6
Findings and Recommendations

Based on its review and analysis in Chapters 1 through 5 of incidents at two operating chemical demilitarization sites, JACADS and TOCDF, the Committee on the Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities developed the following findings and recommendations.

Finding 1. Despite considerable Army security and stewardship activities, the remaining chemical weapons stockpiles are significant hazards to the communities surrounding them. The potential for significant release of agent to the atmosphere, triggered by either accidental or deliberate detonation of agent-loaded munitions within storage igloos, constitutes the greatest risk to the public. Accidental or deliberate release from a chemical demilitarization facility, while potentially serious, is a lesser threat because the agent inside the facility is maintained under stringent and effective engineering controls and because there is substantially less agent present in the demilitarization facility at any given time than there is in the storage facility. While chemical demilitarization operations at both Johnston Atoll Chemical Agent Disposal System and Tooele Chemical Disposal Facility have released small amounts of chemical agent into the environment, these releases were negligible compared with environmental releases from chemical weapons stockpiles (U.S. Army, 2001e). The rate of agent leaks and releases from storage facilities does decrease significantly as the stockpile is processed.

Recommendation 1. The destruction of aging chemical munitions should proceed as quickly as possible, consistent with operational activities designed to protect the health and safety of the workforce, the public, and the environment.

Finding 2. The criteria used by the Army to identify and determine the severity of the impact of an event are site and time specific, and the event classification decision is made at the discretion of the Depot Commander. The recognition of a chemical event is often subjective, and the tendency of personnel to discount initial indicators because of frequent false positive automatic continuous air monitoring system alarms is a persistent problem in declaring a chemical event. The lack of uniform criteria can result in inconsistencies between and among sites that make it difficult to compare and analyze events and that constrain and discourage the application of lessons learned to other locations and situations.

Recommendation 2. The Army should establish a consistent set of criteria to be used by all chemical-agent-processing facilities to ensure uniformity in the classification of events, and to facilitate event analysis and comparison.

Finding 3. Risk assessments, including the quantitative risk assessment and the health risk assessment, are critical inputs to the dialogue necessary to ensure adequate public involvement in, and understanding of, chemical demilitarization activities. A prudent balance between the public’s right to know the risks they face and the need to protect sensitive information will be an ongoing challenge for the chemical demilitarization program. Without adequate risk information available to the public, it will be difficult to develop or maintain the level of public trust necessary for the Program Manager for Chemical Demilitarization to accomplish its mission.

Recommendation 3. The Army should continue its practice of making available to the public the results of its quantitative risk assessments and health risk assessments for each chemical demilitarization site.

Finding 4. Of the wide range of Program Manager for Chemical Demilitarization risk analyses, the quantitative risk assessments (QRAs) are most closely linked with chemical events. They calculate the frequency and consequences of modeled events, and their analysts study real operational



