UC Santa Barbara

Wildfire is a serious threat to millions of people living in the Western United States, yet also an integral part of Southern California ecosystems. Therefore, it is important to quantify fire impacts and patterns of post-fire landscape recovery in order to understand the links between fire events and ecosystems. This research combined Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) remote sensing imagery and Light Detection and Ranging (LiDAR) data to produce a comprehensive, multi-year analysis of the May 2009 Jesusita Fire landscape within the Mission Creek Canyon watershed in Santa Barbara, California, USA. Combining passive and active remote sensing datasets allowed for a more detailed analysis of fire severity and the post-fire landscape recovery. Passive hyperspectral data provided information for a spectrally based assessment of fire severity and for mapping land cover types, while LiDAR provided geometric information such as topography and above ground vegetation structure. The study proposed a new fire severity definition based on multiple hyperspectral and LiDAR metrics: Multiple Endmember Spectral Mixture Analysis (MESMA) fractions and differenced Normalized Burn Ratio (dNBR) from AVIRIS; and a Canopy Height Model (CHM) from LiDAR. The study also examined the topographic effects on fire severity and post-fire recovery, using a LiDAR derived Topographic Wetness Index (TWI) and riparian areas defined from river locations collected from fieldwork. The result showed that the dNBR-MESMA-CHM based severity definition depicted a more detailed severity distribution in the Jesusita fire scar compared to the traditional spectral fire indices, especially for those areas with significant amounts of dead trunks. The riparian zone or areas with high soil water content were less affected by the fire, and the level of green vegetation cover returned to pre-fire status earlier compared to the fire scar average.