SHAP-powered insights into short-term drought dynamics disturbed by diurnal temperature range across China

Short-term drought dynamics are critical for agricultural production and water resource management, yet the complex role of diurnal temperature range (DTR), as a key climate variable influencing surface energy and water cycles, remains poorly understood in drought processes. This study addresses this gap by integrating a high-resolution meteorological dataset (1961–2022) into a SHapley Additive exPlanations (SHAP)-based attribution framework, quantifying DTR’s impacts on short-term drought (the monthly-scale standardized precipitation-evapotranspiration index, SPEI-1) across China’s diverse climatic zones. A significant negative correlation between monthly DTR and SPEI-1 indicates that DTR directly intensifies short-term drought conditions, an effect that has strengthened significantly since 2000. DTR influences drought through dual pathways: direct exacerbation via enhanced evapotranspiration demand and indirect modulation through negative associations with precipitation and relative humidity (RH) and positive links with sunshine duration and wind speed. In arid regions, DTR interacts synergistically with precipitation and RH to exacerbate drought, whereas in humid regions, DTR’s positive association with sunshine duration partially mitigates drought severity. Importantly, DTR is identified as the primary driver of short-term drought, followed by RH and sunshine duration. A 1°C increase in DTR reduces SPEI-1 by −0.04 to −0.26 (worsening drought), while a 1°C decrease in DTR increases SPEI-1 by 0.03–0.28 (alleviating drought), which is particularly pronounced in semi-arid, arid, and hyper-arid regions. This study advances our understanding of DTR’s multifaceted role in short-term drought dynamics and highlights the urgent need for targeted adaptation strategies, such as adaptive irrigation scheduling and water resource allocation, to mitigate drought intensification, particularly in ecologically vulnerable regions.