Association for Public Policy Analysis & Management

The purpose of this study is to test the hypothesis that more accurate and timely National Weather Service (NWS) storm notifications lead to better risk assessments by the public and local governments and hence improve decision-making regarding damage mitigation and avoidance. The hypothesis will be tested in the context of a machine learning model that predicts the economic and safety damages caused by severe weather based on the characteristics of weather events and exposure areas.

A diverse set of data sets from several federal agencies will be merged to gain new insights into the relationships between storm events, NWS storm notifications, and economic and safety outcomes. A new, unique data set (Unplanned K-12 School Closures in the United States, 2011-2016 which was collected by the Centers for Disease Control and Protection) will be used to help estimate lost economic output from winter storms. These data are matched to Department of Education Common Core Data to obtain school characteristics that would be used to estimate direct costs of unscheduled school closures. The resulting data set is matched by location and date to NWS data on storm warnings and to other National Oceanic and Atmospheric Administration (NOAA) data on actual storm events which allows, among other measures, school closings to be categorized by whether they coincide with winter storm warnings and/or verified events. (Preliminary work suggests that one-third of winter storm-related school closings in 2014-15 were in areas and periods without winter storm warnings.) In turn, these data would be matched to the Census Bureau’s American Community Survey demographic data and Longitudinal Employer-Household Dynamics data to estimate potential indirect costs (i.e., the numbers of working parents of school-age children and employees at business establishments in school closure areas).

Other economic damages data as well as injury and fatality data from the Storm Events Database of NOAA’s National Centers for Environmental Information Storm Events database and from the National Highway Traffic Safety Administration’s Fatality Analysis Reporting System will also be used to round out the picture of storm-related damages. Later work will extend the analysis to other storm events including convective storms and related weather hazards.