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 51
6 Findings and Recommendations Based on its review and analysis in Chapters 1 through the discretion of the Depot Commander. The recognition of a 5 of incidents at two operating chemical demilitarization chemical event is often subjective, and the tendency of per- sites, JACADS and TOCDF, the Committee on the Evalua- sonnel to discount initial indicators because of frequent false tion of Chemical Events at Army Chemical Agent Disposal positive automatic continuous air monitoring system alarms is Facilities developed the following findings and recommen- a persistent problem in declaring a chemical event. The lack dations. of uniform criteria can result in inconsistencies between and among sites that make it difficult to compare and analyze Finding 1. Despite considerable Army security and stew- events and that constrain and discourage the application of ardship activities, the remaining chemical weapons stock- lessons learned to other locations and situations. piles are significant hazards to the communities surrounding them. The potential for significant release of agent to the Recommendation 2. The Army should establish a consis- atmosphere, triggered by either accidental or deliberate deto- tent set of criteria to be used by all chemical-agent-process- nation of agent-loaded munitions within storage igloos, con- ing facilities to ensure uniformity in the classification of stitutes the greatest risk to the public. Accidental or deliber- events, and to facilitate event analysis and comparison. ate release from a chemical demilitarization facility, while potentially serious, is a lesser threat because the agent inside Finding 3. Risk assessments, including the quantitative risk the facility is maintained under stringent and effective engi- assessment and the health risk assessment, are critical inputs neering controls and because there is substantially less agent to the dialogue necessary to ensure adequate public involve- present in the demilitarization facility at any given time than ment in, and understanding of, chemical demilitarization ac- there is in the storage facility. While chemical demilitariza- tivities. A prudent balance between the public’s right to know tion operations at both Johnston Atoll Chemical Agent Dis- the risks they face and the need to protect sensitive informa- posal System and Tooele Chemical Disposal Facility have tion will be an ongoing challenge for the chemical demilitari- released small amounts of chemical agent into the environ- zation program. Without adequate risk information available ment, these releases were negligible compared with environ- to the public, it will be difficult to develop or maintain the mental releases from chemical weapons stockpiles (U.S. level of public trust necessary for the Program Manager for Army, 2001e). The rate of agent leaks and releases from Chemical Demilitarization to accomplish its mission. storage facilities does decrease significantly as the stockpile is processed. Recommendation 3. The Army should continue its practice of making available to the public the results of its quantita- Recommendation 1. The destruction of aging chemical tive risk assessments and health risk assessments for each munitions should proceed as quickly as possible, consistent chemical demilitarization site. with operational activities designed to protect the health and safety of the workforce, the public, and the environment. Finding 4. Of the wide range of Program Manager for Chemical Demilitarization risk analyses, the quantitative risk Finding 2. The criteria used by the Army to identify and assessments (QRAs) are most closely linked with chemical determine the severity of the impact of an event are site and events. They calculate the frequency and consequences of time specific, and the event classification decision is made at modeled events, and their analysts study real operational 51

