These data were compiled so that annual wildfire could be modelled across the sagebrush region in the western United States. Our goal was to understand how wildfire probability relates to climate and fuel conditions across the entire sagebrush region. To do this we developed a statistical model that represents the relationship between annual wildfire probability and a small number of climate and fuel variables. Specifically, created predictions of wildfire probability using a biologically plausible logistic regression model that related wildfire probability to mean temperature, annual precipitation, the proportion summer precipitation (PSP), and aboveground biomass of annual herbaceous plants and perennial herbaceous plants. The biomass variables were used as proxies for fine fuel availability. These data represent annual fire occurrence in 1 km pixels (i.e. did a given pixel burn that year), predicted wildfire probability, as well as the three year running average (i.e. average across the current and previous two years) of climate and vegetation variables. These data were collected across the sagebrush region (the extent of the study area is provided by the cell_number_ids.tif file). The climate and vegetation data were compiled using a existing gridded dataset (Daymet) of daily precipitation and temperature, and vegetation data were summaries of annual estimates of aboveground biomass of annual and perennial herbaceous plants from the Rangeland Analysis Platform (https://rangelands.app/). These data can be used to understand spatial and temporal variability in wildfire occurrence and modelled wildfire probability between 1988 and 2019 and how that variability relates to spatial and temporal variability in climate and vegetation.

Data sets used in the analysis presented in the manuscript “Characterizing grassland fire activity in the Flint Hills region and air quality using satellite and routine surface monitor data”. The datasets used for the analysis include data from routine monitor networks located in the central U.S., satellite fire detection data, and burn area estimates. The data supporting each of the Figures in the manuscript are provided in a file specific for that Figure, so there is one file for each Figure. Each file is in comma-separated value (csv) format and contains observation data used to generate that Figure. Meta data on what is included in each file is provided in the Data Dictionary. This dataset is associated with the following publication: Baker, K., S. Koplitz, K. Foley, L. Avey, and A. Hawkins. Characterizing grassland fire activity in the Flint Hills region and air quality using satellite and routine surface monitor data. SCIENCE OF THE TOTAL ENVIRONMENT. Elsevier BV, AMSTERDAM, NETHERLANDS, 659: 1555-1566, (2019).

This dataset includes volatility information derived from various biomass burning samples. Emission factors, volatility distributions, mass fractions, and response factors are reported. A data dictionary is provided to define each variable in the dataset. This dataset is associated with the following publication: Sinha, A., I. George, A. Holder, W. Preston, M. Hays, and A. Grieshop. Development of Volatility Distribution for Organic Matter in Biomass Burning Emissions. Environmental Science: Atmospheres. Royal Society of Chemistry, Cambridge, UK, 0000, (2022).

This product ("Prairie fires") presents burned area boundaries for The Flint Hills Ecoregion (KS and OK), one of the most fire prone ecosystems in the United States where hundreds of thousands of acres burn annually as prescribed fire and wildfire. The prairie fire products provide the extent of larger prairie fires in the Flint Hills to record the occurrence of fire and can be used to identify individual burned areas within the perimeters. This product is published to provide fire information of the most fire prone ecosystems to individuals and land management communities for assessing burn extent and impacts on a time sensitive basis. The methods used to produce the prairie fire products from 2019 to present are different than Monitoring Trends in Burn Severity Program (MTBS) methods. The product is developed by running a classification tree model on Landsat and Sentinel imagery for all available image dates with visible fires and without greater than 80 percent cloud cover in the spring of each year. The model takes each image, uses all Landsat bands 2-7 or Sentinel 2b bands 2-4, 8, 11, and 12, and finds thresholds between burnt and unburnt areas to create perimeters. Fire perimeters are created by the model and no manual editing is performed. Thus, these data are 100 percent (model based) auto-generated, however, analysts do review and remove small polygons less than 3 acres. The Prairie Fire dataset will include multi-part polygons and have one record for each source image date. These new methods are optimized to efficiently map and characterize the large number of fires that occur in this region on an annual basis. Prior to 2019, the standard MTBS fire mapping methods were used. Because of the unique frequency and extent of fire in this prairie biome, these fire products are now delivered through the Burn Severity Portal and are no longer included as part of the MTBS products unless a fire is identified in IRWIN, NFPORS or a legacy federal fire occurrence database. The provided data products will vary slightly based on the mapping methodology applied at the time of fire occurrence (pre-2019 or 2019 and later). This map layer is a vector point shapefile of fires occurring three acres and greater in size between calendar year 2009 and 2024 for the Flint Hills Ecoregion.

