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 7
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities 1 The Chemical Demilitarization Challenge For more than 50 years the United States has maintained an extensive weapons stockpile containing chemical agents, stored primarily in military depots distributed in the continental United States. Largely manufactured 40 or more years ago, the chemical agents and associated weapons in this stockpile are now obsolete. Under a congressional mandate (Public Law 99-145), in 1985 the Army instituted a sustained program to destroy elements of the chemical weapons stockpile and extended this program to destroy the entire stockpile when Congress enacted Public Law 102-484 in 1992. Chemical weapons stored overseas were collected at Johnston Island, southwest of Hawaii, and destroyed by the Johnston Atoll Chemical Agent Disposal System (JACADS), the first operational chemical demilitarization facility. JACADS began destruction activities in 1990 and completed processing of the 2,031 tons of chemical agent and the associated 412,732 munitions and containers in the overseas stockpile in November 2000 (U.S. Army, 2001a). The largest continental U.S. stockpile component, which initially contained 13,616 tons of agent, is stored at the Deseret Chemical Depot (DCD) near Tooele, Utah. This component of the stockpile is being processed by the Tooele Chemical Agent Disposal Facility (TOCDF), which started operation in August 1996 and destroyed 5,320 tons of agent and processed more than 880,000 munitions and containers in its first 5 years of activity. As of September 2002, the first two chemical demilitarization facilities had destroyed over 25 percent of the original chemical agent tonnage (U.S. Army, 2002a). The Army, through its Office of the Program Manager for Chemical Demilitarization (PMCD), now has more than 15 years of cumulative operating experience in chemical weapons demilitarization. PMCD plans to open three additional facilities in the near future to meet Chemical Weapons Convention1 requirements for destruction of the U.S. stockpile. Despite the progress made to date, however, operations at JACADS and TOCDF have not been without incident. Several “chemical events” at the two plants have resulted in either unplanned discharge of significant amounts of agent within the facilities and/or the release of very small amounts of agent to the atmosphere above these plants. The following sections in this chapter discuss the chemical demilitarization challenge: how to safely destroy the stockpile of chemical weapons within the available time constraints imposed by a dangerous and deteriorating stockpile. To put this challenge in context the Committee on Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities describes technology for the chemical stockpile’s disposal, defines and describes chemical events, discusses the significance of risk assessment to the chemical weapons disposal process, and categorizes institutional issues associated with chemical demilitarization. STOCKPILE CONTENT, DISPOSAL DEADLINE, AND DISPOSAL TECHNOLOGY The chemical weapons stockpile contains two types of chemical agents: the cholinesterase-inhibiting nerve agents (GB and VX), and blister agents, primarily mustard (H, HD, and HT) but also a small amount of Lewisite. Both types of chemical agents, which are liquids at room temperature and normal pressures, are frequently, but erroneously, referred 1 Formally known as the Convention on the Prohibition of the Development, Production, Stockpiling and Use of Chemical Weapons and on Their Destruction (P.L. 105-277), the CWC requires the destruction of chemical weapons in the stockpile by 2007 and any non-stockpile weapons in storage at the time of the treaty ratification (1997) within 2, 5, or 10 years of the ratification date, depending on the type of chemical weapon or on the type of chemical with which an item is filled.
OCR for page 8
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities FIGURE 1-1 Location and size (percentage of original stockpile) of eight continental U.S. storage sites. SOURCE: NRC (1997). to as gases. The stockpile contains both bulk (“ton”) containers of nerve and blister agent and munitions, including rockets, mines, bombs, projectiles, and spray tanks loaded with either nerve or blister agents. Many munitions contain both chemical agent and energetic materials (propellants and/ or explosives), a combination that poses particular challenges for safe and efficient destruction. The disposal of stockpiled chemical weapons is a major undertaking. In 1990, the stockpile included 31,496 tons of chemical agents. The current stockpile is stored at eight chemical weapons depots operated by the Army in the continental United States. The location, size, and composition of the original continental U.S. stockpile is presented in Figure 1-1. The U.S. Chemical Stockpile Disposal Program (CSDP) has evolved in parallel with international initiatives to eliminate chemical weapons. After many years of negotiation, the terms of the CWC were agreed upon in 1993 to deal with this issue. As of June 2002, the CWC had been signed by 174 countries and ratified by 145. The convention went into effect on April 29, 1997, after ratification by 65 countries. The CWC requires that signatories, which include the United States, destroy their chemical weapons stockpiles within 10 years of its initiation, making April 29, 2007, the deadline for destruction of the U.S. stockpile. A provision in the treaty allows a 5-year extension of the deadline under some circumstances. As of early October 2001, PMCD released new schedule estimates indicating that chemical demilitarization activities at the three disposal facilities scheduled to commence operation in the near future may not be completed until 2008 at Pine Bluff, Arkansas, and until 2009 at Anniston, Alabama, and Umatilla, Oregon (U.S. Army, 2001b). The disposal technology selected by the Army for storage sites that contain a full range of chemical agents and munitions types is a multifurnace incineration process (NRC, 1999a). In this “baseline” technology approach munitions and containers are drained of agent, which is burned in dual liquid incinerators (LICs). Robotic machinery disassembles munitions containing energetic charges and the separated energetic materials are burned in a rotary kiln-based deactivation furnace system (DFS). Sheared bulk containers and metal munitions parts are fed though a large heated metal parts furnace (MPF) designed to burn off any residual agent or energetic material, decontaminating metal components to the point that they can be recycled as normal scrap metal. The LIC, DFS, and MPF furnaces are all equipped with extensive pollution abatement systems (PASs) designed to substantially eliminate gaseous and particulate exhaust material of potential concern and exhaust remaining gases through a common stack. The first-generation incineration system was deployed at JACADS. The second-generation system, deployed at
