Australia’s first comprehensive severe thunderstorm model launched

Catastrophe risk modeling firm AIR Worldwide has released the industry’s first comprehensive severe thunderstorm model for Australia that explicitly captures all three sub-perils-hail, tornado, and straight-line wind to help companies assess and manage severe thunderstorm risk.

“In Australia, insurance losses from severe thunderstorms are greater than those from other natural perils such as earthquakes, tropical cyclones, bushfires, or floods,” said Dr Eric Robinson, manager and principal scientist, AIR Worldwide. “Because aggregate losses from severe thunderstorms can result in extreme volatility in financial results, a robust view of the risk is critical for organisations developing resilience strategies.”

The AIR Severe Thunderstorm Model for Australia simulates daily severe thunderstorm activity based on historical occurrence rates and local and seasonal weather patterns. The daily simulation enables the model to capture both the large outbreaks that produce insured losses in excess of A$10 million (US$8 million)—the Insurance Council of Australia (ICA) threshold for a catastrophe—and smaller events that may last only one day, but that could still impact a company’s portfolio on an aggregate basis, or a more rural portfolio on an occurrence basis.

Thunderstorm weather systems can last for several days and affect multiple states, but the individual tornadoes, hail swaths, and straight-line wind swaths (the “sub-perils”) that make up an outbreak may last for just minutes and affect highly localised areas. To capture the localised effects, AIR has developed high-resolution event footprints specific to each sub-peril. Additionally, because hailstorms, tornadoes, and straight-line windstorms inflict damage differently, the model’s damage functions are sub-peril-specific to provide more accurate loss estimates.

The model also simulates realistically clustered severe thunderstorm outbreaks using methodology that groups hail, wind, and tornadoes into spatially coherent patterns—patterns that would not be possible using random sampling alone.