The new normal in hurricane modeling

Imagine a coastline devoid of human life, while a few miles inland entire communities are under water. Dystopian, yes, but this is what could happen if we underestimate hurricane risk. Without adequate insurance coverage recovery is difficult, even impossible.

We've seen it happen time and again, albeit on a smaller scale. When a hurricane strikes, residents without insurance or with insufficient coverage have to abandon their homes or cut corners and build back to a lower standard. As a result the local economy suffers and the community is more vulnerable to the next major weather event.

How big is the issue? According to CoreLogic, nearly 33 million homes on or near the Atlantic and Gulf coasts are at risk of hurricane-force wind damage and the estimated cost to replace them is USD 10.5 trillion. The impact of secondary perils like flooding continues to grow, and it would cost more than $2.3 trillion to replace the estimated 7.8 million homes that are vulnerable to storm surge.¹ Many of those homes are uninsured or underinsured, and this protection gap grows larger as storms become more extreme. By failing to make enough use of the forward-looking elements in modeling, the industry would only exacerbate the problem.

The last few years have seen increased hurricane activity in the Atlantic. In 2021 the World Meteorological Organization used all of its 21 names for hurricanes for the second year in a row and just the third time in history. Meteorologists are doubling down on their predictions of another above average hurricane season this year following several years of increased activity and losses.

And impacts from storms are more severe. When Andrew struck 30 years ago, a USD 20 billion loss event had never occurred before. However, in just the past five years we've had four hurricanes with insured losses above USD 20 billion each: Ida, Harvey, Irma and Maria.

This could very well become the new normal, and our industry needs to look at hurricanes through a different, evolved perspective. At Swiss Re we're changing our focus – from relying less on historical experience to adopting more of a forward-looking view that accounts for current conditions and climate change. Our North Atlantic Tropical Cyclone (TCNA) model continuously undergoes revisions and updates to incorporate new learnings, improved scientific understanding and additional loss experience.

The new and improved version presents a forward-looking view of tropical cyclone risk based on recent events and our projections. Those considerations include:

• The current state of the climate: We have observed a 30% and 60% higher occurrence of hurricanes (cat 1-5) and major hurricanes (cat 3-5) respectively over the past 25 years compared to the long-term average.
• Increased risk of rainfall: We are now explicitly modeling flooding from tropical cyclone-induced rainfall, which has been calibrated to a forward-looking view of rainfall probabilities and intensities rather than long term historic rainfall averages.
• Land use/urbanization: A significant amount of time was spent studying changes in exposure due to urban sprawl. The presence of more buildings tends to reduce wind speeds while soil sealing from building projects leads to increased flooding.

At Swiss Re, our mission is to make the world more resilient so we have an obligation to lead on this issue. We're putting more weight on the present and future thanks to improved predictive tools and methods. As a result, our hurricane model is a better representation of the scientific community's view of the future state of the climate combined with learnings from recent events.

Sources:

¹ CoreLogic Report: Nearly 33 Million and 7.8 Million U.S. Properties at risk of Hurricane-force Wind and Storm Surge Damage, respectively - CoreLogic®