Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
A-1 APPENDIX A Chemical Threat Information A.1 POSSIBLE TYPES OF AGENTS There is a wide range of chemicals that are potentially attractive to terrorists, whether at a fixed site or during trans- portation. There are numerous sources that can be consulted for lists of such chemicals, among them being the following: â¢ The EPA Risk Management Program (RMP) 40 CFR Part 68 lists 77 regulated toxic chemicals. See www.epa.gov/ ceppo. â¢ The OSHA Process Safety Management (PSM) Standard 29 CFR 1910.119 provides a list of 136 toxic and highly reactive hazardous chemicals. See (www.osha.gov). â¢ The FBI Community Outreach Program for Manufactur- ers and Suppliers of Chemical and Biological Agents, Materials, and Equipment has a table of 42 industrial chemical materials and agents that âmay be more likely to be used in furtherance of WMD terrorism.â See (http:// www.aiche.org/ccps/pdf/fbi_wmd.pdf) â¢ The Australia Group list of chemical and biological weapons provides a list of 54 chemical weapons pre- cursors. See www.australiagroup.net â¢ The Chemical Weapons Convention (CWC) provides three toxic chemicals and 11 precursors under Sched- ule 2 and 17 toxic chemicals under Schedule 3. See www.cwc.gov. â¢ The Department of Transportationâs (DOTâs) 2004 Emergency Response Guidebook contains guides to responding to dangerous goods/hazardous materials incidents. Each guide is designed to cover a group of materials that have similar chemical and toxicologi- cal properties. This may be obtained online through http://hazmat.dot.gov/pubs/erg/gydebook.htm. The above sources list at most a few hundred chemicals. In actuality there may be âliterally tens of thousands of poiso- nous chemicalsâ that might be of some use to terrorists (Kup- perman and Kamen 1989). Mullen (1978) cites an estimate of âwell over 50,000â for the number different organophosphate (of which sarin is one) alone. Purver (1995) provides a list of chemical agents specifically mentioned in the then-current lit- erature on terrorism: insecticides such as nicotine sulfate, DFP (diisopropylphosphorofluoridate), parathion, and TEPP; herbicides such as 2,4D and 2,4, 5T (against plants), TCDD (dioxin), and benzidine (112-14); âblood agentsâ such as hydrogen cyanide and cyanogen chloride; âchoking agentsâ such as chlorine, phosgene (carbonyl chloride), and chloropi- crin; âblistering agentsâ such as sulfur mustard, nitrogen mus- tard, and lewisite; and ânerve agentsâ such as tabun, sarin, VX, and soman. Other chemicals mentioned include: Prussic acid (hydrocyanic acid), lysergic acid (LSD), aminazin, pheromones, pure nicotine, phosgene oxime (CX), arsenic, Cobalt-60, compound 1080 (sodium fluoroacetate), arsine, nickel carbonyl, and strychnine. Toxic chemicals can be cat- egorized as shown below based upon Cordesman, 1996. NERVE AGENTS: Agents that quickly disrupt the nervous system by binding to enzymes critical to nerve functions, causing convulsions and/or paralysis. Must be ingested, inhaled, and absorbed through the skin. Very low doses cause a running nose, contraction of the pupil of the eye, and diffi- culty in visual coordination. Moderate doses constrict the bronchi, cause a feeling of pressure in the chest, and weaken the skeletal muscles and cause fibrillation. Large doses cause death by respiratory or heart failure. (Can be absorbed through inhalation or skin contact.) Reaction normally occurs in 1â2 minutes. Death from lethal doses occurs within minutes, but artificial respiration can help and atropine and the oximes act as antidotes. The most toxic nerve agents kill with a dosage of only 10 milligram-minutes per cubic meter, versus 400 for less lethal gases. Recovery is normally quick, if it occurs at all, but permanent brain damage can occur. Examples of nerve agents are: â¢ Tabun (GA) â¢ Sarin (GB)ânearly as volatile as water and delivered by air. A dose of 5 mg-min/m3 produces casualties, a respi- ratory dose of 100 mg-min/m3 is lethal. Lethality lasts 1â2 days. â¢ Soman (GD) â¢ GF â¢ VR-55 (Improved Soman) A thick oily substance which persists for some time. â¢ VK/VX/VE/VM/VG/VSâPersistent agents are roughly as heavy as fuel oil. A dose of 0.5 mg-min/m3 produces casualties, a respiratory dose of 10 mg-min/m3 is lethal. Lethality lasts 1â16 weeks. BLISTER AGENTS: Cell poisons that destroy skin and tissue, cause blindness upon contact with the eyes, and which can result in fatal respiratory damage. Can be col- orless or black oily droplets. Can be absorbed through inhalation or skin contact. Serious internal damage if inhaled. Penetrates ordinary clothing. Some have delayed and some have immediate action. Actual blistering nor- mally takes hours to days, but effects on the eyes are much more rapid. Mustard gas is a typical blister agent and expo- sure to concentrations of a few milligrams per cubic meter
over several hours generally causes blisters and swollen eyes. When the liquid falls onto the skin or eyes it has the effect of second or third degree burns. It can blind and cause damage to the lungs leading to pneumonia. Severe exposure causes general intoxication similar to radiation sickness. HD and HN persist up to 12 hours. L, HL, and CX persist for 1â2 hours. Short of preventing exposure, the only treatment is to wash the eyes, decontaminate the skin, and treat the resulting damage like burns. Examples of blister agents are: â¢ Sulfur Mustard (H or HD): A dose of 100 mg-min/m3 produces casualties, a dose of 1,500 mg-min/m3 is lethal. Residual lethality lasts up to 2â8 weeks. â¢ Distilled Mustard (DM) â¢ Nitrogen Mustard (HN) â¢ Lewisite (L) â¢ Phosgene Oxime (CX) â¢ Mustard Lewisite (HL) CHOKING AGENTS: Agents that cause the blood vessels in the lungs to hemorrhage, and fluid to build-up, until the victim chokes or drowns in his or her own fluids (pulmonary edema). Provide quick warning though smell or lung irrita- tion. Can be absorbed through inhalation. Immediate to delayed action. The only treatment is inhalation of oxygen and rest. Symptoms emerge in periods of seconds up to three hours after exposure. Examples of choking agents are: â¢ Phosgene (CG) â¢ Diphosgene (DP) â¢ PS Chloropicrin â¢ Chlorine Gas BLOOD AGENTS: Kill through inhalation. Provide little warning except for headache., nausea, and vertigo. Interferes with use of oxygen at the cellular level. CK also irritates the lungs and eyes. Rapid action and exposure either kills by inhibiting cell respiration or it does notâcasualties will either die within seconds to minutes of exposure or recover in fresh air. Most gas masks have severe problems in pro- viding effective protection against blood agents, examples of which are: â¢ Hydrogen Cyanide (AC)âa dose of 2,000 mg-min/m3 produces casualties. A dose of 5,000 mg-min/m3 is lethal. Lethality lasts 1â4 hours. â¢ Cyanogen Chloride (CK)âa dose of 7,000 mg-min/m3 produces casualties. A dose of 11,000 mg-min/m3 is lethal. Lethality lasts 15 minutes to one hour. BIOLOGICAL TOXINS : Biological poisons causing neu- romuscular paralysis hours or days after exposure. Formed in food or cultures by the bacterium Clostridium botulinum. A-2 Produce highly fatal poisoning characterized by general weakness, headache, dizziness, double vision and dilation of the pupils, paralysis of muscles, and problems in speech. Death is usually by respiratory failure. Antitoxin therapy has limited value, but treatment is mainly supportive. An exam- ple is Botulin toxin (A), of which there are six distinct types Four of these are known to be fatal to man. An oral dose of 0.001 mg is lethal. A respiratory dose of 0.02 mg-min/m3 is also lethal. DEVELOPMENTAL WEAPONS: A new generation of chemical weapons is under development (this statement writ- ten in 1996). The only publicized agent is perfluoroisobutene (PFIB), which is an extremely toxic, odorless, and invisible substance produced when PFIB (Teflon) is subjected to extreme heat under special conditions. It causes pulmonary edema or dry-land drowning when the lungs fill with fluid. Short exposure disables and small concentrations cause delayed death. Activated charcoal and most existing protec- tion equipment offer no defense. Some sources refer to âthirdâ and âfourthâ generation nerve gases, but no technical literature seems to be available. RIOT CONTROL AGENTS: Agents that produce tempo- rary irritating or disabling effects. They cause flow of tears and irritation of upper respiratory tract and skin when in con- tact with the eyes or inhaled. They can cause nausea and vom- iting: can cause serious illness or death when used in confined spaces. CS is the least toxic gas, followed by CN and DM. Symptoms can be treated by washing the eyes and/or removal from the area. Exposure to CS, CN, and DM produces imme- diate symptoms. Staphylococcus produces symptoms in 30 minutes to four hours, and recovery takes 24â48 hours. Treat- ment of Staphylococcus is largely supportive: Tear Gases: â¢ Chlororacetophenone (CN) â¢ O-Chlorobenzyl-malononitrile (CS) Vomiting Gases: (cause irritation, coughing, severe headache, tightness in chest, nausea, and vomiting): â¢ Adamsite (DM) â¢ Staphylococcus INCAPACITATING AGENTS: Agents that normally cause short term illness and psychoactive effects (delirium and hal- lucinations). Can be absorbed through inhalation or skin contact. The psychoactive gases and drugs produce unpre- dictable effects, particularly in the sick, small children, elderly, and individuals who already are mentally ill. In rare cases they kill. In others, they produce a permanent psychotic condition. Many produce dry skin, irregular heart beat, uri- nary retention, constipation, drowsiness, and a rise in body temperature, plus occasional maniacal behavior. A single dose of 0.1 to 0.2 milligrams of LSD-25 will produce profound mental disturbance within a half hour that lasts
