3.3 The recipe for megafires in the Eastern US: the role of the temporal scales

Wednesday, 16 October 2013: 9:50 AM
Meeting Room 1 (Holiday Inn University Plaza)
Renaud Barbero, University of Idaho, Moscow, ID; and J. T. Abatzoglou, C. A. Kolden, and N. K. Larkin

A large fraction of annual area burned occurs during a limited number of exceptionally large wildfires, or megafires, defined as the top 10 percent of large fires. While much prior research has been devoted to macroscale climate-fire relationships in the western half of the United States, this work examines relationships between climatic conditions and the occurrence of megafires across the Eastern United-States (EUS) using fire perimeters from the Monitoring Trends in Burn Severity (MTBS) data set over 1984-2010. Biophysically relevant variables including outputs from the National Fire Danger Rating System and Canadian Forest Fire Danger Rating System, in addition to daily surface meteorological variables and drought indices were acquired for all large wildfires in the EUS. Unlike most prior analysis that examines climate-wildfire linkages through univariate analysis, this analysis considers the influence of independent atmospheric factors on fuel abundance, flammability and spread by incorporating interannual, subseasonal and synoptic variability, respectively. Relationships were examined both using all large wildfires for the EUS and for six objectively defined geographical clusters across EUS that exhibited spatio-temporal commonality in large wildfire spatial statistics. Results show that although most large fires occur under moderate-to-strong fire danger conditions in the high-, medium- and low frequency variability, significantly more-protracted and extreme fire danger was coincident to megafires. A niche-space analysis was used to assess maximum forecast skill for megafire occurrence based on the combination of interannual, subseasonal and synoptic factors. For each cluster, we defined the most relevant variables and the best thresholds to use in megafires detection. The results show that megafires are often a combination of climate variability at different timescales. For example, in fuel-limited grassland ecosystems such as in Great Plains, megafires are favored during a combination of pluvial conditions during the growing season prior to the fire and sustained strong winds following ignition, while megafires along the along the North and South Carolinas are favored by the superimposition of short-, medium- and long-term moisture deficits. This framework may help to better understand the role of the different timescales in the megafire occurrence as well as to better identify megafire hazards giving seasonal climate predictions and climate change scenarios.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner