2
The Chemical/Biological Threat to Air Transportation
The focus of the antiterrorist efforts of the U.S. air transportation system to date has been on the detection of concealed arms or explosives; essentially no capability exists to detect chemical or biological warfare agents effectively and affordably or to mitigate the impact of a terrorist attack involving these agents. Thus, a terrorist runs little risk of being caught or discovered before perpetrating an attack, and the number of people exposed to the agent would depend only on how effectively the perpetrator could disseminate it.
This chapter describes examples of the chemical/biological threat agents that might be used in a terrorist attack on the U.S. air transportation system, as well as some key physical, chemical, and biological characteristics of these agents that would affect the number of people exposed to such an attack. Some plausible scenarios for the release of these agents into the air transportation environment are also discussed.
CHEMICAL/BIOLOGICAL THREAT AGENTS
Many types of threat agents might be used in an attack on the air transportation system. Each has its own set of physical, chemical, and biological characteristics, which would determine how lethal and widely dispersed the effects would be. Four different categories of threat agents can be distinguished:1
-
Fast-acting chemical agents. Individuals exposed to these agents begin displaying symptoms within seconds or minutes. Examples of such agents include the neurotoxic agent sarin, the choking agent chlorine, and the blood agent hydrogen cyanide.
-
Delayed-acting chemical agents and biological toxins. With agents in this category, which includes some chemicals as well as large-molecular-weight toxins produced by certain biological organisms, exposed individuals would not begin exhibiting symptoms for hours or days. Examples of such chemicals include sulfur mustard; biological toxins include ricin and botulinum toxin.
-
Slow-acting, noncontagious biological agents. These agents, which include viruses as well as the bacterial causative agents for anthrax and tularemia, produce no initial symptoms, but cause flu-like symptoms after a few days or weeks. Some, such as Bacillus anthracis, can be disseminated either in the form of spores or as vegetative cells.
-
Slow-acting, contagious biological agents. This category of agents, which includes the virus that causes smallpox and the bacterium that causes pneumonic plague, produces no initial symptoms upon infection but typically causes flu-like symptoms after a few days or weeks. Infected individuals are usually contagious after they are symptomatic.
Table 2-1 lists examples of agents in each of these categories, along with some of their characteristics. In the longer term, biological warfare agents may also be genetically modified to further amplify their lethality, mask their identity, protect them from vaccinates, and protect them from environmental degradation, including the effects of ultraviolet (UV) irradiation, temperature, and humidity.
Fast-Acting Versus Slow-Acting Agents
One common characteristic of many chemical agents is that they tend to be relatively fast acting: that is, victims begin to exhibit symptoms of distress within seconds to minutes after exposure to the agent. This almost-immediate showing of symptoms has implications for defensive strategies based on detection systems, since the chemical agent released in an attack would reach and produce a response
TABLE 2-1 Examples of Chemical and Biological Threat Agents of Concern
Category |
Example Agents |
Initial Symptoms |
Time to Symptoms\ |
Fast-acting chemical agents |
Sarin |
Convulsions, paralysis |
Seconds |
Phosgene |
Coughing, breathing difficulty |
Seconds |
|
Hydrogen cyanide |
Convulsions, respiratory failure |
Minutes |
|
BZ (3-quinuclidinyle benzilate) |
Delirium, hallucinations |
1 hour |
|
Delayed-acting chemical agents and biological toxins |
Sulfur mustard |
Blistering, redness, swelling |
2 to 24 hours |
Ricin toxin |
Breathing difficulty, fever, nausea |
6 to 8 hours |
|
Botulinum toxin |
Descending muscle weakness/paralysis |
18 to 36 hours |
|
Slow-acting, noncontagious biological agents |
Bacillus anthracis (anthrax) |
Flu-like symptoms |
<7 days |
Francisella tularensis (tularemia) |
Flu-like symptoms |
3 to 5 days |
|
Slow-acting, contagious biological agents |
Variola major (smallpox) |
Flu-like symptoms |
12 to 14 days |
Yersinia pestis (plague) |
Flu-like symptoms |
1 to 6 days |
|
Viral hemorrhagic fever (e.g., Ebola) |
Flu-like symptoms, internal bleeding |
2 to 21 days |
|
SOURCE: Centers for Disease Control and Prevention, online at http://www.bt.cdc.gov/agent/. Accessed October 6, 2005. |
from the detection system at about the same time that it began producing symptoms in the exposed population. Identification of the agent involved would still be valuable for forensic purposes, but detectors would not be necessary to establish the fact of the attack itself. The implications of fast-acting agents for defensive strategies are discussed in Chapter 3.
