of chlorinated solvents at Area 6, Dover Air Force
Base: groundwater biogeochemistry." Journal of Contaminant Hydrology, 57(1—
2), 61-80.

Yang, Y. R., and McCarty, P. L. (1998). "Competition for hydrogen within a chlorinated

solvent dehalogenating anaerobic mixed culture." Environmental Science &
Technology, 32(22), 3591-3597.

83

CHAPTER 4

DEVELOPMENT AND APPLICATION OF A NATURAL ATTENUATION
MODEL FOR A VOC CONTAMINATED AQUIFER UNDER
LIMITED ELECTRON DONOR CONDITIONS

4.1 Abstract

A reductive dechlorination model considering the bioavailability and of electron
donors was develop to be included in the mathematical formulation of natural
attenuation. This model was first tested in a hypothetical batch reactor to evaluate the
behavior of the mathematical expression and evaluate mass conservation at the end of the
simulation. Toluene was modeled as the electron donor with an initial concentration of
100ug/L. This concentration was not in stoichiometric excess for complete degradation
of the VOC in the reactor. It was assumed that PCE and TCE were initially present in the
reactor in equilibrium with the solid-phase. Liquid-phase concentration of PCE and TCE
were 100 and SOug/L, respectively. Results showed incomplete degradation of PCE due
to a lack of electron donors capable to sustain an active population of dehalogenators.
Also, accumulation of cis-DCE and VC was observed at the end of the modeling period.

The developed model was applied to a VOC contaminated aquifer in Schoolcraft,
M1 to evaluate the relative importance of reductive dechlorination in the natural
attenuation process occurring at this site. Based on the plume distributions at two
different time periods it was concluded that reductive dechlorination will not play a
significant role in the natural attenuation process due to the lack of electron donor sources

capable of maintaining reduced conditions. The use of Monod-type kinetics in the

84

mathematical formulation of reductive dechlorination makes the model difficult to apply

because of the significant number of constant parameters involved.

4.2 Introduction

Chlorinated aliphatic hydrocarbons are the most frequent detected contaminants
in groundwater due to their widespread use in chemical dry cleaning and as metal
degreasing agents (Ferguson and Pietari 2000; Middledorp et al. 1999; Shouakar-Stash et
al. 2003). Once in the groundwater, these contaminants tend to dissolve in the
groundwater contaminating large volumes of soil and water. Due to their resistance to
chemical and biological breakdown, they tend to persist in groundwaters for long periods
of time posing a risk to the human health and the environment.

Monitored natural attenuation has emerged as a potential remedy for chlorinated
solvent contamination because it is less expensive and in some cases more practical than
engineered cleanup solutions (Richmond et al. 2001). Natural attenuation of chlorinated
solvents has been successfully demonstrated in field and laboratory microcosm studies
(Delvin et al. 2002; Kao and Wang 2001; Ndon et al. 2000).

Witt et al. (2002) used the “lines of evidence” approach adopted by the US
Environmental Protection Agency (1998) to document the occurrence of natural
attenuation of chlorinated solvents at the Dover Air Force Base in Delaware. Biological
and geochemical data collected over a two-year period support the hypothesis that
sequential anaerobic and aerobic degradation of chlorinated solvents are occurring at this
site. Biological destruction of PCB and TCE was most likely due to reductive

dechlorination processes; however, the bioavailability of potential electron donors was

85

not evaluated in this study.

The same protocol was implemented at a Brooklawn hazardous waste site in
Baton Rouge, LO (USA) to evaluate the extent to which natural attenuation is responsible
for the observed decline in contaminant mass (Clement et al. 2002). The hazardous
waste plume generated from the source contained around twelve different chlorinated
organic compounds. Site-specific data indicated that chlorinated ethene and ethane
compounds are being attenuated within 300m downgradient from the source. A swamp
located upstream of the contaminated site is believed to provide the organic carbon
necessary for the biological destruction of the chlorinated compounds. Based on the
site’s conceptual model, it was concluded that reductive dechlorination process is
occurring with excess of electron donor. Although the carbon source in this study seems
to be adequate to support reductive dechlorination for a long period of time, an analysis
of the potential to rely on natural attenuation as a sole treatment technique to reduce
contaminants to regulatory standards was not conducted.

To assess potential future extent of plume migration and the sustainability of the
natural attenuation process it is necessary to include in the conceptual model all the
physical, chemical, and biological processes that play a key role in transport and
degradation of the contaminants. Of those mechanisms, biological destruction has
received broader attention since it is not a reversible process such as sorption and does
not involve a contaminant phase transfer such as volatilization.

Under natural conditions, it has been recognized that reductive dechlorination is
the biological mechanism responsible for the observed degradation of chlorinated

compounds (McCarty 1997) although cis-DCE and VC are susceptible to aerobic

86

degradation (Klier et al. 1999; Vogel et al. 1987). The complete destruction of PCE and
TCE to ethene requires, among other things, a supply of electron donors that
microorganisms can use for growth and maintenance. Conceptual and numerical models
usually ignore the presence of a carbon source needed for reductive dechlorination.

To date, most models of natural attenuation of chlorinated solvents do not assess
the bioavailability of carbon sources to sustain complete depletion of VOC compounds.
Fennel] and Gossett (1998) pointed out that the complex nature of reductive
dechlorination has not been addressed in fate and transport models applied to natural
attenuation and demonstrated that the inclusion of a bio-kinetic term for hydrogen
utilization in the mathematical expression for reductive dechlorination closely matched
data collected in batch reactors. Another study by Haston and McCarty (1999) suggested
that zero-order or Monod kinetics were more appropriate than first-order kinetics to
model reductive dechlorination processes occurring in microcosms amended with PCE as
the electron acceptor and hydrogen as the electron donor. In this study, however, the
electron donor was in excess to ensure complete dechlorination of PCE to ethene.

The objective of this study is to develop a reductive dechlorination numerical
model coupled to oxidation of a carbon source and assess if availability of the electron
donor is limiting the biological component of natural attenuation in a contaminated
aquifer. Toluene was used as the electron donor for the microbial population and PCB
and TCE were used as electron acceptors according to the redox stoichiometry presented.
The model was initially tested in a hypothetical batch reactor to evaluate the behavior of
the mathematical expressions. The model was applied to a VOC contaminated site to

study the natural attenuation process occurring. It is believed that this approach to model

87

natural attenuation will help to assess the sustainability of biological destruction of VOC

compounds at other sites where reductive dechlorination is occurring.

4.3 Model development and methodology
Reductive dechlorination stoichiometry with toluene as electron donor

The stoichiometric equations describing reductive dechlorination with toluene as
the electron donor can be obtained by combining appropriate half-reaction expressions.
The following equations were combined and represent the conceptual model for reductive

dechlorination with toluene as the electron donor:

PCE toluene _ be TCE
micro s + _
18C2Cl4 +14H20+C6HSCH3 ———-——-——-)18C2HC|3 +7C02 +l8H +18Cl (4.1)
TCE . b cisDCE
”(32,103 +14H20+C6H5CH3 M18C2H2C'2 + 7C02 +18H+ +18Cl—(4.2)
cisDCE . b VC
”(3211202 +14H20+C6H5CH3 w13C2H3CI+ 7c02 + 1311* +18C1“(4.3)
VC ethene
microbes

13C2H3C|+14H20+C6H5CH3 ————-)13C2H + 7C02 +18H+ +18Cl- (4.4)

4

A general pathway showing the sequence of degradation of the chlorinated

compounds catalyzed by microbial activity is given in Figure 4.1. Gibbs free energy

88

 

$23

c1 c1 PCE
i

1 1
/ .54. ¢ . {CE
ql H q] 81 H: 91
Cl H H H Cl H
1,1- cis- trans-
\. t ./

H H

c1 ‘ H VC

3:5

H H ethene

Figure 4.1. General pathway for the reductive dechlorination process of
PCB to ethene. Bold arrows indicate the most likely pathway
under the influence of microbial processes (Garant & Lynd,
1998)

 

 

 

calculations show that all these reactions are feasible at standard temperature and

pressure (Table 4.1). However, the reductive dechlorination pathway is a step-by-step

process in a series that is generally; but not always, catalyzed by more than one organism

that ultimately leads to the formation of ethene (Dolfing 2000).

The simplified stoichiometric reactions presented here ignore intermediate steps

in the reductive dechlorination process. It is important to emphasize that those

89

Table 4.1. Gibbs free energy for reductive dechlorination
with toluene as the electron donor.

 

 

Electron Acceptor AG°' kcal/mole a
PCB -425 .6
TCE -404.9
cis-DCE -340.1
VC -331 .5

 

aat 298 K and 1 atm

intermediate steps determine what type of electron accepting process will ultimately take
place in a contaminated environment and consequently certain microbial populations will
outcompete others for preferential consumption of those intermediate metabolites. Such
is the case for hydrogen. Research has shown that anaerobic hydrogen production in
natural environments determines the type of electron accepting processes that will be
favored (Lovley et al. 1994; Lovley and Goodwin 1988). Researchers have tried to
correlate hydrogen concentration with specific reductive dechlorination process (Fennell
and Gossett 1998; Haston and McCarty 1999; Yang and McCarty 1998). In general, it
has been found that hydrogen concentrations above 5 nm produced favorable conditions

for chloroethene dechlorination (Meer et al. 2001).

Kinetic model for electron donor

Reductive dechlorination kinetics is coupled to toluene degradation using Monod
expressions. Although different microbial populations can be involved in this complex
process (McCarty 1997), it was assumed that a single population growing on toluene was
capable of using the chlorinated ethenes as electron acceptors. Competitive inhibition
was not modeled here since those kinetic expressions would tend to complicate further

the application of the model to the aquifer environment and there is no evidence that such

90

a formulation will improve model predictions in natural environments.
Based on Monod kinetics, the rate of toluene utilization by the dehalogenators

would be:

 

 

d[tol] = _q max [x][ [toll ]_ 5M] [pol]- [5%.] (4.5)

lto|l+Ks [tol] Kd[tol]

where [tol] is the concentration of toluene [M L'3]; qmax is the maximum specific
toluene utilization rate [M M" Tl]; [X] is the biomass concentration [M L'3]; Ks [ml] is
the half-velocity coefficient for toluene consumption [M L'3]; [tolLoil is the solid-phase
toluene concentration [M M'l]; Ethel] is the first-order mass transfer coefficient between
liquid and solid phase; Kd [ml] is the toluene partitioning coefficient [L3 M'l]; and t is the
time [T].

Transfer of toluene between solid and liquid phase is modeled assuming first

order mass transfer kinetics. Solid-phase concentration of toluene is given by:

dltOIlsoi _ ¢§[tol] lt0llsoi
dt I P Um]— Kdltoll (4.6)

where d) is the soil porosity [L3 L3]; and p is the bulk density of the soil matrix [M L3];
The net rate of active biomass growing on the dissolved toluene using chlorinated

ethene compounds as electron acceptors can be written as:

91

 

d—D‘lu [x1[ [“1 ]-b[x1 (4.7)

dt [tol]+ K, [ml]

where umax is the maximum specific growth rate of dehalogenators [”1"], and b is the
cell endogenous decay coefficient [Tl]. Limitation on growth with respect to electron

acceptors was not considered since it was assumed that VOC compounds are in excess

and the reductive dechlorination process is controlled by the bioavailability of electron

donors.

The relation between biomass growth rate and substrate utilization is given by:
“max = qmax Y[)(l/[tol] (4-8)
where Y[x]/ [ml] is the yield coefficient for cell synthesis [M M'l].

