Résumé
Different models to estimate the chilling requirements of floral buds in deciduous fruit trees are available, although they have rarely been compared one with the other. Moreover, the mathematical functions and associated parameters involved in chilling models are frequently chosen rather arbitrarily. This work therefore aimed to improve the choice of the chilling model by comparing in same conditions three contrasting models on the basis of biological measurements of completion of chilling requirements. The first one was the sum of hours when temperature is below or equal to the threshold of 7°C (binary model). The second one was the sum of Chill Units considering weighting effects of temperatures (comparable to a normal function). The third one was the sum of Chill Actions calculated according to an exponential function of temperature, proposed by Bidabé in France (1967). The time-course variations of mean growth of apricot floral primordia were established for nine annual cycles in same orchard conditions of the lower Rhône valley for ‘Rouge du Roussillon’. The estimated date of end of the slow growth of floral primordia was assimilated each year to the mean date of completion of chilling requirements. Thus, for each model and each year, sums of chilling effects were calculated from fixed or variable dates (according to model) in the early autumn to the estimated date of chilling completion. A relatively high variance of the annual sum of chilling effects calculated in hours below 7°C was clearly highlighted (coefficient of variation >20%). On the other hand, lower variances were obtained for the annual sums of Chill Units and Chill Actions (coefficient of variation around 6-9%). These results corroborate previous works emphasizing the unreliability of sums of hours below 7°C to estimate chilling requirements in fruit trees, although this model is always mainly used in breeding and orchard management. They therefore suggest that breeders and growers could test other chilling models, as Chill Units and Chill Actions, faced with a probable varietal adaptation to global warming in the near future.