Prediction Model Research Of Sorghum Production

Sorghum is not only an important dryland food crop,but also an indispensable raw material for the brewing industry in Shanxi province.The brewing industry plays a significant role in the economic development in Tunliu,Shanxi.By analyzing the impact of climatic factors on sorghum output,the prediction of sorghum output is obtained,which can prevent the adverse effects of sorghum production fluctuation caused by climate problems on the brewing industry.In this paper,the response of climatic and environmental conditions to sorghum yield in Tunliu district is studied,and the relationship between climatic factors and sorghum yield is discussed.The main contents of this paper are as follows:1.The changing trend of climate factors(average relative humidity,rainfall,duration of light and temperature)in the past 15 years of the Tunliu district(2004~2018)is researched,and the effect of relative humidity,rainfall,light duration and temperature on the sorghum growth is analyzed.The results show that:Relative humidity has a great influence on the jointing and booting stage and heading and flowering stage of sorghum;The time quantum of 50 days before the ears to 40 days after the ears is the period of a significant relationship between precipitation and yield,which dedicates that sorghum yield has longer demand for precipitation;The effect of light duration on yield is mainly in the panicle period of 10 to 40days;The fluctuation range of daily average temperature in Tunliu district is 20~30℃,which is appropriate temperature range for sorghum growth and development.2.According to the sorghum yield and climate data in recent 14 years(2004~2017)in Tunliu district,by using the exponential smoothing method,the sorghum yield is divided into two parts: trend yield and meteorological yield.The climate factors that have great influence on the sorghum meteorological yield are obtained by using the grey correlation algorithm,and then the multiple linear regression model is fitted by using the strong correlation climate factors.1)The exponential smoothing coefficients are selected by using particle swarm optimization algorithm.By comparing the application scope of the first index smoothing method,the second index smoothing method and the third index smoothing method,the third index smoothing method is selected to decompose the sorghum yield into two parts: trend yield and meteorological yield.2)The climate factors that have great influence on sorghum meteorological output are obtained by using the grey absolute correlation method.The factors include precipitation in May,duration of moonlight in May,average relative humidity in June,precipitation in June,precipitation in July,duration of moonlight in July,precipitation in August and duration of moonlight in September.Multivariate linear regression model is fitted by using the strong correlation climate factors.3.By validated with the data of 2018,the accuracy of the regression model is 81.06%,which shows that the model can fit the sorghum yield in the Tunliu area accurately.This modeling process can provide reference for the yield prediction of sorghum or other field crops in other areas.