A common characteristic of biological agents is that they are slow acting: although exposure may occur rapidly, victims’ symptoms may not appear for several hours to several weeks, depending on the agent. Similarly, delayed-acting chemicals and biotoxins would produce no symptoms for several hours to several days. In that amount of time, airport passengers and workers would have dispersed to a wide variety of destinations. Thus, for attacks involving slow- or delayed-acting agents, technologies for early detection become more important—a detector alarm may be the only indicator for several days that an attack has taken place.
Factors Affecting the Potency of an Attack
The interplay between the chemical and biological properties of the threat agent, on the one hand, and the specific attack scenario, on the other, can influence the lethality of the attack. Table 2-2 shows the relative respiratory toxicities (expressed as the lethal concentration of toxin at which 50 percent of test animals are killed, or LCT50, in milligrams per minute per cubic meter) of a variety of toxic gases compared with chlorine gas, which was used as a chemical weapon in World War I. According to Table 2-2, the nerve agent sarin (GB) has a respiratory toxicity approximately 100 times that of chlorine, while sulfur mustard (HD) is about 7 times more toxic. However, the lethality of an attack
TABLE 2-2 Toxicities of Lethal Gases
with an agent depends not only on its inherent toxicity, but also on its chemical and physical characteristics—such as volatility and vapor density—which govern its dispersion in civilian spaces. Other factors affecting the lethality of an attack include the age and overall medical condition of the individuals subject to the attack.
The chemical and physical characteristics affect exposure levels, depending on the specific attack scenario. If, for example, the agent is released at ground level and victim exposure occurs by breathing the vapor while standing, the vola-
tility and density become critical factors, and higher-volatility chlorine becomes a slightly more potent agent than sarin (measured in terms of toxicity times volatility; see Figure 2-1) and much more potent than sulfur mustard, which has a relatively low volatility and high density. If, on the other hand, the agent is released above the victim and exposure occurs by breathing the vapor as it wafts down (Figure 2-2), the inherently higher toxicity of sarin and mustard gas compared with that of chlorine would become dominant in determining the potency of the attack. These considerations explain why the sarin attack on the Tokyo subway system in 1995—while horrific—did not cause more than 12 fatalities. The sarin, which has a relatively high vapor density but low volatility, was released on the floor of the subway car; had it been released from above, far more casualties would have resulted. Of course, in spaces where there are significant air currents, dispersion of the agent would occur primarily by convection rather than by diffusion, and factors such as vapor density become less important. Such air currents may occur in airport concourses where there may be external pressure differences and frequent opening and closing of doors.
In the case of biological agents, there is a range of possible delivery methods that may be used by terrorists (e.g., contamination of the water or food supply, spreading of the agent on surfaces that are touched frequently, and so on). The most effective method for infecting a large number of people in a short time, however, is likely to be that of releasing the agent into the air in the form of aerosol particles.
AIR TRANSPORTATION SPACES
In developing strategies for defending the U.S. air transportation system from chemical/biological attacks, it is important to understand the characteristics of the spaces and their human occupancy patterns. As examples, the committee considers two very different kinds of spaces: the airport terminal and the aircraft itself. Gaining an understanding of the physical configuration of these spaces, as well as entry and exit points and passenger flows, is important in deter-
mining vulnerabilities and potential choke points at which large numbers of people might gather and be exposed to attacks. Since the committee’s focus is on attacks involving releases of agents into the air, understanding the airflow patterns in these spaces is particularly important. By empirically studying and modeling the occupancy patterns, physical characteristics, and airflow patterns in these spaces, one can begin to understand how chemical/biological threat agents might be dispersed and what strategies might be most appropriate for defending against them. The discussion below is intended to identify some of the factors that may be important in the assembly of such models.
