Simultaneous very large wildfires present a unique challenge to fire management and the allocation of firefighting resources. The effect of climate change on simultaneity in large wildfires varies considerably over North America, though most regions show an increase in simultaneity and a longer or later fire season. High levels of simultaneity in multiple regions have a significant impact on fire management, and can result in unfulfilled requests to share resources between regions.

Working at the level of the Geographic Area Coordination Center (GACC, administrative regions used in fire management), we fit generalized linear models (GLMs) with a negative binomial response to observational data to predict megafire simultaneity as a function of GACC-average fire index. We then apply the resulting GLMs to output from 13 regional climate model (RCM) simulations from the NA-CORDEX data archive. These simulations include 7 different RCMs that are driven by global climate models (GCMs) whose equilibrium climate sensitivity roughly spans the range seen in CMIP5, capturing an important dimension of uncertainty about future climate.

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 the GACC and in time using a two-week moving window. The best-performing fire index across the region proves to be CFWI (Canadian Fire Weather Index), except in the Southwest GACC, where FM100 (100-hour Fuel Moisture) is better. The simulations use RCP8.5 greenhouse gas trajectories from 1950-2100.

From the resulting projections, we calculate correlations in simultaneity between GACCs and analyze how they change in the 21st century. To investigate the effect on conditions that interfere with resource sharing, we also calculate changes in the probabilities of two GACCs experiencing high internal simultaneity at the same time. We also evaluate the uncertainties in these changes associated with the choice of simulation.