Kinetic model for reductive dechlorination
Chlorinated compounds are modeled as a kinetic limited liquid-solid phase with

biological degradation in the liquid phase. The concentration of the PCE on the liquid

phase is given by:

 

 

leCE] — ~ [PCE] [PCE]soil
dt — _q[PCEl le[ [PCE]+ Ks [PCE] J" é[PCE] [[PCEl" ml (4.9)

where [PCE ] is the liquid phase PCE concentration [M L‘3]; q[pCE] is the maximum

92

specific utilization rate of PCE [M M" TI]; [(5 [pCE] is the PCE half-velocity coefficient
[M L'3]; §[PCE] is the first-order mass transfer rate coefficient [Tl]; [PCELOH is the
sorbed PCE concentration [M M'l]; and KdlpCE] is the PCB partitioning coefficient

between the solid and the liquid phase [L-3 M-l]. The first term on right hand side of
Equation 4.9 represents the utilization of PCE by dehalogenators and the second term is

the kinetic limited sorption-desorption. A mass balance of PCE on the solid phase yields:

d[PCE]SOi] = 4) §[PCE] [[PCEl- [PCElsoil] (4.10)
(it p Kd[PCE]

Utilization of PCE as the electron acceptor during toluene degradation will yield a
stoichiometric amount of TCE (Equation 4.1). The utilization rate of TCE can be found

by:

 

 

 

d TCE " PCE
[dt ]=YlTCEl/lPCElq[PCE][X][lPCE][+KslPCEl ]- (4 ll)
. [TCE] _ [Elisml .
Q[TCE] [X][ [TCE]+ Ks [TCE] J‘ ngCEl [[TCE] Kd[TCE] ]

where YlTCE]/[pCE] is the stoichiometric yield of PCE to TCE [M M'l]; [TCE ] is the
liquid phase TCE concentration [M L'3]; q[TCE] is the maximum specific utilization rate

of TCE [M M" T']; K is the TCE half-velocity coefficient [M L'3]; é is the
s [TCE] [TCE]

93

TCE first-order mass transfer rate coefficient [1"]; [TCELOH is the sorbed TCE
concentration [M M'l]; and Kd [TCE] is the TCE partitioning coefficient between solid

and liquid phase. The first term on the right hand side of Equation 4.11 is the metabolic
production of TCE due to PCE degradation. The second and third terms represents
biodegradation and sorption-desorption processes, respectively. Solid-phase TCE

concentration is given by:

 

Kd[rc1~:]

d[TCE]SOi] _ ‘l’ ngCE] [[TCE] [TCELoil] (4.12)
dt P

Similarly, expressions for the rate of production and consumption for DCE, VC,
and ethene can be obtained. A mass balance on the liquid and solid phase for DCE, VC,

and ethene will give:

 

 

 

d DCE ~ TCE
-[—dt_1 = YlDCEl/[TCEl qlTCEl [xll [TCE][+ Ks lrce] l (4 13)
. DCE DCE so, i
‘ q [DCE] [X][ [DCEl+ Ks local } ngCE] [[DCEl— IK—dlE’JCE—ll
d[DCE]soil = (l) r;[Dc13] [[DCE]- mil-$21) (4.14)
dt p Kd [DCE]

94

d[VC

dt ] = Y[VC]/[DCE] <1[DCE] le[

 

[DCE] ]

[DCE]+ K s [DCE]

_ fi[vc] [X][ [VC] 1 ]_ §[vc] [[VC]_ [vclsoil ]

lVCEl+ Ks [vc K<1[vc]

(4.15)

 

d[VC]soil = ‘l’ §|vc| [[DCEl— [VClsoil] (4.16)

dt P Kd[VC]

d[ethene]
dt

[VC] ]

[VC]+Ks[vc]

— 2i[vc] (lVCl - [215%]

K d [vc]

= Y[ethene]/[VC] Cl[vc] lxl[
(4.17)

 

d[ethene]soil = ‘l’ €[ethene] [eth e n e]— M] (4.18)
dt P Kd [ethene]

Reaction terms for DCE, VC, and ethene are analogous to the terms in the PCE and TCE

equations.

Batch mode testing

Equations 4.5 to 4.18 were coded in FORTRAN and solved for a hypothetical
batch reactor using RT3D (Clement 1997). For this experiment, a batch reactor with a
toluene concentration of lOOug/L was used. This concentration was not in stoichiometric

excess with respect to PCE or TCE. Starting concentrations of liquid-phase PCB and

95

Table 4.2. Parameters for the reductive dechlorination model.

 

 

Parameter Definition Value Model Value Source

Ks [ml l half-velocity coefficient for 35.0 - 59.0 35.0 Elmen et al.
toluene, mg L' (1997)

K (1 [ml] toluene partitioning 1 6.6x10'7 _ calculated based
coefficient, L mg- 2.23x10'6 on foe (Zhao et

al., 1999)

q max maximum toluene I 0.2 — 3.97 0.3 Elmen et al.
utilization rate, d' (1997)

“max maximum biomasls specific 0.1 — 6.03 0.46 Elmen et al.
growth rate, d' (1997)

Y [x] / [ml] yield coefficient for toluene 0.64 0.64 Reardon et al.
degraders, mg V88 (2000)
mg tol'l

b endogenous cell decay 0.001 — 0.3 0.1 Assumed
coefficient, (1'1

9[PCE] maximum specific PCE 0.0035 — 0.004 EPA (1999)
utilization rate, d' 0-0046

Ks [PCE] half velocity coefficient for 0.008 — 32.8 0.008 Rittman &

PCE , mg L' McCarty (2001)
§[PCE] mass transfer rate 0.0001 - 0.1 0.1 Clement & Jones
coefficient for PCE, d] (1998)
K d [PCE] PCE partitioning I 1_25x1()'6 _ 1.3,(10'6 calculated based
coefficient, L mg' 2.15x10'6 on foe (Zhao et
al,1999)

4) soil porosity 0.2 — 0.4 0.35 measured from
laboratory re-
packed columns

p bulk density of soil matrix, 159 x [o6 1. 59 x 106 Zhao et al.

mg L' (1999)
‘1 [T CE] maximum TCE utilization 0.006 — 0.008 0.007 EPA (1999)
rate, d'1
PCE to TCE stoichiometric 0.79 0.79 from
Y [Tail/[PCE] yield coefficient, mg stoichiometric
TCE mg PCE'l equations
Ks [T CE] half velocity coefficient for 0.18 — 31.2 0.18 Rittman &
TCE, mg L" McCarty (2001)
§[TCE] mass transfer rate 0.0001 — 0.1 0.1 Clement & Jones
coefficient for TCE, d'1 (1993)
K d [T CE] TCE partitioning I 0.52x10°6 _ ()_7x1()'6 calculated based
coefficient, L mg' 0.9x10'6 on foe (Zhao et
al., 1999)
(“003] maximumlDCE utilization 0.058 — 0.547 0.09 Rittman &
rate, d' McCarty (2001)
TCE to DCE stoichiometric 0.57 0.57 from
Y [DCEMTCE] yield coefficient, mg stoichiometric
DCE mg TCE'I equations
Ks [DCE] half velocity copfficient for 0.288 — 0.371 0.290 Rittman &
DCE, mg L' McCarty (2001)
§[DCE] mass transfer rate l 0.0001 — 0.1 0.1 Clement & Jones
coefficient for DCE, d- (1993)

 

96

Table 4.2. (continued)

 

 

Parameter Definition Value Model Value Source
K d [DCE] TCE partitioning 1 0,29x10'6 _ 0,4,(10‘5 calculated based
coefficient, L mg' 0_43x10’6 on foc (Zhao et
al., 1999)
€1[VC] maximuleC utilization 0.037 - 0.285 0.05 Rittman &
rate, d McCarty (2001)
DCE to VC stoichiometric 0.65 0.65 from
Y [VCMDCE] yield coefficient, mg stoichiometric
vc mg DCE'I equations
Ks [VC] half velocity coefficient for 0.161 — 22.3 0.161 Rittman &
VC , mg L' McCarty (2001)
El" C] mass transfer rate 1 0.0001 — 0.1 0.1 Clement & Jones
coefficient for VC, d' (1998)
K d [VC] VC partitlioning coefficient, 2,0,‘10'9 _ 2.0,(10'9 calculated based
L mg’ 3.4,(10'7 on foe (Zhao et
al., 1999)
VC to ethene 0.45 0.45 from
Y [ethenel/[VC] stoichiometric yield stoichiometric
coefficient, mg VC mg equations
DCE"
qethcne] mass transfer rate 0.0001 — 0.1 0.1 Clement & Jones
coleffic1ent for ethene, (1998)
d'
Kd [ethene] ethene partitioning 1 1.0,(10‘9 _ 1.0x10'9 calculated based
coeffic1ent, L mg’ 3.6x10'7 on foc (Zhao et
al., 1999)

 

TCE were 100 and SOug/L, respectively. An initial concentration of toluene degraders
equal to 0.001 mg VSS/L was assumed to be present in the reactor. Total porosity in the
reactor was 0.30 and the soil bulk density was 1.6x104’mg/L. Constant parameters for the
equations were extracted from laboratory and field studies in the literature and are listed
in Table 4.2.

To evaluate mass conservation of the numerical scheme an equivalent PCE
concentration computed on the basis of all chlorinated ethene compounds was compared
at any two selected time steps. Equation 4.19 was used to calculate the equivalent PCE

concentration at any time, t:

97

 

[PCEleq {[pcs]+ [P Elsoup ++[[rc1~:] [“35]...an 1 ]

¢ Y[TCE]/[PCE]

+ [DCE]+ [DCEJSN' p)[ l ]

Y[DCE]/[TCE] Y[TCE]/[PC1~:]

+ [VC]+ [VCL°"p][ l ] (4.19)

Y[vc]/[DC1~:] Y[DCE]/[TCE] Y[TCE]/[PCE]

(
l
+£[ethene-1—]+ [ethenelsoil P)

 

 

 

 

¢

 

l
xYlethene]/[VC]Y[VC]/[DCE] Y[DCE]/[TCE] Y[Tc1-:]/[PCE] ll

Model application to a VOC contaminated aquifer

The developed model was applied to a VOC contaminated aquifer located in
Schoolcraft, MI (USA). Regional and local hydrogeologic conditions for this site have
been described elsewhere (Lipinski 2002; Dybas et al. 1998; Mayotte et al. 1996). The
unconfined aquifer has been contaminated with organic and heavy metal compounds as
result of previous industrial activities (Figure 4.2). A VOC plume extending
approximately 2 km southeast of the suspected source has developed. ARCO Industries,
a former manufacturer of automobile plastic parts was identified as the source of this
contamination, which impacts an estimated 1.3x107m3 aquifer materials.

The existence of metabolic by-products of degradation of PCB and TCE confirm
that biological destruction of these compounds have occurred in the past. However, data
collected over the last 2 years reveal that reductive dechlorination is not a major
component of the current natural attenuation process. Initial soil contamination reports

revealed that not only PCE and TCE were handled at this facility but also toluene and

98

other BTEX compounds. However, none of these compounds have been detected in
recent groundwater and soil samples. It is believed that once the electron donors were
depleted, reductive dechlorination ceased and only physical processes are responsible for
the natural attenuation of VOC compounds at this site. The reductive dechlorination
model developed in this study will be used to gain an understanding of the processes that

lead to the transport and spread of these contaminants.

Reactive transport model for the Plume G site
The following expression (Bear 1979) was used to simulate groundwater flow in

the area of interest:

a ah ah
— K-— + :3 — 4.20
axil ‘axil q, ’6: ( )

where x, is distance along the respective Cartesian coordinate [L]; K; is the principal
component of the hydraulic conductivity tensor [L Tl]; h is the hydraulic head [L]; qS is
the fluid source-sink term [L Tl]; Sy is the specific storage of the aquifer [L'l]; and t is
the time, [T].

The concentration of each chlorinated ethene compounds along with toluene and

dehalogenators can be found by (Clement et al. 1998):

ac of:

a a (Is I
— D -— -C +—-c +R 4.21
at ¢ at axi[ ] (V1 k) (l) k k ( )

fig): axi

99

where Ck is the dissolved concentration of the kth specie [M L’3]; Ck is concentration of

the kth species sorbed on the subsurface solids [M M"]; Dij is the hydrodynamic
dispersion coefficient tensor [L2 T']; C’k is the source-sink flux term concentration for

the kth species [M-L'3]; and R, is the reaction term for the kth specie. The reaction terms
for the dissolved species and the solid-phase concentrations are defined by Equations 4.5
through 4.18.

