6
Opportunities for Improved Chemical Agent Monitoring

The preceding chapters reviewed chemical agent monitoring requirements and challenges, the present systems in use at operating chemical stockpile demilitarization facilities, and the types and capabilities of available monitoring alternatives. It is clear that the present monitoring systems are functioning adequately to protect the health and welfare of workers as well as of the neighboring public. The monitoring systems are distributed throughout the demilitarization facilities and are integrated with plant operating and maintenance procedures, control systems, and systems for the early detection of malfunctions or leakages. Given that the disposal operations will be on-going for some years, at least until 2012 and perhaps beyond, this chapter focuses on the possible use of new, potentially real-time measurement technologies that are compatible with the existing monitoring systems or which provide supplemental capabilities that address some other needs.

PRESENT MONITORING NEEDS AND CAPABILITIES

The chemical agents stored and processed at stockpile sites were designed for use in warfare and therefore have the potential to cause fatal or chronic illnesses, depending on the type of agent, the concentration level, and the time of exposure. In combat, these agents are explosively deployed or sprayed to create aerosols that maximize airborne agent concentrations in target areas.

In stockpile storage, the munitions are stored in igloos and are inspected periodically for leakage. Any munitions that are found to be leaking are overpacked to provide secure containment. Agent stored in bulk (ton containers and spray tanks) has no explosives or propellants present and may be stored in open yards or warehouses. When munitions or bulk containers are transported to a processing facility for disposal, they are generally loaded by workers wearing protective gear into a secure, explosion-resistant container for movement to the unpack area of the facility. Mobile, near-real-time (NRT) airborne agent monitors are used for the periodic checks of igloos for leaking munitions and in areas where munitions or bulk containers are being loaded and transported to processing buildings. After being unloaded from the transport container, the munitions and bulk containers are moved into the processing facility.

The processing facility is designed so that contaminated air within the facility is kept at a pressure below the outside pressure, preventing the leakage of contamination from inside to outside under normal operations. In the receiving area at the plant, the carrier containers are opened and the munitions or bulk containers of agent are moved into the plant. In the unpack area, pallets of munitions are opened and individual munitions are fed into the plant for processing. Demilitarization is conducted in campaigns during which only one type of agent is processed at a time. Agent monitors in the plant are calibrated only for the agent being processed in the current campaign (or, during the changeover periods between agent campaigns, for the previously processed agent and the agent about to be processed). Agent release from a misidentified munition containing a different agent therefore might not be detected by the monitors, although it might be detected by inspection.

Inside the processing area, ventilation systems are designed to keep agent-contaminated air inside the plant, and charcoal filters treat the exhaust air to prevent agent discharge into the atmosphere. However, in the unpack area and the munitions-handling area, agent-contaminated air is not usually present, and workers do not wear masks unless there is a STEL alarm. The STEL alarm is not instantaneous because of the nature of the current monitors—a measurement may take several minutes to register an alarm.

In processing facilities at incineration plants, agent is drained from munitions and containers and burned in a special furnace. This furnace is brought up to temperature using a conventional fuel before the agent feed is started. The combustion gases from the primary furnace chamber pass through a secondary “afterburner” chamber where additional mixing with atmospheric oxygen and lengthened



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 64
Monitoring at Chemical Agent Disposal Facilities 6 Opportunities for Improved Chemical Agent Monitoring The preceding chapters reviewed chemical agent monitoring requirements and challenges, the present systems in use at operating chemical stockpile demilitarization facilities, and the types and capabilities of available monitoring alternatives. It is clear that the present monitoring systems are functioning adequately to protect the health and welfare of workers as well as of the neighboring public. The monitoring systems are distributed throughout the demilitarization facilities and are integrated with plant operating and maintenance procedures, control systems, and systems for the early detection of malfunctions or leakages. Given that the disposal operations will be on-going for some years, at least until 2012 and perhaps beyond, this chapter focuses on the possible use of new, potentially real-time measurement technologies that are compatible with the existing monitoring systems or which provide supplemental capabilities that address some other needs. PRESENT MONITORING NEEDS AND CAPABILITIES The chemical agents stored and processed at stockpile sites were designed for use in warfare and therefore have the potential to cause fatal or chronic illnesses, depending on the type of agent, the concentration level, and the time of exposure. In combat, these agents are explosively deployed or sprayed to create aerosols that maximize airborne agent concentrations in target areas. In stockpile storage, the munitions are stored in igloos and are inspected periodically for leakage. Any munitions that are found to be leaking are overpacked to provide secure containment. Agent stored in bulk (ton containers and spray tanks) has no explosives or propellants present and may be stored in open yards or warehouses. When munitions or bulk containers are transported to a processing facility for disposal, they are generally loaded by workers wearing protective gear into a secure, explosion-resistant container for movement to the unpack area of the facility. Mobile, near-real-time (NRT) airborne agent monitors are used for the periodic checks of igloos for leaking munitions and in areas where munitions or bulk containers are being loaded and transported to processing buildings. After being unloaded from the transport container, the munitions and bulk containers are moved into the processing facility. The processing facility is designed so that contaminated air within the facility is kept at a pressure below the outside pressure, preventing the leakage of contamination from inside to outside under normal operations. In the receiving area at the plant, the carrier containers are opened and the munitions or bulk containers of agent are moved into the plant. In the unpack area, pallets of munitions are opened and individual munitions are fed into the plant for processing. Demilitarization is conducted in campaigns during which only one type of agent is processed at a time. Agent monitors in the plant are calibrated only for the agent being processed in the current campaign (or, during the changeover periods between agent campaigns, for the previously processed agent and the agent about to be processed). Agent release from a misidentified munition containing a different agent therefore might not be detected by the monitors, although it might be detected by inspection. Inside the processing area, ventilation systems are designed to keep agent-contaminated air inside the plant, and charcoal filters treat the exhaust air to prevent agent discharge into the atmosphere. However, in the unpack area and the munitions-handling area, agent-contaminated air is not usually present, and workers do not wear masks unless there is a STEL alarm. The STEL alarm is not instantaneous because of the nature of the current monitors—a measurement may take several minutes to register an alarm. In processing facilities at incineration plants, agent is drained from munitions and containers and burned in a special furnace. This furnace is brought up to temperature using a conventional fuel before the agent feed is started. The combustion gases from the primary furnace chamber pass through a secondary “afterburner” chamber where additional mixing with atmospheric oxygen and lengthened

