Surface Water Quality Management Using Multivariate Analysis Methods (A Case Study: Shafarood River, Gilan Province, Iran)

Surface water management is directly related to determining the correlation between physical, chemical and biological variables as well as identifying their natural and anthropogenic origin.The Shafarood River (Gilan Province, Northern Iran) is subject to the discharge of residential and agricultural pollutants, so it is very important to control its quality and determine the sources of pollutants and its quality status.The present study was conducted to scrutinize the correlation between the major parameters affecting the water quality of the Shafarood River and to monitor the water quality in different areas of the river using canonical correlation analysis and cluster analysis models, respectively.Five physical and four chemical parameters were measured at ive stations based on Standard Methods for the Examination of Water and Wastewater 2015 over six years.The results revealed a signi icant correlation between two categories of response (physical parameters) and predictor (chemical parameters) variables, which mainly originated from anthropogenic pollution (ef luents from residential and garden areas).According to the results of cluster analysis, the stations were grouped into two clusters based on the degree of pollution, and the cluster grouping con irmed the canonical correlation matrix.The research indings revealed the effectiveness of the obtained linear combinations for the physical parameters, including total suspended solids and turbidity, as well as the chemical parameters, including biochemical oxygen demand and nitrate.To conclude, the ef iciency of canonical correlation analysis and cluster analysis methods was con irmed in identifying the determinant variables of water quality and in classifying the water quality monitoring stations in the optimal management of rivers.Considering the importance of different parameters in changing the water quality of the Shafarood River, the multivariate statistical methods have been used in a proper way in classifying the quality and determining the pollution sources.* and ** = respectively indicating between-cluster signi icance at the levels of 5% and 1%