Hydrometeorology-wildfire relationship analysis based on a wildfire bivariate probabilistic framework in different ecoregions of the continental United States

• A probabilistic framework is used to describe wildfire bivariate characteristic. • Soil moisture, precipitation, and temperature dominated wildfire-prone month. • The dominated hydrometeorological factors of wildfire vary with ecoregions. • The Continental United States can be divided into new four wildfire regimes. Wildfires are a natural part of the ecosystem in the U.S.. It is vital to classify wildfires using a comprehensive approach that simultaneously considers wildfire activity (the number of wildfires) and burned area. On this basis, the influence of hydrometeorological variables on wildfires can be further analyzed. Therefore, this study first classified wildfire types using a wildfire bivariate probability framework. Then, by considering six hydrometeorological variables, the dominant hydrometeorological variables for different wildfire types in 17 ecoregions of the United States were quantified. In addition, based on the results of this hydrometeorology-wildfire relationship analysis, we obtained new clusters that simultaneously considered wildfire characteristics and the impact of hydrometeorology on wildfires. In particular, the results were as follows: (1) Through the probability of wildfire bivariate statistical characteristics, wildfires could be classified into five types in this paper: WT-1 (mega-wildfire), WT-2 (joint wildfire-1), WT-3 (joint extremes), WT-4 (joint wildfire-2), and WT-5 (super frequent wildfires); (2) The dominant hydrometeorological variables under different wildfire types were discussed in 17 ecoregions of the United States; and (3) In the four new cluster regions, intensifying droughts are a concern in clusters 1 and 4, while there are multiple concerns in cluster 3, namely, stronger winds, higher temperatures, and more drought.