Figure 4.2 shows the model domain and the wells where pump tests were
conducted to define the hydraulic conductivity field of the unconfined contaminated
aquifer. The constant head boundaries surrounding the area were defined by telescopic
grid refinement (Anderson and Woessner 1992) from a regional groundwater flow model
developed by Lipinski (2002). The regional model considers all the major hydrogeologic
features of the area and extends the boundaries to lakes, rivers, and aquifer no-fiow
boundaries that control groundwater movement in the vicinity.

The flow and transport models were solved numerically using MODFLOW 2000
(Harbaugh et al. 2000) and RT3D (Clement 1997), respectively. The “user-defined
reactions” module was used for the partial differential equations describing the reductive
dechlorination process. The FORTRAN code for the reactions is given in Appendix 111.

Model parameters and initial conditions for the species involved are given in Table 4.3.

100

 

 

LEGEND
O K data from pump test

[:1 VOC plume
model boundary

 

  

 

Figure 4.2. Site location and model boundaries for reactive transport simulation.

Table 4.3. Flow model parameters and initial conditions for the natural
attenuation simulation of the VOC contaminated site.

 

Model parameter Value

F low Model (steaajz state)
hydraulic conductivity, cm s" values kriged from 89 pump tests
effective porosity 0.2 — 0.3
boundary conditions from Lipinski (2002)

recharge, m day"
Reactive Transport Model
stress periods

Starting concentrations
Toluene, pg L"
Toluene sorb, pg Kg"
PCE, pg L"

PCE sorb, pg Kg"
TCE, pg 1."

TCE sorb, pg Kg"
cis-DCE, pg L"
cis-DCE sorb, pg Kg"
VC. pg L"

VC sorb, pg Kg"
dehalogenators, mg VSS L"

6.1x10

Loading period: 1953 ~ 1988 (35 yrs)
no loads: 1988 ~ 2003 (15 yrs)

500
330
150
190
1000
520
0.0
0.0
0.0
0.0
0.001

 

4.4 Results and discussion
Reactions batch mode testing

Initial liquid-phase concentrations of toluene, PCE, and TCE in the batch reactor
were 100, 100, and 50pg/L, respectively. Concentration of cis-DCE, VC, and ethene
were assumed equal to zero since these two compounds were expected by-products of the
utilization of PCE and TCE.

The relation between toluene degradation and growth of dehalogenators is shown
in Figure 4.3. An initial acclimation period of 40 days is required to develop a biomass
concentration that can degrade the available toluene in the theoretical reactor. Biomass
concentration reaches a maximum of 80mg VSS/L at approximately 50 days following

toluene addition. These reactions were modeled for a total of 150 days with a time step

 

 

 

 

 

 

 

 

120 . .
- 1.4 102
100 2
—toluenel ‘ 1'2 10
7" 30 _ ----- XI - 1.0102 :5
d 5
00
3 . - 8.010‘ <
2 6° ' 5‘. E
g 1 '-, - 6.0101 :3.
"‘ 40 - 1 '1,
‘1‘ - 4.0 10‘
20 l - 2.0 10‘
0 v" - "'---L 0.0 100
o 50 100 150

time (days)
Figure 4.3. Toluene and biomass concentration in the hypothetical
batch reactor simulation.

102

of 2.4 hours. Total utilization of toluene is achieved in approximately 60 days. The
specific toluene utilization rate was 0.3d" and the maximum growth rate used was
0.456d".

Figure 4.4 shows the PCE concentration in the liquid and the solid phase.
Utilization of the PCE as the electron acceptor starts approximately at the same time of
toluene utilization. Complete degradation of PCE is not achieved since activity of
dehalogenators decrease when toluene is utilized and there is no alternate substrate that
can promote its growth. At the end of the simulation period, a liquid phase PCE
concentration of 42pg/L is still present in the reactor.

The solid phase PCE concentration starts desorbing from the solid particles as
soon as the concentration in the liquid phase decreases. This desorption process occur to
maintain equilibrium between the solid and the liquid phase. However, the slope of the
liquid-phase PCE degradation curve in the interval where PCE is being utilized is steeper
than its solid counterpart due to a kinetically controlled mass transfer mechanism
between the two phases. At the end of the simulation period, a solid-phase PCE
concentration of approximately 4x10'6pg/mg will remain attached to the solid particles.

PCE degradation will yield a TCE amount according to the stoichiometry given in
Equation 4.1. An initial 50pg/L of TCE was added to the model reactor. Figure 4.5
shows that TCE remains unchanged for approximately 50 days until a decrease in
concentration is observed. An increase in the liquid TCE concentration due to the
utilization of PCE was not observed. The reason could be the slightly higher TCE
utilization rate (Table 4.2) used in the simulation. This model predicts that

approximately 15pg/L of TCE will remain in the reactor at the end of 150 days.

103

120

100

60

-1
PCEmobile (pg L )

40

20

 

 

 

  

 

 

 

 

 

 

 

, , 1.210'5
.............. ~ - 1.0 10‘5
~ ‘ --PCE|
“ ---- PCEsorbI . 8.0 10"
- 6.010'6
‘ .................. . 4.0104S
' - 2.010'6
. - 0.0 100
0 50 100 150
time (days)

Figure 4.4. Liquid and solid phase PCE concentration for the
hypothetical batch reactor simulation.

60

50

30

-l
TCEmobile (l1 g L )

20

10

 

40‘

 

 

 

 

 

 

 

 

. . 5.0 10“
“s. —TCE| - 4.0 10.6
l'. ----- TCEsorb I
- 3.0 10"
. - 2.0 10*5
- - 1.0 10"
. - 0.0 10“
0 50 100 150
time (days)

Figure 4.5. Liquid and solid phase TCE concentration in the batch

reactor.

104

(1-3‘“ 311) Wand

(,.3m 811) "mam.

Complete degradation of TCE is not achieved since the biomass concentration declines
when there is no more toluene to support its growth (Figure 4.3). TCE equilibrium
concentration in the solid phase at the end of the simulation is 1.5x10'6pg TCE/mg soil.

Concentrations of metabolic by-products from the PCB and TCE degradation are
shown in Figures 4.6, 4.7, and 4.8. TCE degradation results in the production of cis-
DCE, cis-DCE degradation results in the production of vinyl chloride, which
consequently degrades to produce ethene. At this stage, the biomass concentration has
declined substantially and the metabolic chlorinated by products start accumulating in the
hypothetical reactor. At the end of the simulation the concentrations in the liquid phase
of cis-DCE, VC, and ethene are 35, 2.1, and 2.8pg/L, respectively. Solid-phase
concentrations of these three compounds are 1.7x10'6pg cis-DCE/mg soil, 4.5x10'9pg
VC/ mg soil, and 2.8x10'9pg ethene/mg soil. Accumulation of these byproducts occurs
due to the lack of electron donors capable of supporting a microbial population that can
perform complete degradation of the chlorinated compounds and its by-products.

An equivalent PCE concentration was computed with Equation 4.19 at 0 and 50
days to ensure mass conservation throughout the simulation period. PCE equivalent
concentration at 0 days was 247pg/L, which compares reasonable well with 244pg/L at
50 days. The equivalent PCE concentration at 50 days yields a 1.3% error with respect to
the initial equivalent PCE concentration. The difference between these values could be
attributed to small numerical errors introduced by the algorithm.

The model showed high sensitivity to several parameters. For example, an order
of magnitude change in the biomass decay coefficient b, resulted in a dramatic change in

concentration profiles for all organic compounds. Also, each particular equation showed

105

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

40 I F 3.0 10.6
35 -
- 2.510‘5
_A 30 - i?"
in 25 " 2.010-6 g
E: b m
a """""""""""""" g
E 20 ' - 1.510‘ 3
ts: .' 9
U I 9°
9 15 I. 0' -6 a
.é .' - 1.010 mg
” 10 - —DCE|
5 ----- DCEsorbI - 5.010'7
' ' 0.0 10°
0 50 100 150
time (days)
Figure 4.6. Liquid and solid phase DCE concentration in the batch
reactor.
2.5 . i 8.0 10‘9
7.010"
2 I. 9
4 6.010'
9 8
- 5.010' §
- 4.010'9 ,1:
5
-1 3.010.9 v.1.
VC - 2.010‘9
""" VCsorbI -9
. - 1.010
' 0.0100
100 150

 

time (days)
Figure 4.7. Liquid and solid phase VC concentration in the batch reactor.

106

3 , . 6.0 10'9

 

 

 

 

 

 

 

 

 

 

 

2.5 - - 5.010"
.7" £12
.4 2 . - 4.0 10" 5'
so ,3
E .. .. '9 '
g 15 ________________ q 3.0 10 9
Q 0'. N
E . 3
E5 1 - - 2.010 V:-
—Ethene|
0.5 - * - 1.0 10"
""" Ethene sorbl
0 - 0.0 10“
0 50 100 150
time (days)
Figure 4.8. Liquid and solid phase ethene concentration in the batch
reactor.

high sensitivity to the maximum substrate utilization rate coefficient. This model has a
total of 27 different reaction parameters of which only the stoichiometric yield
coefficients can be easily determined. Also, field values for all of these parameters are
expected to be different from those determined in the laboratory. Similar Monod type
kinetic models have accurately predicted reductive dechlorination processes in laboratory
reactors; however, the application of these models to field conditions becomes
complicated because measurement of constant parameters is not possible.

These modeling results are representative of what may occur at sites where a
steady supply of electron donors is not available to support complete reductive
dechlorination. It is believed that this condition occurred in a VOC contaminated aquifer

in Schoolcraft, MI where partial reductive dechlorination was observed.

107

Flow model for Schoolcraft VOC plume site

A regular 2-dimensional finite difference numerical grid was constructed for the
model region shown in Figure 4.2. Details of the model domain are given in Table 4.4.
The spatial discretization resulted in 98,900 finite difference cells. A small grid size was
used to minimize artificial (numerical) dispersion effects.

Mean log(K) from 89 pump tests on selected wells (Figure 4.2) across the site was
-1.48 (10"‘48cm/s), which is a typical value for outwash sediments composed of medium
to coarse sands and gravel. The parameters were used to solve Equation 4.20 numerically
using MODF LOW 2000 (Harbaugh et al. 2000).

Figure 4.9 shows the solution of the calibrated steady-state flow model for the
domain. Groundwater flows towards the southeast with a hydraulic gradient of
0.001rn/m. This compares reasonably well with previously reported values obtained in
the vicinity of the contaminated aquifer (Dybas et al. 2002; Hyndman et al. 2000). Head
data collected across the site was used in the model calibration. A plot of the final
observed vs. computed heads is shown in Figure 4.10. Comparison of these results with

the ones presented by Lipinski (2002) confirms the accuracy of the flow model.

Table 4.4. Details of the numerical model domain.

 

 

Model Area Parameter Value
domain 1400 m x 2630 at
no. of columns 230
no. of rows 430
no. of layers 1
cell dimensions
Ax, m 6.1
Ay, m 6.1

 

108

 

 

   
 

; ”7.1
300 m

Leer—2,142

—262.0-' Potentiometric head in meters

 

 

 

Figure 4.9. Groundwater contour map in the VOC contaminated region.

Reactive transport model results for the chemical loading period between 1953 and I 988.
The reductive dechlorination model was coupled to a transport model to simulate
the VOC plume in Schoolcraft, MI. Previous reports indicate that industrial activities at
the facility started approximately in 1953. This VOC plume was discovered in 1986
during a soil and groundwater contamination investigation conducted by personnel of the
Michigan Department of Environmental Quality (formerly Michigan Department of
Natural Resources). Sources of contamination were removed during late 1980’s to early

1990’s.

