DSHW-2012-004045 - 0901a068802ea97e

January 18, 2012

CERTIFIED MAIL
RETURN RECEIPT REQUESTED
Dave P. Gosen, P.E. 7005 -0390 000 7508 5968
Director, Environmental Services
ATK Launch Systems – Promontory
P.O. Box 707
Brigham City, UT 84302-0707

RE: Preliminary Air Dispersion Modeling Assessment Report for Open Burn and Open Detonation Treatment Units
ATK Launch Systems – Promontory Facility
UTD009081357

Dear Mr. Gosen:

The Division of Solid and Hazardous Waste has completed its review of ATK’s Air Dispersion Modeling Assessment Report that was submitted on October 6, 2011. The modeling assessment
is being conducted for the M-136 and M-225 open burn and open detonation treatment units and is a required component of the Human Health Risk Assessment that will be completed for these
units.

Enclosed with this letter are technical review comments as well as two recommendations for reducing the air modeling results for multi-hour averaging periods. The comments primarily
focus on the presentation of the air modeling results and the modeling effort that was performed for both complex terrain and flat terrain.

The results of the air dispersion modeling analysis are based on treatment quantities that far exceed what are currently allowed under the Thermal Treatment Permit Module. In the second
part of the review, two modifications to the air modeling analysis are proposed that should generate more realistic modeling results for multi-hour averaging periods. The first modification
is focused on reducing the amount of waste modeled per event. A further reduction is proposed based on reducing the number of events modeled during a year and is tied to implementation
of the second proposed modification. In configuring the input files for OBODM, ATK has restricted the number of hours OBODM evaluates in calculating annual averages. By eliminating
these restrictions and allowing the amount of waste treated at each source to vary over each hour in the meteorological data file, a significant reduction in annual average air quality
impacts should be realized. An example of this approach (developing an hourly source strength file) is provided on the enclosed CD. Implementing this approach also allows ATK to limit
the number of events per year to the number needed to reach the proposed annual permit limit for the source being modeled.

Neither of the modifications proposed above directly impacts the modeled one-hour peak concentrations. Based on the preliminary model results, it appears that ATK will need to consider
lower daily maximum treatment quantities. One-hour peak concentrations are discussed in a section at the end of the review.

The submittal of the Air Dispersion Modeling Report and the Human Health Risk Assessment Report is required by Permit Conditions IV.M.1 and IV.M.2. Condition IV.M.1 states, “If the
Executive Secretary does not approve the Waste Characterization and Air Dispersion Modeling Report, he shall provide written comments to the Permittee identifying the deficiencies in
the document. The Permittee shall address the comments and submit a revised Waste Characterization and Air Dispersion Modeling Report to the Executive Secretary for approval within
60 days of receipt of written comments.”

The “Waste Characterization” portion of the report was approved with the Air Dispersion Modeling Protocol on June 1, 2011. The “written comments” on the Modeling Report, which include
recommendations for modifying the air dispersion modeling approach are enclosed. Therefore, ATK is required to address the comments and submit a revised Air Dispersion Modeling Report
to the Division for approval within 60 days of receiving this correspondence. The comments and recommendations for modifying the modeling approach should be addressed by ATK by providing
a response for each comment and by providing a response or plan that details how the modeling approach will be modified.

In an effort to ensure that ATK and the Division are in agreement with the path forward, please submit the responses to comments and proposed modifications to the model within seven
working days of your receipt of this correspondence. As stated above, the revised Air Dispersion Modeling Report must be submitted within 60 days of your receipt of this correspondence.

If you have any questions, please contact Jeff Vandel at [phone].

Sincerely,

ORIGINAL DOCUMENT SIGNED BY SCOTT T. ANDERSON ON 1/18/2012

Scott T. Anderson, Director
Utah Division of Solid and Hazardous Waste

STA/JV/tjm

Enclosure

c: Mike Smith, TechLaw
Cheryl Overstreet, US EPA Region VIII
Blair Palmer, ATK Launch Systems
Lloyd Berentzen, MBA, Health Officer, Bear River Health Department

PRELIMINARY REVIEW OF ATK DRAFT AIR DISPERSION MODELING REPORT
OCTOBER 6, 2011

Below is a preliminary evaluation of the ATK Launch Systems’ (ATK’s) Preliminary Air Dispersion Modeling Assessment Report for Open Burn and Open Detonation Treatment Units dated 10/6/2011
(Draft Modeling Report).

