Dataset DOI: 10.5061/dryad.63xsj3vd4

Description of the data and file structure

The WUMI2024a merges wildfire occurrence data provided a number of government sources to produce a comprehensive database of locations, sizes, dates, and maps of western United States (US) wildfire events ≥100 hectares (1 km2) in size from 1984–2024.

Files and variables

File: fire_lists.zip

Description: This directory contains lists of individual fire events. There is a subdirectory for each of the six datasets represented in original_data/. Each subdirectory contains a text-file list of all western US fires ≥10 acres (4.04 ha) that appear in each dataset. Each sub-directory, except for mtbs/ also contains a quality-controlled list (*_qc.txt) of western US wildfires ≥100 ha in area. Quality control mostly consists of eliminating redundancies and also includes correction of some erroneous fire dates and locations. The fire_lists/ directory also contains .txt lists of fires during the process of merging various databases:

wfaip_fpafod_wfigs_calfire_merge.txt: List of western US wildfires ≥100 ha produced by merging the WFAIP (1984–1991), FPA-FOD (1992–2000), and WFIGS (2021–April 2025) with CalFire during 1984–2024.

wfaip_fpafod_wfigs_calfire_mtbs_merge.txt: List of western US wildfires ≥100 ha after merging the list of fires described above (wfaip_fpafod_wfigs_calfire_merge.txt) with those from MTBS.

wfaip_short_wfigs_calfire_mtbs_merge_2.txt: List of western US wildfires ≥100 ha after linking fires in the list described above (wfaip_fpafod_wfigs_calfire_mtbs_merge.txt) with 2018–2020 WFIGS fires.

wfaip_fpafod_wfigs_calfire_mtbs_usgs.txt: List of western US wildfires ≥100 ha after linking fires in the list described above (wfaip_short_wfigs_calfire_mtbs_merge_2.txt) with 1984–2020 USGS fires.

wfaip_fpafod_wfigs_calfire_mtbs_usgs_iafph.txt: Final list of all western US wildfires ≥100 ha after linking fires in the list described above (wfaip_fpafod_wfigs_calfire_mtbs_usgs.txt) with 1984–2023 IAFPH fires. Each fire is assigned a unique identifier (fireid) composed of the start date (yyyymmdd). North latitude multiplied by 10,000 and rounded, and West longitude multipled by 10,000 and rounded (e.g., a fire ignited on Jan 26, 1984 at 33.709°N, 117.608°N would be given a fireid of 19840126_337090_1176180.

File: original_data.zip

Description: This directory contains the original data files that were used to produce the WUMI2024a. Brief descriptions of each data source are:

iafph.zip: The National Interagency Fire Center provides the Interagency Fire Perimeter History (IAFPH) dataset is a database of fire perimeters for the United States beginning in the 1800s and extending through 2023. This database provides information on the year of the fire but not the start date and exhibits a large positive trend in the number of fires represented over time that is due largely to changes in data availability. We accessed the dataset from https://data-nifc.opendata.arcgis.com/datasets/nifc::interagencyfireperimeterhistory-all-years-view/about as a shapefile on October 15, 2024.

mtbs/: The US Forest Services Monitoring Trends in Burn Severity (MTBS) provides Landsat-based maps with 30-m resolution of burned area for large fires in the continental US (≥404 ha in the western US). At the time of production of the WUMI2024a, the MTBS dataset was complete through 2022 and had data for a limited number of fires in 2023–2024. We accessed the data from https://www.mtbs.gov/ and obtained the most recent data available for 2023–2024 on April 26, 2025. For each fire, the MTBS dataset contains a 30-m resolution geotiff of classified burn severity, which can be turned into a 30-m map of burned/unburned classifications, the underlying satellite observations used to assess burn severity, and a shapefile of the burned-area boundaries. Note that this mtbs/ directory contains a .zip file for each US state. The code provided uncompresses the .zip files to state-specific directories and each state’s directory contains a sub-directory for each year with MTBS data.

usgs.zip: The United States Geological Survey (USGS) produced the Combined wildland fire dataset for the United States and certain territories, which is a database of fire perimeters from the 1800s though 2020. This dataset is compiled from many disparate data sources and heavily quality controlled to remove duplicate entries, for each fire archive the metadata from each data source that provided information about that fire, and prodcue a single best-estimate of the fire perimeter based on all the fire perimeter maps available. We accessed the USGS dataset on June 9, 2025 from https://doi.org/10.5066/P9ZXGFY3.

