Hail Storm Risk Assessment Using Space-Borne Remote Sensing Observations and Reanalysis Data

Much of the world is impacted by severe thunderstorms, but whether they become disasters depends upon resilience--our capacity to prepare, mitigate, respond, and recover. Hail is the costliest severe weather hazard for the insurance industry, generating ~70% of severe convective storm losses in 2017 due to damage to assets such as homes, businesses, agriculture, and infrastructure. Most insurance companies do not reserve enough capital to cover catastrophes, so they acquire reinsurance. The reinsurance industry uses catastrophe models (CatModels) to statistically estimate risk to an insurer’s portfolio.

Hail CatModels are developed with climatologies that define hailstorm frequency and severity. Hail-prone areas can be defined using hail reports from trained spotters, the media, and the general public. Extremely severe hail (2+ inch diameter) occurs nearly every day across the world. Weather radars can detect hail because hailstones strongly reflect microwave signals that they emit. However, hail climatologies are difficult to derive because hail covers small areas and there are neither hail reporting mechanisms (e.g. website or mobile app) nor radar networks in most places outside the US and Europe. This lack of ground truth on severe hail puts society and economies at risk.

Hail is generated within storms by strong updrafts. These updrafts exhibit unique signatures in NASA and other agency satellite observations, offering new opportunities for hailstorm analysis. Geostationary (GEO) visible and infrared imagery has been collected for 10-20+ years across the world (region dependent) and methods have been developed at NASA Langley Research Center (LaRC) to detect hailstorm updrafts using GEO imagery. Climatological GEO updraft data has been used by Willis Towers Watson (WTW), a leader in catastrophe risk assessment for the insurance industry, to develop CatModels over Europe and Australia. Hail can also be inferred with passive microwave imagery collected by low-Earth-orbiting sensors such as the GPM GMI, TRMM TMI, AMSR-E, AMSR-2, SSMI, and SSMIS over the last 20+ years. Hailstorms generate enhanced lightning flash rates that can be tracked using new GOES-R series GEO Lightning Mapping (GLM) imagery. Atmospheric reanalysis can be used to define favorable hailstorm environments.

This presentation will describe a framework for developing continental to global hail climatologies and CatModels based on NASA satellite data and capabilities. This is a collaboration between LaRC and Marshall Space Flight Center (MSFC), WTW, and partners in Brazil and Argentina. This project seeks to mitigate hail disasters over South America by aiding development of new satellite-based severe storm nowcasting tools by regional partners and developing climatologies to improve societal understanding of hail frequency. Work with WTW will improve socioeconomic resilience through development of new CatModels. Southern Brazil, Uruguay, Paraguay, and Argentina feature some of the most intense thunderstorms on Earth. South America is a developing insurance market of interest to WTW clients, and is similar to other regions routinely impacted by hail that do not have comprehensive hail reporting or radars to assess hailstorm frequency.

Historical LEO and GEO imagery and reanalysis data will be used to detect hailstorms and estimate their severity. GLM data will be used to further improve hailstorm analyses. We will mature land surface imaging satellite methods for identifying hail damage to aid disaster mapping. Few satellite-based severe storm nowcasting methods have been developed to serve South America. The project team will demonstrate nowcasting capabilities with NASA datasets and provide training to regional partners to help them adopt these methods to improve warnings and mitigate disasters. Datasets will be made available via the LaRC Atmospheric Science Data Center (ASDC) Geospatial Platform and NASA Disasters Portal, which can provide data in multiple formats for use in a wide range of open source and commercial tools.