This product ("Prairie fires") presents burned area boundaries for The Flint Hills Ecoregion (KS and OK), one of the most fire prone ecosystems in the United States where hundreds of thousands of acres burn annually as prescribed fire and wildfire. The prairie fire products provide the extent of larger prairie fires in the Flint Hills to record the occurrence of fire and can be used to identify individual burned areas within the perimeters. This product is published to provide fire information of the most fire prone ecosystems to individuals and land management communities for assessing burn extent and impacts on a time sensitive basis. The methods used to produce the prairie fire products from 2019 to present are different than Monitoring Trends in Burn Severity Program (MTBS) methods. The product is developed by running a classification tree model on Landsat and Sentinel imagery for all available image dates with visible fires and without greater than 80 percent cloud cover in the spring of each year. The model takes each image, uses all Landsat bands 2-7 or Sentinel 2b bands 2-4, 8, 11, and 12, and finds thresholds between burnt and unburnt areas to create perimeters. Fire perimeters are created by the model and no manual editing is performed. Thus, these data are 100 percent (model based) auto-generated, however, analysts do review and remove small polygons less than 3 acres. The Prairie Fire dataset will include multi-part polygons and have one record for each source image date. These new methods are optimized to efficiently map and characterize the large number of fires that occur in this region on an annual basis. Prior to 2019, the standard MTBS fire mapping methods were used. Because of the unique frequency and extent of fire in this prairie biome, these fire products are now delivered through the Burn Severity Portal and are no longer included as part of the MTBS products unless a fire is identified in IRWIN, NFPORS or a legacy federal fire occurrence database. The provided data products will vary slightly based on the mapping methodology applied at the time of fire occurrence (pre-2019 or 2019 and later). This map layer is a vector point shapefile of fires occurring three acres and greater in size between calendar year 2009 and 2024 for the Flint Hills Ecoregion.

This product ("Prairie fires") presents burned area boundaries for The Flint Hills Ecoregion (KS and OK), one of the most fire prone ecosystems in the United States where hundreds of thousands of acres burn annually as prescribed fire and wildfire. The prairie fire products provide the extent of larger prairie fires in the Flint Hills to record the occurrence of fire and can be used to identify individual burned areas within the perimeters. This product is published to provide fire information of the most fire prone ecosystems to individuals and land management communities for assessing burn extent and impacts on a time sensitive basis. The methods used to produce the prairie fire products from 2019 to present are different than Monitoring Trends in Burn Severity Program (MTBS) methods. The product is developed by running a classification tree model on Landsat and Sentinel imagery for all available image dates with visible fires and without greater than 80 percent cloud cover in the spring of each year. The model takes each image, uses all Landsat bands 2-7 or Sentinel 2b bands 2-4, 8, 11, and 12, and finds thresholds between burnt and unburnt areas to create perimeters. Fire perimeters are created by the model and no manual editing is performed. Thus, these data are 100 percent (model based) auto-generated, however, analysts do review and remove small polygons less than 3 acres. The Prairie Fire dataset will include multi-part polygons and have one record for each source image date. These new methods are optimized to efficiently map and characterize the large number of fires that occur in this region on an annual basis. Prior to 2019, the standard MTBS fire mapping methods were used. Because of the unique frequency and extent of fire in this prairie biome, these fire products are now delivered through the Burn Severity Portal and are no longer included as part of the MTBS products unless a fire is identified in IRWIN, NFPORS or a legacy federal fire occurrence database. The provided data products will vary slightly based on the mapping methodology applied at the time of fire occurrence (pre-2019 or 2019 and later). This map layer is a vector point shapefile of fires occurring three acres and greater in size between calendar year 2009 and 2024 for the Flint Hills Ecoregion.

High time resolution (10 s) chamber study burn emission measurements and commercial laboratory fuel analysis reports. This dataset is associated with the following publication: Urbanski, S., R. Long, H. Halliday, A. Habel, E. Lincoln, and M. Landis. Fuel layer specific pollutant emission factors for fire prone forest ecosystems of the western U.S. and Canada. Atmospheric Environment: X. Elsevier B.V., Amsterdam, NETHERLANDS, 0000, (2022).

This data set provides peatland landcover classification maps, fire progression maps, and vegetation community biophysical data collected from areas that were burned by wildfire in 2014 in the Northwest Territories, Canada. The peatland maps include peatland type (bog, fen, marsh, swamp) and level of biomass (open, forested). The fire progression maps enabled an assessment of wildfire progression rates at a daily time scale. Field data, collected in 2015, include an estimate of burn severity, woody seedling/sprouting data, soil moisture, and tree diameter and height of burned sites and similar vegetation characterization at landcover validation sites.

The dataset contains spectral indices values calculated from Landsat imagery, normalized difference vegetation index (NDVI) and delta normalized burn ratio (NBR, deltaNBR). The spectral indices were computed for 1472 forested sites in the Glacier National Park, Montana, over the 1984-2021 period, in areas that were burned or not burned during the Robert Fire of 2003.

These data consist of plot-level plant species cover measurements from numerous targeted post-fire vegetation studies across the western United States. This data release includes two data tables. The first data table: 'postfire_vegplot_dataset.csv', consists of absolute percent live foliar cover measurements of all plant species within plots from targeted post-fire vegetation studies, or 'datasets', across the western United States. The second data table: 'dataset_information_table.csv', lists any citations, links to related publications, or other notes about data collection for specific datasets.