Application of artificial neural networks for leaf wetness duration estimation

Leaf wetness duration (LWD) is an important parameter responsible for the outbreak of plant diseases but, in spite of its importance, the technology for measurements is not rather reliable. For this reason the simulation modelling appears to be a valid support for LWD assessment. In this work an approach rarely used for agro-environmental applications was tested for LWD estimation, the Artificial Neural Network (ANN). The ANN output then was used as an input for an epidemiological model to predict Plasmopara viticola infections. The aim of this work was to carry out an ANN capable to find out the relationships between the agrometeorological input and LWD and to analyse the impact of this estimated LWD on the quality of the epidemiological simulation.