Forecasting olive crop yields based on long-term aerobiological data series and bioclimatic conditions for the southern Iberian Peninsula

In the present study, bio-meteorological models for predicting olive-crop production in the southern Iberian Peninsula were developed. These covered a 16-year period: 1994-2009. The forecasting models were constructed using the partial least-squares regression method, taking the annual olive yield as the dependent variable, and both aerobiological and meteorological parameters as the independent variables. Two regression models were built for the prediction of crop production prior to the final harvest at two different times of the year: July and November. The percentage variance explained by the models was between 83% and 93%. Through these forecasting models, the main factors that influence olive-crop yield were identified. Pollen index and accumulated precipitation, especially as rain recorded during the pre-flowering months, were the most important parameters for providing an explanation of fluctuations in fruit production. The temperature recorded during the two months preceding budburst was another important variable, which showed positive effects on the final yield. The July model that provides accurate predictions of fruit production eight months prior to the final harvest is proposed as an optimal model to forecast fruit produced by olive trees in western Mediterranean areas.