Monitoring of "All-weather" Evapotranspiration Using Multi-source Remote Sensing Imagery in Cloudy and Mountainous Regions in Southwest China

Accurate estimation of surface evapotranspiration (ET) with high quality and fine spatiotemporal resolution is one of the biggest obstacles for routine applications of remote sensing in eco-hydrological studies and water resource management at basin scale. Integrating multi-source remote sensing data is one of the main ideas for many scholars to obtain synthesized frequent high spatial resolution surface ET. This study was based on the theoretically robust surface energy balance system (SEBS) model, which the model mechanism needs further investigation, including the applicability and the influencing factors, such as local environment, heterogeneity of the landscape, and optimized parametric scheme, for improving estimation accuracy. In addition, due to technical and budget limitations, so far, no single sensor provides both high spatial resolution and high temporal resolution. Optical remote sensing data is missing due to frequent cloud contamination and other poor atmospheric conditions. The passive microwave (PW) remote sensing has a better ability in overcoming the influences of clouds and rainy. The accurate "all-weather" ET estimation method had been proposed through blending multi-source remote sensing data acquired by optical, thermal infrared (TIR) and PW remote sensors on board polar satellite platforms. The estimation had been carried out for daily ET of the River Source Region in Southwest China, and then the "All-weather" remotely sensed ET results showed that the daily ET estimates had a mean absolute percent error (MAPE) of 36% and a root mean square error (RMSE) of 0.88 mm/day relative to ground measurements from 12 eddy covariance (EC) sites in the study area. The validation results indicated good accuracy using multi-source remote sensing data in cloudy and mountainous regions.