Spatial controls of historical fire regimes: A multiscale example from theinterior west, USA

Our objective was to infer the controls of spatial variation in historical fire regimes. We reconstructed a multicentury history of fire frequency, si ze, season, and severity from fire scars and establishment dates of 1426 tr ees sampled on grids in four watersheds (similar to 64 plots, over similar to 1620 ha each) representative of the Blue Mountains, Oregon and Washingto n, USA. The influence of regional climate, a top-down control, was inferred from among-watershed variation in fire regimes, while the influence of loc al topography, a bottom-up control, was inferred from within-watershed vari ation. Before about 1900, fire regimes varied among and within watersheds, suggesting that both top-down and bottom-up controls were important. At the regional scale, dry forests (dominated by ponderosa pine), burned twice as frequently and earlier in the growing season in southern watersheds than i n northern watersheds, consistent with longer and drier fire seasons to the south. Mesic forests (dominated by subalpine fir or grand fir) probably al so burned more frequently to the south. At the local scale, fire frequency varied with different parameters of topography in watersheds with steep ter rain, but not in the watershed with gentle terrain. Frequency varied with a spect in watersheds where topographic facets are separated by significant b arriers to fire spread, but not in watersheds where such facets interfinger without fire barriers. Frequency varied with elevation where elevation and aspect interact to create gradients in snow-cover duration and also where steep talus interrupts fuel continuity. Frequency did not vary with slope w ithin any watershed. The presence of both regional-scale and local-scale va riation in the Blue Mountains suggests that top-down and bottom-up controls were both important and acted simultaneously to influence fire regimes in the past. However, an abrupt decline in fire frequency around 1900 was much greater than any regional or local variation in the previous several centu ries and indicates that 20th-century fire regimes in these watersheds were dramatically affected by additional controls such as livestock grazing and fire suppression. Our results demonstrate the usefulness of examining spati al variation in historical fire regimes across scales as a means for inferr ing their controls.