Netherlands urban canals are not designated as official bathing sites for recreation, and their water quality is not monitored routinely. The municipality of Breda has ambitious plans to ensure the recreational water quality in the Breda canals during the bathing season.

The purpose of this study was investigation of microbial water quality of the Breda canals intended for recreational purposes and simulation of water quality in the SOBEK model during the bathing seasons.
The first part of this study focused on analysis of microbial water quality of the study area by water sampling. The study area was divided into five sampling locations. The sampling scheme was defined weekly during dry weather, 3rd day after light rainfall without a CSO discharge and 3rd and 5th days after rainfall event with a CSO discharge. The sampling period was spread over 48 weeks. Also, DNA concentration of markers for four groups of organisms; humans, ruminants, dogs, and birds were monitored by a DNA source trace analysis.

Temporal and spatial variations of E. coli were observed in the Breda canals after rainfall events in the winter and summer seasons, indicating that water quality can be affected by rainfall events. The results of the DNA resource trace of the Breda canals suggested that, in dry weather, humans and ruminants have always been a source of contamination. Depending on circumstances and locations, dogs and birds may also act as contaminants. After rainfall events with a CSO discharge, human faecal material was a source of contamination for most locations.

The second part of this thesis involves testing the sewer model of Breda in the context of discharged overflow volumes. To run the water quality model, overflows volume simulated by the sewer model during the sampling period are used as input. The comparison of simulated and measured water levels at pumping stations indicated that, generally, the sewer model is valid, and the simulated water levels at pumping stations corresponded to the measured values. However, the model overestimated the water level due to inaccurate modelling of the pumps’ operations and capacities.

The last part of this study presents the water quality simulation in SOBEK model. Based on the sensitivity analysis of this model, it can be concluded that there is uncertainty surrounding some essential inputs of the model, including discharged sewer overflow volumes and initial E. coli concentration. To develop a more accurate water quality model, the quality of input data should be improved.
The results of this study indicated that the municipality should adopt a strategy to minimise the discrepancy between simulated and actual discharged sewer overflow volumes. In addition, a better rain–runoff model (RR module) for conducting simulations in SOBEK can be developed by the Water board.

Considering the complicated simulation of pumps’ operational systems in InfoWorks and high expenses of sensor installation for collecting reliable data on all sewer overflows, water quality analysis based on collecting more sample data in different rainfall events sounds more feasible.