Publications

Tang, RL; Li, ZL; Huo, X; Jiang, YZ; Tang, BH; Wu, H (2019). A re-examination of two methods for estimating daily evapotranspiration from remotely sensed instantaneous observations. INTERNATIONAL JOURNAL OF REMOTE SENSING, 40(6-May), 1981-1995.

Abstract
The significance of estimating land surface evapotranspiration (ET) has been widely recognized in the fields of hydrology, meteorology, agriculture, and global change. This article compares two ET conversion methods, the constant decoupling factor () method and the constant reference evaporative fraction (EFr) method, that estimate daily ETs from instantaneous values. The daily ET was estimated either by (1) summing multiple half-hourly or hourly ET estimates that were derived through an application of the two ET conversion methods to half-hourly or hourly meteorological variables (i.e. aggregating the ET outputs) or by (2) directly applying the two ET conversion methods to daily meteorological variables (i.e. aggregating the meteorological inputs). The comparison was made using ground-based eddy covariance (EC) system measurements and the moderate resolution imaging spectroradiometer (MODIS)-based latent heat flux (LE) datasets collected from April 2009 to late October 2011 at the Yucheng station over the North China Plain. The results show that both the constant method and the constant EFr method produced daily latent evaporation (LE) estimates that were in agreement with the ground-based EC measurements. When the two methods were applied to the MODIS-based LE datasets that have a small bias of -8Wm(-2) and an root mean square error (RMSE) <60Wm(-2), the validation results of the estimated daily LE against the ground-based EC measurements showed a relative bias of <7% and a relative RMSE of <20%. For both ET conversion methods, aggregating the ET outputs produced better agreement with the ground-based EC measurements than directly obtaining the daily ET by aggregating the meteorological inputs did. No significant difference was observed in the model performance between the constant method and the constant EFr method.

DOI:
10.1080/01431161.2018.1465615

ISSN:
0143-1161