109

263 .

 

262.5 - .
E
'8
0
"'4 262 - @O .
i s
a
.D
0
261.5 - -

c9

 

 

 

261 l l l
261 261.5 262 262.5 263
computed head (m)

Figure 4.10. Comparison between observed and computed heads
for several wells in the region.

Based on the historical release of VOC compounds from the ARCO facility, two
major stress periods were assumed for modeling purposes. The first stress period
considers a constant loading of chemicals from 1953 to 1988. The second stress period
considers removal of the suspected source of contamination and no further loadings of
chemicals into the aquifer. This corresponds to the years of 1988 to 2003. The total
simulation time was 50 years.

Although suspected sources of contamination within the ARCO facility have been
identified in previous studies (Lipinski 2002), the entire facility was assumed as a source
due to the lack of information on the spatial and temporal distribution of contaminant
releases in those sub-areas. An approximate surface area of 10,000 m2 was used as the

contaminant handling area for the 35 years of loading.

110

Figure 4.11 show the simulated and observed 5pg/L iso-concentration line for the
chlorinated compounds at the end of the 35 year loading period. This concentration value
was chosen because it is the maximum concentration level (MCL) of PCE and TCE in
drinking water in the US (EPA 2003). The model predicts the front of the 5pg/L PCE
contour to be located approximately 1000m southeast from the source at 35 year (Figure
4.1 1(a)). Comparison with the delineated contour using data from monitoring wells
shows a reasonable agreement between both contours lines. The longitudinal extension
of the observed PCE plume is 960m, a difference only of 40m compared with the model
calculated. Aerial extent of the observed PCE plume is 177,600 m2 and the model
predicted contour is 323,000 m2. The predicted PCE plume covers almost twice the area
of the observed plume. A longitudinal dispersion coefficient of 0.20m was used in this
model. This value was selected based on previous tracer studies in an area inside the
plume. However, the longitudinal scale of the tracer study was 1.5% of the total plume
scale. This suggests that dispersivity values for the plume should be larger than the
estimated dispersivity for the tracer studies as this parameter varies with the length sale.

The modeled TCE plume has traveled 200m more than the modeled PCE plume
for the same period (Figure 4.11(b)). This was expected since a lower partitioning
coefficient was used for TCE; and PCE sorbs strongly to the soil compared to TCE. The
longitudinal extent of the modeled TCE plume is in close agreement with the observed
plume at the end of the loading period. The aerial extent of the observed TCE plume is

262,000 m2 and the extent of the simulated plume is 335,000 m2

111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a) (b)

 

((0

Figure 4.11. Simulated (solid) and observed (dashed) 5pg/L isoconcentration line for (a)
PCE, (b) TCE, (c)cis-DCE, and (d) VC at the end of the loading period (35
years).

The front of the 5pg/L cis-DCE contour has not traveled as far as TCE (about
100m less than DCE) in both the modeled and the observed plumes (Figure 4.1 1(c)). A
reason for this could be that TCE was initially present at the source and was not subjected
to biodegradation until an active biomass capable of TCE utilization developed. This can

be seen in both, the modeled and the observed cis-DCE plume. The longitudinal extent

112

of the observed cis-DCE plume is approximately 10% longer than the modeled cis-DCE
plume. The front of the predicted 5pg/L cis-DCE contour is located 927m from the
facility and the observed front is located at an approximate distance of 1 190m. However,
the areas covered by these plumes are very similar, i.e. 223,150 m2 for the model plume
and 221,400 m2 for the delineated.

A discrepancy between the model and the observed contours is revealed in the VC
plume for this period (Figure 4.1 l(d)). The model predicts the VC plume to extend
1,433m from the source whereas the observed front of the 5pg/L VC contour is located
1,150m from the suspected source. The areas covered by these plume are 167,630 m2
and 537,700 m2 for the observed and simulated plumes, respectively. Most likely, the
partitioning coefficient used in the simulation was higher than the real field value. Also,
a higher dispersion coefficient could produce an elongated VC plume similar to the

observed plume.

Model results for the period of 1988 to 2000.

Results from the loading period were used as starting concentrations for a reactive
transport model considering removal of the contaminant sources. Figure 4.12 shows the
simulated and delineated 5pg/L contour line for all VOC compounds for the year 2000.
Panel (a) of Figure 4.12 shows the size of the PCE plume to be significantly reduced as
compared to Figure 4.1 1(a). The extent of the observed plume is smaller than the
simulated one. Possibly, the PCE specific utilization rate for field conditions is higher
than the value used in the model. Also, it might be possible that other carbon sources

such as naturally occurring organic matter served as an electron donor after depletion of

113

 

 

 

toluene since measured concentrations of this compound in 2000 were all below detection
limits. Areas of the PCE plumes for the year 2000 are 150,000 m2 and 23,200 m2 for
model predicted and observed plumes, respectively.

Figure 4.12(b) shows 5pg/L TCE contour lines for the year 2000. The predicted
contour is delayed by 150m; however, the areas covered by the plumes are very similar,
247,000 m2 for the simulated vs. 214,000 m2 for the observed plume. Comparison of this
panel with that of Figure 4.11 reveal that although some biological attenuation of TCE
has occurred, the magnitude is small compared to PCE degradation.

The longitudinal extent of the observed cis-DCE plume for this time period is
approximately 1 1 10m (Figure 4.12(c)), which is about the same size it has at the end of
the loading time period. This indicates that no major biodegradation of cis-DCE has
occurred during the last 15 years and the only attenuation mechanisms that could be
taking place are dilution and dispersion. The model simulated contour for cis-DCE
shows a different shape as compared to the observed one. However, the areas are very
similar; 280,000 m2 for the observed vs. 300,280 m2 for the model predicted plume.

Similarly, in Figure 4.12(d) the model predicts 3 VC plume different than the
observed one. The observed VC plume is elongated and extends through an approximate
area of 206,000 m2 with a longitudinal extension of 1040 m. The model simulated plume
has an approximate area of 194,340 m2 with a length of 440 m. As with the cis-DCE

plume, the aerial extension is reasonable, but the longitudinal extension of the observed

114

 

(b)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(C)

Figure 4.12. Simulated (solid) and delineated (dashed) 5pg/L isoconcentration line for
(a) PCE, (b) TCE, (c)cis-DCE, and (d) VC at 47 years.

plume is about two times what the model predicted. The reason could be that the length

dependency of the dispersion coefficient was not considered in these simulations

Assessment of risk to downgradient groundwater receptors

The impact of the contaminant plume on potential groundwater receptors was

115

evaluated with the calibrated model. This plume has the potential to discharge into a
series of surface water bodies located approximately 4 km southeast of the Schoolcraft
village. The results for a predictive model run are shown in Figure 4.13. The 5pg/L PCE
contour has not changed significantly from the results obtained for the year 2000. The
reason could be that this plume is at “steady—state” where the rates of attenuation are
comparable to the travel times in this sandy aquifer. Also, since PCE tends to sorb to soil
particles, its migration could be subjected to a significant retardation compared to the
groundwater velocity.

The model predicts the TCE plume; delineated by the 5pg/L contour, slightly
downgradient as compared to the 2000 yr position. The behavior of this plume is similar
to the PCE plume.

The 2050 yr cis-DCE plume has migrated considerably as compared to the results
at 2000. However, it has not reach yet the surface water bodies and apparently, dilution
and dispersion mechanisms are sufficient to ensure that concentrations above regulatory
limits do not reach receiving surface water bodies. This simulation shows that these three
plumes, i.e. (PCE, TCE, and cis-DCE), do not represent a risk to the receiving surface
water bodies for the modeling period of 1953 to 2050.

The simulated vinyl chloride plume; however appears to reach the lake and start
discharging into it at concentrations higher than regulatory limits. The fact that vinyl
chloride has migrated significantly more than its parent products is consistent with the

sorption characteristics of this compound.

116

 

 

 

12:)
MC
:3
T4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a
8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a) (b)
....0 1 :_ 0 1 -
”‘3 l
i (
O
/ ‘L‘
r, .

 

 

 

 

 

 

 

 

 

 

 

((1)

Figure 4.13. Simulated 5pg/L isoconcentration line for (a) PCE, (b) TCE, (c)cis-DCE,
and (d) VC for the year 2050.

This results show that of the VOC compounds most likely to be produced in this
aquifer, higher risks are associated with VC production. However, based on analytical
data collected on a previous study, production of VC is not significant and the reductive
dechlorination process does not seem to have the electron donor requirements for the

conversion of cis-DCE to VC.

117

4.5 Summary and conclusions

Results from this study show that the VOC plume at Schoolcraft Plume G site is
not currently undergoing a significant biological attenuation due to depletion of
substrates that can support an active population of dehalogenators. The developed
numerical model showed a reasonable agreement with the observed plumes for PCE,
TCE and cis-DCE. The VC plume was over-predicted by the model; however, not all
wells in Figure 4.2 were analyzed for chlorinated compounds during the sampling events
conducted on 1988 through 1989 and 2000.

Results from Figure 4.12 show a reduction in aerial extent of PCE plumes. Most
likely this reduction is caused by biological degradation processes. The TCE plumes are
similar in shape but the simulated plume is spatially delayed for about 200m. The
computed cis-DCE and VC contour deviates significantly from contours delineated using
the 2000 data. However, this data set is limited and not all the wells were sampled during
this sampling event.

Predictive modeling results show that only vinyl chloride has the potential to
migrate, threatening to discharge in the lakes located at the southeast hydrological
boundary. The other VOC plumes apparent to be in “steady-state” and do not represent a
risk to the surface water bodies in this area.

An important feature of this approach to model natural attenuation is the inclusion
of the electron donor in the mathematical formulation of reductive dechlorination.
Reductive dechlorination has been successfully modeled in laboratory reactors using
Monod-type kinetics (Fennell and Gossett 1998; Garant and Lynd 1998; Haston and

McCarty 1999); however this type of approach has not been used to model sites where

118

reductive dechlorination is occurring. An evaluation of the concentration of electron
donor compounds support the hypothesis that no major biological activity is occurring in
the absence of those sources. This information is important to assess the sustainability of
the natural attenuation process and ultimately will benefit decision makers when selecting
the appropriate remediation strategy for contaminated sites.

This study shows that although reductive dechlorination was a major component
of natural attenuation in the past, most likely it is not a major contributor to the overall
process due to the lack of electron donor sources to create a reduced environment capable
of sustaining a reductive dechlorination process that could result in complete depletion of
the chlorinated ethene compounds.

One potential limitation of this approach is the number of constant parameters
involved in the mathematical expressions for reductive dechlorination. Methods for
measuring those parameters in the laboratory are well established; however, these values
cannot be applied to field conditions. A systematic approach to measure them has not yet

been developed.

4.6 Acknowledgments

Funding for this research was provided by the Michigan Department of
Environmental Quality. Partial support was also provided by fellowships from Graduate
Assistance in Areas of National Needs and a GE Faculty for the Future Fund. Special

thanks to Brian A. Lipinski for providing the groundwater regional model for the area.

119

4.7 Literature Cited

Anderson, M. P., and Woessner, W. W. (1992). Applied Groundwater Modeling:
Simulation of flow and advective transport, Academic Press, INC., San Diego,
Ca.

Bear, J. (1979). Hydraulics of Groundwater, McGraw-Hill.

Clement, T. B., and Jones, N. L. (1998). "RT3D tutorials for GMS users." PNNL-11805,
Pacific Northwest National Laboratory, Richland, WA.

Clement, T. P. (1997). "A modular computer model for simulating reactive multi-species
transport in three-dimensional ground water systems." PNNL-SA-11720, PNNL,
Richland, Washington.

Clement, T. P., Sun, Y., Hooker, B. S., and Petersen, J. N. (1998). "Modeling multi-
species reactive transport in ground water." Ground Water Monitoring &
Remediation, 18(2), 79-92.

