Factor Analysis and Prediction Model Study of Wheat Stripe Rust in Hanzhong City

zhang jichang,wang qingwen,li zhao,li le,zhang yong,wang baotong,wang hui,yang jianjun
DOI: https://doi.org/10.7606/j.issn.1009-1041.2017.12.16
2017-01-01
Abstract:In order to verify the epidemic law of wheat stripe rust in Hanzhong city,and to make the crop disease prediction more scientific and accurate,the DPS data processing system was used to analyze the dominant factors of wheat stripe rust in Hanzhong,according to the local records of recent years.As the consequence,the dominant factors affecting the extent of the disease were screened,including the proportion of wheat susceptible varieties,percentage of diseased fields and the average number of susceptible leaves per 667.7 m2 in last autumn,mean air temperature of last January,rainfall of last November,and percentage of diseased fields in early spring and mid-March.Based on the regression analysis of main factors and the occurrence degree,the prediction model of disease severity was established:y=-6.354 7+0.084 0 x1+0.022 8 x3+0.662 8 x6+0.020 9 x11,R2=0.968 7.Tested by both fitting predictive feedback and actual observation,the historical coincidence rate of this model is 92.31%,and the relative error is 8.96%.Furthermore,the model was used to conduct the prediction of Hanzhong wheat stripe rust during 2014-2016.The predicted results are 2.93,3.21 and1.92,respectively,which coincide with the actual records of 3,3,2.Thus,the model can be applied into practice.
What problem does this paper attempt to address?