effect of abiotic factors such as temperature, relative humidity, sunlight and rainfall. METHODS AND RESULTS: Candida sake (5 × 10(7) CFU ml(-1)), combined with different concentrations of the food additive Fungicover(®), was applied on grape berry clusters. Treated clusters were then exposed to abiotic factors in field or laboratory conditions, recovering populations to describe C. sake population dynamics. The addition of Fungicover significantly increased C. sake multiplication under optimal growth conditions and improved survival under fluctuating abiotic factors. After field applications, significant differences in populations on grape bunches exposed or covered by fine foliage were detected. Simulated rainfall washed off C. sake populations by 0·6-0·9 log units after 20 mm of rain volume. Allowing populations to establish for 24 h or more, prior to a rain event, persistence on grape berries significantly increased and the effect of rain intensity was not observable. CONCLUSIONS: Candida sake demonstrated high survival ability under unfavourable environmental conditions and persistence under intense rain. The study evidenced the importance of the first period just after application for C. sake survival on grape tissues and also the protective effect of the additive Fungicover. SIGNIFICANCE AND IMPACT OF THE STUDY: This research provides new information on the survival of C. sake under field conditions and its practical implications for recommending timing of spray with this antagonist. Our results could be useful for other yeast antagonists applied before harvest. This work, for the first time, defines population dynamics of a yeast BCA using simulated rainfall.