Statistical and geostatistical modelling approach for spatio-temporal assessment of river water quality: a case study from lower stretch of River Ganga.

Present study deals with the application of statistical and geostatistical modelling to analyse spatio-temporal variation in water quality of selected 466 km in the lower stretch of river Ganga (545 km). Six important physicochemical parameters, e.g. water temperature, transparency, pH, dissolved oxygen, alkalinity and specific conductivity, each of thirteen selected sites of monsoon and pre-monsoon seasons are observed for the year 2014. Statistical modelling such as factor analysis is used to identify significant variables for variation in water parameters, whereas geostatistical modelling, e.g. ordinary kriging (OK) is used for spatial distributions of the water parameters in the river. Result showed alkalinity and specific conductivity during monsoon season, while transparency, pH, dissolved oxygen and specific conductivity in pre-monsoon season correlated to maximum variation in water quality. Geostatistical modelling, ordinary kriging (OK)-based interpolation model for water parameters showed more than 85% correlation at validation sites. The generated thematic map of spatial distribution for alkalinity during monsoon season and specific conductivity in both pre-monsoon and monsoon season showed an increasing trend from northern part (upward stream) to southern part (downward stream) of the river stretch. Whereas, in case of transparency, pH and dissolved oxygen, a reverse trend was observed as they increased from southern part (downward stream) to northern part (upward stream) of the river in pre-monsoon season. This combined approach for spatio-temporal water quality assessment in River Ganga can provide useful information for researchers and policy-makers for sustainable management of lower stretch of the River Ganga. [ABSTRACT FROM AUTHOR]