Mutual regularity of spring phenology of some boreal tree species: predicting with other species and phenological models

Phenological models constructed from observations of one species are often extrapolated to predict the phenology of other species. In this study, time series of the flowering and bud burst of several boreal zone trees were collected. The observation series were regressed against each other in pairs to test mutual variation. In addition, two models of phenology, one based on chilling requirement, and the other assuming ontogenetic development starting from a signal from the light climate were fitted to the phenological time series. The root mean square error of the regression models forecasting one observation series with another was quite constant for all event pairs, and the smaller the closer in time the events took place. It seems that different plant species react to climate variables in a similar manner, thus the use of the same models for different species and phenomena is justified. The light climate triggered model, albeit more simple, gave estimates that were better than those of the regression models between the events, while the average residuals of the estimates from the chilling triggered model were considerably larger. It was concluded that the chilling requirement component was redundant for prediction accuracy in the spring phenology models of boreal trees.