PS 16-207 - Quantifying climate-fire relationships in forest ecosystems of the U.S. Northern Rockies, 1902-2008

Monday, August 6, 2012
Exhibit Hall, Oregon Convention Center
Kerry B. Kemp1, Adam M. Young2, Philip E. Higuera3, John Abatzoglou4 and Penelope Morgan3, (1)Forest, Rangeland and Fire Sciences, University of Idaho, Moscow, ID, (2)University of Idaho, Moscow, ID, (3)College of Natural Resources, University of Idaho, Moscow, ID, (4)Geography, University of Idaho, Moscow, ID
Background/Question/Methods

Climate plays a large role in determining the distribution of vegetation and fire occurrence in mountainous regions in the western U.S.  Climate metrics that synthesize temperature and precipitation are important predictors of vegetation type and productivity, and may be more useful than temperature and precipitation alone for predicting fire occurrence over large spatial and temporal scales by providing proxy information about fuel availability.  To understand the influence of decadal-scale climate on fire occurrence in forests of the U.S. Northern Rockies, we modeled the regional distribution of fire as a function of mean annual moisture deficit and mean growing season temperature, and by forest-type classifications compiled from the Landfire database.  Climatologies were derived from climate data downscaled to 4-km resolution for the period 1979-2010.  Perimeters of fires >20 ha were obtained from a regional fire atlas spanning 1902-2008 and encompassing a 9.73 million ha recording area within 12 National Forests in Idaho and western Montana and Glacier National Park. For fires across the entire study area, we analyzed the location and frequency of fires in climate space defined by (1) the ratio of annual evapotranspiration to potential evapotranspiration (AET/PET) and (2) growing degree-days above 0°C. 

Results/Conclusions

From 1902-2008, 6.9 million ha of forests in the Northern Rockies burned within the study area, with 1.92 million ha burned more than once.  At most, a single 120-meter pixel burned six times over the 107-yr record. Area burned within cool-wet forest types was roughly equal to area burned in warm-dry forest types (51 vs. 49%, respectively). Fire occurrence is ubiquitous across the study area; however, fire tends to occur more predominately in warm, dry climate space.  Fire is biased towards regions with annual AET/PET less than 0.6, and where growing degree-days (base 0°C) were greater than 2500 C-days.  In the Northern Rockies, this climate space is currently and historically not very common on the landscape; thus, our results suggest that as moisture deficit and temperatures increase in the Northern Rockies, the likelihood of fire may also increase as more of the landscape is contained within climate space that is favorable for fire occurrence. If true, this will have important implications for understanding the spatial distribution of fire-prone landscapes in the future.