Weibull模型在花生锈病流行预测中的应用

利用花生锈病多年的病害系统调查数据,拟合Weibull方程,计算参数a、b、c。多元回归分析表明,花生锈病流行曲线的Weibull方程的位置参数a和标度参数b可以通过在病害流行初期病情达0.5,1,5,10％的时间或其两点间的时间间距求得,而形状参数c也可以通过上述四点和所求得的a,b参数联合求得。 利用所得的a、b、c参数可按Weibull方程对病害流行全过程预测,若在流行期间持续调查病情,重新估计c值,可提高预测准确度。利用1983年春花生两套数据对本法进行检验。在各210次检验中,平均准确度为89和84％。

THE APPLICATION OF WEIBULL MODEL TO GROUNDNUT RUST FORECASTING

The parameters, a . b . c, of Weibull equation were calculated by using data of development of groundnut rust accurring in many years. Stepwise multiple regression analysis showed that the location parameter a and the scale parameter b can be calculated by the number of days for disease to reach at 0.5, 1, 5, 10% and the time intervals of two points among these ones. The shape parameter c can be calculated by the data (time and disease) of the four points and parameters a and b with the weibull equation. According to parameters a, b and c, disease development can be calculated with Weibull model and, if continuing on making investigation of disease development , we can use more data to estimate the parameter c and thus make the forecasting more accurate.