In this study, daily Escherichia coli measurements at six locations in an urban watershed in Houston, were undertaken over a period of 12 weeks, and were analyzed using time series and fractal analyses. The time series analysis revealed that the E. coli data series were nonrandom in nature and were characterized by a lack of periodicity. Shorter E. coli time series data sets (on the order of 10 days or less) exhibited a fractal structure, suggesting that micro scale time series may be fractal in nature in urban environments, a finding that has significant implications for bacteriological water quality monitoring. Although stormflow E. coli concentrations were significantly higher than baseflow levels, the range of variability in E. coli concentrations both during dry and wet weather conditions was comparable, indicating the residual impacts of rain events on bayou water quality. While other studies in the literature have shown that afternoon E. coli levels were lower than morning levels, the results from this study demonstrate the complexity of this phenomenon and its dependence on flow, turbidity, total suspended solids, temperature and the location/land use of the monitoring point (upstream or downstream and rural/urban). Spatial variability was highly correlated to land use with key differences between grassland and urban uses: urbanized sites exhibited higher overall E. coli concentrations, experienced rebound in E. coli levels during and after a rain event, exhibited no correlations between total suspended solids and E. coli, and exhibited less daily variability in bacteria concentrations.