The excessive and unpredictable growth of cyanobacteria (blue-green algae) in surface waters around the world threatens public health and recreation. This can lead to high societal costs when e.g. intake for drinking water production (DWP) has to be interrupted. Biogeochemical self-organisation in hydrological networks results in either an undesired algae or a desired water plant dominated state. The algae dominated state retains few nutrients and may therefore cause emergent cascade effects of deteriorating water quality downstream. Hitherto, models to predict and prevent cyanobacterial bloom formation are lacking. In the proposed project knowledge institutes NIOO-KNAW, Wageningen University Biometris and Deltares aim to develop such simulation models for complex man-made hydrological networks of surface waters. Specifically, we focus on bloom formation at the Valkenburgse Meer which is important as a bathing location and a potential intake for DWP by Dunea. At daily time scales, we will predict bloom occurrence through a Cyanobacterial Bloom Formation Early Warning System (CBF-EWS) to support short instant mitigation measures to ensure good water quality for DWP and recreation. This will involve real time weather data assimilation and in situ nutrient and algae measurements. At yearly time scales, we will perform scenario analyses to identify management measures acting on the hydrological network and the biogeochemical processes therein that prevent cyanobacterial blooms. This will ultimately lead to the development of Smart Nutrient Retention Networks (SNRNs). The resulting open-source models will be implemented at Dunea and the Waterboard of Rijnland and be available to Witteveen+Bos and RoyalHaskoningDHV for projects worldwide.