目前国内外在热带气旋预报方面所使用的统计模式,概括起来主要有两大类:一类以相似为基础,另一类则以回归方程为基础,在60年代后期,国外一些天气预报中心已分别将这两种模式投入业务使用。在学习国外的先进技术基础上,结合本地区的预报实践经验,我们发现前者由于根据台风的前期路径进行预报,因此前期台风定位的准确程度对预报结果影响很大,而台风中心的准确观测是相当困难的;而后者的关键则在於能否选出适当的因子使得预报量基本上是所选取的这些因子的线性函数。在多数情况下, At present, the statistical models used in tropical cyclone forecasting at home and abroad mainly sum up two major categories: one is based on similarity and the other is based on regression equation. In the late 1960s, some weather forecast centers abroad Respectively, these two modes into business use. Based on the advanced technology from abroad and combined with the practical experience of forecasting in the region, we find that the accuracy of the previous typhoon positioning has a great impact on the forecast results due to the forecast of the former based on the typhoon’s previous path. The accurate observation of the typhoon center Is very difficult; the latter is the key to whether the appropriate factor can be selected so that the forecast is basically a selected linear function of these factors. In most circumstances,