Technical Review Comment 1: Section 2.0, Air Dispersion Modeling Protocol, indicates that complex and flat terrain modeling was performed as part of the air modeling analysis. This
approach was not proposed in ATK’s air modeling protocol. Further, the need to perform both flat and complex terrain modeling was not established in the Draft Modeling Report. Please
revise the Draft Modeling Report to provide additional details on flat terrain modeling. The additional information should establish the need for performing both flat terrain and complex
terrain modeling. When modeling results are discussed, please ensure the text indicates whether the results were obtained under the flat terrain or complex terrain assumption (this
is done in Section 3.2.1.2, M-225 Treatment Unit). In addition, the sensitivity of Gaussian air dispersion models to elevation heights and the impact of the flat terrain assumption
on air modeling results should be addressed in Section 3.6, Uncertainty Analysis.

Technical Review Comment 2: To promote clarity and transparency in the air modeling report, the column entitled Model Quantity in Tables 2-1 and 2-1 should be re-titled Modeled Quantity
per Event.

Technical Review Comment 3: Figure 2-6 depicts the discrete receptors of interest and labels them with a descriptor but does not specify their UTM coordinates. Thus, the figure illustrates
the spatial relationship among the discrete receptors, sources and other points of interest but does not establish their exact location by specifying their UTM coordinates. It appears
the Draft Modeling Report does not include a description of the discrete receptors of interest that matches their location with their name. Please label each discrete receptor of interest
with its UTM coordinates.

Technical Review Comment 4: The majority of the tables in Section 3.0 list the Universal Transverse Mercator (UTM) coordinates for the modeled receptor locations but do not list a descriptor
for the on-site and off-site discrete receptors of interest. Thus, it appears the report does not include a description of the discrete receptors of interest that matches their location
with their name. It is recommended that a comments column be added to these tables so that a descriptor could be provided for discrete receptors of interest.

Technical Review Comment 5: Table 3-5 identifies the discrete receptors of interest by their name but does not list their location in UTM coordinates. Thus, it appears the Draft Modeling
Report does not include a description of the discrete receptors of interest that matches their UTM coordinates (location) with their description (name). It is recommended that columns
be added to Table 3-5 so the UTM coordinates of the discrete receptors of interest can listed.

Technical Review Comment 6: Section 3.5.1 raises concerns regarding the inclusion of non-detected constituents in the suite of constituents addressed in the risk assessment. In conjunction
with this concern, Tables 2-5 and 2-6 should be modified to identify those constituents that were not detected during testing. Consideration should also be given to identifying non-detected
constituents in the results tables presented in Section 3.0.

Technical Review Comment 7: No plots of the air modeling results were found in the report. Important air modeling results should be presented graphically as contour lines overlaying
a figure
that illustrates the source locations, facility boundaries, and both on-site and off-site discrete receptors of interest.

Technical Review Comment 8: The graphical depictions of air modeling results should differentiate between results obtained under the complex terrain assumption and those obtained assuming
flat terrain. Please present the results on the same figure if possible. If not, please provide separate figures.

Technical Review Comment 9: A spot check of the OBODM output files in Appendix D, Gas Modeling Results, indicates that elevations were specified for the modeled receptor locations in
both the complex and flat terrain modeling runs. In flat terrain modeling runs, OBODM treats elevations as flagpole receptor heights (i.e., heights above the ground surface). Further,
the flat terrain runs specify a source elevation of zero meters while the complex terrain runs for M225 specify a source elevation of 1401 meters. In addition, all flat terrain runs
examined in the spot check specified flag pole receptor heights well in excess of the elevation of the sources. It is not clear why this is done as Section 2.1, Air Dispersion Modeling
Protocol, indicates that receptor locations at or below the elevation of the treatment unit are considered as flat terrain receptors. Please review all flat terrain air modeling files
and ensure the flat terrain modeling was performed correctly. Further, please ensure the flat terrain assumption was applied to modeled receptor locations that do not vary significantly
in height with the modeled sources. In addition, please justify the use of flagpole receptor heights in the flat terrain modeling and discuss the impact of using flagpole heights on
the air modeling results.

Technical Review Comment 10: It appears that the OBODM output files (Appendices D through F) and the model-ready meteorological data files (Appendix B) have been submitted. However,
the OBODM input files were not found. Please ensure the OBODM input files corresponding to the submitted output files are also submitted to the Division of Solid and Hazardous Waste.

Technical Review Comment 11: A review of the OBODM output files indicates that the number of hours addressed in the evaluation of the annual average air quality impacts (concentration
and deposition factors) have been limited to hours in which treatment occurred. Thus, the resulting annual averages represent the annual average of the number of hours OB and OD operations
were performed rather than the total number of hours in a year. This approach results in higher annual average impacts than would be realized if all hours in which emissions were zero
were included in the calculation of the annual average.

