Modeling Emergency Warning Systems for Disaster Reduction

Abstract:

Emergency warning systems are the last line of defense for reducing losses from natural hazards. Well-designed systems enable populations-at-risk in impending disasters to take appropriate protective actions. There have been a limited number of research efforts designed to quantify the performance of various types of warning systems and these have been primarily technology-specific. This paper describes the result of research on developing models of warning systems based of the 3 stages of warning: deciding to issue a warning, disseminating the warning to the population-at-risk and initiating a protective action. The warning issuance stage is the time between emergency warning officials being notified of the existence of a threat and reaching a decision to activate the warning system. The warning dissemination stage is the time between that decision and individuals receiving the first warning. The protective action stage is the time between receiving the warning and initiating a protective action. A fourth stage consisting of implementing the protective action is not discussed in this paper.

The basic method used in developing models of each stage of the warning process was to first review previous modeling research. Second, assemble historic data on the timing of the warning process. This included point data as well as cumulative distributions. Third, develop a modeling approach to specify a diffusion equation. Fourth, compare the empirical data to the results of the simulation. The attached figure depicts a comparision between simulated warning diffusion and empirical data

For each stage, best and worst case curves are develop. Several moderate or average curves are also developed. Factors associated with identifying which curves would be appropriate for a large range of potential event scenarios are identified and discussed. Potential application to lose of life modeling and hazard mitigation strategies are also discussed.