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities exposure to high temperature ensure essentially complete combustion of the agent. From the afterburner, gases are further treated and cooled before being exhausted through the stack. In the three third-generation incinerator plants built after the TOCDF, treated and cooled exhaust gases are directed through an activated-carbon filter bed that removes any residual agent or other semivolatile pollutants prior to discharge through the stack. Air monitors in the stack continually test the exhaust gases for agent emissions at the source emission limit (SEL), which was formerly termed the allowable stack concentration (ASC). Although the terminology has changed, the values for emission limits have not. Furthermore, carbon monoxide emission levels in the stack are monitored as a means to check that the furnace combustion processes are operating within design limits. If the stack alarm sounds or if the furnace temperature or carbon monoxide levels are out of normal range, the feed of agent to the furnace is automatically stopped. Other furnaces in the facility that treat energetics, metals, and packing materials, along with some residual agent, also have exhaust gas treatment systems and discharge through the activated-carbon filter bed to the common plant stack. These precautions make it essentially impossible for large amounts of agent to be discharged from the stack. Even if trace amounts should briefly be emitted from the stack, they further disperse, so general population limits (GPLs) for exposure to agents would not be exceeded at the property line. Table 6-1 summarizes major airborne agent monitoring objectives during chemical weapons storage and demilitarization processing of the U.S. stockpile, and relates how the present monitoring systems (described in detail in Chapter 4) are used to assure that these objectives are met. A review of Table 6-1 confirms that most of the monitoring objectives are currently met satisfactorily. However, ACAMS, MINICAMS, and DAAMS all use the same type of gas chromatography-based separation and detection technology. There are multiple monitors that use the same measurement principles in the operational areas of the facilities to provide some redundancy, but there is no independent technology based on different measurement principles in place to pro- TABLE 6-1 Present Goals and Capabilities of Monitoring Systems Objective Monitoring Systems and Capabilities Assurance that non-stockpile workers and the general public are not exposed to low-level agent exposures that might harm their health over the lifetime of operations. Historic (DAAMS) monitors capable of measuring concentrations at or above the GPL are deployed at locations around the property line to provide assurance that concentrations of airborne agents at or above the GPL are not present. Assurance that airborne agent exposures to each worker do not exceed CDC guidelines for worker protection during his or her employment at the facility, whether while working in safe areas or in areas requiring use of appropriate protective gear. Near-real-time (ACAMS or MINICAMS) monitors, calibrated to detect any airborne agent concentrations at or above the STEL, are deployed throughout the facility and are set to alarm at a level at which workers must don appropriate protective gear. Similar NRT monitors are also used to monitor higher agent concentrations in contaminated plant areas. Assurance that the disposal facility is operating reliably and safely, and that any operational upsets or agent leaks are quickly addressed in a protective manner. ACAMS or MINICAMS along with DAAMS at each location are also used to identify any malfunctions leading to airborne agent leakage in the facility. Sometimes an alarm is a false positive. Usually the cause is a lubricant, pesticide, or other substance that also activates the agent detector in the monitor. False positives are treated as real until they are proven false by analysis of the DAAMS sample, so too many false positives can impede plant operation. At baseline plants, ACAMS in the plant common stack are set at the SEL. They alarm and shut down agent feed to the furnaces if an exceedance is detected. Assurance that extreme accident possibilities are controlled to the extent feasible, both in the plant and in the storage area, and that effective emergency response plans are coordinated with surrounding community response services. In addition to the many ACAMS or MINICAMS in the plant, portable units are used in the storage areas to inspect munitions in igloos or bulk storage areas. These instruments are also used whenever stockpile items are being moved to the processing area for demilitarization. Some of these portable monitoring units might also be useful in response to a major release. Assurance that, during closure operations, workers, the environment, and the public are protected as the plant is dismantled and contaminated materials are disposed of and, upon closure, the site is free from any agent residues that might jeopardize public or environmental well-being in the future. Additional NRT monitors are deployed for all agents processed at the demil facility. Additional specialized monitoring techniques are used in closure; these are beyond the scope of this report. NOTE: See Chapter 2 in this report for descriptions of monitoring levels. SOURCE: Committee on Monitoring at Chemical Agent Disposal Facilities, 2004.

