Sustainable Watershed Management Through IoT-Driven AI (SWAIN)

River waters are used in large quantities by many industrial facilities for various purposes such as cleaning or cooling. This carries the continuous risk of a chemical spill into rivers. Recent studies reveal alarming adverse effects of chemicals, particularly micropollutants, on water ecosystems and humans. Therefore, detecting micropollutants in rivers and locating the source of the spills is of utmost importance for environmental sustainability. Existing detection systems are both too costly and unable to identify micropollutants on time. We aim to develop an early warning system for micropollutant spills in rivers based on artificial intelligence techniques. The system will make use of sensors collecting various environmental data continuously. It will then match the previously generated fingerprint of industrial facilities with the collected data to identify the source facility in a matter of minutes after the spill. The decision making will be based on a novel technique that combines human expertise by environmental scientists and artificial intelligence fueled with continuous data. The system will stay current and adapt itself over time based on the changing environmental conditions. Since the sensors will be deployed in remote areas, particular attention will be given to the fault tolerance and energy efficiency of the data collection infrastructure. We will validate the proposed system in Ergene River, Turkey and Kokemäenjoki River, Finland. The ultimate goal of this project is to design and demonstrate the first artificial intelligence based early warning and prediction system for domestic, industrial, and agricultural pollution in European rivers.

The SWAIN project has achieved a major breakthrough in sustainable water quality management through an innovative approach that combines low-energy, long-range sensor networks with advanced AI analytics. By strategically placing sensors at critical points along rivers and watersheds, SWAIN has enabled precise data collection on pollutants even with sparse sampling. This efficient design minimizes environmental impact, as sensors require minimal power and are only deployed where necessary, making it adaptable and scalable for larger water networks. One of the project's standout achievements is its ability to provide nearly real-time pollution tracking and source identification, a significant improvement over traditional methods that are often labor-intensive and slow to produce actionable results. This rapid detection capability empowers decision-makers to respond quickly to pollutant spills, helping to prevent large-scale environmental damage and protect water resources. Furthermore, SWAIN's model integrates data from various sources-such as industrial activity, agricultural practices, and natural water flow patterns-into a unified view, allowing stakeholders to see both immediate and long-term trends in water quality. Conducted in collaboration with partners from TU Wien, Finnish Environment Institute, Istanbul Technical University, Bogazici University, and Università della Svizzera italiana, SWAIN focused its research on the Ergene River in Turkey and the Kokemäenjoki River in Finland. These rivers, both ecologically vital and under significant industrial pressure, served as prime study sites for demonstrating the project's robust approach to pollution detection and water quality management. The project's advancements promise to improve pollution monitoring practices across Europe, providing a cost-effective, sustainable solution that supports cleaner, safer water. In the long term, this model will be instrumental in shaping policies and environmental standards by enabling better resource management and preventive measures, fostering a healthier ecosystem for communities and natural habitats alike.