OCR for page 9
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities TOCDF, is described in detail in the National Research Council report Tooele Chemical Agent Disposal Facility— Update on National Research Council Recommendations (NRC, 1999a)2; this detailed description is reprinted in Appendix A. Third-generation incineration systems are currently close to operational status at the Pine Bluff, Anniston, and Umatilla sites. The new facilities will use basically the same process as that used at TOCDF and JACADS. Weapons will be taken apart in the same way, and there will be the same three lines of incineration: a rotary furnace for destroying propellant and explosive materials (see Appendix A, Figure A-6), a furnace with a moving conveyor primarily for decontaminating metal parts (see Figure A-7), and a furnace for burning liquid agent (see Figure A-8). Improvements to the new facilities have been made compared with TOCDF and JACADS, however; these are noted in Chapter 5. In addition, the Army has selected liquid-phase hydrolysis processes, supplemented by various secondary hydrolysate treatment and/or disposal processes, to destroy the chemical agents contained only in bulk “ton” containers at Newport Chemical Depot in Indiana and at the Edgewood Chemical Activity site on Aberdeen Proving Ground in Maryland. Significant problems in introducing these new technologies to dispose of even the simplest case of “bulk only” chemical agent have been recognized (NRC, 2000a). A disposal technology has not yet been selected for the relatively small stockpile of nerve and mustard munitions at the Blue Grass Chemical Depot in Kentucky. At the Pueblo Chemical Depot in Colorado the Department of Defense has decided to use neutralization, followed by bio-treatment of the secondary waste to dispose of the mustard munitions stored there. CHEMICAL EVENTS During the 10 years of JACADS operation to destroy the chemical weapons stockpile at Johnston Island and the first 5 years of operation of TOCDF, a range of operating incidents occurred that were designated as chemical events by the local Army depot commanders. Army Regulation 50-6 on Chemical Surety (U.S. Army, 1995) defines chemical events very broadly: “The term chemical event encompasses all chemical accidents, incidents and politically/public sensitive occurrences.” The regulation goes on to give specific examples, such as: Confirmed releases of agent from munitions outside a closed containment system, such as a filtered bunker, storage igloo, or overpack container. Discovery of an actual or suspected chemical agent container or munition in a place where it is not supposed to be that may require emergency transportation or disposal. Confirmed detection of agent above the threshold concentration for any period outside the primary engineering control. Actual exposure of personnel to agent above the allowed limits specified in various Army regulations. Loss of chemical agent. Any terrorist or criminal act directed toward a chemical agent storage, laboratory, or chemical demilitarization facility or any deliberate release of chemical agent. Any malfunction or other significant activity at a chemical demilitarization plant that could reasonably be expected to cause concern within the local community or the press, or that in the judgment of the local facility or installation management or leadership could cause embarrassment to the U.S. Army. At the eight continental U.S. storage sites, the Army’s local depot commander has the responsibility to decide whether an upset or incident within the storage yard or at the associated chemical demilitarization facility is a chemical event. Examples 1 through 6 above seem to imply that, in most cases, chemical events are those in which chemical agent ends up where it should not be, i.e., in the ambient atmosphere or under the control of an unauthorized individual. However, no such requirement is inherent in example 7. The wide latitude in judgment about what might “cause concern within the local community or the press, or . . . could cause embarrassment to the Army” that is delegated to the depot commander suggests that some incidents defined as chemical events by one commander may not be considered chemical events by another. Whatever the local Army commander deems to be a chemical event is subject to strict reporting procedures detailed in Army Regulation 50-6 (U.S. Army, 1995). Both telephone reports (within 3 hours) and initial written reports following specified formats (within 24 hours) must be made to both the Army Operations Center and Headquarters–Department of the Army. These reports are usually shared with local authorities and serve as the basis for press releases issued by the local depot and/or PMCD, but there appear to be no general guidelines for the form and the timing of such notification. A tabulation of chemical events is also provided in the Army’s annual reports to Congress summarizing chemical agent storage and chemical demilitarization activities (see, for example, U.S. Army, 2000a). Chemical Events Associated with Disposal The Program Manager for Chemical Demilitarization (PMCD) is required to prepare and provide reports of chemi- 2 This update report also details TOCDF technology and management issues identified by an NRC oversight committee during that facility’s first 3 years of operation.