RECOMMENDATIONS FOR MODIFYING THE AIR DISPERSION MODELING APPROACH AT ATK PROMONTORY

OBODM allows for user configuration of model runs through a series of menus which can lead to relatively quick results for addressing air permitting requirements under the Clean Air
Act (CAA). However, the model developers also recognized the unique requirements for risk assessment-related air dispersion modeling for open burning (OB) and open detonation (OD) units
under RCRA and programmed some modeling options into the code to facilitate the required long-term air quality impacts. Most of these options require additional work on the part of
the model user to achieve the desired results. While it is preferred that initial OBODM modeling analyses satisfy all RCRA permitting requirements, it is not unusual that a second round
of modeling be performed to ensure that all modeled quantities are protective of human health and the environment. The initial modeling performed by ATK provides useful information
on the dispersion of pollutants from OB and OD operations but does not take full advantage of OBODM’s options for obtaining the long-term (e.g., annual) averages required for risk assessment.

In reviewing the OBODM modeling files, it is apparent that two changes in the air modeling approach will produce significant reductions in the results for all averaging times except
the one-hour air concentrations.

As noted throughout the report, the air modeling analysis assumes treatment of waste quantities far in excess of daily and annual permit limits for each source. Reducing the amount
of waste treated per event so that the modeled operations are restricted to the daily and annual treatment amounts will reduce the modeling results for most averaging times (e.g., three-hour,
24-hour, annual).
The number of hours considered in the air modeling analysis was restricted through one of OBODM’s source menus. This eliminated all hours of zero emissions from the annual average calculations.
Lifting the restrictions in the source menu and utilizing an hourly source strength file (which tells OBODM the amount of waste treated by each emission source for every hour in the
meteorological data file) ensures all hours of the year are considered, including those with zero emissions, in calculating the annual average.

Suggestions for implementing these two changes are presented below.

A third section, Possible Approaches for Reducing Modeled 1-Hour Peak Concentrations, presents two approaches for reducing modeled one hour concentrations by reducing the amount of waste
treated during any given hour.

Reducing the Modeled Quantities of Waste to More Realistic Values
As indicated in Table 1, each of the sources considered in modeling M-136 and M-225 has its own daily and annual limit. In addition, no one is allowed to enter a burn area for 16 hours
after a burn occurs. From a practical perspective, this restriction likely applies to detonations as well. Thus, it seems realistic to limit each individual source to no more than
one treatment event per day.

Other assumptions (e.g., treatment event lasts approximately one hour) can be combined with those previously noted to develop a more realistic estimate of the actual quantities of waste
treated and still offer ATK the desired flexibility in scheduling.

Based on the above assumptions, the refined modeling approach should first assume that each source is used once per day and the per event treatment quantity is equal to the daily limit
for the source. Thus M-136, Source 1 would treat 106,500 pounds once per day, M-225, Source 2 would detonate 600 pounds once per day and so forth.

A second reduction using the total annual treatment quantity restriction for each source can be realized by limiting treatment to the Days of Treatment listed in Table 1. The impact
of both reductions can be easily demonstrated if the approach discussed under Using an Hourly Source Strength File is also implemented in the refined modeling analysis. For example,
M-136, Source 1 would treat wastes 70 times during the year while M-225, Source 2 would detonate wastes four times per year.

Table 1. Treatment Limits and Initial Proposal for New Modeled Quantities

For M-136 Hourly or per Total Event Daily Annual Days of Treatment Modeled Burn Waste Limit Limit Treatment Notes Quantity Source Stations Categories lb/day lb/yr days/yr lb/hr
or event 1 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 A, B, C, D, 4, F, G, H 106500 7500000 70 Assume treatment occurs once per day 106500 2 13 50000 496400 10 Assume treatment occurs
once per day 50000 3 14 106500 2000000 19 Assume treatment occurs once per day 106500 4 13 and 14 600 3600 6 Assume treatment occurs once per day 600 Burn/detonation area cannot
be entered for 16 hours after treatment event

