Ultraviolet irradiation (UV) is an important alternative to disinfection for production of potable water. Viable Escherichia coli is a widely used indicator for public health risk. However, UV efficacy is reduced by suspended solids that can act as both UV shielding and UV absorbing, agents. Failure of UV irradiation can lead to an enduring health legacy. Here the probabilistic Fr 13 methodology of Davey et al. (2015) is demonstrated for turbulent flow of feed water with suspended solids irradiated in an annular reactor and, a comparison made with the traditional deterministic method. The aim was to examine the impact of naturally occurring fluctuations in suspended solids concentration on failure to inactivate viable E. coli. A UV failure factor (p) is defined in terms of the design and actual log(10) reduction in viable E. coli. UV irradiation is simulated using (Latin Hypercube) Monte Carlo sampling. Illustrative overall results show some 32.1% and 43.7% of apparent successful operations could unexpectedly fail over the long term due, respectively, to combined impact of random fluctuations in feed water flow (Q), lamp intensity (10) and shielding and absorption of UV by suspended solids [conc]. This translates to four (4) failures each calendar month (the comparison rate without suspended solids is 16% or two (2) failures per month). An unexpected finding however is, although the initial presence of suspended solids as both UV shielding (median particle size 23 mu m) and absorbing agent has a highly significant impact on reducing UV efficacy, fluctuations in concentration of these in the feed water do not meaningfully impact overall vulnerability. UV failure is impacted highly significantly by fluctuation in feed water flow. It is concluded this is strong quantitative evidence to emphasize that solids should be removed prior to the UV reactor, and that an improved flow control be used to reduce variance on feed water flow, rather than increased UV dose. This work will be of benefit to operators of UV equipment and researchers in risk analyses. Crown Copyright (C) 2015 Published by Elsevier Ltd. All rights reserved.