DSHW-2013-006916 - 0901a068803e8034

ATK
Division of
Solid and Hazardous Waste
DEC 11 2013
20 13-00^ |£,
Scott T. Anderson, Director
Department of Environmental Quality
Division of Solid and Hazardous Waste
ATTN: JeffVandel1
P.O. Box 144880
195 North 1950 West
Salt Lake City, Utah 84114-4880
Subject: M-136 Burning Grounds Perchlorate Contamination in Groundwater;
^171/0009081357; DSHW-2013-002295
Dear Mr. Anderson,
In the subject document, ATK Launch Systems Inc. (ATK) was requested by the
Division to evaluate perchlorate concentrations in groundwater at the M-136 burning
ground. It is well understood that Liquid Thermal Treatment Areas (LTTAs) in the M-
136 area were historically used to dispose of waste water containing perchlorate. As a
result, high concentrations of perchlorate in groundwater are found at this location. Over
a dozen monitoring wells are within the immediate area of M-136, and most of these
were installed between 1985 and 1992 prior to the development of a reliable perchlorate
analytical method. After initial monitoring of these wells, the focus for groundwater well
construction and monitoring was the outlying areas. As a result, many of the wells were
not sampled again until 2007 as part of a vegetable oil remediation pilot test. Therefore,
there was a lack of historic and valid perchlorate data. Since 2007 several of the M-136
wells have been routinely sampled based on higher perchlorate concentrations.
The Division has asked ATK to conduct an investigation to determine if the ongoing burn
ground operations could potentially cause an increase in perchlorate concentrations at the
site. To help answer this question, several sets of data were generated.
First, the M-136 monitoring well perchlorate data was summarized in trend charts for
wells C-1, C-2, D-2, D-4, D-5, A-5 and D-6 and provided to JeffVandel and Helge
Gabert. Overall, wells D-4 and C-1 showed an increasing perchlorate concentration trend
and the others show stable or declining concentrations. These trend charts are included in
attachment 1.
December 3, 2013
8200-FY14-052
Second, HYDRUS modeling (which models flow in unsaturated soil) was conducted to
determine if it were possible for precipitation to leach perchlorate from the surface to
groundwater. Conservative assumptions were used such as a 1000 year rainfall event, no
evaporation and a continuous layer of silt, sand or loam. Based on this modeling, storm
water would not reach groundwater. The HYDRUS modeling is included in attachment
2.
Third, the UDSHW asked if the HYDRUS model is applicable for an area with fractured
bedrock. Research was conducted by EarthFax Engineering that demonstrated that the
HYRDUS model would be even more conservative for an unsaturated zone consisting of
fractured bedrock. A summary of the research is included in Attachment 3.
Fourth, the UDSHW requested additional modeling of the M-136 area to better predict
perchlorate movement and connections between wells. This modeling was completed by
creating smaller grid spacing in the modeled zone to refine the modeling predictions.
The results of the modeling predict a travel time and directions between wells and that
water does tend to pool in the M-136 area and that contaminant slug flow would be
anticipated. The refined model results are included in Attachment 4.
If you have questions, or need additional information, please contact Paul Hancock at (435) 863-
3344.
Sincerely,
George Gooch, Manager
ATK Environmental Services
Attachment 1
Trend Charts for M-136 Monitoring Wells
A-5 Perchlorate Sen's Slope - No Trend
10000
9000
c o
"+•• ro
i— +-> c
u c o u <u co
O!
8000
7000
6000
5000
4000
3000
i 2000
1000
0 - —
r\\ r\\ r\\ SS1 <$?
Sample Date
—
I Groundwater
Level
95% Conf.
Min.
•Sen's
Estimate
4295
4294
4293
4292-
4291
4290
4289
4288
4287
4286
>
V +•> CO
<5 •a c
3
o
C-1 Perchlorate Sen's Slope - Increasing Trend
30000
25000
20000
£ 15000
10000
5000
Q_
vs.232.88x + 5552.3
• Concentration
(ug/L)
Sen's Estimate
95% Conf. Max
95% Conf. Min.
939.54x + 2007.4
946.2x- 1537.5
i 1 r i~~—i i
j? \jy jy [jy to? jy jy .jy J?\jy J? jy jy jy jy jy jy jy jy
Sample Date
C-2 Perchlorate Sen's Slope - Decreasing Trend
600
500
400
c o
ro
i-+•>
c
u c o u
0J
ro
300
200
100
<u a.
