Surface water quality analysis using multivariate statistical techniques: a case study of Fars Province rivers, Iran.

Abstract

This study aimed to transform the input of a large dataset into the output of interpretable information. Hence, multivariate statistical methods were carried out to analyze physicochemical parameters in 34 rivers during a 17-year period (1997-2014). Cluster analysis divided the study area into spatially different riverine water quality sub-regions described in ascending order of water quality as severely polluted (SP), highly polluted (HP), polluted (P), moderately polluted (MP), lightly polluted (LP), and not polluted (NP). By diagnosing threats and identifying fragile zones, water contamination sources responsible for impaired water quality in the study area recognized as natural pollutants in LP, municipal wastes in P, discharge of industrial effluents in MP, natural geochemical formations in SP and HP, and superficial flows of agricultural lands in SP, HP, and MP. The dominant water type in each zone wasclassified into Na-Cl, Na-Cl, Na-Mg-Ca-Cl-SO4, Na-Ca-Mg-Cl-SO4, Na-Ca-Cl, and Ca-Mg-HCO3-SO4 groups for SP, HP, P, MP, LP, and NP, respectively. To explore aesthetic aspects of drinking water application, hazard quotient (HQ) was applied for children and adults in terms of ingestion and dermal exposure. Overall health risk assessment revealed the order of impacts of the secondary water quality parameters as Cl- > Na+ > total dissolved solids (TDS) > Ca2+ > SO42- > Mg2+. Furthermore, hazard index (HI) ranged from 0.011 to 31.439 and 0.010 to 30.122 for children and adults, respectively, indicating a potential health risk regarding chloride throughout the whole region excluding NP. To identify significant agents in water quality, principal component analysis extracted 3 varifactors (VFs), with the eigenvalues of 4.74, 1.19, and 0.85, respectively, explained about 83% of the variance. The most important parameters in the first factor were TDS, electrical conductivity, SAR, TH, Na+, Cl-, and SO42- accounting for 58% of the total variance. The most influenced parameters in the second and third factors were pH and HCO3-, respectively, with variance coverage of 26%. These factors indicated that the hydrochemical characteristics of the water originated by natural interactions (existing salt domes, evaporation, weathering, and soil erosion) and anthropogenic activities (fertilizer-rich flows of agro-fields and domestic/industrial disposals), which must be minimized in rivers to supply the population with hygienic water.