Projecting Building Inventory Changes and the Effect on Hurricane Risk

Abstract

This paper introduces a quantitative methodology to model the way that the number, locations, and types of buildings in a region change over time, specifically in order to understand how those changes affect hurricane risk. The methodology is intended ultimately to be incorporated into regional risk-assessment models so they can more accurately estimate expected future hurricane losses. Current regional risk-assessment models assume that a region’s building inventory remains static over time. The methodology begins with population projections, transforms them into household projections, then housing-unit projections, and finally projections of the number of buildings with different structural characteristics. The final output is the number of buildings of a certain type in each census tract and each future year, where the building type is based on units in structure, year built, value, type of roof covering, predominant roof shape, and number of stories. A case study is presented for residential buildings in 15 counties in North and South Carolina. An uncertainty analysis examines the effect of modeling assumptions and data uncertainty on the results.