Skyfora updates AI-powered tropical cyclone model and achieves ILS adoption

Skyfora, a Helsinki-based weather intelligence startup that uses artificial intelligence to derive tropical cyclone intensity forecasts, announced an updated model and also noted that it has integrated a related securities market client. insurance (ILS).

Skyfora believes its forecasting models can provide a new source of improved insight into the potential impacts of hurricanes on the insurance, reinsurance and insurance-related securities (ILS) market.

Today, Skyfora announced an updated version of its seasonal tropical storm forecast model which it says is proven to work, particularly for the important Gulf of Mexico region.

The model provides probabilistic forecasts for the probability of various events in different categories for the upcoming tropical cyclone season.

This covers tropical cyclone (TC) intensity, storm genesis by region and storm impact by region, while Skyfora also provides customers with early April, May and June forecasts for the Atlantic Basin, it publishes also mid-season forecasts in July, August and September, depending on the needs of its customers.

Skyfora’s model makes full use of Bayesian neural networks and probabilistic machine learning, with the model trained using multiple data sources, including historical atmospheric, land and ocean climate variables.

Skyfora claims that cross-validation experiments against the benchmark show that an average absolute prediction error for major hurricanes is reduced by 50% compared to the benchmark, and the prediction of the Gulf of Mexico landfall has a correlation of more than 0.7 with actual values ​​for the years 2011-2021. .

“Seasonal forecasts and particularly qualified landfall forecasts have always been difficult to develop and prove due to the chaotic nature of tropical cyclones,” explained Dr. Svante Henriksson, Founder and CEO of Skyfora. “Last year we published our first seasonal forecasts, but since then we have shifted focus from traditional point predictions and decided to deploy Bayesian deep learning instead, as the probability distribution that result is more useful for our Reinsurance, Catastrophe Modeling and ILS clients.. It turned out to be the right decision.

The company is new to modeling, but it is already gaining traction for some in the area of ​​insurance-linked securities (ILS).

New technological approaches to weather and climate modeling are very important to ILS managers and Skyfora’s seasonal tropical storm prediction model is already in use by investment manager ILS Securis Investment Partners in London.

“The Skyfora team is ambitious and has provided a fresh approach to a very difficult problem. By working closely together, we are able to focus work on the most impactful metrics and the most relevant business issues. By monitoring the evolution of forecasts and risks during the season, we hope that Skyfora will help us manage our risks more accurately and more dynamically, explained Dr. Paul Wilson, Head of Non-Life Analytics. at Securis.

As new technologies become more readily available, scientific inputs into ILS investment decision-making will become richer and more ubiquitous, with next-generation risk modeling techniques constituting a key development for the ILS market. .

Read also : Skyfora launches AI-powered Tropical Storm Tracker intensity forecasts.

Printable, PDF and email version

About Opal Jones

Check Also

Population of endangered Hawaiian monk seals highest in 2 decades

HONOLULU— The population of endangered Hawaiian monk seals has surpassed a level not seen in …