黑龙江稻区稻瘟病流行趋势预测数学模型研究

A study on the Predictive Mathematic Model for Epidemical Tendency of Rice Blast in Rice Districts of Hei-longjiang.

应用因子分析和逐步回归方程对黑龙江省水稻主要产区稻瘟病流行情况和气象因素间的关系进行分析，分别制定了牡丹江地区和佳木斯地区非特殊灾变年份的稻瘟病流行趋势预测数学模型：Ｙ（牡）＝＝－５．６８６＋１０．９６１Ｘ１－２．００１Ｘ２－１０．１７３Ｘ３＋３．５５Ｘ４－６．０６４Ｘ５＋２．６８６Ｘ６－４．８２８Ｘ７－２０．３７１Ｘ８＋Ｍ＋Ｎ±１３；Ｙ（佳）＝－１６．４２－１．０８２Ｘ１－０．５４４Ｘ２＋１．６３７Ｘ３－４．９Ｘ４－１０．６４５Ｘ５＋Ｍ＋Ｎ±１１。用上述方程历史回拟率分别为８４．０％和８１．９％。经两地近年应用预报均准确。。应用该模型在７月１１日发预报时，所用气象资料唯有７月中旬相对湿度不能从气象台获得预测值。因此，根据对历史资料的分析，得出７月中旬相对湿度预测方程：Ｈ７２＝７４．２２５＋０．３４６１Ｈ７１－１．１２６Ｔ７２±６．３Ｈ７２：７月中旬相对湿度（％）Ｈ７１：７月上旬相对湿度（％）Ｔ７２：７月中旬平均气温（℃）±６．３为９０％置信区间。方程的复相关系数，ｒ＝０．６４０。Ｙ（佳）＝－１６．４２＋１．０８２Ｘ１－０．５４４Ｘ２＋１．６３７Ｘ３－４．９Ｘ４－１０．６４５Ｘ５－Ｍ＋Ｎ±１１Ｘ１：７月上旬?

Through analysing the relationship tetween epidemical tendency of rice blast and the climatic factors in the major rice districts of Heilongjiang by using factorial analysis and stepwise regression equation,epidemical tendency predictive mathematic models of rice blast for a nonparticular calamity year in Mudanjiang and Jiamusi region were established respectively. They are: Y(Mu)-25.686+ 10. 961X1-2.001X2-10. 173X3+ 3. 55X4-6. 064X5+2.686X6-4.828X7-20. 371X8+M+N±13;Y(Jia)=-16. 42+1.082X1-0.544 X2+1. 637X3-.9X4-10. 645X5+M+N±11. The models were also evaluated.Their historical false co-incidence percentages are 84.0%and 81.9%respectively.A better fitness was oberved between the actual and theorebical values in recent years.