10 hours. The lethal dose is 100 to 200 milligrams. Examples of incapacitating agents are: â¢ BZ â¢ LSD â¢ LSD Based BZ â¢ Mescaline â¢ Psilocybin â¢ Benzilates A.2 MEASURES OF TOXICITY The routes of exposure for chemicals can be one or more of inhalation, ingestion, or absorption through the skin. Inhalation In the context of accidental or deliberately engineered releases of toxic chemicals, there are several toxicity indica- tors. Examples are: â¢ LC50: the âLethal Concentrationâ in air that will prove fatal to 50% of the people exposed to it. This quantity is a function of exposure time. There are other, similar measures, such as the LC10 (10% probability of fatal- ity) and LC01 (1% probability of fatality). â¢ The probit Pr is a concise summary of the available data on a specific chemical. For a given exposure concentra- tion and duration, there is a chemical-specific equation from which the probit can be calculated. There is then a one-to-one relationship between the value of the probit and the probability that the exposed individual will suf- fer fatality. â¢ The Emergency Response Planning Guidelines (ERPGs) are published by the American Industrial Hygiene Asso- ciation (AIHA). As of 2002, AIHA had published ERPGs for 100 chemicals. There are three levels: (1) ERPG-1âmild irritation; (2) ERPG-2âa threshold for irreversible health effects; and (3) ERPG-3âa threshold for fatalities. Each ERPG is tied to an exposure time of one hour. EPA uses the ERPG-2 or equivalent as the toxic endpoint for hazards analysis in its Risk Manage- ment Program. â¢ Temporary Emergency Exposure Limits (TEELs) are published by DOE for 2234 chemicals at http://tis.eh.doe. gov/web/chem_safety/teel.html. Where ERPGs exist, the TEELs are the same. Otherwise, TEELs are intended to be temporary surrogates for ERPGs. â¢ Acute Exposure Guideline Limits (AEGLs) are essen- tially time-dependent ERPGs and are applicable to five emergency exposure periods (10 and 30 min, 1 h, 4 h, and 8 h). A five-minute AEGL is available for some chemicals. The latest information on AEGLS is at http:// www.epa.gov/oppt/aegl/chemlist.htm. They are essen- A-3 tially the most current attempts to develop toxicity lim- its that are useful in the context of emergency response. â AEGL-1 is the airborne concentration (expressed as parts per million or milligrams per cubic meter (ppm or mg/m3)) of a substance above which it is predicted that the general population, including susceptible indi- viduals, could experience notable discomfort, irrita- tion, or certain asymptomatic nonsensory effects. However, the effects are not disabling and are transient and reversible upon cessation of exposure. â AEGL-2 is the airborne concentration (in ppm or mg/m3) of a substance above which it is predicted that the general population, including susceptible individ- uals, may have irreversible or serious long-lasting health effects, or impaired ability to escape. â AEGL-3 is the airborne concentration (expressed as ppm or mg/m3) of a substance above which it is pre- dicted that the general population, including suscep- tible individuals, could experience life-threatening health effects or death. Table A-1 displays the relative inhalation toxicity of some of the chemicals mentioned above. The measure provided is the AEGL-2 for one hour, the ERPG-2, or the TEEL-2, depending on which is available. Dermal Exposure Some chemicals can produce hazardous health effects by being absorbed through the skin. For example, the blister agent sulfur mustard is specifically designed to do this. As noted above, it is persistent so that people handling contam- inated clothing or touching contaminated surfaces can also be affected hours or days after the release. The most effec- tive way in which percutaneous absorption might occur is Chemical 1-hr AEGL-2/ERPG-2/ TEEL-2 (mg/m3) Ammonia 105.0 Hydrogen chloride 30.0 Hydrogen Fluoride 16.4 Vinyl Chloride 12.5 Hydrogen Cyanide 11.1 Sodium Cyanide 5.0 Nicotine Sulfate 9.0 Chlorine 7.5 Benzidene 3.5 Chloropicrin 2.0 Parathion 1.0 TEPP 1.0 Phosgene 0.8 Phosphine 0.7 Sulfur Mustard 0.02 Sarin 0.006 VX 0.00027 TABLE A-1 Relative Measures of Acute Toxicity of Chemical Agents
when there is liquid on the skin. Both sarin and sulfur mus- tard are effective in this way. In both cases, vapor can also be absorbed through the skin, but this is a much less efficient mechanism. Ingestion Both sarin and sulfur mustard can be fatal if ingested. However, on the face of it, this does not seem a promising avenue for a terrorist who wishes to cause large numbers of casualties. It is conceivable that persons whose skin is cont- aminated by liquid sarin or sulfur mustard could ingest some of the agent, but this particular exposure pathway is not con- sidered further here. Factors Potentially Influencing Terroristsâ Choice of Chemical Agents All other things being equal, one would choose to use chemicals near the bottom of the Table A-1 for maximum effect. However, there are many other factors that influence the likelihood of successfully achieving terrorist objectives. Examples are as follows: â¢ The vapor pressure of VX at 20 oC is 0.00063 mm/Hg. If spilled on the floor, it will evaporate extremely slowly. Sarin is about 20 times less toxic, but its vapor pressure is 1.48 mm/Hg at 20 oC so that, under given conditions it will evaporate about 200 times more rapidly than VX. Thus, Sarin spilled on the floor of a railcar in the Tokyo Subway attack evaporated rapidly enough to cause fatal- ities and severe injuries: VX would not have. On the other hand, VX is a persistent agent. If the terroristsâ intention had been to cause severe cleanup problems, VX would have been a better choice than sarin. â¢ Both VX and sarin would be more effectively dispersed if they were aerosolized by some kind of spray device or explosion. Both of them cause severe health effects if absorbed through the skin. Dispersing them as fine droplets in a crowded area would be an effective way of ensuring skin contact. However, building effective aerosolizing devices is more difficult than simply arranging for a liquid to be spilled. â¢ On its face, chlorine is much less attractive as a poten- tial WMD. Its toxicity as measured by the 1-hr AEGL- 2 is one thirty thousandth that of VX. However, chlorine is available in extremely large quantities on the nationâs transportation networks (e.g. 90 ton railcars). Its vapor pressure at 20 oC is approximately 8 atm. When released catastrophically, as explained below, all of the chlorine may well become and remain airborne and cover many square miles with potentially dangerous concentrations within a few minutes. From the terrorist perspective, A-4 this could well be attractive. Similar arguments apply even in the case of ammonia, which has the highest of all the toxicity levels in the above table. In short, the terroristâs choice of a chemical will be driven by several factors: â¢ Toxicity â¢ Persistence â¢ Vapor pressure (i.e., ease of getting airborne) â¢ Ease of availability or manufacture of the chemical â¢ Mass available, and (last but not least) â¢ Objectives (e.g., maximize fatalities, maximize economic disruption, maximize cleanup problems). A.3 ATMOSPHERIC DISPERSION OF CHEMICAL AGENTS There is a huge variety of ways in which toxic chemicals can be released and subsequently dispersed in the atmosphere. The following are some examples pertinent to possible ter- rorist activities in the transportation system. Gases Liquefied Under Pressure and Transported in Bulk Ammonia and chlorine are transported as gases liquefied under pressure in amounts up to 90 tons in railcars. The attractiveness of these toxic materials from the terrorist per- spective is that it is possible to release large quantities very quickly, as is explained below. Chlorine In its usual transportation containers, (150lb or 1 ton cylin- ders, 17 ton road tankers or 90 ton railcars) chlorine is a gas liquefied under pressure. For example, at 77 oF (25 oC) the vapor pressure of chlorine is 8 atm1. When the liquid is released, part of it (about 20%) flashes to vapor and the rest can be fragmented into fine liquid droplets. Unless the emerging release encounters an obstacle (e.g. impinges on the ground or adjacent structures), all of the vapor and droplets can remain airborne. If the size of the rupture in the chlorine container is sufficiently large, most if not all of the whole contents can become airborne in a time that varies from virtually at once to a few minutes2. Thus, a terrorist could put a large quantity of highly toxic chlorine into the air in a time so short that measures to minimize the release can- not be implemented. As air is entrained, the liquid droplets 1 Physical properties of chlorine (vapor pressure, boiling point, flash fraction, fraction of liquid droplets remaining airborne, etc.) obtained from Appendix A of CCPS (1996). 2 Appendix A of CCPS (1996) calculates the rate of release of chlorine through a hole of diameter 3â in the liquid space of a chlorine vessel at 25 oC as 44.16 kg/s.~97 lb/s. This would empty a 17 ton road tanker in about 6 minutes. If terrorists used explosives to blow a one-foot diameter hole in the side of the road tanker, the vessel could empty in about 20 seconds.
