Using PCA and ANN in the explanation of groundwater chemism. Case of the Messaad Plateau. South Algerian Steppe Region

Abstract

The objective of this work is to study the chemistry of groundwater in the messaad plateau which is located 300km south of the capital of Algiers. The  results obtained showed that the waters are very strongly mineralized overall, with a predominance of calcium sulphate facies and calcium chloride  facies. a high degree of mineralization in the South and South-East part of the study area near Oued Messaad. The Principal Component Analysis (PCA)  has shown that the increase in mineralization is mainly due to the dissolution of evaporative mineral-rich salt-bearing formations. The hydro-chemical  model based on Artificial Neuron Networks (ANN) of the multilayer perceptron (MLP) type has shown that the SO₄ ^2-,Cl^- ,Ca²⁺,Na⁺ ions are the most  important factors influencing the electrical conductivity of water. These ions are the result of the dissolution of gypsum and halite continuously in the  clays and marls of the Barremian formations.