Phenology modelling and forecasting in olive (Olea europaea L.) using artificial neural networks.

The phenological behaviour of seven olive cultivars from south and central Italy was studied in the environment of inland Tuscany. The potential use of artificial neural networks (ANN) in olive phenology modelling was investigated. A backpropagation neural network was trained and tested for predicting the date of occurrence of defined olive phenophases in inland Tuscany. The network was fitted using meteorological data by ten-day periods as inputs and phenological event dates as outputs. The model tested was also used to predict the responses of olive phenology to future hypothetical climate change by simulating alterations in temperature and light conditions.