The spatiotemporal dynamics of terrestrial water storage and the main driving factors in the drylands of northern China based on PLS-SEM

Understanding the mechanisms by which both anthropogenic activities and climate change influence terrestrial water storage (TWS) anomalies is crucial for effective regional water resource management. However, as a critical element of the water cycle, the multi-effects of the natural and anthropogenic factors on TWS were still not fully understood and quantified. This study systematically explores the spatiotemporal variations of TWS, vegetation cover, climate variables, and anthropogenic activities in northern China from 1992 to 2020. The spatial variations of TWS were analyzed using the Theil-Sen slope method and Mann–Kendall test for the periods 1992–1999 and 2000–2020, while the overall temporal trend was examined through linear least squares regression. Partial least squares structural equation modeling (PLS-SEM) was further used to quantify the causal effects of anthropogenic activities and climate change on TWS and to identify the pathways of influence. We found a continuous decline in TWS, with a significant depletion rate of 0.34 mm/yr, while normalized difference vegetation index (NDVI) exhibited a greening trend (0.0006/yr) across northern China during the same period. From a hydrological balance perspective, TWS depletion was primarily driven by precipitation and actual evapotranspiration during the study period. Notably, we observed a reversal in the correlation between TWS and runoff before and after the implementation of the ecological restoration. The PLS-SEM model further revealed that climate change and anthropogenic activities affect TWS through both direct and indirect pathways, influencing variables such as NDVI, runoff, soil moisture, and groundwater. Among these factors, anthropogenic activities (absolute total effect value of 0.34 and 0.37 before and after the implementation of ecological restoration)—including sectoral water withdrawal and land use—were the dominant drivers of TWS depletion, surpassing the impact of climate variables (0.23 and 0.29 during the same period). Our findings demonstrate the complex role of hydrological processes in mediating the effects of climate, vegetation cover, and human activities. This study offers valuable insights for improving regional water resource management and assessment strategies.