| 摘要: |
| 利用建立在计算机系统的热带气旋历史资料,经反复试验提出了一套集热带气旋发生的日期相似、地理位置相似、路径相似和强度相似等相结合的挑选"最相似"热带气旋的方案;并利用所挑选的路径和强度"最相似"历史样本,分别对各预报时效的移动方向、移动速度和强度进行分级概率值统计,根据最大概率值选取相应的移向、移速和强度作为热带气旋路径和强度的预报。本方法已在计算机系统成功开发并投入业务试用,检验结果表明:本方法具有一定的预报能力,其输出的未来24、48、72小时的路径和强度预报对预报员有一定参考价值。 |
| 关键词: 热带气旋 相似和概率集成 预报方法应用 |
| DOI： |
| 分类号: |
| 基金项目:广州中心气象台科研基金资助 |
|
| METHODS OF ANALOGY AND MAXIMUM PROBABILITY ENSEM-BLE IN THE FORECAST OF TROPICAL CYCLONES ON AND OPERATIONAL APPLICATION |
|
LIN Liang-xun, CHENG Zheng-quan, HUANG zhong, HE Xia-jiang
|
|
Guangdong Meteorological Observatory, Guangzhou 510080, China
|
| Abstract: |
| In this paper,a method for tropical cyclone(TC) track and intensity forecast is proposed.Based on the similarity in date,location,track and intensity,the method retrieves the most similar TCs from the historical samples in computer database.The retrieved samples are classified and analyzed statistically in terms of their future direction,speed and intensity.And the TC track and intensity with maximum probability assembly are chosen as forecast results.The method has already been implemented and utilized in operational application.The forecast ability of this method is demonstrated with test results. |
| Key words: TC forecast analogy and maximum probability assembly method and application |