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities vide timely confirmation. Analytical instruments based on different measurement principles generally respond quite differently to potential interferent species. Since most false-positive detections by the current monitoring systems are due to interferents, employing a second detection method, unlikely to be “fooled” by these interferents, can significantly enhance overall system reliability. Areas for potential improvements in airborne agent monitoring systems include the following: Multiagent monitoring capabilities. ACAMS are calibrated to detect one agent at a time. Although there is no evidence of an agent type within a munition having been misidentified in operations to date, there is a remote possibility that this might happen. Multiagent monitoring capability, even if at a much higher level than the STEL, might be of most use in the unpack area of the facility. It would also be useful during the closure operations of chemical demilitarization plants at multiagent stockpile sites when it is necessary to monitor simultaneously for all agents processed during plant operations (NRC, 2002b). Shorter time intervals to setting off an alarm or to confirming or denying an alarm. Shorter time intervals would be particularly desirable at the low-concentration monitoring levels required to meet recently revised CDC guidelines for agent airborne exposure limits. Rapid detection of agent releases within unpack areas might enhance worker safety, even at levels above the STEL. Better capabilities for detecting an incipient major accident and tracking the resulting agent release plume. Accidents so severe that they have potential for off-site impact are limited to scenarios in which large quantities of agent might be released in a short period of time. Site risk assessments identify the main causes and nature of such events, which are discussed later in this chapter as release scenarios. Depending on the scenario, air monitors might play a role in the early identification of such events and the notification of emergency services. By detecting or tracking the agent release plume, air monitors may also be useful in helping to identify effective emergency responses. In certain plant areas with the potential for serious accident events, air monitoring would be most useful around the higher levels, AEGL-1 or AEGL-2, which can be tolerated for an hour without irreversible health effects. If this monitoring used a real-time (nonchromatographic) technique, it would also provide both a prompt and an independent means of detecting, and, if deployed on a mobile platform, of tracking a major release. The following sections present a discussion of release scenarios that might lead to a significant agent release beyond the boundaries of the Army depots containing chemical weapons stockpile storage areas and demilitarization facilities and describe the potential utility of systems to detect and track such releases. On the basis of its analyses of the current airborne agent monitoring systems (see Chapter 4) and the monitoring requirements within and near the demilitarization facilities (summarized in Table 6-1), the committee concluded that the basic agent air monitoring system currently in place at stockpile storage and processing facilities seems capable of meeting the goals of protection for the public, workers, and the environment. Plant operability and worker safety might be further enhanced by improvements in sensitivity, selectivity, response time, and reliability of air monitors if such improvements are justified within the overall program goals. There may be a useful role for developing faster-warning higher-alarm-level monitors for certain critical work areas in order to provide more rapid warning to workers of a serious release of agent. Such monitors would not have to be as sensitive or selective as the existing monitors. If this higher-level-alarm monitor were to use a different technology from the existing gas chromatography-based instruments, it could provide an independent means for rapidly detecting or confirming a major release. Finding 6-1. The current airborne agent monitoring systems are adequate to safely protect the chemical demilitarization workforce, the public, and the environment, although potential incremental improvements that enhance sensitivity and specificity to reduce the rate of false-positive alarms and/or shorten cycle times might improve plant efficiency and safety. Recommendation 6-1. Continued incremental improvements in the current airborne chemical agent monitoring systems at chemical stockpile storage and demilitarization sites, as discussed in Chapter 4 of this report, should be pursued by the Army. Finding 6-2. The unpack area is an area in chemical demilitarization facilities that process multiple munitions in which enhanced monitoring that features faster alarm response and/ or multiagent capability might significantly enhance worker safety. An analysis of historic STEL alarms in such areas may indicate that worker protection could be enhanced by a more rapid alarm at a higher level, above the STEL but at or below the IDLH level or AEGL-1, that would allow faster masking of workers in the event of a large leak. Recommendation 6-2. The Army should analyze whether the addition of real-time and/or multiagent monitoring in the unpack area of chemical demilitarization facilities that process multiple munitions would significantly reduce risk to workers who unpack and stage munitions for processing. If the risk analysis indicates a significant enhancement of worker safety, the Army should investigate whether other,

