Water quality assessment of river using RBF and MLP methods of artificial network analysis (case study: Karoon River Southwest of Iran)

Karoon River as water supply provider of 16 cities, several villages, and thousands of hectares agricultural land is a very important river in Iran. The present research was conducted at the Mollasani hydrometer station in Karoon River using statistical data of a 49-year period. The aim of the present study is to evaluate the quality of the Karoon River using artificial neural network analysis of multilayer perceptron (MLP), radial basis function (RBF), and regression methods. The studied parameters included: TDS, pH, HCO₃, Cl, SO₄, Ca, Mg, Na, K, total anions and cations, sodium absorption ratio (SAR), total hardness (TH), and electrical conductivity (EC). The study results showed that the MLP neural network with a hidden layer and R² = 0.903 gives a more accurate estimation of SAR compared to the RBF model and regression method. Based on these parameters, the water quality was classified as good. In general, with due regard to the precautions currently in place, the basis of the Kendall test, in recent years, despite decreasing the pH of water, the amount of salts in the water has increased, which indicates a decrease in river water quality.