An integrated wildfire event database for improved fire science research for the continental United States, 1992–2021

Abstract. This paper describes a wildfire event dataset that was initially created by mining information from three existing national datasets that have provided the foundation for numerous wildfire studies in the continental United States over the past two decades. The goal for creating this dataset was to address existing uncertainties and ambiguities in the the existing datasets by using information from regional reports and datasets and searches of the world wide web. Wildfire events not contained in the three datasets were added as new records. Approaches were developed to harvest information for records where there were significant differences in the size of the same fire from different datasets. Specific attention was paid to fires named as complexes, those contained multiple fires, smaller events that had merged to form larger events, events that had joined, events where the size was incorrectly reported, and duplicate events. This new wildfire event dataset provides more accurate information on the number and sizes of individual wildfires and improved information for quantifying annual area burned and changes in extreme wildfire activity over a three-decade period (1992 to 2021) at different spatial scales. Results are presented that illustrate the differences in wildfire characteristics developed from the new dataset and those currently being used by the fire science community.