Risk Assessment and Management at Deseret Chemical Depot and the Tooele Chemical Agent Disposal Facility (1997)

Public Law 99-145, enacted by Congress in 1985, authorized the Army to initiate the process of eliminating the aging U.S. chemical weapons stockpile and led to the establishment, in 1987, of the National Research Council (NRC) Committee on Review and Evaluation of the Army Chemical Stockpile Disposal Program (Stockpile Committee) to provide the Army with technical advice and program oversight. Subsequently, Public Law 104-484 directed the Army to dispose of the entire unitary chemical agent and munitions stock-pile by December 31, 2004.

The Army selected incineration as the baseline destruction technology for the Chemical Stockpile Disposal Program (CSDP), and in 1990, a prototype destruction facility was completed at Johnston Island in the Pacific Ocean. Incineration had been endorsed as the preferred technology in a 1984 report by a predecessor of the Stockpile Committee, the Committee on Demilitarizing Chemical Munitions and Agents. At that time, incineration was selected from several alternatives as the most mature and proven technology for the destruction of agents and munitions. Concurrent with the construction of the Johnston Atoll Chemical Agent Disposal System (JACADS), the Army developed and issued its Final Programmatic Environmental Impact Statement (FPEIS), which included a programmatic risk assessment.

A comprehensive understanding of the full spectrum of risks is fundamental to sound risk management. Risks to the public, workers, and the environment have been and continue to be a central theme of NRC oversight. The Stockpile Committee reviewed the FPEIS and, in a January 1993 letter report, noted that the programmatic risk assessment ''was not directed at managing risk at any specific site.'' In that same report, the committee recommended that the CSDP should include site-specific risk assessments for each of the eight continental U.S. disposal sites as a basis for overall risk management of the disposal program. The report laid out a specification for site-specific risk assessments. Subsequently, in a 1994 report, the Stockpile Committee expanded on the nature of and need for site-specific assessments.

In response to the 1993 letter report, the Army Program Manager for Chemical Demilitarization (PMCD) directed that a quantitative risk assessment (QRA) and a risk management program be developed for each of the continental U.S. storage and disposal sites, beginning with the Deseret Chemical Depot (DCD) (formerly Tooele Army Depot, South) in Tooele, Utah, and its associated Tooele Chemical Agent Disposal Facility (TOCDF).¹ The TOCDF, the first full-scale chemical agent and munitions disposal facility in the continental United States, is a second-generation incineration system that incorporates design modifications to improve performance over the initial full-scale prototype incineration system used at JACADS. The stockpile at DCD contains neurotoxic (nerve) agents (GB, VX, GA) and mustard (blister) agents (H, HD, HT) in bulk (ton) containers, rockets, projectiles, mines, bombs, cartridges,

Figure ES-1

Schematic illustration of TOCDF risk elements.

and spray tanks. The DCD contains 45 percent of the remaining U.S. stockpile, some 13,000 tons of agent in more than one million inventory items. Agent GB, the most hazardous agent stored at DCD, constitutes about half of the agent and is contained in 75 percent of the inventory items.

A health risk assessment (HRA) for TOCDF was conducted by the state of Utah as part of the environmental permitting process.² Figure ES-1 is a schematic illustration of the two major categories of risks, i.e., risks arising from accidents (analyzed in the QRA) and risks arising from emissions during normal and mild upset conditions (analyzed in the HRA). These first site-specific risk assessments have been completed and are primary information for this report.

The Army's site-specific QRA for DCD/TOCDF was conducted, under contract, by Science Applications International Corporation, Inc. (SAIC). An independent group of experts in risk assessment and engineering, the Risk Assessment Expert Panel on the Tooele Chemical Agent Disposal Facility Quantitative Risk Assessment (Expert Panel), oversaw the QRA and provided the Army with a separate, detailed technical review. Thus, the committee's role included both general oversight of the risk assessment/risk management process and oversight of the Expert Panel review process. One or two members of the committee attended most Expert Panel review meetings. The committee also took advantage of many opportunities to examine the technical details of the risk assessment work. A 1996 Stockpile Committee report reviewed the

methodology of the QRA and the Expert Panel' s oversight. The committee found that:

• The QRA methods met the recommendations of the committee's earlier reports.
• The SAIC QRA team was responsive to the committee's questions and the Expert Panel's comments. New analytical tools for first-of-a-kind QRA calculations were developed; outside experts were retained to give advice in areas where the literature was incomplete; tests and new mechanistic analyses to answer technical questions were conducted; and the QRA analyses were being revised based on that new information.

The current report continues the Stockpile Committee's oversight of risk considerations. The report encompasses the program-wide and site-specific definition of the documented CSDP risk management process and evaluates the results of risk assessments of the Tooele storage and disposal facilities and the overall risk management process being implemented for the TOCDF.

