The use of weather types for the definition of seasons is examined. Specifically, an objective definition of seasons is attempted for the southern Balkans based on the intra-annual variations of the frequencies of the objectively defined weather types. The methodology includes a combined application of Principal Component Analysis and k-means Cluster Analysis on daily grid point meteorological data obtained from the ERA5 database, for the 40-year period 1979–2018. The methodology is used at first for the classification of weather types and then for the definition of seasons based on the intra-annual variations of the weather types’ frequency. The application of the methodology results in nine weather types and four seasons. The characteristics of the defined weather types are connected to the prevalence of specific atmospheric circulation patterns, the windward or leeward character of the various sub-regions, and the thermal characteristics of the earth’s surface. The four seasons generally correspond to the conventional ones, but they present differences regarding their onset and cessation dates and their duration.

检查了使用天气类型定义季节。具体而言，尝试根据客观定义的天气类型频率的年内变化，对巴尔干南部地区的季节进行客观定义。该方法包括将主成分分析和k-均值聚类分析相结合，应用于从ERA5数据库获得的1979-2018年40年期间的每日网格点气象数据。该方法首先用于天气类型的分类，然后用于基于天气类型频率的年内变化来定义季节。该方法的应用导致了九种天气类型和四个季节。定义的天气类型的特征与特定大气环流模式的普遍性有关，各个子区域的迎风或背风特性，以及地球表面的热特性。四个季节通常与常规季节相对应，但是它们在发作和停止日期以及持续时间方面存在差异。