Comparison of different models predicting the date of beginning of flowering in pea (Pisum sativum L.)

In pea, the time of flowering is mainly related to the photoperiod (P) and mean temperature (Tm) during the vegetative period. In field conditions, both variables depend mainly on the latitude (LAT) and the date of sowing (RDS). On the basis of these four variables, several empirical models simulating the time to flowering either in days or in degree-days were calibrated (the parameters determined) and compared for pea (cv. Solara). Data were from trials in various locations throughout France over 8 years and with several sowing dates (from mid-November till mid-April). Surprisingly, the model with the more explicative variables (P and Tm) did not give the most reliable results in field conditions as assessed with a validation sample of situations including many years and locations. The best fit and MSEP (mean square error of prediction) were obtained by combining P, RDS and LAT in the model. Models can be constructed to use days or degree-days: models in days are very useful for crop management, whereas models in degree-days are well-adapted for crop modelling. The case of autumn sowings was analysed separately: models were recalibrated to give a good account of the whole range of sowing dates.