Research On Integrated Technology Of Fine Yield Dynamic Prediction Of Winter Wheat Based On The Combination Of Impetus And Statistics

Providing objective and quantitative yield forecast by carrying out study of fine yielddynamic prediction of our country’s main crops, could greatly improve the level and qualityof agricultural services of meteorological department. Therefore, newly developed statisticalmethod, cluster analysis of statistical test (CAST), was used for regionalization of winterwheat of Shandong Province, based on which, models for predicting winter wheat yields ineach sub-region were established according to the historical good or bad harvest ofmeteorological impact index of crop yield, the key meteorological factors impact index, theclimatic suitable index and the WOFOST crop growth model and so forth. Based on theresults of the above four methods, using computer simulation technology, a winter wheatyield dynamic prediction method was developed by using weighted method. The mainconclusions of the study are:(1) The division results of the newly cluster analysis of statistical test (CAST) showthat, the CAST method could divide the study area objectively, and the divisions are objectiveand unique.(2) The prediction results are relatively better in North region of Shandong and Southregion of Shandong based on the dynamic prediction method of winter wheat historical goodor bad harvest impact index, that only few of forecasting accuracy did not reach90.0%.(3) The prediction results of the method of the key meteorological factor impact indexwhich combine with the biological characteristics of the crop and consider the meteorologicalfactors in the growth stage of winter wheat that having a great impact on the yield, show that,the average prediction accuracy are relatively higher in North region of Shandong and Southregion of Shandong, while the prediction results in other regions are unstable.(4) Combined with physiological characteristics of winter wheat and suitability indexmodel, a winter wheat yield dynamic prediction model was constructed based on winterwheat climatic suitability index. The trend forecast of this method is unstable, but, theaverage prediction accuracies in each sub-region are generally higher, nearly all above90.0%. (5) The yield dynamic forecast model of winter wheat at different growing times wasestablished using the WOFOST crop growth model combining with measured data andmeteorological material under average climatic conditions. The simulation results are poordue to the yield harvest trend. While the simulation results of the four areas are good due tothe average prediction accuracy of average yield.(6) Based on the historical good or bad harvest of meteorological impact index of cropyield, the key meteorological factors impact index, the climatic suitable index and theWOFOST crop growth model, the dynamic integrated prediction method of winter wheatyield was established using accuracy weighting method. Integrated prediction method, eitherthe yield trend forecast, or the average yield forecast accuracy has improved, and the stabilityis improved too. Indicating that integrated method may be more suitable for application onforecasting business.