关键词: 风场, 低空急流, 自动识别, 聚类, 拟合

Abstract: It is too inefficient to draw low level jet artificially and there are a lot of subjective factors. In order to solve this problem effectively, a method that can achieve automatic identification of low-level jet is presented. Conventional meteorological data, intensive Automatic Weather Stations（AWS） observation data, wind field information of MICAPS and feature of low level jet are used in this method. Automatic recognition of low-level jet on 850 hPa constant-pressure surface in coastal areas of east China has been realized by pretreatment, clustering and fitting. A large number of experiments show that the automatic recognized jet stream satisfies the meteorological requirement. In addition, the efficiency is also improved.

Key words: wind field, low-level?jet, automatic recognition, clustering, fitting