OCR for page 10
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities cal events for incidents so designated that occur within its facilities, as specified in Army Regulation 50-6.3 PMCD provided to the committee a chronicle of 81 incidents that occurred over the 10 years of JACADS operation and the initial 5 years of TOCDF operation (U.S. Army, 2001c). As shown in the Army’s document (see Appendix B), a significant number reported under example 7 listed above did not involve chemical agent. Of those events involving chemical agent, only a few resulted in release of agent outside of engineering control and into the atmosphere. The total mass of chemical agent released to the environment in these incidents was almost certainly less than a gram (U.S. Army, 2001d), which is equivalent to no more than a few drops. The committee’s analysis of PMCD-reported chemical events at the two chemical demilitarization facilities is presented in Chapter 2. In addition to a list of PMCD-reported chemical events, the committee also received lists of possible chemical-agent-related incidents from local officials and concerned citizens groups (see Appendixes C and D). These are also addressed in Chapter 2. Chemical Events Occurring During Storage A listing and analysis of the chemical incidents involving leaking containers and munitions at the Johnston Island storage site from 1990 until the end of disposal operations in 2000 and at the Tooele site (now the DCD) from 1990 through 2000 were provided to the committee by the Army’s Soldier Biological and Chemical Command’s (SBCCOM’s) Stockpile Management Team (U.S. Army, 2001e). As a result of its continuous stockpile inspection program SBCCOM has records of the frequency of chemical agent leaks occurring in stockpiled munitions and containers. Most of the incidents listed by SBCCOM involved a single leaking munition or container, although incidents involving more than one leaking munition discovered in a storage igloo were not uncommon. One incident involved 20 leaking munitions treated over the course of a month. The most serious incidents, including all those known to have discharged a significant amount of agent outside of engineering control, were designated as chemical events and reported as required by Army Regulation 50-6. According to the SBCCOM statistics on stockpile leakage at Johnston Island, 13 incidents involving leaking munitions were reported from 1990 through 2000. Ten of these occurred from 1990 through 1992, with only 3 occurring in 1993 or later. Some of the later falloff is likely due to reduction of the stockpile as chemical demilitarization proceeded, with the most problematic munitions scheduled for the earliest destruction, according to the overall risk mitigation strategy outlined below in this chapter. Inspection and remediation of corroded or leaking munitions prior to their shipment to Johnston Island probably also contributed to the fact that so few munitions and containers leaked in the Johnston Island storage site. The statistics for stockpile leakage at the Tooele site (DCD) for the same period from 1990 to 2000 differ considerably from those for JACADS. SBCCOM tabulated 31 incidents for 1990, 34 for 1991, 40 for 1992, 37 for 1993, 38 for 1994, 33 for 1995, 26 for 1996, 14 for 1997, 14 for 1998, 10 for 1999, and 11 for 2000, for an 11-year total of 288 (U.S. Army, 2001e). Again the record of events suggests that as stockpile destruction proceeds, with the most problematic weapons and containers scheduled for early destruction, the stockpile leakage statistics improve. From 1990 through 1995, 72 percent (159) of the 213 tabulated stockpile leakage incidents involved GB, which was the agent type destroyed at TOCDF through March 2002. Some of the 288 storage chemical agent incidents at DCD from 1990 through 2000 involved only relatively low storage igloo vapor readings with most, possibly all, of the released agent captured in the carbon air filters present in igloos storing the highest-risk munitions (generally those containing GB) (U.S. Army, 2001e). In more that 100 different incidents, however, storage personnel who entered an igloo reported observing liquid agent leaks with volumes estimated to range between 1 teaspoon and 2 quarts, with one leak from a GB ton container totaling 10 gallons. In dozens of additional incidents smaller amounts of leaked liquid were observed. Since most of these incidents involved high-vapor-pressure GB, many resulted in release of small amounts of agent to the atmosphere, if only when the igloo door was opened to allow entry. Many of these leaks were initially detected by monitoring igloo air drawn through sampling ports. Mobile powered air filtration units were often used to minimize agent migration out of the igloo. In addition, in 11 incidents at DCD from 1990 through 2000, ton containers of mustard, stored outside, were found to be leaking agent directly into the environment (U.S. Army, 2001e). When these incidents occurred, storage site personnel attempted to quantify the amount of agent lost by estimating the volume of contaminated gravel or soil underneath the leaking container, or, in the case of large leaks, measuring the agent remaining in the container. For most of the outdoor incidents documented at DCD, relatively small amounts of agent (a few drops to a few cups) were estimated by depot workers to have been lost. However, in the most serious event, a leak of distilled mustard estimated at 78 gallons (~375 kg) was discovered on September 9, 1993. The volume of agent released in this incident alone swamped the total mass of known emission of agent from chemical demilitarization facilities by at least a factor of several hundred thousand. Large outdoor releases from storage facilities are an ongoing concern. In fact, while the committee was gather- 3 Detailed guidance on the preparation and distribution of these reports and on associated record keeping is presented in a periodically updated document designated PMCD Regulation 385-3, “Accident and Chemical Event Notification, Investigation, Reporting and Record Keeping” (U.S. Army, 1999a).
