Optimization of canopy resistance models for estimating evapotranspiration on summer maize in a semi-arid condition of China.

Accurate estimation of evapotranspiration (ET) is crucial for improving crop water-use efficiency and facilitating precision irrigation. The Penman–Monteith (P-M) model is a widely used approach for ET calculation, with canopy resistance (rc) acting as a critical determinant of accuracy that is greatly affected by meteorological, crop, and soil factors. This study focused on the calibration of two rc models: Jarvis (JA) and coupled resistance (CO), utilizing ant colony optimization (ACO) and the least square method (LSM). Data from an eddy covariance system and a meteorological station in the Huailai summer maize field (115°47'E, 40°20'N) for the year 2013 were employed for model parameterization. Performance evaluation of these models was conducted using ET measurements obtained through the eddy covariance system in 2014. Path analysis was conducted to elucidate the impact of net radiation (Rn), temperature (T), vapor pressure deficit (VPD), soil water content (θ) and leaf area index (LAI) on rc. Results revealed a decrease in rc during growing season, with influencing factors ordered as Rn, LAI, VPD, T, and θ. The CO model optimized by ACO exhibited superior accuracy, with R2 of 0.89. Integration of the ACO-optimized the CO model into P-M model for ET estimation also demonstrated the highest accuracy with R2 of 0.72. JA model optimized by ACO provided satisfactory results in both parameters calibration and ET estimation. Therefore, the ACO-optimized CO model was recommended as the optimal approach for rc calculation of summer maize in semi-arid area. Additionally, the JA model optimized by ACO was proposed as a credible alternative to the CO model. These findings contribute to the refinement of ET estimation methodologies, offering valuable insights for precision agricultural water management practices. [ABSTRACT FROM AUTHOR]