COS 218-1 - What Drives Periods of Rapid Fire Spread?

Abstract: Extreme wildfires threaten lives, property, and ecosystems. The direct cost of US fire suppression for the 2021 fire season was over 4 billion dollars. Is extreme fire behavior driven by weather, or by fuels? Large fires can experience “spread-days” where fire perimeters expand quickly, and “stay-days” where fires continue to burn but the perimeter expansion is slow. How significant spread-days are to a region’s fire regime, if spread-days affect vegetation differently than stay-days, and the conditions driving shifts between stay-days and spread-days are not well understood. Improving our understanding of fire weather impacts on fire behavior could support the development of improved forecasts to anticipate spread-day conditions. Regardless of if fire spread-days are caused by fuels or weather, variability in fire spread rates and fire persistence likely contribute to heterogenous impacts within the perimeter of multi-day fires.

We investigated spread-days in the Western United States 2020 fire season using sub-daily information on fire area and mean Fire Radiative Power (FRP) measured by the VIIRS sensors. We classified spread-days within the multi-day time series of fire expansion for each fire event, and then compared our fire-spread time series to a gridded Fire Weather Index (FWI) product based on satellite precipitation data (IMERG). Finally, we looked at the relationship between FRP, spread, and post-fire classification of severity by the MTBS program for the 2021 Caldor fire.

We found that on average 60% of a fire’s final area accumulates during spread-days. Across all large fires (n = 405), 85% of the total fire area came from a spread-day. Ignitions (new fire starts) were associated with the highest FWI values, and spread-days were associated with higher FWI values than stay-days. Ignitions and spread-days for ongoing fire events were synchronized across the western US during the 2020 fire season. In the Caldor fire, we found that periods of rapid fire spread were also more intense, leading to higher soil burn severity.

These findings suggest that most of the total burned area for a given fire event occurs during days when a fire is spreading quickly, and that fires are more likely to spread quickly under high FWI values. The temporal clustering of ignitions and spread-days across the Western US suggests that FWI may explain aspects of fire-behavior better than local conditions like fuels. Spread-days may also explain differences in fire-severity within the footprint of a single fire, providing important context for assessments of post-fire impacts.