Concentration
(ug/L)
•Sen's Estimate
•95% Conf. Max
•95% Conf. Min.
= -22.737x + 539.42
^=-22.925x + 449.15
_y_=-23.113x + 358.88
^ «fy <<v & & r> A A A &> r?» c$ \> x> O O J§> J§> .rgP .^P v~cP .~oN ^
Sample Date
D-4 Perchlorate Sen's Slope - Increasing Trend
180000
160000
140000
120000
c
•2 IOOOOO
ro
+•>
c <U 80000 u c o
60000
+->
(0
^ 40000
(V CL 20000
y = 5029.8X + 108640
y = 5076.4X + 103451
y = 5122.9x +98261
sir ^ ^ .nov ~ov ~ov
vnov ~o <$P J> ^ J? J? ^
^\ vo\ <cA & # ^ ^ <f ^ #
Sample Date
• Concentration
(ug/L)
Sen's Estimate
95% Conf. Max
95% Conf. Min.
D-5 Perchlorate Sen's Slope - No Trend
300000 -i
250000
200000
c o
'+•> ro i_ +-> c 0J u c o u OJ +J
ro
OL
150000
100000
50000
/ / /
<v? "S5 \> \> \> O O
^ ^ ^ ^ .J§»
r / / X / / ^
Sample Date
y = 2591.3x + 212328
y = 3119.9x + 172641
y = 3648.5X + 132953
• Concentration
(ug/L)
Sen's Estimate
95% Conf. Max
95% Conf. Min.
D-6 Perchlorate
160000
140000
120000
100000 c o
ro 80000
c 1 u
§ 60000
u m -t->
£ 40000
o
OJ 20000
i3
Sample Date
Sen's Slope - Decreasing
3<
• Concentration
(ug/L)
Sen's Estimate
95% Conf. Max
95 %Conf. Min.
366.9x + 156748
549.4x + 136321
5732x + 115895
3>
Attachment 2
HYDRUS Modeling for M-136
May 8, 2013
EarthFax
Mr. Paul Hancock
ATK Space Launch Systems
PO Box 707
Brigham City, Utah 84302-0707
EarthFax
Engineering, Inc.
Engineers/Scientists
7324 So. Union Park Ave.
Suite 100
Midvale, Utah 84047
Telephone [phone]
Fax [phone]
www. earthlax. com
Subject: Potential for continued leaching of perchlorate
from the surface to groundwater at M-136
Dear Paul:
Pursuant to your request, we have evaluated the potential for continued leaching of perchlorate
from the surface to groundwater at the M-136 burning ground. Our investigation involved
modeling of flow in the unsaturated zone and an assessment of perchlorate concentration trends
in groundwater at M-136. The results of this evaluation are presented below.
HYDRUS Modeling
We modeled flow in the unsaturated zone using the one-dimensional, finite-element model
HYDRUS. In this model, we assumed that the underlying bedrock could be modeled as a
continuous layer of either silt, loam, or sand. We did not model that portion of the site that has a
clay cap, since those results would be more conservative than the remaining models.
Under each of the modeled soil conditions, we assumed that the site would be subjected to
rainfall for 60 days from each of the following events:
• 50-year, 60-day event: 8.80 inches
• 100-year, 60-day event: 9.44 inches
• 1000-year, 60-day event: 11.3 inches
We obtained the appropriate precipitation amounts from the online National Weather Service
Hydrometeorological Design Studies Center Precipitation Frequency Data Server for Utah
(http://hdsc.nws.noaa.gov/hdsc/pfds/pfds map cont.html?bkmrk=uf). We assumed that
precipitation would fall at a constant rate during the 60-day period and modeled the advance of
the wetting front for a period of 1 year thereafter. The minimum depth to groundwater at M-136
is approximately 230 feet (70.1 m), which depth was set as the lower limit of the HYDRUS
model.
The results of this modeling effort are presented in Attachment A and summarized in Table 1.
As indicated, the maximum depth of the wetting front under the modeled conditions is 12.6 m
(41.3 ft) under both the silt and sand assumptions one year after the beginning of the 1000-year,
60-day precipitation event. The maximum depth of moisture change (indicating some water
movement in the unsaturated zone) is 19.6 m (64.3 ft) under the silt assumption one year after
the beginning of the 1000-year, 60-day precipitation event. With a minimum depth to
groundwater at the site of 230 feet, it is reasonable to conclude that leaching of perchlorate from
the surface to groundwater at M-136 is not currently occurring.
