APPLICATION OF THE SELF ORGANIZING MAP METHOD FOR THE CLASSIFICATION OF THE ENVIRONMENTAL QUALITY OF THE LAKE SYSTEMS IN THE MOROCCAN MIDDLE ATLAS: LAKES CASES OF IFRAH, IFFER AND AFOURGAGH

Abstract

Self-Organising Maps (SOM) and other methods of computational intelligence are effective tools to model and solve environmental problems. In this study, we proposed a classification approach based on self-organizing maps to highlight the seasonal classification of physicochemical parameters of the waters of lakes systems in the Moroccan Middle Atlas: Ifrah, Iffer and Afourgagh lakes.

The self-organizing map's results define four clusters of samples. The cluster I and II samples were significantly characterized by high concentrations of bicarbonates (HCO₃^- (mg / l)), Sodium (Na ⁺ (mg / l)), Magnesium (Mg^{2 +} (mg / l)), (TDS (mg / l)) and the suspended solids (MES (mg / l)). In the other side, those of cluster III have high concentrations of calcium (Ca^{2 +} (mg / l)), Potassium (K ⁺ (mg / l)), Sulfates (SO₄^2- (mg / l)) and Nitrates (NO₃^- (mg / l)). Indeed, the samples of cluster IV are characterized by low values of all parameters except the Total Dissolved load and the electrical conductivity (EC (µs / cm) are high compared to other groups. The results show that the classification by self-organizing Kohonen maps (SOM) is an excellent and fundamental tool for understanding and visualizing spatial and temporal distribution of the samples of the three lakes of the Middle Atlas.