NC0003468_DRSS CAP Part I_Appx C_Final_20151112

Appendix C
UNCC Groundwater
Flow and Transport
Model

This page intentionally left blank

Memorandum

October 8, 2015

TO: Ed Sullivan and Tyler Dubose
FROM: Bruce Hensel
SUBJECT: DAN RIVER MODEL REVIEW

Summary
EPRI has reviewed the Dan River model report and files provided by Duke Energy, HDR
Engineering, and the University of North Carolina-Charlotte. The review was performed by
John Ewing (Intera), with input by Chunmiao Zheng and myself. Based on this review, it is
our opinion that, subject to the caveats below, this model is set-up and meets its flow and
transport calibration objectives sufficiently to meet its final objective of predicting effects of
corrective action alternatives on groundwater quality. The caveats associated with this
opinion are:
 Constant heads used to represent some boundaries may provide an unmitigated source
of water for simulation of any corrective action alternatives that potentially involve
pumping near those boundaries. If corrective actions involve pumping, drains may be
a more appropriate boundary condition choice for the Dan River and unnamed
streams that use constant head boundaries.
 Boundary conditions representing the unnamed streams penetrate to layers beneath
the surficial layer 7. We understand that this will be corrected prior to use of the
model for evaluation of corrective actions.

Specific Comments
Model Report, Setup, and Calibration
a) Is the objective/purpose of modeling clearly defined?
Yes. The objective and purpose of the modeling is clearly defined in Section 1.2 as
consisting of three main activities: development of a calibrated steady-state flow model of
current conditions, development of a historical transient model of constituent transport
calibrated to current conditions, and predictive simulations of different corrective action
options.

b) Is the site description adequate?
Yes. Section 1.1 of the report provides a description of the site that is adequate for
evaluating the model for both flow and transport purposes.
Dan River Model Review
October 8, 2015
Page 2

c) Is the conceptual model well described with appropriate assumptions?
Yes. The conceptual model section contains subsections discussing the aquifer system
framework, the groundwater flow system, hydrologic boundaries, hydraulic boundaries,
sources and sinks, water budget, modeled constituents of interest, and constituent transport.
The CSA report is referenced in describing aquifer properties. The method used to estimate
recharge is discussed along with references. The rationale behind the choice of constituents
to be modeled is discussed.
d) Is the numerical model properly set up (steady state or transient; initial condition;
boundary conditions; parameterization; etc.)?
i) Appropriateness of the simulator
The use of MODFLOW-NWT and MT3DMS to simulate groundwater flow and contaminant
transport, respectively, are both appropriate for the modeling objectives and represent an
industry standard in choice of simulators.
ii) Discretization: temporal and spatial (x-y notably z)
The spatial discretization for this model appears more than adequate to delineate differences
in simulated heads and concentrations for all reasonable calibration and predictive purposes.
For transport simulations, the Courant Number and the Grid Peclet Number can be used to
determine whether discretization is numerically appropriate for a given model. In this model,
transport time steps are automatically calculated by the simulator based on a specified
Courant Number constraint of 1. This ensures that temporal discretization is appropriate for
the spatial discretization of the model.
The Grid Peclet Number (Pe_grid = ∆l/α) can be used to evaluate whether numerical
dispersion dominates constituent transport based on spatial discretization and physical
dispersion. Grid Peclet numbers preferably less than 2 and no more than 10 are generally
recommended to minimize numerical dispersion.
The lateral discretization (∆x = ∆y = 20 feet) for the entire model area, coupled with the
longitudinal dispersivity of 10 feet in the Dan_trace.dsp file results in a Grid Peclet number
of 2 which is adequate. The transverse lateral dispersivity of 1 foot results in a Grid Peclet
number of approximately 20 which may result in some unintended additional transverse
lateral transport of constituents however this is unlikely to influence model predictions in a
meaningful way.
The vertical discretization is variable in contrast to the horizontal discretization. Because a
considerable portion of the early transport in the model is vertical, special consideration of
this discretization is warranted. Spot checks of grid cells beneath the ash pond indicate cell
thicknesses (∆z) in layers 11 and 12 of approximately 40 feet. The transverse vertical
dispersivity (αv) is 1 foot in the model. This results in a grid Peclet number (Pe_grid = ∆z/αv)
equal to 40/1 = 40. The effect of this relatively large grid Peclet number is that numerical
dispersion will dominate over physical dispersion and constituent concentrations will
transport deeper in the model than intended purely due to the relatively large vertical
discretization. Since the model was calibrated to match shallow concentrations, any vertical
Dan River Model Review
October 8, 2015
Page 3

transfer of mass to deeper layers via numerical dispersion results in more mass input to the
model to make up for the mass transfer from shallow layers to deep layers than would have
otherwise occurred; therefore, this is conservative. As a result, predictions of time required to
achieve corrective action goals may also be conservatively long.
iii) Hydrologic framework – hydraulic properties
Hydraulic properties were grouped by material type (ash, dike, upper unconsolidated zone,
transition zone, and fractured bedrock zone) which are based on the hydrostratigraphy of the
site described in the CSA report. This approach is reasonable.
iv) Boundary conditions
The rate of areal recharge originating as precipitation is based on ranges for the Piedmont
ranges from a referenced study. This is appropriate. Flow to and from surface water
boundaries such as the Dan River to the south and an unnamed stream and service settling
water pond to the west are represented by constant head boundaries. The drain package was
used to simulate an unnamed stream along the eastern model boundary. Beneath these
surficial boundaries, a groundwater divide is assumed which is represented as no flow. A
topographical groundwater divide defines the northern boundary which is represented as no
flow. These boundaries are reasonable.
The recharge in the shallow flow system conceptualized as primarily discharging to the Dan
River along the southern boundary and the unnamed streams and the service settling water
pond along the eastern and western boundaries of the model. The constant heads may also
provide an unmitigated source of water for simulation of any corrective action alternatives
that potentially involve pumping near those boundaries, however, personal communication
with the UNCC team indicates that corrective actions involving pumping are unlikely. If
corrective actions do involve pumping, drains may be a more appropriate boundary condition
choice for the unnamed streams that use constant head boundaries.
It was noted that the boundary conditions representing the unnamed streams penetrate to
layers beneath the surficial layer 7 and personal communication with the UNCC team
indicates that this will be corrected.
v) Initial conditions in transient simulations
The flow model is run in steady state, where initial conditions are not relevant apart from
numerical convergence, which was achieved. The transient transport model assumes a zero
concentration in 1983 with the ash ponds acting as a constant concentration sources
thereafter. These initial conditions are reasonable.
vi) Convergence criteria and mass balance errors
The head tolerance in the NWT packages of 1e-3 is more than adequate in ensuring head
precision for the purposes of the model. The flow mass balance discrepancy in the
MODFLOW Listing file of -0.03 percent is more than adequate in ensuring minimal flow
mass balance errors. The cumulative concentration mass balance discrepancy from
Dan River Model Review
October 8, 2015
Page 4

MT3DMS is consistently around 6e-2 percent for the transport simulation which is more than
adequate in ensuring minimal constituent mass balance errors.
e) Is the calibration done properly and adequately?
For flow calibration to heads, the Mean Absolute Error over the range in observed values is 9
percent, which is below the industry standard of 10 percent (i.e., below 10 percent is
acceptable). Comparison of simulated and observed arsenic concentrations shows reasonable
agreement. Beneath the ash storage units and up-gradient of the ash basin simulated boron
concentrations are zero with observed concentrations above background. The model results
are because no source concentrations were applied to the ash storage units. Between the ash
basin and the Dan River, simulated boron concentrations are consistent with observed values
at all depths. This indicates that the model is adequately calibrated to assess corrective
actions between the ash basin and the Dan River but will not provide information on
corrective actions up-gradient of the ash basin.
i) Property/boundary condition correlation – parameter bounds
Recharge and hydraulic properties were adjusted simultaneously during calibration. These
parameters are correlated with various parameter combinations potentially being able to
match observed heads. However, hydraulic properties were confined to the ranges of
measured values from the CSA and a hierarchy in recharge rates, whereby infiltration was
higher in the ash basins than the areas outside the basins, was enforced. Based on the
recharge package file (Dan_trace.rch), the calibrated recharge rates were 6 inches per year
within the ash basins and 5 inches per year outside the ash basins as shown in Figure 3. This
approach appears reasonable.
ii) Discretization of calibration parameters
Hydraulic properties were grouped by material type (ash, dike, upper unconsolidated zone,
transition zone, and fractured bedrock zone). This is a reasonable approach. Recharge is
grouped by infiltration within the ash basins and outside the basins. This is also a reasonable
approach.
iii) Appropriateness of target as a metric of simulation objectives (e.g., calibrating
primarily to heads when transport is the primary purpose)
The flow model calibration to observed heads is adequate. Several plots also show simulated
constituent concentrations in the context of several observed concentrations. The
concentrations of arsenic and boron are adequately calibrated for the area between the ash
basin and the Dan River and can be used to assess corrective actions in that area.
f) Is the sensitivity analysis conducted and if so, correctly?
i) Sensitivity Analysis approach (look for parameters which may be insensitive to flow
but not to transport)
A flow model sensitivity analysis was conducted whereby recharge rates and hydraulic
conductivities in the transport zone were perturbed up and down. It is stated that water levels
rise with increasing recharge and decreasing transport zone hydraulic conductivity and the
Dan River Model Review
October 8, 2015
Page 5

results are depicted in figures 14 and 15 for each of the monitoring wells. These figures are
absent in version 2 of the report but were included at figures 7a and 7b in version 1 of the
report. Figure 7b shows units of hydraulic conductivity of cm/s however, the discretization
package (Dan_trace.dis) parameters ITMUNI and LENUNI indicate the time and length units
are days and feet, respectively. It is suggested that these figures be included and with units
consistent to that of the model. This shows the sensitivity of the head targets to 2 sets of flow
parameters.
For the transport model, two Kd values were simulated to show the sensitivity of Arsenic to
Kd. This could be considered a minimal sensitivity analysis compared to simulating the
sensitivity of the model to every model parameter. However, it does concisely summarize
the sensitivity of the model to the key flow and transport parameters.
Model Files:
a) Can the model be run with the input files provided by the developer?
Yes.
b) Do the model results match those presented in the report?
i) Independent check of input data vs. conceptual model/report
Model inputs are consistent with what is described in the report with one notable exception.
Section 5.3 of the report indicates a transport simulation period for Boron of 64 years and a
period for Arsenic of 32 years. The same section discusses Arsenic results with a sorption
coefficient of Kd=1 at 64 years. The basic transport package delivered (Dan_trace.btn)
indicates a transport period of 32 years.
ii) Check of water balance vs. conceptual model
The MODFLOW listing file indicates that 92% of the inflow to the flow model enters as
recharge with the remaining 8% of inflow entering through constant head boundaries. 57%
of the outflow exits through constant heads representing the Dan River along with unnamed
streams and the service settling water pond along the western boundary. The remaining 43%
of the outflow exits through drains representing surficial ditches and streams. This is all
consistent with the conceptual model.
iii) Independent check of model results vs. those reported
Independently generated model results are consistent with what is described in the report.

This page intentionally left blank

Groundwater Flow and Transport Model
Dan River Steam Station
Rockingham County, NC

Prepared for:
HDR Engineering, Inc.
Hydropower Services
440 S. Church St, Suite 1000
Charlotte, NC 28202
Investigators:
William G. Langley, Ph.D., P.E.
Dongwook Kim, Ph.D.