Clement, T. P., Truex, M. J ., and Lee, P. (2002). "A case study for demonstrating the
application of US EPA's monitored natural attenuation screening protocol at a
hazardous waste site." Journal of Contaminant Hydrology, 59(1-2), 133-162.

Delvin, J. F ., McMaster, M., and Barker, J. F. (2002). "Hydrogeologic assessment on in
situ natural attenuation in a controlled field experiment." Water Resources
Research, 38(1), 1 ~11.

Dolfing, J. (2000). "Energetics of anaerobic degradation pathways of chlorinated
aliphatic compounds." Microbial Ecology, 40(1), 2-7.

Dybas, M. J ., Barcelona, M., Bezborodnikov, 8., Davies, S., F omey, L., Heuer, H.,
Kawka, O., Mayotte, T., Sepulveda-Torres, L., Smalla, K., Sneathen, M., Tiedje,
J., Voice, T., Wiggert, D. C., Witt, M. E., and Criddle, C. S. (1998). "Pilot-scale
evaluation of bioaugmentation for in-situ remediation of a carbon tetrachloride
contaminated aquifer." Environmental Science & Technology, 32(22), 3598-3611.

Dybas, M. J., Hyndman, D., Heine, R., Tiedje, J., Linning, K., Wiggert, D., Voice, T.,
Zhao, X., Dybas, L., and Criddle, C. (2002). "Development, operation, and long-

term performance of a full-scale biocurtain utilizing bioaugmentation."
Environmental Science & Technology, 36(16), 3635-3644.

Elmen, J ., Pan, W., Leung, S. Y., Magyarosy, A., and Keasling, J. D. (1997). "Kinetics of

toluene degradation by a nitrate-reducing bacterium isolated from a groundwater
aquifer." Biotechnology and Bioengineering, 55(1), 82-90.

120

 

U.S. Environmental Protection Agency (EPA). (1998). "Technical protocol for evaluating
natural attenuation of chlorinated solvents in groundwater.” EPA/600/R-98/128,
USEPA, Cincinnati, OH.

U.S. Environmental Protection Agency (EPA). (1999). "Anaerobic biodegradation rates
of organic chemicals in groundwater: A summary of field and laboratory studies."
U.S. Environmental Protection Agency, Washington, DC.

U.S. Environmental Protection Agency (EPA). (2003). "National primary drinking water
standards." EPA 816-F-03-016, EPA, Washington, DC.

Fennell, D. E., and Gossett, J. M. (1998). "Modeling the production of and competition

for hydrogen in a dechlorinating culture." Environmental Science & Technology,
32(16), 2450-2460.

Ferguson, J. F., and Pietari, J. M. H. (2000). "Anaerobic transformations and

bioremediation of chlorinated solvents." Environmental Pollution, 107(2), 209-
215.

Garant, H., and Lynd, L. (1998). "Applicability of competitive and noncompetitive
kinetics to the reductive dechlorination of chlorinated ethenes." Biotechnology
and Bioengineering, 57(6), 751-755.

Harbaugh, A. W., Banta, E. R., Hill, M. C., and McDonald, M. G. (2000). "Modflow-
2000, The U.S. Geological Survey modular ground-water model-User guide to
modularization concepts and the ground-water flow process." ofi00-92, United
States Geological Survey, Reston, VA.

Haston, Z. O, and McCarty, P. L. (1999). "Chlorinated ethene half-velocity coefficients
(K-s) for reductive dehalogenation." Environmental Science & Technology, 33(2),
223-226.

Hyndman, D., Dybas, M. J ., F omey, L., Heine, R., Mayotte, T., Phanikumar, M. S.,
Tatara, G., Tiedje, J ., Voice, T., Wallace, R., Wiggert, D., Zhao, X., and Criddle,
C. S. (2000). "Hydraulic characterization and design of a full-scale biocurtain."
Ground Water, 38(3), 462-474.

Kao, C. M., and Wang, Y. S. (2001). "Field investigation of the natural attenuation and
intrinsic biodegradation rates at an underground storage tank site." Environmental
Geology, 40(4-5), 622-631.

Klier, N. J ., West, R. J ., and Donberg, P. A. (1999). "Aerobic biodegradation of
dichloroethylenes in surface and subsurface soils." Chemosphere, 38, 1175-1188.

121

Lipinski, B. A. (2002). "Estimating natural attenuation rates for a chlorinated
hydrocarbon plume in a glacio-fluvial aquifer, Schoolcraft, Michigan," M.S.
thesis, Michigan State University, East Lansing, MI.

Lovley, D. R., Chapelle, F. H., and Woodward, J. C. (1994). "Use of dissolved H2
concentrations to determine distribution of microbially catalyzed redox reactions
in anoxic groundwater." Environmental Science & Technology, 28(7), 1205-1210.

Lovley, D. R., and Goodwin, S. (1988). "Hydrogen concentrations as an indicator of the
predominant terminal electron-accepting reaction in aquatic sediments."
Geochimica Cosmochimica Acta, 52, 2993-3003.

Mayotte, T. J., Dybas, M. J ., and Criddle, C. S. (1996). "Bench-scale evaluation of
bioaugmentation to remediate carbon tetrachloride-contaminated aquifer
materials." Ground Water, 34(2), 358-367.

McCarty, P. L. (1997). "Breathing with chlorinated solvents." Science, 166, 1521-1522.

Meer, J. T., Eekert, M., Gerritse, J ., and Rijnaarts, H. (2001). "Hydrogen as indicator for
redox conditions and dechlorination." Natural attenuation of environmental
contaminants: The sixth international in situ and on-site bioremediation
symposium, A. Leeson, M. E. Kelly, H. S. Rifai, and V. S. Magar, eds., Battelle
Press, San Diego, CA.

Middledorp, P. J ., Luijten, M. L. G. C., van de Pas, B. A., van Eekert, M. H. A., Kengen,
S. W. M., Schraa, G., and Stams, A. J. M. (1999). "Anaerobic microbial reductive
dehalogenation of chlorinated ethenes." Bioremediation Journal, 3(3), 151-169.

Ndon, U. J ., Randall, A. A., and Khouri, T. Z. (2000). "Reductive dechlorination of
tetrachloroethylene by soil sulfate-reducing microbes under various electron
donor conditions." Environmental Monitoring and Assessment, 60(3), 329-336.

Richmond, S. A., Lindstrom, J. E., and Braddock, J. F. (2001). "Assessment of natural
attenuation of chlorinated aliphatics and BTEX in subarctic groundwater."
Environmental Science & Technology, 35(20), 4038-4045.

Reardon, K. F., Mosteller, D. C., and Bull Rogers, J. D. (2000). " Biodegradation kinetics
of benzene, toluene, and phenol as single and mixed substrates for pseudomonas
putida F1." Biotechnology and Bioengineering, 69(4), 385-400.

Rittman, B. E., and McCarty, P. L. (2001). Environmental Biotechnology: Principles and
Applications, McGraw-Hill, San Francisco, CA.

Shouakar-Stash, O., Frape, S. K., and Drimmie, R. J. (2003). "Stable hydrogen, carbon

and chlorine isotope measurements of selected chlorinated organic solvents."
Journal of Contaminant Hydrology, 60(3-4), 211-228.

122

 

 

Vogel, T. M., Criddle, C., and McCarty, P. L. (1987). "Transformation of halogenated
aliphatic compounds." Environmental Science & Technology, 21, 722-736.

Witt, M. E., Klecka, G. M., Lutz, E. J., Ei, T. A., Grosso, N. R., and Chapelle, F. H.
(2002). "Natural attenuation of chlorinated solvents at Area 6, Dover Air Force
Base: groundwater biogeochemistry." Journal of Contaminant Hydrology, 57(1-
2), 61-80.

Yang, Y. R., and McCarty, P. L. (1998). "Competition for hydrogen within a chlorinated
solvent dehalogenating anaerobic mixed culture." Environmental Science &
Technology, 32(22), 3591-3597.

Zhao, X., Szafranski, M. J ., Maraqa, M. A., and Voice, T. C. (1999). "Sorption and

bioavailability of carbon tetrachloride in a low organic content sandy soil."
Environmental Toxicolog/ and Chemistry, 18(8), 1755-1762.

123

 

 

 

 

CHAPTER 5

SUMMARY AND CONCLUSIONS

5.] Introduction

In recent years natural attenuation has been widely applied, either in conjunction
with engineered technologies, or as a sole remediation strategy to clean up contaminated
sites. According to EPA data, natural attenuation use in the Superfund program
developed during 1990’s from application at 6% to more than 25% of groundwater
contamination sites (Macdonald 2000). At some sites natural attenuation is being used as
a sole remediation technology while at other sites, a combination with engineered
treatment technologies has being adopted.

Scientists have demonstrated that natural attenuation can destroy certain
contaminants, primarily fuel hydrocarbons (NRC 2000). However, it has been
recognized, that the surge in use of natural attenuation has outpaced the development of
adequate guidelines for its use (Renner 2000). Yet, this treatment technology has been
approved as a formal remedy, despite the limitations in scientific understanding.

Chlorinated compounds are among the most detected contaminants in soil and
groundwaters. Their presence in the environment poses a high risk to the human health
and the environment due to their high toxicity and mobility. Treatment sites
contaminated with chlorinated solvents by natural attenuation has been a focus of
intensive research during the last decade. However, the current level of understanding of

the process leading to their destruction has been described as moderate (Macdonald

124

 

 

2000), in part because an agreement among researchers on how to evaluate and document
natural attenuation of chlorinated solvents does not exist.

Guidelines for documenting natural attenuation of chlorinated solvents are
evolving rapidly as the scientific understanding of the processes progress. These
guidelines have been used to document natural attenuation at several contaminated sites
(Alleman and Leeson 1999; Eganhouse et al. 2001; Witt et al. 2002). However, it has
been recognized that these protocols are oftentimes misused, leading to wrong
conclusions about the natural attenuation process at contaminated sites.

In general, it has been recognized that natural attenuation protocols should be
replaced by methods (NRC 2000) which assign more weight to the specific conditions of
the site being investigated. The goal of this research was to evaluate natural attenuation
of chlorinated solvents by studying the processes that most likely influence this treatment
strategy. This chapter summarizes, in a systematic way, the process employed during this
research to understand and document the natural attenuation process occurring at

Schoolcraft Plume G site.

5.2 Hydrogeological characterization of the VOC contaminated area
Conceptual hydrogeological model development

It is widely recognized that a site’s hydrogeological characteristics influence
dramatically the transport and distribution of compounds in aquifer systems (Hyndman et
al. 2000). The characterization of the plume G site groundwater flow was accomplished
through a study of the general regional groundwater flow (Lipinski 2002) and developing

a local conceptual model for the chemically impacted regions. However, local

125

hydrogeological characteristics at the plume’s scale resolution were not derived from this
analysis. A series of cores were drilled in a location impacted by the VOC plume and the
extracted material was analyzed to identify local heterogeneities in the saturated aquifer
zone that could have influence the migration of VOC compounds. From this analysis two
distinctive preferential flow zones were discovered.

A series of tracer tests were used to test the hypothesis that contaminants in this
region have a tendency to migrate to downstream groundwater receptors through the
zones identified in the soil core analysis. Breakthrough curves generated from the data
collected at downstream monitoring wells confirmed the influence that these zones exert
on migration patterns of contaminants in this aquifer region. Figure 5.1 show the

stratigraphy of the site with typical tracer breakthrough curves obtained in each zone.

 

 

 

 

 

 

 

 

 

 

 

 

Legend 36.5 m 0 I—é—‘wa-
I:] well graded sands 0 8 16
12:1 sorted sands w/ gravel time, days
well graded gravel
- clay

Figure 5.1. Cross section of the Plume G site with distinctive stratigraphic tracer
breakthrough curve.

