Advancing Evapotranspiration Modeling With Optimized Soil and Canopy Resistance Combinations

Abstract Dual‐source remotely sensed evapotranspiration (ET) models require accurate separation of soil evaporation (Es), plant transpiration (Ec), and precipitation interception (Ei) based on soil and canopy resistances. Despite the availability of several ET products and algorithms, comprehensive...

Full description

Saved in:
Bibliographic Details
Main Authors: Jinfeng Zhao, Shikun Sun, Yali Yin, Yihe Tang, Chong Li, Yongshan Liang, Yubao Wang, Alexander Winkler, Shijie Jiang
Format: Article
Language:English
Published: Wiley 2025-06-01
Series:Water Resources Research
Subjects:
Online Access:https://doi.org/10.1029/2024WR039252
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Dual‐source remotely sensed evapotranspiration (ET) models require accurate separation of soil evaporation (Es), plant transpiration (Ec), and precipitation interception (Ei) based on soil and canopy resistances. Despite the availability of several ET products and algorithms, comprehensive evaluations of resistance configurations remain scarce. This study systematically evaluates various combinations of five soil resistance methods, eight canopy resistance methods, and two precipitation interception algorithms within the Shuttleworth‐Wallace (S‐W) framework. Using eddy covariance data from 119 FLUXNET sites and the latest ET products, we find that the Ball‐Berry‐Leuning method, unified stomatal method, and RL empirical method provide comparable and top‐ranked performance across plant functional types (PFTs) and climate zones, with only a single free parameter calibrated by genetic algorithm. The power function method (S2), sensitive to soil surface water content proves to be the most effective for modeling Es, particularly in water‐limited regions. The performance of best‐performing but unexplored combinations (S2‐C1, S2‐C2, S2‐C5) is consistent with PML‐V2, GLEAM4, and underlying water use efficiency model, explaining 56% of the variation in daily ET and achieving an root mean square error as low as 1.02 mm day−1. However, these models show reduced accuracy in arid zones, where prolonged water stress led to a 38% reduction in R2. This highlights the need for a more accurate representation of soil moisture stress in arid regions, which is often overlooked in existing models. Our study offers robust, parsimonious, and broadly applicable models for ET estimation across PFTs and climate zones.
ISSN:0043-1397
1944-7973