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities shorter response time and/or multiagent deployment modes for current NRT monitors or the development and/or procurement of real-time, multiagent monitors based on innovative technology are feasible and practical. PROSPECTS FOR IMPROVED AMBIENT RELEASE MONITORING Fence-Line Monitoring for Community Protection The present systems for fence-line monitoring of chemical stockpile demilitarization facilities at the GPL appear adequate for providing a record that chronic agent emissions are not reaching the site property lines at levels that would involve a health threat to a member of the public, even for a lifetime of exposure. When false positives are eliminated, there has been only one documented indication that fence-line concentrations have ever exceeded the GPL level. In that case, a ton container of HD in the outdoor storage area at the Deseret Chemical Depot in Utah leaked about 80 gallons of agent, and a signal at the GPL level was later detected by plant perimeter monitors about a mile from the spill (NRC, 2002a). The spill had been detected and cleaned up by the time the small signal from the perimeter monitor was analyzed and confirmed.1 In the absence of other indications of actual agent emissions penetrating beyond a site perimeter, the use of additional perimeter air monitors capable of measuring at such low levels may yield no significant benefit. Detection closer to potential release points is much more reliable and faster, given the capabilities of existing monitoring equipment. Demilitarization Facility and/or Storage Area Monitoring Should a major agent release occur in a demilitarization plant or storage area, the fence-line monitors, as noted, are not a suitable means of warning, because they operate on long time scales and require laboratory analysis of collected samples to detect agent concentrations. At present, the airborne agent monitors in the plant and portable monitors in the storage area that are designed to protect workers are also used as the main way of detecting any abnormal agent release. However, the alarm levels for these instruments are at the STEL, the appropriate limit for managing worker risks. Such STEL alarms from monitors in the working areas of the facility, or from the portable monitors where work is being performed in the storage area, do not represent an incident large enough to reach a site property line unless the alarms are associated with an accident that causes a release creating much higher concentrations. In a storage area, portable monitors are only used when some operations are in progress. At other times, the storage area igloos are monitored for leaks only on a quarterly or annual basis, depending on the type of munitions being stored.2 Even though the integrity of the stockpile is good, there is some possibility of releases that might not be detected promptly (NRC, 2004). Emergency Response to Major Events In assessing the overall efficacy of current monitoring systems, it is necessary to look at accident possibilities that might cause extensive damage and perhaps impact the neighboring public. An examination of incidents that have occurred in the past at the operational chemical demilitarization facilities is one means to approach such an assessment. The NRC report Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities examines a series of agent-related incidents that have occurred at stockpile facilities (NRC, 2002a). While that report examines a variety of incidents presenting potential risk to workers and makes a number of recommendations to improve worker safety, there were no reported incidents in which enough airborne agent crossed a facility boundary line to cause public risk. Another way to identify possible accident scenarios is to hypothesize what might go wrong, how likely it is to happen, and what the resulting consequences might be under various weather conditions. To understand what sorts of accidents might have off-site impacts, it is useful to review the site-specific quantitative risk assessments (QRAs) prepared for the Army (U.S. Army, 2002a, 2003b, 2003c). These assessments include possible causes and estimates of the likelihood of potential major accidents, as well as estimates of the extent of the consequences for each scenario. Consequence models consider how much agent is released, release conditions, and the subsequent gradual dispersal of the plume through dilution by atmospheric mixing as the plume is carried downwind. Wind direction, speed, and atmospheric stability levels also determine rates of dispersal and the hazard zone for the plume. The QRAs for the combined storage and disposal facilities at Anniston, Alabama; Tooele, Utah; and Umatilla, Oregon (ANCDF, TOCDF, and UMCDF) are the basis for the risk estimates shown in Table 6-2 for storage and processing during the estimated weapons-processing period (U.S. Army, 2002a, 2003b, 2003c). The numbers shown are a measure of overall public risk. It is evident that the risks of storage, even during the processing period, significantly exceed the risks 1   Personal communication from Cheryl Maggio, Senior Project Engineer, CMA, to the committee at its November 22, 2004, meeting. 2   Igloos containing nonoverpacked mines, projectiles, or bombs are required to be inspected for leaks on an annual basis according to the Army’s storage monitoring and inspection (SMI) program. All other igloos, such as those containing M55 rockets and overpacked munitions, are inspected quarterly (NRC, 2004).

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities TABLE 6-2 QRA Public Risk Estimates for Three Sites Site Duration of Processing Period Chance of Having One or More Public Fatalities Over Processing Period Duration Storage Processing Anniston, Alabama (ANCDF) 7 years 1 in 750 1 in 1,800 Umatilla, Oregon (UMCDF) 6 years 1 in 2,100 1 in 3,300 Tooele, Utah (TOCDF) 4 years 1 in 10,000 1 in 1,100,000 NOTE: The risk estimates for TOCDF have been updated to reflect that all GB weapons have already been processed there; consequently, the overall level of remaining risk is much less than the level of risks at ANCDF and UMCDF. Risk estimates are for accidental agent releases and do not include releases due to sabotage or terrorist activities. (See discussion in text.) SOURCE: U.S. Army, 2002a, 2003b, 2003c. of processing. This is reasonable because of the smaller quantities in the processing area at any time, and because of the major layers of protection built in to the processing plant to contain any internal releases. The Army has recently asked Science Applications International Corporation (SAIC) to use its QRA capabilities to establish “design basis” accident scenarios for the stockpile facilities (U.S. Army, 2003a). Since each site is different because of its location, weather, and surrounding populations, the scenarios were adapted to site-specific conditions. Table 6-3 shows recommended design basis accident source terms (that is, how much agent must be vaporized in the facility or storage area to cause a significant health risk to the nearest population). The likelihood of each design basis accident is in the range of 1 in 1,000 to 1 in 10,000 years; these are unlikely to occur during the lifetime of the plant but are not beyond credibility. The SAIC report (U.S. Army, TABLE 6-3 Airborne Source Terms for Stockpile Sites from Design Basis Accident Scenarios Site Evaluation Distance (km) GB Vapor (kg) VX Vapor (kg) HD Vapor (kg) Tooele, Utah 3.4 — 2 100 Anniston, Alabama 3.8 20 1 50 Umatilla, Oregon 2.9 30 2 90 Pine Bluff, Arkansas 0.9 5 0.3 20 Blue Grass, Kentucky 1.8 10 1 40 Aberdeen, Maryland 2.0 — — 40 Pueblo, Colorado 2.5 — — 90 Newport, Indiana 1.1 — 1 —   SOURCE: U.S. Army, 2003a. TABLE 6-4 Airborne Exposure Limits (AELs) (2005 values) and Vapor Pressure of Agents   Concentration Level of Agent (mg/m3) GB VX HD GPL 1 × 10−6 6 × 10−7 2 × 10−5 STEL 1 × 10−4 1 × 10−5 3 × 10−3 Stack SEL 3 × 10−4 3 × 10−4 3 × 10−2 1-hr AEGL-2 3.5 × 10−2 2.9 × 10−3 1 × 10−1 IDLH limit 1 × 10−1 3 × 10−3 7 × 10−1 Saturated vapor 25°C 2.2 × 104 1 × 101 7 × 102 Saturated vapor/IDLH limit 220,000 3,000 1,000   SOURCE: Based on data provided to the NRC Committee to Assess Designs for the Pueblo and Blue Grass Chemical Agent Destruction Pilot Plants, and data from Abercrombie, 2003. 2003a) also develops design basis release source terms for workers within a 600 meter radius; the worker source terms generally are around 1 to 10 times higher than those for the public. Note that the design basis release probabilities were developed only for accidental agent releases; the probabilities of deliberate releases triggered by sabotage or terrorist attack are not included. POTENTIAL MAJOR RELEASE SCENARIOS In this section, the information presented earlier is used to consider several spill scenarios, to examine how current monitoring would respond, and to identify what additional monitoring might be useful. Table 6-4 summarizes information (discussed in detail in Chapter 2) on agent exposure limits and vapor pressure parameters for GB, VX, and HD. This information is important in assessing the consequences of different sorts of releases. In risk assessment studies, the focus is on accidents that have the potential to cause fatalities among the public. In assessments of risk, the Army sometimes uses a dose criterion (concentration multiplied by exposure time) that would cause 1 percent lethality in exposed populations.3 Table 6-5 presents the 1 percent lethality dose for the agents of interest and also shows the exposure time at the IDLH limit that would give the 1 percent lethality dose. Scenario 1: Liquid Agent Spills Simple spills of agent generate vapor by evaporation from the liquid pool. The vapor pressure is important because it 3   The Army more recently is using the AEGL-1 and AEGL-2 values in its emergency response planning activities.

