Wildfire extent and severity correlated with annual streamflow distribution and timing in the Pacific Northwest, USA (19842005)

Climate change effects on wildfire occurrence have been attributed primarily to increases in temperatures causing earlier snowpack ablation and longer fire seasons. Variability in precipitation is also an important control on snowpack accumulation and, therefore, on timing of meltwater inputs. We evaluate the correlation of total area burned and area burned severely to snowmelt-induced streamflow timing and total annual streamflow metrics across the Pacific Northwest region from 1984–2005. Principal component scores on total annual water year flow and date of 50th percentile flow (PC1T) in the Pacific Northwest were used as predictors of satellite-inferred area burned and area burned severely in forested settings. Both annual area burned and burned severely are significantly correlated with mean annual flow and streamflow timing. PC1T alone explains 24% of the variability in annual area burned. Path analysis suggests that a substantial amount of the variability in annual area burned, previously attributed solely to temperature effects on melt timing, may be primarily driven by trends in precipitation and total annual streamflow. Principal component analysis scores on mean annual streamflow explain as much as 46% of the variability in annual area burned from 1984–2005. Thus, although streamflow timing may be a better single correlate of annual wildfire activity, timing is, in turn, strongly dependent on precipitation. These results suggest that recent fire activity in forests of this region are influenced more by precipitation variability than temperature-induced shifts in snowmelt timing, with significant implications for our ability to predict wildfire activity in the future. Published in 2011. This article is a US Government work and is in the public domain in the USA.