摘要
Precipitation anomalies have greatly affected eco-hydrological processes in the upper reach of the Mekong River. In this research, an integrated spatiotemporal decomposition analysis method was proposed for the identification of precipitation trends under climate change at the basin scale. This method was a comprehensive use of multiple environmental statistical analysis approaches, including the S-mode empirical orthogonal function analysis, the inverse distance weighted interpolation, the Mann-Kendall trend test, the linear regression and the weighted moving average. Results indicated that precipitation trends were mainly obtained by two spatial-temporal variation patterns resulting from the decomposition of the original precipitation field. The primary spatial pattern (more than 40% variances) indicated that the annual and seasonal precipitation in the entire basin had a common variation trend. Moreover, variability of precipitation increased to the largest in the central basin and gradually decreased from north to south. Accordingly, precipitation-sensitive areas were mainly located in the east-central basin. The associated temporal trends showed that annual precipitation increased slightly over 1960–2009, and decreased significantly at the 5% significance level since 2000. Spring precipitation increased significantly over the entire study period. The secondary spatial pattern extracted at least 15% of the variance and represented a north–south inverse-variation trend. The northern mountainous region was more sensitive to climate change. Temporal trend analysis showed that annual precipitation had an increasing trend in the northern region and a decreasing trend in the southern region. Precipitation in the northern region increased significantly at the 1% significance level in winter. The research results could form a basis for supporting basin-scale water resources management, especially in the mountainous basin.