Can Indirect Evaluation Methods and Their Fusion Products Reduce Uncertainty in Actual Evapotranspiration Estimates?

Accurately estimating actual evapotranspiration (ET) across global land surface is one of the key challenges in terrestrial hydrological cycles and energy flux balance studies. Gridded ET products have the potential for application in ungauged basins, but their uncertainties are possibly large and i...

Full description

Saved in:
Bibliographic Details
Published in:Water resources research 2022-06, Vol.58 (6), p.n/a
Main Authors: Shao, Xingmin, Zhang, Yongqiang, Liu, Changming, Chiew, Francis H. S., Tian, Jing, Ma, Ning, Zhang, Xuanze
Format: Article
Language:English
Subjects:
Balance studies
basin
Basins
Bayesian analysis
climate change
Energy flux
Energy transfer
ensemble
Estimates
Evapotranspiration
Evapotranspiration estimates
global
Hydrologic cycle
Hydrologic data
Hydrological cycle
Hydrology
Methods
Probability theory
Products
River basins
Stream discharge
Stream flow
Streamflow data
Terrestrial environments
Uncertainty
Water balance
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Accurately estimating actual evapotranspiration (ET) across global land surface is one of the key challenges in terrestrial hydrological cycles and energy flux balance studies. Gridded ET products have the potential for application in ungauged basins, but their uncertainties are possibly large and it remains unclear which one is best for a given basin. The water balance (WB) method provides a direct estimate of basin scale ET, but it cannot be used in ungauged basins where streamflow data are unavailable. Here, we first assess the performance of ET from 10 global ET products against WB ET estimates in 43 large river basins. The paper then uses three indirect evaluation methods [Three Cornered Hat (TCH), Arithmetic Average (AA), and Bayesian Three Cornered Hat] to identify the optimal ET products without the need of prior information, and to generate fusion products combining the ET from multiple products. Using the evaluation results derived from the WB method as the reference, the results show that the three methods have great success in identifying poorer products, suggesting that they are useful in filtering poor ET products in applications. However, the ability of such methods in identifying better ET products degrades slightly. The AA fusion product, which combines ET outputs from multiple products, is marginally better than the best single ET product in many of the 43 basins. Because of its simplicity, it could be used to reduce the uncertainty in ET estimates from multiple products for ungauged basins and regions. Key Points A comprehensive study on reducing the uncertainty in evapotranspiration (ET) estimates from 10 global ET products for application in ungauged river basins The indirect evaluation methods without using prior information, particularly Arithmetic Average (AA) and Bayesian Three Cornered Hat can successfully identify the poor ET products The AA fusion product, which combines ET outputs from multiple products, is generally marginally better than the best single ET product
ISSN:0043-1397
1944-7973
DOI:10.1029/2021WR031069