Comprehensive Evaluation of Terrestrial Evapotranspiration from Different Models under Extreme Condition over Conterminous United States
Xingjiao Yu,Long Qian,Wen'e Wang,Xiaotao Hu,Jianhua Dong,Yingying Pi,Kai Fan
DOI: https://doi.org/10.1016/j.agwat.2023.108555
IF: 6.7
2023-01-01
Agricultural Water Management
Abstract:Terrestrial evapotranspiration (ET) is a key process in the water, energy, and carbon cycles. ET products are becoming more widely accessible with the development of remotely sensed inversions and availability of land surface assimilation information. However, the response of these products to extreme weather is not yet clear and challenging to evaluate. This study evaluated and compared the consistency and uncertainty of nine evaporation products (including GLEAM3.6b, reliability ensemble averaging (REA), CLSM, ERA5-Land (ERA5), MOD16A2, PML-V2 (PML), NOAH, FLDAS, and Synthesized (Syn)) for the period 2003–2015. Eddy covariance (EC) flux observations and the three-cornered hat (TCH) method were used to validate and assess the uncertainty of these products at both the point-scale and grid-scale under normal climatic conditions. More importantly, we conducted uncertainty assessments of ET products at daily and monthly scales under extreme climatic conditions (including high temperature (Temp), high vapor pressure deficit (VPD), and drought), referencing 50 EC measurements over the conterminous United States (COUNS). The results showed that the accuracy of different ET products varied with cover type. Compared to other products, GLEAM3.6b, REA, and Syn showed superior results for most land cover types. Specifically, GLEAM3.6b showed higher performance for the grasslands; REA showed improved performance in mixed forests, open shrubs, and wetlands, and Syn fared better in forests and croplands. The point-scale indicated that the lowest root mean square deviations (RMSD) come from Syn grid products (mean 16.13 mm/month), followed by GLEAM3.6b (mean 18.53 mm/month) and REA (mean 19.42 mm/month). The highest RMSD (mean 49.49 mm/month) originated from ERA5, while the lowest R was found in MOD16A2. The uncertainty ranking calculated using the TCH method is similar to the findings of the point-scale evaluation based on EC measurements, with Syn recording the lowest uncertainty of 5.13 mm/month, after GLEAM3.6b (5.72 mm/month), followed by REA (6.24 mm/month), and ERA5 (12.26 mm/month) with the largest uncertainty. When extreme weather conditions were encountered, the uncertainty of the nine ET products showed an increasing trend, and their consistency varied. Under high-temperature and high-VPD conditions, the consistency decreased, while during drought, the consistency showed a slight increase. Among the ET products, GLEAM3.6b performed best under high-temperature and high-VPD conditions, closely followed by REA. However, ERA5 exhibited the largest error in its estimations. On the other hand, when drought conditions occurred, GLEAM3.6b, REA, and Syn showed better performance, while CLSM, NOAH, and FLDAS ranked second, ERA5 had the largest uncertainty, and MOD16A2 and PML exhibited the worst consistency. The findings of this study offer a foundation for selecting appropriate ET products for hydrologic analysis and agricultural water resource management in the CONUS and give developers suggestions for lowering the uncertainty of ET products in the face of extreme climatic conditions.