Actual evapotranspiration (ET) with high spatiotemporal resolution is very important for the research on agricultural water resource management and the water cycle processes, and it is helpful to realize precision agriculture and smart agriculture, and provides critical references for agricultural layout planning. Due
[...] Read more.
Actual evapotranspiration (ET) with high spatiotemporal resolution is very important for the research on agricultural water resource management and the water cycle processes, and it is helpful to realize precision agriculture and smart agriculture, and provides critical references for agricultural layout planning. Due to the impact of the clouds, weather environment, and the orbital period of optical satellite, there are difficulties in providing daily remote sensing data that are not contaminated by clouds for estimating daily ET with high spatial-temporal resolution. By improving the enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM), this manuscript proposes the method to fuse high temporal and low spatial resolution Weather Research and Forecasting (WRF) model surface skin temperature (TSK) with the low temporal and high spatial resolution remote sensing surface temperature for obtaining high spatiotemporal resolution daily surface temperature to be used in the estimation of the high spatial resolution daily ET (ET_WRFHR
). The distinction of this study from the previous literatures can be summarized as the novel application of the fusion of WRF-simulated TSK and remote sensing surface temperature, giving full play to the availability of model surface skin temperature data at any time and region, making up for the shortcomings of the remote sensing data, and combining the high spatial resolution of remote sensing data to obtain ET with high spatial (Landsat-like scale) and temporal (daily) resolution. The ET_WRFHR
were cross-validated and quantitatively verified with MODIS ET products (MOD16) and observations (ET_Obs) from eddy covariance system. Results showed that ET_WRFHR
not only better reflects the difference and dynamic evolution process of ET for different land types but also better identifies the details of various fine geographical objects. It also represented a high correlation with the ET_Obs by the R2
amount reaching 0.9186. Besides, the RMSE and BIAS between ET_WRFHR
and the ET_Obs are obtained as 0.77 mm/d and −0.08 mm/d respectively. High R2
, as well as the small RMSE and BIAS amounts, indicate that ET_WRFHR
has achieved a very good performance.