Login Register Subscribe
Current Issue

Help

BI’s Article search uses Boolean search capabilities. If you are not familiar with these principles, here are some quick tips.

To search specifically for more than one word, put the search term in quotation marks. For example, “workers compensation”. This will limit your search to that combination of words.

To search for a combination of terms, use quotations and the & symbol. For example, “hurricane” & “loss”.

Catastrophe modelers respond to growing wildfire threat

Reprints

As insured losses from wildfires continue to grow in California and elsewhere, demand for better information and modeling is rising, and catastrophe modelers are responding with new and updated products.

Insured losses from wildfires have topped $1 billion thus far in 2018, according to Aon Benfield, after totaling about $14 billion in insured losses in 2017.

“In general, there is increasing market demand for models that produce reliable estimates for the higher-frequency, lower-severity events, and most of these are generated by weather-related perils, such as hail, localized windstorms and wildfires,” said Karen Clark, president and CEO of Boston-based catastrophe modeler Karen Clark & Co.

“The sizable fires that occurred in 2017 and earlier have driven demand,” said Peter Fallon, senior vice president at brokerage Risk Strategies Co. Inc., in an email. “Demand is also being driven by the concern that with global climate changes, there could be an increase in the frequency and severity of wildfires.”

KCC scientists and engineers are working on a wildfire model, she said. “We’re currently in the first phase of developing the intensity footprints and producing accurate loss estimates for significant historical events,” with the initial focus on the United States.

“The demand is incredible — and fully justified by carrier loss experience,” said Chris Folkman, senior director of product management, models and data at Risk Management Services Inc. in Newark, California. “Wildfire is now a major catastrophe concern, where multibillion-dollar event losses are not uncommon.”

“As a result of large-scale fires in California and Oregon, parts of western Canada and Australia, the insurance industry has been looking for better tools to help them understand how wildfires turn into major catastrophic events, underwrite and price their exposed risks more appropriately, and manage significant portfolios over large geographic areas more effectively,” Mr. Fallon said.

AIR and RMS, two “best in class” modeling organizations, are amongst those who are helping the industry estimate potential losses at the policy and portfolio level, Mr. Fallon added.

RMS has a wildfire model under active development that will focus on the contiguous U.S. and Canada, according to Mr. Folkman.

AIR Worldwide recently released an updated wildfire model for the western U.S., though it had been in development even prior to the 2017 and 2018 California wildfires.

“We’ve been working on this model release for about three and a half to four years,” said Tammy Viggato, senior scientist with AIR Worldwide in Boston.

AIR did have a model for California only that was released in 2008, so the model was due for an update, she said. “We really had seen the need to expand beyond just California,” Ms. Viggato said. “There have been a lot of wildfires beyond California,” she said, citing events in Colorado and Texas.

“We had started talking to clients, and there were some fires that hit some of these insurers pretty strongly,” Ms. Viggato said. “So there was already discussion about expanding the model beyond California.”

Capabilities have increased over time, she said.

“From 2008 to 2018, there is a lot more we can do computationally, and we have increased access to data sources such as satellite imagery and weather data,” Ms. Viggato said.

Data is a key element of modeling performance, Mr. Folkman said.

“A wildfire model is only as good as the exposure data that it is fed,” he said. “The more comprehensive and accurate the input data, the better the result. The industry still has to overcome data collection challenges for wildfire modeling — just like it did with hurricanes and earthquakes over the last few decades.”

One of the main areas of focus for the new AIR model is the wildland-urban interface, “where human development meets and begins intermingling with wildlands, so we have a change in fuels and additional suppression efforts affecting the behavior of the fires,” Ms. Viggato said.

“We’ve been seeing a significant increase in the population that’s living there,” Ms. Viggato said of the interface. “The fires we saw in 2017 and the ones earlier this year, these are areas which historically have wildfire hazard, butpeople haven’t always lived there in the numbers that they are now.”

“As urban sprawl expands into areas where there is undeveloped terrain, such as forests and grasslands, there is the expectation that more risks will be exposed to wildfires and bushfires,” Mr. Fallon said.

Efforts are underway to bring regulation to some models associated with fires, according to California’s insurance commissioner.

The California Department of Insurance has recommended that lawmakers enact legislation to require the filing and approval of fire risk score models used by insurers, Insurance Commissioner David Jones said.

“We’ve recommended that the legislature require fire risk score models be filed with the Department of Insurance because we have discovered that many of the models do not take into account significant measures taken by homeowners and communities at the recommendation of fire officials to reduce the risk fire damage,” Mr. Jones said as he surveyed some of the fires raging in Northern California.