UNC Charlotte / Lee College of Engineering
Department of Civil and Environmental Engineering
EPIC Building 3252
9201 University City Blvd.
Charlotte, NC 28223

November 6, 2015

TABLE OF CONTENTS
1 Introduction ............................................................................................................................ 1
1.1 General Setting and Background ................................................................................... 1
1.2 Study Objectives ............................................................................................................. 1
2 Conceptual Model .................................................................................................................. 2
2.1 Geology and Hydrogeology (HDR 2015) ........................................................................ 2
2.2 Groundwater Flow System ............................................................................................. 2
2.3 Hydrologic Boundaries ................................................................................................... 4
2.4 Hydraulic Boundaries ..................................................................................................... 4
2.5 Water Sources and Sinks ............................................................................................... 4
2.6 Water Budget .................................................................................................................. 4
2.7 Modeled Constituents of Interest (COI) .......................................................................... 5
2.8 Constituent Transport ..................................................................................................... 5
3 Computer Model .................................................................................................................... 6
3.1 Model Selection .............................................................................................................. 6
3.2 Model Description ........................................................................................................... 6
4 Groundwater Flow and Transport Model Construction .......................................................... 6
4.1 Model Hydrostratigraphy ................................................................................................ 7
4.2 GMS MODFLOW Version 10 ......................................................................................... 8
4.3 Model Domain and Grid .................................................................................................. 9
4.4 Hydraulic Parameters ................................................................................................... 10
4.5 Flow Model Boundary Conditions ................................................................................. 10
4.6 Flow Model Sources and Sinks .................................................................................... 11
4.7 Flow Model Calibration Targets .................................................................................... 11
4.8 Transport Model Parameters ........................................................................................ 11
4.9 Transport Model Boundary Conditions ......................................................................... 12
4.10 Transport Model Sources and Sinks ............................................................................ 12
4.11 Transport Model Calibration Targets ............................................................................ 13
5 Model Calibration to Current Conditions .............................................................................. 13
5.1 Flow Model Residual Analysis ...................................................................................... 13
5.2 Flow Model Sensitivity Analysis .................................................................................... 14
5.3 Transport Model Calibration and Sensitivity ................................................................. 14
6 Predictive Simulations of Source Removal Scenarios ......................................................... 16
6.1 Existing Conditions Scenario ........................................................................................ 17
6.2 Cap-in-Place Scenario .................................................................................................. 20
6.3 Excavation Scenario ..................................................................................................... 23
7 Summary ............................................................................................................................. 26
7.1 Model Assumptions and Limitations ............................................................................. 26
7.2 Model Predictions ......................................................................................................... 26 
TABLES
Table 1 MODFLOW-NWT and MT3DMS Input Packages Used
Table 2 Model Hydraulic Conductivity
Table 3 Measured vs. Modeled Water Levels
Table 4 Flow Parameter Sensitivity Analysis
Table 5 Measured vs. Modeled Arsenic, Boron, Sulfate, and Chromium Concentrations
Table 6 Arsenic Transport Parameter Sensitivity Analysis-Sorption
Table 7 Boron Transport Parameter Sensitivity Analysis-Porosity
Table 8 Boron Transport Parameter Sensitivity Analysis-Dispersivity
Table 9 Transport Model Parameters and Calibration Results

FIGURES
Figure 1 Conceptual Groundwater Flow Model/Model Domain
Figure 2 Model Domain Cross section A-A’
Figure 3 Model Domain Cross section B-B’
Figure 4 Flow model boundary conditions
Figure 5 Recharge and constant concentration boundary condition
Figure 6 Shallow Observation Wells
Figure 7 Deep Observation Wells
Figure 8 Bedrock Observation Wells
Figure 9 Hydraulic Conductivity Zonation in S/M1/M2 Model Layers
Figure 10 Measured versus modeled water levels
Figure 11 Water head contour map at Layer 7
Figure 12 Water head at cross section C-C’
Figure 13 Water head cross section D-D’
Figure 14 Predicted Arsenic (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 15 Predicted Arsenic (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 16 Predicted Arsenic (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 17 Initial (2015) Arsenic Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 18 Initial (2015) Arsenic Concentrations (μg/L) in the Deep Groundwater Zone
Figure 19 Initial (2015) Arsenic Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 20 Existing Conditions Scenario 1 - 2115 Predicted Arsenic Concentrations (μg/L) in
the Shallow Groundwater Zone
Figure 21 Existing Conditions Scenario 1 - 2115 Predicted Arsenic Concentrations (μg/L) in
the Deep Groundwater Zone
Figure 22 Existing Conditions Scenario 1 - 2115 Predicted Arsenic Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 23 Cap-in-Place Scenario 2 - 2115 Predicted Arsenic Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 24 Cap-in-Place Scenario 2 - 2115 Predicted Arsenic Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 25 Cap-in-Place Scenario 2 - 2115 Predicted Arsenic Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 26 Excavation Scenario 3 - 2115 Predicted Arsenic Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 27 Excavation Scenario 3 - 2115 Predicted Arsenic Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 28 Excavation Scenario 3 - 2115 Predicted Arsenic Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 29 Predicted Boron (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 30 Predicted Boron (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 31 Predicted Boron (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 32 Initial (2015) Boron Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 33 Initial (2015) Boron Concentrations (μg/L) in the Deep Groundwater Zone
Figure 34 Initial (2015) Boron Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 35 Existing Conditions Scenario 1 - 2115 Predicted Boron Concentrations (μg/L) in
the Shallow Groundwater Zone
Figure 36 Existing Conditions Scenario 1 - 2115 Predicted Boron Concentrations (μg/L) in
the Deep Groundwater Zone
Figure 37 Existing Conditions Scenario 1 - 2115 Predicted Boron Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 38 Cap-in-Place Scenario 2 - 2115 Predicted Boron Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 39 Cap-in-Place Scenario 2 - 2115 Predicted Boron Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 40 Cap-in-Place Scenario 2 - 2115 Predicted Boron Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 41 Excavation Scenario 3 - 2115 Predicted Boron Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 42 Excavation Scenario 3 - 2115 Predicted Boron Concentrations (μg/L) in the Deep
Groundwater Zone
Figure 43 Excavation Scenario 3 - 2115 Predicted Boron Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 44 Predicted Chromium (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 45 Predicted Chromium (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 46 Predicted Chromium (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 47 Initial (2015) Chromium Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 48 Initial (2015) Chromium Concentrations (μg/L) in the Deep Groundwater Zone
Figure 49 Initial (2015) Chromium Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 50 Existing Conditions Scenario 1 - 2115 Predicted Chromium Concentrations (μg/L)
in the Shallow Groundwater Zone
Figure 51 Existing Conditions Scenario 1 - 2115 Predicted Chromium Concentrations (μg/L)
in the Deep Groundwater Zone
Figure 52 Existing Conditions Scenario 1 - 2115 Predicted Chromium Concentrations (μg/L)
in the Bedrock Groundwater Zone
Figure 53 Cap-in-Place Scenario 2 - 2115 Predicted Chromium Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 54 Cap-in-Place Scenario 2 - 2115 Predicted Chromium Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 55 Cap-in-Place Scenario 2 - 2115 Predicted Chromium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 56 Excavation Scenario 3 - 2115 Predicted Chromium Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 57 Excavation Scenario 3 - 2115 Predicted Chromium Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 58 Excavation Scenario 3 - 2115 Predicted Chromium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 59 Predicted Chromium VI (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 60 Predicted Chromium VI (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 61 Predicted Chromium VI (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 62 Initial (2015) Chromium VI Concentrations (μg/L) in the Shallow Groundwater
Zone
Figure 63 Initial (2015) Chromium VI Concentrations (μg/L) in the Deep Groundwater Zone
Figure 64 Initial (2015) Chromium VI Concentrations (μg/L) in the Bedrock Groundwater
Zone
Figure 65 Existing Conditions Scenario 1 - 2115 Predicted Chromium VI Concentrations
(μg/L) in the Shallow Groundwater Zone
Figure 66 Existing Conditions Scenario 1 - 2115 Predicted Chromium VI Concentrations
(μg/L) in the Deep Groundwater Zone
Figure 67 Existing Conditions Scenario 1 - 2115 Predicted Chromium VI Concentrations
(μg/L) in the Bedrock Groundwater Zone
Figure 68 Cap-in-Place Scenario 2 - 2115 Predicted Chromium VI Concentrations (μg/L) in
the Shallow Groundwater Zone
Figure 69 Cap-in-Place Scenario 2 - 2115 Predicted Chromium VI Concentrations (μg/L) in
the Deep Groundwater Zone
Figure 70 Cap-in-Place Scenario 2 - 2115 Predicted Chromium VI Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 71 Excavation Scenario 3 - 2115 Predicted Chromium VI Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 72 Excavation Scenario 3 - 2115 Predicted Chromium VI Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 73 Excavation Scenario 3 - 2115 Predicted Chromium VI Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 74 Predicted Cobalt (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 75 Predicted Cobalt (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 76 Predicted Cobalt (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 77 Initial (2015) Cobalt Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 78 Initial (2015) Cobalt Concentrations (μg/L) in the Deep Groundwater Zone
Figure 79 Initial (2015) Cobalt Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 80 Existing Conditions Scenario 1 - 2115 Predicted Cobalt Concentrations (μg/L) in
the Shallow Groundwater Zone
Figure 81 Existing Conditions Scenario 1 - 2115 Predicted Cobalt Concentrations (μg/L) in
the Deep Groundwater Zone
Figure 82 Existing Conditions Scenario 1 - 2115 Predicted Cobalt Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 83 Cap-in-Place Scenario 2 - 2115 Predicted Cobalt Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 84 Cap-in-Place Scenario 2 - 2115 Predicted Cobalt Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 85 Cap-in-Place Scenario 2 - 2115 Predicted Cobalt Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 86 Excavation Scenario 3 - 2115 Predicted Cobalt Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 87 Excavation Scenario 3 - 2115 Predicted Cobalt Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 88 Excavation Scenario 3 - 2115 Predicted Cobalt Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 89 Predicted Sulfate (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 90 Predicted Sulfate (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 91 Predicted Sulfate (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 92 Initial (2015) Sulfate Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 93 Initial (2015) Sulfate Concentrations (μg/L) in the Deep Groundwater Zone
Figure 94 Initial (2015) Sulfate Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 95 Existing Conditions Scenario 1 - 2115 Predicted Sulfate Concentrations (μg/L) in
the Shallow Groundwater Zone
Figure 96 Existing Conditions Scenario 1 - 2115 Predicted Sulfate Concentrations (μg/L) in
the Deep Groundwater Zone
Figure 97 Existing Conditions Scenario 1 - 2115 Predicted Sulfate Concentrations (μg/L) in
the Bedrock Groundwater Zone
Figure 98 Cap-in-Place Scenario 2 - 2115 Predicted Sulfate Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 99 Cap-in-Place Scenario 2 - 2115 Predicted Sulfate Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 100 Cap-in-Place Scenario 2 - 2115 Predicted Sulfate Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 101 Excavation Scenario 3 - 2115 Predicted Sulfate Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 102 Excavation Scenario 3 - 2115 Predicted Sulfate Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 103 Excavation Scenario 3 - 2115 Predicted Sulfate Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 104 Predicted Thallium (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 105 Predicted Thallium (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 106 Predicted Thallium (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 107 Initial (2015) Thallium Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 108 Initial (2015) Thallium Concentrations (μg/L) in the Deep Groundwater Zone
Figure 109 Initial (2015) Thallium Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 110 Existing Conditions Scenario 1 - 2115 Predicted Thallium Concentrations (μg/L)
in the Shallow Groundwater Zone
Figure 111 Existing Conditions Scenario 1 - 2115 Predicted Thallium Concentrations (μg/L)
in the Deep Groundwater Zone
Figure 112 Existing Conditions Scenario 1 - 2115 Predicted Thallium Concentrations (μg/L)
in the Bedrock Groundwater Zone
Figure 113 Cap-in-Place Scenario 2 - 2115 Predicted Thallium Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 114 Cap-in-Place Scenario 2 - 2115 Predicted Thallium Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 115 Cap-in-Place Scenario 2 - 2115 Predicted Thallium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 116 Excavation Scenario 3 - 2115 Predicted Thallium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 117 Excavation Scenario 3 - 2115 Predicted Thallium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 118 Excavation Scenario 3 - 2115 Predicted Thallium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 119 Predicted Vanadium (μg/L) in Monitoring Well A for Model Scenarios 1-3
Figure 120 Predicted Vanadium (μg/L) in Monitoring Well B for Model Scenarios 1-3
Figure 121 Predicted Vanadium (μg/L) in Monitoring Well C for Model Scenarios 1-3
Figure 122 Initial (2015) Vanadium Concentrations (μg/L) in the Shallow Groundwater Zone
Figure 123 Initial (2015) Vanadium Concentrations (μg/L) in the Deep Groundwater Zone
Figure 124 Initial (2015) Vanadium Concentrations (μg/L) in the Bedrock Groundwater Zone
Figure 125 Existing Conditions Scenario 1 - 2115 Predicted Vanadium Concentrations (μg/L)
in the Shallow Groundwater Zone
Figure 126 Existing Conditions Scenario 1 - 2115 Predicted Vanadium Concentrations (μg/L)
in the Deep Groundwater Zone
Figure 127 Existing Conditions Scenario 1 - 2115 Predicted Vanadium Concentrations (μg/L)
in the Bedrock Groundwater Zone
Figure 128 Cap-in-Place Scenario 2 - 2115 Predicted Vanadium Concentrations (μg/L) in the
Shallow Groundwater Zone
Figure 129 Cap-in-Place Scenario 2 - 2115 Predicted Vanadium Concentrations (μg/L) in the
Deep Groundwater Zone
Figure 130 Cap-in-Place Scenario 2 - 2115 Predicted Vanadium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 131 Excavation Scenario 3 - 2115 Predicted Vanadium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 132 Excavation Scenario 3 - 2115 Predicted Vanadium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 133 Excavation Scenario 3 - 2115 Predicted Vanadium Concentrations (μg/L) in the
Bedrock Groundwater Zone
Figure 134 Conceptual Groundwater Flow Model/Model Domain-Complete Excavation
Scenario