OCR for page 11
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities ing data for this report during a visit to TOCDF on July 26, 2001, a leaking plug in a ton container of HD in the DCD storage site produced a vapor plume large enough to force workers at adjacent TOCDF to don respirators. That leak, according to chemical event reports and related memoranda supplied by SBCCOM, was determined to be about 9 pounds (~4 kg). This incident delayed the committee’s access to TOCDF for several hours. While chemical demilitarization operations at JACADS and TOCDF have released small amounts of chemical agent into the environment, these releases are negligible compared with releases to the environment from associated chemical weapons storage sites. The rate of agent leaks and releases does decrease significantly as the stockpile is processed. TOOLS FOR ASSESSING HAZARDS IN THE OPERATION OF CHEMICAL STOCKPILE DISPOSAL FACILITIES The Army has developed a suite of risk assessment and risk management tools to permit analysis of potential risks in terms of the scenarios that can contribute to risk, the likelihood of those scenarios, and the consequences associated with them. Those consequences are as follows (NRC, 1997, p. 16): For humans (both workers and the public) there are three potential measures of risk either from the stockpile or from stockpile destruction: acute lethality; acute and latent noncancerous health effects; and latent cancer. The potential adverse consequences for the environment are the contamination of land and/or water and adverse effects on native or endangered species. These tools can be used to evaluate the risk associated with specific chemical events. Real-world events can also then be used as a check on the analyses, enabling revision of risk analyses to include new classes of events when surprises occur. The variety of analysis tools is useful because of the differing needs of various program elements. To understand how they are related, the committee first groups these tools into two large classes: prospective (or predictive) tools and retrospective (or documentation) tools.4 Prospective Risk Analysis Tools Health Risk Assessment A health risk assessment (HRA) is a compliance-oriented analysis that examines the risk to a set of stylized receptors (e.g., the subsistence fisherman) associated with routine releases (intended to be conservative upper bounds based on tests and performance of other units) and mild upset conditions (assumed to lead to release of a multiple of the conservative routine release for a specific fraction of the year). Accidents, specific systems failures, and specific human actions are not considered. The HRA is an upper-limit risk estimate for routine operations. Because it does not provide a realistic estimate (accounting for uncertainty), does not consider accidents, and does not address worker risk, it is not helpful in evaluating chemical events, other than providing a baseline, of sorts, against which the consequences of chemical events can be evaluated. For an example HRA see the analysis of TOCDF sponsored by the Utah Division of Solid and Hazardous Waste (Utah DEQ, 1996). Systems Hazards Analysis A systems hazards analysis (SHA) is a systematic and comprehensive search for and evaluation of all significant failure modes of facility systems components that can be identified by an experienced team. The hazards assessment often includes failure modes and effects analysis, fault tree analysis, event tree analysis, and hazards and operability studies. Generally, the SHA does not include external factors (e.g., natural disasters) or an integrated assessment of systems interactions. However, the tools of SHA are valuable for examining the causes and the effects of chemical events. They provide the basis for the integrated analysis known as quantitative risk assessment. For an example SHA see the TOCDF Functional Analysis Workbook (U.S. Army, 1993-1995). Quantitative Risk Assessment A quantitative risk assessment (QRA) is an integrated, quantitative analysis (including uncertainty) of accident scenarios, their likelihood, and possible consequences. Current QRAs examine human actions as well as systems failures, external events as well as internal failures, and worker risk as well as public risk. A salient feature of a QRA is that it is integrated, in that it: considers the interactions of systems and their effects on each component, considers common causes of failures, and considers all forms of system dependencies considers the integrated impact of multiple system and human failures on the potential for releases considers the impacts of weather and emergency protection on public consequences Thus, the QRA provides an effective tool for evaluating the significance of chemical events. In fact, scenarios leading to chemical events and the frequency and consequences of these events are exactly what a QRA describes and calculates. Real-world events provide a check on the analysis. If 4 The committee uses the Army’s names and acronyms for these methods. Use of these names is not consistent with language in other environments.
