Evaluating the Risk and Complexity of Pluvial Flood Damage in the U.S.

Abstract

Flood hazard in the U.S. is typically separated into two categories: storm surge and inland. Inland flooding is frequently represented as fluvial, stemming from overflow of rivers and streams. However, significant inland flood risk also exists in the form of pluvial flooding, resulting from ponding of water, and often occurring due to insufficient drainage in urban areas. Pluvial flood risk can be complex to model and is often not shown on floodplain maps, leading to significant uninsured risk. While pluvial and fluvial flooding are sometimes lumped together for purposes of modeling, insurance, and risk management, we hypothesize that key differences exist in damage associated with these types of inland flooding. We investigate and quantify these differences through study of three datasets: (i) Tropical cyclone (TC) flood insurance claims; (ii) non-Tropical Cyclone (non-TC) flood insurance claims; and (iii) Hurricane Harvey flood insurance claims in the Houston area. From our data analyses, we find that fluvial claims involve greater damage on average and are more likely to occur within the mapped Special Flood Hazard Area (i.e., 100-year floodplain) than pluvial claims. However, pluvial flooding is more frequent overall being the cause of 74% of TC and 60% of non-TC flood insurance claims in the U.S. TC floods have a higher percentage of pluvial flood insurance claims and higher average pluvial damage than non-TC floods. Analysis of Hurricane Harvey claims shows that zip codes with higher-than-expected losses tend to have high percentages of pluvial claims. Pluvial flooding and its local nuances exhibit the potential to reduce the accuracy of catastrophe models of expected flood damage.