A Novel Model of Canadian Wildfire Risk for Climate Assessments

Concerns surrounding climate-driven shifts in wildfire risks in North America have motivated a new model of wildfire risk projections for Canada. Canada has experienced an acceleration of wildfire activity in the last decades, with significant impacts on infrastructure, resources, and livelihoods. However, current and future wildfire exposure and its associated financial costs are still poorly quantified. First Street has computed the first climate-adjusted, asset-specific estimates of Canadian wildfire risk. These estimates include the effects of a changing climate, and projected exposure estimates for today as well as 30 and 100 years into the future. These estimates were constructed using large numbers of Monte Carlo simulations using the ELMFIRE fire behavior model, driven by a time series of hourly NOAA RTMA surface weather observations and a novel fuels dataset for Canada derived by First Street from both satellite and in situ data sources. Statistical methods were used to adjust the weather time series using WCRP CMIP6 future climate projections for 2055 and 2100 under the SSP126, SSP245, and SSP585 scenarios to reflect future air temperature, humidity, and precipitation conditions. Historic ignition locations were used as the basis for an ignition density field to initiate modeled fires for all scenarios. The resulting wildfire hazard exposure estimates are expressed as burn probability, flame length (mean and maximum), and ember exposure at 30m horizontal resolution for all of Canada below the Arctic Circle. The resulting exposure levels enable the climate-driven risk evaluation of each of the approximately 17M homes and businesses in Canada, including those within the Wildland Urban Interface. The model predicts a larger number of fires, increases in wildfire burn probability and a greater number of exposed assets under future CMIP6 scenarios, driven mainly by changes to fuels states in a warming climate.