1 

1 Introduction
Duke Energy owns and formerly operated the Dan River Steam Station, located on a 380-acre
tract on the Dan River in Rockingham County near Eden, North Carolina. DRSS began
operation as a coal-fired generating station in 1949 and was retired from service in 2012. The
Dan River Combined Cycle Station (DRCCS) natural gas generating facility was constructed at
the site and began operations in 2012. Historically, coal ash residue from DRSS’s coal
combustion process was disposed of in the ash basin system located northeast of the station
and adjacent to the Dan River as shown in Figure 2-1 from the Comprehensive Site
Assessment Report (CSA) for the Dan River Steam Station (HDR 2015). The CSA report
contains extensive detail and data related to most aspects of the site groundwater model
reported on in this document. (References to figures and tables from the CSA cited in this report
are indicated as in the following example: HDR Figure 2-1 refers to Figure 2-1 from the CSA).

1.1 General Setting and Background

DRSS is a former coal-fired electricity generating facility along the Dan River. The three-unit
station began commercial operation in 1949 with operation of a single coal-fired unit (Unit 1)
with a second unit (Unit 2) being added in 1950. A third unit was added by 1955 and resulted in
a total installed capacity of 276 MW. All three coal-fired units, along with three 28 MW oil-fired
combustion turbine units, were retired in 2012. The DRCCS, a 620 MW combined cycle natural
gas facility, began commercial operations on December 10, 2012.

The natural topography at the DRSS site generally slopes from northwest to southeast and
ranges from an approximate high elevation of 606 feet near the northern property boundary just
west of Edgewood Road to an approximate low elevation of 482 feet at the interface with the
Dan River. Ground surface elevation varies about 124 feet over an approximate distance of 0.7
miles. Surface water drainage generally follows site topography and flows from the northwest to
the southeast across the site except where drainage patterns have been modified by the ash
basins or other construction. A site features map is shown on HDR Figure 2-4.

Groundwater in the shallow aquifer under the DRSS ash basin and under the ash storage areas
flows horizontally to the southeast and discharges to the Dan River. This flow direction is away
from the direction of the nearest public or private water supply wells. The Dan River serves as a
lower hydrologic boundary for groundwater within the shallow aquifer, and serves as a
discharge feature for shallow groundwater flow from the ash basin. Regional groundwater flow
in the vicinity of the DRSS is south/southeast toward the Dan River.

1.2 Study Objectives

The purpose of this study is to predict the groundwater flow and constituent transport that will
occur as a result of different possible corrective actions at the site. The study consists of three
2 

main activities: development of a calibrated steady-state flow model of site conditions observed
in June and July 2015 (HDR 2015), development of a historical transient model of constituent
transport that is calibrated to conditions at this time, and predictive simulations of the source
removal options.

2 Conceptual Model
The site conceptual model for DRSS is primarily based on the CSA Report (HDR 2015).

2.1 Geology and Hydrogeology (HDR 2015)

Geology at the DRSS site is comprised of two interconnected layers, or mediums: 1) residual
soil/saprolite and weathered fractured rock (regolith) overlying 2) fractured bedrock (HDR Figure
5-3). The regolith layer is a thoroughly weathered and structureless residual soil that occurs
near the ground surface with the degree of weathering decreasing with depth. The residual soil
grades into saprolite, a coarser grained material that retains the structure of the parent bedrock.
Beneath the saprolite, partially weathered/fractured bedrock occurs with depth until sound
bedrock is encountered. The site is underlain by rocks of the Pine Hall and Cow Branch
Formations. Alluvial and terrace deposits consisting of unconsolidated sand, silt, and clay with
occasional sub-rounded to well-rounded pebbles occur along the Dan River and major
tributaries. A transition zone (TZ) at the base of the regolith has been interpreted to be present
in many areas of the Piedmont and is present in the Dan River Basin. The TZ zone consists of
partially weathered/fractured bedrock and lesser amounts of saprolite that grade into bedrock
and may serve as a conduit of rapid flow and transmission of water (HDR 2015).

The groundwater system is a two medium system generally restricted to the local drainage
basin. The groundwater occurs in a system composed of two interconnected layers described
above: residual soil/saprolite and weathered rock overlying fractured metamorphic rock
separated by a transition zone (TZ). Typically, the residual soil/saprolite is partially saturated
and the water table fluctuates within it. Water movement is generally preferential through the
weathered/fractured bedrock of the TZ (i.e., zone of higher horizontal permeability). The
character of the system results from the combined effects of the rock type, fracture system,
topography, and weathering. Topography exerts an influence on both weathering and the
opening of fractures, while the weathering of the crystalline rock modifies both transmissive and
storage characteristics. Ranges of hydrostratigraphic layer properties measured at DRSS are
provided in HDR Tables 11-6, 11-7, and 11-8.

2.2 Groundwater Flow System

Groundwater is recharged by infiltration where the ground surface is permeable, including the
dikes and ash of the ash basin system where exposed at the ground surface. After infiltrating
the ground surface, water in the unsaturated zone percolates downward to the unconfined water
3 

table, except where ponded water conditions exists on a portion on the secondary cell of the
ash basin system. From the water table, groundwater moves downward and laterally through
unconsolidated material (residual soil/saprolite) into the weathered, fractured rock, then into
fractured bedrock. Mean annual recharge in the Piedmont ranges from 4.0 to 9.7 inches per
year (Daniel 2001).

Historical and current information about the DRSS ash basin system assembled by HDR (2015)
is relevant to developing the conceptual and numerical groundwater flow models. Refer to HDR
Figures 2-1 and 2-2 for locations of ash basin system components.

The current ash basin was constructed in three phases beginning in 1956 and ending in 1976.
Based on a review of historical construction drawings, it appears that the dams were
constructed of earthen materials borrowed from within the ash basin footprint.

The original ash basin was constructed in 1956 with a footprint that lies within the current
Primary Cell. The dam was constructed as a homogenous sandy silt embankment with an
approximate crest elevation of 525 feet. In 1968 and 1969, the footprint of the ash basin was
expanded to the footprint of the current Primary and Secondary Cells by constructing an earthen
dam with a crest elevation of 530 feet. At this period, the ash basin was a single impoundment.

In 1976, the single basin was divided into the current Primary and Secondary Cells by
construction of an interior dike on top of existing ash. This divider dike was constructed as an
earthen embankment at elevation 540 feet. The embankment for what is now the Primary Cell
was raised to 540 feet.

The area contained within the waste boundary for the Primary and Secondary Cells currently
extends over an area of approximately 53 acres. The Primary Cell extent is approximately 28.1
acres and contains approximately 1 million tons of ash material. The Primary Cell has a crest
elevation of approximately 540 feet and has an impoundment surface area of approximately
21.8 acres. The Secondary Cell extent is approximately 15.3 acres and contains approximately
200,000 tons of ash material. The Secondary Cell has a crest elevation of approximately 530
feet and has a current impoundment surface area of approximately 12.2 acres. The elevation of
the Dan River adjacent to the ash basin is approximately 482 feet.

During operation of the coal-fired units, the ash basin received variable inflows of fly ash, bottom
ash, pyrites, stormwater runoff (including runoff from the coal pile), cooling water, powerhouse
floor drains, sanitary waste effluent, station yard drainage sump, and boiler chemical cleaning
wastes. Flow was historically routed from the Primary Cell to the Secondary Cell through a
concrete discharge tower. Since the February 2, 2014, all inflows into the ash basin are now
routed directly to the Secondary Cell. Effluent from the Secondary Cell is routed to the Dan
River via a concrete discharge tower located in the Secondary Cell. The water surface in both
the Primary and Secondary Cells is controlled by the use of stop logs.
4 

Given that it has not been measured or estimated in the CSA or other studies of the site,
recharge to the ash basin system has been estimated by considering it as a calibration
parameter in the groundwater flow model.

Groundwater under the DRSS site flows horizontally generally toward the south and discharges
to the Dan River. Lesser amounts of groundwater discharge to unnamed streams and the
service settling water pond to the east and west of the site. According to the Piedmont Slope
Aquifer System of LeGrand (2004) depicted in HDR Figure 5-5, bedrock fractured density
decreases with depth, limiting deep groundwater flow.

2.3 Hydrologic Boundaries

The major discharge locations for the groundwater system at DRSS, the Dan River, and
unnamed streams and the service settling water pond to the east and west, serve as hydrologic
boundaries. Local ditches and streams also serve as local, shallow hydrologic boundaries. In
the flow model, these smaller features are treated as internal water drains.

2.4 Hydraulic Boundaries

The groundwater flow system at the DRSS study area does not contain impermeable barriers or
boundaries with the exception of bedrock at depth where fracture density is minimal. Natural
groundwater divides exist along topographic divides, but are a result of local flow conditions as
opposed to a barriers.

2.5 Water Sources and Sinks

Recharge, including that to the ash basins, is the major source of water to the groundwater
system. Most of this water discharges to the hydrologic boundaries described above. Three
private water supply wells and no public water supply wells within a half-mile radius of the site
were identified in the CSA (HDR 2015). The DRSS study area is not considered to be within the
capture zone or zone of influence of any extraction well.

2.6 Water Budget

Over an extended period of time, the rate of water inflow to the study area is equal to its rate of
water ouflow. That is, there is no change in groundwater storage. Water enters the groundwater
system through recharge and ultimately discharges to the Dan River, unnamed streams and the
service settling water pond to the east and west, and other small-scale discharge locations.
Recharge to the ash basins represents the summation of precipitation, evaporation,
evapotranspiration, plant wastewater discharge, and discharge through the outlet structures.

5 

2.7 Modeled Constituents of Interest (COI)

As defined in the CSA, constituents are those chemicals or compounds that were identified in
the approved groundwater assessment plans for sampling and analysis (HDR 2015). If a
constituent exceeded its respective regulatory standard or screening level in the medium in
which it was found, the constituent was then termed a Constituent of Interest (COI). The CSA
found the constituents of interest (COIs) in ash basin porewater are antimony, arsenic, barium,
beryllium, chromium, cobalt, iron, lead, manganese, thallium, and vanadium.

The COIs attributable to ash handling (arsenic, boron, chromium, and sulfate) were considered
in the transport simulations for Existing Conditions. Boron and sulfate occur at elevated levels in
water infiltrating from ash basin systems, and are considered mobile in groundwater as neither
precipitates or adsorb to soils to any significant extent. Arsenic also occurs at elevated levels in
infiltrating water, but is adsorbed to commonly occurring soil types. Chromium adsorbs to a
lesser extent than arsenic.

