Résumé
Despite experimental evidence of the individual and interactive effects of photoperiod and temperature on bud growth, photoperiod has not yet been effectively accounted for in models of budburst. However, in some tree species, such as Betula pubescens (birch), photoperiod has an important role in phenological control, and its inclusion in process-based models of budburst might affect phenological projections under climate change scenarios. The aim of the present study was to integrate photoperiod into a process-based phenological model (Chuine 2000; J Theor Biol 207: 337-347; Unified model), using experimental findings in which photoperiod was found to significantly affect budburst in B. pubescens (Caffarra et al. 2011; Clim Res 46: 147-157, this issue). The effect of photoperiod was integrated into the model at 2 levels. Firstly, photoperiod, in interaction with temperature, affects the course of dormancy induction. Secondly, photoperiod modifies the response to temperature during the phase of forcing. The resulting model (DORMPHOT) for the simulation of birch budburst was fitted to a large phenological dataset, including data from different latitudes, and validated with 7 datasets from 4 different European countries. Besides giving more biological realism to the model, the newly introduced mechanisms improved its predictive performance. The DORMPHOT model outperformed the Unified model, the linear regression model (budburst date vs. spring average temperature), and the UniForc model. It also proved to be more accurate at predicting budburst in extremely warm years, which suggests it might be more reliable than previous models when using future climate change scenarios.