wfaip.zip: The Wildland Fire Application Portal (WFAIP) hosts a lists of fire incidents addressed by the US Forest Service (FS) and Fish and Wildlife Service (FWS) from 1972–2020, and by the Bureau of Land Management (BLM), Bureau of Indian Affairs (BIA), Bureau of Reclamation (BOR), and National Park Service (NPS) from 1972–2017. We accessed these data from https://www.wildfire.gov/page/zip-files on December 15, 2024 in the form of text files. Within the wfaip directory there is a sub-directory containing the text file(s) for each agency listed above.

wfgis.zip: The National Interagency Fire Center’s Wildland Fire Interagency Geospatial Services (WFIGS) provides a comprehensive dataset of wildfire boundaries that begins in 2018 and is updated in real time, though is not complete prior to 2021. We accessed the WFIGS dataset as a shapefile from https://data-nifc.opendata.arcgis.com/datasets/nifc::wfigs-interagency-fire-perimeters/about on April 26, 2025. The final fire in the WFIGS dataset that we processed began on April 21, 2025.

File: code_wumi2024a.zip

Description: This directory contains the matlab code used to read the data files provided in original_data/, perform quality control, and ultimately merge the datasets to produce the final list of western US wildfire events, maps of area burned by each event, and shapefiles of fire boundaries. Note that some of the code in the files below call to original database files that could not be shared publicly here due to copyright. Those datasets are the Calfire FRAP database of fire perimeters in California, the FPA FOD v6 list of fire occurrences, and the MODIS FIRED dataset of event-specific burned area maps. The code files that call these datasets describe to the user how to access the data from the original data sources.

File names and descriptions are:

01_make_fire_lists.m: Produces a list of western US wildires ≥10 acres (4.04 ha) in .txt format for each of the six datasets represented in original_data/.

02_quality_control_WFAIP_dataset.m: Make quality controlled list of western US wildfires ≥100 ha from 1984–1991 from the WFAIP dataset.

03_quality_control_fpa_fod_dataset.m: Make quality controlled list of western US wildfires ≥100 ha from 1992–2020 from the FPA-FOD dataset.

04_quality_control_calfire_dataset.m: Make quality controlled list of California wildfires ≥100 ha from 1984–2024 from the CalFire dataset.

05_quality_control_wfgis_dataset.m: Make quality controlled list of California wildfires ≥100 ha from 2018–2024 from the WFGIS dataset.

06_quality_control_iafph_dataset.m: Make quality controlled list of western US wildfires ≥100 ha from 2000–2018 from the IAFPH dataset.

07_quality_control_usgs_dataset.m: Make quality controlled list of western US wildfires ≥100 ha from 2000–2018 from the USGS dataset.

08_merge_wfaip_fpafod_and_nifc_with_calfire.m: Merge the list of CalFire fires with the WFAIP, FPA-FOD, and WFIGS datasets in California, prioritizing CalFire fires in cases of duplicate and, in the case of complex fires in the CalFire dataset (large fires composed of multiple smaller fires that merged), keeping track of the smaller fires.

09_merge_wfaip_fpafod_nifc_and_calfire_with_mtbs.m: Merge the list of WFAIP, FPA-FOD, WFIGS, and CalFire fires produced above with the list of MTBS fires, prioritizing MTBS in cases of duplicates and, in the case of complex fires in the MTBS dataset (large fires composed of multiple smaller fires that merged), keeping track of the smaller fires.

10_link_fires_in_master_dataset_to_wfigs_pre2021.m: Identify matches between the fires in the list produced above and WFIGS fires from the pre-2021 period when WFIGS is incomplete, as, in cases of matches with FPA-FOD fires, the WFIGS dataset provides unique fire-boundary data.

11_link_fires_in_master_dataset_to_usgs.m: Identify matches between the fires in the list produced above and USGS fires, as, in cases of matches with WFAIP and FPA-FOD fires, the IAFPH dataset provides unique fire-boundary data.

12_link_fires_in_master_dataset_to_iafph.m: Identify matches between the fires in the list produced above and IAFPH fires, as, in cases of matches with FPA-FOD fires, the IAFPH dataset provides unique fire-boundary data.

