Swiss Re Insurance-Linked Fund Management

Original Risk: A Society for Change Agents

U.S. tornado forecasting tool shows promise


Meteorologists have long studied the unpredictability and devastation caused by tornado outbreaks in the U.S. in an effort to gain a greater understanding into the predictability, severity and potential location of events.

Supercell and tornado severe convective weatherAnd now, new research by Dr. Victor Gensini has shown some promising signs of predicting favourable tornado conditions up to three weeks in advance, a tool that could prove valuable to insurance, reinsurance, and ILS market participants.

Meteorology professor at the College of DuPage, Illinois, U.S., Dr. Victor Gensini has conducted research that shows strong potential in its ability to predict favourable conditions for severe weather events, including tornadoes, across the U.S. two to three weeks in advance.

Via numerous media reports, seminar discussions, television appearances and conference speeches over the last two months or so, Gensini has emphasised that while the new approach is showing some promising results, the ability to tell exactly when a tornado is going to hit and where is very unlikely.

However, the advanced prediction tornado work undertaken by Gensini does appear to be able to predict when weather conditions are favourable for tornado activity across the U.S., based on primarily on jet stream patterns.

Catastrophe modelling for a range of perils is improving all the time as insurance, reinsurance, ILS players and risk managers continue to work to improve their understanding of peril regions and potential exposures.

Speaking to online publication, Gensini explained how the tool could be beneficial to the risk transfer industry.

“Who would really be interested in a product like this would be an insurance company looking to buy catastrophe bonds for a big set-up that could influence a large chunk of the country and the emergency managers who no matter where you live, educate people and bring awareness,” said Gensini.

The ability to warn communities in advance that conditions are favourable for tornado outbreaks will assist with evacuations and help to mitigate loss of lives.

For insurers, reinsurers, and ILS markets having a long-range view of potential outbreaks could help with more accurate expected loss and pricing for tornado exposures, mitigating any overexposure from unexpected events.

“We are never going to be able to say right a tornado is going to hit your house, or hit your community 2-3 weeks from now because of the small scale of these events. But what we can say is we’re moving into a regime and our weather pattern, our jet streams that is going to be either favourable or not favourable,” said Gensini addressing an audience at a Fermlab Tornado and Severe Storms Seminar in early April 2016.

Gensini explains that jet stream patterns across the Pacific is what he’s been exploring to determine an ability to long-range predict tornado activity or likelihood.

Specifically, methodology concerning High Angular Momentum (AAM), which is a measure of how fast wind speeds are, where the jet stream travels in a relatively straight line from East to West, is unfavourable for tornado activity in the U.S. explains Gensini.

However, when this jet stream, which occurs at roughly the same height as commercial airliners fly, (approximately 33,000 feet) can be seen to be wavy (Low AAM), so the pattern still moves East to West but with Northerly and Southerly peaks and troughs, Gensini’s research shows that conditions become much more favourable for severe weather activity across the U.S.

The research involved looking at the pattern of the jet stream two to three weeks before a recorded event to determine whether there is any correlation between differing weather regimes, so Low and High AAM, and the frequency of tornado outbreaks.

And the results so far are somewhat promising.

Gensini examined the largest U.S. tornado outbreaks of the last 50 years, which revealed that each occurred during a transitional weather pattern phase (specifically a transition out of El Nino), similar to what the U.S. is expected to experience soon.

As Stressed by Gensini the tool is still in its infancy and further work is required to tackle uncertainties surrounding the formation of tornadoes across the U.S., but the fact that he’s discovered some correlation between jet stream activity and tornado outbreaks is a positive step.

Insurers, reinsurers, and ILS markets utilise a variety of modelling tools to attain their own view of the risk, so any advances or new approaches that show accuracy will be welcomed for risk assessment and ultimately the pricing of transactions.

The small scale of tornado events means it’s very unlikely that there will be a time when the exact location and time of a tornado outbreak is possible to predict, but knowing when conditions are favourable 2-3 weeks in advance is certainly an improvement on current tornado models.

It will be interesting to see how accurate the tool is at predicting tornadoes in the coming months, with the U.S. tornado season running April through June, although outbreaks can occur anytime.

Should the tool prove accurate in predicting favourable conditions then it could become a more common feature of the risk transfer market in the future.

The goal, says Gensini, is to one day be able to say that we are entering an average, above average, or below average tornado probability, similar to what the NOAA issues for other peril regions.

Any improvement in forecasting tornado and severe convective storm activity in the U.S. will be welcomed by insurers, reinsurers, ILS managers and ILS investors alike. The ability to foresee when inclement weather is coming could assist with hedging and investment decisions, portfolio management and even underwriting.

As granularity, availability and accuracy of weather forecast products improves it could also stimulate the need for more short-term weather hedging products. Sometimes it may be more efficient to buy a short-term weather hedge, rather than shift out of positions in the portfolio. Better access to data could also stimulate the need for risk transfer exchanges as well, again as short-term risk transfer becomes more necessary.

Whether this research could result in a better view of tornado risk is not yet certain, but any advancements that help to clarify the threat of tornadoes weeks out will certainly be welcomed. It may be especially welcomed given the loss-heavy start to the 2016 convective storm season, with already $6.17 billion of insurance and reinsurance losses attributed to the peril.

Register today for ILS NYC 2023, our next insurance-linked securities (ILS) market conference. Held in New York City, February 10th, 2023.

Artemis ILS NYC 2023 - Insurance-linked securities conference in New York City

Get a ticket soon to ensure you can attend. Secure your place at the event here!

Print Friendly, PDF & Email

Artemis Newsletters and Email Alerts

Receive a regular weekly email newsletter update containing all the top news stories, deals and event information

  • This field is for validation purposes and should be left unchanged.

Receive alert notifications by email for every article from Artemis as it gets published.