热带气旋气候数学模型的预报应用

使用西太平洋海温格点资料，选取若干个因子，组成多个复合因子，建立权重方程，使得单因子的相关系数信度检验0．05提高到复合因子的信度检验0．01，权重方程的信度检验提高到0．001。用权重方程产生的突变的高阶非线性预报方程，其Y与X的相关系数比1阶线性方程提高5％左右。自1999年至今，热带气旋年、月频数气候预测的模型投入到实际预报应用，其预报准确律达到75%～90％。使用非线性预报模型作了逐日气压、逐日雨量的气候预测。将沿海气压场、雨量场的气候预测结果用于分析、制作热带气旋登陆中国以及广东地区的时段、地段的气候预报，准确率达80％～90％。

Forecast of tropical cyclones using climatic mathematical model

Several factors selectedand combined from Western Pacific surface seawater data wereused to construct weighted equations. As a result, reliability of correlation coefficients was improved from 0.05 for single- factor to 0.01 for complex-factors, with the reliability of weighted equations up to 0.001. The Y-X correlation coefficient of a high order non- linear equation through mutation was higher than that of a simple linear equation by about 5 percent. The model forecast has been employed since 1999; during this time the accuracy rate has been verified at between 75 to 90 percent. The non-linear forecast model has been used for daily climate pressure and precipitation prediction. By analyzing the climatic forecasting results of coastal pressure and precipitation, the non-linear forecast model has shown an accuracy rate of between 80 to 90 percent in the forecasting of tropical cyclone climatic landing time and landing spots in Guangdong and other coastal prov- inces of China.