2.8 Constituent Transport

Chemical constituents enter the ash basin system in the dissolved phase and solid phase of the
station’s wastewater discharge. Some constituents are also present in native soils and
groundwater beneath the basin. In the ash basin, constituents may incur phase changes
including dissolution, precipitation, adsorption, and desorption. Dissolved phase constituents
may incur these phase changes as they are transported in groundwater flowing downgradient
from the basin. In the fate and transport model, chemical constituents enter the basin in the
dissolved phase by specifying a steady state concentration in the ash pore water. Phases
changes (dissolution, precipitation, adsorption, and desorption) are collectively taken into
account by specifying a linear sorption coefficient Kd. The accumulation and subsequent
release of chemical constituents in the ash basin over time is a complex process. In the
conceptual fate and transport model, it was assumed that the entry of constituents into the ash
basin is represented by a constant concentration in the saturated zone of the basin, which is
continually flushed by infiltrating recharge from above.

At the ash storage areas, leachable constituents enter the dissolved phase during transient
infiltration events through the stored ash. Infiltration with constituents in solution moves
downward through the stored ash and underlying, unsaturated soils and finally into groundwater
at the water table. Dissolved phase constituents may incur phase changes also as they migrate
through the stored ash and native soils to the water table. In the conceptual fate and transport
model, it was assumed that the entry of constituents into groundwater beneath the ash storage
areas is represented by a constant concentration at the water table beneath the ash storage
areas, which is continually flushed by infiltrating recharge from above.

6 

3 Computer Model
3.1 Model Selection

The computer code MODFLOW solves the system of equations that quantify the flow of
groundwater in three dimensions. MODFLOW can simulate steady state and transient flow, as
well as confined and water table conditions. Additional components of groundwater can be
considered including pumping wells, recharge, evapotranspiration, rivers, streams, springs, and
lakes. The information assembled in the conceptual site model is translated into its numerical
equivalent from which a solution is generated by MODFLOW.

3.2 Model Description

The specific MODFLOW package chosen for the study is NWT - a Newton formulation of
MODFLOW-2005 that is specifically designed for improving the stability of solutions involving
drying and re-wetting under water table conditions (Niswonger, et al. 2011). The numerical code
selected for the transport model is MT3DMS (Zheng and Wang 1999). MT3DMS is multi-
species 3-dimensional transport model that can simulate advection, dispersion/diffusion, and
chemical reaction of COIs in groundwater flow systems and has a package that provides a link
to the MODFLOW codes. The MODFLOW-NWT and MT3DMS input packages used to create
the groundwater flow and transport model, and a brief description of their use, are provided in
Table 1.

4 Groundwater Flow and Transport Model Construction
The flow and transport model for the study was developed through a multi-step processes. First,
a 3D model of the site hydrostratigraphy was constructed based on historical site construction
drawings and field data. Next, the model domain was determined, from which a numerical was
produced. Flow parameters, assigned to the numerical grid, were adjusted during the steady-
state flow model calibration process. Once the steady-state flow model was calibrated, a
transient transport simulation for the selected COIs was calibrated by adjusting transport
parameters to best match the current day conditions.

Three terrain surface models for RBSS were created using geographic information systems
(GIS) software: 1) current existing surface, 2) pre-construction surface without ash and ash
basin dikes, and 3) pre-construction surface with dikes but without ash. An interpolation tool in
ArcGIS 10.3 software was used to generate the terrain surfaces as a raster datasets with 20-
foot cells. Each surface was created to cover the extent of the groundwater model domain.

7 

4.1 Model Hydrostratigraphy

The model hydrostratigraphy was developed using historical site construction drawings and
borehole data to construct three-dimensional surfaces representing contacts between
hydrostratigraphic units with properties provided in HDR Tables 11-6 through 11-10.

1) Existing Ground Surface
Topographic and bathymetry elevation contours and spot elevations were produced from
surveys conducted in 2014. Since these surveys did not cover the entire model extent, elevation
data extracted as spot elevations from the North Carolina Floodplain Mapping Program’s 2010
LiDAR elevation data were used for the areas surrounding the surveys. At DRSS, simplified
elevation contours were digitized along the river channels to depress the surface a small
amount below water level.

2) Pre-construction Surface
Elevation contours of the original ground surface were digitized in CAD from engineering
drawings supplied by Duke Energy. These data were imported into GIS, and georeferenced.
These contours were trimmed to the areas underlying ash basins, dams, and dikes and ash
storage areas. The source data used in the existing surface were then replaced by the original
surface data where there was overlap. Elevation data from coal storage areas were removed.
The pre-construction surface was then created using the combination of original surface
elevations, 2014 survey elevations, and 2010 LiDAR elevations.

3) Pre-construction Surface with Dikes
Surface models of the ash basin dams and dikes were constructed from crest elevations as
determined from the 2014 survey and slopes given on the engineering drawings. Only the
sections of the dams and dikes facing the ash basins were modeled in this way. The 2014
survey data were used for dike/dam crests and outwardly facing surfaces. These surfaces were
merged with the pre-construction surface. These GIS data sets were exported into formats
readable by RockWorks and GMS MODFLOW.

4) 3-D Hydrostratigraphic Grids
The natural materials in the CSA boreholes and existing boreholes were assigned a
hydrostratigraphic layer using the above classification scheme and judgment and the borehole
data entered into RockWorks 16™ for 3-D modeling. In the portions of the area to be modeled
for which borehole data is not available, dummy boreholes were used to extend the model to the
model boundaries. These boreholes were based on the hydrostratigraphic thickness of the
existing boreholes and the elevation of the existing boreholes based on the assumption that the
hydrostratigraphic layers are a subdued replica of the original topography of the site and
geologic judgment.

8 

A grid of the pre-construction ground surface (described above) was used to constrain the
modeling of the natural layers. For gridding the data on a 20 ft x 20 ft grid across the area to be
modeled, hybrid algorithm was used with inverse distance weighted two (2) and triangulation
weighted one (1) and declustering, smoothing, and densifying subroutines. The declustering
option is used to remove duplicate points and de-cluster clustered points. The option creates a
temporary grid with a z-value assigned based on the closet data point to the midpoint of a voxel.
The smoothing option averages the z-values in a grid based on a filter size. For this modeling,
the z-value is assigned the average of itself and that of the eight nodes immediately surrounding
it. One smoothing pass is made. The densify option adds additional points to the xyz input by
fitting a Delaunay triangulation network to the data and adding the midpoint of each triangle to
the xyz input points. The net result is that the subsequent gridding process uses more control
points and tends to constrain algorithms that may become creative in areas of little control. Only
one densification pass is made. The completed model grids were exported in spreadsheet
format for use in the groundwater flow and transport model.

4.2 GMS MODFLOW Version 10

The conceptual model approach to construct a MODFLOW simulation in GMS MODFLOW
consists of employing GIS tools in a Map module to develop a conceptual model of the site
being modeled. The location of sources/sinks, layer parameters (such as hydraulic conductivity),
and all other data necessary for the simulation can be defined at the conceptual model level.
Once this model is complete, the grid is generated and the conceptual model is converted to the
grid model and all of the cell-by-cell assignments are performed automatically
The following table presents the sequence of the steps used for the groundwater modeling.
Steps 1 through 6 describe the creation of 3D MODFLOW model.
Step 1. Creating raster files for the model layer
‐ 3 surface layers (pre-construction, pre-construction with dike, and existing surface
including dike and ash) using GIS and AutoCad
‐ 2 subsurface layers (transition (TZ) and bedrock (BR)) by converting 3D scatter data
Step 2. Creating the Raster Catalog to group the raster layers
‐ Assigning Horizons and materials for each layer
Step 3. Creating horizon surfaces (i.e. TIN) from raster data
‐ Used exiting surface and bedrock rasters
Step 4. Building Solids from the Raster Catalog and TINs
‐ Used raster data for the top and bottom elevations of the solids
Step 5. Creating the conceptual model
9 

‐ Building model boundary, specified head boundary, and drain
‐ Defining zones and assigning hydraulic conductivity and recharge rate
‐ Importing observation wells and surface flow data
Step 6. Creating the MODFLOW 3D grid model
‐ Converting the solids to 3D grid model using boundary matching
‐ Mapping the conceptual model to 3D MODFLOW grid
Step 7. Flow model calibration/Sensitivity Analysis
‐ Initializing the MODFLOW model
‐ Steady-state calibration with the trial-and error method
‐ Parameters: hydraulic conductivity and recharge rate
‐ Used observation well and surface flow data
Step 8. Setting the transport model (MT3DMS)
‐ Species
‐ Stress periods
‐ Porosity and dispersion coefficient
‐ Distribution coefficient (Kd) from the lab experiments
‐ Recharge concentrations
Step 9. Performing model simulations
‐ Model scenarios - Existing Conditions, Remove ash

4.3 Model Domain and Grid

The model domain encompasses the DRSS site, including a section of the Broad River and all
site features relevant to the assessment of groundwater. The model domain extends beyond the
ash management areas to hydrologic boundaries such that groundwater flow and COI transport
through the area is accurately simulated without introducing artificial boundary effects.

The bounding rectangle around the model domain extends 5,000 feet north to south and 6,000
feet east to west and has a grid consisting of 255,442 active cells (Figure 1). In plan view, the
DRSS model domain is bounded by the following hydrologic features of the site:

 the northern shore of the Dan River to the south,
 the unnamed stream to the east,
 the presumed groundwater divide along a topographic divide north of the site, and
 the unnamed stream and the service settling water pond to the west.

The domain boundary was developed by manually digitizing 2-foot LiDAR contours in ARCMAP.
The lower limit of the model domain coincides with an assumed maximum depth of water
yielding fractures in bedrock. This was assumed to be 80 feet below the base of the transition
zone across the site upper limit based on a review of boring logs contained in the CSA (HDR
2015).
10 

There are a total of 12 model layers divided among the identified hydrostratigraphic units to
simulate curvilinear flow with a vertical flow component (Figures 2 and 3). The units are
represented by the model layers listed below:

• Model layers 1 through 3 Ash Material
• Model layers 4 through 6 Dike and Ash Material
• Model layer 7 M1 Saprolite and Alluvium where present
• Model layer 8 M1 Saprolite
• Model layer 9 M2 Saprolite
• Model layer 10 Transition Zone
• Model layers 11 and 12 Fractured Bedrock

4.4 Hydraulic Parameters

Horizontal hydraulic conductivity and the ratio of horizontal to vertical hydraulic conductivity,
which are specific for each hydrostratigraphic unit, are the primary determinants of groundwater
flow for a given set of boundary conditions. Measurements of these parameters from the CSA
(HDR Tables 11-6 through 11-10) provide guidance for the flow model calibration.

4.5 Flow Model Boundary Conditions

Boundary conditions for the DRSS flow model are one of two types: constant head or drain. No-
flow boundaries of the site have no prescribed boundary in the model. The outer boundary of
the model domain was selected to coincide with physical hydrologic boundaries at rivers and
drainage features, and no flow boundaries at topographic divides (Figures 5 and 6).

At the model boundary defined by the Dan River and the cooling water settling pond, a constant
head boundary was applied at layers those layers above fractured bedrock (layers 11 and 12)
with bottom elevations below the water surface (Figures 5 and 6). The head was interpolated
digitally from photogrammetry recorded in 2014.

External and internal drain boundaries were applied at the unnamed streams east and west of
the site and at selected low areas in the model interior to limit the rise of the modeled water
table above ground surface (Figure 5). Drainage features within small catchments in the upland
area of the site and other selected low areas act as shallow hydrologic boundaries. Physically
they represent locations where the water table intersects the ground surface and groundwater is
discharged. Drain boundaries were applied with a conductance calculated using the hydraulic
conductivity of the adjacent model layer, an assumed bed thickness of one foot, and width equal
to the model cell width. It was assumed that this surface drainage is ultimately conveyed to an
outfall at the Catawba River.

11 

Constant heads used to represent the Catawba River may provide an unmitigated source of
water for simulation of any corrective action alternatives that potentially involve pumping near
those boundaries. If corrective actions involve pumping, drain boundaries may be a more
appropriate boundary condition selection than constant heads for this boundary.

