OPTIMIZATION OF ENDODORMANCY RELEASE MODELS, USING SERIES OF ENDODORMANCY RELEASE DATA COLLECTED IN FRANCE

Abstract

Evaluation of endodormancy release models in temperate fruit tree species has generally been indirect, as a component of the global evaluation of bud break-date models that include an endodormancy release module. Due to the lack of data, their direct evaluation based on the knowledge of the dynamics of bud dormancy states has been very succinct. Consequently, these models have not been optimized, except for the 'amount of chill units required to break endodormancy' taken as the genotype specific parameter. As part of different studies on bud state during the rest period, near 30 yearly dynamics of the dormancy state of 'Redhaven' peach vegetative buds under natural conditions have been recorded, based on the biological 'one node cuttings' test under standard conditions (25°C). These data were used to optimize classic models (Weinberger-W; Utah-U, etc.) of endodormancy release in fruit tree species with temperature as the only input variable, in order to compare their goodness of fit and their predictive capacity. The fitted parameters were the 'amount of chill units required to break endodormancy', those of the temperature functions f(T) describing the rate of endodormancy development and, when relevant, the starting date of endodormancy release. The results and practical conclusions for predicting dormancy release date are presented. The optimization method used was the minimization of the mean square error (MSE) between the observed (biological test) and the computed dates of endodormancy release and was carried out using the nls module of R software. As main results: the optimization of the W model resulted in an increase of the standard threshold temperature; the endodormancy release effect of negative temperatures (up to freezing injury) in the optimized form of U model that was not taken into account in the standard form; the sharp transition from positive to negative f(T) values for the temperatures above 12°C , in the optimized U model; and the fact that the model of Weinberger is a particular form of smoothed Utah model.