Analysis of the effects of uncertainties on agrometeorological models

We analyze how the uncertainties in meteorological data affect the outcome of an agrometeorological model. In particular we focus on a model simulating the development of primary infections of downy mildew on grapevine. This model takes as inputs hourly measurements of temperature, relative humidity and rainfall and it simulates all the phases that eventually lead to the downy mildew primary infections. In order to assess the robustness of the model against the uncertainties of the input data we set up a Monte Carlo simulation. This procedure enhances the potentiality of the model since allows to extract, as an additional valuable piece of information, the probability of the infection event. The results we obtain are highly significant and suggest that the effects of the uncertainties are actually important and should be carefully evaluated.