Microbial quality of reclaimed water can be adequately interpreted by graphical adjustment of experimental results to a lognormal probability distribution; the method serves to verify the degree of conformity to the proposed distribution and to visualize the time variability of the results. A conventional physico-chemical treatment can inactivate up to 1.0 μlog of common bacterial indicators, while maintaining time variability below 1.0 (μlog/100 ml). Dissolved organic matter in the influent will significantly alter the performance and the reliability of disinfection process, highlighting the need for improved adjustments of disinfectant dose to actual quality of treated water. Although those process may reach inactivation rates as high as 6.0 μlog for common bacterial indicators, their reliability may be considerably affected, forcing standard deviations beyond 3.0 (μlog/100 ml). Natural processes can provide inactivation rates from 3.0 to 3.5 for common microbial indicators, under the weather conditions studied. Although those values are lower than those of physical and chemical disinfectants, the reliability of natural processes is more stable, with standard deviations ranging from 0.65 to 1.1 (μlog/100 ml). Natural processes are particularly sensible to external inputs of microbial indicators, due to the presence of wildlife.