• SAFETY & DISASTER

The cost and difficulty to perform continuous sampling of faecal contamination in aquatic environments leads to a deficit of water quality information in protected areas close to discharges from sanitation systems. Therefore, bathing water quality is commonly assessed by process-based models. However, process-based modelling demands a high computational cost: hydrological, hydraulics, hydrodynamic and water quality simulations. This fact rules out process-based modelling as a robust decision making system when an early response is required (< 24 hours). These limitations are the starting point of CYCLOPS, which aims to develop an operational system based on deep neural networks (DNN) for monitoring and warning the impact of sanitation systems.

Due to the scarcity of field information for training DNNs, a lumped hydrological model, a storm water management model and a coupled hydrodynamic and water quality model will be calibrated and validated to provide a sufficiently extensive record of information to feed and train DNNs. These models will be fed using the information from CMEMS reanalysis products. Finally, CYCLOPS will forecast (2-3 days) river flows, sanitation system discharges and evolution of faecal contamination in the aquatic environment and will operationally assess the compliance of Directive 2006/7/EC by using CMEMS forecasting products.

The Copernicus Marine Service (CMEMS) and Copernicus Atmosphere Monitoring Service (CAMS) provide regular and systematic reference information on the physical state and dynamics of the ocean and marine ecosystems for the European Atlantic Ocean and atmospheric information (total precipitation and solar radiation, wind velocity), respectively. This capacity encompasses the description of the current situation (nowcast), the prediction of the situation a few days ahead (forecast), and the provision of consistent retrospective data records for past years (hindcast). CMEMS hindcasting products have been used to calibrate process-based models and train the Deep Neural Networks, whereas CMEMS forecasting products are used to provide the forecasting healthy risk assessment.

CMEMS products http://marine.copernicus.eu/ are key elements of the CYCLOPS service. CYCLOPS makes use of the following products:

• Atlantic-Iberian Biscay Irish- Ocean Physics Reanalysis (IBI_MULTIYEAR_PHY_005_002)
• Atlantic-Iberian Biscay Irish- Ocean Physics Analysis and Forecast (IBI_ANALYSISFORECAST_PHY_005_001)

CAMS product: CAMS global atmospheric composition forecasts

The CYCLOPS Service will create the availability of, until now, non-existent tools that would enhance healthy in these aquatic spaces by reducing human infections:

• Aquatic Space Public Users: Through a CYCLOPS web space, each user will be able to have detailed and real-time information of health risks and their variability throughout the day. It should be mentioned that lack of knowledge of bathing areas is one of the risk factors that contributes to the increase cases on infections at beaches. Providing information in a predictive manner would therefore help to minimize this risk factor, enabling the selection of the healthiest spaces at every moment.
• Lifeguard Services: CYCLOPS offers those in the lifeguard service to carry out a health identification and evaluation of the aquatic space in which they are working. In addition, they will be able to predict future risks for their beaches, enabling them to make decisions that are more efficient on preventive strategies for their environment, helping to reduce infections.
• Administration: CYCLOPS provides the administration, such as owners and managers of aquatic spaces, with the tools for the management and improvement of sanitation networks and bathing areas. It is necessary to remember that everyone who comes into contact with the inherent risks in these areas, are vulnerable to suffer an accident, by which reducing infections grants the administration important savings in healthcare spending.