Paul Hancock
May 31,2013
Page 2
Trends in Perchlorate Concentrations
Groundwater samples are periodically collected by your staff from monitoring wells installed at
the M-136 burning ground. Trends in perchlorate concentrations at those monitoring wells were
evaluated using Sen's slope estimator, a method that selects the median slope among all lines
through pairs of sample points. In this case, the samples pairs consisted of perchlorate
concentration versus time.
The resulting Sen's slope evaluation is presented in Attachment B. These data indicate that
concentrations of perchlorate in groundwater are generally decreasing (wells C-2, D-2, and D-6)
or not statistically trending (wells A-5, A-6, and D-5) in wells that are upgradient or cross
gradient from the majority of the burning ground. Wells that are immediately downgradient
from the burning ground (C-1 and D-4) are exhibiting increasing trends in perchlorate
concentrations.
With particular reference to data collected from wells D-6 and C-1, these data suggest that
perchlorate is moving through the area as a slug. At D-6, perchlorate concentrations have
generally decreased from a range of about 120,000 to 140,000 ug/L to a range of about 40,000 to
80,000 ug/L during the period of October 2007 through October 2012. During the period of May
2000 through October 2012, concentrations at C-1 (located about 500 feet downgradient from D-
6) have increased from a range of about 5,000 to 10,000 ug/L to a range of about 20,000 to
25,000 ug/L.
In the past, perchlorate-contaminated water was disposed of at the burning ground. During the
time that this water remained in surface sumps prior to burning, it infiltrated the underlying
fractured bedrock and impacted the groundwater. However, disposal of contaminated water at
the burning ground was ceased several years ago. Thus, the above-noted data (together with the
HYDRUS model results presented above) indicate that groundwater has not received any direct
inputs of perchlorate during the period of evaluation. It is therefore reasonable to conclude from
the groundwater monitoring data that perchlorate is migrating beneath the burning ground as a
slug of contamination rather than through the input of additional perchlorate from the surface.
Please feel free to contact me or Kris Blauer of our office if you have any questions.
Sincerely,
Richard B. White, P.E.
President
Attachments
Paul Hancock
May 31, 2013
Page 3
TABLE 1
Summary of M-136 HYDRUS Model
Modeled
Structure
Silt
Loam
Sand
60-Day Precipitation
Return Period
50
100
1000
50
100
1000
50
100
1000
Amount
(in)
8.80
9.44
11.3
8.80
9.44
11.3
8.80
9.44
11.3
Percolation Depth After 1 Year
Wetting Front
(m)
11.2
11.9
12.6
11.2
11.2
11.9
10.5
11.2
12.6
Moisture Change
(m)
18.2
18.2
19.6
15.4
16.1
16.8
11.9
12.6
14.0
ATTACHMENT A
HYDRUS Model Results
Figure 1. Water content change with depth at different time periods
(Silt; 50-yr, 60-day storm event)
0.35 0.36 0.37 0.38 0.39 0.4
—I—
0.41 0.42
-Day 0
-After 60 days
-After 1 year
Water content (theta)
Figure 2. Water content change with depth at different time periods
(Silt; 100-yr, 60-day storm event)
0.35 0.36 0.37 0.38 0.39 0.4 0.41 0.42 0.43
-10
-1010
-2010
-3010
f
Q -4010
-5010
-6010
-Day 0
-After 60 days
- After 1 year
-7010
Water content (theta)
Figure 3. Water content change with depth at different time periods
(Silt; 1000-yr, 60-day storm event)
0.35 0.36 0.37 0.38 0.39 0.4 0.41 0.42 0.43
-10
-1010
-2010
-3010
-4010
-5010
-6010
-•—Day 0
-•—After 60 days
-A— After 1 year
-7010
Water content (theta)
-10
-1010
-2010
_ -3010
E
f
° -4010
-5010
-6010
Figure 4. Water content change with depth at different time periods
(Loam; 50-yr, 60-day storm event)
0.05 0.1 0.15 0.2 0.25 0.3 0.35
-Day 0
-After 60 days
- After 1 year
-7010
Water content (theta)
Figure 5. Water content change with depth at different time periods
(Loam; 100-yr, 60-day storm event)
0.05 0.1 0.15
-10 +
-1010
-2010
-3010
Q.