13_make_1km_maps_of_circular_burned_area.m: For each fire in the final wildfire list, make a shapefile with the hypothetical fire perimeter assuming the fire is circular and centered on the ignition location as well as a netcdf containing a 1-km resolution gridded map of fractional area burned assuming all area within the perimeter is uniformly burned.

14_make_1km_maps_of_calfire_burned_area.m: For each CalFire fire in the final wildfire list, make a shapefile with the fire perimeter as well as a netcdf containing a 1-km resolution gridded map of fractional area burned assuming all area within the perimeter is uniformly burned. Perimeter shapes and sizes and the and grids of area burned are adjusted so that areas of open water, bare soil, rock, or ice are not included in the burned areas.

15_make_1km_maps_of_iafph_burned_area.m: For fire in the final wildfire list that is linked to an IAFPH fire, make a shapefile with the fire perimeter as well as a netcdf containing a 1-km resolution gridded map of fractional area burned assuming all area within the perimeter is uniformly burned.

16_make_1km_maps_of_wfigs_burned_area.m: For each WFIGS fire in the final wildfire list, make a shapefile with the fire perimeter as well as a netcdf containing a 1-km resolution gridded map of fractional area burned assuming all area within the perimeter is uniformly burned.

17_make_1km_maps_of_usgs_burned_area.m: For fire in the final wildfire list that is linked to an IAFPH fire, make a shapefile with the fire perimeter as well as a netcdf containing a 1-km resolution gridded map of fractional area burned assuming all area within the perimeter is uniformly burned.

18_make_1km_maps_of_mtbs_burned_area.m: For each MTBS fire in the final wildfire list, make a shapefile with the fire perimeter as well as a netcdf containing a 1-km resolution gridded map of fractional area burned, calculated from the 30-m resolution MTBS maps of fire-severity classifications (summing up the area represented by grid cells classified as mild, moderate, or high severity). The maps of burned area and fire perimeters are written to the fire_lists/ directory.

19_make_final_lists_of_wildfires.m: This code produces final WUMI2024a lists of wildfires ≥100 ha and writes them to WUMI2024a_final_wildfire_lists.zip.

20_make_monthly_1km_maps_area_burned.m: Make a netcdf of monthly 1-km resolution maps of wildfire area across the western US for 1984–2024 associated with each distinct fire event. In the case of complex fires, only the event representing the final fire complex is mapped and listed, not the data associated with the smaller fire events that that merged.

21_zip_fire_maps.m: This code archives the many directories with maps of individual fire events into annual zip files within fire_maps/.

File: projection.zip

Description: Geotiff containing the map projection used for the 1-km gridded maps of area burned and the shapefiles with fire perimeters. Projection is NAD83/Conus Albers (EPSG:5070).

File: study_region.zip

Description: This directory contains a netcdf file with a 1-km gridded map of the western United States study region, which is composed of the 11 westernmost states in the continental United States.

File: WUMI2024a_final_wildfire_lists.zip

Description: This directory contains the final lists of distinct wildfire events:

WUMI2024a_wildfires_1984_2024_no_subfires.txt: Final list of distinct western US wildfire events ≥100 ha from 1984–2024. In the cases of fire events composed of multiple smaller sub-fires (e.g., fire complexes composed of multiple smaller fires that merged), only the parent event is listed but not the sub-fires.

WUMI2024a_wildfires_1984_2024_with_subfires.txt: Final list of distinct western US wildfire events ≥100 ha from 1984–2024, where parent events composed of multiple sub-fires are replaced by the sub-fire events. In these cases, the areas burned by sub-fires are rescaled to sum up to equal the area burned indicated in the parent event’s burned area map and ignition locations of sub-fires are adjusted, if necessary, to the nearest location indicated as burned by the parent event’s burned area map.

subfires/: For each fire listed in WUMI2024a_wildfires_1984_2024_no_subfires.txt that is found to be a parent fire composed of multiple subfires, there is a text file named after the parent fire’s fireid. Each of these text files provides in the first row the information about the parent fire, which also appears in WUMI2024a_wildfires_1984_2024_no_subfires.txt. Below the top row, a line is skipped and the information about each sub-fire is provided with a skipped line between each sub-fire. In some cases, a sub-fire is composed of its own sub-fires (sub-sub-fires). In these cases, the sub-sub fires are listed immediately below their parent sub-fire without any line skips.

