Temporal and spatial variations in evapotranspiration on the northern slope of the Kunlun Mountains based on multi-source datasets

By combining eight types of evapotranspiration datasets, the spatial and temporal variations in the evapotranspiration (ET) on the northern slope of the Kunlun Mountains were analyzed in uninhabited areas that lack observational data. The order of the average annual ET was ERA5_Land (312.32 mm/a) > CR (239.80 mm/a) > MOD16STM (211.87 mm/a) > GLADS (119.02 mm/a) > ETM (111.88 mm/a) > EB-ET (109.90 mm/a) > GLEAM (100.84 mm/a) > MERRA-2 (100.81 mm/a). The ET value from the ERA5_Land dataset was three times higher than that of the other five datasets. The ET values of the CR and MOD16STM datasets were twice that of the other five datasets. In terms of time, the correlation coefficient between the GLEAM and MERRA-2 datasets was the highest ( R = 0.82). In terms of space, GLDAS and MERRA-2 had the highest multi-year average ET correlation coefficient ( R = 0.80). The reduction in spatial scale resulted in clear differences in the multi-year average ET correlations among different products in the same region. In terms of time, the average annual ET of the basins on the northern slope of the Kunlun Mountains exhibited an overall increasing trend for all data sources, and the overall annual average change in the study area estimated by the eight datasets was 1.09 mm/a. The most rapid rates of increase were obtained from GLDAS (1.38 mm/a) and GLEAM (1.38 mm/a). In the CR, ERA5_Land, GLEAM, GLDAS, MERRA-2, ETM, MOD16STM, and EB-ET datasets, 46.46%, 41.47%, 87.30%, 40.30%, 49.10%, 47.13%, 57.16%, and 45.12% of the watersheds, respectively, showed a significant increasing trend. The ET value of the Yarkand River Basin showed a significant increasing trend for all eight data sources. The results of this study provide a scientific reference for the allocation of water resources on the northern slope of the Kunlun Mountains.