evaporate. The cloud is denser than air, partly because the molecular weight is 70, partly because the flashing release is initially at the boiling point of chlorine (about â30 oF), and partly because initially entrained air is cooled as it evaporates the droplets. One characteristic of heavy vapor clouds is an initial slumping phase when they behave much as would a liquid. They become very broad and can even back up against the wind. For 20 ton or 90 ton releases, the backup can be sev- eral hundred meters and the cloud can become a kilometer or more 3in width within a few minutes. Another consequence of the pseudo-liquid-like behavior of heavy vapor clouds is that they will tend to run into depressions, ditches, drains, and basements of buildings (if released in an urban area). They can persist there for several hours if the windspeed is low or if internal volumes are not well ventilated. This can be dangerous for unwary members of the public and emer- gency rescue personnel. Once the initial slumping phase is over, during which the chlorine is considerably diluted by air entrained due to tur- bulence generated within the cloud itself, the vapor begins to behave more like a passive plume in which the predominant dilution mechanism is the entrainment of air by the action of atmospheric turbulence. Conservative atmospheric disper- sion models, such as that used by EPA in its guidance on atmospheric dispersion modeling for Risk Management Pro- grams (EPA 1999) as embodied in RMP*Comp (EPA 2004), show that a 17 ton release of chlorine in unfavorable weather conditions can propagate up to 5.8 miles at an urban site or 12 miles at a rural site before falling below the ERPG-2 of 3 ppm. The corresponding distances for a 90 ton railcar are 14 miles and > 25 miles respectively. While it is exceedingly unlikely that such distances will actually be achieved, never- theless it is entirely possible that a broad chlorine vapor cloud would cover many square miles with concentrations that could prove highly injurious or fatal, and that this could happen within a few minutes. Anhydrous Ammonia Anhydrous ammonia is also transported as a gas liquefied under pressure in road tankers or 90 ton railcars. Many peo- ple believe that, because ammonia is less dense than air, it will rise when it is released. This is true if the release con- tains only pure vapor. However, it is well established that sudden, large releases of ammonia from pressurized contain- ment behave in much the same way as do large releases of chlorine. There is an initial flashing phase when typically ~ 20% of the ammonia vaporizes and fragments the remain- ing liquid. Because the situation is highly turbulent, much air A-5 is entrained virtually instantaneously (as much as 10â20 times as much air as the mass of ammonia released). This evaporates the liquid droplets and, within less than a minute, the cloud consists of air at the boiling point of ammonia mixed with a few percent of ammonia. This mixture is heav- ier than air and slumps as was described for chlorine above, becoming very broad and backing up against the wind (Kaiser 1989)4. The principal difference between chlorine and ammonia as an effective means of causing mass casualties lies in the very different levels of toxicity. From Table 7.2.2, the AEGLs for Ammonia are 50â150 times as large as those for chlorine. Therefore, a release of ammonia will propagate to a much smaller distance than would the release of the same mass of chlorine before falling below levels such as AEGL-2 or ERPG-2. This is evidenced by results from RMP*Comp, which show that 17 tons of ammonia would propagate up to 2.1 miles at an urban site or 3.3miles at a rural site before falling below the ERPG-2 of 150 ppm. The corresponding distances for a 90 ton railcar are 4.9 miles and 7.8 miles respectively. As for chlorine, it is extremely unlikely that such conservative distances would actually be realized as a result of an ammonia release. A study (Kaiser et al. 1999) that compiled data from 12 accidental releases of ammonia, including railcars and road tankers (Markham 1986), large- scale anhydrous ammonia release experiments (Goldwire et al. 1985), and dispersion models showed that concentrations of 20,000 ppm (which is the LC50 for durations of cloud pas- sage of a few minutes) are never exceeded beyond a few hun- dred meters from the point of release. Concentrations in the range 1,000â10,000 ppm are not seen or predicted beyond about 2 km. This observation is pertinent for emergency responders in that, in practice, distances predicted by con- servative models with conservative endpoints like those in RMP*Comp are extremely unlikely to be encountered. Materials That Are Liquids at Typical Ambient Temperatures Many toxic chemicals have boiling points that are above any reasonably anticipated ambient temperature. For exam- ple, sarin has a boiling point of 316.4 oF5. Therefore, if spilled, it will initially form a slowly evaporating pool on the ground. For such pools, the subsequent rate of evaporation is limited by mass transfer from the pool to the atmosphere and is dependent on vapor pressure (1.4 mm/Hg at 68 oF ~ 1,800 3 For the 3â hole described in footnote 2, Appendix A predicts that the chlorine vapor cloud will be 3.4 km wide at a distance downwind of 1 km in atmospheric stability cate- gory F with a windspeed of 2 m/s. However, this predicted width is reduced by a factor of 5 in atmospheric stability category D with a windspeed of 5 m/s. 4 On May 11 1976, a road tanker carrying 19 metric tons of anhydrous ammonia crashed through a barrier on an elevated section of motorway near Houston, TX. The pressurized tank burst on falling to the roadway below and the contents were rapidly released. (McMullen 1976). After about one minute the observed breadth of the cloud was 400â600m. The maximum upwind distance, measured by observing burnt grass, was 200m. These observations are substantiated by photographs in Fryer and Kaiser (1979). In 1977, a train derailment in Pensacola, Florida punctured the tank head on a rail car so that 50% of the contents, amounting to some 40 metric tons, quickly vaporized. (NTSB, 1978). After about five minutes, the cloud was already âabout a mile across.â 5 Physical data on Sarin from the National Institute of Occupational Safety and Health (NIOSH) Emergency Response Card for Sarin, available at http://www.bt.cdc.gov/ agent/sarin/erc107-44-8pr.asp.
ppm), the area within which it is confined or over which it spreads, and the windspeed (CCPS 1996, Section 4.2.5). Sarin is likely to be available to terrorists only in small quantities. For example, on March 20, 1995 a terrorist cult released a few liters of sarin in commuter trains on three different Tokyo subway lines. They concealed the sarin in lunch boxes and soft-drink containers and placed it on subway train floors. It was released as terrorists punctured the containers with umbrellas before leaving the trains so that the sarin spilled onto the floor of the subway cars and began to evaporate (Ohbu et al. 1997). Therefore, it is more likely to be used in an attack on the transportation system than in an open-air release. In point of fact, pouring agents like sarin and sulfur mus- tard on the ground or floor is a relatively inefficient way of dispersing them. If terrorists were able to obtain or manufac- ture chemical munitions, they could cause a more devastat- ing impact. For example, as quoted by Purver (1995) from Berkowitz et al. (1972): âIn the open, six pounds of Sarin distributed by a three pound burster charge at a height of 15 feet creates a dosage of 3500mg-min/m3 20 yards from the burst within 10 sec- onds; in 25 seconds the cloud expands to a 50-yard radius with a minimum dosage of 100 mg/m3 (Robinson, 1967). A minute after the burst, anyone in an area of over 70,000 square feet around the burst will have received at least a median lethal dose, and probably much more than that. In a confined space (banquet hall, auditorium), the effects will be even greater.â Similar comments about effects could be applied to a con- fined space such as a crowded airport terminal. The impact would be further enhanced by the use of a persistent agent such as sulfur mustard. Note that Purver (1995) states that âChemical weapons such as nerve agents are generally credited with being capa- ble of causing casualties in the range of hundreds to a few thousand.â Thus the magnitude of the consequences would not be as great as could potentially be obtained from biolog- ical or nuclear weapons. HVAC Systems Gases could potentially be introduced into the HVAC sys- tems of large buildings such as airport terminals or large vessels such as passenger liners. For example, it might be possible to release the contents of a 150 lb cylinder of chlo- rine into the HVAC intake. It would also be possible to envisage pouring the contents of a container of sarin the size of a milk bottle into the HVAC system. Releases into or onto Water As an example of potential terrorist interest in the trans- portation of toxic chemicals, consider the transportation of anhydrous ammonia by barge. This occurs extensively on inland and coastal waterways in the U.S. in barges typically containing 2,500 tons of refrigerated liquid ammonia. If the A-6 tanks containing this ammonia were to be ruptured (perhaps by a USS Cole-type attack), up to 40% of the refrigerated ammonia could be vaporized (Raj et al. 1974). What happens is that, as ammonia dissolves in the water, heat is released, which is responsible for vaporizing the aforementioned 40%. Raj et al. also show that the resulting vapor cloud is slightly buoyant (i.e., the liquid droplet effects described for the pres- surized release are not important in this instance). However, this does appear to be a potential mechanism for rapidly vaporizing up to a thousand tons of ammonia. A.4 METHODS FOR INTRODUCING AND DISPERSING CHEMICALS The means of maliciously introducing and dispersing chemical agents discussed above can be summarized as follows: â¢ Hijack a railcar or road tanker of chlorine or ammonia, move it to an optimum location and rupture the con- tainment with an explosion. Within a short time, an area potentially encompassing many square miles could be covered by a vapor cloud that is potentially dangerous or fatal to those exposed. â¢ Cause a highly toxic chemical such as sarin to be spilled in a confined space such as a subway car. It will evapo- rate and cause adverse or fatal health effects through inhalation. â¢ Obtain a chemical munition containing an agent such as sarin and explode it in a large confined volume such as an airport terminal. Individuals would be affected both through inhalation and by liquid droplets impinging on exposed skin. The adverse consequences could be enhanced by using a persistent agent such as sulfur mus- tard whereby surfaces would be dangerously contami- nated by deposited liquid for hours or days. â¢ Introduce a toxic gas into the HVAC system of a build- ing such as an airport or marine terminal, or the HVAC system of a ship. â¢ Hijack and blow up a barge or ship containing a large quantity of material such as ammonia. Note that these are only meant to be examples of a much larger number of scenarios that can be imagined. A.5 EXAMPLES OF CHEMICAL INCIDENTS The following subsections provide summaries of chemical events that are examples of the potential effects of a terrorist attack using a chemical agent. Some of the summarized events are hypothetical, most are actual accidents that have occurred. These examples are grouped as: chemical weapons events, chemical transportation incidents, and releases from fixed facilities.
