The authors found that future fire probabilities increased with increasing temperature; however predictions for each of the climate models diverged for the southwestern U.S. The CGCM data resulted in decreased future fire probability (0 to ?30%) while the GFDL data resulted in increased fire probability (near 0 to greater than 40%). The authors suggest this discrepancy is due to the limitations in predicting precipitation and moisture conditions and their effect on fuel production.

The authors found that fire was highly asynchronous between the three sites despite exhibiting similar fire regimes, suggesting that at the scale of the study, fine-scale factors were more influential on fire occurrence than regional climate. At long-term scales, ENSO, specifically La Niña events (drought), were correlated with higher fire occurrence at the study sites until approximately 1830.

The authors found that fire activity/area burned and fire severity across the western US increased with increasing actual evapotranspiration (AET) which represented fuel amount in the study. However, area burned increased with increasing water deficit (fuel moisture) to a threshold before fire activity subsequently decreased. Fire severity decreased with increasing water deficit. This suggests that fire activity is both fuel and moisture limited. Regional areas across the West that have higher fuel moistures and lower water deficits rarely burn because fuel is not dry enough to ignite, whereas areas with low fuel moisture and high moisture deficits are not productive enough to support fire spread.

The authors found that treated sites prior to the 2011 Wallow Fire resulted in lower tree mortality, smaller patches of high severity, and significantly higher understory herbaceous cover post-fire suggesting that fuel treatments imbue resiliency to uncharacteristically severe fire in mixed conifer ecosystems.

The authors predicted an increase in wildfire area burned across most of the western U.S. based on a 1–3°C increase in temperature, with larger increases in some areas, Pacific Northwest and Rocky Mountains Forest ecoregions, than others, almost no increase predicted for Nevada Nevada Mountains/Semidesert and Eastern Rocky Mountains/Great Plains ecoregions. Along with predicted increases in area burned, the authors modeled a 40% increase in summertime organic carbon (OC) concentrations and an 18% increase in elemental carbon (EC) concentrations by 2050. Areas with the highest predicted future increases in carbonaceous aerosol concentrations are concomitant with the areas of greatest increases in wildfire.

Across regional climate models (RCMs) and general circulation models (CGMs) the authors found that the number of days with a high Haines Index Value (? 5) increased across most of the contiguous U.S. during the summer season. The moisture component of the Haines Index (HI) was more influential in predicting future projected changes in conditions conducive to increased fire activity than the stability component of HI.

The authors found that anomalously warm conditions occurred during a period of intense fire activity in the summer of 1910. A strong La Niña pattern in the year of 1910 likely contributed to warming and drought during the spring and summer prior to the fires. Furthermore, similar anomalous conditions in 1998 and 2012 occurred during other significant fire years across areas of the western U.S.

The authors found that total area burned in the Southwest has increased since 1984 at a rate of 10.2% per year, and that the rate of increase is greater for forests at higher elevations. Furthermore, March through August VPD, a measure of the ability of the atmosphere to extract moisture from surface vegetation, is more strongly correlated with total area burned and area burned at high severity than temperature alone for the Southwest region. The authors explain that air temperature is commonly correlated with severity, however, it is often used as a proxy for atmospheric water demand as high temperatures in a low-humidity environment cause dry fuel conditions in the Southwest. Extended periods when the value of VPD remains high can directly affect the rate of transpiration or water loss from the foliage and can lead to plant desiccation and mortality causing increased flammability. More importantly, vapor pressure at saturation increases exponentially with increasing temperature, thereby small increases in temperature will exponentially increase VPD, potentially resulting in more frequent drought and increased wildfire risk until fuels become limiting.

The authors found that a single entry, low severity fire was not effective in reducing tree densities in ponderosa pine trees with white fir encroachment, although a second entry fire did reduce white fir populations. Low severity fire, however, reduced fuel load density in the dry mixed-conifer stands. They also found that aspen regeneration after high severity fire was abundant. For ponderosa pine, seedlings did increase after fire, however, due to the dense overstory canopy, survival into adulthood was uncertain for both stand types.

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.