Close Window

Using remotely sensed data to asses the burn severity of wildland fires in Alaska’s interior boreal forest.

Elizabeth E Hoy, Department of Geography, University of Maryland, elizabeth.hoy@gmail.com (Presenting)
Eric S Kasischke, Department of Geography, University of Maryland, ekasisch@geog.umd.edu
Nancy HF French, Altarum Institute, nancy.french@altarum.org
Merritt R Turetsky, Department of Plant Biology, Michigan State University, mrt@msu.edu

The interior Alaskan boreal region has the potential for warm and dry summer conditions resulting in the increased possibility of fires, such as in 2004 when over six million acres of land burned. This complex system can have important implications for carbon sequestration and emissions - both on local and global scales. This study uses broad scale remote sensing mapping techniques to more accurately quantify the fraction of area burned and the burn severity of wildland fires in an effort to gain a better understanding of the carbon emissions of this region. First, the accuracy and feasibility of two Landsat TM and ETM+ remotely sensed indices are compared in the analysis - the differenced normalized burn ratio (dNBR) and the tasseled cap transformation. While the dNBR is traditionally used to map burn severity, it is believed that the tasseled cap transformation, based on information extending over a wider range of the electromagnetic spectrum than the dNBR, could be an alternative. The usefulness of these indices as methods to examine the fraction of area burned within each fire is then explored within the study. Additionally, MODIS Hotspot data is used to examine the progression of each fire event; this analysis is then used to better understand the results from the dNBR and tasseled cap in light of the time of year in which the fire occurred. These image analysis techniques, coupled with detailed vegetation maps of the interior boreal region, give greater insight into this region through accounting for and analyzing unburned islands within each fire scar and through quantifying variations within the perimeter of the burned area.

Presentation Type:  Poster

Abstract ID: 122

Close Window