Biologically operated filters are known to reduce elevated levels of biodegradable organic matter produced by the ozonation process, but they also consistently release high-levels of heterotrophic bacteria. An acceptable disinfection strategy for biofilter effluent bacteria is necessary to maintain distributed water quality. The efficiencies of various strategies for disinfecting naturally occurring heterotrophic bacteria in the effluent from an ozone/biofiltration plant treating California surface waters were compared. Bench-scale tests were conducted in the laboratory, using ultraviolet (UV) light (low-pressure and medium-pressure lamps), and the results were compared with those for the disinfection efficiency of free chlorine and chloramines. Inactivation efficiencies provided by the low- and medium-pressure UV lamps were similar (k⋍0.47 cm2 mJ−1). To consistently reduce biofilter effluent bacteria levels to <10 colony-forming units per millilitre (CFU ml−1) prior to distribution, this study showed that contact with either 1 min of free chlorine or 60 min of monochloramine (both at 2.5 mg l−1, pH=8, 20°C) would be necessary, compared with a UV dose of 15 mJ cm−2 from either lamp. To evaluate bacterial regrowth potential following UV treatment, samples were incubated at 20°C in a dark environment to mimic transport/storage in the distribution system. Without residual disinfection, bacteria treated by 60 and 140 mJ cm−2 of UV were able to regrow to untreated biofilter effluent levels within 3 and 7 days, respectively. Samples treated with 20 mJ cm−2 of UV and 2.5 mg l−1 of chloramines maintained bacteria at levels of <10 CFU ml−1 for up to 7 days. To investigate the plausibility of bench-scale results, a full-scale UV reactor (treating 3 million gallons per day (0.13 m3 s−1) was operated to treat the biofilter effluent at the large-scale treatment plant for a period of 8 months. Continuous UV treatment was observed to provide long-term control of the biofilter effluent bacteria to levels similar to those afforded by bench-scale treatment. However, towards the end of the study, the number of bacteria leaving the UV reactor increased to levels exceeding 102 CFU ml−1. It was observed that 2.5 mg l−1 monochloramine residual (no free-chlorine contact) applied at the inlet of a 60-min clearwell downstream from the UV reactor kept bacteria at a level of <10 CFUml−1 at the clearwell effluent. Based on these results, it is recommended that utilities using UV to control biofilter effluent bacteria apply a residual disinfectant (such as chloramines) to maintain low bacteria levels during distribution.