Spatiotemporal analysis of temperature-variation patterns under climate change in the upper reach of Mekong River basin

Occurrence of temperature anomaly has greatly affected natural cycles of water resources in Lancang River basin in China, which is the upper reach of Mekong River. An integrated spatiotemporal decomposition and analysis method was proposed for the identification of temperature-variation patterns under changing climatic conditions in the basin. This method was based on the combination of S-mode empirical orthogonal function analysis, IDW interpolation, liner regression, weighted moving average and Mann Kendall methods. Results indicated that the first two modes extracted nearly 80% of spatiotemporal variations in temperature. Temperature in the whole basin followed the same variation trend through the first mode analysis. Sensitive areas were mainly located in the southwest of the basin, which occupied nearly half of the basin. The associated time series presented that the basin appeared transition from cold periods to warm periods. Temperature increased significantly over the period of 1960 to 2009 at annual and seasonal scales, particularly over 1990s. At the same time, the most significant rising occurred in winter and the least in summer. In the second mode, a west–east inverse phase pattern of temperature variations was a distinct feature in most of the basin. Temporal trend indicated that the increasing trend in the west region was slightly stronger than that in the east. This was particularly the case of edge areas almost vertical juncture with monsoons. This research is not only helpful in improving understanding of temperature response to global warming in the basin but also provides a basis for basin management.

► Temperature-variation patterns were analyzed in the upper reach of Mekong River basin.
► The phase consistency of spatial temperature variations was a distinct feature through the first leading mode analysis.
► Temperature increased significantly during 1960 to 2009, particularly over 1990s.
► A west–east inverse pattern of temperature variations was found through the second leading mode analysis.
► The increasing trend of temperature variations in the west region was slightly stronger than that in the east.