Impacts of Urbanization on Compound Heat and Drought Events in the Beijing–Tianjin–Hebei Region Based on Explainable Machine Learning

Compound heat and drought events (CHDEs) are occurring more frequently due to climate change. However, the role of urbanization as a key driver of climate change in modulating these events remains poorly quantified. Hence, this study evaluates the impacts of urbanization on CHDE frequency, duration, and severity using a daily scale analytical framework, and further quantifies the key influencing factors contributing to these impacts using explainable machine learning. The results revealed that: (1) the urban expansion rates ranged from 0.06% to 22.92% per decade in the Beijing–Tianjin–Hebei (BTH) region. Concurrently, the frequency, duration, and severity of CHDEs exhibited overall upward trends, with rates of 0.11, 1.45, and 0.06 per decade, respectively. (2) Urban stations exhibited more apparent upward trends in the frequency, duration, and severity of CHDEs than rural stations. Urbanization contributed to these increases in CHDEs, with its greatest contribution to duration (44.0%), followed by severity (39.1%) and frequency (33.6%). (3) The explainable machine learning analysis revealed that the attributes of CHDEs are driven by multiple urban and climatic factors under urbanization. While enhanced Tmean and UHI consistently intensified CHDEs, changes in urban underlying surfaces manifest complex non-linear relationships with CHDE attributes. Notably, BV and UrbF amplified the duration and severity of CHDEs more strongly than frequency. These findings suggest that urbanization may primarily amplify the duration and severity of CHDEs through enhanced thermal conditions and surface modification, highlighting the need to prioritize heat mitigation and urban land-use regulation in climate adaptation planning.