OCR for page 51
52 EVALUATION OF CHEMICAL EVENTS AT ARMY CHEMICAL AGENT DISPOSAL FACILITIES events to help ensure the completeness of the models. The instituted without significant increases in false positive alarm QRAs, and an understanding of their results, provide a rates or unless health risk assessments demonstrate that lower framework for managing the risk from chemical events. As thresholds are necessary to protect workers or the public. concern has focused on worker risk as well as risk to the public, recent QRAs have added extensive human reliability Finding 6. Relatively frequent false positive ACAMS alarms analysis—analysis of the human actions that can lead to a for airborne agent and the lack of true real-time (<10 s) moni- chemical event. While hazard analysis is a qualitative analy- toring for airborne agent have long been a concern of Na- sis primarily of single-failure events, it provides a catalog of tional Research Council (NRC) committees assessing and possible failures that QRA analysts can use to improve the examining the chemical demilitarization program (NRC, completeness of their models. Hypothesized chemical 1994, 1999a, 2001a). Improvements in the sensitivity, speci- events, including those that might arise from sabotage, ter- ficity, and time response of the ACAMS system and the de- rorism, and war, can also be incorporated into the QRA sce- velopment of an additional airborne-agent-monitoring tech- narios to determine their range of consequences. Actual nology robust at the parts-per-trillion level have previously events can be used to test and improve the completeness of been recommended. (Recent NRC reports have also noted the QRA and continue the effort to improve the human re- the poor state of agent-monitoring technology for liquid liability analysis and focus on causal factors. waste streams and solid materials suspected of possible agent contamination (NRC, 2000a, 2001a).) Although the Pro- Recommendation 4. The quantitative risk assessment gram Manager for Chemical Demilitarization has made some (QRA) for each chemical demilitarization site should be it- efforts to develop better agent-monitoring technology, re- erative. Actual chemical events should be used routinely to sults to date have been disappointing. Development and de- test the completeness of the QRA, which should be routinely ployment of airborne-agent monitors with shorter response utilized to hypothesize the frequency and consequences of times and lower false alarm rates would enhance safety and chemical events. The Program Manager for Chemical De- reduce the tendency to discount agent alarms. militarization and the U.S. Army Soldier and Biological Chemical Command should use the QRAs to evaluate mea- Recommendation 6. To reduce the rate of false positive sures to control future chemical events. The Army should alarms for both airborne and condensed-materials agent con- also consider using QRAs to examine scenarios associated tamination, the Program Manager for Chemical Demilitari- with sabotage, terrorism, and war. zation and the relevant Department of Defense research and development agencies, such as the Army Research Office, Finding 5. Alarm thresholds for airborne agent monitoring the Army Research Laboratory, the Defense Advanced Re- used in the Army’s chemical demilitarization program are search Projects Agency, and the Defense Threat Reduction very conservative (i.e., 20 percent of the applicable control Agency, should invigorate and coordinate efforts to develop limit, resulting in alarm thresholds that, depending on the chemical agent monitors with improved sensitivity, speci- agent, are either below or only moderately above the level of ficity, and time response. These efforts should be coordi- agent deemed safe for continuous exposure of an unmasked nated with, and take advantage of, the increased level of in- worker over an 8-hour shift.) These alarm thresholds are near terest in and increased resources available for developing the detection limits for the automatic continuous air moni- chemical weapons detectors for homeland defense. toring system (ACAMS). As a result, there are frequent false positive alarms as well as actual alarms for events that pose Finding 7. Chemical demilitarization facility and process no threat to workers or the public (NRC, 2001a). These design has contributed to the mitigation of incident severity in conservative stack-monitoring thresholds ensure that no sig- that, for most of the incidents examined by the committee, nificant amounts of agent can be released into the ambient engineering controls functioned as designed. These incidents air without the facility alarming and the agent incineration have been investigated honestly and thoroughly using straight- feed automatically terminating. In-plant air breathed by un- forward techniques, but investigation could benefit from the masked workers and the output of the scrubbing system for use of other methodologies such as event tree analysis and air exiting the chemical demilitarization plant are monitored human factors engineering to aid in understanding the com- at similarly conservative thresholds. plete set of causal factors associated with each incident. Recommendation 5. The Army should maintain conserva- Recommendation 7. Incident investigation teams should tive chemical demilitarization exhaust stack and in-plant air- use modern methodologies of incident investigation rou- borne agent exposure thresholds. If current limits for expo- tinely at all chemical demilitarization sites to help uncover a sure to stockpiled chemical agents are further reduced, the broader set of causal and contributing factors, and to enable Army should not further reduce existing monitoring thresh- greater understanding of the interrelationships between and olds unless chemical agent monitors can be made both more among these factors. Experts in human performance model- sensitive and more specific so that lower thresholds can be ing should be included on any incident investigation team. A