A.5.1 Chemical Weapons Events The following are examples of actual or potential attacks on transportation systems. The Tokyo Subway AttackâSarin The most notorious chemical attack on an element of the transportation system was the use of sarin in an assault on the Tokyo subway system on March 20, 19956. Members of a Japanese millenarian cult, Aum Shinrikyo released sarin gas on several lines of the Tokyo subway. As a result, 12 people died and thousands were injured. Aum Shinrikyo was a Japanese millenarian cult centered on the charismatic leader Asahara Skoko, whose teachings combined elements of Buddhism and Hinduism as well as millenarian Christianity. Central to the groupâs teachings is that the apocalypse is near. The Aum cult attracted people from all walks of life, and had as many as 10,000 followers at its peak. Members lived in communes, cut off relations with outsiders, and gave all their savings to the cult. It was believed that the cult was to become more powerful than the state, and needed the most advanced weapons of mass destruction to achieve this end. Monday, 20 March, 1995 was for most a normal workday, though the following day was a national holiday. The attack came at the peak of the Monday morning rush hour on one of the worldâs busiest commuter transport systems. The liquid sarin was contained in plastic bags which each team then wrapped in newspapers. Each of the five perpetrators carried two packets of sarin totaling approximately 1 liter of sarin, except for one who carried three bags. Thus, there was a total of only about 5 liters used in the attack. A single drop of pure sarin the size of the head of a pin can kill an adult. However the sarin used was diluted, possibly to cause slower effects and thereby allow more people to be exposed. Carrying their packets of sarin and umbrellas with sharp- ened tips, the perpetrators boarded their appointed trains; at prearranged stations, each perpetrator dropped his package and punctured it several times with the sharpened tip of his umbrella before escaping to his accompliceâs waiting get- away car. The Tokyo subway system transports millions of passengers daily. During rush hour trains are frequently so crowded that it is impossible to move. As noted above, 12 people were killed in the attacks. Symptoms included choking, coughing, foaming at the mouth, and fading vision as victims staggered from the trains. Those that received the highest doses fell to the ground, writhing in convulsions. At one subway station in particular, the subway entrance was described as resembling a battlefield, where the injured simply lay on the ground, struggling for breath. In other areas, many of those affected A-7 by sarin went to work in spite of their symptoms. Most of these left and sought medical treatment as the symptoms worsened. Several of those affected were exposed to sarin only by helping passengers from the trains (these include passengers on other trains, subway workers and health care workers). Recent surveys of the victims show that many are still suffering, particularly from post-traumatic stress disor- der. In one survey, 20% of 837 respondents complained that they feel insecure whenever riding a train, while 10 percent answered that they try to avoid any gas-attack related news. Over 60 percent reported chronic eyestrain and said their vision has worsened (Mainichi Online 2001). Purver (1995) provides a detailed review of the then-current information about the subway attack. He suggests that Aum Shinrikyo experimented with poison gas before March 20, 1995. For example, in the mountain resort of Matsumoto, 125 miles northwest of Tokyo, late in the evening of 27 June 1994, a substance later identified as sarin seeped through the open windows of apartments and houses, killing or injuring every living thing inside an area 500 yards long by 100 yards wide. Seven people were killed and 264 sought hospital treatment. A report in the London Sunday Times of 19 March 1995 (just one day before the subway attack) reported that there had been âan intensive investigation involving a special sarin unit of Tokyoâs metropolitan police department criminal investiga- tion laboratory, the national police agency, and the security services that had concluded that âthe attack was a trial run by terrorists of the delivery systemâ of a chemical agent. Purver also indicated that commentators expressed sur- prise that, given the toxicity of sarin and the nature of the target, the casualty toll had not been much higher. Had the terrorists implemented a means of aerosolizing the sarin, rather than simply letting it evaporate from the floor of the subway cars, many more people would have been killed, although the rapidly falling bodies would have prevented others from entering the train, thus possibly exposing fewer people to the sarin. Police raids on Aum Shinrikyo com- pounds uncovered the capacity to make much larger quanti- ties of sarin, other agents such as tabun (which is ten times as toxic as sarin), evidence of experimentation with biolog- ical weapons including botulism, and evidence of attempts to obtain the Ebola virus and Q-fever. An Aborted Attack on the Tokyo Subwayâ Hydrogen Cyanide On Friday May 5, 1995, Tokyo subway guards responded to a fire in a public restroom, and averted a potential mass- casualty incident (Purver 1995). The incident it attributed to Aum Shinrikyo cult. Two plastic bags, one containing two liters of powdered sodium cyanide already in flames and the other containing 1.5 liters of diluted sulphuric acid, were found side-by-side on the floor of a menâs bathroom in Tokyoâs busiest subway station, Shinjuku. The bags reportedly were arranged so that a reaction producing hydrogen cyanide gas would have occurred if contents from the two bags had mixed. 6 The primary source for this summary, except as otherwise noted, is âProtecting Pub- lic Surface Transportation Against Terrorism and Serious Crime: Continuing Research and Best Security Practices,â published by the Mineta Transportation Institute (2001), and available at. http://www.transweb.sjsu.edu/publications/terrorism_final.pdf.
Four subway guards who doused the flames with water were overcome by fumes and briefly hospitalized, but otherwise there were no casualties. Chemical experts later estimated that the amount of cyanide gas that could have been released may have been sufficient to kill between 10,000 and 20,000 people. Hypothetical ScenarioâPersistent Agent Several workers move drums labeled as cleaning agents into a large shopping mall, large public facility, subway, train station, or airport. They dress as cleaners and are wearing what appear to be commercial dust filters or have taken the antidote for the agent they will use. They mix the feedstocks for a persistent chemical agent at the site during a peak traf- fic period. Large-scale casualties result, and draconian secu- rity measures become necessary on a national level. A series of small attacks using similar âbinaryâ agents virtually para- lyze the economy, and detection is impossible except to iden- tify all canisters of liquid (Cordesman 1996). A.5.2 Chemical Transportation Incidents Bulk chemicals are transported by truck, rail, barge, and ocean-going vessels. For chemicals categorized as âtoxic by A-8 inhalationâ, 59% of the annually transported tonnage moves by rail, amounting to 5,766,000 tons. Based on ton-miles, a much larger proportion of chemicals (i.e., 95% or, 4,940x106 ton-miles) are transported by rail (GAO, 2003a). To exem- plify the variety of hazardous bulk chemicals shipped in the US, Table A-2 shows the volumes of hazardous materials shipped by rail from 1998â2001. The authors are unaware of any deliberate, large-scale release of a highly toxic chemical from the transportation system. The scenarios presented after Table A-2 are examples of accidental releases that could potentially be replicated by terrorists. Ammonia Release from Road Tank Truckâ Houston, TX Between 11.09 am and 11.15 am on May 11, 1976, a tank truck crashed through a guard rail at the Southwest Freeway and Loop 610 West Interchange, landing on a street some 30 feet below. The tank ruptured on impact, releasing 19 tons of anhydrous ammonia. The accident resulted in the imme- diate death of four people and injury to more than one-hun- dred others. Two other victims died later as a result of their injuries. (McMullen, 1976). One of the most striking fea- tures of this accident was that the ammonia cloud exhibited Hazardous Material Estimated Total Carloads, 1998-2001b Estimated Annual Average No. of Carloads Freight Forwarder Trafficc 1,188,109 297,027 All freight rate shipments, not elsewhere coded (NEC) or trailer on flatcar shipments, commercial, except where identified by commodity. 716,177 179,044 Sulfur liquid or molten nonmetallic minerals except fuels 273,005 68,251 Liquefied Petroleum Gas, NEC, compressed 253,234 63,308 Sodium (soda), caustic (sodium hydroxide) 236,455 59,114 Asphalt pitches or tars, from petroleum, coal tar, coke ovens, or natural gas 222,163 55,541 Sulfuric acid or oil of vitriol 200,875 50,219 Anhydrous ammonia 163,057 40,764 Chlorine 128,600 32,150 Gasolines, blended, consisting of motor fuels containing 50% or more of gasoline 97,192 24,298 Ethyl alcohol, anhydrous denatured in part with petroleum products and/or chemicals (not to exceed 5 wt%) 95,333 23,833 Phosphate fertilizer solution, containing not more than 77 wt% of phosphoric anhydride 90,779 22,695 Chemical, NEC 86,854 21,713 Vinyl Chloride (chloroethane or chloroethylene) 73,033 18,258 Methanol (methyl or wood alcohol liquid) 67,903 16,976 Propane gas, liquefied 65,702 16,425 Carbon dioxide gas, liquefied or carboic acid gas 63,020 15,755 Ammonium nitrate fertilizer 62,563 15,641 Muriatic (hydrochloric) acid 58,165 14,541 Styrene (liquid) 55,910 13,977 Footnotes for Table A-1 a. Reproduced from GAO (2003a), Appendix III, Table 4 b. Extrapolated from a 1% sample of waybills c. Non-bulk, mixed TABLE A-2 The Top 20 Hazardous Materials Shipped by Rail By Volume 1998â2001a
