DSHW-2019-000862 - 0901a06880999484

Div of Waste Management
and Radiation Control
January 21, 2019 JAN 2 2 2019
8200-FY19-003
sY-Alifil-2401,-000e62-
Scott T. Anderson, Director
Department of Environmental Quality
Division of Waste Management and Radiation Control
ATTN: Jeff Vandel
P.O. Box 144880
195 North 1950 West
Salt Lake City, Utah 84114-4880
RE: Post-Closure Permit, ATK Launch Systems Inc. - Promontory Facility, EPA ID#
UTD009081357, Response to DWMRC Groundwater Modeling Comments
Dear Mr. Anderson,
ATK Launch Systems Inc. Promontory facility is submitting with this letter in response to the
UDWMRC November 8, 2018 comments on the updated groundwater flow and contamination
transport model.
If you have questions, or need additional information, please contact Paul Hancock at (435)
863-3344.
Kris H Blauer, Manager,
Environmental Services
ATK Launch Systems, Inc. • P.O. Box 707 Brigham City, Utah 84321 • [phone]
RESPONSE TO
NOVEMBER 8, 2018 COMMENTS
FROM THE UTAH DIVISION OF WASTE MANAGEMENT
AND RADIATION CONTROL CONCERNING THE
UPDATED GROUNDWATER FLOW AND
CONTAMINANT TRANSPORT MODE,
ATK PROMONTORY, UTAH FACILITY
DWMRC Comment #1: The text states, "The agency has recently indicated that they are now
comfortable with the model showing separate aquifers if needed to provide better predictive
results. This current model update was performed with the separate aquifers in mind." Please
explain how the separate aquifers are set up in the six-layer model. Was the flow model
modified? Is a specific layer set up to represent the perched aquifer everywhere in the model
domain, followed by a low conductivity zone to represent an aquitard? Does its thickness vary
over the model domain? Please elaborate.
Response: Layers in the 2018 model update were consistent with the 2008 model (EarthFax
Engineering, 2008). These layers are shown schematically in Figure 2-1 (attached) from that
report. As shown, groundwater flow and contaminant transport at the Promontory facility were
modeled using six layers generalized as follows:
• Layer 1 - Fine-grained, low-permeability unconsolidated materials overlying the regional
groundwater system
• Layer 2 - Unconsolidated coarser-grained sediments of the regional groundwater system
• Layer 3 - Generally unconsolidated sediments containing perched groundwater
• Layer 4 - Low-permeability, generally unconsolidated, sediments underlying perched
groundwater
• Layer 5 - Fractured bedrock
• Layer 6 - Unfractured bedrock
At the request of the Division, the 2013 model combined the perched aquifer with the regional
aquifer. This modification was reversed for the 2018 model, which again evaluated the perched
aquifer and its underlying low-permeability layer consistent with the 2008 model. The flow
model was recalibrated to account for this change.
Consistent with the 2008 model, the thickness of the low-permeability layer underlying the
perched aquifer was set at a constant value of 1 foot. Since the perched aquifer is unconfined,
the thickness of this layer is defined by the elevation of the groundwater surface relative to the
elevation of the underlying low-permeability layer.
DWMRC Comment #2: The text states that calibration targets for the mean average error,
normalized root mean square error, and mass balance error were all achieved. Is it possible to
supply calibration charts for the flow model (e.g., observed vs. modeled heads) and the transport
model (e.g., observed vs. modeled TCE and perchlorate concentrations)? These charts seem to
1
be missing in the appendices. What is the rationale for performing linear interpolation from
measured values in model cells to estimating perchlorate mass concentrations, as concentration
data can be non-linear? Please elaborate.
Response: Calibration charts for head, perchlorate, and TCE are attached as Figures 1, 2, and 3,
respectively. Calibrated steady-state heads ageed well with the observed heads in each layer
except Layer 3 (the perched groundwater layer). This was not unexpected given the steep
hydraulic gradient of groundwater in the perched aquifer, particularly in the area southeast of the
burning ground (see Plate 2-2 of EarthFax Engineering, 2008).
Figures 2 and 3 show excellent agreement between observed and calibrated perchlorate and TCE
concentrations. These concentrations were predicted by the transient model at a time of 22,265
days (61 years), which represents the approximate time to 2018 since site operations began in
1957.
The mass of perchlorate in groundwater underlying M-136 was estimated at 61 points based on
data collected from 11 monitoring wells. This required that estimates be made in multiple
coordinate directions based on the closest set of monitoring wells. Although perchlorate
concentrations may vary non-linearly at the site, the data are insufficient to determine the degree
of that non-linearity for all of the directional combinations that were needed. Thus, the data did
not justify estimates by any means other than linear.
DWMRC Comment #3: The text states that two alternatives were considered:
Alternative 1: Full and instantaneous remediation of all source areas.
Alternative 2: No remediation of or continued contaminant input at any source area.
It is unclear which alternative is recommended, or if there is any alternative in between these
two bounds which is preferred. Were the predictions in both cases run without any perchlorate
emanating from M-136-associated source areas entering the model domain? Please elaborate.
Response: These were alternative model scenarios, chosen to represent the range of anticipated
outcomes represented by full source remediation and no source remediation. Other scenarios
between these two bounds could have been modeled, but these scenarios were considered
adequate for representing the range of potential outcomes. The purpose of the alternative model
scenarios was not to develop a recommended remediation approach but rather to estimate the
range of potential outcomes that might be expected from natural attenuation with and without
source remediation.
Alternative 1 assumed that all perchlorate was removed instantaneously from the source area,
while Alternative 2 assumed that no perchlorate was removed from the source area. In each
case, it was assumed that no active remediation of groundwater downgradient from the source
area would occur. Thus, the alternative scenarios represent the range of potential outcomes
expected from natural attenuation with and without source remediation.
2
Reference
EarthFax Engineering, Inc. 2008. Groundwater Flow and Contaminant Transport Model Report
for the ATK Promontory Facility. Revised project report submitted to ATK Launch
Systems, originally submitted in December 2005. Midvale, Utah.
3
Calculated vs. Observed Head : Steady state
• Layer *I
• Layer*2
• Layer *3
✓ Layer *4
95% confidence interval
95% inteival
4237.1 4337.1
Observed Head (ft
Max. Residual: -187.843 (ft) at J-3/A
Min. Residue 0.256 (ft) at EW-511
Residual Mean : 0.476 (ft)
Abs. Residual Mean : 13.317 (ft)
Num. of Data Points : 88
Standard Error of the Estimate : 3.237 (11)
Root Mean Squared : 30.196 (11)
Normezed RMS : 13.387 ( % )
Correlation Coefficient : 0.625
FIGURE 1. HEAD CALIBRATION CHART, ATK PROMONTORY GROUNDWATER MODEL
==.=
I 41,11\11 I I • • kVA
EarthFax
Calculated vs. Observed Concentration : Time = 22265 days
• Layer #2 : Conc001
• Layer #3 : Conc001
A Layer #4 : Conc001
9E% confidence interval
9E% interval
9E.3
Observed Concentration mg/L)
Max. Residual: 2.887 (mg/L) at B-4/CONCENTRATION
Min. Residual: 0 (mg/L) at LF-4/A
Residual Mean : o.osa (rngiL)
Abs. Residual Mean : 0.421 (mg/Li
Num. of Data Points : E2
Standard Error of the Estimate : 0.11 (mg/1..;
Root Mean Squared : 0.788 MO.::
Normalized RMS : 0.37 %
Correlation Coefficient : 1
FIGURE 2. PERCHLORATE CALIBRATION CHART, ATK PROMONTORY GROUNDWATER MODEL.

