AIR Worldwide Releases Updated Severe Thunderstorm Model for the U.S.

Catastrophe risk modeling firm AIR Worldwide (AIR) announced that it has updated its Severe Thunderstorm Model for the United States. The comprehensive update to the model features significant enhancements to all three model components: hazard, engineering, and financial. The improvements are based on a decade’s worth of new data and scientific research, including damage data collected and analyzed by AIR scientists and engineers following major outbreaks in 2008, 2011, and 2013; approximately $3 billion in insurance company claims; and additional analysis of billions of dollars of claims data from AIR sister company Xactware.

“The culmination of this new data, research, and 15,000 person-hours of development time is a model that provides a more comprehensive view of U.S. severe thunderstorm risk,” said Scott Stransky, manager and principal scientist, AIR Worldwide. “The model update reflects a unique blend of statistical and physical modeling combined with the latest meteorological research to help overcome many of the limitations and reporting biases in the historical data. Such a model serves to provide companies with the high-resolution, detailed view of the risk needed to assess the impact of severe thunderstorm losses on their portfolios.”

The updated model underwent a peer review by world-renowned severe thunderstorm experts.

“The AIR model is realistic and represents the likely variability of intensity that will allow users to make better decisions about their exposure over a range of time horizons,” said Dr. Harold Brooks, senior scientist, National Severe Storms Laboratory. “AIR researchers have a very good understanding of the state of the scientific understanding and the uncertainties of the community’s knowledge about severe thunderstorm hazards and have utilized that knowledge in the model.”

“AIR has developed a comprehensive Severe Thunderstorm Model for the United States,” said Timothy Marshall, PE, meteorologist and principal engineer, Haag Engineering. “The model has damage functions that are based on the scientific relationship between building damage and wind speed/hail impact energy. For hail, the model includes both the vertical fall speed as well as the horizontal component of wind speed to calculate impact energy. This is especially useful when estimating the amount of damage to building exteriors, such as siding and windows.”

“To effectively capture the risk that severe thunderstorms pose to insured properties across the United States, a catastrophe model should differentiate the risk to various assets by subperil (tornadoes, hailstorms, and straight-line winds) and by primary and secondary building characteristics,” said Cagdas Kafali, assistant vice president and principal engineer, AIR Worldwide. “The AIR model explicitly accounts for not only primary building features, construction type, occupancy, number of stories, and year-built but also many secondary features, such as hail-resistance category of the roof cover and other roof and building envelope characteristics.”

To create a spatially complete catalog of simulated events, AIR “smart-smoothed” historical reports of tornadoes, wind, and hail from the Storm Prediction Center (SPC) to physically realistic locations. The smart-smoothing methodology combines statistical and physical methods that leverage high-resolution meteorological parameters to determine when and where conditions were favorable for severe thunderstorm formation. Smart-smoothing allows the model to better account for risk in rural areas where there may have been historical under-reporting and in areas that may not have experienced major activity in the brief historical record.

To produce a more complete picture of risk from the severe thunderstorm peril, the AIR model not only captures the large outbreaks that produce insured losses in excess of $25 million but also the smaller events that may last only a day and produce much lower losses – but still impact a company’s portfolio on an aggregate basis. In addition to the standard 10,000-year simulation, AIR will release 50,000- and 100,000-year stochastic simulations and a historical catalog containing several key recent events.

AIR also released today its Crop Hail Model for the United States, a probabilistic model that captures the effects of hail on insured crops, providing companies with a comprehensive view of their crop hail risk. The new Crop Hail Model uses the 10,000-year stochastic catalog from AIR’s Severe Thunderstorm Model for the United States, in which hailstorms are a modeled peril.

Because hail affects various crops differently, the AIR Crop Hail Model for the United States features crop-specific damage functions – for corn, soybeans, wheat, cotton, rice, and barley – that account for the unique damage mechanisms that hail imposes at various stages of each crop’s growth. Lastly, the new model has been extensively validated against loss estimates issued by the National Crop Insurance Services (NCIS), data provided by crop insurers, and published research.

The Severe Thunderstorm Model for the United States is available in Version 2.0 of the Touchstone and Version 16 of the CATRADER catastrophe risk management systems, and the Crop Hail Model for the United States is available in Version 16 of CATRADER.

AIR is a member of the Verisk Insurance Solutions group at Verisk Analytics.

Source: AIR Worldwide