the ground-hugging features of a dense vapor cloud. Figure A-1 was taken about a minute after the after the crash from the Transco tower in the Houston Galleria complex. One can see that as the vapor cloud propagates downwind (from right to left), it appears to decrease in height. This is characteristic of a slumping, heavy vapor cloud. In addition, the cloud is backing up against the wind towards the right of the picture. This is reinforced by Figure A-2, which was taken from a Houston Air Pollution Control Program Enforcement helicopter after the ammonia had dispersed. The photograph shows the area where the grass had been burnt by the ammonia. The wind was blowing from left to right. One can clearly see that the plume spread a consider- able distance both across the wind and upwind. The upwind boundary is at about 200m and the cloud is about 600m across just downwind of the overpass (Kaiser 1979). As was briefly discussed above, it is well established that sudden, large releases of ammonia from pressurized contain- ment behave as denser-than-air vapor clouds. There is an initial flashing phase when typically ~ 20% of the ammonia vaporizes A-9 and fragments the remaining liquid. Because the situation is highly turbulent, much air is entrained virtually instantaneously (as much as 10â20 times as much air as the mass of ammonia released). This evaporates the liquid droplets and, within less than a minute, the cloud consists of air at the boiling point of ammonia mixed with a few percent of ammonia. This mixture is heavier than air (Haddock and Williams 1978) and slumps as was described for chlorine above, becoming very broad and backing up against the wind (Kaiser 1989). Depending on the weather conditions and the amount of ammonia released, the vapor cloud can be considerably larger than that observed in Houston. See the discussion of the Pen- sacola accident below for an example. This incident is also relevant to potential large-scale delib- erate or accidental releases of chlorine. In fact, McMullen (1976) explicitly states âHad the truck been loaded with chlo- rine the potential for disaster would be further com- pounded.â To the authorsâ knowledge, there has not been a large chlorine spillage during transportation that has resulted in the immediate or near-immediate release of the contents of the transportation vessel and which resulted in reasonably good documentation of the size of the vapor cloud. However, there is no doubt that chlorine would behave in much the same way as ammonia. If anything, the slumping effects for chlorine would be even greater because chlorine, even as a pure vapor, is already denser than air. Section 184.108.40.206 cites a calculation Appendix A of CCPS (1996) calculates the rate of release of chlorine through a hole of diameter 3â in the liq- uid space of a chlorine vessel at 25 oC as 44.16 kg/s.~97 lb/s. This would empty a 17 ton road tanker in about 6 minutes7. Appendix A predicts that the chlorine vapor cloud will be 3.4 km wide at a distance downwind of 1 km in atmospheric stability category F with a windspeed of 2 m/s8. Ammonia Release from RailcarâPensacola About 6.06 pm on November 9, 1977, 2 SD-45 locomotives and 35 cars of Louisville and Nashville freight train No. 407 derailed at Pensacola, Florida (NTSB 1978). The adjacent tank heads of the 18th and 19th cars were punctured and this released anhydrous ammonia into the atmosphere. Two persons died and 43 were injured. NTSB established that, within ten minutes, about 50 per- cent of the contents of the 19th car quickly vaporized, so that about 40 tons became airborne. The air traffic controller at the Pensacola Airport first observed the ammonia cloud on radar at about 6.10 pm (i.e., only 4 minutes after the accident). At that time, âIt appeared to be about one mile in diameter and about 125 feet high.â This substantiated the statement, made above, that clouds that are considerably broader than that observed in the aftermath of the Houston road tanker crash are possible. Figure A-1. Cloud of ammonia fumes spreading over the west loop 610 overpass, Houston, TX, May 11, 1976. Photo reproduced from Fryer and Kaiser (1979). Photograph taken by Carrol S. Grevemberg. Figure A-2. Photo of the Houston ammonia tank truck crash site after the ammonia had dispersed. 7 If terrorists used explosives to blow a one-foot diameter hole in the side of a 17-ton road tanker, the vessel could empty in about 20 seconds. 8 Note, however, that this predicted width is reduced by a factor of 5 in atmospheric stability category D with a windspeed of 5 m/s.
Other Ammonia Transportation Accidents Crete, Nebraska, February 18, 1969: At about 6.30 am on February 18, 1969, Chicago, Burlington, and Quincy (CB&Q) train No. 64 derailed the 72nd to the 90th cars as the train was entering Crete, Nebraska (NTSB, 1971). The derailed cars struck Train 824, standing on a track north of the main track. A tank car in train 824 was completely fractured by the impact and rapidly released 29,200 gallons (about 76 metric tons) to the atmosphere. NTSB states that âA cloud was formed which blanketed the immediate area. The cloud extended westward beyond the (nearby) Blue River and for several blocks north and south of the railroad. The concentrated cloud of ammonia vapor was retained in the area for a considerable period of time. Unfortunately, the NTSB report provides no further data on the size of the cloud. Three trespassers riding on rain 64 were killed as a result of the derailment. Six people were killed and 53 were injured as a result of exposure to the ammonia cloud. The ambi- ent temperature was 4 oF and the wind was calm. Minot, North Dakota, January 18, 2002: On January 18, 2002, a 112-car train derailed one mile west of Minot, ND in the Souris River valley. Eleven anhydrous ammonia cars and two granular urea cars released their contents. In excess of 300,000 pounds (about 136 metric tons) of anhydrous ammo- nia was released from the ruptured tanker cars. One tanker car was propelled over 600 feet in the air and skidded another 500 feet on the ground, crashing through a corner of a nearby house. Part of another tanker car landed on the ice of the Souris River. A dense anhydrous ammonia cloud drifted through low-lying areas of the city. A temperature inversion, with sub-zero temperatures and no wind, prevented the cloud from dissipating. One fatality occurred, 15 people were hos- pitalized, and eventually over 1600 people sought medical treatment for anhydrous ammonia exposure (Radig 2003). GAO (2003a) states that the vapor plume was 5 miles long and 21/2 miles wide. It caused 1 death and more than 300 injuries and affected 15,000 people. This was in a relatively lightly populated rural area. NTSB has not yet released its report on this accident. Chlorine Release from RailcarâMississauga Just before midnight on Saturday, November 10, 1979, a derailment occurred involving 24 railway tanker cars of a 106- car Canadian Pacific Railway Train in the city of Mississauga, Ontario. One railcar containing 90 tons of liquid chlorine began to leak. More than 200,000 Mississauga residents were evacuated for their protection against potentially harmful exposure to chlorine drifting away from the site, and as a pre- caution against the possibility of further explosions, which might have released more chlorine. (SCIEX 1980). SCIEX performed numerous measurements of airborne chlorine concentrations from November 11, 1979 at 1 pm until 10 am on November 19 at numerous locations ranging from about 0.5 km to 5 km from the point of release. The highest recorded concentrations were about 400 g/m3 and the most serious health effects noted were âEye irritation in A-10 plume.â For comparison, the chlorine AEGL-2 for a 1-hour exposure is 2 ppm (6,000 g/m3âsee Table 7.2.2). No chlo- rine concentrations above 1 g/m3 were observed after 8 am on November 15. It is apparent that the chlorine was leaking slowly over a period of days. Thus, the Mississauga derailment and subse- quent chlorine leak is not an example of a catastrophic release that might be engineered by a terrorist, although presumably a terrorist would be gratified by the dislocation caused by the need to evacuate nearly a quarter of a million people for sev- eral days. Howard Street Tunnel FireâBaltimore, MD At 3:04 pm on Wednesday, July 18, 2001, the 60-car CSX freight train L412-16 entered the Howard Street Tunnel in downtown Baltimore, MD (Carter et al. 2002)9. The train car- ried 29 loaded and 31 empty cars, including several tanker cars. At 3.07 pm, the train lurched and came to a rough stop as several cars derailed. The engineers uncoupled the three diesel engines and exited the tunnel in order to report the event. From the amount of smoke exiting the tunnel, it was clear that there was a fire somewhere among the cars. Baltimore City firefighters received notification of the event between 3.35 pm and 4.15 pm. After reviewing the bill of lading firefighters discovered that the freight train was car- rying a variety of hazardous materials including tripropylene and hydrochloric acid. One of the immediate problems was to determine the poten- tial environmental impact from the hazardous materials and whether downtown Baltimore needed to be evacuated. This problem was solved by the Maryland Department of the Envi- ronmentâs (MDEâs) Emergency Response Division (ERD). Following a review of the bill of lading, ERD personnel con- tacted members of the South Baltimore Industrial Aid Plan (SBIMAP), a voluntary consortium of manufacturers, emer- gency response personnel, Baltimore City environmental and emergency management personnel, and MDE. SBIMAP pro- vided two chemists, who quickly determined that, individually or in combination, the hazardous chemicals involved in the fire would not present a serious environmental hazard and that it was not necessary to evacuate downtown Baltimore. At 10.30 am on July 19, 2001, firefighters found small leaks in one of the hydrochloric acid cars. This necessitated the arrangement of acid transfer activities. The damaged HCl car was finally removed from the tunnel at 11.30 am on July 22. This incident, like the one at Mississauga, is thus more notable for the disruption it caused over a period of several days and the indications it gives of the potential danger. Had there been a car load of chlorine and had the crash caused a significant rupture, the released chlorine would have flowed down the 4.8% grade of the tunnel and poured out into down- town Baltimore. This accident is interesting because of its 9 This description of the Howard Street Tunnel derailment draws freely upon and quotes from the work by Carter et al. (2002).