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities TABLE 6-5 One Percent Lethality Doses for Agents and Exposure Times at IDLH Limit Agent Type 1% Lethality Dose (mg-min/m3) Time (min) at IDLH to Give 1% Lethality Dose GB 10 100 VX 4.3 1,400 HD 150a 200a aLevel corresponding to maximum exposure with no permanent skin injury, rather than 1 percent lethality. SOURCE: NRC, 1984. represents the maximum concentration of agent vapor in the air above a surface of liquid agent. Both HD and VX have relatively low vapor pressures at ambient temperature. GB has higher volatility, similar to that of water, but is less toxic than VX. The last row of Table 6-4 shows the ratio of the vapor pressure to the IDLH AEL for the relevant chemical agent. Thus, the higher vapor pressure of GB has consequences in terms of the extent of the hazard zone that can result from a spill. For example, if the contents of ton containers of HD or VX spill, the Army has estimated radial hazard zones for workers of about 80 feet or less (NRC, 1984). For a similar release of GB, one would expect a higher hazard zone because of its higher volatility—perhaps by a factor of about 10, since its ratio of vapor pressure to IDLH limit is about 100 times greater.4 An estimate for GB is also described in an early NRC report (NRC, 1984, p. 37): As an example of a calculation, the panel asked the personnel at Umatilla to calculate the 1% lethality distance for a large spill of GB resulting from the rupture of a 750 pound bomb. The assumptions of the calculations were as follows: 70°F temperature, 3 m/s wind speed, stability class F, 25% of the agent in the bomb spilled, and 30 minutes elapsed until the spill was covered. (Umatilla personnel noted that a plastic sheet is kept in readiness when such munitions are handled and that 10 minutes is more realistic as the time needed to cover the spill.) The resulting 1% lethality distance was computed to be 200 meters (about 600 feet). The Army procedure is to rapidly cover any liquid agent spill with emergency tarps to further reduce vaporization. The spills are then treated with decontamination solutions to destroy the agent. Mobile air monitoring units are available to identify concentrations in the work areas around the spill so that workers are protected from exposures adverse to health. In the QRAs for ANCDF, TOCDF, and UMCDF, simple spills of liquid appear significant only with respect to worker risk, since the vaporization rates from spills produce concentrations of airborne agents that will be highly diluted before they can reach a location where the public could be exposed. Scenario 2: Explosive Dispersion of Agent The more hazardous events associated with the stockpile are those that involve explosive dispersion of agent. In early studies of stockpile risks, a “maximum credible event” was taken to be an explosion of an igloo full of M55 GB rockets. The igloos are spaced so that explosions in one igloo are very unlikely to set off an explosion in an adjacent igloo. Igloos with M55 rockets currently exist at four of the stockpile sites. The Army calculated 1 percent lethality distances for an explosion of an igloo containing GB M55 rockets. It was assumed that about 97 percent of the agent was consumed in the explosion and subsequent fire, based on some evidence from tests conducted at the Black Hills Army Depot in South Dakota and the Dugway Proving Ground in Utah. Those tests also showed that for M55 rockets, sympathetic detonations could spread throughout rockets stored in the same igloo. The Army also found that the explosion of one land mine in a three-mine container might detonate the other two mines, but that sympathetic detonation in other containers of munitions does not occur.5 Using the explosion of an igloo of GB M55 rockets as an extreme accident scenario, the Army computed downwind dispersion distances for 1 percent lethality and found that hazard zones extended about 5 miles from the source under typical daytime conditions and up to about 27 miles at night, when the height of the mixing layer is much lower (Irving et al., 1970; Lloyd, 1994). Other major scenarios included detonation of a 155 mm GB round, creating 1 percent lethality at a distance of 3,000 feet, and detonation of a 155 mm HD round, creating irreversible skin damage up to 260 feet. Therefore, the risks to the public from the stockpile are associated with major explosions that might be caused by accidental detonation of an M55 rocket or rockets during handling or by detonation as a result of some natural disaster (e.g., lightning strike, earthquake), a plane crash, or sabotage. For the Tooele facility, for example, the site risk assessment shows that over 82 percent of all the average public fatality risk from the storage facility is initiated by seismic events, with 11 percent from a GB container leak6 and 4 percent 4   As vapor disperses, air mixes in from the side and top of the plume, so concentrations drop more than linearly with downwind distance. 5   Reference material sent by Marilyn Daughdrill (then Tischbin), Chemical Materials Agency, to Don Siebenaler, NRC Senior Program Officer, May 9, 1997, described as information to accompany an Army videotape, Project ID 7VA-1334-0004-SP86. 6   All GB ton containers have now been demilitarized at TOCDF. There are no more GB ton containers in the stockpile.

