With the use of different multivariate statistical analysis methods, spatio-temporal fluctuations in the water parameters of Tiru reservoir located at the Marathwada drought-prone area of Maharashtra, India, have been analysed and reported in this case study. Tiru reservoir, situated on the tributary of the Godavari River, was regularly monitored at five different sites from August 2017 to January 2019 for the estimation of 20 water quality parameters. Various multivariate methods such as pattern reorganisation using cluster analysis (CA), factor analysis/principal component analysis (FA/PCA), and discriminant analysis (DA) were used for handling complex datasets. CA extracted three different clusters from five sampling sites with similar water quality characteristics. FA/PCA extracted thirteen factors (65% of 20 measured) required to explain 74% of the data variability and identified the factors accountable for variation in water quality and also evaluated the prevalence of each cluster on the overall dissimilarity at five different sampling sites. Discriminant analysis extracted a total of 16 parameters with 97.7% right assignations. Varifactors (VFs) acquired by factor analysis recommended that the water quality parameters accounted for variation were linked to two groups. The first group included water quality parameters like T, DO, SDD, turbidity, TDS, PA, and MA, whereas the second group covered most of the nutrients Cl-, silicates, PP, TP, NO3-N, NO2-N, and NH3-N; hardness; and CHL-a and mainly entered the reservoir during surface runoff from agriculture fields and the surrounding area containing domestic as well as animal waste. Thus, the present work showed the efficiency of multivariate methods for the assessment of spatial as well as a temporal variation in the water quality of a small reservoir.
Keywords: Cluster analysis; Discriminant analysis; Factor analysis; Principal component analysis; Reservoir; Surface water analysis.