This thesis is writen by Steffen Skaarup and Pernille Wahlberg, two graduate students on the 3. and 4. semester of the master program: Water and Environmental Engineering by the department of Civil Engineering at Aalborg University.

The location of the project is the lake, Knudsø, situated by the city of Ry. The goal of this project is to develop a data-driven model which is able to predict the current concentration and position of bacteria in the lake and use this knowledge to inform about temporary pollution at the lake’s three bathing areas.

The system for forecasting a temporary pollution consists of two parts: an offline part and an online part. In the offline part of the forecasting model the flow patterns of Knudsø is generated in the DHI software MIKE 3 FM. In the process of generating the flow patterns of Knudsø variables like wind direction, wind speed and stratification are taken into account. By generating the flow patterns in the offline part of the forecasting system all the heavy and time consuming computations of the flow patterns are done before moving on to the online part of the system.

The online part of the forecast system is the data-driven part of the model which consists of two models. The first model is an overflow model calculating the volume of water discharged into the lake based on the measured rain intensity in the catchment area. The volume of discharged water together with the flow pattern for the specific wind direction and wind speed are used as input to the bacteria model calculating the transport and decay of the bacteria when they enter the lake. Based on the bacteria’s transport and decay it is possible to estimate the concentration of bacteria in the three bathing areas in Knudsø. If the concentration of bacteria exceed the quality requirements for excellent water quality the public is warned against bathing by the beaches until the concentration of bacteria is back under the quality requirements.