基于马尔科夫链理论对贵州锦屏白背飞虱发生程度的预测

Forecasting the Occurrence Degree of Sogatella furcifera(Horváth) in Jinping of Guizhou Province by Applying Markov Chain Theory

白背飞虱是我国水稻上的重要害虫,以贵州一季中稻区白背飞虱为研究对象,运用马尔科夫链理论,以代表站点锦屏1990-2011年的白背飞虱田间发生程度的时间序列资料,构建了1~5阶的概率转移矩阵,以预报年前5年的连续发生级别预测第6年的发生状态.回报锦屏17年的结果,得到了较高的历史符合率,达到了82%,对锦屏2012年白背飞虱田间发生级别的预测结果也符合当年的实际发生情况.

White-backed planthopper,[Sogatella furcifera（Horváth）],is an important rice pest insect in China.In a study reported in this paper,the time series data of S.furcifera field occurrence level from 1900 to 2011 in Jinping,aquaternary middle rice-growing area in Guizhou,were selected to establish a probability transfer matrix of 1 to 5 steps according to Markov chain theory,and to forecast,based on the continuous occurrence level of the previous five years,the occurrence of the pest in the sixth year.When used to reversely forecast the occurrence of S.furciferain Jinping in the preceding 17 years,this matrix gave a satisfactory historical coincident rate（82%）,and the results based on prediction in 2012 of Jinping were also consistent with the real situation of the year.