Mapping surface fuels using LIDAR and multispectral data fusion for fire behavior modeling

摘要

Fires have become intense and more frequent in the United States. Improving theaccuracy of mapping fuel models is essential for fuel management decisions and explicitfire behavior prediction for real-time support of suppression tactics and logisticsdecisions. This study has two main objectives. The first objective is to develop the useof LIght Detection and Ranging (LIDAR) remote sensing to assess fuel models in EastTexas accurately and effectively. More specific goals include: (1) developing LIDARderived products and the methodology to use them for assessing fuel models; (2)investigating the use of several techniques for data fusion of LIDAR and multispectralimagery for assessing fuel models; (3) investigating the gain in fuels mapping accuracywith LIDAR as opposed to QuickBird imagery alone; and, (4) producing spatiallyexplicit digital fuel maps. The second objective is to model fire behavior usingFARSITE (Fire Area Simulator) and to investigate differences in modeling outputs usingfuel model maps, which differ in accuracy, in east Texas.Estimates of fuel models were compared with in situ data collected over 62 plots.Supervised image classification methods provided better accuracy (90.10%) with thefusion of airborne LIDAR data and QuickBird data than with QuickBird imagery alone (76.52%). These two fuel model maps obtained from the first objective were used to seethe differences in fire growth with fuel model maps of different accuracies. Accordingto our results, LIDAR derived data provides accurate estimates of surface fuelparameters efficiently and accurately over extensive areas of forests. This studydemonstrates the importance of using accurate maps of fuel models derived using newLIDAR remote sensing techniques.