Time Series Analysis of Historical Surface Water Quality Data of the River Glen Catchment, U.K.

Abstract

Solute concentrations in surface waters have increased in the recent past in many developed countries, and are close to or above WHO and EC drinking water limits in many British rivers. A time series analysis of surface water quality as indexed by total oxidized nitrogen (TON) is carried out for the River Glen catchment, U.K. for the period 1976–1990. Significant seasonal trends are observed in water quality, with solute concentrations increasing during winter and decreasing in summer. In addition, a strong increasing trend is observed in the nitrogen records using both graphical outputs and statistical analysis. Standard time series analyses (auto-correlation and partial auto-correlation functions) are not applicable to irregularly sampled water quality data sets provided by river monitoring public/statutory bodies, and there is a need for an established method capable of handling such data sets that would be available for most rivers. An approach based on multiple regression analysis has therefore been developed to quantify both seasonal and annual trends in the TON water quality data. It has been demonstrated that this approach is useful for the analysis of historical water quality data sets.