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities from lightning (NRC, 1997). At Umatilla, 97 percent of the storage risk to the public is from seismic events (U.S. Army, 2002b). For Anniston, public risks from storage are dominated by lightning strikes (84 percent), seismic events (9 percent), and errant rounds from the Pelham firing range (6 percent) (U.S. Army, 2002c). While air monitors are used to check for leakages in the stockpile, it is unlikely that air monitors would be of relevance to improved detection of such an extreme igloo detonation scenario. Portable or mobile units might be of use in detecting and tracking downwind concentrations of agent in the aftermath of a major release. If the Army has not already done so, it should evaluate how useful its current monitoring equipment might be in such an accident response mode and in conjunction with the response plans of local emergency services. While fixed-site NRT monitors might provide useful indications of agent-contaminated areas, the actual tracking of a dispersing plume requires real-time monitors deployed on mobile ground or air vehicles. Scenario 3: Major Releases in a Processing Facility In the original QRA for TOCDF, over 97 percent of the public risk from processing was due to seismic events that caused collapse of major portions of the processing facility (U.S. Army, 1996). In the more recent QRAs for TOCDF (updated after the completion of GB campaigns), ANCDF, and UMCDF, the leading cause of public risk (around 90 percent) comes from a number of fire-initiated major accidents that spread within the facility. Some of these events included fire damage to ventilation systems and the carbon filter beds. In both fire and natural disaster scenarios, the usefulness of air monitoring equipment is likely to be compromised. At other sites where a QRA is not currently available, the causes of risk may shift to other sorts of serious initiating events. Nonetheless, rapid reporting of fires within a plant needs to be a major part of any emergency response program on site and should include off-site emergency services. Here again, the availability of portable real-time or near-real-time or mobile real-time monitoring equipment might be useful in the aftermath of a release. Also, since fire is a major contributor to risk in both disposal and storage facilities, fire detection, alarm, and suppression equipment are additional elements of protection, and fire prevention activities may also help prevent significant agent release events. Summary of Major-Release Scenario Analyses To pose an acute risk to the public, the atmospheric release of sufficient chemical agent vapor or aerosol would require a major accident. A careful review of portable air monitoring equipment now used to protect storage area workers should be conducted to evaluate the role that such equipment might play in responding to major disasters with the potential for off-site impact. The development of innovative, dispersed portable or mobile real-time agent monitors operating between the STEL and IDLH levels would allow rapid detection and, potentially, the tracking of a significant release plume. The development of such capabilities might be coordinated with the Department of Homeland Security, which may have similar chemical agent monitoring requirements. Finding 6-3. To pose an acute risk to the public, the atmospheric release of sufficient chemical agent vapor or aerosol would require a major accident, almost certainly involving explosion and/or fire. The ability to confirm dispersion model predictions that an agent plume has penetrated the depot boundary and threatens the public or to track the agent plume would require fast-response monitors operating at levels between the STEL and the IDLH that are either widely dispersed or are mounted on a suitable ground or air mobile platform. Recommendation 6-3. The Army and other relevant stakeholders should assess whether public protection would be significantly enhanced by the development and deployment of dispersed fixed or portable fast-response agent sensors or the development of a mobile fast-response agent sensor platform capable of detecting and tracking a large release plume. Applicability of Identified Advanced Monitoring Technologies for Agent Release Scenarios As stated in Chapter 5 and noted in the release scenarios analyses presented in this chapter, the primary reasons to consider using advanced agent monitoring technologies are to shorten measurement response times significantly without unduly sacrificing either sensitivity or specificity, and to deploy these instruments where rapid detection or confirmation of agent would provide significant protection to workers or the public. Two potential venues for enhanced agent monitoring have been identified by the committee: (1) faster response and, possibly, multiagent monitoring in the unpack areas of the demilitarization facilities; and (2) detection and tracking of a major-agent-release plume at agent concentration levels somewhere above the STEL but below the IDHL. Of the advanced agent monitoring technologies evaluated in Chapter 5, only two appear to offer multiagent capability, real-time detection, and sufficient sensitivity and specificity for relevant chemical agents coupled with a significant demonstrated utility within the air quality/atmospheric chemistry community to warrant consideration for possible deployment at chemical disposal and storage sites. These are open-path or folded-path Fourier transform infrared spectroscopy and chemical ionization mass spectrometry. As demonstrated in Chapter 5, in order to have fast enough response times and sufficient sensitivity for the relevant chemical agents, Fourier transform infrared systems must be deployed with either long open-path mirror systems

