The quality of meteorological dataset has a considerable impact on the performance of land surface models.Two meteorological datasets,GSWP3 and WFDE5,were used to drive the CLM4.5 to simulate the surface evapotranspiration process across the Yellow River Basin(YRB).The performance was systematically assessed applying various metrics based on multiple space-time scales.The RMSE of multi-year averaged simulated ET for GSWP3 and WFDE5 were 61.5 and 49.5 mm/a,and KGE values were 0.83 and 0.87 against reference ET,respectively.Temporally,both simulated ET and reference ET showed significant increasing trends.Spatially,the best performance of GSWP3 and WFDE5 occurred in the middle reaches of the YRB with the smallest relative errors,while the poorest performance occurred in the source area of the YRB.Seasonally,the GSWP3 is superior to WFDE5 in spring and winter,whereas WFDE5 is superior to GSWP3 in summer and autumn.Overall,the performance of CLM4.5 driven by WFDE5 was better than that of GSWP3.The study can provide motivation and direction for the improvement of the ET module in land surface models.
evapotranspiration simulationland surface modelmeteorological forcingYellow River basinvarious datasetsGSWP3WFDE5
NETL
NSTL
维普
万方数据