4.6 Flow Model Sources and Sinks

Recharge is the only source considered in the model. (Drainage boundaries are described
above as flow boundary conditions.) No pumping wells or other sources and sinks were
identified in the CSA (HDR 2015). Mean annual recharge in the Piedmont ranges from 4.0 to 9.7
inches per year (Daniel 2001). Recharge applied in the model is shown in Figure 4.

4.7 Flow Model Calibration Targets

The steady state flow model calibration targets are the 53 water level observations made in
June 2015. These wells include 21 wells screened in the ash, ash dike, and shallow zone
(S/M1/M2), 17 wells in the transition zone, and 15 wells fractured bedrock. Observations were
assigned by layer as shown in Table 3.

4.8 Transport Model Parameters

The calibrated, steady state flow model was used to apply flow conditions for the transport
models at the primary and secondary cells of the ash basin and the ash storage areas (HDR
Figures 2-1 and 2-2). Although their approximate dates of operation are known, the sluiced ash
loading histories for these locations are not available. In order to calibrate the transport model to
Existing Conditions, constant concentration source zones were applied at the ash model layers
in the ash basin cell and the layers where the water table occurs beneath the ash storage areas
starting from the date when each was placed in service (Figure 4). The relevant input
parameters were the constant concentration at the source zone, the linear sorption coefficient
Kd for sorptive constituents including arsenic, the total and effective porosity, and the
dispersivity tensor.

Arsenic sorption was represented by a linear isotherm in which the sorption coefficient Kd, with
units of ml/gram, is multiplied by the bulk density of the soil, with units of grams/ml. Sorption
studies on soil samples obtained during the CSA at DRSS indicate that the arsenic Kds for
native soils surrounding the ash basin and ash storage areas are greater than 20 ml/gram
(Langley and Oza 2015). Preliminary models runs with this Kd and an estimated bulk density of
2.12 gr/ml indicated that arsenic was essentially immobile over the simulation period.

The conceptual transport model specifies that COIs enter the model from the shallow saturated
zone in the ash basin and beneath the ash storage areas. When the measured Kd values are
applied in the numerical model to arsenic migrating from the source zones, arsenic does not
12 

reach the downgradient observation wells where it was observed in June 2015 at the end of the
simulation period. The most appropriate method to calibrate the transport model in this case is
to lower the Kd values until an acceptable agreement between measured and modeled
concentrations is achieved. Thus, an effective Kd value results that likely represents the
combined result of intermittent activities over the service life of the ash basin and storage areas.
These may include pond dredging, dewatering for dike construction, and ash grading and
placement.

The velocity of COIs in groundwater is directly related to the effective porosity of the porous
medium. A single effective porosity value of 0.10 was assigned to the ash, dike, and S/M1/M2
layers, which is within the range of estimated values from the CSA (HDR 2015). For the
transition zone and fractured bedrock, the porosities applied were 0.01 and 0.0005,
respectively. Dispersivity quantifies the degree to which mechanical dispersion of COIs occurs
in advecting groundwater. Dispersivity values of 80 ft, 8 ft, and 0.8 ft (longitudinal, transverse
horizontal, transverse vertical) were applied in this model. Traditionally, dispersivity is estimated
to be some fraction of the scale, or plume length (Zheng and Bennett [2002]). The commonly
applied estimate is ten percent of the observation scale.

In order to avoid artificial oscillation in the numerical solution to the advection dispersion
equation, the grid Peclet number, or the ratio of grid spacing to longitudinal dispersivity, should
be less than two (Zheng and Bennett [2002]). Directly beneath the ash basin system, shallow
groundwater flow is vertical downward. In this case, the grid Peclet number criteria will not be
met due to the relatively small value for vertical dispersivity and the relatively large grid spacing,
or thickness of the model layers at depth. The effect of numerical oscillation on modeled
concentration and mass transport is indeterminate. However, any effect is considered to be
limited as vertical groundwater flow transitions to horizontal over a short distance beneath the
ash basin system.

4.9 Transport Model Boundary Conditions

The transport model boundary conditions are fixed at zero concentration where water leaves the
model. Initial concentrations and concentrations in recharge are zero. No background
concentration is specified.

4.10 Transport Model Sources and Sinks

The primary and secondary ash basin cells and ash storage areas are the sources for COIs in
the model. These sources are modeled as fixed concentrations in the layers at the water table
of these areas and their constant concentrations are shown in Table 9. Their magnitudes are
based transport model calibration to COI measurements in the shallow zone from the June 2015
sampling event, as described in Section 5.3.

13 

The transport model sinks correspond to the constant head and drain boundaries of the flow
model. Water and COI mass are removed as they enter cells comprising these boundaries.

4.11 Transport Model Calibration Targets

The calibration targets are the measured arsenic and boron concentrations for June/July 2015
shown in HDR Tables 7-5, and 10-9 to 10-11. Specific calibration target wells are highlighted in
Section 5.3.

5 Model Calibration to Current Conditions
5.1 Flow Model Residual Analysis

The flow model was calibrated to the water level observations taken during July 2015 in the
shallow, deep, and bedrock wells (Table 3). The water table elevations in the ash basin cells
and selected drainage features and depressions across the site were also considered as
calibration targets. The locations of the observation wells are provided in Figure 14. The initial
trial-and-error calibration assumed homogeneous conditions in each model layer.

Throughout the flow model calibration process, the assumption of a homogeneous transition
zone was retained, and its horizontal conductivity was assigned to give a reasonable, initial
model result for the water table and water level observations. Recharge was also fixed at
reasonable values early in the calibration process, then refinements were made by adjusting
hydraulic conductivity in the upper layers (S/M1/M2) zonally as shown in Figure 9. The goal of
delineating the zones in this way was to obtain the best local calibration using conductivity
values within the range of measurements made during the CSA.

The model was sensitive to increasing horizontal hydraulic conductivity of the shallow zone
(residual soil/saprolite layers S/M1/M2) where the elevation of the water table was inversely
proportional to this parameter. The flow model was also sensitive to changes in recharge
outside the ash basins and increased anisotropy of the shallow zone.

The calibrated flow model parameters are provided in Table 2. Measured and modeled water
levels (post-calibration) are compared in Table 3 and Figure 10. The calibrated flow model is
considered to represent steady-state flow conditions for the site and the ash basin system under
a long-term, average condition. The average root mean squared error (RMS) of the measured
versus modeled water levels for wells gauged in July 2015 is also provided in Table 3 for
comparison with total measured head change across the model domain. The ratio of the
average RMS error to total measured head change (the normalized root mean square error
NRMSE), is 9.95%.

14 

Contours of hydraulic heads for the calibrated steady state flow model are shown in Figures 11
to 13. From the upland area of the site, groundwater in the shallow, deep, and fractured bedrock
zones diverges from a mound between the two ash storage areas and flowed to the southwest,
south, and southeast. In the shallow zone, head contours reflect the zones of variable hydraulic
conductivity applied there. All groundwater flow ultimately discharges to the Dan River, the
unnamed stream to the east, and the unnamed stream and the service settling water pond to
the west. Head gradients steepen in the direction of the discharge boundaries. Most
groundwater flow passing beneath and originating from the ash basins and ash storage areas
discharges to the Dan River. The remainder discharges to the unnamed stream to the east.

5.2 Flow Model Sensitivity Analysis

Sensitivity of the flow model was considered by varying selected parameters by the amounts
shown in Table 4 above and below their respective calibration values and calculating the
NRMSE for comparison with the calibration value. The model was more sensitive to increasing
horizontal hydraulic conductivity (Kh) of the shallow zone relative to the transition zone. A
slightly lower error was noted for decreased Kh in both zones. Increasing recharge outside the
ash basin yielded a lower NRMSE. The opposite was true for an increase. Increasing anisotropy
of the shallow zone yielded an increase in NRMSE, while the opposite was true for an increase.
The model NRMSE was relatively insensitive to changes in transition zone anisotropy and
recharge within the ash basin.

5.3 Transport Model Calibration and Sensitivity

For the transport model calibration, the constant source concentrations and their lateral extents
were adjusted to minimize residual concentrations in target wells. Varying sorption coefficients
for sorptive COIs was essential to achieve a reasonable calibration as discussed in Section 4.8.
Dispersivity and porosity changes can also be applied for transport calibration, but their
simultaneous effects on each COI must be considered.

Targeted observation wells for the arsenic transport model are listed in Table 5. The calibrated
transport model results for arsenic at the end of the 59-year simulation (2015) are shown in
Figures 17 to 19. From the source zones at the ash basins and ash storage areas, arsenic
migrated horizontally and vertically through the shallow zone and into the transition zone and
upper fractured bedrock. The general direction of arsenic transport from the source zones was
to the southeast. Arsenic with a sorption coefficient Kd of 2 ml/gram applied in the model moved
at less than the pore velocity of groundwater. The vertical transport of arsenic was limited due to
sorption. Lateral arsenic transport from the ash storage areas did not reach the model
boundaries at the Dan River and unnamed stream to the east, or the ash basins. Arsenic
sourced from the middle third of two ash basins reached the Dan River. The residuals
(measured – modeled concentration) for the targeted wells are presented in Table 5. The
transport model focused on attempting to match the higher measured arsenic concentrations to
15 

their corresponding modeled concentrations. Shallow zone groundwater that discharges to
surface water at the drains carried arsenic out of the model.

Targeted observation wells for the boron transport model are listed in Table 5. The calibrated
transport model results for boron at the end of the 59-year simulation (2015) are shown in
Figures 32 to 34. From the source zones at the ash basins and ash storage areas, boron
migrated horizontally and vertically through the shallow zone and into the transition zone and
fractured bedrock. The general direction of boron transport from the sources zones was to the
southeast. Boron without sorption applied moved at the pore velocity of groundwater. Due to the
low conductivity of the downgradient dike, boron at higher concentrations exits the model at
depth with groundwater discharging to the Dan River and the unnamed stream on the eastern
boundary of the site. The highest boron concentrations exiting at the model boundary are
downgradient of the southern half of the ash basin primary cell. The residuals (measured –
modeled concentration) for the targeted wells are presented in Table 5. The transport model
focused on attempting to match the higher measured boron concentrations to their
corresponding modeled concentrations. Shallow zone groundwater that discharges to surface
water at the drains carried boron out of the model.

Targeted observation wells for the chromium transport model are listed in Table 5. The
calibrated transport model results for chromium at the end of the 59-year simulation (2015) are
shown in Figures 62 to 64. From the source zones at the ash basins and ash storage areas,
chromium migrated horizontally and vertically through the shallow zone and into the transition
zone and upper fractured bedrock. The general direction of chromium transport from the source
zones was to the southeast. Chromium without sorption applied moved at the pore velocity of
groundwater. The vertical transport of chromium was limited due to sorption. Lateral chromium
transport from the southwestern ash storage area reached the ash basins. Chromium sourced
from the two ash basins reached the Dan River. The residuals (measured – modeled
concentration) for the targeted wells are presented in Table 5. The transport model focused on
attempting to match the higher measured chromium concentrations to their corresponding
modeled concentrations. Shallow zone groundwater that discharges to surface water at the
drains carried chromium out of the model.

Targeted observation wells for the sulfate transport model are listed in Table 5. The calibrated
transport model results for sulfate at the end of the 59-year simulation (2015) are shown in
Figures 92 to 94. From the source zones at the ash basins and ash storage areas, sulfate
migrated horizontally and vertically through the shallow zone and into the transition zone and
fractured bedrock. The general direction of sulfate transport from the sources zones was to the
southeast. Sulfate without sorption applied moved at the pore velocity of groundwater. Due to
the low conductivity of the downgradient dike, sulfate at higher concentrations exits the model at
depth with groundwater discharging to the Dan River and the unnamed stream on the eastern
boundary of the site. The constant source concentrations and their lateral extents were adjusted
to minimize residual concentrations in the target wells. The residuals (measured – modeled
16 

concentration) for the targeted wells are presented in Table 5. The transport model focused on
attempting to match the higher measured sulfate concentrations to their corresponding modeled
concentrations. Shallow zone groundwater that discharges to surface water at the drains carried
sulfate out of the model.

