Minimum temperature prediction in agricultural area using artificial neural networks

The motivation for this study is derived from the potential problem of frost damage to citrus trees in the region surrounding Catania on the Island of Sicily Italy. The frost prediction technique developed during the investigation is based on the application of a parallel and distributed processing methodology to solve a time series weather forecasting problem. Past meteorological data observed and recorded in the region around Catania is used to train a neural network so that, given readings for a particular day, a prediction on whether frost formation is imminent can be made