Study On Ecological Evaluation And Yield Prediction Model Of Rice Production Area In Shandong Province

At present,the production of rice in China still faces a series of issues such as food security,climate change,non-point source pollution and other problems.From the perspective of rice growth,it depends strongly on the natural conditions like climate,soil and the management measures such as fertilization,irrigation and cultivation and other.In Shandong Province,both the quality and the yield of rice are good in Shandong Province,especially in the area of lakes,swamps,the areas along the Yellow River and coastal saline alkali areas.Rice is an ecological suitable crop,which has the advantages that other crops cannot replace.In order to explore the advantages of rice production in Shandong Province further,the environment-friendly green production mode of rice was explored.Based on the systematic evaluation of the eco-environmental quality of rice production area,the research of rice potential productivity and yield prediction model was carried out.The main research results are as follows:1.Ecological evaluation of rice producing area in Shandong Province.The ecological environment(production conditions,climate resources,soil fertility and environmental quality)of rice production area in Shandong Province has good overall suitability.The photoperiod during the rice growth season is 1172.8 hours,and the effective accumulated temperature reaches 4094.4?.The conditions of light and temperature are superior.Soil organic matter and nutrient contents are relatively high while heavy metal elements are low.These superior ecological environment conditions make Shandong an important rice production area in North China with good quality and high yield.However,due to the limitation on the total amount of water resources,the rice planting area in Shandong Province has been decreasing year by year.In addition,the amount of chemical fertilizer application is relatively high,and there is imbalance between nitrogen and phosphorus application in the Inter Cropping and rice growing season,and the risk of non-point source pollution is outstanding.These are the main problems in rice production in Shandong Province,especially in two maturing area of rice and wheat(Linyi irrigation area library).2.Study on rice potential productivity in Shandong Province Based on improved potential reduction method.On the basis of the potential of climate production and the consideration of factors related to soil fertility,the potential of rice production in Shandong Province was explored by the improved potentiality reduction method.The results showed that the potential productivity of rice in Shandong Province under natural precipitation accounted for 50% of the rice potential under irrigation.Therefore,irrigation is an important measure to ensure the stable and high yield of rice.At present,the yield per unit area of rice accounts for only 67.5% of the paddy soil potential productivity.The main factors that restrict the sustainable development of Shandong rice are the disorder of varieties,the inadequate prevention and control technology of the pest and grass damage,the imperfect mechanized production technology and the lag of the brand creation of rice products.3.Research on production prediction model of time series based on artificial neural network optimized by intelligent algorithm.The prospect of rice yield in Shandong Province was predicted by the production prediction model of time series based on artificial neural network optimized by intelligent algorithms.The research combines genetic algorithm,artificial bee colony algorithm and particle swarm optimization algorithm and optimizes the BP neural network respectively.Four rice yield prediction models including BP(1),GA-BP,PSO-BP and ABC-BP have been established.The experiment results show that the MAPE values of the four yield prediction models are 5.72%,5.15%,4.93% and 5.32% respectively,and the RMSE% values are 7.42%,6.74%,6.14% and 6.6%,and the Theil IC values are 3.48%,3.16%,2.88% and 3.15% respectively.The results of the study show that the best performance is the PSO-BP model after optimized by intelligent algorithms with the obvious improvement of the prediction efficiency,prediction accuracy and prediction stability compared with the BP(1)model.The prediction of rice yield in the next 3 years by the PSO-BP model shows that the rice yield in Shandong Province still has a downward trend,which needs to be paid more attention by the relevant departments.4.Study on prediction model of rice yield in Shandong Province Based on rough set optimization neural network.Rough set is used to optimize the structure of neural network,and many factors that influence the yield of rice in Shandong Province are reduced.By eliminating redundancy and simplifying structure,rice yield prediction models BP(2)and RS-BP were set up respectively.The experimental results show that the MAPE values of the BP(2)and RS-BP are 10.22% and 4%,the RMSE% values are 12.05% and 4.28%,and the Theil IC values are 5.70% and 2.02%.The prediction speed and accuracy of RS-BP model is much better than that of BP model and time series prediction model GA-BP,PSO-BP and ABC-BP.5.Design and development of rice information management and yield prediction system in Shandong.The rice information management and yield prediction system of Shandong Province was designed and developed based on the survey data of the ecological environment in the rice production area and research results of rice yield prediction model.The system applies the principle of object-oriented programming,implements the system architecture and interface customization by programming language,and realizes the coupling of rice information management,production potential calculation and production prediction model and computer technology.The comprehensive service platform of rice can provide data query,potential estimation and output prediction for the production management and research departments.