OCR for page 51
53 FINDINGS AND RECOMMENDATIONS standing incident review board at each site should be estab- dict the extent of an agent emission plume are representative lished to identify chemical events requiring in-depth inves- of the state of the art as of the late 1970s. Adoption of more tigation and to ensure that the lessons learned appropriately modern and accurate emission plume models seems to have influence ongoing operations. These boards would meet been delayed by the failure to integrate better plume models regularly to review accidents and incidents, including into standard Chemical Stockpile Emergency Preparedness chemical events, and would be fully informed of any find- Program emergency response models. ings and recommendations made by chemical event investi- gation teams. Recommendation 9a. Stockpile sites that still use the D2PC computer model should, at a minimum, upgrade their emer- Finding 8a. Repeating patterns of causal factors evident in gency response models to take advantage of the improved the incidents at Johnston Atoll Chemical Agent Disposal capabilities available in the D2-Puff model. Consideration System and Tooele Chemical Disposal Facility reviewed by should be given to testing and possibly optimizing the D2- the committee included, in particular, deficiencies in stan- Puff model at each site by performing tracer release experi- dard operating procedures, design failures, and understand- ments under a variety of meteorological conditions. able, although inappropriate, assumptions made by opera- tions personnel. In part, these inappropriate assumptions Recommendation 9b. The Chemical Stockpile Emergency grew from development of dangerous mind-sets associated Preparedness Program should undertake a continuing evalu- with frequent false-positive alarms. Repeating patterns of ation of alternative approaches to modeling the release and causal factors in most incidents do not appear to have been impact of chemical agents. used by management to generalize incident findings beyond the immediate context of each incident. Recommendation 9c. Accurate agent plume dispersion modeling capability should be coupled with timely commu- Finding 8b. The programmatic lessons learned (PLL) data- nication of results and appropriate responses to the stockpile base is a large undertaking, and the Program Manager for site and surrounding communities. Chemical Demilitarization is to be commended for creating it. However, if the data were organized in a different manner Finding 10a. Communications during and after chemical that included a priority system and the operators were aware events have not always occurred as intended between and of its uses, the database would be more useful. “Mining” of among the various stakeholders. The lack of a call-forward- data might yield patterns in events and information that ing mechanism for getting information directly to people or a would allow broader generalization and understanding of hot line dedicated to notification that an event has occurred causes derived from specific information on individual inci- has contributed to an inadequate communication process dur- dents. To this end, experts in each area of use have to specify ing chemical events. The lack of notification and warning exactly what they need to find in the data, before program- between the Deseret Chemical Depot (DCD), Tooele County, mers develop software to search and set priorities. and other Utah responsible agencies reflects in part a lack of coordination between components of the two programs (Fed- Recommendation 8a. The Program Manager for Chemical eral Emergency Management Agency / Chemical Stockpile Demilitarization should analyze all chemical-agent-related Emergency Preparedness Program and the Army’s Emer- incidents at chemical demilitarization plants for patterns of gency Operations Centers) and in part the DCD’s perspective causal factors and should institute program-wide actions to that its emergency management responsibilities “end at the address the causes found. fence.” This perspective, if carried to other communities where chemical demilitarization facilities are to be operated, Recommendation 8b. Any improvements made in investi- can endanger an effective coordinated emergency response gation procedures should become part of a systematically to incidents. The memorandum of understanding recently organized programmatic lessons learned (PLL) database that agreed to by the DCD and Tooele County (Appendix G) for makes information easier for the non-expert to find and/or information exchange could serve as a model for every com- use. This can include prioritization and developing a drop- munity with a chemical weapons stockpile, to ensure very down “tree” list. Lastly, the Program Manager for Chemi- close oversight of the disposal plant operations. cal Demilitarization should ensure that, at the plant level, the data are available to, known by, and useful to operations Finding 10b. The Army’s recent and sincere effort to build personnel. The proposed contractor for the PLL program public trust in its actions has not been sufficiently successful, should address these issues. For the program to be useful all although the degree of trust or mistrust has not been effec- stakeholders need to buy into its use and structure. tively measured. Of equal or greater importance is public trust in the governmental agencies that monitor the Army’s Finding 9. Gaussian puff/plume dispersion modeling tech- activities. It is essential that these agencies be seen by the niques embedded in the D2PC computer model used to pre- public as being fully cognizant of the Army’s actions and of

