Publications

Xu, TR; Guo, ZX; Xia, YL; Ferreira, VG; Liu, SM; Wang, KC; Yao, YJ; Zhang, XJ; Zhao, CS (2019). Evaluation of twelve evapotranspiration products from machine learning, remote sensing and land surface models over conterminous United States. JOURNAL OF HYDROLOGY, 578, UNSP 124105.

Abstract
Evapotranspiration (ET) products were evaluated over the conterminous United States (CONUS). These products include the following: 1 product from machine learning model (GFET), 3 products based on remotely sensed observations (SSEBop, MOD16 and GLEAM) and 8 products from land surface models in NLDAS-2 (Mosaic, Noah28, SAC and VIC403) and NLDAS-3 (CLSM25, Noah36, NoahMP36 and VIC412). The AmeriFlux observations and water balance derived ET (WBET) were used to validate these products at point and basin scales. The three-corned hat (TCH) method was employed to quantify ET uncertainties over the basin scale and in each grid over the whole CONUS. The ET interannual variability and the impacts of drought on ET were analyzed over the basin scale and Texas. The results indicate that all models are able to capture ET seasonal variations compared to AmeriFlux observations. Over basin scale, all ET products are closely related to WBET with high correlation coefficient values (larger than 0.83). Noah28 (VIC412) has smallest root-mean-square difference (RMSD) of 27.32 mm/year (42.78 mm/year). The uncertainties calculated from TCH method indicate that NLDAS-3 monthly ET products have lower uncertainties (4-7 mm/month) than those from NLDAS-2 ET (7-8 mm/month) and two MODIS ET (MOD16 and SSEBop) (10-15 mm/month). Specifically, ET uncertainty is reduced 47% (19%) for Noah36 (NoahMP36) compared to Noah28; it is reduced 19% for VIC412 compared to VIC403; it is reduced 19% for CLSM25 compared to Mosaic. The GFET (4.37 mm/month) and GLEAM (6.44 mm/month) have comparable low uncertainties with NLDAS-3 ET products. This study provides an important basis for the selection of proper ET data sets for the hydrological analysis over CONUS.

DOI:
10.1016/j.jhydrol.2019.124105

ISSN:
0022-1694