Q -4010
-5010
-6010
-7010
0.2
—I—
0.3 0.25 0.35
-•—Day 0
-•-After 60 days
-*— After 1 year
Water content (theta)
Figure 6. Water content change with depth at different time periods
(Loam; 1000-yr, 60-day storm event)
-10
-1010
0.05 0.1 0.15 0.2
—H—
0.3 0.25 0.35
-2010
-3010
E
£
Q.
5
Q -4010
-5010
-•—Day 0
-•—After 60 days
-A—After 1 year
-6010
-7010
Water content (theta)
Figure 7. Water content change with depth at different time periods
(Sand; 50-yr, 60-day storm event)
-10
-1010
0.02 0.04 0.06 0.08 0.1
—h
0.12 0.14
-2010
_ -3010
I
a
Q -4010
-5010
-Day 0
-After 60 days
-After 1 year
-6010
-7010 Water content (theta)
Figure 8. Water content change with depth at different time periods
(Sand; 100-yr, 60-day storm event)
-10
-1010
0.02 0.04 0.06 0.08 0.1 0.12 0.14
-2010
-3010
?
i
Q -4010
-5010 -Day 0
-After 60 days
- After 1 year
-6010
-7010
Water content (theta)
Figure 9. Water content change with depth at different time periods
(Sand; 1000-yr, 60-day storm event)
0.02 0.04 0.06 0.08 0.1 0.12 0.14
-•—Day 0
-•-After 60 days
-*—After 1 year
Water content (theta)
ATTACHMENT B
Results of Sen's Slope Estimator Evaluation
A-3
C-6
HAZARDOUS
WASTE
AREA
9
• •
9
A-1
D
9 c3
•
M 36
M186 M381
A-8
LEGEND
S MONITORING WELL
0' 400'
M-136 LOCATION MAP
Sample Date Concentration (mp/l) Groundwater Levels
4/26/2000 2300 4291 46
A-5 Perchlorate
9/20/2000
4/19/2001
10/16/2001
5/13/2002
9/25/2002
4/30/2003
5/10/2004
5/3/2005
5/8/2006
9/6/2007
10/18/2007
11/15/2007
6/13/2008
9/24/2008
6/22/2009
6/28/2011
6/7/2012
10/30/2012
3100
3790
4180
4160
4780
6210
4360
4720
3340
3290
2210
2310
2500
9360
3080
2930
4291.21
4290.76
4290 22
4290 1
4289 75
4289 58
4289 03
4290 26
4291.21
4293 96
4293 96
4293 96
4292 91
4292 45
4292 03
4291 51
4291 9
n
E
S
c
8
t
s .
loco
Sen's Slope • No Trend
—Sans Ci!ira!f
ill*
• •
an $
3 • >5
///////////////////
Sample Dete
Sample
Date
9/27/2000
4/26/2001
10/18/2001
5/15/2002
4/28/2003
5/12/2004
5/3/2005
5/8/2006
10/8'2008
10/25/2011
Concentration
W)
22000
21000
17000
15000
13000
15500
11800
32700
23100
21000
Groundwater
Levels
4292 67
4292.17
4291 94
4291 3S
4291 48
4290 51
4291 6
4295 08
4294 26
A 6 Perchlorate Sen's Slope - No Trend
If. Gwrt.Min.
fell tftiMI
- if, Cent m
-JJ'.Ccnf.Mifi
J / § S / J / / /
Sample
Date
5/10/2000
9/19/2000
4/24/2061
10/15/2001
5/15/2002
4/30/2003
Concentration
(mg/L)
6050
Groundwater
Levels
4290 01
C-1 Perchlorate Sen's Slope-Increasing Trend
7800
7300
6430
lm
4930
5'12'2004 4670
4/13/2005
6/14/2006 '
7/30/2007
10/16/2007
11/15/2007
6/16/2008 '
9/23/2008 '
10/14/2008
6/16/2009 '
6/28/2011 '
10/27/2011
5/30/2012
10/30/2012
6340
0
11800
14900
14700
14900
13800
12600
19500
26300
20800
21500
20600
4289.83
4289.17
4288 72
4288 33
4288 22
428746
4288.41
4293 76
4293 18
4292 34
4292 34
4291.57
4291 08
4291 08
4290.63
9/29'1911
4290 32
////////////////////
Sample Concentration
(mg/L)
Groundwater Level
5/11/2000
9/25/2000
4/24/2001
10/25/2001
5/13/2002
9/26/2002
4/28/2003
5/10/2004
5/3/2005
5/8/2006
7/19/2007
10/18/2007
11/15/2007
6/13/2008
9/23/2008
6/19/2009
6/28/2011
6/7/2012
10/30/2012
300
500
400
397
361
371
392
103
462
278
72
50
44
20
38
64
93
93
84
4291
4290
4290
4289
4289
4290
4288
4289
4293
4294
4293
4293
4292
4291
4291
4291
4292
C 2 Perchlorate Sen's Slope - Decreasing Trend
/ // / // / // / / / J / / .