Airport Terminals
Airport terminals are typically large, open spaces with relatively uniform physical configurations that include, for example, ticketing/check-in areas, baggage claim areas, security checkpoints, concessions, restrooms, and departure gates. These large spaces require heating, ventilation, and air conditioning (HVAC) systems to maintain acceptable air quality. Some areas are open to the general public, and others are restricted to passengers, employees, and/or security personnel. Access to the various spaces is less well controlled in other transportation venues such as bus, railway, and shipping terminals, presumably because there has been no history of attacks against such facilities in the United States.
Terminals generally accommodate the large numbers of the traveling public that move through them with residence times of about 1 to 2 hours. There are also large numbers of visitors who are not traveling but are dropping off travelers at check-in areas or waiting to meet travelers in the baggage claim areas. Their stay at the facility would typically be at least 30 minutes, although such assumptions should be checked against actual data. In addition, there are a significant number of airline, concessions, ground transportation, and security personnel working within airport terminals around the clock. In passenger ticketing/check-in areas, there are numerous entry and exit points. Data and models are needed for flow behavior of people both during normal
airport operations and after being alerted to a potential hazard.
Most passengers are carrying at least one bag, if not more, and a very large number of checked bags are passing through the facility from check-in areas to aircraft. There is a constant movement of materials and supplies throughout all areas of the facility, and large amounts of cargo and shipping containers are also moved through on their way to the aircraft. From the point of view of a terrorist, these relatively invariant characteristics of airport terminals may suggest logical points of attack. Similarly, they can provide the basis for a rational design of a defense against such an attack.
Aircraft
The aircraft itself comprises space very different from that of the terminal. The cabin features a relatively small, confined space with a very high density of passengers, crew, and carry-on bags. The minimum passenger residence time on an aircraft is about 1 hour, with maximum times stretching to 14 hours for very long distance flights. Well-publicized penalties for deviant passenger behavior and for failure to obey crew instructions have conditioned passengers to behave in a more compliant manner than can be expected in terminals. Airflow is also better controlled in an aircraft.
Virtually all aircraft have the same basic physical configuration. The Environmental Control System (ECS) is crucial for maintaining air quality during a flight, and the vast majority of aircraft have similar localized airflow patterns (Figure 2-3). During a flight on a typical airliner, 50 percent of outside air is mixed with 50 percent filtered, recirculated air, with complete exchange of the cabin air volume every 2 to 3 minutes.2 Special consideration is given to air supplied to the cockpit.
Although there are few access points to an aircraft before a flight and although passenger access is carefully controlled, a significant number of airport personnel have access to the aircraft between flights. These include baggage handlers, cleaners, food service personnel, maintenance personnel, and refuelers. In addition, while the aircraft is on the ground, it is connected to an external HVAC system. These various factors suggest that aircraft face a significant vulnerability to chemical/biological attacks while they are on the ground.
ATTACK SCENARIOS
The committee was not given specific attack scenarios to consider in this study, either in terms of the spaces attacked, the agents involved, or the manner of agent release. Accordingly, the discussion below is qualitative, intended to highlight some of the factors that need to be considered in forming appropriate defensive strategies against such attacks. As noted above, only an air release of agent is considered here.3
Point Release Versus Attack on the Heating, Ventilation, and Air Conditioning System
In a point release attack, the agent would be released in a plume from a single point (or perhaps several discrete points simultaneously). Individuals near the point of release would be exposed to agent at high levels, whereas those farther away would be exposed at lower doses. Initially, the spread of agent would be confined to the space in which it was
2 |
Information s available online at http://www.boeing.com/commercial/cabinair/. Accessed October 5, 2005. |
3 |
Examples of potential chemical attack scenarios are identified in a Federal Aviation Administration (FAA)-funded study undertaken at Johns Hopkins University: Chemical Sensing and Mitigation Options for Commercial Airliners, Final Report, STD-01-189, Laurel, Md.: Johns Hopkins University Applied Physics Laboratory, July 2001. |
originally released; subsequently, it might spread considerably via air drift down corridors and from level to level through mezzanines and stairwells. If HVAC systems continued to operate, the agent would be drawn into the ductwork and spread to connected rooms or spaces, albeit with considerable dilution.
