Ecological Systems and Devices Annotation << Back A Method for Classifying Surface Waters and Determining Regional Thresholds for Hydrochemical Indicators Tunakova Yu.A., Novikova S.V., Valiev V.S., Baibakova E.V. The method of classifi cation of surface water bodies belonging to the same basin by multidimensional neural network clustering with fuzzy elements and expert assessment was substantiated. The data sets on the values of hydrochemical indicators obtained in the system of state environmental monitoring of .surface waters of the Volga-Kama basin in the period 2014–2021 were studied. Clustering of the available sets of measured hydrochemical indicators of water bodies was carried out using Kohonen neural networks, the advantage of which is the ability to retrain when new data arrive, and the selection of heterogeneous groups with a given degree of accuracy using an algorithm similar to the k-means algorithm. Six classes of surface waters covering the whole variability of the background hydrochemical composition of the Volga-Kama basin waters were identifi ed, qualitative assessments were given to them and classifying indicators were identifi ed by the factor analysis method. The threshold values of classifying indicators were statistically determined. An algorithm for assigning an arbitrary water sample to one of the selected classes of surface waters is proposed. Keywords: surface waters, hydrochemical indicators, classifi cation, representative classifying indicators, threshold values. DOI: 10.25791/esip.1.2025.1496 Pp. 15-23.