Search for a command to run...
Wildfires are a major natural hazard causing extensive ecological damage and endangering human survival. Previous studies on wildfires in China have mostly focused on specific regions or individual drivers, with limited systematic assessments at the long-term and national scales. The spatiotemporal patterns of wildfires and their multiple driving mechanisms under China’s diverse climatic regimes remain insufficiently understood. To bridge this gap, we combined MCD64A1 burned area data (2001–2023) with multi-source natural (meteorological, vegetation, and topographic) and anthropogenic factors, using random forest models at both the national and regional scales to examine the spatiotemporal patterns, dominant drivers, and response mechanisms of wildfires in China. The results revealed that: (1) Spatially, wildfires were concentrated in northeastern and southern China, which accounted for 86.20% of the total burned area. Temporally, northern wildfires were primarily a spring-dominated fire regime, with peak activity in March and April, whereas southern wildfires were winter-dominated, peaking in February. (2) At the national scale, elevation was the key topographic factor influencing wildfire occurrence (relative importance = 0.49), with low-elevation and gentle-slope areas being more fire-prone. At the regional scale, the driving factors exhibit spatial differentiation, forming a spatial pattern of topography-dominated and climate-dominated. (3) Partial dependence plot analysis revealed nonlinear and threshold responses. Fire probability increases rapidly when the soil moisture is below 20 mm, while extremely high land surface temperatures in arid regions suppress fire occurrence due to fuel limitations. This study enhances the understanding of spatially heterogeneous wildfire drivers in China and provides a scientific basis for region-specific wildfire prevention and management strategies.