In the Southwest GACC, reference evapotranspiration explained the greatest percentage of variance between climate and area burned for both forested and non-forested areas. In general, biophysical variables that include a direct link to fuel moisture conditions performed better than any single individual variable, such as temperature or precipitation.
Negative relationships with increased snow-water equivalent (SWE) were also observed in montane ecosystems where late-winter/early-spring precipitation delays the onset of fire season and increases fuel moistures into the fire season.

In the Northern Rockies, they found that winter and fire season precipitation was the most important factor in limiting fire size and occurrence.

The authors reconstruction of past fire activity suggests that until the late 18th century fire was driven mostly by climate. Specifically, they found that variation in global precipitation had the strongest influence on global fire activity. Post Industrial Revolution, anthropogenic activities had a stronger influence on global fire trends. Finally, predictions in future warming due to climate change suggest an imminent shift to temperature-driven global fire activity over the next century.

The authors found that globally, changes in temperature and precipitation may result in a rearrangement of fire probabilities where fire activity will increase in some areas and decrease in others under climate change.

Southwestern deserts and semiarid desert ecoprovinces tend to be fuel limited. Wetter conditions in the winter prior to the fire season followed by dry condition during the year of the fire resulted in greater area burned. The author’s suggest that climate change may lead to one of two paths: first a reduction in fuels due to persistent drought, limiting area burned, or second, wetter conditions leading to an increase in vegetation biomass along with an earlier onset of warmer temperatures that dry these fuels and increase the area burned. The mountainous ecoprovince of Arizona and New Mexico also did not correlate to year-of-fire precipitation, but did have a relationship with precipitation from the winter previous.

The authors found that the 1 April snow water equivalent (SWE) is correlated with the number of lightning-ignited fires and area burned. High levels of persistent snowpack later into the year decreases the possibility of lightning-ignited fires during the fire season. Conversely, low levels lead to an increase in fire activity earlier in the fire season. The type and duration of winter precipitation greatly affects the fuel moisture conditions of the early fire season, and to a lesser extent, the late fire season as fuels are typically dry regardless of the previous winter’s precipitation. However, the relationship between SWE and area burned was weaker as bottom-up controls were stronger predictors of fire behavior on the landscape. However, they found that fire season length is not directly related to fire severity in that a longer fire season does not necessarily result in more severe fire across the landscape, especially in more drought tolerant ecosystems. Convective instability associated with high temperatures may also lead to increased lightning occurrence and thus increased ignition rates. Should spring SWE decrease, landscape flammability may increase and become more variable.

The authors found that during the period of deglaciation, there was an increase in the level of burning and fire occurrence, although fire activity varied across the continent likely due to spatially complex climate controls and their consequent effects on vegetation changes.

The authors found that across vegetation types, annual area burned was significantly correlated with the frequency and intensity of springtime rain events. Burn severity was significantly correlated with lack of precipitation represented by total number of days without rain and maximum consecutive number of days without rain. Snowpack was only marginally related to area burned and fire severity in the upper elevation forest types spruce-fir and mixed conifer.

Fire size in the lower and middle elevations are related to rainfall in the season previous to the fire season, but at higher elevations, this pattern does not hold.

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.