Estimation of Faecal Indicator Bacteria in Stormwater by Multiple Regression Modelling and Microbial Partitioning to Solids

Concerns about the contamination of sources of drinking water by stormwater motivated a sanitary survey of several urban catchments in the City of Östersund (Northern Sweden). A data subset from these surveys, comprising of faecal indicator bacteria (FIB) concentrations (E. coli and enterococci), measured for six storm events in three catchments, was used for investigating the feasibility of developing a FIB estimation procedure for the studied catchments by two approaches: (a) Multiple regression models, and (b) microbial partitioning to solids. In regressions, five explanatory variables (associated constituents) were derived from the literature and measured data: stormwater temperature and flow rate, and measurements of total suspended solids (TSS), total phosphorus (TP) and electric conductivity (EC). The obtained regression models were satisfactory for enterococci (regression of modelled FIBs on measured FIB was described by R² = 0.7), but less acceptable for E. coli (R² = 0.2). Microbial partitioning to stormwater solids from gully pots was found infeasible; the sediment sampled contained very low FIB counts. Hence, the former method is recommended for further refinement and applications.