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities or a fairly large multipass gas cell into which high-volume air samples are drawn from multiple sampling lines. Note, however, that if contaminated air is only drawn from one of these sampling lines, as may be expected to be the case in the early stages of a release, the air in the cell is diluted by air drawn from the other lines, significantly increasing the limit of detection. In general, these systems will be too large for mobile deployment, and a single system can cover only a limited area, even with multiple open-path target mirrors or sampling lines into a multipass gas cell. Its utility for large-plume-release detection and quantification seems limited, since FT-IR systems would need to be deployed relatively near potential release points to avoid using a large number of units, although one organization has promoted this application.7 An open- or folded-path FT-IR system might be of utility as a fast-response detector in a demilitarization facility unpack area, an application briefly explored with some promise by the Army nearly a decade ago (Stedman and McLaren, 1996). As noted above, an unpack area FT-IR monitoring system could be deployed either in an open-path configuration, above the level where people or equipment would routinely block the infrared beam, or in a closed, multipath cell configuration with multiple sampling lines. Since the molecular weight of agent molecules is much greater than that of oxygen or nitrogen, they would tend to have higher concentration at ground level. Thus, their concentration at ground level may be dangerously underestimated by an elevated beam. As demonstrated in Chapter 5, chemical ionization mass spectrometry (CIMS) is a far more sensitive analytical technique than is FT-IR spectroscopy for most semivolatile pollutants. CIMS is a point sampling analytical technique, although its very high sensitivity and rapid measurement capability can allow it to be coupled to multiple sampling lines or mounted on a mobile platform to provide significant spatial coverage quickly and accurately. An unpack area deployment would presumably monitor a number of sampling lines sequentially. A release plume detection and tracking deployment would most likely involve a mobile platform. Finding 6-4. Open- or folded-path FT-IR and CIMS technology have some promise for providing enhanced, fast-response chemical agent monitoring capability to chemical weapons storage and demilitarization facilities. The most likely effective use for FT-IR spectroscopy is to provide fast-response, multiagent monitoring for a relatively restricted space such as a demilitarization facility’s unpack area. CIMS instruments are likely to be far more sensitive chemical agent detectors than are FT-IR instruments. Potential CIMS applications include monitoring a restricted space, such as a demilitarization facility’s unpack area, through multiple sampling lines, and detecting and tracking a large-release plume mounted onboard a mobile van or small aircraft. TECHNICAL AND COMMERCIAL MATURITY OF PROSPECTIVE MONITORING TECHNOLOGIES Potential supplemental chemical agent monitoring roles at chemical demilitarization facilities have been identified for two specific innovative real-time monitoring technologies: open-path Fourier transform infrared spectroscopy and chemical ionization mass spectrometry. In order to significantly impact the chemical demilitarization process, commercial versions of these instruments capable of sustained and effective operation in an industrial environment need to be available within 2 to 3 years. The commercial status of these two airborne chemical agent monitoring technologies is briefly reviewed below. Open-Path Fourier Transform Infrared Spectroscopy OP/FT-IR spectroscopy is a relatively mature measurement technology that has been recognized by the U.S. Environmental Protection Agency as a reference method for airborne pollutant quantification. As early as 1997, an open-path FT-IR product review identified and evaluated commercial systems from five vendors (Newman, 1997). FT-IR systems are now widely used in both industrial process and emissions monitoring and in environmental field measurements. Ongoing improvements in optical, electronic, and computer technologies are prompting the continuous development of compact, rugged, and capable FT-IR systems, many of which can be deployed in open-path configurations. Several commercial FT-IR vendors and custom-instrument companies are capable of producing a version customized for chemical agent detection within a year; although it is unlikely that these instruments would have practical detection limits much below 1 ppb for any of the relevant chemical agents (see Box 5-1 in Chapter 5). Chemical Ionization Mass Spectrometry As detailed in Chapter 5, CIMS instruments must be tailored for specific target molecules by demonstrating that the precursor ion species effectively ionizes the target agent, producing a unique product ion that is not compromised by large signals at the same m/z values produced from common potential interferents. While CIMS work in Army laboratories showed that various proton transfer reactive ions were efficient chemical agent ionizers, the more recently developed negative ion precursor systems, which produce extremely high sensitivities for many atmospheric pollutants, have not been tested with chemical agents. As noted in 7   Optical Remote Sensing to Detect and Map Low Levels of Chemical Threat, presentation by Ram A. Hashmonay, Arcadis, to the Technologies for Chemical Agent Detection Workshop, Washington, D.C., August 24, 2004.