OCR for page 12
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities potentially risk-significant events occur that were not previously modeled, the QRA can and should be updated to account for that event and any similar events that could occur. For an example QRA see the Army’s TOCDF QRA (U.S. Army, 1996a). The committee notes that the TOCDF QRA was the first PMCD QRA and does not include all the features in the current analyses being finalized for the facilities at Anniston, Umatilla, and Pine Bluff. At the time of this writing the TOCDF QRA is the only one that has been published. Similar QRAs are being completed at the remaining sites. It is possible that portions of these may be unavailable publicly because of security concerns. Key elements of the Army’s approach to quantitative risk assessment are summarized in Appendix E for the interested reader. More details are available in the NRC Stockpile Committee’s risk report (NRC, 1997). Retrospective Analysis Tools Monitoring Systems Monitoring systems detect releases of hazardous chemicals, providing warning of hazardous conditions and a record of their occurrence and extent. They can also measure the burden of chemicals on the human body. They are not predictive, but instead provide real-time observations. For a description of monitoring schemes, see Box 1-1 and the NRC Stockpile Committee’s report Occupational Health and Workplace Monitoring at Chemical Agent Disposal Facilities (NRC, 2001a). Chemical Event Investigations When chemical events occur, investigations identify what actually happened and when, the reasons, and the consequences; they usually suggest corrective actions for the future. An investigation (separate from possible corrective action) is most effective when it focuses on what actually happened from the viewpoint of those involved (i.e., why the actions of people involved made sense to them at the time, what they could see and what they knew, how they viewed their alternatives) (Weick and Sutcliffe, 2001). Too often these investigations are biased by hindsight and focus on what the operators might have done rather than why they did what they did. An effective investigation identifies the organizational and management issues that made the actions seem reasonable to those involved, and it can provide a basis for real improvement. Chemical event investigation and analysis are the subjects of Chapter 2 of this report. Putting It All Together From this brief introduction, it is clear that the QRA, chemical agent monitoring, and event investigation are the key tools for addressing the safety issues associated with chemical events. In its published QRA the Army performed a detailed assessment of the risk of public fatalities and cancers associated both with the stockpile storage sites and chemical weapons processing activities at Tooele (U.S. Army, 1996a), and it has performed ongoing risk assessments for the planned third-generation incineration system chemical demilitarization facilities and associated stockpile storage areas at Pine Bluff, Anniston, and Umatilla. In its QRA for TOCDF, the Army’s analyses indicated that, over the facility’s projected operating schedule, the risk associated with accidental releases of agent due to disruption of the stockpile, most likely due to earthquake or leaks from ton containers of GB, greatly outweighed the risk of release of agent due to chemical demilitarization activities (U.S. Army, 1996a). This risk assessment does not examine potential terrorist activities, threats that are addressed by other federal agencies in addition to the Army. The Army’s risk assessment for TOCDF and its associated storage facility was reviewed by the NRC and found to be sound (NRC, 1997). Even in the event of an earthquake or plane crash that damages the disposal plant, the risk of public fatalities due to release of agent from the disposal facility is calculated to be about 5 percent of the expected risk of fatality due to releases of agent from the storage yard (U.S. Army, 1996a; NRC, 1997). A more detailed discussion of the TOCDF QRA and of advances incorporated in subsequent QRAs is presented in Appendix E. Until the last few days of the disposal schedule, the amount of agent in the storage yard greatly exceeds the amount in the chemical demilitarization plant; as the stockpile is depleted, the risk posed by the storage facility drops proportionally. A key risk management strategy adopted by the Army is to order the stockpile destruction so that the most volatile, highly toxic agent and associated munitions are processed first (those containing the nerve agent GB), while less volatile and/or deadly agents are processed later. Finally, it is important to note that the original TOCDF QRA focused on public risk, and little effort was devoted to examining worker risk. One consequence of this limitation in scope was that very little modeling of human performance was done in the TOCDF QRA. As attention in the program shifted to include worker risk, more significant modeling of human action has been performed. None of these improved analyses have yet been published. A variety of human reliability analysis methods have been used (Gertman and Blackman, 1994). For ongoing work, new approaches that account for details of context and human cognitive function are being adapted (Hollnagel, 1998; USNRC, 2000). With more careful and complete analysis, new scenarios especially important to worker risk are being developed. These methods, integrated into the risk assessments, can be used to quantify the impact of human actions on situations posing risk. Human performance not only is a significant component in risk assessment, but also, as the committee learned in its study, is directly involved in most chemical events. In
OCR for page 13
