The fallacy of area burned: Climate impacts on unburned area within fire perimeters

For decades, wildfire studies have utilized fire occurrence as the primary data source for investigating the causes and effects of wildfire on the landscape. Fire occurrence data fall primarily into two categories: ignition points and perimeter polygons that are used to calculate a ‘burned area' for a fire. The challenge in understanding ecological and climatological relationships to fire occurrence is that fires do not generally burn homogeneously across the landscape, nor do they generally consume 100 percent of the area bound by the perimeter. This research investigates the importance of unburned area within the fire footprint, which provides an ecological refugia and seed source for post-fire regeneration. It utilizes differenced Normalized Burn Ratio (dNBR) data from the Monitoring Trends in Burn Severity project to look at patterns of unburned area in three national parks across the western U.S. (Glacier, Yosemite, and Yukon-Charlie Rivers). We found that unburned pixels comprised at least 7 percent of the pixels within the fire perimeter for any given year, and up to 92 percent within the fire perimeter for a low fire year in Yosemite NP. From 1984-2009, the total area within the fire perimeters that was classified as unburned from dNBR was 35 percent for Yosemite, 14 percent for Glacier, and 13 percent for Yukon-Charlie Rivers. These results raise questions about the validity of relationships found between external forcing agents, such as climate, and ‘burned area' values derived solely from polygon fire perimeters.