Authors: Douglas Stow*, San Diego State University, Philip Riggan, United States Forest Service, Janice Coen, National Center for Atmospheric Research, Gavin Schag, San Diego State University
Topics: Remote Sensing, Geographic Information Science and Systems, Biogeography
Keywords: wildfire spread, thermal infrared sensing, topography, fire fuels
Session Type: Paper
Presentation File: No File Uploaded
Spatial data on wildfire rate of spread (ROS) is of interest for: (1) fire suppression and emergency management purposes, (2) reference data for validating fire model simulations, and (3) studying fire behavior. Observations and measurements of fire spread are typically limited to laboratory or controlled burn experiments over relatively small spatial scales, or airborne or satellite observations of fire perimeter expansion at relatively large space and time scales. Through a customized strategy for repetitive capture and geoprocessing of aerial thermal infrared (ATIR) imaging, we are studying fire behavior, estimating ROS, and analyzing landscape covariates with ROS for several large-scale wildfires. We will provide an overview of the processing flow, frequency distributions of ROS estimates, results from statistical relationships between ROS and slope and fuel covariates, and 3-D visualizations for several wildfires that burned in mostly shrubland vegetation in California between 2016 and 2018. Key findings are: (1) ATIR images can be co-located within the size of a pixel, enabling reliable estimates of ROS, (2) frequency distributions of our ROS estimates are useful for characterizing fire behavior and (3) for these study fires, directional topographic slope is a strong predictor of ROS, while image-derived surrogates for fuel load are not, likely due to limited spatial variation in fuel loads within the study areas.