Hydro-chemical analysis of the Edough springs
using statistical and ANNs approaches, northeast Algeria
H. Majour*, A. Hani, K. Khanchoul, L. Djabri
Département de Géologie. Université Badji Mokhtar - Annaba, B.P. 12, Sidi Amar, 23000 Annaba, Algérie
* Corresponding author. E-mail: email@example.com
Received: 04 February 2008; revised version accepted: 25 July 2008
The metamorphic Edough massif, located in the northeast of Algeria has been chosen to analyse the physico-chemical elements contained in the springs waters whose source is generated from the hydrogeological system. The supply of the aquifer is due to the rainfall infiltrations. In this work the artificial neural networks have been used to simulate the flows and the mineralization of the springs waters. Consequently, we have shown the effect of the physico-chemical and climatic factors and more particularly the structural one on the hydrological behaviour of the springs. The study of the physico-chemical characteristics of the aquifer was based on the analytical results carried out on a set of samples every 15 days. On the basis of a statistical treatment, the main processes ruling the observed chemical evolution within the aquifer were defined. The infiltration of meteoric water governs the dilution of the chemical elements and the increase of potassium contents. The waste water increases the mineralization of groundwaters. The rock matrix is submitted to dissolution according to the carbon dioxide contribution, which explains mineralization.
Keywords: Crystalline medium; Artificial neurons networks; Flows; Mineralization.