Temporal Variation Analysis of Rice Yield in the Jiangsu Province, China: Application of Decision Support System for Agrotechnology Transfer Model

Abstract

The accuracy of grain yield estimation is critical for national food security. Because of the comprehensive influence of spatial differentiation conditions, such as temperature, precipitation, soil, rice variety, and irrigation, yield estimation requires integrated modeling that is based on dynamic conditions. These dynamic conditions include geographical background, biological factors, and human impact. Most existing studies focus on the observation and analysis of external factors; only a few reports on yield simulations are coupled with nature, management, and crop growth mechanism. Our study incorporates the crop growth mechanism of rice, along with data of rice varieties, soil, meteorology, and field management, to determine the rice yield in Jiangsu province, China. In addition, we have used a decision support system for the agrotechnology transfer model, along with Coupled Model Intercomparison Project data and geographic information system technology. Our results showed that: (1) A calibrated variety genetic coefficient could simulate rice biomass value (flowering stage, maturity stage, and yield) reasonably. The values of NRMSE (Normalized Root Mean Square Error) between the simulated and measured values after parameter calibration are all less than 10%, the values of d(index of agreement) are all close to 1, the simulated value of yield is in good agreement with the measured value. (2) A linear correlation between the meteorological elements and yield was observed. The linear correlation had regional differences. Notably, an increase in precipitation was conducive to the increase in yield. Except at the Huaiyin site, the other sites showed that the temperature rise could potentially lead to reduced production. We found that an increase in solar radiation was unfavorable to the production of rice in the northern and western sites in the Jiangsu province, whereas it was conducive in the southern and eastern sites. (3) Our study predicted the rice yield from typical sites in the Jiangsu province from 2019 to 2060 in the wake of climate change while excluding the extreme effects of diseases, pests, typhoons, and floods. The order of average yield per unit area is as follows: Xinghua site (8212.76 kg/ha) > Huaiyin site (7912.70 kg/ha) > Gaoyou site (7440.98 kg/ha) > Gaochun site (7512.29 kg/ha) > Ganyu site (7460.88 kg/ha) > Yixing site (7167.00 kg/ha). Notably, the average yields from the Xinghua and Huaiyin sites were higher than that from the Jiangsu province (7617.77 kg/ha). The fluctuation of the yield per unit area at each site was generally consistent with the fluctuation in the overall yield, showing a downward trend and tends to be stable. The dispersion of yield per unit area indicates that Gaochun had the most stable yield per unit area followed by Xinghua, Ganyu, Yixing, Huaiyin, and Gaoyou. The yield per unit area of the Huaiyin and Gaoyou sites was unstable and portrayed the biggest fluctuations. Future studies need to focus on how to deal with spatial variation and carry out adaptive verification to make the simulation results applicable to more dimensions.