Multiobjective optimization of regional irrigation and nitrogen schedules by using the CERES-Maize model with crop parameters determined from the remotely sensed leaf area index

Abstract

Using crop models to develop reasonable irrigation and nitrogen schedules (INSs) is critical for achieving high yields and maximizing water use efficiency (WUE) and nitrogen use efficiency. However, the INSs developed based on crop parameters determined through field experiments have low robustness at the regional scale. Remote sensing data indicate the growth conditions of regional crops, and representative regional crop parameters can be determined using remote sensing data. The objective of the present study was to develop a regional INS by adopting a crop model with crop parameters calibrated using the remotely sensed leaf area index (LAI). First, the crop parameters of a field-scale crop model were calibrated on the basis of field-scale trial data for four summer maize growing seasons. Sensitive crop parameters at the regional scale were then identified using an assimilation algorithm based on the remotely sensed LAI. The rationality of using the remotely sensed LAI for determining regional crop parameters was demonstrated by simulating regional maize growth. Finally, a regional-scale crop model was developed and coupled with a multiobjective genetic algorithm to optimize the regional INS for different typical-rainfall years. The results indicated that the regional-scale crop model was more accurate than was the field-scale crop model in simulating the regional LAI, plant height, soil water content, and yield and the corresponding RMSE decreased by 0.50 m2/m2, 0.15 m, 0.02 cm3/cm3, 344.11 kg/ha. Compared with the INSs currently used by farmers, the optimized INS was associated with a marginally higher irrigation quantity (average increase of 2.66%) but a significantly lower nitrogen application rate (average decrease of 31.6%). The crop yield, WUE, and partial factor productivity for nitrogen were 6.35–12.51%, 2.64–10.41%, and 61.8–66.2% higher, respectively, in different typical-rainfall years when using the optimized INS than when using an unoptimized INS. The results of this study indicate that a regional INS established on the basis of regional crop model parameters estimated using remote sensing data can aid irrigation and nitrogen application scheduling.

Keywords

• Regional-scale CERES–Maize model
• Remote sensing
• Water use efficiency
• Partial factor productivity of nitrogen
• Multiobjective optimization