Is Climate Dominating the Spatiotemporal Patterns of Water Yield?

Climate and underlying-surface changes give rise to governance challenges in water resource management in arid and semi-arid regions, and knowledge of the spatiotemporal patterns of water yield is critical in policy and stakeholder engagement. The Soil and Water Assessment Tool (SWAT) model was integrated using spatial autocorrelation analysis, empirical orthogonal functions (EOF), rotated EOF (REOF), and wavelet analysis to explore variations in water yield and factors influencing the Jinghe River watershed from 1980 to 2010. The well-calibrated SWAT-based methodological framework, which considers the influence of climatic and human activities using dynamic modeling and statistical decomposition, is a reliable tool for evaluating and visualizing the spatiotemporal patterns of water yield at the watershed scale. The water yield of the Jinghe River watershed showed an overall downward trend over time and from southeast to northwest. The variation in water yield from the 1980s to the 1990s was mostly attributed to climate, though human activities became the dominant factor from the 1990s to the twenty-first century. Three spatial modes of water yield were detected by EOF: “consistent in the whole watershed,” “North–South,” and “central in middle and North.” The variance contributions were 77.68%, 12.92%, and 6.67%, respectively. Four regional modes of water yield were obtained through the REOF: north–south, central, eastern, and western. The variance contributions were 36.2%, 32.35%, 23.42%, and 7.58%, respectively. These findings will help deepen the understanding of the patterns of water yield and its response to climate and underlying surface in arid and semi-arid regions, which has practical significance for hydrologists, environmentalists, water resources scientists, and policymakers in assessing available water resources for a proposed or existing development.