If the agent were released directly into the inlet of an HVAC system that had no associated air treatment technologies, it would be pumped through the ductwork into all of the connected spaces, creating the potential for simultaneous, widespread exposures to relatively high concentrations of agent. Although this scenario is of particular concern owing to the potential for mass casualties, the confined nature of an HVAC system offers a number of possibilities for defense, as discussed in Chapter 3.
Aerosol particles in the 1- to 10-micrometer (µm) size range are removed from the air with fairly high efficiency (>50 percent) by common building HVAC air filters, and with very high efficiency (>99.97 percent) by the high-efficiency particulate air (HEPA) filters in the ECS of a modern airliner (Figure 2-4).4 Nevertheless, in an attack with a large quantity of aerosolized agent, these removal efficiencies might not be high enough to prevent widespread exposure and symptoms, particularly with highly infectious agents. Chemical agents are typically small organic molecules (not associated with particles) that would not be removed by common air-filtration systems. Air can be cleansed of organic chemical contaminants by passing it through a bed of highly adsorbent material such as activated carbon with appropriate additives, or other sorbents, and/or through an active (e.g., plasma) sterilization chamber.
Attacks Involving Fast- Versus Slow-Acting Agents
Some of the characteristics of fast- and slow-acting agents have been discussed above. An attack on the air transportation system with a fast-acting (chemical) agent would likely occur in an area in which large numbers of people were gathered, either in a terminal or in an aircraft. The fact of the attack would quickly become obvious as victims began to collapse or exhibit other symptoms of distress. It would be important for the authorities to recognize quickly that an attack had occurred in order to facilitate evacuation, get medical help for the victims, and limit the access of nonessential personnel to the contaminated area. If the attack occurred in an aircraft passenger cabin during a flight, the aircraft might be brought down (or taken over by terrorists) if the pilots became incapacitated. Similarly, attacks on critical nodes within an airport terminal (control rooms, emergency-response centers, and so on) could incapacitate key decision makers or deny authorities the use of these spaces, so as to prevent an effective response.
In contrast, in the absence of an effective detection or air-treatment system, an attack involving a slow-acting agent might go unrecognized for days, until the exposed victims began exhibiting symptoms of disease. Since the incubation periods for the appearance of symptoms of illness caused by slow-acting agents are typically long compared with the residence time of travelers in an airport terminal or aircraft, victims would be geographically dispersed by the time symptoms had appeared, and it might be difficult to locate them. If the disease were communicable, it could be spread by infected passengers to far-flung parts of the world in a very short time. An intentionally infected passenger (i.e., a suicide terrorist) traveling on a long flight could be one efficient means of infecting other passengers with a slow-acting biological agent. Unlike the situation created by a fast-acting agent, however, the problems caused by the release of a slow-acting agent in flight would not be an effective way of bringing down an aircraft. In the specialized case of very large releases of anthrax spores, the attacked aircraft could provide a means of infecting passengers over the course of multiple flights owing to the high survival capability of spore-forming bacteria.
Quantity and Rate of Release of Agent
The quantity of agent used in an attack and its rate of release are also factors that a terrorist, and thus a defender, must consider. In manufacturing, transporting, or preparing for the release of a large quantity of agent, the terrorist would risk being discovered before the attack could be perpetrated. Similarly, if a slow-acting agent were released at a high rate, there would be a higher probability that the attack would be observed or would cause a detector (if present) to alarm, since the ambient concentration would likely be well above the detector threshold. By comparison, a slow release rate (“trickle attack”) and a highly infectious agent might produce ambient agent levels that are below the threshold of available detection systems, yet sufficient to infect exposed individuals. This would be balanced against the increased time required to carry out the attack, with the associated increased risk of discovery.