Spatial variation of climatology monthly crop reference evapotranspiration and sensitivity coefficients in Shiyang river basin of northwest China

Crop reference evapotranspiration (ET0) is often used to determine crop water requirement. ET0 maps are useful for regional agricultural and water resources management, and also play an important role in the distributed hydrological modeling. For generating spatial ET0 surfaces, 'Interpolate-then-calculate (IC)' approach is powerful in principle and is recommended especially for sparse weather station networks. The partial thin-plate smoothing spline incorporated in ANUSPLIN for interpolating climatic variables has been accepted widely across the world. In this paper, the climatology monthly ET0 data of Shiyang river basin, one of the three inner basins in northwest China, are developed by spatially modeling the input climatic parameters with ANUSPLIN, and from the interpolated climate and ET0 datasets, sensitivity coefficients of ET0 to the climatic variables of selected months are also spatially distributed. In the cool months (January, February, November and December), the spatial variability of ET0 is small and the value is rather low, whereas the warm season (May, June, July and August) is characterized by high values of ET0 and large spatial variations in the river basin. Vapor pressure deficit is the most sensitive variable during the cool months and in the mountainous area with lower temperature; mean air temperature is the least sensitive one during the year and a little variation is observed at the basin scale. In summer, available energy primarily forces ET0 as expected, and in winter, wind speed plays an important role and affects ET0 greater at the northern plain region where deserts are dominated by dunes and low shrubs. We conclude that for regions with isolated climate stations, 'IC' procedure by including topographic and geographic factors can effectively model spatially distributed ET0.