Joint uncertainty analysis in river water quality simulation: a case study of the Karoon River in Iran

Estimating rivers’ responses to natural and man-made changes is of high importance in river water quality management. It is also necessary to incorporate uncertainties in planning and management of these important systems. The aim of this paper is simultaneously analyzing uncertainties of inputs and parameters of a river water quality simulation model using a novel Markov chain Monte Carlo technique, namely differential evolution adaptive metropolis (DREAM). This technique helps to consider the interaction between the input variables and parameters in water quality simulation results. A DREAM-based uncertainty analysis model is linked with the QUAL2K simulation model to incorporate uncertainties of several inputs and parameters such as headwater quantity and quality, pollution loads and reaeration parameters. A part of the Karoon River located at the southwestern part of Iran is selected for the case study. The results illustrate that the interactions among the parameters and inputs should be taken into account in river water quality simulation. The methodology presented in this paper can accurately provide the statistical characteristics of parameters of the water quality simulation model using the observed values of several water quantity and quality variables.