Evaluating the accuracy of nine canopy resistance models in estimating winter wheat evapotranspiration using the Penman-Monteith equation.

Abstract

Accurate estimation of farmland evapotranspiration (ET) is crucial for agricultural production. The accuracy of the widely used Penman-Monteith (PM) equation for estimating crop ET depends on the quality of input data and their ability to accurately model the canopy resistance (r _c). In this study, we evaluated the PM equation in estimating winter wheat ET using nine r _c models, with both original and recalibrated parameters, including the Farias (FA), Monteith (MT), Garcίa-Santos (GA), Idso (IS), Jarvis (JA), Katerji-Perrier (KP), Stannard (ST), Todorovic (TD), and Coupled surface resistance (CO) models. We used long-term measurements (2018 to 2023) from the Bowen ratio energy balance method at both daily and seasonal scales. Parameterization was performed using data from the 2020-2021 growing season, while the remaining 4 years were used for verification. The results showed that the FA, KP, and ST models performed better in estimating daily ET with original parameters, achieving a root mean square error (RMSE) of 1.07-1.16 mm d^-1 and a mean bias error (MBE) of -0.59-0.02 mm d^-1. After parameterization, the performance of acceptable r _c models based on RMSE (ranging from 1.07 to 1.22 mm d^-1, averaged 1.16 mm d^-1) ranked as follows on the daily scale: FA > CO > KP > ST > IS > GA > JA > MT. The r _c models were more accurate in simulating ET on a seasonal scale than on the daily scale. Before calibration, the acceptable FA, KP, and MT models overestimated seasonal ET with the MBE ranging from 2.83 to 75.32 mm and RMSE from 29.79 to 82.38 mm. After correction, the suitable r _c models based on RMSE values decreased by FA > CO > KP > IS > ST > GA > JA on the seasonal scale, which ranged from 29.79 to 76.35 mm. The performance of the revised r _c models improved on both daily and seasonal scales, with RMSE reductions of 29.03% and 68.18%, respectively. Considering both the accuracy and calculation complexity, the FA and KP models were recommended to be used in the PM equation to estimate daily and seasonal ET in semiarid regions. The CO, GA, ST, IS, and JA models can also be used as alternatives, depending on the availability of meteorological parameters.