Classification of annual precipitations and identification of homogeneous regions using K-Means method [pp. 1609-1622]

Abstract

Reliable and correct estimation of hydrological and meteorological processes is one of the major problems in regions with insufficient hydrologic information and data. The classification of the hydrological variables and determination of homogeneous regions are the most important steps of regional studies. The purpose of this study is to classify the annual total precipitation series and to identify the homogeneous regions by K-Means method. The K-means method, which is the simplest and most commonly used clustering method, divides a data set into clusters by minimizing the sum of the Euclidean distance between each feature vector and its closest cluster centre. The annual precipitation records and longitude, latitude and altitude values obtained of 188 stations operated by the National Meteorology Works (DMI) in Turkey were considered for clustering analysis. The number of clusters was determined as 7. Moreover, the regional homogeneity test was applied for testing the homogeneity of regions identified by clustering analysis.