Tropical Cyclones and Thunderstorms

An afternoon with the Willis Research Network

November 23, 2017

By Geoffrey Saville

On 1 November, we held our second London-based Willis Research Network (WRN) seminar of the year. After the last seminar in February, which focussed on seismic and volcanic risk, our latest meeting was a chance to showcase our latest WRN research into weather and climate related risks, following on from last year’s similar event. Our ongoing research partnerships have given us plenty of inspiration to build a relevant agenda.

Relevant research topics

This past year has seen a high number natural catastrophes, including a sequence of major hurricanes affecting the U.S. coastline, leading to some significant losses in the reinsurance market. Hurricanes Harvey, Irma and Maria (referred to as HIM in the market) accounted for much of this loss.

Tropical Cyclone related flooding

Given that Harvey was a prodigious rain-maker, causing severe flooding over a large part of Texas and neighbouring states, we chose to highlight the research we are supporting at the Wharton Risk Management and Decision Processes Center, through Dr Jeffrey Czajkowski and Dr Gina Tonn. This work investigates the National Flood Insurance Program (NFIP) claims database in the U.S. and allows us to understand the differences between coastal and inland flooding driven by Tropical Cyclones, and the character of the spread of losses in different types of storms. For more information you can view the slides presented here.

Footprinting a storm

We are also supporting a project which develops a new database of tropical cyclone peak wind footprints using a sophisticated method developed in partnership between the WRN Fellows at the National Center for Atmospheric Research, Dr James Done and Dr Greg Holland and Willis Re. These footprints are intended for comparison and validation of existing footprints used in the market. They provide an open and credible alternative to currently available products, and early evaluations suggest that they can improve our understanding of the distribution of the strongest winds during a storm’s passage. This data set is also global, which allows a consistent view of tropical cyclone impacts in multiple territories. Slides from this section of the seminar can be found here.

Making a thunderstorm

Changing tack a little, the remaining part of the seminar looked at new applications of climate models from Columbia University. Our WRN partners, Dr Michael Tippett and team, have been developing a new seasonal forecast methodology to identify areas that are likely to have a higher number of extreme tornados or hail events over the next month. Forecast issues at the beginning of the month can provide skilful forecasts for up to 4 weeks ahead. The predictability is based largely on the how the climate naturally varies according to the El Niño-Southern Oscillation (basically the influence of warmer or cooler than average sea surface temperatures in the tropical Pacific Ocean). This relationship can be captured by the climate models used by Dr Tippett, and expressed as an index combining the ingredients required to make severe thunderstorms which can bring losses associated with tornados and hail events. These slides can be viewed here, and more description can be found in a previous blog here.

These are always exciting events and attract industry-wide interest. Our visiting academics presented their work and fielded an insightful array of questions from the audience. Whether you attended or not, the WRN or your Willis Re contacts would be happy to discuss this research if it could be of use to your business. We will soon be uploading some video highlights from the event on our WRN page, but in the meantime feel free to browse our hot off the press 2017 WRN brochure, which highlights some of our recent and key research projects.


Related content

Latest Natural Catastrophe Updates from the Willis Towers Watson Wire blog