effects on transportation systems in and around Baltimore. It had a major effect on rail transportation, road transportation and transit. It had some effect on marine transportation (the inner harbor was briefly closed to traffic) and some effect on air (to the extent that transportation to and from BWI airport was affected. These are documented by Carter et al. (2002). The short-term transportation impacts of the July 18, 2001 Howard Street Tunnel fire lasted up to 36 hours and included the following: Closing of Major Highways into Baltimore: at the request of the Incident Commander, Maryland State authorities closed major highways into the city, including I-83 southbound, MD- 295 northbound (the Baltimore-Washington Parkway), Route 40 eastbound and I-395 northbound. These roadways were reopened on the morning of July 19. Closing of City Streets in the Vicinity of the Tunnel and the Rerouting of Passenger, Bus, and Commercial Vehicle Traf- fic: Howard Street and the surrounding area were closed to traffic, cutting Baltimoreâs central business district in half and closing off east-west traffic flows (Howard Street runs North-South for 1.7 miles over the tunnel). This resulted in gridlock, but once traffic management was put in place the City was cleared of traffic within 2 hours of the normal end of rush hour (8pm instead of 6pm). Closing of the Metro Subwayâs State Center Station: the station was closed due to smoke accumulation from the fire. However, Metro officials conducted an inspection of the Metro tunnel running under the Howard Street tunnel and determined that no damage had occurred. They were able to keep the trains running. The State Center Station itself was reopened on July 21, 2001. The Disruption of Light Rail Service: light rail track runs along Howard Street. In the immediate vicinity of the fire a water main ruptured and washed away the track bed, neces- sitating the closure of the light rail service. Metro set up a bus bridge between nearby stations to carry passengers around the break. The Disruption of Maryland Commuter (MARC) rail and Oriole Game Day Service: MARC trains were stopped at the Dorsey Station near BWI Airport and bus bridge was set up by the MTA to bring passengers into the city. The bus bridge was only needed for July 18. 2,000 Orioles employees and between 2,500 and 5,000 fans were evacuated. The Disruption of Bus Services: disruptions were system- wide. The Closing of the Inner Harbor: the U.S. Coastguard (USCG) closed the inner harbor to boat traffic at 5.00 pm and set up booms to minimize potential contamination from chemicals seeping from leaking rail cars. The Disruption of Rail Freight Along the East Coast: the Howard Street Tunnel is one of only two direct northeast- southeast freight lines along the East Coast. Losing access to the tunnel required CSX to divert or delay a significant por- tion of rail traffic along the Eastern Seaboard. Freight moving from the northeast to Florida was advised to expect delays of A-11 24 to 36 hours. Some freight that would normally have used the tunnel was diverted as far west as Ohio. Medium-term transportation impacts of the July 18, 2001 Howard Street Tunnel fire continued until July 19 to 23, 2001. These impacts included the following: Suppression of and Initial Clean-up from the Tunnel Fire took approximately 5 days. All cars were removed from the tunnel and inspected for damage, and all haz- ardous materials were off-loaded and removed. The tunnel was inspected for structural damage and reopened to rail traffic on July 23. Miscellaneous: On Monday morning, July 23, when city and state employees returned to work after taking advantage of liberal leave at the end of the previous week, traffic was backed up for more than a mile on northbound I-95 before the junction with thI-395 spur that takes traffic downtown. Street closures in the vicinity of Howard and Lombard Streets caused MTA officials to divert approximately 23 bus routes. Light rail service continued to rely on buses to trans- port riders between the Patapsco and North Avenue Station stops around the section of track damaged by the water main break. For five days following the accident, streets in the vicinity of the tunnel and the water main break remained closed, and all vehicle traffic was diverted. On July 24, nearly all of the streets were reopened to traffic. Only a two- block stretch of Howard Street and a portion of Lombard Street remained closed. Metro: commuters took advantage of Metro services to travel into Baltimore during this time. Ridership on Monday July 24 was 7,000 higher than normal. East Coast Rail Network: the East Coast network became increasingly constrained with each day that the major north- south artery through the Howard Street tunnel remained closed. Freight trains were delayed, cancelled, or diverted hundreds of miles throughout the Middle Atlantic States. MARC Rail Service on the Camden Line was also disrupted until the fire was suppressed. Service ended at the Dorsey Street Station near BWI airport. However, there was not a sig- nificant decrease in ridership because MARC commuters to and from Washington D.C. took advantage of free parking at the Dorsey Street station. Longer-term transportation impacts of the July 18, 2001 Howard Street Tunnel fire included only the 55 days it took to repair the damage to the Central Light Rail Line caused by the bursting water main. Relatively speaking, there were no long-term impacts as might have been the case in the event of a radiological incident. Table A-3 summarizes the agencies that were involved and their roles and responsibilities. A.5.3 Releases from Fixed Installations There is a vast range of releases of toxic chemicals from fixed installations. The following are a couple of examples.
Methyl Isocyanate ReleaseâBhopal The Union Carbide India Ltd. (UCIL) pesticide plant in Bhopal, India produced Methyl Isocyanate (MIC) by reacting monomethylamine with phosgene in the plantâs MIC produc- tion unit. The MIC was used to make SEVIN carbaryl and sev- eral other carbamate pesticides. The MIC was stored in two horizontal, mounded, 15,000-gallon, stainless steel tanks. On the night of December 2â3, 1984, the 41 metric tons of MIC in one of the tanks underwent a chemical reaction which was caused by the introduction of water into the tank (Kalelkar A-12 1988). A toxic cloud of MIC drifted over the hundreds of dwellings in a crowded shanty town outside the plant, killing more than 3,800 people and leaving 11,000 more with perma- nent disabilities (figures provided by the Indian Government in 1991). Investigations proved that the introduction of water was deliberate (Browning 1993). Once the toxic cloud had dispersed and drifted away, there appeared to be no lasting impact on transportation systems around the plant. The water was introduced by a disgruntled employee who may not have been aware that the conse- quences would be so catastrophic. However, it does illustrate the potential for terrorists to work with an insider to engineer catastrophic releases of toxic chemicals. Hydrogen Fluoride ReleaseâTexas City At 5.20 pm on October 30, 1987, a crane was moving a 50-foot, multi-ton convection section from a vertical heating Jurisdiction Modal Administration or Agency Role in Incident Response Area of Concern Fire Department Incident Command Fire Suppression. Police Department Traffic Enforcement Closing of streets crossing over the Howard Street Tunnel. Department of Public Works Infrastructure Repairs Traffic Management Repairs to water main and street surface at Howard and Lombard Streets. Traffic control in Baltimore. Baltimore City Office of Emergency Management Interagency Coordination & Public Information Media information. Headquarters Coordination of DOT Response Activities Worked with Baltimore Department of Public Works (DPW) to establish a plan on how to repair the infrastructure damage once the fire was extinguished (procurement issuesâhaving a contractor in place, developing a plan on how repair work would be implemented once the âgreen lightâ was received, plans for site survey, traffic diversion, etc.). State Highway Administration Traffic Management on Interstate System Through CHART system10, posted notices on fixed and mobile DMS advising that major routes into the city were closed. Mass Transit Administration Rail and Bus Transit Operations in Baltimore City Light rail and bus operations. Establishing bus bridge between north and south segments of light rail. MARC operations. Metro subway operationsâtunnel inspection. Maryland Department of Transportation Maryland Transportation Authority Traffic management on I-95 approaches to Ft. McHenry Tunnel & I-395 Ensured that I-395 route into Baltimore was closed off during initial incident response activities. Maryland Department of the Environment (MDE) Emergency Response Division Air Quality, Water Quality, Hazardous Materials, Leaks/Discharges Obtained information possible environmental impact of train fire (hazardous materials). Monitored air and water quality in area around the tunnel and the Inner Harbor. Checked railcars pulled from tunnel for structural integrity. Coordinated removal and disposal of hazardous materials from the train. Maryland Emergency Management Agency N/A Coordination of State Government Emergency response and Incident Management Activities Coordinating activities of state agencies. Media relations and rumor control. U.S. Coast Guard (USCG) USDOT Supported MDEI Implemented waterway safety measures, including closing of Inner Harbor. Supported hazardous material detection and containment. U.S. Environmental Protection Agency (EPA) N/A Supported MDE Assisted with monitoring of air quality and water TABLE A-3 Incident Response and Agency Responsibilities (Table 5 from Carter et al 2000) 10 The CHART (Chesapeake Highways Advisories Routing Traffic) program started in the mid-1980s as the âReach the Beachâ initiative, focused on improving travel to and from Marylandâs eastern shore. It has become so successful that it is now a multi-jurisdictional and multi-disciplinary program extending statewide. This comprehensive, advanced traffic management system is enhanced by a newly con- structed state-of-the-art command and control center called the Statewide Opera- tions Center (SOC). The SOC is the âhubâ of the CHART system, functioning 24 hours-a-day, seven days a week with satellite Traffic Operations Centers (TOCs) spread across the state to handle peak-period traffic.
