Chemometric Application on Surface River Water Quality: A Case Study of Linggi River, Malaysia

Chemometric methods were applied to Linggi River water quality data sets to evaluate spatial - temporal variations and identify sources of pollutants. In 2011, the data sets consist of 11 variables, monthly collected from 26 sampling stations. Three clusters (C1; 6, C2; 10 C3; 10 stations) were generated using a cluster analysis method; four latent factors were identified by principle component analysis with factor analysis method. The scree plot was used to identify the number of principle components and show pronounced change after 4 eigenvalues. Four principle components explained about 80.8 % of the total variance in the water data sets from eigenvalue > 1. Using a discriminant analysis method, the discriminant function weighed with 5 (pH, biochemical oxygen demand, suspended solids, iron and ammonia nitrogen) and 7 (pH, conductivity, biochemical oxygen demand, suspended solids, iron, manganese and ammonia nitrogen) variables capable to distinguish between 84.6 % spatial and 83.3 % temporal variation, respectively. Linggi River water quality status were explained in that pollution mainly originated from agriculture run off or aquaculture residue (nutrients), soil weathering or industrial discharge (metals), and solid waste disposal site or municipal discharge (biochemicals). This study successfully demonstrated that chemometric methods demonstrate useful information for further monitoring strategies.