OCR for page 51
54 EVALUATION OF CHEMICAL EVENTS AT ARMY CHEMICAL AGENT DISPOSAL FACILITIES being, in effect, in command—a result that will require an tives from all stakeholder groups and agencies, including extraordinary level of communication between the Army members of the public who can effectively represent and and relevant government oversight agencies and can lead communicate with local officials and the affected public. to contradictory advice and requirements that will have to Incident findings should be documented in a single compre- be resolved. hensive report that incorporates the findings, proposed cor- rective actions, and concerns of the various oversight agen- Recommendation 10a. Chemical demilitarization facilities cies. should develop site-specific chemical event reporting proce- dures and an accompanying training program that tests and Finding 12. Safety programs and performance appear to be improves the implemented procedures and communication adequate to ensure that chemical demilitarization operations system. are being conducted safely. Even so, there is considerable opportunity for improvement. Many of the incidents at Recommendation 10b. The standing incident review board Johnston Atoll Chemical Agent Disposal System (JACADS) recommended by the committee for each site should include and Tooele Chemical Disposal Facility (TOCDF) could have a qualified member of the public who can effectively repre- been significantly mitigated—if not prevented—had a true sent and communicate public interests. “safety culture” been in place and functioning. Recommendation 10c. Each chemical demilitarization site Recommendation 12a. Much of the needed improvement should consider the establishment of a reporting and com- in safety at chemical weapons facilities can come from in- munication memorandum of understanding (MOU), of the creased attention to factors that contribute to and/or cause sort developed between the Deseret Chemical Depot and chemical events. For example, the Program Manager for Tooele County, which specifies reliable and trusted means Chemical Demilitarization and chemical demilitarization fa- of alerting and informing local officials about chemical cility managers should ensure that standard operating proce- events. These MOUs should be designed to permit ready dures are in place, up to date, and effective, performing haz- evaluation and updating of the terms of the MOU to take full ard operations analyses on new process steps and design advantage of learning across the array of chemical demilita- changes even when such changes are viewed as trivial and rization sites. recognizing that chemical hazards are posed by things other than agent (e.g., waste). Recommendation 10d. The Army Emergency Operations Centers and the Chemical Stockpile Emergency Prepared- Recommendation 12b. Management at the Tooele Chemi- ness Program should establish a stronger capability and ca- cal Agent Disposal Facility (TOCDF) and the new third-gen- pacity for the coordination of training, equipment, and plans eration facilities should develop or identify and implement necessary to respond effectively to an emergency incident, programs that will result in the establishment of a pervasive, and the commitment to do so in a coordinated and coopera- functioning safety culture as well as improved safety perfor- tive fashion. Additionally, the Army should continue its pro- mance. In doing so, TOCDF and the new chemical demilita- gram of outreach—including listening to public concerns and rization sites should draw on experience in the chemical in- responding to them, as well as engaging in more conven- dustry, obtained through industry associations or other tional public information efforts—to both the public and the appropriate venues. The Army should revise the award fee relevant government oversight agencies to enhance general criteria to encourage each new chemical demilitarization site understanding of the chemical demilitarization program. operator to demonstrate better safety performance than that at the older sites. Finding 11. A major chemical event can result in several months of lost processing time at chemical demilitarization Finding 13. It is probable that conditions will arise in plant plants. This delays the destruction of the chemical agents, operation for which no standard operating procedure has requiring that they remain in the stockpiles where they could been written. Operators need an in-depth knowledge of their create a hazard. When incidents have led to plant shutdown, equipment and its limitations to handle these unusual condi- multiple investigations and responses have contributed to tions and maintain plant security. New plant start-up can be additional delays in restarting operations. All aspects of a difficult learning experience for new operating crews. They chemical incident investigations and resumption of opera- need to know how and why procedures are to be performed. tions should be accelerated, consistent with safe operations. It is common practice in other industries to include engi- neers with “design” knowledge and experience in the start- Recommendation 11. All stakeholders and involved regu- up crew for new plants. latory agencies should agree that a single team will investi- gate chemical events requiring outside review. This investi- Recommendation 13. A generous allotment of time should gation team should comprise already-appointed representa- be given to training and retraining chemical demilitarization

OCR for page 51
55 FINDINGS AND RECOMMENDATIONS plant operating personnel to ensure their total familiarity with operational risk. Each facility should develop training pro- the system and its engineering limitations. All plant person- grams using the newly designed in-plant simulators to nel should receive some education on the total plant opera- present challenges that require knowledge-based thinking. tion, not just the area of their own special responsibility. The The training programs should include a process for judging extent of this overall training will be a matter of judgment the effectiveness of the training. Including “design” experts for plant management, but the training should focus on how in the start-up crew for new plants could be helpful in iden- an individual’s activities affect the integrated plant and its tifying latent failures in process and facility design.

OCR for page 51