vessel to a semi-truck trailer in Marathon Petroleum Com- panyâs Texas City refinery. The crane was located immedi- ately east of the HF acid vessel, which was part of the plantâs HF alkylation unit. The crane dropped the convection section while it was above the HF vessel (Ryan 1988). The convec- tion section severed two lines attached to the HF vessel. Marathon reported that 53,200 pounds of HF were released over a 44-hour period. Shortly after the accident, Marathon directed a water spray at the HF cloud. As of November 16, 1987, local area hospital reported a total of 1,037 patients who were treated for HF exposure. Of these, 97 were hospitalized, two of whom were in critical condition. There were no fatalities. 85 square blocks and approximately 4,000 residents were evacuated. Reports of the incident do not indicate that the HF cloud hampered the activities of emergency responders. A.6 CHEMICAL DETECTION Chemical detection equipment (CDE) is an essential com- ponent of hazardous material (HAZMAT) emergency response. This equipment should detect the harmful agent, correctly identify the agent, and define the area of exposure. Rapid detection is essential so that responders and military targets can recognize a threat and don protective gear (ide- ally in 9 s). It also is important to know the extent of con- tamination. During several documented chemical attacks, first responder casualties have been vast enough to delay the rescue. During the Tokyo subway sarin attack in 1995, 9% of emergency medical services (EMS) providers suffered the affects of acute exposure. Effective CDE may help prevent these occurrences. Several different technologies are used today to detect chemical agents (CAs). CAs are defined as chemicals intended to kill or seriously injure human beings. CDE usu- ally detects the most common CAs: nerve agents, blister agents, and arsenical vesicants. A large variety of equipment is available that is capable of identifying liquid droplets of CAs on surfaces and in vapors. Laboratory-based equipment can detect agents in water. The main challenges with these technologies are ensuring an appropriate sample for analysis and filtering out nonhazardous environmental chemicals that may be present. This article focuses on the technologies and devices that may be used by first responder teams in the field. Laboratory detection techniques are beyond the scope of this discussion. Chemical Detection Paper Chemical detection paper is a very sensitive technique for detecting CAs. It is one of the least sophisticated and thus least expensive methods of detection. It is used to detect liquids and aerosols and is a common means for defining a contaminated area. Chemical detection paper is composed of 2 dyes soluble in CAs and a pH indicator integrated into cellulose fibers. A-13 When exposed to CAs, it can change color according to the type of agent. If an aerosolized droplet encounters the paper, the diameter and density of the spot can be used to determine the droplet size of the agent and the degree of contamination. Chemical detection paper lacks specificity and is prone to error because it reacts with contaminants such as brake fluid, antifreeze, and insect repellent, resulting in false-positive readings. False readings are especially undesirable in civilian situations because they may lead to mass panic. Therefore, chemical detection paper should always be used with another modality for accuracy of detection. M8/M9 chemical detection paper M8 and M9 CA detection papers, commonly used by the military, are available commercially to HAZMAT response teams. M8 paper is packaged in 25 perforated sheets, 2.5 in by 4 in, and is blotted on liquids that arouse suspicion. It iden- tifies CAs by changing colors within 30 seconds of exposure: dark green for vesicants, yellow for nerve agents, and red for blister agents. M9 paper has adhesive backing that allows it to be attached to clothing and equipment. M9 paper detects the same agents as M8 paper but does not change color to enable identifica- tion. M9 paper tends to react faster than M8 paper and can be attached to vehicles that are entering areas filled with vapor to determine contamination. Vehicles thus equipped are limited to a speed of 30 km/h. M256A1 chemical agent detection kit The M256 CA detector kit originally was released in 1978 and was modified in 1987 to the M256A1, which is sensitive to lower concentrations of nerve agents. It was used exten- sively during the Gulf War but also is available commercially. It is another common component of CDE provided to civilian response teams. This portable kit detects nerve gas, mustard gas, and cyanide and usually is used to define areas of conta- mination. The M256A1 contains a package of M8 paper, detailed instructions, and a vapor sampler (12 enzymatic tick- ets that contain laboratory filter paper for detecting CA vapors). The vapor sampler employs wet chemistry technol- ogy, in which ampoules containing different substrates are crushed so that the liquids interact with strips of filter paper, chromatographic media, and glass fiber filter. These substrates then are exposed to the vapor under suspicion. The reaction causes a color change, alerting the user to the presence of a CA. The reactions typically take 15 minutes to occur. The M256A1 can detect nerve gas concentrations of 0.005 mg/m3, hydrogen cyanide concentrations of 11mg/m3, and mustard gas concentrations of 0.02 mg/m3. This is one of the militaryâs most sensitive devices for detecting CAs and detects all agents at levels below those that can kill or injure people. It is prone to false-positive results, similar to other enzymatic detection techniques, but has not been demon- strated to produce false-negative results in real situations.
Colorimetric tubes Colorimetric tubes such as those available from Draeger and RAE systems use enzymatic techniques to identify CAs. A hand pump is used to draw a sample into a specific tube, and the concentration of the substance is read from the tube. This is another simple and inexpensive way of detecting and identifying a CA. It is used extensively in civilian response units for this reason, but it has some disadvantages. Avail- able are 160 substance-specific reagent tubes identifying dif- ferent agents. For each agent, a different tube must be used. Efficient use of this system demands knowledge of which CA is likely to be present in a given environment. If a tube for vesicants is used to sample the air and the CA is a nerve agent, the tube reports a false-negative result. A tube for each possible CA must be used for thorough detection. Ion Mobility Spectroscopy Ion mobility spectroscopy (IMS) is used in many handheld and stand-alone detection devices that can be used to scan equipment, surfaces, and people for contamination. This tech- nology involves drawing a gaseous sample into a reaction chamber using an air pump. The air molecules then are ionized, most commonly using radioactive beta emitters such as nickel- 63 or americium-241. The ionized particles then are passed through a weak electrical field toward an ion detector. Conta- minants are identified according to the time it takes to traverse the distance to the detector. This time is proportional to the mass of the molecule. The pattern is compared to a sample of clean air; if the pattern is markedly different and unique to cer- tain types of agents, the alarm sounds. These systems are capa- ble of detecting and distinguishing between nerve gas, mustard gas, and vesicants. Its sensitivity ranges from 0.03 mg/m3 for nerve gases such as sarin to 0.1 mg/m3 for mustard gas. IMS has certain advantages. It is less sensitive to contami- nants, because it relies on a clean air sample for calibration. Thus, if an area has a certain baseline nonhazardous environ- mental vapor present, it is not detected. Stand-alone detectorsâM8A1, Automatic Chemical Agent Alarm, and Fixed Site/Remote Chemical Agent Detector Many stand-alone detectors also use IMS technology. The military employs the M8A1 detector that consists of a stand- alone detector, which continuously monitors the environment for hazardous vapors and aerosols, and up to 5 alarms that can be dispersed throughout an area. The M8A1 detects nerve agents and blister agents when the concentration is 0.1 mg/m3 or greater and alarms within 1â2 minutes. M8A1 is an ideal device for protection from off-target attacks, in which a vapor is released upwind from the targets. However, it is less effec- tive for on-target attacks, in which the CA is released in large amounts within seconds. In this situation, the alarm sounds after the personnel have been exposed. This system was used during the Gulf War and has been upgraded to the Automatic Chemical Agent Alarm (ACAA) system. The ACAA is A-14 slightly larger and has a communications interface that is use- ful in combat. ETG provides a commercial version of an IMS stand- alone detector called the Fixed Site/Remote Chemical Agent Detector. This system detects and identifies nerve and blister agents and offers superior reliability from interference. The alarm information can be transmitted via radio, satellite, or hardwiring. This system can be useful if placed in hospital wards or at victim collection sites to detect contamination Infrared Detection Infrared radiation (IR) is employed in several CA detec- tors, including long-range detectors and point detectors. IR can be used to excite molecules, and each agent has a unique infrared pattern referred to as a fingerprint. Several different detection techniques use IR, including photoacoustic infrared spectroscopy, filter-based infrared spectroscopy, forward-looking infrared spectroscopy (FLIR), and Fourier transform spectroscopy. These include photoacoustic infrared spectroscopy, filter-based infrared spectroscopy, differential absorption light detection and ranging, and pas- sive infrared detection The military uses the M21 Remote Sensing Chemical Agent Alarm (RSCAAL) based on passive infrared detec- tion. It is the first fielded standoff chemical detection device. This system can detect a vapor cloud from 5 km with an 87% detection rate. The M21 RSCAAL continuously monitors a background and notes the change in spectral information if a vapor cloud obstructs the background. It automatically scans along a 60Â° angle, allowing the operator to monitor horizon- tal movement. The M21 can be set up in 10 minutes and is unaffected by low light conditions. However, the M21 is lim- ited in that it must be stationary and can be obstructed by snow and rain. CDE in Civilian response to Terrorist Attacks CDE technology has advanced primarily as a result of mili- tary necessity. More recently, the need for civilian preparedness for terrorist attacks with CA has been recognized. Civilian response is different from military response in many ways, and the choice of CDE must take this into account. Key differences include the following: â¢ Civilian responders tend to be less experienced in chem- ical attacks. â¢ Civilian responders have less information concerning the origin and type of attack and may not recognize that it is a CA attack initially. â¢ Civilian responders have more stringent budget restraints and thus must use cost-effective equipment. â¢ Civilian responders have less latitude in incorrectly identifying a CA. â¢ Civilian responders are deployed primarily to provide medical care, leaving detection as a secondary goal.
