Modeling Orchard Evapotranspiration and Its Components by Combining a Simplified Canopy Resistance Algorithm and Penman–Monteith-Based Models

Accurate estimation of evapotranspiration (ET) and transpiration (T) is crucial in enhancing irrigation schedules in agricultural ecosystems. A simplified canopy resistance (rsc) algorithm grounded in the Ball-Berry model was integrated into Penman–Monteith (PM)-based models and assessed in a humid-region kiwifruit orchard. The Shuttleworth–Wallace (SW) and Clumping (CL) models agreed well with eddy covariance ET (ETEC) throughout the growth season (R2 = 0.79 and 0.89, RRMSE = 0.73–0.75 and 0.22, at the sub-daily and daily scales), outperforming the Two-Patch (TP) and topography- and vegetation-based surface energy partitioning (TVET) models. SW and CL also best reproduced sap-flow-based T (TSF) (R2 = 0.72 and 0.71–0.72, RRMSE = 0.85–0.87 and 0.38–0.39), primarily due to their higher accuracy during the mid stage. In this ecosystem, canopy interception evaporation had only a limited influence on the simulation performance of the SW and CL models. ET and T were most sensitive to changes in rsc and related environmental factors, including gross primary productivity (GPP), the empirical parameter (a1), and soil water content (θ). The sensitivity of T to θ was higher during the early stage but lower during the mid and late stages due to seasonal drought. ET was more sensitive to θ than T, due to its direct effect on soil surface resistance (rss). T simulated by TP, TVET, and CL models showed greater sensitivity to leaf area index (LAI) than SW, while contrasting T and soil evaporation (E) responses to LAI caused ET to remain relatively insensitive. Both ET and T were highly sensitive to net radiation (Rn), air temperature (Ta), and vapor pressure deficit (VPD), all of which directly affect the energy balance. Overall, integrating the simplified canopy resistance algorithm with PM-based models improves ET and T estimation in humid-region orchards, supporting more efficient water management.