Thursday, November 9, 2006
Spatial variability in forest wildfire regime response to warming temperatures and an earlier spring snowmelt, and applications for seasonal forecasting and climate change impact assessment
Analysis of a comprehensive wildfire history for Western United States forests indicates the incidence of large forest wildfires since the mid-1980s is four times the level that prevailed in the 1970s and early 1980s. Total area burned in these large fires has increased to more than six and a half times its previous level. Interannual variability in western United States large forest wildfire frequency and total area burned over this period is found to be strongly correlated with mean spring and summer temperatures for the region, and the increased fire activity is associated with trends towards warmer temperatures and an earlier spring melt. The result of these trends has been a longer dry season, drier soils and vegetation in summer, a longer large-fire season, more and larger fires, longer-burning fires, and greater and more variable fire suppression expenditures.
These results have been most pronounced in the forests of the Northern Rocky Mountains, where elevation appears to play a strong role in modulating the sensitivity to wildfire. Here, we find that at mid elevations (6000-8000 feet), there is a large forested area that is usually buffered from summer drought by moisture reservoirs in snow and soils that has experienced anomalous moisture deficits in recent warm, early springs. Conversely, forests in the Southwest and Southern California appeared to be least affected by warming, and wildfire activity there was not as strongly associated with an earlier spring snowmelt in the region.
Using the VIC hydrologic model, we examine climatologies of the local water balance (i.e. actual evapotranspiration and moisture deficit) and the length of time snow is on the ground to explain the observed spatial variation in forest wildfire regime sensitivity to warming and snowmelt timing. The same climatic factors that control the geographic distribution of forest types also appear to be determining factors for fire regime sensitivity to warming and snowmelt timing. Categorizing western forest areas by elevation and location, more than 80% of the spatial variability in wildfire activity associated with snowmelt timing can be explained by related changes in moisture deficit. These findings have important applications in modeling wildfire for climate change impact assessment and seasonal forecasts.
Anthony Westerling is Assistant Professor at the School of Engineering & School of Social Sciences, Humanities, and Arts at the University of California, Merced.