Single-family housing inventory projection method for natural hazard risk modeling applications

Today’s regional natural hazards loss models rarely incorporate changes in a region’s built environment over time, and thus likely misestimate a region’s natural hazard risk. Of the existing natural hazard loss models that incorporate changes in the built environment, none are developed at an adequately granular spatiotemporal scale that is appropriate for regional (multi-county) natural hazards loss modeling. This work presents the new Housing Inventory Projection (HIP) method for estimating changes in a region’s housing inventory for natural hazards loss modeling purposes. The method includes two modules: (1) the Regional Annual County-Level Housing module, which estimates the annual number of housing units per county over a multi-county region and multi-decadal projection period, and (2) the Single-family Location Estimation module, which estimates the likely location of future single-family housing units across a subcounty grid space. While the HIP method can be applied over a range of spatiotemporal scales, we present a case study that estimates the number of single-family houses per 1 km 2 grid cell in the state of North Carolina for each year from 2020 to 2049. We then used these projections to estimate how a future housing stock would experience a Hurricane Florence-type event. Future housing projections suggest that between 2020 and 2049, nearly 2900 new houses will be built, each year, in areas that experienced at least two feet of flooding following Hurricane Florence.