For M-225 Hourly or per Total Event Daily Annual Days of Treatment Modeled Burn Waste Limit Limit Treatment Quantity Source Stations Categories lb/day lb/yr days/yr lb/hr
or event 1 1, 2, 3, 4, A, B, C, D, E, F, G, H 4500 52500 12 Assume treatment occurs once per day 4500 2 1 C, D, G, H 600 2500 4 Assume treatment occurs once per day 600 Burn/detonation
area cannot be entered for 16 hours after treatment event
Using an Hourly Source Strength File
OBODM’s source data menu screens display the data entry fields for the required source inputs. The first of these menus allows the user to specify the fuel or explosive and the pollutant
species of interest. The last option on the first source menu is a switch that can be used to direct the model to read a data file containing hourly source data for the same hours as
the hourly meteorological inputs file. If [Y] is entered in the first menu field for “source data vary with hourly met data,” the next menu will ask the user to enter the source input
filename containing the chronologically ordered hourly inputs. Source inputs read from this file will override any value(s) set in the previous source input menus. Also, source parameters
that are not contained in the hourly inputs set must have be specified in the other source data menus. All source data units for the values listed in the hourly source data file must
be consistent with those specified in the source menus. Sources that vary hourly are identified by their source number. Further, there must be a source data record for each source included
in the model run for each hour in the meteorological data file. Thus, the file for M-136 would require four rows of data per hour while the file for M-225 would require only two. The

user can press the F4 key while in OBODM for more information on the contents and format of this file.

In modeling impacts from ATK’s open burn and open detonation operations, it is desirable to allow the source strength (i.e., amount of waste treated) to vary by hour. Thus, an hourly
source strength file should be developed for each modeled unit. An example hourly source strength file for M-225 is shown on the CD that was enclosed with this document. Note that
the first data field specifies the source number while the second data field specifies the amount of waste treated by that source during the hour. As noted above, each hour requires
a row entry for each source. Thus, the file for M-225 should include 17,520 row entries.

This example file is presented for illustration only. While it reflects a variation in source strength with time as well as the annual treatment limit for the listed sources, no effort
has been made to ensure that hours of nonzero source strength are based on typical operations at M-136 and M-225 and the operational restrictions applicable to those units or correspond
to hours during the year demonstrated as generating worst-case results in the initial modeling analysis. Development of a suitable file for each unit is left to ATK Promontory to ensure
the scheduling flexibility desired by the facility is reflected in the OBODM model runs.

As discussed above, OBODM’s source menus are used to configure a modeling run to use the hourly source strength file. Figure 1 illustrates the menu used to indicate that some source
characteristics (e.g., source strength) will vary with the hourly meteorological data. As seen a Y (in bold font) has been entered at the bottom of the menu. Figure 2 illustrates the
next source menu where the path and name of the hourly source strength file should be entered. Note that Figure 2 shows the name of the example file entered in this menu.

For this option to work effectively, all restrictions on the number of hours of the meteorological data file evaluated in the model run must be removed. For ATK, this is done by removing
the N’s and Y’s from the hour entry table in the bottom right corner of the source menu shown in Figure 3. Note that the restrictions would also be eliminated if Y was entered for all
hours.

Additional information on these menus and configuring an OBODM model run to use the hourly source strength file can be found in Volume 1 of the OBODM User’s Guide, Sections 2.3, 3.4,
and Appendix A.

While these modifications to the modeling approach are expected to reduce multi-hour average results, no change to one-hour peak concentrations is expected as the maximum amount of waste
treated during an hour remains unchanged. If one-hour modeling results continue to be problematic, other approaches including changes to daily permit limits may need to be considered.

Figure 1. OBODM Source Menu 1

OBODM (Version 01.3)
SOURCE POLLUTANT/SPECIES
{F6 displays common fuels and explosives with the respective}
{quantity of pollutant/species per quantity of fuel or explosive}

FUEL OR EXPLOSIVE LABEL ----------- = [ ]
POLLUTANT/SPECIES LABEL ----------- = [ ]
POLLUTANT/SPECIES IS GASEOUS OR PARTICULATE (G/P) ---------------------- [G]
Pollutant/species molecular weight (grams/gram-mole) ------- =[ ]
(Required for print output units of PPM or PPB)
Pollutant/Species density {units (g/cm**3, etc.)=[ ]}- =[ ]
(Required for PARTICLES)
Poll/species half-life (0=infinite) {units (s,mn,h)=[ ]} -- =[ ]

{SPECIFY EITHER}
Ratio of mass of pollutant/species {units (g,kg,lb,etc.)=[ ]}=[ ]
per mass of fuel or explosive burned {units= (g,kg,etc.)=[ ]}=[ ]
{OR}
Fraction of exhaust cloud constituting pollutant/species --- =[ ]
Some source data vary with hourly input met. data (N/Y) -----------------[Y]

Figure 2. OBODM Source Menu 2

OBODM (Version 01.3)
SOURCE POLLUTANT/SPECIES

Note: Source data read from an hourly chronologically-ordered data set override any value(s) set in the above source input data menus. Also, parameters that are not read from the hourly
data set must be specified in the above source data menus. All source data units must be specified. The source input data file cannot be the same as the hourly meteorological input data
file. Sources that vary hourly are identified by their source numbers. There must be a source data record(s) for each of these sources for each hour in the meteorological data file.
Use F4 for more information on the contents and format of this file.