In the civilian setting, EMS or other medical providers are the first to arrive. Most EMS providers do not carry CDE to detect CAs and thus initially must recognize the potential threat in order to notify specialized HAZMAT response teams. These teams exist in many cities and are at a minimum equipped with pH paper and combustible gas indicators. This equipment is inadequate in identifying most CAs. Other teams now are equipped with colorimetric tubes Colorimetric tubes are much less expensive than more technical devices, such as the ICAM, and can be distributed generally. Major cities in the US have a Metropolitan Medical Strike Response System (MMRS) orga- nized by the Public Health Service. These are highly special- ized, fully equipped, deployable teams to combat civilian threats from weapons of mass destruction. They are primarily medical providers who provide EMS services, decontamina- tion, detection, and treatment. The first such team was orga- nized in 1995 in Washington, DC, and a second was organized for the 1996 Olympics in Atlanta. MMRS teams are often bet- ter equipped to respond to CA attacks than HAZMAT response teams. Even so, wide variability exists in the type of detection devices used. A recent study by the National Guard recognized that no standards regulate the detection devices among differ- ent civilian emergency response units. MMRS teams can employ any of the devices and technologies described above. They commonly use inexpensive CDE such as SAW detectors and enzymatic techniques such as M9 paper and the M256 kit. Some teams also use IMS devices such as the APD2000 and a modified ICAM for domestic preparedness. A.7 REFERENCES FOR APPENDIX A Berkowitz, B.J. et al. (1972). Superviolence: The Civil Threat of Mass Destruction Weapons, ADCON (Advanced Concepts Research) Corporation, Report A72-034-10, Santa Barbara, CA. Browning, J.B., Retired Vice President, Health, Safety, and Environ- mental Programs, Union Carbide Corporation (1993). Union Car- bide: Disaster at Bhopal. Available at: http://www.bhopal.com/ pdfs/browning.pdf. Reprinted from Crisis response: Inside Stories on Managing Under Siege, Edited by J.A. Gottschalk, Visible Ink Press, a Division of Gale Research, Detroit, MI. Carter, M.R., M.P. Howard, N. Owens, D. Register, J. Kennedy, K. Pecheux, and A. Newton (2002). Effects of Catastrophic Events on Transportation System Management and Operations, Howard Street Tunnel Fire, Baltimore City, Maryland, July 18, 2001 Final Report: Findings, prepared for U.S. Department of Transportation Intelligent Systems Joint Program Office by Science Applications International Corporation, McLean, VA. CCPS (Center for Chemical Process Safety) 1996. Guidelines for Use of Vapor Cloud Dispersion Models, Second Edition, ISBN 0-8169-0702-1, American Institute of Chemical Engineers, New York, NY. Cordesman, A.H. (1996). Terrorism and the Threat From Weapons of Mass Destruction in the Middle East â The Problem of Paradigm Shifts (Working Draft), Canadian Security Intelligence Service Middle East Studies Program. Available at http://www.csis.org/ mideast/reports/terror.html. A-15 Davis, G. and Kelen, P. (2001). CBRNE â Chemical Detection Equip- ment, emedicine, http://www.emedicine.com/emerg/topic924.htm EPA (U.S. Environmental Protection Agency) 1999. Risk Man- agement Program Guidance for Offsite Consequence Model- ing Analysis, EPA 550-B-99-009, Washington, D.C. Available at http://yosemite.epa.gov/oswer/ceppoweb.nsf/vwResources- ByFilename/oca-all.PDF/$File/oca-all.PDF EPA (U.S. Environmental Protection Agency), 2004. RMP*Comp Modeling Program for Risk Management Plans. Available at http:// yosemite.epa.gov/oswer/ceppoweb.nsf/content/rmp-comp.htm. Fryer, L.S., and G.D. Kaiser (1979). DENZ â A Computer Program for the Calculation of the Dispersion of Dense Toxic or Explosive Gases in the Atmosphere, United Kingdom Atomic Energy Author- ity Report SRD R 152, Culcheth, Cheshire, UK. GAO (U.S. General Accounting Office) 2003a. Rail Safety and Security: Some Actions Already Taken to Enhance Rail Security, but Risk-Based Plan Required, GAO-03-435, Washington, D.C. Goldwire, Jr., H.C., T.G. McRae, G.W. Johnson, D.L. Hipple, R.P. Koopman, J.W. McLure, L.K. Morris, and R.T. Cedarwall, (1985). Desert Tortoise Series Data report â 1983 Pressurized Ammonia Spills, Lawrence Livermore National Laboratories Report UCID- 20562, Livermore, CA. Haddock, S.R. and R.J. Williams (1978). The Density of an Ammonia Cloud in the Early Stages of its Atmospheric Dispersion, United Kingdom Atomic Energy Authority Safety and Reliability Direc- torate Report SRD R103, Culcheth, Warrington, Cheshire, UK. Kaiser, G.D. (1979). Examples of the Successful Application of a Simple Model for the Atmospheric Dispersion of Dense, Cold Vapors to the Accidental Release of Anhydrous Ammonia from Pressurized Containers, United Kingdom Atomic Energy Author- ity Safety and Reliability Directorate Report SRD 150, Culcheth, Warrington, Cheshire, UK. Kaiser, G.D. (1989). A Review of Models for Predicting the Disper- sion of Ammonia in the Atmosphere, Plant/Operations Progress Vol. 8, No. 1, pp 58â64. Kaiser, G.D., J.D. Price and J. Urdaneta (1999). Technical Back- ground Document for Offsite Consequence Analysis for Anhydrous Ammonia, Aqueous Ammonia, Chlorine, and Sulfur Dioxide, pre- pared for U.S. Environmental Protection Agency, Washington, D.C. Available at http://yosemite.epa.gov/oswer/ceppoweb.nsf/ vwResourcesByFilename/backup.pdf/$File/backup.pdf. Kalelkar, A.S. (1988). Investigation of Large-Magnitude Incidents: Bhopal as a Case Study, Presented at The Institution of Chemical Engineers Conference On Preventing Major Chemical Accidents, May 1988: Arthur D. Little, Inc., Cambridge, MA. Available at: http://www.bhopal.com/pdfs/casestdy.pdf Kupperman, R.H. and J. Kamen (1989). Final Warning: Averting Disaster in the New Age of Terrorism, Doubleday, New York, NY. Kupperman, R.H. and D.M. Trent (1979). Terrorism: Threat, Real- ity, Response, Stanford, CA: Hoover Institution Press. Mainichi Online (June 18 2001). Survey: Subway Sarin Attack Haunts More Survivorsâ Available at: http://www12.mainichi.co.jp/news/ mdn/search-news/896300/aum-60-71.html Markham, R.S. (1986). A Review of Damages from Ammonia Spills, presented at the 1986 Symposium on Safety in Ammonia Plants and Related Facilities, Boston, MA. American Institute of Chem- ical Engineers, New York, NY. Mullen, R.K. (1978). Mass Destruction and Terrorism, Journal of International Affairs 32:1 (SpringâSummer 1978), pp. 62â89.
McMullen, G. (1976). A Review of the 11th May Ammonia Truck Acci- dent, City of Houston Health department Report (unnumbered). NTSB (U.S. National Transportation and Safety Board), 1971. Rail- road Accident Report: Chicago, Burlington, and Quincy Train 64 and Train 824 Derailment and Collision with Tank Car Explo- sion, Crete Nebraska, February 18, 1969. NTSB-RAR-71-2, Washington, D.C. NTSB (U.S. National Transportation and Safety Board), 1978. Rail- road Accident Report: Louisville and Nashville Railroad Company Freight Train Derailment and Puncture of Anhydrous Ammonia Tank Cars at Pensacola, Florida, November 9, 1977. NTSB-RAR- 78-4, Washington, D.C. Ohbu, Sadayoshi, Akira Yamashina, Nobukatsu Takasu, Tatsuo Yamaguchi, Tetsuo Murai, Kanzoh Nakano, Yukio Matsui, Ryuzo Mikami, Kenji Sakurai, and Shigeaki Hinohara (June 1997). Sarin Poisoning on Tokyo Subway, Southern Medical Journal, Birming- ham, AL. Available at http://www.sma.org/smj/97june3.htm. Purver, Ron (1995). Chemical and Biological Terrorism: The Threat According to The Open Literature, Canadian Security Intelligence Service Unclassified report (updated November 2000). A-16 Radig, S.A. (2003) Catastrophic Anhydrous Ammonia Release, Minot, North Dakota, Scott A. Radig, North Dakota Department of Health, Division of Water Quality. Raj, P.K., J. Hagopian, and A.S. Kalelkar (1974). Predictions of Hazardous Spills of Anhydrous Ammonia Onto Water, CG-D-74- 74, prepared for the U.S. Coastguard by Arthur D. Little, Inc., Cambridge, MA. Robinson, J.P. (1967). Chemical Warfare, Science Journal (Lon- don), Vol.3 No.4. Ryan, R.M., U.S. Environmental Protection Agency (1988). OSC Report to the National Response Team: Major Air Release of Hydrofluoric Acid, Marathon Petroleum Company, Texas City, Galveston County, Texas â October 30 to November 1, 1987, Dallas, TX. SCIEX Inc. (1980). Mississauga Derailment â Ground Level Con- centration Measurements of Chlorine and Detection of Other Compounds, Report Prepared for the Ontario Ministry of the Environment, MOE Report ARB-TDA 05-80, Thornhill, Ontario, Canada.