Simultaneous very large wildfires present a unique challenge to fire management and the allocation of firefighting resources. All else being equal, a single large fire is easier to fight than multiple separate fires with equivalent area.
The effects of climate change on the degree of simultaneity in large wildfires is unknown. Although climate change is expected to increase temperatures, heightening and extending the fire season, many studies also project changes in the duration and intensity of precipitation. These effects vary considerably over North America, and their net impact on fire depends on whether the ecosystem is fuel-limited or moisture-limited.
We fit generalized linear models (GLMs) to observational data to predict various measures of simultaneity as a function of different fire indexes. We then apply the resulting GLMs for the best predictors to output from regional climate model (RCM) simulations. We compare the distributions of the resulting projected simultaneity values, to evaluate how well the simulations capture fire-related climate factors and their interrelationships in the historical period, and assess projected changes in simultaneity in the future.
We relate wildfires from the MTBS (Monitoring Trends in Burn Severity) database to climate data from the gridMET observational dataset. We aggregate data spatially over GACCs (Geographic Area Coordination Centers, administrative regions used to coordinate firefighting) and use a two-week moving window in time. The fire indexes we use include KBDI (Keetch-Byram Drought Index), CFWI (Canadian Fire Weather Index), mFFWI, (modified Fosberg Fire Weather Index), ERC (Energy Release Component), BI (Burning Index), FM100, and FM1000 (100- and 1000-hour Fuel Moisture). For future projections, we use six 25-km simulations from the North American branch of the CORDEX downscaling program: two RCMs (WRF and RegCM4) driven by 3 GCMs (GFDL, MPI, and HadGEM2) using RCP8.5 greenhouse gas trajectories from 1950-2100.