Earthquakes as initiating events pose the greatest risk to the public from continued storage of the DCD stockpile. Earthquakes can have widespread, often severe effects, leading to significant adverse consequences. Seismic events that contribute to public risk have mean accelerations higher than 0.2 g (1 g equals the acceleration of gravity) and recurrence intervals of 1,000 years or more. Such earthquakes significantly exceed normal building code design values and thus can lead to failures of equipment, structures, and stored munitions. Earthquakes account for 82 percent of the average risk of fatalities to the public. Significant contributors to the public risk from storage are illustrated in Figure ES-2.

Public risk from disposal processing during the 7.1-year disposal period is substantially lower than the risk from continued storage during the same period. Again, earthquakes are the most significant initiating events, accounting for 97.4 percent of the public risk during processing. Other contributors to risk are summarized in Figure ES-3. Figure ES-4 depicts the magnitude of risk associated with continued storage and agent/munitions processing and shows the decrease in storage risk as the stockpile is processed. This figure provides the most complete picture of the risks during processing and the most thorough comparison of the risks of continued storage with the risks associated with processing (including the diminishing contribution of storage risk during the processing period).

Figures-ES-2

Contributors to the average public fatality risk from continued storage at DCD.

For the 500 workers at the TOCDF, there is about a one in seven probability that there will be one fatality

Figure ES-3

Contributors to the average public fatality risk from processing at DCD and TOCDF.

Figure ES-4

Comparison of public risks during processing at DCD and TOCDF.

in the 7.1 years of disposal processing (expected number of fatalities is 0.13). This is equivalent to the probability of fatality to an individual worker of one in 25,000 (or 4 × 10^-5) per year, which is consistent with the average occupational risk for all occupations in the United States. However, because the risk levels to disposal workers from agent exposure must be added to normal occupational risk levels, the committee believes that emphasis on job safety related to both agent and nonagent activities is extremely important. The sources of risk for workers are shown in Figure ES-5.

The HRA evaluated the effects of incinerator emissions under various operating scenarios for an adult resident, a child resident, a subsistence fisherman, and three farmers at various locations with various durations of exposure during the year. Conservative modeling assumptions maximizing anticipated concentrations of emissions and maximizing exposure values at points of maximum off-site impact were used to derive the upper-end estimates of risk. The calculated human health risk was measured against the 1 × 10^-5 carcinogenic risk level established under EPA exposure assessment guidance protocols. This threshold was not exceeded in any scenario.

Figure ES-5

Contributors to the average risk of fatality to disposal-related workers at DCD and TOCDF.

Risk management typically involves the following steps: understanding the risk; suggesting ways to reduce risk; evaluating the alternatives; and selecting preferred alternatives. Each affected party plays a role in the risk management process. The Army is responsible for managing the chemical stockpile and its destruction in a manner that maintains safety for the public, workers, and the environment. The current CSDP risk management program is a multilevel program that delineates policy, defines requirements, provides guidance for implementation, defines specific requirements for the facility, and defines the processes that should be used.

The Army recently published (in draft) A Guide to Risk Management Policy and Activities (the Guide), which includes a description of the "management of change process," a significant aspect of the risk management process at the TOCDF and other sites. The management of change process may involve changes or modifications to equipment, operating and maintenance procedures, or agent destruction schedules.

The first full scale application of the management of change process is expected to be an evaluation of adding a carbon filter system to the pollution abatement system (PAS). Consideration of the PAS filter system (PFS) was prompted by a committee finding in a 1994 report that adding a carbon filter system downstream of the existing PAS would add further protection against emissions of agent and trace organics, even in the unlikely event of a substantial system upset.

The Stockpile Committee has followed the DCD/TOCDF QRA project closely since its inception and has maintained oversight of the Expert Panel independent peer review process. The QRA has achieved the goals set out in the committee's 1993 letter report and subsequent reports. The findings in the DCD/TOCDF QRA are consistent with interim findings in the committee' s review of systemization of the TOCDF. The committee concurs with the following findings of the Expert Panel:

• The methodology was sound and has extended the state of the art in several areas.
• The methodology was well implemented.
• Despite some reservations concerning a few technical aspects of this QRA, the panel was reasonably satisfied that these did not affect the overall conclusions of the QRA.

The committee finds that the interactive independent review process was effective and that the Expert Panel played a significant role in ensuring that the QRA met or exceeded state-of-the-art standards in all significant respects.

The committee believes that the HRA performed by the Utah Division of Solid and Hazardous Waste, which is based on many assumptions and follows EPA-mandated protocols, is appropriate at this stage of TOCDF operations because it approximates a worst case for the public for all evaluated parameters. The greatest uncertainty in the HRA is about the magnitude and composition of actual TOCDF emissions (emissions in the HRA were based on adjusted data from JACADS). As actual TOCDF operating parameters are established and data on the nature and magnitude of actual emissions become available, they can be incorporated into the HRA. The HRA does not provide the more realistic and detailed estimation of risk sources, impacts, and distribution provided by the QRA. However, it does screen latent cancer risk to "maximally exposed" individuals, impose an acceptability criterion (1 × 10^-5 carcinogenic risk level over a 70-year lifetime), and infer that the exposure of multiple individuals at or below the screening level is acceptable.

