Simulation of dryland maize growth and evapotranspiration using DSSAT‐CERES‐Maize model

Abstract

AbstractEvapotranspiration (ET) is a crucial component of plant water use and ecosystem energy balance. This study aims to simulate maize (Zea mays L.) ET using a crop modeling system and compare that to ET measured using an eddy‐covariance (EC) system. An EC system was used to collect growing season ET data from a rainfed maize field in 2017, 2018, and 2019. Supporting soil and plant growth data were also collected. The Decision Support System for Agrotechnology Transfer (DSSAT)‐Crop Environment Resource Synthesis (CERES)‐Maize model was used in this study. Results from DSSAT's ET simulation methods, FAO‐56 and Priestley–Taylor, were compared to measured ET. Results indicated a low coefficient of determination (R2) between simulated and measured ET, due to overestimation of simulated ET in all 3 yr. The average percent root mean square error for the daily ET simulation across all 3 yr was 34.6% for the Priestley–Taylor method and 30.4% for the FAO‐56 method. For the FAO‐56 simulation, the average seasonal overestimation compared to measured ET was 59 mm and the average R2 was .59. For the Priestley–Taylor method, the average seasonal overestimation was 72 mm and the average R2 was .57. Possible sources of error contributing to this overestimation include DSSAT's potential ET estimations, crop coefficient estimation, and the simulation of soil water balance in vertisols with high swell‐shrink clay minerals. The mixed simulation results, particularly compared to those seen in studies of irrigated crops, indicate a greater need to improve ET estimation capabilities in rainfed systems.