126

Hydrogeological Characterization: quantitative description

A quantitative description of the hydrogeologic features identified in the
conceptual model was obtained by the methods described in Chapter 2. Field and
laboratory hydraulic conductivity experiments revealed that samples with the higher
conductivity values correlate well with the conceptual model description of the site. In
general, outwash sediments composing the upper unconfined aquifer have hydraulic
conductivity values ranging from 1x10'3 to 4x10"cm/s.

The hydraulic conductivity and tracer experiments revealed a different transport
behavior among the stratigraphic units. To quantitatively describe those differences, a
methodology to optimize the estimate of dispersion parameters using all possible
collected data was developed. Estimates of depth-specific average linear velocity using
this procedure revealed that for high conductivity zones, the average linear velocity was
considerably higher than the rest of the aquifer. For example, the average linear velocity
at 23.2m bgs was 99.5 d: 9.81cm/day, and at 21 .3m bgs was 33.12 i 2.99 cm/day. The
model developed showed that optimization on a depth-specific basis validates the
observations better than using a single dispersivity value for the entire model domain.
Figure 5.2 shows two simulation scenarios; one with the layer optimum dispersivity and

one with the entire grid domain optimum dispersivity.

127

 

C/Co

 

 

 

 

 

time (days)

Figure 5.2. Tracer breakthrough curves for the simulated
scenarios. The solid line represents the depth specific
case, the dashed line represents the entire model grid
case, and the solid circles are the observed data.

5.3 Geochemical analyses: insight into the natural attenuation process

Generally, footprints of chemical reactions are used as indicators of processes
leading to natural attenuation of contaminants. Measurements of footprints give a
preliminary indication whether or not contaminants at a particular site are being
attenuated.

The most widely accepted destruction mechanism for chlorinated solvents is
reductive dechlorination (McCarty 1997). Figure 5.3 illustrates a conceptual model of
the breakdown of perchloroethene via reductive dechlorination (PCE). A supply of an

electron donor capable of maintaining reduced conditions in the aquifer is an essential

128

C1' C1' C1‘ C1'
PCE —L-> TCE —L> cis-DCE —L> VC —l—> ETH

anaerobic anaerobic anaerobic anaerobic

Figure 5.3. PCB to ethene breakdown through reductive dechlorination (Clement et al.
2000)

component of reductive dechlorination.

Analysis of geochemical and biological indicators performed during this
investigation revealed that biological attenuation at this site has occurred in the past. The
presence of metabolic by-products form the PCE and TCE degradation confirm that
reductive dechlorination processes have occurred. However, environmental conditions
that will ensure complete destruction of PCE, TCE and their metabolic by-products are
not favorable. Analysis of geochemical and biological parameters led to the conclusion
that reductive dechlorination occurred at this site up to the point where all sources of
electron donors were depleted. From that point on, the attenuation that is occurring can
be attributed only to physical processes.

Complete destruction of chlorinated solvents has been observed in places where co-
contaminant plumes of fuel hydrocarbons exist. Typically, the necessary reduced
environment is provided by the degradation of the compounds associated with these co-
contaminants. Figure 5.4 shows a typical scenario where different levels of chlorinated
solvent attenuation can be achieved based on the location and distribution of the
petroleum hydrocarbon plume. Panel (c) of figure 5.4 could have been the scenario that
occurred at the Plume G site. However, the co-contaminant plume was not sufficient for

complete degradation of the chlorinated compounds.

129

 

PH plume
Strongly reduced zone

 
        

  

Aerobic [ b No natural attenuation of any
groundwater _ - g \ - . __ _ portion of the solvent plume
7 {>\ \I \\ \ x.
w - . we
CS plume
(a)
PH plume Significant natural attenuaiton
/ of this portion of the solvent plume

  
          

  

     
  
  

 

     

/ /’ l 0.3:?“ '06:”. o
0'... so a... " no
0 o o o It so...‘o.o,o o o 1 ~

CS plume

Aerobic p“ \

groundwater

 
  

  
  
 

 

No natural attenuation of this
portion of the solvent plume

(b)

Significant natural attenuaiton
of all the chlorinated solvents plume

 

. 1
328513....l—Q

(C)

Figure 5.4. Illustration of three different scenarios that can be found in co-contaminated
environments. (a) Non-interacting petroleum and chlorinated solvents
plumes, (b) partly interacting plumes, and (c) completely interacting plumes.
PH = petroleum hydrocarbon, CS = chlorinated solvents. (NRC 2000)

130

 

5.4 Reductive dechlorination model for the VOC contaminants plume

Numerical computer-based models can be useful when the complexity of the
hydrogeology and the bio-geochemistry need to be capture in solute transport models.
Based on the analysis of the geochemical and biological indicators at the Plume G site, a
reactive transport model incorporating Monod kinetics in reactions terms for chlorinated
solvents was developed. It has been recognized that there is a need for models of natural
attenuation of chlorinated solvents to address the complex nature of reductive
dechlorination (Fennell and Gossett 1998; Haston and McCarty 1999). However, this
type of model should be developed only when the underlying processes are studied and
understood well enough that they can be represented by mathematical formulations and
when data is available to generate a reasonable matrix of parameter estimates.

The numerical model developed during this research simulated the plumes of each
of the VOC compounds. Reasonable agreement was found between observed and
simulated plumes. A predictive model run showed that only VC has the potential to
migrate and contaminate a series of surface water bodies located at the southeast by the

year 2050.

5.5 Summary

This research explores a methodology for evaluating natural attenuation of
chlorinated solvents using conceptual and numerical models considering the
hydrogeology and biogeochemistry intrinsic to the contaminated environment. A
hydraulic characterization of the site revealed the existence of geologic features that

influence dramatically the transport and distribution of solutes in a particular region.

131

Footprints of natural attenuation identified that biological destruction of the VOC
compounds occurred in the past. However, biogeochemistry of the aquifer confirms that
these mechanisms are not contributing significantly to the natural attenuation process at
this site.

A numerical model was developed incorporating Monod type kinetics in the
description of the reaction terms for the chlorinated compounds found in this aquifer.
This model incorporated all the information collected at the site. Two different loading
periods were simulated and the results showed reasonable agreement with plumes
delineated with observed data.

Natural attenuation processes of chlorinated solvents at Plume G site
methodology were evaluated by means of site specific data coupled with the development
of conceptual and numerical models. This novel approach deviates from the traditional
screening systems employed in most technical protocols to date, criticized for
overestimating the magnitude of natural attenuation processes. Also, the series of
experiments in this dissertation could serve as the basis for new protocols and guidelines

for evaluating natural attenuation of chlorinated solvents.

5.6 Conclusions

1. Field tracer experiments coupled with numerical methods and optimization techniques
proved to be effective in the hydrogeologic characterization of contaminated
environments. The existence of preferential flow pathways was revealed by the

characterization study performed.

132

 

2. Solid and liquid phase analysis of biogeochemical parameters revealed the importance

 

of considering both phases when evaluating the extent of naturally occurring
degradation processes. This experiment revealed that biological components of the
natural attenuation process do not contribute significantly to reduce the contaminant

mass and concentration at this site.

3. A model coupling the bioavailability of electron donors in reductive dechlorination
processes was successfully employed to describe natural attenuation of chlorinated

solvents in a contaminated aquifer. This evaluation confirmed that reductive

 

dechlorination was important as some point in the plume’s life-time. However, due to
the lack of electron donor sources, reductive dechlorination does not appear to be a
major contributor for the natural attenuation process at this site. Also, the predictive
model showed that only vinyl chloride represents a risk to downgradient groundwater

receptors during the simulated time period.

 

4. The methodology used in this research could be the foundation for a new approach to
evaluate natural attenuation of chlorinated solvents using techniques that deviate from

traditional approaches.

5.7 Literature cited

Alleman, B. C., and Leeson, A. (1999). Natural attenuation of chlorinated solvents,

petroleum hydrocarbons, and other organic compounds, Battelle Press,
Columbus, OH.

133

 

Eganhouse, R. P., Cozzarelli, I. M., Scholl, M. A., and Matthews, L. L. (2001). "Natural
attenuation of volatile organic compounds (VOCs) in the leachate plume of a

municipal landfill: Using alkylbenzenes as process probes." Ground Water, 39(2),
192-202.

Fennell, D. E., and Gossett, J. M. (1998). "Modeling the production of and competition
for hydrogen in a dechlorinating culture." Environmental Science & Technology,
32(16), 2450-2460.

Haston, Z. G, and McCarty, P. L. (1999). "Chlorinated ethene half-velocity coefficients

(K-s) for reductive dehalogenation." Environmental Science & Technology, 33(2),
223-226.

Hyndman, D., Dybas, M. J., F omey, L., Heine, R., Mayotte, T., Phanikumar, M. S.,
Tatara, G., Tiedje, J., Voice, T., Wallace, R., Wiggert, D., Zhao, X., and Criddle,
C. S. (2000). "Hydraulic characterization and design of a full-scale biocurtain."
Ground Water, 38(3), 462-474.

Lipinski, B. A. (2002). "Estimating natural attenuation rates for a chlorinated
hydrocarbon plume in a glacio-fluvial aquifer, Schoolcraft, Michigan," M.S.
thesis, Michigan State University, East Lansing, MI.

Macdonald, J. A. (2000). "Evaluating natural attenuation for groundwater cleanup."
Environmental Science & Technology, 34(15), 346A-353A.

McCarty, P. L. (1997). "Breathing with chlorinated solvents." Science, 166, 1521-1522.

NRC. (2000). Natural attenuation for groundwater remediation, National Academy
Press, Washington, DC.

Renner, R. (2000). "Natural attenuation's popularity outpaces scientific support, NRC
finds." Environmental Science & Technology, 34(9), 203A-204A.

Witt, M. E., Klecka, G. M., Lutz, E. J., Ei, T. A., Grosso, N. R., and Chapelle, F. H.
(2002). "Natural attenuation of chlorinated solvents at Area 6, Dover Air Force

Base: groundwater biogeochemistry." Journal of Contaminant Hydrology, 57(1-
2), 61-80.

134

 

 

 

 

 

APPENDICES

 

 

135

APPENDIX A - BORING LOGS FOR THE WELLS INSTALLED FOR THE

PILOT SCALE STUDY

Well id: MPA-l

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 18.9 No recovery
18.9 20.9 Medium to Coarse Sand w/ Fine Gravels SW
20.9 21.0 Graded Gravels, Mixtures of Medium Conglomerate
and Fine Gravels w/ Little or no Fines
21.0 22.3 No recovery
22.3 22.6 Medium Sands SP
22.6 24.4 Medium to Coarse sands with big GP
pebbles and cobbles
24.4 25.6 Gray clay, stiff, dry, cohesive CL

 

Well id: MFA-2

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.6m
14.6 14.9 Blind drill to 14.6m
14.9 15.8 Medium to Coarse sands/ Well Graded SW
Sand
15.8 16.8 Graded Gravels, Mixtures of Medium Conglomerate
and Fine gravels w/ little or no fines
16.8 22.4 No recovery
22.4 23.0 Medium to Coarse Sands/ well graded SW
sands
23.0 23.6 Graded Gravels, Mixtures of Medium Conglomerate
and fine gravels w/ little or no fines
23.6 25.1 Medium to Coarse Sands w/ big pebbles GP
and Cobbles
L 25.1 25.6 Gray clay, stiff, dry, cohesive CL

 

136

 

 

Well id: MFA-3

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.1 No recovery
17.1 19.4 Medium to coarse sands/ well graded SW
sands
19.4 21.0 Gravel-Sand Mixtures/ Poorly graded GW
Gravels
21.0 21.9 No recovery
21.9 24.6 Medium to Coarse Sand w/ big pebbles GP
and cobbles
24.6 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

Well id: MFA-4

 

 

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 17.4 No recovery

17.4 20.1 Medium to coarse sands/ more medium SW

20.1 23.5 Gravel-Sand Mixtures/ Poorly graded GW

gravels
23.5 24.4 No recovery
24.4 25.6 Gray clay, stiff, dry, cohesive CL

 

