Improving, Validating, and Teaching ABI Fire Detection

Since the launch of GOES-16, the Advanced Baseline Imager (ABI) Fire Detection and Characterization Algorithm (FDCA) has been simultaneously undergoing algorithm refinement and validation as well as been a key topic in multiple GOES-R trainings for a wide variety of users from NWS Incident Meteorologists (IMETs) to aerosol researchers to broadcasters and more. The FDCA is essentially the Wildfire Automated Biomass Burning Algorithm (WFABBA) that processes data from the generation of geostationary satellites prior to GOES-R and inherited its strengths and weaknesses. The algorithms allow for characterization of the highest confidence fires as well as providing other confidence levels for those with a higher tolerance for false alarms. The refinement of the algorithm has greatly improved initial performance in terms of reducing false alarms, and validation of the fire characteristics, specifically fire radiative power (FRP), has been proceeding. Validation of ABI FDCA FRP data is complicated by the lack of ground truth data, so comparisons are made to the polar orbiting satellites that provide FRP, specifically those carrying the VIIRS and MODIS instruments. Those comparisons need to account for viewing angles and conditions, and adjust for the particular characteristics of ABI radiance data. Several case studies of both well-known fires and run-of-the-mill situations will be presented. The techniques used to teach about fire detection with satellites will also be described. Those training efforts are part of ongoing outreach to the user community, as well as part of the effort to expand the GOES-R L2 product user community beyond the traditional users at NOAA and NWS.