Global burn severity in forest ecoregions: trends, climate drivers, and predictive insights
Kang He, Xinyi Shen, Emmanouil N. Anagnostou
Abstract
Forest fires play an important role in shaping ecosystems but increasingly threaten biodiversity and carbon dynamics. Advancements in remote sensing and machine learning have facilitated the monitoring and prediction of burn severity, yet a comprehensive global assessment of severity trends and their climatic drivers remain limited. In this study, we analyzed burn severity trends across global ecoregions from 2003 to 2023 using a newly developed 30-m resolution Global Burn Severity Dataset. We observed significant increases in forest fire burn severity in tropical and subtropical regions, while boreal zones exhibited declines. For ecoregions showing significant trends, we developed predictive models using XGBoost and 14 climate variables from the TerraClimate dataset to identify key drivers. Our analysis highlighted the influence of regional climate variables on wildfire dynamics. These findings enhance our understanding of wildfire–climate interactions and support management strategies aimed at mitigating wildfire impacts on biodiversity and carbon cycling.