Cases in which non-quantitative data (including fire identification codes) are not relevant for a given fire, entries of "N/A" appear.

In text files of fire lists, column-label descriptions are:

fireid: unique identifier composed of the start date (yyyymmdd) ‘’ North latitude multiplied by 10,000 and rounded ‘’ and West longitude multipled by 10,000 and rounded (e.g., a fire ignited on Jan 26, 1984 at 33.709°N, 117.608°N would be given a fireid of 19840126_337090_1176180.

dataset: The primary dataset used to determine the fire’s name and start date (mtbs, calfire, wfaip, fpa_fod, or wfigs).

agency: The government agency, as indicated by the dataset, as the original source of the fire report.

name: fire name

year: year of fire start

month: month of fire start

day: day of month of fire start

lat: North latitude of ignition location

lon: West longitude of ignition location

lat_ll: North latitude of lower-left (southwest) corner of the bounding box of fire perimeter

lon_ll: West longitude of lower-left (southwest) corner of the bounding box of fire perimeter

lat_ur: North latitude of upper-right (northeaset) corner of the bounding box of fire perimeter

lon_ur: West longitude of upper-right (northeaset) corner of the bounding box of fire perimeter

poly_area_ha: Area (hectares) within the fire perimeter

burn_area_ha: Area burned (hectares) within the fire perimeter according the MTBS Landsat-derived classification of burn severity (we consider any grid cell classified as mild, moderate, and severe to have burned)

mtbs_name: Name of fire according to MTBS

mtbs_ID: MTBS fire identification code

irwinid: IRWIN ID identification code

FOD_ID: Fire occurrence dataset identifier provided by FPA-FOD

FOA_ID: US Forest Service’s Fire Program Analysis identifier provided by FPA-FOD

object_ID_wfaip: This number corresponds to the object_ID column in fire_lists/wfaip/wfaip_fires_qc.txt.

object_ID_fpafod: This number corresponds to the object_ID column in fire_lists/fpa_fod/fpa_fod_fires_qc.txt, which corresponds to the OBJECTID attribute in the shapefile of fire occurrences we produced from the FPA-FOD version 6 geodatabase of United States wildfire occurrences.

object_ID_wfigs: This number corresponds to the object_ID column in fire_lists/wfigs/wfigs_fires_qc.txt, which corresponds to the OBJECTID attribute in the WFIGS shapefile of fire perimeters we accessed from NFIC.

object_ID_calfire: This number corresponds to the object_ID column in fire_lists/calfire/calfire_fires_qc.txt, which corresponds to the OBJECTID attribute in the shapefile of wildfire perimeters we produced from the CalFire FRAP version 2024 geodatabase.

object_ID_usgs: This number corresponds to the object_ID column in fire_lists/usgs/usgs_fires_qc.txt, which corresponds to the OBJECTID attribute in the shapefile of wildfire perimeters we produced from the USGS combined wildfires geodatabase.

object_ID_iafph: This number corresponds to the object_ID column in fire_lists/iafph/iafph_fires_qc.txt, which corresponds to the OBJECTID attribute in the IAFPH shapefile of fire perimeters we accessed from NFIC.

cause_human_or_natural: General cause of ignition (Human, Natural, or Unknown). Note that "N/A" indicates that the fire report came from a database that does not provide information on fire cause. In contrast, an entry of "Unknown" indicates that the fire report came from a database that does provide fire-cause information, but that the cause was unknown.

cause_specific: Specific cause of ignition (from FPA-FOD and CalFire). Note that "N/A" indicates that the fire report came from a database that does not provide information on fire cause. In contrast, an entry of "Unknown/Unidentified" indicates that the fire report came from a database that does provide fire-cause information, but that the cause was unknown.