The committee finds that the TOCDF risk management plan has progressed and that positive steps have been taken, e.g., the development and limited use of guidance and implementation documents. The Army's draft Guide on risk management provides an overview of the overall risk management program, incorporating references to subsidiary risk management documents and activities. The Guide defines interrelationships among Army offices, contractor offices, and public entities that are or should be involved in risk management activities. The Guide is a significant step by the

Army toward following NRC recommendations on risk management and public involvement, particularly with respect to using risk analysis in the management of change process. The Guide, however, has not yet been completed and finalized. The committee's comments in this report reflect the expectation that the Guide will be improved, adopted, and implemented. The committee may evaluate its implementation in the future, but at present, it falls short of expectations in several respects:

• The Guide does not describe the contributions of risk assessment to changes that led to the current Established Configuration, against which proposed future changes will be evaluated.
• The focus of the Guide is primarily on agent-related safety rather than on developing and institutionalizing a comprehensive safety program (i.e., establishing a safety culture).
• The Guide acknowledges that more work must be done to shift the focus from public information to public involvement. The role of public involvement should be extended and integrated beyond the management of change process.
• The Guide is not specific enough about how to ensure that workers, Chemical Stockpile Emergency Preparedness Program (CSEPP) personnel, and the public understand the risk analyses.
• Although the Guide indicates that there are functional relationships among PMCD organizations with regard to risk management, it does not detail how management roles and communications across all the Army groups, contractors, subcontractors, and other agencies involved in the program will be integrated.
• The Guide does not indicate plans to track CSEPP responses to changes in risk or to document public involvement and Army responses to public input regarding risk.

On the positive side, the Guide presents a framework for managing changes in the configuration of a facility or changes in operations that may significantly affect risk levels. The framework allows for public input on significant changes through a comment process, which is followed by formal feedback to the public explaining the basis for a decision. The committee finds the proposed management of change process satisfactory and encourages its use. However, there may be opportunities to further expand public participation as the Army develops a more comprehensive public involvement program.

Concerning the evaluation of adding carbon filters to the PAS (an earlier recommendation of the Stock-pile Committee), the committee finds that:

• The proposed methodology, if well implemented, is appropriate for evaluating whether or not to install a PFS on a site-specific basis.
• The proposed methodology for the PFS evaluation is consistent with the Army's proposed management of change process, as described in the Guide.
• Carbon filters appear to be effective in reducing the levels of dioxins/furans to below the limits of detection and have a useful life of at least one year. Because these levels are too low to be measured, however, credit only for reducing them to the detection limit appears in the HRA.
• The QRA calculations for the PFS must account for a potential sudden release of accumulated agent (based on HRA-assumed emissions at the lower detection limit) in case of a PFS malfunction.

Recommendation 1

The Army should update both the QRA and HRA at the TOCDF whenever changes to system design or operations occur that could affect QRA or HRA calculations to ensure that estimates of risk are current and reflect changes in operating conditions and experience, assumptions, and program status (current Established Configuration). The process for updating the QRA and HRA should be included in the Guide.

Recommendation 2

The Army should continue the site-specific QRA and HRA processes at all PMCD sites. The development of assessments for sites other than the DCD will be greatly simplified because much

of the methodology has already been established. The Army should continue to obtain interactive, independent expert reviews of all site-specific risk assessments. The Army should heed the lessons learned from development of the TOCDF QRA and should incorporate the changes recommended by the Expert Panel.

Recommendation 3

The QRA methodology manual should be updated to reflect the significant improvements that have been made.

Recommendation 4

The Army should expand its draft report on risk management policy, A Guide to Risk Management Policy and Activities, to encourage the establishment of a "safety culture" within the PMCD and its field offices and among contractors and other government agencies. The Guide should elucidate the Army's policy on industrial safety, including the responsibilities of individuals and managers in the field and the definitions of acceptable performance.

Recommendation 5

The Army should develop a management program (and include it in the Guide) that defines the integration of management roles, responsibilities, and communications across activities by risk management functions (e.g., operations, safety, environmental protection, emergency preparedness, and public outreach).

Recommendation 6

The Army should review and expand the current draft risk management plan to include public involvement in appropriate areas beyond the management of change process.

Recommendation 7

The Army should institutionalize the management of change process developed in the Guide. The Army should track performance of the change and document public involvement and public responses to decisions. The Army should use this experience to improve the change process.

Recommendation 8

The Army should expand implementation of the risk management program to ensure that workers understand the results of the risk assessments and risk management decisions. The Army should also ensure that CSEPP and other emergency preparedness officials understand the QRA and how their activities might affect risk. CSEPP activities should be tracked by the Army as part of the risk management program.

Recommendation 9

The Army should implement the risk management plans and update them whenever necessary to ensure that they reflect current practices and lessons learned.

Recommendation 10

The Army should proceed with the application of its proposed methodology for evaluating the use of PAS carbon filters on a site-specific basis. For consistency with the HRA assumptions, the QRA should take into account the possible sudden release of agent that may have accumulated on the filter at a gas concentration equal to the lower detection limit.