Assessment of the Water Quality of Karun River Catchment Using Artificial Neural Networks-self-Organizing Maps and K-Means Algorithm

Abstract

Analyzing the water quality status is one of the most important issues to control the pollution discharged to the river. This helps to obtain the effective environmental management. However, the conventional water quality indices cannot accurately indicate the water quality status, particularly in the presence of the various water quality parameters. In this study 15, water quality parameters of Karun River were categorized based on the similar variation pattern by Self-Organizing Map (SOM) for a 6-year period between 2013 and 2018. Data were then clustered using the k-means nonhierarchical algorithm for determination of the spatio-seasonal pattern of Karun River's water quality and discrimination of the sources of water pollution all along Karun River catchment. The results indicated the dissimilar variation patterns of some water quality parameters which were might due to the different sources of pollutions around the Karun River catchment. Moreover, data clustering by k-means specified 5 different groups collected data. Spatio-seasonal assessment identified the critical stations in the Karun River during a 6-year period. The result of the applied methodology showed an integration of SOM and k-means algorithms gives an insight into the similarities of water quality parameters and analysis of the water quality conditions in critical locations. Therefore, it can be used as a tool for efficient decision making in environmental management.