Wildland fires emit significant amounts of greenhouse gases, particulate matter, and ozone precursors which have significant negative effects on public health at multiple scales. In order to mitigate these impacts, state agencies require daily air quality forecasts to minimize exposure risk. Air quality analyses are also necessary to quantify the contribution of fires to regional air pollution and thereby support the development of effective and efficient emission controls for industrial, power generation, and transportation sources. In addition to air quality forecasting and analyses, burned area maps are invaluable tools used by emergency response teams, which often include hydrologists, wildlife biologists, soils scientists, geologists, ecologists, engineers, foresters, botanists, and GIS specialists, and which assess threats to life, property, and natural resources in the days and weeks immediately following a fire. The availability of timely, comprehensive, and consistent burned area estimates can improve fire and forest management decisions and lead to better fire emission estimates and subsequent air quality forecasts and air regulatory strategies.
Currently, the MODIS sensor on the polar-orbiting Terra and Aqua satellites provides burned area products (the satellites’ orbits provide two local overpasses each day - one nighttime and one afternoon). However, the aging MODIS sensors have exceeded their expected lifetime and a longer-lasting data solution is needed. The Visible Infrared Imaging Radiometer Suite (VIIRS) sensor onboard the Suomi-National Polar-orbiting Partnership satellite (S-NPP) is the first of the next generation of sensors that will replace MODIS. To address this need a near real-time burned area detection algorithm has been developed for the VIIRS sensor. The algorithm combines VIIRS observations of surface reflectance and thermal anomalies (also called “hot spots” or “active fire detections”) to provide a map of burned area for the western United States shortly after the sensor’s afternoon overpass. The algorithm’s parameters were optimized using Monitoring Trends in Burn Severity (MTBS) annual fire boundaries from 2013 as “ground truth”. The performance of the algorithm was evaluated against MTBS data using the spatial metric “union overlap” as our figure of merit (FOM), as illustrated in Figure 1 for the Beaver Creek Fire (Sawtooth National Forest, Idaho, August 2013). The possible range of FOM is 0 – 1.0, where 1.0 would be perfect overlap between the ground truth and the VIIRS burned areas. Overall, the optimized algorithm performed extremely well with FOM = 0.69 for a combined 212 fire events. Additionally, the overall omission and commission errors were only 15% and 21%, respectively.
The USDA Forest Service Remote Sensing Application Center (RSAC) will begin running the VIIRS algorithm operationally for the 2018 western US fire season. The VIIRS based burned area maps will be included as a regular fire support product offered by the RSAC Active Fire Mapping Program. The RMRS Fire, Fuel, and Smoke Program’s Smoke Emission and Dispersion research team will use the VIIRS near real-time burned area to provide daily inventories of pollutant emissions from western wildfire for use in the smoke dispersion and air quality forecasting activities of fire incident air resource advisors, state environmental agencies, and scientific researchers. The timely and comprehensive burned area estimates provided by this new VIIRS algorithm will improve fire and forest management decisions and lead to better air quality forecasts and air regulatory strategies.
