Improved Estimation of Hail Damage Integrating Radar and Ground Truth Data

Radar Maximum Estimated Size of Hail (MESH) provides a high spatiotemporal resolution of hail storm intensity. However, analysis of major hail storms shows that MESH tends to underestimate maximum observed hail size at the ground level. The purpose of this study is to develop an algorithm that combines ground truth hail size measurements, utilizing Understory’s network of hail sensors, with radar MESH to estimate the spatiotemporal distribution of hail stone size. The study incorporates an analysis of over 24,000 home insurer claims across 15 major hailstorms in North America from 2017-2018. Models relating each hail size data source to claim probability, across nearly half a million policy holders, are evaluated. Out of sample validation shows that the model with mean hail stone size has the greatest predictive performance, with superior calibration and discrimination as compared to the radar MESH model. Hail damage is among the costliest weather related events. The number and severity of hail claims has been increasing at a growing rate over the past decade. By taking into account ground-truth measurements of hail stone diameters, we are able to provide a spatially detailed and accurate model of the distribution of hail stone sizes, helping to identify at-risk areas with greater precision.