Source input data file name ----------------------- [M225_example.txt]

Figure 3. OBODM Menu 3

OBODM (Version 01.3)
GASEOUS OR PARTICULATE SOURCE DATA
SOURCE NUMBER ( )
{F5 displays units abbreviations}

SOURCE IDENTIFICATION NAME ---------- [ ]
Source type (Volume (point), Line) --------------------------------------- [V]
Source emission type (Instantaneous,Quasi-continuous) ------------------ [I]

Total mass of material burned or released
{units (g,kg,lb,etc.)=[ g]} ------------------------ = [ 1.0]
Fuel material heat content {units(CA/g,J/lb,etc.)=[ CA/g]}= [ ]
Fuel material burn rate (quasi-continuous sources only)
{units (g/s,lb/s,etc.)=[ g/s]} -------------------- = [ ]

Hours in a day during which this source can be burned or detonated, starting
with hour 00 and ending with hour 23. Enter Yes or No (blank is default
Yes)
hour - 00 01 02 03 04 05 06 07 08 09
[ ][ ][ ][ ][ ][ ][ ][ ][ ][ ]
10 11 12 13 14 15 16 17 18 19
[ ][ ][ ][ ][ ][ ][ ][ ][ ][ ]
20 21 22 23 2
[ ][ ][ ][ ] ]

Possible Approaches for Reducing Modeled 1-Hour Peak Concentrations

It is given that one-hour concentrations can be reduced by reducing the amount of waste treated during any single hour. A number of approaches can be taken to reduce the hourly treatment
amount and two possible approaches are presented below. Note that changes to the daily limit for a source also change the hourly limit as the 16 hour restriction on entering a source
after treatment effectively restrict an individual source to one, 1-hour treatment event per day.

If 106,500 pounds is the daily limit for OB operations at unit M-136, the pounds of waste treated per hourly event should be reduced so the daily amount treated for all OB sources does
not exceed 106,500. ATK’s initial modeling assumed a daily limit of 106,500 pounds for Source 1, a daily limit of 106,500 pounds for Source 3, and a daily limit of 50,000 pounds for
Source 2. If the 106,500 pound daily limit actually represents the daily limit for the sum of Source 1 and Source 3, half can be treated at Source 1 and half can be treated at Source
3.

New daily limits (and the related per event limits) can be specified/requested based on recent process history and the desired level of operational flexibility. The limits assumed in
the initial air modeling analysis are likely indicative of past treatment requirements but may be well in excess of what is currently needed. Lower treatment limits can be requested
and demonstrated as protective of human health and the environment in the refined air modeling analysis reducing all modeling results including one-hour concentrations. If a need for
increased treatment amounts arises during the life of the permit, the increase can be addressed through a permit modification.

Neither of these suggested approaches may be attractive if the daily limits addressed in the initial modeling are needed for operational flexibility (i.e., must be able to treat the
full daily limit at one source). However, it is preferred that ATK, in cooperation with the Division of Solid and Hazardous Waste, propose desired permit limits that can be demonstrated
as protective of human health and the environment. If such a demonstration can be provided, the Division will not have to propose permit limits which may not reflect the operational
flexibility needed at the M-136 and M-225 units.

TN201101296

195 North 1950 West • Salt Lake City, UT
Mailing Address: P.O. Box 144880 • Salt Lake City, UT 84114-4880
Telephone [phone] • Fax [phone] • T.D.D. [phone]
www.deq.utah.gov
TN201101296 Printed on 100% recycled paper

GARY R. HERBERT
Governor
GREG BELL
Lieutenant Governor

Solid and Hazardous Waste Control Board
Kevin Murray, Chair
Kory Coleman, Vice-Chair
Brian E. Brower
Scott Bruce

Jeff Coombs, MPH, LEHS
R. Ryan Dupont, Ph.D.
Larry A. Ellertson
Brad Mertz
Brett Mickelson
Gary Mossor
Dennis Riding
Dwayne Woolley
Amanda Smith
Scott T. Anderson
Executive Secretary

Department of Environmental Quality

Amanda Smith
Executive Director

DIVISION OF SOLID AND
HAZARDOUS WASTE
Scott T. Anderson
Director

State of Utah