Sensitivity analyses for porosity, dispersivity, and sorption are provided in Tables 6 to 8. When
the sorption coefficient for arsenic was decreased by factors of 1/20 and 1/200 from the
calibration value, the modeled concentrations increased significantly (Table 6). The retardation
effect of sorption was still significant when Kd was reduced from one to 0.1 ml/gram. Sensitivity
of modeled results to sorption will be greater for transient plumes where their leading and
trailing edges will shift in response to the Kd value.

In general, boron concentrations increased significantly, as porosity was decreased from 0.3 to
0.1 depending on location with respect to a boron source zone (Table 7). At lower porosity,
boron will travel faster from the constant concentration source zones, enabling higher
concentrations to be realized sooner at a given distance from the zone. The effect is reduced or
reversed at locations near the source zone.

In general, increasing dispersivity resulted in increased boron concentration at locations where
the leading edge of the plume was situated at the end of the 59-year simulation (Table 8). The
results in Table 8 show dispersivity effects are limited, which is characteristic of a plume that is
near steady state.

6 Predictive Simulations of Source Removal Scenarios
The groundwater model, calibrated for flow and constituent fate and transport under Existing
Conditions, was applied to evaluate three corrective action scenarios at DRSS: the As-Is
scenario, the Cover-in-Place scenario, and the Excavation scenario. Being predictive, these
simulations produce flow and transport results for conditions that are beyond the range of those
considered during the calibration. Thus, the model should be recalibrated and verified over time
as new data becomes available in order to improve its accuracy and reduce its uncertainty. The
model domain developed for Existing Conditions was applied without modification for the As-Is
and Cover-In-Place scenarios.

For the Excavation scenario, the primary and secondary cells of the ash basin, the ash basin
dikes, and the ash storage areas were removed as shown Figure 134. The flow parameters for
this model were identical to the Existing Conditions models except for the removal of ash related
layers, and the same recharge being applied on the ash basin as the remainder of the site.

Transport model parameters and supplemental transport calibrations for predictive simulations
are provided in Table 9.

17 

6.1 Existing Conditions Scenario

The Existing Conditions scenario consists of modeling each COI using the calibrated model for
steady-state flow and transient transport under the Existing Conditions across the site to
estimate when steady state concentrations are reached across the site and at the compliance
boundary. COI concentrations can only remain the same or increase initially for this scenario
with source concentrations being held at their constant value over all time. Thereafter, their
concentrations and discharge rates remain constant. This scenario represents the worst case in
terms of groundwater concentrations on and off site, and COIs discharging to surface water at
steady state.

The time to achieve a steady state concentration plume depends on the source zone location
relative to the compliance boundary and its loading history. Source zones close to the
compliance boundary will reach steady state sooner. The time to steady state concentration is
also dependent on the sorptive characteristics of each COI. Sorptive COIs will be transient for a
longer time period as their peak breakthrough concentration travels at a rate that is less than the
groundwater pore velocity. Lower effective porosity will result in shorter times to achieve steady
state for both sorptive and non-sorptive COIs. Lower total porosity will result in longer times for
sorptive COIs.

Figures 14 through 16 show predicted arsenic concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Existing Conditions scenario, arsenic concentrations increase at each well shown
over the modeled time period of 250 years. Arsenic concentrations remain above the 2L
Standard, which is 10 μg/l, at these wells under the Existing Conditions scenario. Figures 17
through 19 show arsenic concentrations in the shallow, deep and fractured bedrock zones in
2015. Figures 20 through 22 show Existing Conditions arsenic concentrations in the shallow,
deep and fractured bedrock zones in 2115. Arsenic exits the model with groundwater
discharging at the Dan River to the southeast of the site in all zones in 2015 and 2115. In 2015
and 2115, the shallow groundwater zone has the highest arsenic concentrations at the ash
basins. Horizontal and vertical advective transport of arsenic is slower than groundwater pore
velocity due to sorption.

Figures 29 through 31 show predicted boron concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Existing Conditions scenario, boron concentrations at these wells have reached or
are approaching steady state in 2015. Concentrations at these wells relative to the 2L standard
depend on the upgradient source zone concentrations. Figures 32 through 34 show boron
concentrations in the shallow, deep and fractured bedrock zones in 2015. Figures 35 through 37
show Existing Conditions boron concentrations in the shallow, deep and fractured bedrock
zones in 2115. In 2015 and 2115, boron has reached the shallow, deep, and bedrock zones
beneath the ash basins and ash storage areas. Boron exits the model with groundwater
18 

discharging at the Dan River to the southeast of the site from all zones in 2015 and 2115.
Horizontal and vertical advective transport of boron is the same as groundwater pore velocity as
it is modeled without sorption.

Figures 44 through 46 show predicted chromium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Existing Conditions scenario, chromium concentrations at these wells have reached
or are approaching steady state in 2015. Concentrations at these wells relative to the 2L
standard depend on the upgradient source zone concentrations. Figures 47 through 49 show
chromium concentrations in the shallow, deep and fractured bedrock zones in 2015. Figures 50
through 52 show Existing Conditions chromium concentrations in the shallow, deep and
fractured bedrock zones in 2115. In 2015 and 2115, chromium has reached the shallow, deep,
and bedrock zones beneath the ash basins and ash storage areas. Chromium exits the model
with groundwater discharging at the Dan River to the southeast of the site from all zones in
2015 and 2115. Horizontal and vertical advective transport of chromium is the same as
groundwater pore velocity as it is modeled without sorption.

Figures 59 through 61 show predicted chromium VI concentrations versus time at
representative observation wells under all three model scenarios. The concentration versus time
curves show that for the Existing Conditions scenario, chromium VI concentrations at these
wells have reached or are approaching steady state in 2015. Concentrations at these wells
depend on the upgradient source zone concentrations. Figures 62 through 64 show chromium
VI concentrations in the shallow, deep and fractured bedrock zones in 2015. Figures 65 through
67 show Existing Conditions chromium concentrations in the shallow, deep and fractured
bedrock zones in 2115. In 2015 and 2115, chromium has reached the shallow, deep, and
bedrock zones beneath the ash basins and ash storage areas. Chromium exits the model with
groundwater discharging at the Dan River to the southeast of the site from all zones in 2015 and
2115. Horizontal and vertical advective transport of chromium VI is the same as groundwater
pore velocity as it is modeled without sorption.

Figures 74 through 76 show predicted cobalt concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Existing Conditions scenario, cobalt concentrations increase at each well shown
over the modeled time period of 250 years. Cobalt concentrations remain above the IMAC
Standard, which is 1 μg/l, at these wells under the Existing Conditions scenario. Figures 77
through 79 show cobalt concentrations in the shallow, deep and fractured bedrock zones in
2015. Figures 80 through 82 show Existing Conditions cobalt concentrations in the shallow,
deep and fractured bedrock zones in 2115. Cobalt exits the model with groundwater discharging
at the Dan River to the southeast of the site from all zones in 2015 and 2115. Horizontal and
vertical advective transport of cobalt is slower than groundwater pore velocity due to sorption.

19 

Figures 89 through 91 show predicted sulfate concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Existing Conditions scenario, sulfate concentrations at these wells have reached or
are approaching steady state in 2015. Concentrations at these wells depend on the upgradient
source zone concentrations. Figures 92 through 94 show sulfate concentrations in the shallow,
deep and fractured bedrock zones in 2015. Figures 95 through 97 show Existing Conditions
chromium concentrations in the shallow, deep and fractured bedrock zones in 2115. In 2015
and 2115, chromium has reached the shallow, deep, and bedrock zones beneath the ash
basins and ash storage areas. Sulfate exits the model with groundwater discharging at the Dan
River to the southeast of the site from all zones in 2015 and 2115. Horizontal and vertical
advective transport of sulfate is the same as groundwater pore velocity as it is modeled without
sorption.

Figures 104 through 106 show predicted thallium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Existing Conditions scenario, thallium concentrations increase at each well shown
over the modeled time period of 250 years. Thallium concentrations remain above the IMAC
Standard, which is 0.2 μg/l, at these wells under the Existing Conditions scenario. Figures 107
through 109 show thallium concentrations in the shallow, deep and fractured bedrock zones in
2015. Figures 110 through 112 show Existing Conditions thallium concentrations in the shallow,
deep and fractured bedrock zones in 2115. Thallium exits the model with groundwater
discharging at the Dan River to the southeast of the site from the shallow and deep zones in
2015 and from all zones in 2115. Horizontal and vertical advective transport of thallium is slower
than groundwater pore velocity due to sorption.

Figures 119 through 121 show predicted vanadium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Existing Conditions scenario, vanadium concentrations increase at each well shown
over the modeled time period of 250 years. Vanadium concentrations remain above the IMAC
Standard, which is 0.3 μg/l, at these wells under the Existing Conditions scenario. Figures 122
through 124 show vanadium concentrations in the shallow, deep and fractured bedrock zones in
2015. Figures 125 through 127 show Existing Conditions vanadium concentrations in the
shallow, deep and fractured bedrock zones in 2115. Vanadium exits the model with
groundwater discharging at the Dan River to the southeast of the site from all zones in 2015 and
2115. Horizontal and vertical advective transport of vanadium is slower than groundwater pore
velocity due to sorption.

20 

6.2 Cap-in-Place Scenario

The Cap-in-Place model simulates the effects of covering the ash basins and ash storage areas
at the beginning of this scenario. In the model, recharge and source zone concentrations at the
ash basins and ash storage areas are set to zero. Groundwater flow is affected by this scenario
as the water table may be lowered and groundwater velocities may be reduced beneath the
covered areas. In the model, non-sorptive COIs will move downgradient at the pore velocity of
groundwater and will be completely displaced by the passage of a single pore water volume of
clean water. Sorptive COI migration will be slowed relative to the groundwater pore velocity as
they are desorbed by clean water.

Figures 14 through 16 show predicted arsenic concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Cap-in-Place scenario, arsenic concentrations increase until 2015 when the
scenario is implemented. Thereafter, arsenic concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones. Arsenic
concentrations remain above the 2L Standard, which is 10 μg/l, at these wells under the Cap-in-
Place scenario. Figures 23 through 25 show Cap-in-Place arsenic concentrations in the shallow,
deep and fractured bedrock zones in 2115. Arsenic concentration are reduced in each zone
relative to the Existing Conditions scenario at 2115 (Figures 20 through 22). Arsenic exits the
model with groundwater discharging at the Dan River to the southeast of the site in all zones in
2115 for the Cap-in-Place scenario. In 2115, the deep groundwater zone has the highest
arsenic concentrations at the ash basins. Horizontal and vertical advective transport of arsenic
is slower than groundwater pore velocity due to sorption

Figures 29 through 31 show predicted boron concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Cap-in-Place scenario, boron concentrations increase until 2015 when the scenario
is implemented. Thereafter, boron concentrations decline with a shape that is a function of the
well location relative to the upgradient source zones. Concentrations at these wells relative to
the 2L standard depend on the upgradient source zone concentrations. Figures 38 through 40
show Cap-in-Place boron concentrations in the shallow, deep and fractured bedrock zones in
2115. Since boron is modeled without sorption, essentially all boron in the shallow, deep, and
bedrock zones has exited the model with groundwater discharging at the Dan River to the
southeast of the site by 2115. The predicted boron concentrations versus time at the
observation wells (Figures 29 through 31) suggest that the approximate year for boron depletion
in the model is 2050 for the Cap-in-Place scenario.

Figures 44 through 46 show predicted chromium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Cap-in-Place scenario, chromium concentrations increase until 2015 when the
scenario is implemented. Thereafter, chromium concentrations decline with a shape that is a
21 

function of the well location relative to the upgradient source zones. Concentrations at these
wells relative to the 2L standard depend on the upgradient source zone concentrations. Figures
53 through 55 show Cap-in-Place chromium concentrations in the shallow, deep and fractured
bedrock zones in 2115. Since chromium is modeled without sorption, essentially all chromium in
the shallow, deep, and bedrock zones has exited the model with groundwater discharging at the
Dan River to the southeast of the site by 2115. The predicted chromium concentrations versus
time at the observation wells (Figures 44 through 46) suggest that the approximate year for
chromium depletion in the model is 2040 for the Cap-in-Place scenario.

