Relationships between Daily Fire Danger and Daily Fire activity Characterized across the Continental U.S. Using the MODIS Active Fire and Burned Area Products

The computational process of the National Fire Danger Rating System (NFDRS) inputs fuels, weather, and topographic variables, and outputs a suite of daily components and indexes that reflect the "worst-case" burning conditions averaged over large areas. NFDRS components and indexes were formulated to be physically interpretable with regard to fire occurrence and fire behavior, and as such have measurement units that coincide with a fire's rate of spread, heat release per area, and flame length. Fire occurrence and fire behavior in the U.S. have long been summarized using the daily number of newly discovered fires and their end-of-the-season final fire size. Unfortunately the use of just two metrics to characterize fire activity, and the historical lack of spatial, temporal and descriptive information recorded on agency fire reports, has limited the prospective ability of the NFDRS to indicate more than just the probability of fire occurrences and total burned area. Until now associations between NFDRS indexes and alternative descriptors of fire activity have not been possible. Since 2001, however, the Moderate Resolution Imaging Spectroradiometer (MODIS), flown aboard NASA's Terra and Aqua satellites, has been collecting multi-spectral images of landscape fire activity across the continental U.S. Here we relate daily values of the Energy Release Component (ERC) with an expanded set of daily fire metrics retrieved from 10 years of the MODIS Direct Broadcast (DB) burned area products (MCD64A1) and MODIS active fire products (MCD14ML). In general the results reveal that an increase in fire danger is almost universally accompanied by an increase in (i) the probability of detecting a burned area pixel or an active fire pixel, (ii) the daily burned area, (iii) the daily sum of fire radiative power, FRP, (iv) the mean FRP per active fire pixel, and (v) the nighttime fraction of FRP. Moreover, the form and strength of the relationships between fire danger and fire activity are geographically dependent, suggesting that different fuels and vegetation complexes (stratified by ecoregions) are differentially sensitive to in-season fire weather conditions. Interpreting these relationships, however, is complicated by (i) an expanding window of realized fire activity as fire danger increases, and (ii) differences in the detection strategies and performance of the two satellite-based fire products. It is anticipated that results herein will add a finer temporal dimension to our understanding of cross-scale, ecoregion dependent fire-climate relationships.