A day-night population exchange model for better exposure and consequence management assessments

Abstract

Emergency response in the event of a hazardous chemical or biological material release in an urban area needs to occur on two time scales: 1) real time response to a known event and 2) delayed response to latent events which are not expressed at the point of release. The latter is particularly important in the case of chemical or biological material releases which may go undetected until patients begin showing up in hospitals due to the delayed pathogenic or toxic response to the organism or chemical in question. In order to reconstruct such undetected events, trace exposed individuals and provide adequate treatment and prophylaxis to quell any burgeoning epidemic, emergency responders need to be able to estimate who was exposed at the time of event and where they may expect to see them enter the public health system – i.e., determine which hospital they will eventually arrive at. Definition of expected hospital arrivals of new cases requires being able to estimate migratory patterns of urban citizens under normal and emergent conditions. There are many simulation systems capable of the simulation of urban mobility patterns. While these systems are capable of high spatial and temporal resolution urban mobility simulation, they often require significant computing resources and are difficult to use in a near real time environment. In this research, we present a model that fuses US Census Bureau population mobility data with a raster based day-night population data model in a decision support system for fast turnaround analysis of public health emergencies. This approach is an improvement over our prior work in which we created static representations of day-night, indoor-outdoor population for the continental USA and Hawaii at 250 m grid resolution (McPherson and Brown, 2003; McPherson and Brown, 2004). Our previous static data models were developed to improve exposure assessments for hazardous airborne contaminant releases and were akin to other research by Lo (2001), Yuan et al. (1997), Dobson et al. (2000), Harvey (2002), and Langford and Harvey (2001). In this research, we use our static population models with data on population dynamics to support consequence assessment and emergency response. To construct such a model, we decompose human population mobility patterns into three primary migratory forcing functions: the journey to work, the associated journey to home, and the journey to the hospital. The former two migratory patterns are the two major patterns that most adults conduct throughout their lives. The latter defines where affected people may enter the public health system.