World Journal of Engineering Research and Technology

*Ndoh Mbue Innocent and Bitondo Dieudonné

ABSTRACT

This longitudinal study assesses the seasonal variations in water quality for pollution source identification in the west coast of Cameroon using a combination of statistical techniques. A total of 13 physico-chemical parameters were collected in both the dry and wet seasons of 2017 from ten locations constituting three sampling sites (sea, rivers and springs) for analysis. Principal component analysis retained 12 significant parameters, yielding three principal components with ?>1, explaining 81.92 % of the total variance in the data sets. All the parameters were significantly correlated with seasons (p<.01) except pH. Mann-Whitney U-values for the parameters were found to be statistically significant with large differences between seasons: F- (U = 26.000 (Z = -3.343: r = -.68), p = .00); pH (U = 14 (Z = 3.860: r = .73)), p = .000; SO42- (U = 22 (Z = 3.509: r = .66), p = .000); and TDS (U = 15 (Z = 3.814: r = .72), p = .000). Discriminant analysis (independents entered together) further revealed a significant association between seasons and all predictors (Wilks ? = .328, Chi-square = 23.981, df = 9, Canonical correlation = .82, p =.004). Overall, the results revealed that while some parameters might be crucial in determining the fluctuation of pollutants for one season, the same parameter might be less crucial for another season. EC, Na+, Mg2+, F-, Cl-, NO3- and TDS were relatively higher in concentration than when compared with the WHO (2004) guidelines for drinking water quality and these abnormalities were common in in sea and rivers. This study demonstrates the usefulness of both descriptive and multivariate statistical techniques for analysis and interpretation of complex data sets. The results are important for necessary management decisions to control current and future pollution of receiving water bodies.

[Full Text Article] [Download Certificate]