Before an outbreak is detected, researchers and system builders use dispersion models to simulate outbreaks to understand the most likely tactics of a terrorist or military opponent. They also generate simulated outbreak data for use in measuring the outbreak detection performance of algorithms and systems.
Terrorists or military adversaries who wish to create an outbreak with maximal impact are likely to release a biological agent from locations or under weather conditions that maximize the lethality of the attack. Simulating releases from various locations and under various weather conditions can identify scenarios with maximal morbidity and mortality. Planners and system builders can then direct resources at preventing, detecting, and preparing responses to these scenarios. For example, one could design detection algorithms with heightened sensitivity to these scenarios. The BioWatch program conducts analyses about where to place air samplers to maximize the probability of detecting an aerosol release of biological agent. Atmospheric dispersion modeling is a critical component of these analyses.
A number of researchers have used the Gaussian plume model to simulate windborne outbreaks of inhalational anthrax. For example, Buckeridge et al. describe a model of anthrax attacks that generates simulated office visit data (Buckeridge et al., 2004). It uses the Gaussian plume model to simulate the windborne spread of B. anthracis spores. Evaluators can add the simulated outbreak data to historical baseline data and measure the false alarm rate, sensitivity, and timeliness of outbreak detection (Chapter 20 discusses evaluating algorithms).
Was this article helpful?