Assessment of water quality and Algae growth for the Ganwol reservoir using multivariate statistical analysis

Abstract

ABSTRACTComprehensive multivariate statistical techniques (i.e. analysis of variance, correlation analysis, principal component analysis and factor analysis, and multiple linear regression model) were applied to evaluate both temporal and spatial variations in 13 water quality parameters of eutrophic Ganwol reservoir collected on monthly basis for three years (2014–2016). From the results of comprehensive multivariate statistical techniques, both temporal and spatial variations in nutrient concentrations (N and P) inside the Ganwol reservoir were found to be substantial. Also, the water quality of each monitoring site was affected by variations in loadings of natural and anthropogenic factors from various pollution sources. Both principal component analysis and factor analysis were successfully applied to identify important components/factors accounting for most of the variance of whole water quality of Ganwol reservoir, and to generate different numbers of varifactors (VFs) of latent pollution sources/factors for each monitoring sites. Finally, multiple linear regression analysis using VFs as independent variables reasonably estimated the eutrophic state (Chl-a) of Ganwol reservoir. Therefore, comprehensive multivariate statistical techniques can identify both temporal and spatial variations in complex water quality parameters and in different loadings of natural and anthropogenic factors, convert huge water quality parameter structures into simpler factor structures (i.e. VFs), and offer a valuable site-specific solution for reliable management of agricultural reservoir.