Association for Public Policy Analysis & Management

All localities endure occasional extreme weather events such as floods or droughts, but some areas also may experience intermediate- to long-term term variations in weather patterns where precipitation, temperature, or storms may deviate sharply from long-term normals. The variations may include exceptionally busy hurricane seasons, long-term droughts, El Niño or La Niña events, or other extreme seasonal weather. Anticipation of such seasonal anomalies holds promise for improving flood, drought, and hurricane preparation, and some local officials have effectively used seasonal climate forecasts to improve emergency management practice.  However, prior research suggests that even when signals are clear and extreme events appear strongly correlated with seasonal climate signals, the application of climate information to improve local disaster planning poses difficulties.  

With support from the NSF and NOAA, we have conducted a multi-year, mixed methods study using focus groups, simulation exercises, and a survey to examine climate forecast use in emergency planning and management.  We emphasize the ways in which local emergency managers (EMs) make decisions with uncertain forecasts and whether these decisions differ when forecasts are presented differently.  Our principal hypothesis is that these public officials apply a variety of heuristics, or decision shortcuts, in using information to decide on a course of action for an uncertain future.  While a large literature in risk and decision analysis has emerged on the influence of these heuristics and the potentially perverse outcomes that result from their application in decision making of private individuals, much less has appeared on the role of the heuristics in public decision making under uncertainty, the target of our work.  Drawing on a sample of the general population, we also explore whether the difficulties that these EMs face in decision making under uncertainty differ from those of the average person on the street.  

In this paper, we present the results of a national-level survey of county emergency managers engaged in flood and drought planning and emergency management. In addition to collecting background information on respondents (e.g., gender, professional experience, educational background, etc.), the survey questionnaire presents eight climate and flood emergency management scenarios to investigate the role of several decision heuristics (e.g., numeracy, anchoring, and prospect theory) in influencing the choice of actions.  Each of the eight scenario has four treatments, but each of the respondents sees only one of these treatments.  As noted above, we also have imposed the same structure in a survey of the general population.  To test for differences across different treatments and with respect to EMs vs. the general population, we employ ANOVA analysis.

In addition to contributing to understanding of decision-making processes, our project will develop general practical mechanisms to improve risk communication and the incorporation of scientific information in flood and drought planning, mitigation, and management.  Moreover, although centered on climate variability, the work applies to a much larger class of problems associated with decision making under uncertainty.