 

Well id: MPA-S

 

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 17.1 No recovery

17.1 19.5 Medium to Coarse Sands, more medium SW

19.5 21.0 Gravel-Sand Mixtures/ Poorly Graded GW
Gravels

21.0 22.6 No recovery

22.6 24.5 Medium to Coarse Sands w/ big pebbles GP
and cobbles

24.5 25.6 Clay CL

 

 

 

137

 

 

 

Well id: MPA-6

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.] Medium to Coarse Sand/ more medium SW
17.1 18.6 Gravel-Sand Mixtures/ poorly graded GW
gravel
18.6 22.6 No recovery
22.6 23.2 Medium to coarse sands SW
23.2 25.0 Medium to coarse sands w/ big pebbles GP
and cobbles
25.0 25.6 Clay CL
Well id: MPA-7
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 19.7 Brown fine sand, saturated SW
19.7 20.9 Medium to coarse sand/ more medium SW
20.9 24.4 Brown medium to coarse sand, SW
saturated
24.4 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

Well id: MPA-8

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.7 No recovery
17.7 19.4 Medium to coarse sand/ more medium SW
19.4 21.0 Gravel-Sand Mixtures/ poorly graded GW
gravel
21.0 22.6 No recovery
22.6 25.0 Medium to coarse sands w/ bib pebbles GP
and cobbles
25.0 25.6 Clay CL

 

 

138

 

 

 

Well id: MPA-9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 18.0 No recovery
18.0 20.1 Medium to coarse sand/ more medium SW
20.1 21.0 Gravel-Sand mixtures/ poorly graded GW
gravel
21.0 22.3 No recovery
22.3 24.4 Medium to coarse sand w/ big pebbles GP
and cobbles
24.4 25.6 Gray clay, stiff, dry, cohesive CL
Well id: MPA-10
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 16.8 No recovery
16.8 19.8 Medium to coarse sand/ more medium SW
19.8 21.0 Gravel-Sand Mixtures/ poorly graded GW
gravel
21.0 22.3 No recovery
22.3 24.4 Medium to coarse sand/ w big pebbles GP
and cobbles
24.4 25.6 Gray clay, stiff, dry, cohesive CL
Well id: MPA-ll
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.7 No recovery
17.7 19.8 Medium to coarse sand/ more medium SW
19.8 21.0 Gravel-Sand mixtures/ poorly graded GW
gravel
21.0 21.6 No recovery
21.6 22.7 Medium to coarse sand w/ gravel and GP
pebbles and cobbles / more coarse sand
w/ fine gravel
22.7 24.4 Medium to coarse sand w/ pebbles and GP
cobbles
24.4 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

 

139

Well id: MFA-12

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 18.6 No recovery
18.6 20.1 Medium to coarse sand/ more medium SW
20.1 21.0 Gravel-Sand mixtures/ poorly graded GW
gravel
21.0 25.6 No recovery
Well id: MPA-l3
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9
14.6 14.9 Blind drill to 14.9
14.9 17.4 No recovery
17.4 19.9 Medium to coarse sand/ more medium SW
19.9 20.8 Gravel-Sand mixtures/ poorly graded GW
gravel
20.8 21.0 Medium to coarse sand SW
21.0 22.3 No recovery
22.3 24.5 Medium to coarse sand w/ big pebbles GP
and cobbles
24.5 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

Well id: MFA-14

 

 

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 19.8 Brown fine sand, saturated SW

19.8 20.7 Brown coarse sand and gravel GW

20.7 21.5 Brown coarse sand, trace gravel, SW
saturated

21.5 24.8 Mixture of medium to coarse sand w/ GP
big pebbles and cobbles

24.8 25.6 Clay CL

 

 

140

 

 

 

Well id: MPA-IS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 16.2 Medium to coarse sand/ more medium SW
16.2 17.5 Gravel-Sand mixtures/ poorly graded GW
gravel
17.5 22.6 No recovery
22.6 22.8 Coarse gravel GP
22.8 25.0 Medium to coarse sand w/ big pebbles GP
and cobbles
25.0 25.6 Gray clay, stiff, dry, cohesive CL
Well id: MPS-l
Depth (m) Description Classification
From To
0.0 14.9 Blind drill to 15.2m
14.9 15.2 Blind drill to 15.2m
15.2 17.7 No recovery
17.7 19.7 Medium to Coarse Sand/ more medium SW
19.7 20.4 Gravel-Sand Mixtures/ more gravel, GW
well graded gravel
20.4 21.0 Medium to coarse Sand/ well mixed SW
sand
21.0 22.9 No recovery
22.9 23.9 Medium to coarse sand /w big pebbles GP
and cobbles
23.9 25.3 Gray clay, stiff, dry, cohesive CL
Well id: MPS-2
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 15.2m
14.6 14.9 Blind drill to 15.2m
14.9 18.0 No recovery
18.0 18.6 Fine to medium sand/ more fine SW
18.6 19.6 Fine to medium sde more medium SW
19.6 20.2 Gravel-Sand Mixtures/ more gravel, GW
well graded gravel
20.2 21.0 No recovery
21.0 23.2 No recovery
23.2 23.5 Medium gravel w/ large pebbles and GP
cobbles, mixed material
23.5 24.8 Medium to coarse sand /w large pebbles GP
and cobbles
24.8 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

141

 

 

 

Well id: MPS-3

 

 

 

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 15.2m
14.6 14.9 Blind drill to 15.2m
14.9 16.5 Medium to coarse sand/ more medium SW
16.5 18.0 No recovery
18.0 18.6 Medium to coarse sde more medium SW
18.6 19.2 Gravel-Sand Mixtures/ more gravel, GW
well graded gravel
19.2 21.0 No recovery
21.0 21.9 No recovery
21.9 24.5 Gravel-Sand Mixtures, Coarse sand w/ GP
gravel at the top, Coarse sand w/ big
pebbles and cobbles at bottom
24.5 25.3 Gray clay, stiff, dry, cohesive CL
Well id: MPS-4
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 49'
14.6 14.9 Blind drill to 49'
14.9 16.8 No recovery
16.8 21.0 Available on Dr. Zhao's log book, not l
found in the field
21.0 23.8 Medium to coarse sand/ Probable SW
mixture of gravel
23.8 25.1 Gravel-Sand Mixtures/ Poorly graded GP
gravel
25.1 25.6 Clay CL

 

 

 

 

 

Well id: MPS-S

 

 

 

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 15.2m

14.6 14.9 Blind drill to 15.2m

14.9 18.3 Medium to coarse sand/ more medium SW

18.3 19.2 Gravel-Sand mixtures/ more gravel, GW
well graded gravel

19.2 21.0 Medium to coarse sand, well-mixed SW
sands

21.0 22.3 No recovery

22.3 22.7 Gravel-Sand mixtures/ poorly graded GP
gravel

22.7 25.1 Medium to coarse sand/ w big pebbles GP
and cobbles

25.1 25.6 Gray clay, dry, stiff, cohesive CL

 

 

142

Well id: MPS-6

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 16.8 No recovery
16.8 19.4 Medium to coarse sand, more medium SW
19.4 21.0 Gravel-Sand Mixtures/ well graded GW
gravel
21.0 22.7 No recovery
22.7 23.2 Medium to coarse sand SW
23.2 23.8 Gravel-Sand Mixtures/ poorly graded GP
gravel
23.8 25.3 Medium to coarse sand / w big pebbles GP
and cobbles
Well id: MPS-7
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 18.7 No recovery
18.7 20.6 Medium to coarse sand/ more medium SW
20.6 21.0 Gravel-Sand Mixtures/ well graded GW
gravel
21.0 21.8 No recovery
21.8 22.1 Gravel-Sand Mixtures/ well graded GW
gravel
22.1 25.1 Medium to coarse sde w big pebbles GP
and cobbles

 

 

 

 

Well id: MPS-8

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.4 No recovery
17.4 19.8 Medium to coarse sand/ more medium SW
19.8 21.0 Gravel-sand mixtures/ well graded GW
gravel
21.0 22.6 No recovery
22.6 23.0 Medium to coarse sand/ more coarse SW
sand
23.0 23.3 Gravel-Sand mixtures/ poorly graded GP
gravel
23.3 25.0 Medium to coarse sand /w big pebbles GP
and cobbles
25.0 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

143

 

 

 

 

Well id: MPS-9

 

 

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 17.1 No recovery

17.] 19.8 Medium to coarse sand/ more medium SW

19.8 20.4 Gravel-Sand mixtures/ well graded GW
gravel

20.4 21.0 Medium to coarse sand / more coarse SW

21.0 22.6 No recovery

22.6 22.9 Med to coarse sand/ more coarse SW

22.9 23.2 Gravel-sand mixtures/ poorly graded GP
gravel

23.2 25.0 Medium to coarse sand /w big pebbles GP
and cobbles

 

 

Well id: MPS-10

 

 

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 17.1 No recovery

17.1 19.8 Medium to coarse sand/ more medium SW

19.8 21.0 Gravel-Sand mixtures/ more gravel; GW
well graded gravel

21.0 22.9 No recovery

22.9 23.2 Medium to coarse sand SW

23.2 25.3 Medium to coarse sand /w big pebbles GP
and cobbles; material gravelly at the
bottom

25.3 25 .6 Gray clay, stiff, dry, cohesive CL

 

 

Well id: MPS-ll

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 18.0 No recovery

18.0 19.8 Medium to coarse sand/ more medium SW

19.8 21.0 Gravel-Sand mixtures/ well graded GW
gravel

21.0 22.4 No recovery

22.4 25.1 Medium to coarse sand/ with big GP
pebbles and cobbles

25.1 25.6 Gray clay, stiff, dry, cohesive CL

 

 

 

 

 

144

 

 

 

Well id: MPS-12

 

 

 

 

 

 

 

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.7 No recovery
17.7 19.7 Medium to coarse sand/ more medium SW
19.7 20.4 Gravel-Sand mixtures/ more gravel; GW
well graded gravel
20.4 21.0 Medium to coarse sand/ more coarse SW
21.0 22.4 No recovery
22.4 25.1 Medium to Coarse sand w/ fine gravel; GP
big pebbles and cobbles
25.1 25.6 Gray clay, stiff, dry, cohesive CL
Well id: MPS-13
Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 17.4 No recovery
17.4 19.7 Medium to coarse sde more medium SW
19.7 21.0 Gravel-Sand mixtures/ more gravel; GW
well graded graved
21.0 22.9 No recovery
22.9 25.6 Medium to coarse sand /w fine gravel; GP
big pebbles and cobbles

 

 

 

 

Well id: MPS-l4

 

 

 

Depth (m) Description Classification

From To

0.0 14.6 Blind drill to 14.9m

14.6 14.9 Blind drill to 14.9m

14.9 17.1 No recovery

17.1 19.8 Medium to coarse sand/ more medium SW

19.8 20.4 Gravel-Sand Mixtures / well graded GW
gravel

20.4 23.5 No recovery

23.5 24.1 Coarse Sand SW

24.1 25.6 Gravel Sand mixtures / poorly graded GP
gravel

 

 

 

 

 

145

 

 

 

Well id: MPS-lS

 

 

 

Depth (m) Description Classification
From To
0.0 14.6 Blind drill to 14.9m
14.6 14.9 Blind drill to 14.9m
14.9 18.3 No recovery
18.3 20.7 Medium to coarse Sand w/ fine gravel SW
20.7 21.0 Gravel-Sand Mixtures/ well graded GW
gravel
21.0 22.9 No recovery
22.9 24.7 Medium to coarse sand, more medium SW
24.7 25.3 Gravel-Sand mixtures; poorly graded GP
gravel
25.3 25.6 Medium to coarse sand/ w gravel and GP
big pebbles

 

 

 

 

146

 

 

APPENDIX B - MatLab SCRIPT FOR THE OPTIMIZATION OF AQUIFER
PARAMETERS

**********************************************************************

This function optimizes aquifer physical parameters given a series of tracer observations
in an aquifer

