Main content area

Simulation of cotton growth and soil water content under film-mulched drip irrigation using modified CSM-CROPGRO-cotton model

Li, Meng, Du, Yingji, Zhang, Fucang, Bai, Yungang, Fan, Junliang, Zhang, Jianghui, Chen, Shaoming
Agricultural water management 2019 v.218 pp. 124-138
arid zones, cash crops, cotton, crop yield, data collection, decision support systems, field experimentation, irrigation management, leaf area index, microirrigation, models, phenology, plastic film mulches, soil types, soil water, soil water content, China
Cotton is one of the major cash crops in Xinjiang, an arid region in northwestern China. Most cotton fields in this region are drip irrigated with plastic film mulch. However, few studies have tested the performance of the DSSAT (Decision Support System for Agrotechnology Transfer) model on cotton under film-mulched drip irrigation. The objective of this study is to evaluate the performance of the modified CSM-CROPGRO-Cotton model and determine the genetic coefficients of cotton using data from different irrigation treatments conducted at the irrigation experiment station in Bayingol Mongolian Autonomous Prefecture, Korla, China. Based on the selected genetic coefficients, the model was calibrated using field experiment datasets in 2016 and validated using datasets in 2015. The simulation results of soil water content, leaf area index, phenological period and cotton yield were compared with their corresponding measurements. Absolute relative error (ARE) of seeding emergence date, flowing date, maturity date and cotton yield simulated by the original model were 5.8%, 7.3%, 5.0% and 70.6%, respectively, while those of the modified model were 1.7%, 1.1%, 1.1% and 0.1%, respectively. Simulated leaf area index agreed well with the field observations with the coefficient of determination (R2) > 0.739, the index of agreement (d) > 0.910 and the root mean square error (RMSE) < 0.965. Simulated soil water content in the 20–40 cm soil layer were generally consistent with the measured values with R2 = 0.724, d = 0.639 and RMSE = 0.087 cm3/cm3, while the simulated soil water content in the 40–60 cm was slightly worse than the observations with R2 = 0.708, d = 0.519 and RMSE = 0.109 cm3/cm3. Overall, the modified CSM-CROPGRO-Cotton model can be used as a practical tool to simulate the cotton growth under film-mulched drip irrigation in this region. Potential opportunities for the model improvement in cotton growth include the influence of various scenarios such as soil types, climates and irrigation strategies.