Reducing the wildfire protection gap

Stephen Lathrope, global head of insurance at ICEYE, highlights the need to plug the data gap as insurers contend with increasing frequency and severity of wildfire events

With climate change impacting the frequency and severity of wildfire events worldwide, the unpredictability of how and when wildfires ignite and spread is becoming a multi-billion dollar problem – and a risk that insurance carriers and insureds alike are struggling to contend with.

In some of the worst-affected regions there is a growing protection gap. Insurance cover in wildfire-exposed regions has experienced significant shifts in premiums and deductibles, with terms becoming increasingly restrictive, while in California temporary bans have been introduced to prevent insurers from withdrawing cover from highly exposed areas.

Meanwhile, insurers that are willing to provide affirmative coverage are struggling to quantify potential losses from wildfire events, despite the growing availability of risk models, making it harder to price coverage and manage their own exposures effectively.

Between 1980 and 2021, the US alone had 20 catastrophic wildfires that caused more than $1bn in damage each, 16 of which occurred in the past 20 years.

California, which has experienced significant habitat and property loss from wildfires in recent years, has suffered 13 of its 20 most destructive wildfires to date (in terms of structures destroyed) in the five-year period from October 2017 to October 2022, according to the California Department of Forestry and Fire Protection (Cal Fire).

Of these, the three most costly in insurance terms all occurred within that same five-year period – the 2018 Camp Fire (which cost an estimated $10.75bn), the 2017 Tubbs Fire ($9.56bn), and the 2018 Woolsey Fire ($4.52bn), according to Aon.

Plugging the data gap

When assessing the extent of property damage caused by a wildfire event, insurers have typically either relied upon on-site inspections, often conducted weeks post event, or needed to scan the available aerial imagery and annotate the images themselves to determine which individual properties have been impacted – a laborious and time-consuming process, and one that impacts insurers’ reaction time when responding to the needs of their customers and other stakeholders.

Recent developments in satellite technology and AI capabilities now offer insurers a completely new, near real-time perspective in wildfire events, allowing access to building-level impact data for any US wildfire event as it is developing.

In particular, advances in synthetic-aperture radar (SAR) satellite imagery and data analytics and machine learning will mean that insurers can measure the extent of wildfire damage in a matter of hours, rather than days, for any built-up area within the wildfire impact area.

The development of SAR satellite constellations now means it is possible to monitor events on the ground irrespective of atmospheric conditions with high-resolution imagery of building damage captured through clouds, at night, and – most importantly for wildfire events – through smoke.

This gives insurers the ability to monitor and target wildfire events that have a high probability of becoming a significant loss. A further advantage of 24/7 satellite monitoring is that users can have eyes-on for all active wildfire events in the US at any given time, enabling them to target those wildfires likely to impact policyholders, before initiating in-depth analysis of a specific event.

When a wildfire event exceeds a given threshold (based on the size of the fire and its proximity to a built-up area), data capture of satellite imagery can be initiated. This imagery can be overlaid with building footprints and analysed to map the locations of buildings which have been destroyed and those which are undamaged within the wildfire zone.

Monitoring of the event concludes when the fire is not progressing any further or threatening any other adjacent population centres, and hence considered to be contained.

Improved outcomes for insurers and clients

The resulting hazard information can be made available to stakeholders, including insurers and government agencies, within 24 hours of the initial wildfire warning.

For insurers, this takes out the guesswork involved in assessing news reports and aerial optical imagery data. Crucially, it also provides proactive communication of wildfire hazard data as the event is unfolding. This enables insurers to more efficiently manage event response activities and, post event, improve claims processing and speed up payment to policyholders.

ICEYE has recently announced the beta release of its Wildfire Insights product, the first of its kind to provide building-level impact data for a wildfire event in near real time based on satellite imagery and a combination of advanced analytics and machine learning. The product captures data for any wildfire that impacts communities throughout the US.

As part of the beta phase, ICEYE has invited select insurers to participate in the final testing of the product functionality prior to the planned general availability release in Q3 2023.

This approach is also being validated during this phase using building damage data from Cal Fire, which is typically released some weeks after an event. By overlaying the Cal Fire damage assessment with our own output we have been able to demonstrate 80-90 percent accuracy for events monitored using the Wildfire Insights product.

The product will initially be rolled out across the US, with the ability to capture every wildfire catastrophe in the country, with plans to extend the geographic scope of the product to include other wildfire-prone regions.

Access to accurate, granular hazard data as an event is developing will help to make the modelling of wildfire risks more consistent, enabling more accurate and stable pricing of insurable risks, and ultimately reducing the insurance gap for this growing peril.