File: fire_maps.zip

Description: This directory contains annual .zip file that uncompress to annual directories containing a sub-directory for each fire event from that year listed in wfaip_short_wfigs_calfire_mtbs_iafph_final.txt. The fire-specific sub-directory names correspond to each fire’s fireid. Each fire’s sub-directory contains a 1-km resolution gridded map (netcdf) of fractional area burned and a shapefile with the fire perimeter assuming the fire is a circle centered on the ignition location. If the fire was represented by a database that includes fire perimeters (MTBS, CalFire, WFIGS, AIFPH, USGS) and/or high-resolution maps of area burned within perimeters (MTBS), then the sub-directory also contains a pair of netcdf and shapefile files that corresponds to each of those datasets. The fire_maps/ directory also contains the netcdf (burnarea.nc) of monthly 1-km resolution maps of western US area burned 1984–2024 from the fires listed in WUMI_wildfires_1984_2024_no_subfires.txt, which was produced with 17_make_monthly_1km_maps_area_burned.m. Map projection is NAD83/Conus Albers (EPSG:5070).

Code/software

The code.zip directory contains the Matlab 2024a code used to produce the WUMI2024a dataset.

Access information

Data was derived from the following sources:

• CalFire FRAP: https://www.fire.ca.gov/what-we-do/fire-resource-assessment-program/fire-perimeters
• PFA-FOD: https://www.fs.usda.gov/rds/archive/catalog/RDS-2013-0009.6
• Interagency Fire Perimeter History: https://data-nifc.opendata.arcgis.com/datasets/nifc::interagencyfireperimeterhistory-all-years-view/about
• MTBS: https://www.mtbs.gov/
• USGS Combined Fire Perimeters: https://doi.org/10.5066/P9ZXGFY3
• The Wildland Fire Application Portal interagency fires: https://www.wildfire.gov/page/zip-files
• The National Interagency Fire Center’s Wildland Fire Interagency Geospatial Services (WFIGS) fire perimeters: https://data-nifc.opendata.arcgis.com/datasets/nifc::wfigs-interagency-fire-perimeters/about
• USGS National Land Cover Database: https://www.usgs.gov/centers/eros/science/annual-national-land-cover-database

NOTE: Original datasets that could not be provided here

The code used to produce the WUMI2024a called to a number of publicly available data files, but not all of those data files could be shared in this database due to licensing.

calfire.zip: The California Department of Forestry and Fire Protection (CalFire) maintains a database of fire occurrences and burned-area perimeters in California through its Fire and Resource Assessment Program (FRAP). The dataset is updated annually and we accessed the most recent version, which is complete through 2024, from https://www.fire.ca.gov/what-we-do/fire-resource-assessment-program/fire-perimeters as a geodatabase on April 26, 2025. The data file downloaded was called fire241gdb.zip. We uncompressed the file and used the geodatabase to produce a shapefile of wildfire perimeters in QGIS called firep24_wf.shp with an equal area NAD83/Conus Albers (EPSG:5070) projection. The shapefile was then compressed into a .zip file, calfire.zip, and put in the original_data/ directory.

fpa_fod.zip: The US Forest Service maintains a database of wildfire occurrence data (FOD) for the United States in support of the Fire Program Analysis (FPA). At the time of production of the WUMI2024a the dataset covered 1992–2020, which we accessed from https://www.fs.usda.gov/rds/archive/catalog/RDS-2013-0009.6 as a geodatabase on November 4, 2022. We downloaded the FPA FOD v6 geodatabase file (RDS-2013-0009.6_Data_Format2_GDB.zip) from https://doi.org/10.2737/RDS-2013-0009.6, uncompressed, used QGIS to convert the geodatabase file (FPA_FOD_20221014.gdb) to a shapefile (short_1992_2020.shp) with a geographic projection of WGS 84 (EPSG:4326), compressed into a zip file (fpa_fod.zip), and put in the original_data/ directory.

fired.zip: Fire events database (FIRED) developed from the satellite-derived MODIS v6.1 burned area product, beginning in November 2000 and extending through August 2024. This database provides information on fire event start dates and perimeters based on gridded binary fire detections with a ~500-m spatial resolution. We accessed the dataset from https://doi.org/10.25810/37by-q597 on July 15, 2025. This dataset was only used to make figures for the paper describing the WUMI2024a, published in Earth System Science Data. The downloaded file was uscan.zip. We uncompressed, used QGIS to reproject the shapefile called fired_uscan_2000_to_2024_events_reproj.shp to the WGS 84 (EPSG:4326) projection, saved as fired_uscan_2000_to_2024_events_reproj.shp, compressed into a zip file (fired.zip), and put in the original_data/ directory.