ffff/fwf * * §/ rvf
InistiMIHli ti IK
Gout Mi
Sample Date Concentration (ug/L) Groundwater Level
D 2 Perchlorate Sen's Slope - No Trend
7/19/2007
10/18/2007
11/12/2007
5/10'2008
6/12/2008
9/12/2008
6/16/2009
6'28'2011
10/25/2011
6/7/2012
1260
235
838
706
588
476
387
~73T
411
662
4295 32
4294 27
4294 27
4293 54
4293 54
42932
4292.9
4291 86
4291 86
4292.27
...
/ / / / / / /
<jto»d.ateil£ d
I . - i - -.v L
n<M.n
K1 (oil n.
Sample Date Concentration (mg/L) Groundwater Level D-4 Perchlorate
07/31/2007
10/15/2007
10/07/2008
5/13/2009
6/2/2010
12'1/2010
7/7/2011
10/25/2011
5/30/2012
10/30/2012
112000
115000
124000
135000
130000
150000
139000
149000
157000
429743
4297 06
4295 19
4294 69
429344
Sen's Slope - Increasing Trend
4293.25
4294 65
4294 4 5
c o
5 ROM
c
s
if i ton Mil
Mil Erwtt
lf< »M M|.
—P9S<M i-'r
Sample Date
10/15/2007
10/07/2008
5/13/2009
6/2/2010
12/1/2010
6'28/2011
10/25/2011
5/30'2012
10/30'2012
Concentration (mg/l) Groundwater Level
174000
206000
162000
200000
168000
155000
244000
181000
162000
4296 01
4294 02
4293 48
4292 51
10/1/1911
4293 21
4293 16
D 5 Perchlorate Sen's Slope-No Trend
/ 4 i
Sample Date Concentration (mg/L) Groundwater Level
10/16/2007
12'11/2007
3/10'20C8
6/12'2008
9/24/2008
5/14/2009
8/23/2010
6'28/2011
10/25/2011
5/30/2012
10/30/2012
128000
135000
129000
112000
108000
91600
73400
124000
104000
67600
41400
4294 83
4293 65
4292 69
4292 69
4292 35
4291 84
4291 28
4291 97
4291 72
D-6 Perchlorate Sen's Slope Decreasing
S
' / / / / / / / / / / / / rfN ^ t f f / / / /
Attachment 3
Research on Applicability of HYDRUS Modeling with a
Fractured Bedrock Unsaturated Zone
May 8, 2013
EarthFax
Mr. Paul Hancock
ATK Space Launch Systems
PO Box 707
Brigham City, Utah 84302-0707
Subject: Appropriateness of HYDRUS to model flow in the
unsaturated zone at Ml 36
EarthFax
Engineering, Inc.
Engineers/Scientists
7324 So. Union Park Ave.
Suite 100
Midvale, Utah 84047
Telephone [phone]
Fax [phone]
www. earth/ax. com
Dear Paul:
In comments received from the Utah Division of Solid and Hazardous Waste, the agency has
raised a concern regarding the appropriateness of using HYDRUS to model flow in the
unsaturated zone at the Ml 36 burning grounds at ATK's Promontory facility. In particular, the
agency raised the concern of whether or not HYDRUS (which models flow in unsaturated soil)
would be applicable in an area of fractured bedrock. The purpose of this letter is to address that
concern.
