In studies of the biological effects of UV radiation, ozone depletion can be mimicked by performing the study under ambient conditions and adding radiation with UV-B lamps. We evaluated this methodology at three different locations along a latitudinal gradient: Rimouski (Canada), Ubatuba (Brazil) and Ushuaia (Argentina). Experiments of the effect of potential ozone depletion on marine ecosystems were carried out in large outdoor enclosures (mesocosms). In all locations we simulated irradiances corresponding to 60% ozone depletion, which may produce a 130–1900% increase in 305 nm irradiance at noon, depending on site and season. Supplementation with a fixed percentage of ambient irradiance provides a better simulation of irradiance increase due to ozone depletion than supplementation with a fixed irradiance value, particularly near sunrise and sunset or under cloudy skies. Calculations performed for Ushuaia showed that, on very cloudy days, supplementation by the square-wave method may produce unrealistic irradiances. Differences between the spectra of the calculated supplementing irradiance and the lamp for a given site and date will be a function of the time of day and may become more or less pronounced according to the biological weighting function of the effect under study.