Microbial Risks in Surface Water Sources

Doctoral thesis, 2011

Microbial risks need to be properly handled for the provision of healthy drinking water. The mitigation of pathogens in the water source lowers the risk for infection associated with nominal and suboptimal drinking water treatment. Several recent waterborne outbreaks have been related to point discharges of sewage in surface waters, involving pathogenic bacteria, viruses and parasites. Diffuse sources further contribute with faecal matter, although these are difficult to estimate compared to point discharges. As rainfall often results in pathogen peak concentrations in surface waters, the microbial levels during rainfall events need particular consideration. Different approaches have been used in this thesis to assess the microbial release from point and diffuse faecal sources in surface waters under dry and wet weather conditions. Wastewater discharges as a point source from the urban sewer systems along the River Göta älv were investigated, and several log-unit higher microbial loads were found in wet weather compared to dry weather. Results from a pathogen sampling program combined with quantitative microbial risk assessments indicated the highest infection risk for noroviruses and enteroviruses. The ability to close the river raw water intake for Gothenburg prevented pathogen peak concentrations from reaching the water treatment plants. To regulate this pathogen barrier, information on upstream discharge events, rainfall, indicator bacteria and turbidity was found more appropriate than the reliance on a single threshold value for E. coli. Bacteroidales human (BacH) and ruminant (BacR) genetic markers, targeting faecal matter from humans and ruminants, were quantified in water samples collected around the Lake Rådasjön. Positive correlations with cultivable faecal indicators in event samples provided support for the impact of human faecal matter from on-site sewers. The human and ruminant markers were detected in soil at several log-units lower levels compared to faecal matter. In microcosm trials, the inactivation of the genetic markers agreed with the faecal indicator bacteria and somatic coliphages, and decay coefficients were used in hydrodynamic modelling for the Lake Rådasjön. A Bayesian Monte Carlo approach including expert judgements on faecal sources was successfully applied to provide a reliability measure for findings of genetic markers at specific sampling sites.