Multi-source data’s critical role in responding to the LA wildfires

When wildfires erupted across Los Angeles on 7^th January, few could have predicted the catastrophic scale of destruction that would follow. Within just two weeks, these fires were being regarded as some of the most severe natural disasters in US history, with some forecasters estimating total damage and economic losses of $250bn+.

Despite California’s long-standing vulnerability to wildfires, the unique confluence of hurricane-force winds and arid conditions fuelled the rapid spread of the Palisades, Eaton, and Hurst Fires, tragically claiming more than 25 lives and leaving widespread devastation in their wake.

As the world watched in horror, geospatial specialists were working tirelessly to gather critical data to provide insurers with the intelligence to assess the damage, with timely insights in the crucial early stages of the disaster to help them to anticipate claims and the impact on their portfolios.

Navigating complex challenges with a multi-source data approach

Wildfires present unique challenges for data collection and analysis. The dynamic nature of the LA event, coupled with obstructive smoke plumes and rapidly changing conditions, hindered traditional assessment methods.

However, by employing a multi-source data strategy it was possible for industry loss assessment experts to overcome these obstacles and deliver actionable intelligence to insurers.

In the first 24 hours following the wildfire outbreak, specialists focused on integrating data from emergency response agencies, satellite-based sensors, and open-source reporting.

This initial phase resulted in an Exposure Layer, delineating fire perimeters and identifying observed hotspots across the affected areas. This real-time intelligence provided insurers with immediate situational awareness, enabling them to not only anticipate potential claims, but also allocate claims response resources more effectively.

Balancing speed and accuracy

While rapid response is critical in any catastrophic event, it is not always the sole determinant of success. The LA wildfires underscored the importance of balancing speed with precision.

Initially, the thick smoke plumes posed a significant challenge for high-resolution optical satellite imagery, a key component in building-level damage assessments.

For some insurers this was a major issue, as they had to wait for the heavy smoke to clear before understanding the impact of the fires. In response, Synthetic Aperture Radar (SAR) imagery was used by some Nat Cat intelligence specialists, and their insurer clients, to supplement data sources.

Although SAR can penetrate smoke and provide valuable insights, its interpretation remains complex, particularly in distinguishing building-level damage. Recognising the limitations of relying on SAR alone, building-level assessments were delayed until higher-quality optical data became available. This approach ensured the accuracy and reliability of assessments, ultimately benefiting insurers and their policyholders.

By January 10th, 40cm satellite imagery was available, enabling the first release of the Claims Layer within a 72-hour timeframe. This provided insurers with a broad understanding of the damage extent in affected areas, even though individual building-level details were not yet available.

Actionable intelligence

Much has been written about the scale of the LA fires as an industry event, with insurance loss estimates of between US$20bn and up to US$45bn. This information, while of interest, does not give insurers the data they need. The difference lies between modelled data and observed intelligence.

Prioritising high quality observed intelligence data ahead of speed proved invaluable. By 11^th January, additional high-resolution satellite imagery was available, enabling McKenzie Intelligence Services’ expert human analysts to provide 17,000 initial building-level damage assessments, offering insurers a comprehensive view of the impact on their portfolios.

Within a week we had completed assessments for nearly 57,000 unique properties across LA. This meticulous approach identified 12,362 buildings as destroyed, with an additional 17,500 properties at risk of significant internal damage. The accuracy of these findings enabled insurers to act swiftly and confidently, providing critical support to affected policyholders without unnecessary delays.

Lessons learned and the future of disaster response

The response to the LA wildfires highlights several key lessons for the insurance and intelligence sectors.

First, the importance of adopting a multi-source data strategy cannot be overstated. By integrating diverse data streams – ranging from satellite and SAR imagery to ground reports – the industry was able to benefit from a holistic view of the disaster’s impact, well in advance of single source data or on the ground only assessments.

Second, a measured approach that balances the precision brought by multi-data sources with human intelligence analysts leads to better long-term outcomes. ML/AI tools are not yet sufficient to handle the task alone, and rushing to conclusions based on incomplete or lower-quality data can result in costly errors and delays down the line. Experience shows the value of strategic patience in the face of complex challenges.

Finally, technological advancements such as AI-driven analytics continue to enhance our industry’s capabilities in responding to catastrophic events. However, the technology is yet to provide the sector with all of the actionable insights it needs for major Nat Cat events.

Conclusion

The recent LA wildfires served as a powerful reminder of the devastating impact natural disasters can have on both communities and economies. By leveraging a multi-source data approach, supported by expert human specialists, insurers were able to access timely, accurate intelligence that enabled them to make well-informed decisions when they mattered most.

These events have reaffirmed our belief that assessment accuracy still relies on thorough analysis by specialists, as ML/AI tools, while valuable, have yet to match the depth and nuance of human expertise.

That said, as technology continues to evolve, it will play an increasingly vital role in helping insurers optimise their disaster response. In the future, AI-driven insights, used in tandem with human judgment, will provide critical support in delivering aid to those affected by catastrophic events.

By David Heathcote, intelligence delivery manager, McKenzie Intelligence Services