Prediction Of Typical Grain Crop Yield In Northwest Hebei Based On Meteorological Factors

On a global scale,climate change has always existed and is uncertain.The stable development of grain output and stable output level play a very key role in the harmonious and stable development of China.Therefore,the research of grain yield forecast can provide the basis for the government to determine grain prices,and can also remind farmers to adopt appropriate agricultural technology measures to cope with the climate change.Zhangjiakou is the central city of northwest Hebei.In this thesis,the characteristic crop naked oats in Zhangjiakou is taken as the main research object.The three main meteorological factors including temperature,precipitation and sunshine duration are taken as the main influencing factors.Specific work is as follows:(1)This thesis sorts out the existing problems in grain yield forecast and demonstrates the feasibility of constructing the combined forecasting model of grain yield in Zhangjiakou.(2)The selection and pretreatment of meteorological and naked oats production data have been completed.The time trend production prediction model has been established,and then the meteorological production prediction model has been established by stepwise regression analysis method and BP neural network algorithm.(3)By combining the stepwise regression prediction model and BP neural network prediction model,the combined prediction model is designed to forecast the grain yield in northwest Hebei.The grain yield data of Zhangjiakou in 2017 were selected to verify the prediction model.The experimental results showed that the combined forecasting model was more accurate than the single forecasting model to predict the production of naked oats in Zhangjiakou.The innovations of this thesis are as follows:(1)Based on the validity criterion,the stepwise regression model and BP neural network model are combined,and the model can overcome the shortcomings of the common combination prediction model.(2)Forecast the output of naked oats in northwest Hebei using meteorological data.Combined with the previous research status of grain yield prediction models in different regions,Zhangjiakou was taken as the research area to build a combination prediction model based on meteorological factors to forecast grain yield.To sum up,the following conclusions are reached in this study: the combined prediction model based on the validity criterion has a high prediction accuracy,which can reach 82%.The model is relatively stable,the fluctuation range is relatively small,and the error rate is low,which can better predict the grain yield in northwest Hebei.