A Multi-Index Examination of Future Fire Season Length and Severity Over the United States

Our goal is to better understand the potential effects of climate change on fire danger for the near (2030-2060) and far (2069-2099) future.  We examine changes in fire season length and severity using multiple regional climate model (RCM) simulations and multiple fire indexes for the entire United States.  As different fire indexes make different assumptions about fuel and soil conditions, and different fire indexes are favored for different applications in different regions, we use 8 indexes, with 2 variations on 3 of those, for 11 measures of fire behavior.  We examine changes in the number of days above the 80^th, 90^th, and 97^th percentiles of these indexes, corresponding to thresholds used to assign descriptors of fire danger as high, very high, and severe.  To define fire season length, we use the number of days above the 80^th percentile for each given index.     

We employ thirteen simulations produced for the North American component of the Coordinated Regional climate Downscaling Experiment (NA-CORDEX), leveraging RCP8.5 emission scenario simulations for the future projections.  From these simulations, we calculate the KBDI (Keetch-Byram Drought Index), mFFWI (modified Fosberg Fire Weather Index), CFWI (Canadian Fire Weather Index), FM100 and FM1000 (100- and 1000-hour Fuel Moisture), ERC (Energy Release Component), BI (Burning Index), and SFDI (Severe Fire Danger Index).  Two fuel scenarios, G and LAF, are input into ERC, BI, and SFDI.  

By mid-century, most regions are projected to see an increase in the length of fire season, though the magnitude of this projected change varies considerably by region and fire index (and the latter’s sensitivity to precipitation).  Agreement on a lengthening of the season by up to 50% is strongest across simulations and fire indexes over the Southern Plains and Southwest U.S.  Changes in fire season severity are largest and most consistent across indexes for the U.S. west of the Mississippi River, excluding the intermountain West, where uncertainty is higher across the indexes and RCMs.  An approximately 2- to 6-fold increase in the number of days that reach the severe fire danger threshold is projected by mid-century.  Simulations for the far future are qualitatively similar, but the projected changes are quantitatively worse and more widespread.