Climate and Fire Interactions
How does drought relate to annual area burned and fire size?
The interannual variation of each of the eight fuel aridity metrics, reference potential evapotranspiration (ETo), VPD, climatic water deficit (CWD), Palmer drought severity index (PDSI), fire weather index (FWI) from the Canadian forest fire danger rating system, energy release component (ERC) from the US national fire danger rating system, McArthur forest fire danger index (FFDI), and the Keetch–Byram drought index (KBDI), were all significantly correlated to annual area burned of forested lands from 1984-2015. Furthermore, the authors observed a 3.3-fold difference in the area burned between 1984 to 1999 and 2000 to 2015.
The authors state that the influence of anthropogenic climate change will continue to increase areas burned due to project increases in aridity unless fuel become limiting.
The authors found that fire season length has significantly increased across approximately 25% of the Earth’s vegetated area resulting in a doubling of the global burnable area. Areas with the greatest increase in fire season length occurred where there were the greatest changes in especially temperature, but also humidity, length of rain-free periods, and wind speeds.
Despite occurring during time periods of similar short-term climate conditions, differences in fire severity and fire size were observed between areas of differing vegetation types. This is likely the result of differences in species composition and fuel amount and condition associated with longer-term climate effects.
The authors found that in flammability-limited ecoregions, specifically the Western Cordillera and Appalachian forests, long-term drought conditions were strong predictors of very large fire occurrence. Prior year anomalously wet conditions preceded very large fire occurrence in fuel-limited ecoregions, Semi-arid Prairies and Sierra Madre Piedmont of southern Arizona.
The authors found that the Keetch-Byram Drought Index (KBDI) is expected to increase, driven primarily by increasing temperature across the Southwest, Rocky Mountains, northern Great Plains, Southeast, and Pacific coast, thereby increasing future fire potential across most of the continental U.S.
Climate-induced stress leading to increased tree mortality interacts with disturbance such as fire further exacerbating mortality. There is uncertainty how future climate change may affect this feedback loop between physiological stress and fire disturbance.
The authors found that most, but not all, large fires occurred during the La Niña phase of ENSO. For the Cerro Grande fire, which occurred during an extreme La Niña event, topography had a stronger influence on reducing severity in some areas than the regional effects of climate.
Climate change projections suggest a shift toward hotter and drier conditions across the west resulting in an approximately 50 to 500% increase in area burned across the western U.S.
Across the continental U.S., drought conditions, as measured by PDSI, are strongly correlated to the number of acres burned. The strength of the correlations between PDSI and acres burned vary across the U.S. at one, two, and current year time lags.
The authors found that longer fire seasons measured by an increase in the total and consecutive number of days without rain may have increased area burned by providing longer periods of weather favorable for fire activity.
The authors found that area burned was strongly related to PDSI across all regions of the interior West, but that PDSI had a varied relationship with phases of AMO and PDO in the different regions. For example, in Arizona and New Mexico, area burned was most strongly related to PDSI during the warm phase of the AMO (long-term drought conditions typical) or the cool phase of PDO (cooler and wetter winters typical), whereas the relationship was either weak or not significant in the other regions during these same period.
The authors found that across the gradient, total area burned is significantly correlated to moisture-related (except actual precipitation) factors in the year(s) previous to the fire season. The authors suggest that this is related to the increase in fine fuel production. The lags between high and low elevation sites, however, differed. Low elevation sites were associated with wet antecedent conditions up until the fire season, whereas high elevation sites were associated with wet condition up to 3 years prior. However, the authors did not find any correlations between total precipitation and fire at any yearly or seasonal lags. They also found that drought was not necessarily associated with increased fire activity and that in certain vegetation types, fuels dry sufficiently to burn under natural conditions.
Although they vary widely across regional and temporal gradients, throughout the West moisture anomalies from two years up to the spring proceeding the fire season are widely linked to fire activity.
The authors found that, generally, in shrub- and grassland-dominated ecosystems, there was a strong relationship between area burned and increased moisture in the year before the fire season, suggesting these systems are fuel limited and only burn after moisture (inferred from PDSI) increases fuel production. For open-canopy forests, they found a strong relationship between area burned and decreased moisture immediately prior or during the fire season, suggesting that the moisture conditions of the heavier fuels drive fire activity.
The authors found a significant correlation between annual area burned and fire intensity directly related to the weather variable frequency. Years where large area burned were associated with extreme fire weather and subsequent high intensity fire.
The authors found strong relationships between area burned and highly positive values of the Southern Oscillation index (SOI), which is associated with reduced rainfall and severe winter-spring drought in the Southwest. They also found regional synchrony of large fires during the high-SO phase, which is associated with limited spring precipitation resulting in reduced tree growth. This suggests that seasonal climate, and not just local weather conditions, affect fire activity at a regional scale.