This study presents an systematic approach to water quality assessment, hybrid modelling and decision support for eutrophication management in deep reservoirs. It is found that during the summer monsoon the catchment runoff into the Yongdam reservoir induces a transfer of pollutants from a middle stratified layer to the surface layer. Although the transport mechanism limits nutrient accumulation on the bottom of the reservoir, it also offers an opportunity for on-going algae production in the surface water. Physically based modelling is used to understand the process of micro-scale turbulent mixing and its impact on the nutrient uptake by algae. Further, a data-driven model using clustering and partial least squares regression which uses results from a physically based model of the reservoir successfully predicts Chlorophyll-a concentrations.