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities BOX 1-1 Details on Airborne Chemical Agent Monitoring Methods and Standards at Chemical Demilitarization Facilities Two systems are currently used to monitor concentrations of airborne chemical agents at chemical demilitarization facilities. One system, the automatic continuous air monitoring system (ACAMS), is designed for “near-real-time” monitoring (currently ~3- to 8-minute cycle time, dependent on agent, for a single instrument). The ACAMS consists of an air sampling system connected to a gas chromatograph (GC) equipped with a flame photometric detector (FPD). Specific columns and detector filters are used for each agent. The nerve agents, GB and VX, are detected by phosphorus oxide chemiluminescence, due to their P content, excited in the FPD detector, while mustard is detected by sulfur dimer chemiluminescence, from its sulfur content. Since VX has high molecular weight (298 amu), it is catalytically cleaved at the entrance to the GC column to shorten its detection time. This detection scheme relies on the characteristic GC column transit time of the agent or agent fragment (in the case of VX) plus the P or S spectrally specific flame chemiluminescence detection signal to identify the agents. The method is quite sensitive; ACAMS are often run at threshold detection volume-mixing ratios of a part per trillion (pptv) or lower. However, at these low threshold levels false positive alarms often occur because other chemical species can “interfere” by producing chemiluminescent signals that overlap the time gate and spectral band pass associated with the agents. For time-critical applications, like exhaust stack monitoring, the GC cycle time can be mitigated by time phasing two or more ACAMS sampling the same gas stream. ACAMS alarms must be verified to ensure that they are not a false positive due to an “interferent” species or instrument malfunction. This verification is done using a depot area air monitoring system (DAAMS) deployed near an ACAMS. DAAMS is a passive system that draws an air stream through a sorbent tube. The tubes are collected and replaced periodically if there are no ACAMS alarms or shortly after an alarm occurs. They are transported to a laboratory and thermally desorbed onto a sample tube and analyzed on a laboratory scale GC/FPD system. Without confirmation by the more sensitive and specific laboratory GC/FPD system, the ACAMS alarm is not confirmed. If the laboratory GC/FPD system does not show a chromatogram consistent with agent, a second DAAMS sample may be run on a laboratory GC equipped with a mass spectrometric detector (MSD). The GC/MSD analysis is designed to identify interferent compounds that may have caused a false positive ACAMS alarm. The Army currently mandates very conservative alarm thresholds for it chemical demilitarization facilities (U.S. Army, 1997a; NRC, 2001a). Current exhaust stack alarms are set at 0.2 of the allowable stack concentration (ASC), for GB the ASC is just three times the time weighted average (TWA), which serves as the worker population limit (WPL) for the demilitarization workforce. The TWA is the level of agent an unmasked person can breathe for an 8-hour shift without harm. Thus, GB, the most volatile agent and therefore the greatest airborne threat to surrounding communities, is monitored at stack concentrations equivalent to 0.6 of the level currently deemed safe for a worker to breathe for a full shift without protection. Stack exhaust monitoring levels for the less volatile and less threatening HD and VX are monitored at levels a factor of 2 and 6 above their TWA (WPL) levels, respectively. The 0.2 ASC stack level for GB is a factor of 10,000 below the “immediately dangerous to life and health” (IDLH) level for this agent. In-plant air levels breathed by unmasked workers and the output of the scrubbing system for air exiting the demilitarization plant are monitored at similarly conservative levels (generally 0.2 TWA). Since any agent in either the stack exhaust gas or scrubbed plant air will be greatly diluted before reaching the facility’s fence line, air flowing into the surrounding communities will be well below the “general population limit” (GPL) defined as the level believed to pose no threat to the public. The GPL for the three stockpile agents is set at 30 to 33 times lower than their TWA (U.S. Army, 1997a; NRC, 2001a). The current ASC, TWA, and GPL levels for GB are 3 × 10−4, 1 × 10−4, and 3 × 10−6 mg/m3. For VX these values are 3 × 10−4, 1 × 10−5, and 3 x 10−6 mg/m3, while for HD they are 3 × 10−2, 3 × 10−3, and 1 × 10−4 mg/m3 (NRC, 2001a; U.S. Army, 1997a). Chapter 2, the committee examines the more significant of the chemical events at JACADS and TOCDF to determine their characteristics with respect to facility performance and human performance. How these events are related to safety performance is not a simple question. In his widely referenced book (Reason, 1997), in a chapter devoted to the relationship between frequent, low-consequence events and the risk of high-consequence events, James Reason concludes that: If both individual and organizational accidents have their roots in common systemic processes, then it could be argued that . . . personal injury statistics are indicative of a system’s vulnerability (or resistance) to organizational accidents. The number of personal injuries sustained in a given time period must surely be diagnostic of the “health” of the system as a whole. Unfortunately, this is not so. The relationship is an asymmetrical one. An unusually high [personal injury rate] is almost certainly the consequence of a “sick” system that could indeed be imminently liable to an organizational accident. But the reverse is not necessarily true. A low . . . rate (of the order of 2-5 per million man hours)—which is the case in many well-run hazardous technologies—reveals very little about the likelihood of an organizational accident.
