We illustrate the seasonal cycle of wildfire and climate in the western United States using a combination of observed, reanalysis, and simulated climate data sets, along with the 1986-1996 National Fire Occurrence data base described in Hardy et al. (2001, Intl. J. Wildland Fire 10:353-372; available at http://www.fs.fed.us/fire/fuelman/).

The total number of fires from all ignition sources for each day of the year, plotted as a function of day number, resembles a Gaussian curve that reaches an annual maximum during the first half of August. Fires started by lighting are almost entirely restricted to the interval from late April through October, while those from human causes occur during any month of the year. There is a prominent singularity in the frequency of human-set fires around the Fourth of July, when daily fire frequencies attain levels three times higher than those just before or after the holiday. High summertime levels of human-set fires tend to taper off gradually during the autumn. In general, the shape of the distribution of the total number of fires resembles that of the lightning-set fires more than that of the human-set fires; the latter acts mainly to determine the shape of the general distribution during the late autumn, winter, and early spring.

The time of year of peak incidence of fires from all sources varies across the western U.S. and among ecoregions. Fires occur earlier in the southwestern desert and steppe ecoregions (SW), and progress, in order, into the intermountain west and southern Rocky Mountains, the Pacific Northwest (PNW), northern Rocky Mountains (NRM), and California. Wintertime fires in the western U.S. are almost exclusively restricted to the Mediterranean ecoregions.

The temporal and spatial variations of fire incidence reflect atmospheric circulation and moisture flux, and surface energy- and water-balance variations, which jointly produce the low fuel- and soil-moisture conditions and high ignition probabilities that typify summers in the western U.S. The seasonal variations of key climate variables are illustrated by examining their intermonthly (month-to-month) changes. This approach removes the large-amplitude summer/winter difference and reveals considerable spatial heterogeneity in the month-to-month progression of climate during the year. The principal regional contrasts among the SW (with an earlier maximum in the incidence of fire) and the PNW and NRM (with later maxima) reflect the antiphasing of precipitation generated by mid-latitude cyclones (in winter) and the southwestern monsoon (in summer), which is superimposed on the insolation-driven depletion of soil moisture in summer. Ignition by lighting is favored by the summertime development of thermal lows in the interior that are associated with high temperature, dry atmospheric conditions, high evaporation/evapotranspiration rates, onshore flows of moisture, steep environmental lapse rates, and support for uplift from weak upper-level troughs that together trigger dry convective storms and associated windy conditions.