With respect to flow in unsaturated, fractured media, Preuss and Wang1 refer to the "role reversal
between fractures and matrix." These authors point out that "although the saturated permeability
of the rock matrix is several orders of magnitude smaller than the saturated permeability of the
fractures, it is likely that during desaturation, the effective permeability of fractures will become
smaller than that of the matrix. An interesting consequence of this role reversal between
fractures and matrix in transporting liquid is that water will tend to flow across the fractures at
asperity contacts from one matrix block to another instead of flowing along the fractures. The
flow lines may be expected to circumvent drained portions of the fractures. . . . The flow lines
bypass the drained portions of the fractures, going from one matrix block to another normal to
the fracture planes." This occurs due to the fact that the suction pressure (or capillarity) of the
matrix is more negative than that of the fracture. As a result, the water cannot enter the fracture,
effectively causing the fracture to serve as a barrier to flow in the unsaturated zone rather than a
conduit of flow. Since the opposite condition occurs in the saturated zone, this may seem
counterintuitive (and therefore has been termed as a "role reversal" by Preuss and Wang). This
fact of soil physics is the reason that capillary breaks have been used to successfully isolate fine-
grained waste from its overlying cover soil in some waste-isolation designs.
Air also serves to block the flow of liquid in an unsaturated fracture. According to Preuss and
Wang, "within a partially saturated fracture, the remaining water is held in sections with small
apertures near the contacts, and the liquid phase may be surrounded by air. The presence of a
relatively continuous air phase produces an almost infinite resistance to liquid flow parallel to the
1 Preuss, K. and J.S.Y. Wang. 2001. Numerical Modeling of Isothermal and Nonisothermal Flow in
Unsaturated Fractured Rock: A Review, pp. 19-32 in Flow and Transport Through Unsaturated Rock,
Second Edition. Geophysical Monograph 42. American Geophysical Union. Washington, D.C.
Paul Hancock
May 8,2013
Page 2
fracture plane. Therefore, as the fracture begins to desaturate, its effective permeability declines
abruptly by many orders of magnitude as the pressure head decreases."
It could be argued that the conservative assumptions of the M136 HYDRUS model (a 25-year,
24-hour precipitation event without the benefit of evaporation) would result in ponding that
would cause the fractures to become saturated, thereby negating the "role reversal" of the
unsaturated zone. However, citing Travis et al.,2 Preuss and Wang point out that under
conditions of intense infiltration events, "slugs of water injected into fractures will penetrate into
the [bedrock] only over short distances before being sucked into the matrix."
Citing modeling work performed by Wang and Narasimhan in a fractured, unsaturated tuff,
Preuss and Wang noted that "fluid flow in the fractured tuff column was nearly identical to
simulation results for the same column without taking the fractures into account. As soon as the
fractures are drained, the transport of fluid will be through the matrix and will be controlled by
the characteristic curves of the matrix. At a given elevation, the pressure values in the fractures
are nearly equal to the pressure values inside the matrix blocks. If the fracture pressures are the
same as the matrix pressures, there is no need to model separately the fractures and the matrix."
Given these factors, it is my opinion that the HYDRUS model adequately simulates flow in the
unsaturated zone at Ml 36. In fact, these factors point to the added conservativeness of the
predictions made using HYDRUS.
Please let me know if you have any questions regarding this information.
Sincerely,
Richard B. White, P.E.
President
EarthFax Engineering, Inc.
2 Travis, B.J., S.W. Hodson, H.E. Nuttall, T.L. Cook, and R.S. Rundberg. 1984. Numerical Simulation of
Flow ad Transport in Fractured Tuff. Materials Research Society, Symposium Proceedings. 26:1039-
1047. Elsevier. New York.
3 Wang, J.S.Y. and T.N. Narasimhan. 1985. Hydrologic Mechanisms Governing Fluid Flow in a Partially
Saturated, Fractured, Porous Medium. Water Resources Research. 21 (12): 1861 -1874.
Attachment 4
Refined Groundwater Modeling for M-136
M-136 Groundwater Modeling
At the request Division of Solid and Hazardous Waste, ATK tasked EarthFax
Engineering to complete a refined modeling effort at the M-136 (burning grounds) area of the
Promontory Utah facility. The request was for additional groundwater modeling predictions at
the M-136 site. The existing model grid at the Promontory facility was set at a uniform spacing
of 200 ft by 200 ft. This refined modeling included reducing the grid spacing in the M-136 area
only to 25 feet by 25 feet. The intent of this change in grid spacing was to gain a more accurate
portrayal of groundwater movement within the burning grounds and specifically between wells.
The M-136 area is approximately 1,000 feet by 1,200 feet and would be covered by
approximately 30 grids. The new grid spacing would put approximately 478 x 497 grid cells in
the M-136 area. Grid spacing is shown on the attached Figure 1.
