Implementation of the BBCH coding system for winter oilseed rape (OSR) phenology simulation can allow detailed description of crop ontogeny necessary for crop management and crop growth modelling. We developed such a BBCH model using an existing approach (Habekotté 1997).

The new model describes winter OSR development by a combination of differential and conversion equations based on the structure of the BRASNAP-PH model (Habekotté 1997). Six phenological phases were reproduced daily according to the BBCH codes (00–89): emergence (00–09), leaf development (10–19), stem elongation (30–39), inflorescence emergence (50–59), flowering (60–69) and pod development-maturation period (70–89). The model takes into account temperature (including vernalisation) and photoperiod as the main environmental forces affecting crop phenology. The macro stages of leaf development and shooting were reproduced considering the rates of leaf appearance and internode extension. Model calibration and validation were performed using an extensive database of phenological observations collected from several experimental sites across France (n = 144), Germany (n = 839) and Italy (n = 577). The stability of the parameterisation was checked by a cross-calibration procedure.

Applied to the independent datasets used for validation and cross-validation, the model was able to predict the whole-crop cycle with a root mean square error (RMSE) of 2.8 and 3.2 BBCH stages, respectively. Particularly accurate predictions of winter OSR development were obtained with the Italian datasets (RMSE: 2.1 and 2.3 BBCH stages for validation and cross-validation, respectively). Considering the phenological phases separately, emergence, leaf development, flowering and the pod development–maturation period were simulated with RMSE of 1.0, 2.4, 2.9 and 3.2 BBCH stages, respectively (validation datasets). Slightly higher uncertainty emerged in the prediction of stem elongation and inflorescence emergence phases (RMSE: 3.5 and 4.1 BBCH stages, validation datasets).

The model reproduced winter OSR development with a sufficient degree of accuracy for a wide range of years, locations, sowing dates and genotypes, resulting in an efficient and widely applicable prediction tool with relevant practical purposes in the crop management scheduling.