Figures 59 through 61 show predicted chromium VI concentrations versus time at
representative observation wells under all three model scenarios. The concentration versus time
curves show that for the Cap-in-Place scenario, chromium VI concentrations increase until 2015
when the scenario is implemented. Thereafter, chromium VI concentrations decline with a
shape that is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the 2L standard depend on the upgradient source zone
concentrations. Figures 68 through 70 show Cap-in-Place chromium VI concentrations in the
shallow, deep and fractured bedrock zones in 2115. Since chromium VI is modeled without
sorption, essentially all chromium VI in the shallow, deep, and bedrock zones has exited the
model with groundwater discharging at the Dan River to the southeast of the site by 2115. The
predicted chromium VI concentrations versus time at the observation wells (Figures 59 through
61) suggest that the approximate year for chromium VI depletion in the model is 2030 for the
Cap-in-Place scenario.

Figures 74 through 76 show predicted cobalt concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Cap-in-Place scenario, cobalt concentrations increase until 2015 when the
scenario is implemented. Thereafter, cobalt concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones. Cobalt
concentrations remain above the 2L Standard, which is 1 μg/l, at these wells under the Cap-in-
Place scenario. Figures 83 through 85 show Cap-in-Place cobalt concentrations in the shallow,
deep and fractured bedrock zones in 2115. Cobalt concentrations are reduced in each zone
relative to the Existing Conditions scenario at 2115 (Figures 80 through 82). Cobalt exits the
model with groundwater discharging at the Dan River to the southeast of the site in all zones in
2115 for the Cap-in-Place scenario. Horizontal and vertical advective transport of cobalt is
slower than groundwater pore velocity due to sorption.

Figures 89 through 91 show predicted sulfate concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Cap-in-Place scenario, sulfate concentrations increase until 2015 when the scenario
is implemented. Thereafter, sulfate concentrations decline with a shape that is a function of the
well location relative to the upgradient source zones. Concentrations at these wells relative to
the 2L standard depend on the upgradient source zone concentrations. Figures 98 through 100
22 

show Cap-in-Place sulfate concentrations in the shallow, deep and fractured bedrock zones in
2115. Since sulfate is modeled without sorption, essentially all sulfate in the shallow, deep, and
bedrock zones has exited the model with groundwater discharging at the Dan River to the
southeast of the site by 2115. The predicted sulfate concentrations versus time at the
observation wells (Figures 89 through 91) suggest that the approximate year for sulfate
depletion in the model is 2030 for the Cap-in-Place scenario.

Figures 104 through 106 show predicted thallium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Cap-in-Place scenario, thallium concentrations increase until 2015 when the
scenario is implemented. Thereafter, thallium concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the 2L standard depend on the upgradient source zone
concentrations. Figures 113 through 115 show Cap-in-Place thallium concentrations in the
shallow, deep and fractured bedrock zones in 2115. Thallium concentrations are reduced in
each zone relative to the Existing Conditions scenario at 2115 (Figures 110 through 112).
Thallium exits the model with groundwater discharging at the Dan River to the southeast of the
site in all zones in 2115 for the Cap-in-Place scenario. Horizontal and vertical advective
transport of thallium is slower than groundwater pore velocity due to sorption.

Figures 119 through 121 show predicted vanadium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Cap-in-Place scenario, vanadium concentrations increase until 2015 when the
scenario is implemented. Thereafter, vanadium concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the IMAC standard depend on the upgradient source
zone concentrations. Figures 128 through 130 show Cap-in-Place vanadium concentrations in
the shallow, deep and fractured bedrock zones in 2115. Vanadium concentrations are reduced
in each zone relative to the Existing Conditions scenario at 2115 (Figures 125 through 127).
Vanadium exits the model with groundwater discharging at the Dan River to the southeast of the
site in all zones in 2115 for the Cap-in-Place scenario. Horizontal and vertical advective
transport of vanadium is slower than groundwater pore velocity due to sorption.

23 

6.3 Excavation Scenario

The Excavation model simulates the effects of removing the ash basins, the dikes, and ash
storage areas at the beginning of this scenario. In the model, source zone concentrations at the
ash basins and ash storage areas are set to zero while recharge is applied at the same rate as
other surrounding areas. Groundwater flow beneath the ash basin is affected by this scenario as
the basins are completely drained. In the model, non-sorptive COIs will move downgradient at
the pore velocity of groundwater and will be completely displaced by the passage of a single
pore water volume of clean water. Sorptive COI migration will be retarded relative to the
groundwater pore velocity as they are desorbed by clean water.

Figures 14 through 16 show predicted arsenic concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Excavation scenario, arsenic concentrations increase until 2015 when the
scenario is implemented. Thereafter, arsenic concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the 2L standard depend on the upgradient source zone
concentrations. Figures 26 through 28 show Excavation arsenic concentrations in the shallow,
deep and fractured bedrock zones in 2115. Arsenic concentration are reduced in each zone
relative to the Existing Conditions scenario at 2115 (Figures 20 through 22). Arsenic exits the
model with groundwater discharging at the Dan River to the southeast of the site in all zones in
2115 for the Excavation scenario. In 2115, the deep groundwater zone has the highest arsenic
concentrations at the ash basins. Horizontal and vertical advective transport of arsenic is slower
than groundwater pore velocity due to sorption

Figures 29 through 31 show predicted boron concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Excavation scenario, boron concentrations increase until 2015 when the scenario is
implemented. Thereafter, boron concentrations decline with a shape that is a function of the well
location relative to the upgradient source zones. Concentrations at these wells relative to the 2L
standard depend on the upgradient source zone concentrations. Figures 41 through 43 show
Excavation boron concentrations in the shallow, deep and fractured bedrock zones in 2115.
Since boron is modeled without sorption, essentially all boron in the shallow, deep, and bedrock
zones has exited the model with groundwater discharging at the Dan River to the southeast of
the site by 2115. The predicted boron concentrations versus time at the observation wells
(Figures 29 through 31) suggest that the approximate year for boron depletion in the model is
2060 for the Excavation scenario.

Figures 44 through 46 show predicted chromium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Excavation scenario, chromium concentrations increase until 2015 when the
scenario is implemented. Thereafter, chromium concentrations decline with a shape that is a
24 

function of the well location relative to the upgradient source zones. Concentrations at these
wells relative to the 2L standard depend on the upgradient source zone concentrations. Figures
56 through 58 show Excavation chromium concentrations in the shallow, deep and fractured
bedrock zones in 2115. Since chromium is modeled without sorption, essentially all chromium in
the shallow, deep, and bedrock zones has exited the model with groundwater discharging at the
Dan River to the southeast of the site by 2115. The predicted chromium concentrations versus
time at the observation wells (Figures 44 through 46) suggest that the approximate year for
chromium depletion in the model is 2060 for the Excavation scenario.

Figures 59 through 61 show predicted chromium VI concentrations versus time at
representative observation wells under all three model scenarios. The concentration versus time
curves show that for the Excavation scenario, chromium VI concentrations increase until 2015
when the scenario is implemented. Thereafter, chromium VI concentrations decline with a
shape that is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the 2L standard depend on the upgradient source zone
concentrations. Figures 71 through 73 show Excavation chromium VI concentrations in the
shallow, deep and fractured bedrock zones in 2115. Since chromium VI is modeled without
sorption, essentially all chromium VI in the shallow, deep, and bedrock zones has exited the
model with groundwater discharging at the Dan River to the southeast of the site by 2115. The
predicted chromium VI concentrations versus time at the observation wells (Figures 59 through
61) suggest that the approximate year for chromium VI depletion in the model is 2030 for the
Excavation scenario.

Figures 74 through 76 show predicted cobalt concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Excavation scenario, cobalt concentrations increase until 2015 when the scenario
is implemented. Thereafter, cobalt concentrations reach a peak flow by a decline whose shape
is a function of the well location relative to the upgradient source zones. Cobalt concentrations
remain above the 2L Standard, which is 1 μg/l, at these wells under the Excavation scenario.
Figures 86 through 88 show Excavation cobalt concentrations in the shallow, deep and
fractured bedrock zones in 2115. Cobalt concentrations are reduced in each zone relative to the
Existing Conditions scenario at 2115 (Figures 80 through 82). Cobalt exits the model with
groundwater discharging at the Dan River to the southeast of the site in all zones in 2115 for the
Excavation scenario. Horizontal and vertical advective transport of cobalt is slower than
groundwater pore velocity due to sorption.

Figures 89 through 91 show predicted sulfate concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that for the Excavation scenario, sulfate concentrations increase until 2015 when the scenario is
implemented. Thereafter, sulfate concentrations decline with a shape that is a function of the
well location relative to the upgradient source zones. Concentrations at these wells relative to
the 2L standard depend on the upgradient source zone concentrations. Figures 101 through
25 

103 show Excavation sulfate concentrations in the shallow, deep and fractured bedrock zones
in 2115. Since sulfate is modeled without sorption, essentially all sulfate in the shallow, deep,
and bedrock zones has exited the model with groundwater discharging at the Dan River to the
southeast of the site by 2115. The predicted sulfate concentrations versus time at the
observation wells (Figures 89 through 91) suggest that the approximate year for sulfate
depletion in the model is 2030 for the Excavation scenario.

Figures 104 through 106 show predicted thallium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Excavation scenario, thallium concentrations increase until 2015 when the
scenario is implemented. Thereafter, thallium concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the 2L standard depend on the upgradient source zone
concentrations. Figures 116 through 118 show Excavation thallium concentrations in the
shallow, deep and fractured bedrock zones in 2115. Thallium concentrations are reduced in
each zone relative to the Existing Conditions scenario at 2115 (Figures 110 through 112).
Thallium exits the model with groundwater discharging at the Dan River to the southeast of the
site in all zones in 2115 for the Excavation scenario. Horizontal and vertical advective transport
of thallium is slower than groundwater pore velocity due to sorption.

Figures 119 through 121 show predicted vanadium concentrations versus time at representative
observation wells under all three model scenarios. The concentration versus time curves show
that under the Excavation scenario, vanadium concentrations increase until 2015 when the
scenario is implemented. Thereafter, vanadium concentrations reach a peak flow by a decline
whose shape is a function of the well location relative to the upgradient source zones.
Concentrations at these wells relative to the IMAC standard depend on the upgradient source
zone concentrations. Figures 131 through 133 show Excavation vanadium concentrations in the
shallow, deep and fractured bedrock zones in 2115. Vanadium concentrations are reduced in
each zone relative to the Existing Conditions scenario at 2115 (Figures 125 through 127).
Vanadium exits the model with groundwater discharging at the Dan River to the southeast of the
site in all zones in 2115 for the Excavation scenario. Horizontal and vertical advective transport
of vanadium is slower than groundwater pore velocity due to sorption.

26 

7 Summary
7.1 Model Assumptions and Limitations

The model assumptions include the following:

 The steady state flow model was calibrated to heads measured at observation wells in
June and July of 2015.
 Steady state flow conditions were assumed from the time that the ash basins and ash
storage areas were placed in service through the current time until the end of the
predictive simulations.
 COI source zone concentrations at the ash basins and ash storage areas were assumed
to be constant with respect to time.
 Boron, chromium, chromium VI, and sulfate were assumed to be non-sorbing.
 Effective sorption coefficients for arsenic, cobalt, thallium, and vanadium were applied in
the model to allow for transport model calibration that is consistent with the conceptual
transport model and measured COI concentrations.
 The model does not account for varying geochemical conditions such as pH and redox
potential that will affect COI mobility.

7.2 Model Predictions

The model predictions are summarized as follows:
 For the Existing Conditions scenario, boron, chromium, chromium VI, and sulfate
concentrations will be at steady state after 2015. Arsenic, cobalt, thallium, and vanadium
concentrations will increase after 2015 years under this scenario.
 For the Cap-in-Place and Excavation scenarios non-sorptive COIs will be depleted from
the model before 2060.
 The Cap-in-Place is predicted to result in lower concentrations over time relative to the
Excavation scenario for sorptive constituents.

