The use of monitoring network for estimating early infestation of Bactrocera oleae at large scale

Abstract: Bactrocera oleae (Rossi) population dynamics strongly depend, like other pest, onweather pattern, in particular air temperature. In Tuscany (central Italy), the regionalextension service has set a monitoring network at farm level to follow the B. oleae infestationon olive fruits from July to October. Here, we present a synthesis based on the last 14 years ofmonitoring activity. Starting from the agrometeorological network database, we havecalculated indices based on temperature and precipitation to describe the weather patternduring three periods of the year. These periods were selected according to the B. oleae annualcycle. We implemented a prediction model relating the percentage of infestation observed atthe field level to selected agrometeorological indices. To choose representative indices andavoid collinearity, we applied Principal Component Analysis (PCA). Only temperature-basedindices were statistically significant in predicting B. oleae infestation, in particular minimumand average temperature of winter preceding the summer infestation. Warmer winterssustained high infestation level in the following summer. Temperatures of the previous winterand spring explained 66% of variance of early-season infestation. Although a correctapplication of integrated pest management requires long-term monitoring of B. oleaeinfestation at the field level, the development of predictive models can provide early-warningsignals of severe outbreaks of this pest and a well-timed set of control strategies.