OCR for page 14
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities The problem of human-caused events, how to control them, and how to discern the difference between high- and low-risk events continues to be studied in many industries (Reason, 1997; Hollnagel, 1998; IOM, 2000). Monitoring Methods The occurrence and the extent of a release of chemical agent are tracked through PMCD’s workplace chemical agent monitoring system as described in NRC (2001a). Monitoring for airborne agent is a major activity at each chemical agent disposal facility. Box 1-1 provides details on monitoring. Sensitivity requirements for the near-real-time automatic continuous air monitoring system (ACAMS) for airborne agent are demanding. This is because the allowable stack concentration and time-weighted-average levels used for exhaust stack and in-plant action levels are quite low and because the ACAMS alarms are currently set at 0.2 of the relevant action level. This demand for sensitivity results in relatively frequent false positive alarms, particularly for the ACAMS monitoring the individual incinerator exhaust flows and the common exhaust stack (NRC, 1999b). Previous NRC reports have noted that the frequency of false positive ACAMS alarms disrupts plant operations, particularly when stack alarms trigger an automatic shutdown of agent feed to the liquid incinerator, and can lead to an unsafe “crying wolf” mind-set that tends to discount ACAMS alarms (NRC, 1999b, NRC, 2001a). In fact, one of these studies found evidence that the May 8-9, 2000, agent stack release at TOCDF was exacerbated by an expectation that what proved to be real exhaust system ACAMS alarms were instead just false positives (NRC, 2001a). While previous NRC reports have urged PMCD to improve both the reliability and time response of its airborne agent monitoring systems (see NRC, 1999b for a summary), progress in this area has been modest. Another weakness of the airborne monitoring system is the lack of real-time (<10 seconds) agent detection. The NRC has previously recommended that the Army develop a real-time system that uses a measurement technology independent of the gas chromatography with flame photometric detector methods used by the ACAMS and the depot area air monitoring system (DAAMS) (NRC, 1994). To date, the Army’s attempts to develop and demonstrate such a real-time system have not been successful (NRC, 1999b, 2001a). New interest in chemical agent detection as a key component of antiterrorism activities has spurred government and commercial activities focused on developing better sensors for airborne agent (IOM, 1999). The NRC has also urged the Army to continue to monitor technological advances in trace gas detection and to consider implementing any that are appropriate for monitoring agent in chemical munitions disposal facilities (NRC, 1999b). Renewed interest in chemical agent detection and monitoring methods spurred by homeland defense concerns may lead to better and more robust technology. The committee urges PMCD to vigorously seek out and exploit any suitable developments arising from these activities. Previous NRC reports have also noted the lack of robust techniques for rapidly measuring agent and agent breakdown products present in liquid waste streams and associated with solid materials (NRC, 2000a; NRC, 2000b; NRC 2001a). These reports recommend vigorous efforts to develop better methods to measure agent contamination in these media. Event Analysis and Significance The committee notes the importance of chemical event analysis that focuses on the viewpoint of the operators during the sequence of events. Understanding why their actions seemed appropriate to them, at the time, is the key to effecting real improvement in performance. Gaining an understanding of the factors within their work environment—training, equipment, and operational indications, as well as goals and rewards—which led them to conclude that their actions were appropriate is an essential element of developing a real safety culture at the facility. An associated effort is to ensure that the QRA includes the class of events that actually have occurred. Mapping real event scenarios onto scenarios modeled in the QRA allows one to see a particular action integrated into the larger system for each chemical event and thus determine its effect on safety. CHEMICAL DEMILITARIZATION INSTITUTIONAL ISSUES Trust and Institutional Arrangements The chemical demilitarization program necessarily depends on a combination of trust and institutional arrangements to accomplish the destruction of the chemical stockpile. Because extremely hazardous materials and complex technologies are involved, those seeking destruction of chemical agent and munitions must rely on agencies and firms expert in these processes to carry out the chemical demilitarization program. In essence, legislation and regulatory agency rule making establish institutional and contractual arrangements for the chemical demilitarization program, stipulating what is to be accomplished and (in some cases) how it is to be done. As in any contract, the “principal” relies on an “agent”5 to accomplish a task or service, and provides 5 It is unfortunate that use of the term “agents” to indicate those that carry out tasks for “principals” might in this report be a source of confusion in the context of the chemical demilitarization program (where “agent” usually refers to chemical agent). Where agent is used in this report in the institutional sense, it is italicized to reduce the potential for confusion.
