Study On Direct-seeding Rapeseed Growth Simulation Model

This study takes Huayouza9as the main test material, which is a representative hybrid rapeseed variety of the Yangtze River, combines with other varieties of rapeseed, and then constructs four dynamic simulation sub-models, such as phenological developmental stages, green area index, photosynthesis and dry matter accumulation model, dry matter partitioning and yield. Each sub-model’s algorithm and their effective coupling are advanced by database SQL combining with advanced programming language C#. The paper builds a direct-seeding rapeseed growth simulation model ultimately, which has mechanism and practicability. The main contents and conclusions are as follows:1. Rapeseed phenological developmental stages were used the theory of crop physiological development time (PDT) to simulate, and logistic equation was used to simulate the rapeseed photoperiod effectiveness, which can accurately describe the rule of rapeseed photoperiod. The model introduced temperature sensitivity (TS), physiological vernalization time (PVT), photoperiod susceptibility factor(PSF), pod period factor (PPF) as four genetic parameters of varieties, it fully quantified the effects of temperature, vernalization, photoperiod and pod maturing rate for the growth and development of oilseed rape, the practice indicated that four parameters reflected the differences between the different varieties, and TS reflected the effectiveness of temperature sensitivity in different varieties, PVT reflected the length of time that varieties required to complete vernalization, PSF reflected the effectiveness of photoperiod susceptibility in different varieties, PPF reflected the rate of varieties’ pods maturing. The prediction errors root mean square errors (RMSE) for development stages showed that:the result of four sowing dates was2.67d, the result of different varieties was3.74d. A method of application through meteorological data and the model was presented, and it was used to determine the suitable sowing date, the results of validation showed that the suitable sowing dates of Huayouza9were from September10th to October27th in2011, and the rapeseed yields was higher than other sowing date in the sowing dates, in addition, taking into account of the impact of previous crops harvest date, direct-seeding rapeseed planting in the appropriate date range had better sow late.2. Leaf area index was used specific leaf area (SLA) to simulate. According to the law of leaf growth and experimental data, the paper put forward a new rule of SLA variation, and used a logarithmic and cosine function to simulate the relation of SLA and growth process. The simulation R2was0.9260, the results showed that the method simulated leaf area index of rapeseed accurately. Meanwhile using specific pod area (SPA) simulated pod area index. The model which had fewer parameters simulated the change rule of LAI and PAI with physiological development time accurately and effectively. Validation results showed that average RRMSE (relative root mean square error) of simulated and measured LAI, PAI with different sowing dates were14.16%,10.46%.3. According to the specifically characteristic of rapeseed canopy, the canopy was divided into two levels (pod layer and leaf layer),and the canopy structure and light interception were using the Gaussian integration to simulate. Considering the diversity of three leaf types in different pheno logical developmental stages, the paper used photosynthetic rate of different leaf types to reflect the change of leaf photosynthetic rate. Also calculated consumption of growth respiration and maintenance respiration. Then builded rapeseed’s photosynthesis and dry matter accumulation model. Validation results showed that the average simulation error RRMSE of different sowing dates was6.05%.4. The dry matter partitioning and yield model adopted dry matter partitioning index to simulate dry matter partitioning dynamic changing process, and through the seed-pericarp ratio and pods dry weight to calculate the yield of rapeseed. The model simulated the relationship of rapeseed’s different organs’ dry matter partitioning and growth process, and used a set of mathematical functions with certain physiological significance to simulated the relationship. The coefficient of determination of roots, stems, green leaves, flowers and pods dry matter weight were0.9895、0.9876、0.9935、0.8555、0.9561respectively. Validation results were that the average simulating errors RRMSE of roots, stems, green leaves, flowers and pods dry matter weight were9.37%,9.10%,11.87%,14.97%,4.77%respectively; Yield validation results showed that RRMSE of simulated and measured theoretical yield was3.40%.The results showed that the direct-seeding rapeseed growth simulation model had better mechanism, reliability and predictability, and laid a good foundation for digitized cultivation, management and production of rapeseed.