REFERENCES
Daniel, C.C., III, 2001, Estimating ground-water recharge in the North Carolina Piedmont for
land use planning [abs.], in 2001 Abstracts with Programs, 50th Annual Meeting, Southeastern
Section, April 5-6, 2001: Raleigh, N.C., The Geological Society of America, v. 33, no. 2, p. A-80.
Haven, W.T. Introduction to the North Carolina Groundwater Recharge Map-Groundwater
Circular Number 19, North Carolina Department of Environment and Natural Resources Division
of Water Quality, Groundwater Section.
HDR. Comprehensive Site Assessment Report, Dan River Steam Station Ash Basin, August
2015.
Langley, W.G. and Oz, Shubhashini. Soil Sorption Evaluation for Dan River Steam Station,
UNC-Charlotte, October 2015.
LeGrand, H. E. 2004. A Master Conceptual Model for Hydrogeological Site Characterization in
the Piedmont and Mountain Region of North Carolina, A Guidance Manual, North Carolina
Department of Environment and Natural Resources Division of Water Quality, Groundwater
Section.
Niswonger, R.G., Panday, S., and Ibaraki, M. 2011. MODFLOW-NWT, A Newton formulation for
MODFLOW-2005: U.S. Geological Survey Techniques and Methods 6-A37, 44 p.
Zheng, C. and Bennett, G. Applied Contaminant Transport Modeling Second Edition, Wiley
Interscience, 2002.
Zheng, C. and P. Wang. 1999. MT3DMS, A modular three-dimensional multi-species transport
model for simulation of advection, dispersion and chemical reactions of contaminants in
groundwater systems, Documentation and Users Guide, U.S. Army Engineer Research and
Development Center Contract Report SERDP-99-1, Vicksburg, MS, 202 p.

Fi
g
u
r
e
 1.
 Co
n
c
e
p
t
u
a
l
 Gr
o
u
n
d
w
a
t
e
r
 Fl
o
w
 Mo
d
e
l
/
M
o
d
e
l
 Do
m
a
i
n

Fi
g
u
r
e
 2.
 Mo
d
e
l
 Do
m
a
i
n
 Cr
o
s
s
 se
c
t
i
o
n
 A ‐A’

Fi
g
u
r
e
 3.
 Mo
d
e
l
 Do
m
a
i
n
 Cr
o
s
s
 se
c
t
i
o
n
 B ‐B’

Fi
g
u
r
e
 4.
 Fl
o
w
 mo
d
e
l
 bo
u
n
d
a
r
y
 co
n
d
i
t
i
o
n
s

Fi
g
u
r
e
 5.
 Re
c
h
a
r
g
e
 an
d
 co
n
s
t
a
n
t
 co
n
c
e
n
t
r
a
t
i
o
n
 bo
u
n
d
a
r
y
 co
n
d
i
t
i
o
n

Fi
g
u
r
e
 6.
 Sh
a
l
l
o
w
 Ob
s
e
r
v
a
t
i
o
n
 We
l
l
s

Fi
g
u
r
e
 7.
 De
e
p
 Ob
s
e
r
v
a
t
i
o
n
 We
l
l
s

Fi
g
u
r
e
 8.
 Be
d
r
o
c
k
 Ob
s
e
r
v
a
t
i
o
n
 We
l
l
s

Fi
g
u
r
e
 9.
 Hy
d
r
a
u
l
i
c
 Co
n
d
u
c
t
i
v
i
t
y
 Zo
n
a
t
i
o
n
 in
 S/
M
1
/
M
2
 Mo
d
e
l
 La
y
e
r
s

Fi
g
u
r
e
 10
.
 Me
a
s
u
r
e
d
 ve
r
s
u
s
 mo
d
e
l
e
d
 wa
t
e
r
 le
v
e
l
s

Fi
g
u
r
e
 11
.
 Wa
t
e
r
 he
a
d
 co
n
t
o
u
r
 ma
p
 at
 La
y
e
r
 9  

Fi
g
u
r
e
 12
.
 Wa
t
e
r
 he
a
d
 at
 cr
o
s
s
 se
c
t
i
o
n
 C ‐C’

Fi
g
u
r
e
 13
.
 Wa
t
e
r
 he
a
d
 cr
o
s
s
 se
c
t
i
o
n
 D ‐D’

0
20
0
40
0
60
0
80
0
10
0
0
12
0
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
A
r
s
e
n
i
c

u
g
/
l
Fi
g
u
r
e
 14
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
S
L
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L

St
a
n
d
a
r
d
/
I
M
A
C

020406080
10
0
12
0
14
0
16
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
A
r
s
e
n
i
c

u
g
/
l
Fi
g
u
r
e
 15
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐5D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L

St
a
n
d
a
r
d
/
I
M
A
C

020406080
10
0
12
0
14
0
16
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
A
r
s
e
n
i
c

u
g
/
l
Fi
g
u
r
e
 16
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐25
S
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L

St
a
n
d
a
r
d
/
I
M
A
C

Fi
g
u
r
e
 17
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 18
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 19
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 20
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 21
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 22
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 23
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 24
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 25
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 26
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 27
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 28
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ar
s
e
n
i
c
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

0
20
0
40
0
60
0
80
0
10
0
0
12
0
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
B
o
r
o
n

u
g
/
l
Fi
g
u
r
e
 29
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐25
D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0
10
0
20
0
30
0
40
0
50
0
60
0
70
0
80
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
B
o
r
o
n

u
g
/
l
Fi
g
u
r
e
 30
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐30
B
R
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0
10
0
20
0
30
0
40
0
50
0
60
0
70
0
80
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
B
o
r
o
n

u
g
/
l
Fi
g
u
r
e
 31
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 MW
‐22
B
R
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

Fi
g
u
r
e
 32
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 33
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 34
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 35
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 36
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 37
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 38
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 39
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 40
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 41
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 42
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 43
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Bo
r
o
n
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

051015202530
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
h
r
o
m
i
u
m

u
g
/
l
Fi
g
u
r
e
 44
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐25
B
R
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

024681012
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
h
r
o
m
i
u
m

u
g
/
l
Fi
g
u
r
e
 45
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 MW
‐10
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0510152025
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
h
r
o
m
i
u
m

u
g
/
l
Fi
g
u
r
e
 46
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 MW
‐31
1
B
R
 fo
r

Mo
d
e
l
 Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

Fi
g
u
r
e
 47
.
 In
i
t
i
a
l
 (2
0
1
5
)

Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 48
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 49
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 50
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 51
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 52
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d

Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 53
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 54
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 55
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 56
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 57
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 58
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
h
r
o
m
i
u
m

V
I

u
g
/
l
Fi
g
u
r
e
 59
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d
‐NA

0123456
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
h
r
o
m
i
u
m

V
I

u
g
/
l
Fi
g
u
r
e
 60
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
S
L
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d
‐NA

0
0.
2
0.
4
0.
6
0.
8
1
1.
2
1.
4
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
h
r
o
m
i
u
m

V
I

u
g
/
l
Fi
g
u
r
e
 61
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐30
D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d
‐NA

Fi
g
u
r
e
 62
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 63
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 64
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 65
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 66
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 67
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 68
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 69
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 70
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 71
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 72
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 73
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Ch
r
o
m
i
u
m
 VI
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

020406080
10
0
12
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
o
b
a
l
t

u
g
/
l
Fi
g
u
r
e
 74
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

020406080
10
0
12
0
14
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
o
b
a
l
t

u
g
/
l
Fi
g
u
r
e
 75
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
S
L
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

05101520253035
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
C
o
b
a
l
t

u
g
/
l
Fi
g
u
r
e
 76
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐30
S
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

Fi
g
u
r
e
 77
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 78
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 79
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 80
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 81
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 82
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 83
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 84
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 85
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 86
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 87
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 88
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Co
b
a
l
t
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

0
50
0
0
0
10
0
0
0
0
15
0
0
0
0
20
0
0
0
0
25
0
0
0
0
30
0
0
0
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
S
u
l
f
a
t
e

u
g
/
l
Fi
g
u
r
e
 89
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐25
B
R
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0
50
0
0
0
10
0
0
0
0
15
0
0
0
0
20
0
0
0
0
25
0
0
0
0
30
0
0
0
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
S
u
l
f
a
t
e

u
g
/
l
Fi
g
u
r
e
 90
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐30
B
R
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0
50
0
0
0
10
0
0
0
0
15
0
0
0
0
20
0
0
0
0
25
0
0
0
0
30
0
0
0
0
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
S
u
l
f
a
t
e

u
g
/
l
Fi
g
u
r
e
 91
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐35
B
R
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

Fi
g
u
r
e
 92
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 93
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Su
l
f
a
t
e

Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 94
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 95
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 96
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 97
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d

Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 98
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 99
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 10
0
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 10
1
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 10
2
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 10
3
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Su
l
f
a
t
e
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

0
0.
2
0.
4
0.
6
0.
8
1
1.
2
1.
4
1.
6
1.
8
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
T
h
a
l
l
i
u
m

u
g
/
l
Fi
g
u
r
e
 10
4
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
S
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0
0.
1
0.
2
0.
3
0.
4
0.
5
0.
6
0.
7
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
T
h
a
l
l
i
u
m

u
g
/
l
Fi
g
u
r
e
 10
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐5S
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

0
0.
5
1
1.
5
2
2.
5
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
T
h
a
l
l
i
u
m

u
g
/
l
Fi
g
u
r
e
 10
6
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 GW
A
‐10
D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

Fi
g
u
r
e
 10
7
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 10
8
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 10
9
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 11
0
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
1
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
2
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
3
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
4
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
5
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
6
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
7
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 11
8
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Th
a
l
l
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

0510152025
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
V
a
n
a
d
i
u
m

u
g
/
l
Fi
g
u
r
e
 11
9
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AB
‐10
S
L
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

02468101214161820
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
V
a
n
a
d
i
u
m

u
g
/
l
Fi
g
u
r
e
 12
0
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 AS
‐4D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

012345678910
1
9
5
0
2
0
0
0
2
0
5
0
2
1
0
0
2
1
5
0
2
2
0
0
2
2
5
0
2
3
0
0
V
a
n
a
d
i
u
m

u
g
/
l
Fi
g
u
r
e
 12
1
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 (μ g/
L
)
 in
 Mo
n
i
t
o
r
i
n
g
 We
l
l
 MW
‐9D
 fo
r
 Mo
d
e
l

Sc
e
n
a
r
i
o
s
 1 ‐3
Sc
e
n
a
r
i
o
 1  Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
Sc
e
n
a
r
i
o
 2  Ca
p
‐in
‐Pl
a
c
e
Sc
e
n
a
r
i
o
 3  Ex
c
a
v
a
t
i
o
n
2L
/
I
M
A
C
 St
a
n
d
a
r
d

Fi
g
u
r
e
 12
2
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 12
3
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 12
4
.
 In
i
t
i
a
l
 (2
0
1
5
)
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zo
n
e

Fi
g
u
r
e
 12
5
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 12
6
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 12
7
.
 Ex
i
s
t
i
n
g
 Co
n
d
i
t
i
o
n
s
 Sc
e
n
a
r
i
o
 1  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 12
8
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m

Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Sh
a
l
l
o
w
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 12
9
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m

Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 De
e
p
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 13
0
.
 Ca
p
‐in
‐Pl
a
c
e
 Sc
e
n
a
r
i
o
 2  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 13
1
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 13
2
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 13
3
.
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o
 3  ‐

21
1
5
 Pr
e
d
i
c
t
e
d
 Va
n
a
d
i
u
m
 Co
n
c
e
n
t
r
a
t
i
o
n
s
 (μ g/
L
)
 in
 th
e
 Be
d
r
o
c
k
 Gr
o
u
n
d
w
a
t
e
r
 Zone  

Fi
g
u
r
e
 13
4
.
C
o
n
c
e
p
t
u
a
l
 Gr
o
u
n
d
w
a
t
e
r
 Fl
o
w
 Mo
d
e
l
/
M
o
d
e
l
 Do
m
a
i
n
‐Co
m
p
l
e
t
e
 Ex
c
a
v
a
t
i
o
n
 Sc
e
n
a
r
i
o