Example Of Using An Atmospheric Dispersion Model To Prove Wind As Route Of Outbreak Transmission

Meselson et al. (1994) demonstrated—using the Gaussian plume model of atmospheric dispersion as a critical component of their analysis—that the 1979 outbreak of anthrax in Sverdlovsk, Soviet Union, was due to an aerosol release of B. anthracis spores from a military microbiology facility (Meselson et al., 1994). Furthermore, they determined the time of the release of spores to within 2.5 hours (Guillemin, 2000) and the quantity of the release (Meselson, 2001, Meselson et al., 1994).Their analysis, when combined with the following two facts established that the route of transmission of this outbreak was the wind. During the course of their investigation, Russian president Boris Yeltsin admitted officially that the outbreak was related to military production of biological weapons (The Record, 1992). Prior to their study, analysis of autopsy specimens determined that the victims suffered from inhalational anthrax and not gastrointestinal anthrax as was the official Soviet explanation for over a decade (Abramova et al., 1993).

The analytic procedure that Meselson et al. (1994) used follows. After first obtaining as complete a list of cases as possible, they interviewed surviving victims and friends and relatives of deceased victims. They determined the whereabouts of each

Handbook of Biosurveillance ISBN 0-12-369378-0

Elsevier Inc. All rights reserved.

case in the days prior to the onset of illness and dates of symptom onset, entry into the healthcare system, and death. Of 77 victims identified, they were able to determine the home and work locations for 66. They mapped the daytime locations during the week of April 2,1979 for 57 victims (Figure 19.1). The area encompassing the 57 locations was a narrow, 4-km zone extending south-southeast from a Soviet military microbiology facility. Of the nine patients whose daytime whereabouts were unknown, three lived inside this zone and another three, although they lived and worked outside the zone, had occupations (truck driver, pipe layer, and telephone worker) that might have taken them inside the zone.

They also identified and mapped animal populations that experienced anthrax outbreaks at the same time. They identified

Anthrax Plume Modeling

figure 19.1 The high-risk zone for humans for the Sverdlovsk outbreak of inhalational anthrax in 1979. The ellipse-like lines are lines of equal concentration produced by the Gaussian plume model. The numbers are the daytime locations of victims. (From Meselson, M., Guillemin, J., Hugh-Jones, M., et al. (1994). The Sverdlovsk anthrax outbreak of 1979. Science 266:1202-8, with permission.)

figure 19.1 The high-risk zone for humans for the Sverdlovsk outbreak of inhalational anthrax in 1979. The ellipse-like lines are lines of equal concentration produced by the Gaussian plume model. The numbers are the daytime locations of victims. (From Meselson, M., Guillemin, J., Hugh-Jones, M., et al. (1994). The Sverdlovsk anthrax outbreak of 1979. Science 266:1202-8, with permission.)

six villages that experienced outbreaks of anthrax in animals at the same time that humans were ill in Sverdlovsk. On April 2, 1979, these villages lay downwind of Sverdlovsk along the axis (when extended to 50 km) of the high-risk zone for humans (Figure 19.2). The compass bearing of the centerline of human and animal anthrax outbreaks was 330° ± 10°.1

They obtained historical weather data from the Koltsovo airport (located 10 km east of the high-risk zone). Records of surface observations every three hours showed that the only time the wind direction was consistent with both the known locations of victims and the compass bearing of 330° was April 2, when the wind direction ranged from 320° to 350° from 4:00 am to 5:00 pm local time (Figure 19.3). The date of April 2 and the time delay to onset of illness in the victims were consistent with the incubation period of inhalational anthrax.

Finally, Meselson et al. (1994) used the Gaussian plume model in combination with data about breathing rates to show that the meteorological data and spatial distribution of cases were entirely consistent with an aerosol release of spores from the military microbiology facility as the cause of the outbreak. The ellipse-like lines in Figures 19.1 and 19.2 show contours of constant dosage (airborne concentration times the breathing rate) computed from the Gaussian plume model with the microbiology facility as the release location.

60"30h 61J0Q'

60"30h 61J0Q'

Plume Modeling Anthrax
figure 19.2 The high-risk zone for animals, which is an extension of the high-risk zone for humans to 50 km along the same axis. (From Meselson, M., Guillemin, J., Hugh-Jones, M., et al. (1994). The Sverdlovsk anthrax outbreak of 1979. Science 266:1202-8, with permission.)

1 The meteorological convention for wind directions is to report the direction from which the wind is blowing, and 0°/360° is due North.

Anthrax Plume Modeling

figure 19.3 Weather data from April 2-4,1979. Wind directions are reported as the direction from which the wind was blowing. Only on April 2 is the wind direction in alignment with the daytime locations of victims. Reprinted with permission from (From Meselson, M., Guillemin, J., Hugh-Jones, M., et al. (1994). The Sverdlovsk anthrax outbreak of 1979. Science 266:1202-8, with permission.)

figure 19.3 Weather data from April 2-4,1979. Wind directions are reported as the direction from which the wind was blowing. Only on April 2 is the wind direction in alignment with the daytime locations of victims. Reprinted with permission from (From Meselson, M., Guillemin, J., Hugh-Jones, M., et al. (1994). The Sverdlovsk anthrax outbreak of 1979. Science 266:1202-8, with permission.)

Using the Gaussian plume model, Meselson (2001) estimated the number of spores released. Given the data available about the victims, he calculated that the release amount ranged from a few milligrams to a gram, depending on assumptions about the LD502 of anthrax spores and a difficult-to-estimate meteorological parameter used by many atmospheric dispersion models, the atmospheric stability class (Meselson, 2001). This quantity of release is consistent with one explanation for the outbreak put forward by at least two former Soviet officials: workers in the military microbiology facility mismanaged a ventilation system with subsequent discharge of B. anthracis spores into the air outside the facility (Guillemin, 2000).

In summary, Meselson reasoned from (1) the spatial and temporal distribution of cases of anthrax and (2) meteorological conditions present in Sverdlovsk prior to the outbreak, to the fact that an aerosol release had occurred and the location, time, and quantity of release that caused the outbreak.

Meselson's analysis, however, took months to complete, and therefore would not provide timely results for the prospective detection and characterization of an outbreak. In the following sections, we provide background about atmospheric dispersion models and then a method for automating Meselson's analysis to enable prospective detection of outbreaks.

0 0

Post a comment