EarthFax
Calculated vs. Observed Concentration : Time = 22265 days
Layer *2 : Con c001 4, Layer *3 : Conc001
A Layer #4 : Conc001
V Layer #E : Conc001
Layer:* : Conc001
9E341 confidence interval
95% interval
Observed Concentration (mg/L

Max. Residual: -0.45 (mg,'L'i at B-6/CONCENTRATION
Min, Residual: 0 (mg/L) at M-636B1/A
Residual Mean : 0.003 (mg/12.;
Abs. Residual Mean : 0.081 (mg;L
Num. of Cate Points : 55
Standard Error of the Estimate : 0.018 (mg/L)
Root Mean Squared : 0.131 (mg/L)
Normalized RMS : 1.671 ( % )
Correlation Coefficient : 0.996
FIGURE 3. TCE CALIBRATION CHART, ATK PROMONTORY GROUNDWATER MODEL.
II=0 MINIM =MP IMMOMM 4.16. NEM W rM 1• '
EarthFax
Layer 3
erched Aquifer
Ground Surface Potentiometric Surface
P ik er 1
•
Confining Clay
Layer:4 C6nfirfiri9. !Layer:.
Layer 5
Fractured Bedrock
Layer 6
Unfractured Bedrock
NOT TO SCALE
kVA
FIGURE 2-1. GENERALIZED HYDROGEOLOGIC CROSS-SECTION EarthFax