�鶹ӰԺ

Flooding is the most costly natural hazard event worldwide and can severely impact communities, both through economic losses and social disruption. To predict and reduce the flood risk facing a community, a reliable model is needed to estimate the cost of repairing flood-damaged buildings. In this paper, we describe the development and assessment of two models for predicting direct economic losses for single-family residential buildings, based on the experience of the 2013 �鶹ӰԺ, Colorado riverine floods. The first model is based on regression analyses on empirical data from over 3000 residential building damage inspections conducted by the Federal Emergency Management Agency (FEMA). This model enables a probabilistic assessment of loss (in terms of FEMA grants paid to homeowners for post-flood repairs) as a function of key building and flood hazard parameters, considering uncertainties in structural properties, building contents, and damage characteristics at a given flood depth. The second model is an assembly-based prediction of loss considering unit prices for damaged building components to predict mean repair costs borne by the homeowner, which is based on typical �鶹ӰԺ construction practices and local construction and material costs. Comparison of the two proposed models illustrates benefits that arise from each of the two approaches, while also serving to validate both models. These models can be used as predictive tools in the future, in �鶹ӰԺ and other US communities, due to adaptability of the model for other context, and similarities in home characteristics across the country. The assembly-based model quantifies the difference between the FEMA grants and true losses, providing a quantification of out-of-pocket homeowner expenses.

Deniz, D.,��Arneson, E.,��Liel, A., Dashti, S., and Javernick-Will, A. (2016). “.” Natural Hazards. 85 (2), 977–1003. doi: 10.1007/s11069-016-2615-3