Western US MTBS-Interagency (WUMI) wildfire dataset

Area burned is an important variable for measuring wildfire activity. In the western United States (US), the timing and magnitude of area burned can be associated with meteorological and human activity to find the drivers of wildfire activity, but this type of research is dependent on the spatial and temporal resolution of available wildfire datasets. The Western US MTBS-Interagency (WUMI2) database is a dataset of wildfire events in the western United States (US) larger than 1 km2 for 1984 to 2020. WUMI2 includes the important Monitoring Trends in Burned Severity (MTBS) project (Eidenshink et al., 2007)—a Landsat satellite-based dataset of large fires (>4.04 km2)—and adds small (>1 to 4.04 km2) and large fires from government agency databases, including from the Fire Program Analysis (FPA) fire-occurrence database (Short et al., 2022). We performed extensive quality control to merge the datasets together and remove errors. The result is a western US-wide dataset with accurate fire frequency, timing, and area burned that can be used for analyses and modeling of wildfire activity. The current version of this data is WUMI2. The first iteration of the dataset (WUMI1) was published and described in Juang et al. (2022).

Methods
Version WUMI2
Updated August 1, 2024: Our WUMI2 fire database consists of 21,693 western US fire events from 1984 through 2020. A text file (west_US_fires_1984-2020_WUMI2.txt) provides a list of each fire event, including the fire’s name, discovery date, point location, total area burned, and forested area burned (see the corresponding readme.txt file for column labels). We also include NetCDF files of the 1-km map of forest fractional coverage (forest_type_frac.nc) and the 1-km maps of monthly burned area over 1984–2020 (burnarea_1984-2020_WUMI2.nc). Fires included in this database are from the Monitoring Trends in Burned Severity Product (MTBS) (Eidenshink et al., 2007), the Fire Program Analysis fire-occurrence database (FPA FOD 6th edition) of interagency fires (Short, 2022), and interagency fires from local databases (CalFire, ST/C&L, TRIBE), and interagency fires from government agency databases (BIA, BLM, BOR, DOD, DOE, NPS, FWS, FS, NPS). More information on methodology can be found in the Supporting Information in Juang et al. (2022). In addition to this methodology, the Fire Program Analysis fire-occurrence database (FPA FOD 6th edition) (Short, 2022) replaces our WUMI1 (Juang et al. (2022)) methodology for the government interagency fires from 1992-2020 for version WUMI2. As in WUMI1, we performed extensive quality control across all included datasets to remove errors in the various wildfire databases and merge the datasets together. 

Version WUMI1 (older)
Updated August 16, 2021: Our WUMI1 fire database consists of 18,368 western US fire events from 1984 through 2019. A text file (west_US_fires_1984_2019.txt) provides a list of each fire event, including the fire’s name, discovery date, point location, total area burned, and forested area burned (see the corresponding readme.txt file for column labels). We also include NetCDF files of the 1-km map of forest fractional coverage (forest_type_frac.nc) and the 1-km maps of monthly burned area over 1984–2019 (burnarea_1984_2019.nc). Fires included in this database from the Monitoring Trends in Burned Severity Product (MTBS), fires from a state database (CalFire), fires from government interagency databases (BIA, BLM, BOR, NPS, FWS, FS). More information on methodology can be found in the Supporting Information in Juang et al. (2022).