Water samples from sites potentially impacted by septic tanks, cattle, sewage treatment plant (STP) and natural forests were collected at regular monthly intervals and within 48 h of rainfall events between October 2004 and June 2006. All samples (n=296) were analysed for faecal coliforms and faecal sterols including coprostanol, epicoprostanol, cholestanol, cholesterol and 24-ethylcoprostanol. Faecal sterol ratios were used to assign human and/or herbivore contamination sources and to estimate their percentage relative contributions in water samples. The catchments had significantly different profiles of designated contamination origins (p<0.05), which were consistent with land use patterns. The STP impacted site had the highest incidence of human contamination assignations and the highest mean levels of coprostanol, whilst the forested site had the highest incidence of uncontaminated samples and the lowest mean concentration of coprostanol. Coprostanol concentrations were not always correlated with faecal coliform counts.