Synoptic weather patterns for large wildfires in the northwestern United States—a climatological analysis using three classification methods

Wildfire regimes respond to atmospheric variability on multiple time scales from interannual variations of drought to daily fluctuations of humidity and wind. Synoptic weather patterns effectively link both short- and long-term atmospheric variability, and thus, understanding fire-prone synoptic patterns is an integral part of fire management programs. Different analysis procedures of varying complexity have been used to identify the most common synoptic weather patterns conducive to wildfires. In this study, we examine the utility of three synoptic-pattern classification methodologies, namely Composite, Empirical Orthogonal Function (EOF), and Self Organizing Map (SOM). We focused on the synoptic weather patterns associated with 203 wildfires that burned 50,000 acres (20,250 ha) or more and 80 wildfires that burned 100,000 acres (40,500 ha) or more in the Northwestern United States between 1984 and 2014. The basic 500-hPa geopotential height patterns identified by the three analysis methods were found to be similar for the two fire-size categories, but the strengths of the geopotential height anomalies were larger for the larger size fires. All three methods were able to consistently identify the dominant pattern with only small differences in strength and position of the anomalies, but the EOF and SOM methods yielded other less dominant, but still important, patterns. Particularly, the SOM method captured transitional synoptic patterns that may account for 20–30% of these large wildfires.