Modeling the Impact of Wildfire-Related Air Pollution on Mortality

November 2025

Authors

Eve Titon
Flora Auter
Tinhinane Talbi
Marie Ganon

Executive Summary

This report examines the impact of wildfires and broader air pollution on mortality, with an emphasis on modeling approaches relevant to life insurers.

Although wildfires can cause immediate fatalities—such as caused by burns and smoke inhalation—their more substantial mortality impact arises from increased air pollution. Smoke from wildfires raises concentrations of harmful pollutants, especially fine particulate matter (PM2.5 and PM10), which contribute to chronic health conditions and long-term excess mortality not always directly linked to wildfire events.

Capturing this indirect, delayed relationship between pollution and mortality is the primary modeling challenge. This report addresses that challenge through three modeling approaches, each suited to different actuarial applications and time horizons:

Stochastic Mortality Framework An adaptation of the Lee-Carter model incorporates climate variables to estimate short- to medium-term impacts. This method is most effective where strong empirical relationships exist between pollution indicators and mortality and is compatible with insurers’ existing stochastic frameworks.
Prevalence Scenario Approach Mortality is projected based on expected increases in disease prevalence attributable to pollution. This intuitive method aligns with morbidity modeling practices but depends on reliable disease prevalence data and Global Burden of Disease mortality estimates, which are derived from statistical modeling rather than direct counts.
WHO AirQ+ Methodology This epidemiological model uses concentration-response functions (CRFs) to estimate deaths attributable to pollutants such as PM2.5 and ozone. Although well suited for scenario testing, it requires careful calibration and may not fully reflect local conditions because of assumptions embedded in the CRFs.

Across all approaches, several limitations are identified:

Data granularity: Accurate modeling requires geographically and temporally detailed mortality data.
Historical data gaps: Long-term pollutant exposure modeling is limited by sparse historical measurement data. It is challenging to obtain a comprehensive historical record of reliable air pollution data, due to either an insufficient number of monitoring stations or inadequate frequency of data collection, particularly for older years.
Population dynamics: Calibration on historical data may not reflect future changes in vulnerability or exposure.
Spatial averaging: State-level data may obscure local pollution hotspots and topographic effects.
Interaction effects: Existing models do not capture combined impacts of pollution and other climate stressors, such as heatwaves or behavioral adaptations.

Despite these challenges, the models provide valuable tools for actuaries assessing environmental risks. They can inform stress testing, pricing adjustments, and the quantification of climate-related mortality shocks. For example, region-specific pollution risk factors could support the development of mortality zoning or adjusted pricing for wildfire-prone areas.

Materials

Acknowledgements

The researchers’ deepest gratitude goes to those without whose efforts this project could not have come to fruition: the Project Oversight Group for their diligent work overseeing, reviewing, and editing this report for accuracy and relevance.

Project Oversight Group members:
Sam Gutterman, FSA, CERA, MAAA, FCAS, FCA, HonFIA
Charlie Mathews
Donna Megregian, FSA, MAAA
Rebecca Owen, FSA, MAAA, FCA
Sandra Said
Aadit Sheth, FSA, CERA, FCIA
Cirhan Truswell, BSc, MSc
Georgiana Willwerth, MD

At the Society of Actuaries Research Institute:
Kara Clark, FSA, MAAA, Senior Research Actuary
Rob Montgomery, ASA, MAAA, FLMI Consultant-Research Project Manager
Ronora Stryker, ASA, MAAA, Senior Research Actuary
Barbara Scott, Senior Research Administrator

Questions or Comments?

Give us your feedback! Take a short survey on this report. Take Survey

If you have comments or questions, please send an email to Research@soa.org