OCR for page 15
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities the means to ensure that the task is accomplished according to the principal’s needs (Wood, 1992; Scholz and Wei, 1986). The U.S. Congress and the public it represents rely on agencies of the U.S. Army, state and federal regulators, private contractors, and a host of other entities to carry out the chemical demilitarization program. At specific chemical demilitarization sites, the local public and the officials who represent them similarly depend on these agents to carry out the task safely and effectively. When principals delegate complex tasks, they create a relationship in which the agents on whom they rely are more knowledgeable about the task than are the principals. Agents design, test, construct, operate, and modify the chemical demilitarization facilities and have intimate knowledge of these steps, while principals often rely on agents for such knowledge. This kind of information asymmetry may place the principal at a disadvantage in overseeing the safety and effectiveness of the program, and necessitates monitoring and control mechanisms that are specified in the relevant laws and contracts. Monitoring mechanisms include permitting and reporting requirements, inspections, investigations, and rules governing whistle-blowers, while control mechanisms include arrays of incentives (such as contract fee structures) and sanctions (civil and criminal punishments, fee deductions, and so on). A trade-off implicit in this relationship is that of the principal’s control over agents versus the scope of the agent’s discretion, some of which is typically necessary for complex and demanding tasks that require the agent to push the boundaries of known processes and technologies. Greater trust reduces the need to rely on formal monitoring and control, and conversely loss of trust increases the need for monitoring and controls. The Institutional Setting of Chemical Demilitarization The U.S. government’s approach to chemical demilitarization involves a complex amalgam of institutional stakeholders. The Army’s SBCCOM is the operator of the eight remaining stockpile storage facilities. PMCD is responsible for the construction, operation, and subsequent closure of JACADS, TOCDF, the three new incinerator system facilities at Anniston, Umatilla, and Pine Bluff, and the two-bulk only, hydrolysis-based facilities under construction at Newport and Edgewood (Aberdeen). By law, evaluation of alternative (non-incineration) technologies that may be used to dispose of the stockpiles located at Pueblo, Colorado, and Blue Grass, Kentucky, has been delegated to an independent Program Manager for Assembled Chemical Weapons Assessment (PMACWA) within the Army. Protection of the public from harm due to accidental releases of agent near storage depots and associated chemical demilitarization facilities is the responsibility of the Chemical Stockpile Emergency Preparedness Program (CSEPP), which is funded by the Army but administered by the Federal Emergency Management Agency (FEMA). Thus, at any of the six continental sites where a chemical demilitarization facility is either operating or under construction, a concerned citizen needs to receive a consistent and accurate message from a range of state and federal entities including PMCD, SBCCOM, FEMA, and CSEPP. In the past a consistent message from these Army or Army-related entities has been hard to achieve (NRC, 1999a). In addition, chemical demilitarization facilities must obtain environmental permits from state environmental agencies in order to commence operations and must seek permit amendments and renewals from these same agencies in order to sustain operations. Permit conditions may vary widely from state to state even though the state environmental agencies operate largely under authority delegated from and overseen by the federal Environmental Protection Agency. State-to-state discrepancies in chemical demilitarization facility operating permits or amendments to existing operating permits may raise public concerns. Hearings related to environmental permit applications and amendments give citizens an important opportunity for input into the operation of chemical demilitarization plants. The PMCD receives guidance from the Army Medical Services (U.S. Army Center for Health Promotion and Preventive Medicine), working in conjunction with the Department of Health and Human Services’ Centers for Disease Control and Prevention, about the levels of exposure to chemical agent that are considered safe for both workers and the public (U.S. Army, 1990, 1991). Recent reevaluations have led to proposals for significantly lower recommended standards related to exposure to chemical agents. This possibility has raised citizen concern about the safety of Army stockpile storage and chemical demilitarization operations designed to meet the current exposure limits. Chemical events have raised questions about the safety of the stockpile storage and the demilitarization process. Understanding whether an event results from flaws in design, fundamental problems with technologies, organizational failures, or personnel lapses is essential for determination of appropriate responses. Because the answers to these questions materially affect the circumstances of the agent, concerns about whether agents are sufficiently forthcoming and responsive are inevitable. The U.S. Congress has responded to such concerns with diligent oversight (including the request for this report), requirements that whistle-blowers be protected from retaliation, and requests for formal annual reports from the Army on the progress of chemical demilitarization and the occurrence of any chemical events associated with either chemical demilitarization or storage facilities. REPORT ROADMAP The chemical events that have occurred at JACADS and TOCDF are characterized and a selected subset analyzed in Chapter 2. Chapter 3 discusses protocols and pro-
OCR for page 16
Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities cesses for reporting chemical events, outlines how selected events were reported at both facilities, and discusses how these events affected plant operator interactions with other stakeholders, including environmental regulators, elected state and local officials, and the public. Chapter 4 discusses the implications of lessons learned from past chemical events and their impact on continuing operations at TOCDF and future operations at Anniston, Umatilla, and Pine Bluff. Prudent preparations to minimize the occurrence and impact of future chemical events at incineration system chemical demilitarization facilities are discussed in Chapter 5. Chapter 6 contains focused findings and recommendations drawn from material presented in the first five chapters.
Representative terms from entire chapter: