Linking chilling with forcing: ontogenetic competence in process-based tree phenology models.

Process-based models are widely used for predicting the timing of spring phenological events, such as vegetative bud burst or flowering, in extratropical trees. These models address both the effects of low (chilling) temperatures on rest break (endodormancy release) and those of high (forcing) temperatures on the microscopic ontogenetic development inside the buds which leads to the visible bud burst. In this Methods paper we discuss the use of the theoretical quantity ontogenetic competence, Co, in linking chilling with forcing in such process-based models. In the first part of the study, with the aid of a methodological comparison of two modelling approaches we show that Co provides a biologically more realistic option than the more frequently used empirical alternating model for linking chilling with forcing. This is because Co makes it possible to use an explicit state variable for the ontogenetic development leading to bud burst. The alternating model does not allow that, but its strength is simplicity and its straightforward relation to empirical data. In the second part of the study not related to the alternating model we revised the calculation of the empirical values of Co in our experimental method for the process-based spring phenology modelling. This revision was done to take variation of bud burst percentage in the experimental results into account in a biologically more realistic way. In general, the revision did not cause any consistent improvement in the accuracy of the model in predicting spring phenology, nor any major change in the predicted values of Co. The predicted timing of spring phenology under current climatic conditions and the acceleration rate of it under climatic warming, too, were only marginally affected by the revision. However, use of the revised experimental method is recommended because it further improves the biological realism of the models by allowing a more realistic addressing of the bud burst percentage observed in the experiments carried out for the modelling.