Synthetic aperture radar (SAR)-based mapping of wildfire burn severity and recovery

New radar based techniques for efficient identification of forest damage caused by wildfire and subsequent recovery are applied to data acquired over the 2002 Rodeo-Chediski, Arizona and 1988 Yellowstone National Park wildfire complexes. Fully polarimetric C-, L- and P-band airborne synthetic aperture radar data were acquired in approximately east-west and north-south swaths over the northern half of the Rodeo-Chediski wildfire scar on August 1, 2002, 25 days after the last active fire was contained. The AIRSAR instrument also acquired fully polarimetric data in C-, L- and P-band over Yellowstone on October 11, 1994, approximately 6 years after the event. We combine single frequency polarimetric parameters in composite images for optimal identification of fire-induced damage to the forests. For the Rodeo-Chediski fire, we demonstrate that polarimetric parameters - such as average scattering mechanism - readily identify bare surfaces, pine trees with intact needles, trees with exposed branches, and trees with residual burned trunks. In particular, a single-pass of polarimetric SAR provides identification of the perimeters and within-burn variability. This work shows that fully polarimetric SAR is sensitive to scattering changes wrought by wildfire in two types of predominantly coniferous forests. By extension, this technique should help quantify different degrees of burn severity in ecosystems where the canopy is altered. We anticipate that these results will provide a new method of responding to wildfires, perhaps even at the tactical fire-fighting level.