Vol. 126, No. 1 * Pages 1–157 * January - March 2022

The aim of this work was the regionalization of low flow for chosen catchments located in the upper Vistula river basin using non-hierarchical cluster analysis. Next, with such creative clusters, the regional relationships were determined between the specific low flow discharge q₉₅ and the meteorological and physiographic parameters of the catchment. The study evaluated regional regression models for low flow (specific q₉₅ discharge) in selected, 30 catchments located in the upper Vistula river basin. The data for calculations were a series of observations of daily discharge from the multiannual period of 1963–2016 and were obtained from the Institute of Meteorology and Water Management – National Research Institute in Warsaw. The study showed, that the k-means method can be used for regional regression determination. The parameters which influenced the catchments grouping in clusters were the specific low flow discharge q₉₅, precipitation, median catchment altitude, mean catchment slope, soil, and land use. The study indicated that k-means method may be an effective tool for evaluating low flow in rivers of the southern parts of Poland.