Use of multivariate statistical and geographic information system (GIS)-based approach to evaluate ground water quality in the irrigated plain of Tadla (Morocco)

摘要

Twenty five chemical, physical and bacteriological features of water samples from 25 wells were analyzed by multivariate statistical tools to provide the characterization of the ground water distribution in the basin of Tadla province (Morocco). The 25 parameters determined include: temperature, pH, conductivity, IP (permanganate index), dryresidue (RS), total hardness, total iron (FeT), several bacteriological residues (faecal, total coliforms and faecal streptococcus) and several cations and anions (Ca2+, Na+, K+, Mg2+, SO42−, Mn2+, NH4+, NO3, HCO-3, NO-2,Cl−). All sampling was performed in the period between December 2007 and February 2007(rain season). Principal Component Analysis together with the GIS approach(kriging methods) which provided a description of the area investigated with respect to the characteristics ofthe ground water was used.It is demonstrated that the water, quality in this region is critical. Nitrate levels are situated between 11.3 and 100 mg/L, with 73% of the observations exceeding the critical level of 50 mg/L fixed by the standards of the World Health Organization (WHO) for drinking water. However, bacteriological residues (faecal, total coliforms, and faecal streptococcus) are retrieved in nearly all water samples. Principal Component Analysis indicates that Bacteriological contamination is merely correlated with nitrite and ammonia amount rather than with nitrate amount, indicating a possible contribution of local pollution sources to ground water deterioration. The variability of the nitrate and bacteriological pollution is important and spatially correlated.A significant difference in watercomposition has been highlighted between water table of Beni Amir and water table of Beni Moussa and also a difference between ground water near cities (with a probable human polluting effect) and zones far from the built-up areas. Cluster analysis was also performed in order to evaluate the different wells similarity and to confirm the results obtained by Principal Component Analysis.