Mechanisms and model performance evaluation of evapotranspiration partitioning in summer maize in Northwest China

Le résumé fourni par la source

Under the context of climate change in Northwest China, accurately partitioning the components of actual crop evapotranspiration (ET c act ) is critical for enhancing water use efficiency and formulating precise irrigation strategies. However, the regulatory mechanisms of evaporation (E act ) and transpiration (T act ) in summer maize remain unclear in this region, and a locally applicable evapotranspiration partitioning model is yet to be developed. To address this, the present study was conducted at the Yangling station using three consecutive years of field measurements during the summer maize growing season. ET c act , T act , and E act were respectively obtained using an eddy covariance system, wrapped stem sap flow meters, and micro-lysimeters, enabling full-process observation of evapotranspiration components. Based on these observations, the control mechanisms of evapotranspiration composition were systematically analyzed in conjunction with leaf area index (LAI) and meteorological factors. Three representative models—Priestley–Taylor Jet Propulsion Laboratory (PT-JPL), Shuttleworth–Wallace (S-W), and FAO-56 Dual Crop Coefficient (Dual-FAO)—were optimized and evaluated. The results showed that the energy balance measured by the eddy covariance system was reliable during the observation period, with regression slopes of energy closure ranging from 0.67 to 0.89 over the three growing seasons and R 2 values exceeding 0.80, confirming the accuracy of the flux measurements. In 2022, the year with the highest soil moisture, the average daily T act /ET c act ratio for summer maize was 0.71, which was 0.02 and 0.03 higher than in 2023 and 2024, respectively, indicating a greater proportion of water consumption through transpiration and thus more efficient water use in that year. Net radiation consistently emerged as the dominant driver of variations in ET c act and its components. LAI played a central mediating role in evapotranspiration partitioning: when LAI < 3 m 2 ·m −2 , increases in LAI significantly promoted T act while suppressing E act ; when LAI > 3 m 2 ·m −2 , this effect was further enhanced, reflecting the stage-specific dominance of canopy structure in regulating water flux pathways. After parameter optimization, all three models were capable of reproducing the dynamic changes in ET c act and its components with reasonable accuracy. Among them, the S-W model exhibited the highest precision and best error control in simulating summer maize water fluxes, with R 2 values of 0.81, 0.86, and 0.92 and RMSEs of 0.61 mm day −1 , 1.62 mm day −1 , and 1 .42 mm day −1 for E act , T act , and ET c act , respectively. In contrast, the PT-JPL model showed certain limitations in capturing water dynamics in complex agricultural systems due to its simplified structure. This study provides practical insights and methodological support for improving water management strategies, enhancing drought resilience in cropping systems, and guiding model selection in arid agricultural regions.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.