Variability and Attribution of the Surface Water Budget Under the Background of Warming and Wetting in Northwest China

The climate warming and humidification trend across Northwest China has been extensively validated and widely recognized by the global meteorological research community. However, the spatiotemporal evolution characteristics, variation patterns and driving mechanisms of the terrestrial surface water cycle under the backdrop of this distinctive climatic regime shift still lack systematic and in‐depth elucidation, posing uncertainties for regional water resource assessment and ecological security governance. In the study, leveraging the high‐quality MERRA‐2 atmospheric reanalysis dataset and the classical terrestrial surface water balance equation as the theoretical framework, we conduct a comprehensive, systematic investigation into the spatiotemporal heterogeneity characteristics, long‐term variation trends and multi‐scale influencing factors of the surface water budget across Northwest China. The main findings are summarized as follows: The surface water budget in Northwest China manifests pronounced spatiotemporal variability across multiple scales. Spatially, the annual mean surface water budget presents a distinct heterogeneous distribution pattern: persistent water surplus dominates northern Xinjiang, southeastern Qinghai and the majority of Shaanxi Province, whereas widespread water deficit prevails across the remaining regions of Northwest China. Intra‐annually, a sharp east–west contrast pattern emerges with the 90° E as the natural boundary: the eastern part of Northwest China is characterized by water surplus in the summer half‐year and water deficit in the winter half‐year, while the western part exhibits the exact opposite seasonal variation rhythm. In terms of long‐term linear trends, the annual surface water budget across the entire study region depicts a significant increasing trend towards water surplus. Seasonally, most areas experience a deteriorating water deficit trend during the winter half‐year, with the sole exceptions of northern and eastern Xinjiang; conversely, the summer half‐year witnesses a widespread water surplus enhancement trend across the region. From a decadal variation perspective, the annual surface water budgets for the whole region were basically balanced, while showing a surplus trend before 2000 and then a deficit trend in the winter half‐year; it is exactly the opposite in the summer half‐year. Regarding direct driving factors, precipitation acts as the sole water input term in the surface water balance equation, while actual evapotranspiration and surface runoff constitute the core water output terms. Despite this binary input–output structure, all three hydrological components exert positive contributions to the regional surface water budget, which is predominantly governed by the dominant increment of precipitation over evapotranspiration and runoff losses. Climatic factors including surface net radiation, air temperature, near‐surface wind speed and relative humidity indirectly modulate the surface water budget by altering the intensity of land‐atmosphere water exchange and actual evapotranspiration efficiency; their contribution intensities vary drastically across subregions due to the complex heterogeneity of climatic backgrounds and underlying surface properties. Quantitative attribution analysis reveals that precipitation, actual evapotranspiration and surface runoff contribute 50.8%, 44.7% and 4.6% to the overall variation of the surface water budget, respectively, with remarkable discrepancies across different temporal scales. For indirect climatic drivers, relative humidity and meridional wind velocity impose relatively dominant impacts, with contribution rates reaching 24.6% and 17.0%, respectively; in comparison, surface net radiation, air temperature and zonal wind velocity yield marginal contributions to the budget variation. Furthermore, the contribution intensities of these indirect driving factors also exhibit significant temporal scale‐dependent heterogeneity.