Assessing Spatiotemporal Water Quality Variations in Polluted Rivers with Uncertain Flow Variations

Abstract

Many rivers in developing countries receive high pollution loads from large cities. Such rivers also experience extreme flow variations due to overexploitation of freshwater resources, poor management practices, and climate change impacts. Consequently, levels of water quality parameters (e.g. biochemical oxygen demand, dissolved oxygen, unionized ammonia, and coliforms) considerably change with the extreme flow variations. The morphology of a river segment also influences the levels of water quality parameters by changing the residence time. For a river, a water quality index is a robust assessment tool to indicate the overall water quality for different segments (e.g. having natural freshwater, receiving single point source loads, and cumulative loads from several outfalls) along the length. The index also captures the impacts of different uncertainties on water quality due to seasonal variations in river flows, hydrodynamics, and pollution loads. Water quality data are scarce in developing countries due to the absence of planned periodic monitoring programs. Type-2 fuzzy sets, an extension of the ordinary fuzzy sets, directly model these uncertainties by providing an additional degree of freedom. Type-2 fuzzy sets improve the specific kind of interface that exists due to increasing uncertainties associated with imprecision in knowledge and vagueness in information due to limited water quality data. Using the triangular type-2 fuzzy sets approach, the index developed in the present work effectively caters for these uncertainties and has proved to be a reliable water quality assessment measure for highly polluted rivers with extreme flow variations in developing countries and elsewhere with similar conditions.