Created by: Jaime A Graulau (2001)

Please send comments to: graulaus@egr.msu.edu
*********************************************************************

function f = funtest(alpha)
persistent Cobs iter disper restfile layers

if (iter > 0)

else
% load the observed data file only one time
% and read relevant data for the first time from
% the dispersion file for MT3Dms....
iter = 0;
load observeddat
Cobs = observed;
%read and write to the dispersion file
fid = fopen('optim.dsp','rt');

%---change for the no. of layers in the model---
layers = 1
for k = 1:1ayers

disper(k,:)=fscanf(fid,'%f %f‘, [layers 2]);

end
restfile = fscanf(fid,'%c');
fclose(fid);

end

iter = iter + l;
%--- substitute the values of the x array and run MT3DMS_recirc2
for k=1 :layers

%if (9 <= k < 15) --- remove this coment when running the optimization for more
layers
disper(k,2) = alpha(1);
end

fid = fopen('optim.dsp','wt');
for k=1:layers
if (k==layers)

147

 

fprintf(fid,'%10.0f %9.7f,disper(k,:));
else
fprintf(fid,'%10.0f %9.7f\n',disper(k, :));
end
end
fprintf(fid,"’/oc',restfile);
fclose(fid);

% ----- Run MT3DM -----

!MT3D_recirc_imp optim.mts
OI.

modeldata = (read_obs_file('MT3D001.0BS INITIAL CO',1,-1))';
Ccalc = interp1(modeldata(:,2), modeldata(:,3),Cobs(:, 1 ));

sumtot = 0.0;

for i=1 :5
F conc(i)=abs(Cobs(i,2) - Ccalc(i));
sumtot = sumtot + F conc(i);

end
f = sumtot;
if (iter = 1)

fid = fopen('optim_rout.out','wt');
fprintf(fid, 'Iteration long dispers(ft) f \n');
fprintf(fid, ' --------- --- \n');
fclose(fid);
end
fid= fopen('optim_rout.out', 'at');
fprintf(fid, '%i %f %6.4f\n', iter, alpha(l), f);
fclose(fid);

 

148

 

APPENDIX C - RT3D USER DEFINED REACTION CODE FOR THE
EVALUATION OF REDUCTIVE DECHLORINATION LINKED
TO CARBON SOURCE DEGRADATION

SUBROUTINE rxns(ncomp,nvrxndataj,i,k,y,dydt,

& poros,rhob,reta,rc,nlay,nrow,ncol,vrc)
CIIIBlock l:*********#********************************************III*******
c List of calling arguments
c ncomp - Total number of components
c nvrxndata - Total number of variable reaction parameters to be input via RCT file
c J, I, K - node location (used if reaction parameters are spatially variable)
c y - Concentration value of all component at the node [array variable y(ncomp)]
c dydt - Computed RHS of your differential equation [array variable dydt(ncomp)]
c poros - porosity of the node
c reta - Retardation factor [ignore dummy reta values of immobile species]
c rhob - bulk density of the node
C re - Stores spatially constant reaction parameters (can dimension upto 100 values)
c nlay, nrow, ncol - Grid size (used only for dimensioning purposes)

c vrc - Array variable that stores spatially variable reaction parameters
ClllEnd OfBlOCk l****t**$**¢******###**********#*#******¥************t**it

CtBlock 2.**************************************************************

c* *Please do not modify this standard interface block“
!DEC$ ATTRIBUTES DLLEXPORT :: rxns
IMPLICIT NONE
INTEGER ncol,nrow,nlay
INTEGER ncomp,nvrxndata,j,i,k
INTEGER, SAVE :: First_time=1
DOUBLE PRECISION y,dydt,poros,rhob,reta
DOUBLE PRECISION rc,vrc
DIMENSION y(ncomp),dydt(ncomp),rc( 1 00)

DIMENSION vrc(ncol,nrow,nlay,nvrxndata),reta(ncomp)
CIkEnd OfblOCk 218********#*****#**********#*#**#*#***##*******#tt**Illll'flltlll

C*BlOCk 3.***********II#*****#*******************************************

c *Declare your problem-specific new variables here“

c INTEGER

cjgs DOUBLE PRECISION pce ,tce,dce,vc,kpce,ktce,kdce,kvc
c j gs DOUBLE PRECISION ytcepce,ydcetce,yvcdce

 

C

C Comment previous declaration statements and declare pertinent
C variables for the Plume G Model here

c

DOUBLE PRECISION tol, qtol, X, KsTOL

149

DOUBLE PRECISION tolsorb, Etol, KdTOL

DOUBLE PRECISION miu, Yxtol, b

DOUBLE PRECISION PCE, qPCE, KsPCE, Epce, PCEsorb, KdPCE

DOUBLE PRECISION TCE, qTCE, KsTCE, thepce, TCEsorb, KdTCE, Etce
DOUBLE PRECISION DCE, qDCE, KsDCE, Edce, DCEsorb, KdDCE, chetce
DOUBLE PRECISION VC, qVC, KsVC, Evc, VCsorb, KdVC, chdce

C end of declaration statements for Plume G particular Model
c

 

ClllEnd 0f blOCk 3************It*1!*****************************************

C*BlOCk 4.**********************************#************#**************

c *Initilize reaction parameters here, if required“
IF (First_time .EQ. 1) THEN
C
C Plume G particular variables
KsTOL = 3.50E+01 !half-veloc coeff. for toluene degradation [mg/L]
qtol = 3.00E-01 !Max spec. toluene utilization rate [l/d]
Yxtol = 1.52E+00 !Yield coefficient for pce degraders [mg VSS/ mg tol]
b = 1.00E-01 !Decay coefficient for toluene degraders [1/d]
Etol = 1.00E-01 !Toluene mass transfer coefficient [l/d]
KdTOL = 6.60E-07 !Toluene partitioning coefficient [L/mg]
qPCE = 4.0E-2 !Max specific PCE utilization rate [l/d]
KsPCE = 8.00E-03 !Half-veloc. coefficient for PCE dehalogenation [l/d]
Epce = 1.00E-01 !PCE mass transfer coeff [l/d]
KdPCE = 1.00E-07 !PCE partitioning coefficient [L/mg]
qTCE = 6.00E-02 !Max. specific TCE utilization rate [l/d]
thepce = 7.90E-01 !PCE to TCE yield coefficient [mg TCE/mg PCE]
KsTCE = 1.8E-01 !Half-velocity coeff. for TCE dehalogenation [mg/L]
Etce = 1.00E-01 !TCE mass transfer coefficient [l/d]
KdTCE = 9.00E-08 !TCE partitioning coefficient [L/kg]
qDCE = 9.00E-02 !Max. specific DCE utilization rate [l/d]
chetce = 5.70E-01 !TCE to DCE yield coefficient [mg DCE/mg TCE]
KsDCE = 2.88E-01 !Half velocity coefficient for DCE dehalogeation [mg/L]
Edce = 1.00E-01 !DCE mass transfer coefficient [l/d]
KdDCE = 5.00E-08 !DCE partitioning coefficient [L/kg]
qVC = 5.00E-03 !Max. specific VC utilization [l/d]
chdce = 6.50E-01 !DCE to VC yield coefficient [mg VC/mg DCE]
KsVC = 1.6lE-01 !Half-velocity coefficient for VC dehalogenation [mg/L]
Evc = 1.00E-01 !VC mass transfer coefficient [l/d]
KdVC = 2.00E-09 !VC partitioning coefficient [L/kg]
miu = qtol*Yxtol

 

! KsTOL = rc(l) !Half-veloc coeff. for toluene degradation [mg/L]
1 qtol = rc(2) !Max spec. toluene utilization rate [l/d]

150

 

C*BIOCk 5-******************#*******************************************

C
C

C

Yxtol = rc(3) !Yield coeff. for pce degraders [mg VSS/ mg tol]

b = rc(4) !Decay coefficient for microorganisms [l/d]

Etol = rc(5) !Toluene mass transfer coefficient [l/d]

KdTOL = rc(6) !Toluene partitioning coefficient [L/kg]

qPCE = rc(7) !Max specific PCE utilization rate [l/d]

KsPCE = rc(8) !Half-veloc. coefficient for PCE dehal. [l/d]

Epce = rc(9) !PCE mass transfer coeff [l/d]

KdPCE = rc(10) !PCE partitioning coefficient [L/kg]

qTCE = rc(l 1) !Max. specific TCE utilization rate [l/d]

thepce = rc(12) !PCE to TCE yield coefficient [mg TCE/mg PCE]
KsTCE = rc(13) !Half-velocity coefficient for TCE dehalog. [mg/L]
Etce = rc(14) !TCE mass transfer coefficient [l/d]

KdTCE = rc(15) !TCE partitioning coefficient [L/kg]

qDCE = rc(16) !Max. specific DCE utilization rate [l/d]

chetce = rc(l7) !TCE to DCE yield coefficient [mg DCE/mg TCE]
KsDCE = rc(18) !Half velocity coefficient for DCE dehalog. [mg/L]
Edce = rc(l9) !DCE mass transfer coefficient [l/d]

KdDCE = rc(20) !DCE partitioning coefficient [L/kg]

qVC = rc(21) !Max. specific VC utilization [l/d]

chdce = rc(22) !DCE to VC yield coefficient [mg VC/mg DCE]
KsVC = rc(23) lHalf-velocity coefficient for VC dehal. [mg/L]
Evc = rc(24) !VC mass transfer coefficient [l/d]

KdVC = rc(25) !VC partitioning coefficient [L/kg]

miu = rc(2)*rc(3) !Maximum specific bacterial growth rate [l/d]

end of variable initialization for Plume G model

 

First_time = 0 !reset First_time to skip this block later
END IF

CtEnd Of blOCk 43****************#****************#****************#****

*Assign or compute values for new variables, if required"
Assign the Plume G model statements in this sub-block

to] = y(l)

X = y(2)

PCE = y(3)
TCE = y(4)
DCE = y(5)

VC = y(6)
tolsorb = y(7)
PCEsorb = y(8)
TCEsorb = y(9)
DCEsorb = y(10)
VCsorb = y(l 1)

end of assignment statements for plume G model

151

C

 

CIFEnd OfblOCk 5**#*****************************************************

C*BlOCk 6-**************************************************************

c *Ditferential Reaction Equations"
C Plume G particular reactions here

dydt(] )=-qtol*X *(tol/(tol+KsTOL))-(Etol *rhob/poros)*
& (KdTOL*tol-tolsorb)

dydt(2)=miu*X*(tol/(tol+KsTOL))-b*X

dydt(3)=-qPCE*X*(PCE/(PCE+KsPCE))-(Epce*rhob/poros)*
& (KdPCE*PCE-PCEsorb)

dydt(4)=thepce*qPCE*X*(PCE/(PCE+KsPCE))-
qTCE*X*(TCE/(TCE+KsTCE))-
& (Etce*rhob/poros)*(TCE*KdTCE-TCEsorb)
dydt(5)=chetce*qTCE*X*(TCE/(TCE+KsTCE))-
qDCE*X*(DCE/(DCE+KsDCE))-
& (Edce*rhob/poros)*(DCE*KdDCE-DCEsorb)

dydt(6)=chdce*qDCE*X*(DCE/(DCE+KsDCE))-qVC*X*(VC/(VC+KsVC))-
& (Evc*rhob/poros)*(VC*KdVC-VCsorb)

dydt(7)=Etol*(tol*KdTOL-tolsorb)

dydt(8)=Epce*(PCE*KdPCE-PCEsorb)

dydt(9)=Etce*(TCE*KdTCE-TCEsorb)

dydt( 1 0)=Edce*(DCE* KdDCE-DCEsorb)
' dydt(l 1 )=Evc*(VC*KdVC-VCsorb)

C*End OfblOCk 6********************#****************Ill******************

RETURN
END

152

   

lllllllllllllllllljlllllIlllljllllll