Assessment of surface water quality of Ain Zada dam (Algeria) using multivariate statistical techniques

• Published in: International journal of river basin management 2017-04, Vol.15 (2), p.133-143
• Main Authors: Bouguerne, Ammar, Boudoukha, Abderrahmane, Benkhaled, Abdelkader, Mebarkia, Abdel-Hafid
• Format: Article
• Language: English
• Subjects: Ain Zada dam; Algeria; Assessments; Cluster analysis; Discriminant analysis; factor analysis; Freshwater; Parameters; principal component analysis; Principal components analysis; Statistical methods; Surface water; Variance; Water quality
• ISSN: 1571-5124; 1814-2060
• DOI: 10.1080/15715124.2016.1215325

Multivariate statistical techniques, such as cluster analysis (CA), principal component analysis (PCA), factor analysis (FA) and discriminant analysis, have been applied for the assessment of temporal variations of surface water quality in Ain Zada dam, Algeria, for 10 years by monitoring 16 parameters. The different parameters indicate that the data are homogeneous. As against this record an annual variation is more important than the monthly change in connection with climate change. The facies of these waters is Cl-Na especially in connection with human actions. Values of the Water Quality Index classified the surface water as medium to good quality. The Pearson correlation analysis revealed a significant positive relationship between salinity and all variables and negative relationship between water volume of dam and all variables. The CA in R mode grouped the 16 variables into 4 clusters of similar water quality characteristics and in Q mode, 160 sampling are grouped into 2 statistically groups where total dissolve solids and capacity seem to be major distinguishing factors between variables and years. The CA has classified the data into two groups, one formed by the dry years and the other formed by wet years. The PCA and the FA applied to the datasets have resulted in two significant factors which represent 69.92% of total variance. The first factor as salinization factor explained 58.68% of the total variance. The second factor, can be called organic pollution factor, explained 11.24% of the total variance. The results of discriminant analysis showed only 11 parameters were necessary in the temporal variations analysis, affording more than 90% correct assignations.