Estimation of Forest Fuel Load From Radar Remote Sensing

Understanding fire behavior characteristics and planning for fire management require maps showing the distribution of wildfire fuel loads at medium to fine spatial resolution across large landscapes. Radar sensors from airborne or spaceborne platforms have the potential of providing quantitative information about the forest structure and biomass components that can be readily translated to meaningful fuel load estimates for fire management. In this paper, we used multifrequency polarimetric synthetic aperture radar (SAR) imagery acquired over a large area of the Yellowstone National Park by the Airborne SAR sensor to estimate the distribution of forest biomass and canopy fuel loads. Semiempirical algorithms were developed to estimate crown and stem biomass and three major fuel load parameters, namely: 1) canopy fuel weight; 2) canopy bulk density; and 3) foliage moisture content. These estimates, when compared directly to measurements made at plot and stand levels, provided more than 70% accuracy and, when partitioned into fuel load classes, provided more than 85% accuracy. Specifically, the radar-generated fuel parameters were in good agreement with the field-based fuel measurements, resulting in coefficients of determination of R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =85 for the canopy fuel weight, R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2 </sup> =0.84 for canopy bulk density, and R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> =0.78 for the foliage biomass