摘要
Soil moisture (SM), an indicator of the amount of water entering into an ecosystem, is an important component of hydrological cycling. In this study, the Soil and Water Assessment Tool (SWAT) model was used to explore spatiotemporal variations in SM and its responses to precipitation (P) and actual evapotranspiration (ET) in the Yellow River basin (YRB), a large water-limited basin in China. Results showed that model simulations performed well while drawing several important conclusions: (i) Annual P, ET, and SM exhibited a decreasing trend. On a seasonal scale, SM exhibited a negative trend with the exception of summer. On a monthly scale, SM was highest in September and lowest in June. Greater than half of the total area within the three precipitation gradients (PGs) exhibited a decline in SM, ranging between 53.16 and 74.63 %, indicating that YRB has been experiencing an increasingly severe drought period over the past 50 years. (ii) SM was positively correlated and more sensitive to P in descending order from arid regions, semi-arid regions, and semi-humid regions. SM was negatively correlated to ET in YRB but positively correlated to ET for each PG investigated. (iii) SM lagged behind P and ET by 0–3 months for YRB and all PGs, and time lags were relatively shorter in more arid areas. This study provides useful information for early warning of land water shortages. It also offers scientifically-based suggestions on ecosystem recovery and sound soil and water management practices for water-limited basins.