The calibration goal for the refined model was to achieve similar results as the original
model. In the original site model, the maximum acceptable Mean Average Error (MAE) was
defined to be 10 feet. The maximum acceptable Root Mean Squared (RMS) goal was to be less
than 5 ft. Both of these criteria were met. The Normalized Root Mean Squared value was higher
than the target of 5%. This was caused by modeling the perched aquifer with the regional aquifer
and it presented some challenges trying to achieve model convergence during calibration.
Calculated vs. Observed Head for the model is shown in Figure 2. Figures 3 and 4 show
calculated vs. observed concentrations of perchlorate and TCE.
Following calibration of the tighter grid spacing at M-136, the model was allowed to run
to determine a travel time from well D-5 to well D-4. At 100 days, the particles start to move
from the D-5 region. At 390 days, they reach D-4. So the travel time is about 290 days from D-5
to D-4.
Similarly, the water particles at D-6 begin to move at 290 days into the simulation. They
quickly travel through the fractures toward C-1 and can reach there at 460 days. (The tracking
particles would presumably not actually pass through well C-1 because fractures in the limestone
would intercept the flow). The final particle track shows no particles at C-1. The travel time
ATK Launch Systems
Promontory UT M-136 Model Summary
should be around 170 days.
A particle track exercise was undertaken with the model to show flow paths in the M-136
area. The attached map shows the particle track for particles from the two aforementioned areas
only with an overlay of the 2012 water surface showing flow paths.
ATK Launch Systems
Promontory UT M-136 Model Summary
•••MM
nt TO
FIGURE 1
ATK Launch Systems
Promontory UT M-136 Model Summary
Calculated vs. Observed Head : Steady state
52 _
t A
OJ LU
U CO
CO -
o
A
V
Layer #1
Layer #2
Layer #3
Layer #4
95% confidence interval
95% interval
-4.4351838E14 9.55648162E15
Observed Head (ft)
1.95564816E16
Max. Residual: 2.217592E16 (ft) at F-4/1
Min. Residual: 0.256 (ft) at EW-5/1
Residual Mean : 3.400547E14 (ft)
Abs. Residual Mean : 3.400547E14 (ft)
Num. of Data Points : 97
Standard Error of the Estimate : 2.532854E14 (ft)
Root Mean Squared : 2.50487E15 (ft)
Normalized RMS : 7.294109E14 ( % )
Correlation Coefficient: -0.076
FIGURE 2
ATK Launch Systems
Promontory UT M-136 Model Summary
Calculated vs. Observed Concentration : Time = 18980 days
o
m -
o
O
A
Layer #2: ConcOOl
Layer #3 : ConcOOl
Layer #4: ConcOOl
95% confidence interval
95% interval
i i
95.8
Observed Concentration (mg/L)
195.8
Max. Residual: -16.144 (mg/L) at A-6/CONCENTRATION
Min. Residual: 0 (mg/L) at FISH SPRING/A
Residual Mean : -0.348 (mg/L)
Abs. Residual Mean : 1.546 (mg/L)
Num. of Data Points : 84
Standard Error of the Estimate : 0.415 (mg/L)
Root Mean Squared : 3.799 (mg/L)
Normalized RMS : 1.844 ( % )
Correlation Coefficient: 0.992
FIGURE 3
ATK Launch Systems
Promontory UT M-136 Model Summary
Calculated vs. Observed Concentration : Time = 18980 days
//
i i
-0.42 9.58
Observed Concentration (mg/L)
—\—
19.58
• Layer #2 : ConcOOl
• Layer #3 : ConcOOl
A Layer #4 : ConcOOl
95% confidence interval
95% interval
Max. Residual: 3.853 (mg/L) at A-2/A
Min. Residual: 0 (mg/L) at M-636B1/A
Residual Mean : 0.031 (mgA.)
Abs. Residual Mean : 0.394 (mg/L)
Num. of Data Points : 84
Standard Error of the Estimate : 0.09 (mg/L)
Root Mean Squared : 0.817 (mg/L)
Normalized RMS : 6.333 ( % )
Correlation Coefficient: 0.905
Figure 4
ATK Launch Systems
Promontory UT M-136 Model Summary
c 8
S> 37
3 &
9 A
A-3
A 6 D-l
D B-9 4 C-4 5
C B-10
A B
C 5 B 2
ro 4293 6p(j) 0' \
EarthFax POTENTIOMETRIC SURFACE AND MODEL PARTICLE TRACK
FIGURE 5