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities Chapter 5, computational and experimental studies to determine the extent which various negative precursor ions could improve real-time chemical agent detection levels could be quickly performed. There is one commercial CIMS system currently available which uses H3O+ as the precursor ion. It is produced and marketed by Ionicon Analytik in Austria. This instrument is widely used in airborne field measurements to map airborne organic pollutants without significant modification, and a moderately ruggedized version has been successfully deployed in a van-mounted vehicle laboratory to map ground-level organic pollutants and identify organic emissions plumes (Kolb et al., 2004). Based on the experience of committee members in developing and commercializing mass spectrometric instruments of similar complexity, the committee estimates that a commercial CIMS instrument optimized for GB, VX, and HD detection could be customized and produced for sale in moderate quantities within 2 years. ABILITY TO MEET REGULATORY REQUIREMENTS There are no U.S. federal or state environmental regulations that set emissions standards for chemical agents or specify air monitoring requirements. For instance, since chemical agents are not designated by the EPA as either criteria pollutants or controlled air toxic species, maximum achievable control technology (MACT) rules do not apply to their emission control or emission monitoring technologies. Rather, these agent emissions are regulated under the RCRA permits that are issued for each site by the appropriate state environmental authority. As the Army adapts to the new AELs recommended by the CDC, it has modified its operational monitoring procedures throughout the chemical demilitarization program. The Army is now monitoring to the STEL (which is numerically equivalent to the former WPL/TWA) as the alarm point at which workers mask and operations stop until the source of the alarm is identified. Further, measurements are made at fractions of the STEL that are closer to the new WPLs for informational purposes. The new WPLs are monitored using the DAAMS equipment, which does not provide near-real-time information. Health records are maintained for all worker exposures that are significant so that longer-term health impacts from multiple exposures are controlled. The implementation of the new AELs basically maintains the levels of protection that had been provided to workers prior to the revisions to the AELs promulgated by the CDC. However, the new nomenclature is not consistent with that found in existing RCRA permits at the sites. Therefore, the Army is seeking the necessary permit modifications at each of the sites and is updating its own procedures and documents accordingly. The meetings with the regulators are ongoing as of this writing and are progressing satisfactorily. In situations in which the new AELs are enacted before the permit issues have been satisfactorily resolved, the Army will request temporary authorization to operate in the interim. Permits for sites not yet ready for operation will be based on the new requirements. SUMMARY OF OPPORTUNITIES FOR IMPROVED AIRBORNE AGENT MONITORING The issue of whether or not to enhance the current airborne chemical agent monitoring systems at chemical weapons storage and demilitarization facilities is complex. The current monitoring systems are generally adequate and well integrated into facility operations. The addition of new technology would impose new costs, both for hardware and for additional staff, as well as added complexity to systems that are already difficult to manage and maintain effectively. Thus, the addition of any advanced monitoring capabilities need to provide benefits that are greater than the monetary and operational challenges they will impose. In addition, new technology must be developed, demonstrated, acquired, and integrated into operations in a relatively short time if its benefits are to be exploited. Chemical demilitarization facilities have already destroyed over a third of the original stockpile and are not likely to operate more than 7 to 12 years longer. In the committee’s opinion, this means that new monitoring technology needs to be developed and convincingly demonstrated within 2 to 3 years and acquired and integrated within 4 years if it is to have a substantial impact. Recommendation 6-4. The Army should only deploy advanced chemical agent monitoring equipment after a thorough risk/benefit analysis shows that the risk reduction to the workforce and/or public justifies the monetary and opportunity costs. Recommendation 6-5. If worker or public risk reduction analyses indicate significant benefit at acceptable cost from deployment of fast-response, multiagent monitoring capabilities, systems using FT-IR or, more likely, CIMS should be considered. REFERENCES Abercrombie, P.L. 2003. Final Physical Property Data Review of Selected Chemical Agents and Related Compounds. Updated Field Manual 3-9, ECBC-TR-294, September. Aberdeen Proving Ground, Md.: Chemical Materials Agency. Irving, S., M.D. Shavit, D.D. Halvey, and O.F. Haase, Jr. 1970. Methods for Estimating Hazard Distances for Accidents Involving Chemical Agents. Operations Research Group Report No. 40, February. Edgewood Arsenal, Md.: Operations Research Group. Kolb, C.E., S.C. Herndon, J.B. McManus, J.H. Shorter, M.S. Zahniser, D.D. Nelson, J.T. Jayne, M.R. Canagaratna, and D.R. Worsnop. 2004. Mobile laboratory with rapid response instruments for real-time measurements of urban and regional trace gas and particulate distributions and emission source characteristics. Environmental Science and Technology 38(21): 5694−5703.

OCR for page 64
Monitoring at Chemical Agent Disposal Facilities Lloyd, J.D. 1994. Memo from Director, U.S. Army DARCOM Field Safety Activity, Charlestown, Indiana, to LTC Leideritz, “Maximum Credible Events (MCE’s) and 1% Lethality Distances for DARCOM Chemical Agent Storage Sites,” March 2. Newman, A.R. 1997. Open-path FT-IR takes the long view. Analytical Chemistry A 69(1): 43A−47A. NRC (National Research Council). 1984. Disposal of Chemical Munitions and Agents. Washington, D.C.: National Academy Press. NRC. 1997. Risk Assessment and Management at Deseret Chemical Depot and the Tooele Chemical Agent Disposal Facility. Washington, D.C.: National Academy Press. NRC. 2002a. Evaluation of Chemical Events at Army Chemical Agent Disposal Facilities. Washington, D.C.: The National Academies Press. NRC. 2002b. Closure and Johnston Atoll Chemical Agent Disposal System. Washington, D.C.: The National Academies Press. NRC. 2004. Effects of Degraded Agent and Munitions on Chemical Stockpile Disposal Operations. Washington, D.C.: The National Academies Press. Stedman, D.H., and S.E. McLaren. 1996. Detection of Chemical Agents by Open-Path FTIR Spectroscopy. Final Report for MDM Test Area Project, Final Report for U.S. Army Contract DAAM01-94-C-0068. U.S. Army. 1996. Tooele Chemical Agent Disposal Facility Quantitative Risk Assessment. Summary Report. Aberdeen Proving Ground, Md.: Program Manager for Chemical Demilitarization. U.S. Army. 2002a. Anniston Chemical Agent Disposal Facility Quantitative Risk Assessment. Summary Report. June. Aberdeen Proving Ground, Md.: Program Manager for Chemical Demilitarization. U.S. Army. 2002b. Umatilla Chemical Agent Disposal Facility Quantitative Risk Assessment, Volume I. December. Aberdeen Proving Ground, Md.: Program Manager for Chemical Demilitarization. U.S. Army. 2002c. Anniston Chemical Agent Disposal Facility Quantitative Risk Assessment. Revision 0, June. Aberdeen Proving Ground, Md.: Program Manager for Chemical Demilitarization. U.S. Army. 2003a. Evaluation Guidelines for Design Basis Accident Analysis Associated with Establishing a Safety Basis for PM ECW Facilities. RM-03-003 Revision 3, July 29. Aberdeen Proving Ground, Md.: Program Manager for Elimination of Chemical Weapons Risk Management and Quality Assurance Office. U.S. Army. 2003b. Tooele Chemical Agent Disposal Facility Quantitative Risk Assessment Results for VX and HD Disposal Processing. Summary Report, February. Aberdeen Proving Ground, Md.: Program Manager for Chemical Demilitarization. U.S. Army. 2003c. Umatilla Chemical Agent Disposal Facility Quantitative Risk Assessment. Summary Report, September. Aberdeen Proving Ground, Md.: Program Manager for Elimination of Chemical Weapons.