Implementation of a near real-time burned area detection algorithm calibrated for VIIRS imagery

Abstract

There is a need to implement methods for rapid burned area detection using a suitable replacement for Moderate Resolution Imaging Spectroradiometer (MODIS) imagery to meet future mapping and monitoring needs (Roy and Boschetti 2009, Tucker and Yager 2011). The Visible Infrared Imaging Radiometer Suite (VIIRS) sensor onboard the Suomi-National Polar-orbiting Partnership satellite (S-NPP) presents an alternative raw data source (www.nsof.class.noaa.gov) with resolutions comparable to that of MODIS (Boken and others 2010, Murphy and others 2001, Tucker and Yager 2011), and it has already proven capable of standing in for MODIS in active fire detection work (Schroeder and others 2014). Given the specifications of the sensor, it should be possible to replicate the output from MODIS-based burned area detection models using VIIRS imagery (figure 1). With sponsorship from the Remote Sensing Steering Committee (RSSC) and additional funding from the Missoula Fire Sciences Laboratory (FiSL), we have developed an application that applies a MODIS-based model for burned area detection to VIIRS imagery. Given current data access timelines, a burned area product can be delivered shortly (within a few hours) after the S-NPP/VIIRS afternoon overpass. Using the MODIS algorithm as a foundation, VIIRS-specific thresholds have been calibrated to create VIIRS-derived burned area products. These data will soon be made available on the U.S. Department of Agriculture, Forest Service, Remote Sensing Applications Center Active Fire Mapping website.

Citation

Schwert, Brenna; Albury, Carl; Clark, Jess; Schaaf, Abigail; Urbanski, Shawn; Nordgren, Bryce. 2016. Implementation of a near real-time burned area detection algorithm calibrated for VIIRS imagery. RSAC-10092-TIP1. Salt Lake City, UT: U.S. Department of Agriculture, Forest Service, Remote Sensing Applications Center. 5 p.