基于四套不同数据集的西北太平洋热带气旋快速增强现象统计对比分析

Comparison of the Rapid Intensifications of Tropical Cyclones over Western North Pacific Based on Four Best-Track Datasets

利用中国气象局(China Meteorological Administration, CMA)上海台风研究所、美国联合台风警报中心(Joint Typhoon Warning Center, JTWC)、日本气象厅(Japan Meteorological Agency, JMA)和中国香港天文台(Hong Kong Observatory, HKO)发布的四套热带气旋(Tropical cyclone, TC)最佳路径集资料,对比分析了1985—2014年间发生于西北太平洋上TC快速增强(Rapid intensification, RI)现象。根据TC记录24 h强度变化的上95%累积概率分位确定四套资料RI阈值分别为26 kn(CMA)、30 kn(JTWC)、25 kn(JMA)和25 kn(HKO)。四套RI样本时空分布特征差异不大,TC多发于菲律宾以东的西太平洋暖池区,以9、10月份最多。对比发现,JTWC资料中1~3级TC发生RI的频数和比例均远高于其他资料,经历RI后多增强为4~5级TC,从而导致强台风频数高于其他资料。考虑到四套资料在确定台风强度(定强)时使用的风速标准不同,利用CI指数与近地面最大风速转换表将其统一为1 min风速平均标准后发现,在RI频数以及4~5级TC频数方面,转换后的JMA资料与JTWC资料较为一致,CMA资料和HKO资料相近,但较前述两家明显偏少。这表明RI数量的差异不仅是因为采用的风速平均标准不同,也是因为对强台风的定强方面存在差异,但该差异在2005年后明显缩小。

This paper compares the characteristics of rapid intensification of tropical cyclones(TCs) over the western North Pacific during 1985—2014 based on four TC best-track datasets which issued by China Meteorological Administration(CMA), Joint Typhoon Warning Center(JTWC), Japan Meteorological Agency(JMA), and Hong Kong Observatory(HKO), respectively. The thresholds of RI are examined firstly according to the 95% of total samples of 24 h TC intensity change. They are 26 kn in CMA, 30 kn in JTWC, 25 kn in JMA and 25 kn in HKO, respectively. Using the different thresholds, the spatio-temporal characteristics of RI are compared to demonstrate similar features which show the high frequencies over the east ocean of Philippines in September and October. But the frequencies of RI in JTWC dataset are higher than those in other three datasets obviously, resulting in the highest numbers of super typhoon. Considering the wind-averaged thresholds are different, it is recomputed after standardized by the relationship of CI index and the maximum of surface wind. The results indicate that the RI frequencies are similar in JMA and JTWC which are higher than that in CMA and HKO. It suggests that the differences are not only caused by the use of different wind-averaged thresholds, but also by the technologies in determining the intensity of strong TCs. But the RI discrepancies have been decreased rapidly since 2005.