Olive oil is a major economic resource of the Mediterranean region. Olive crop management can be improved by models that forecast the variable reproductive biology of olive tree. However, the processes controlling olive harvest are complex on large scales. Here, we study the parameters that influence olive fruit production for developing accurate forecasting models. Seventeen aerobiological sampling points have monitored olive pollen grains in Spain, Italy and Tunisia from 1993 to 2012. Six crop models have been developed at two provinces and country scales. The modelling has been done in two steps: (1) typification and (2) modelling by partial least square regression. Results show that higher pollen indexes and water availability during spring are related to an increase of final fruit production in all the studied area. Higher pollen indexes are also positively correlated with air temperature during early spring and autumn. Furthermore, a decrease of fruit production is related with increasing air temperature during winter and summer. To conclude, we have designed accurate models that allow accurate predictions of olive production.

Olive trees have a plentiful bloom but a low percentage of normal fruit set. To improve fruit set, numerous investigations have sought to identify the obstacles that prevent full production. In this work, flower development in five DOP Umbria cultivars (Leccino, Frantoio, Moraiolo, Dolce Agogia and San Felice) was studied throughout different developmental phases, from before microsporogenesis and megasporogenesis to post-anthesis, by morphological and cytohistological observations. Dolce Agogia was the most precocious cultivar, while full flowering was simultaneous in Leccino, Frantoio, Moraiolo and San Felice. Frantoio and Leccino were also good pollen producers, having the highest percentage of pollen viability and germinability. Dolce Agogia can also be considered a good pollen producer in terms of the high quantity of released pollen, but it had the lowest levels of pollen viability and germinability and the highest percentage of aborted flowers and ovaries. Morphological and cytohistological observations on the number of flowers per inflorescence and the number of aborted flowers and ovaries suggest that fruit set was not influenced by the number of flowers per inflorescence, but rather by the number of inflorescences, which depends on the global fruiting potential of the tree.

The aim of the present study was to develop pheno-meteorological models to explain and forecast the main olive flowering phenological phases within the Mediterranean basin, across a latitudinal and longitudinal gradient that includes Tunisia, Spain, and Italy. To analyze the aerobiological sampling points, study periods from 13 years (1999-2011) to 19 years (1993-2011) were used. The forecasting models were constructed using partial least-squares regression, considering both the flowering start and full-flowering dates as dependent variables. The percentages of variance explained by the full-flowering models (mean 84 %) were greater than those explained by the flowering start models (mean 77 %). Moreover, given the time lag from the North African areas to the central Mediterranean areas in the main olive flowering dates, the regional full-flowering predictive models are proposed as the most useful to improve the knowledge of the influence of climate on the olive tree floral phenology. The meteorological parameters related to the previous autumn and both the winter and the spring seasons, and above all the temperatures, regulate the reproductive phenology of olive trees in the Mediterranean area. The mean anticipation of flowering start and full flowering for the future period from 2081 to 2100 was estimated at 10 and 12 days, respectively. One question can be raised: Will the olive trees located in the warmest areas be northward displaced or will they be able to adapt their physiology in response to the higher temperatures? The present study can be considered as an approach to design more detailed future bioclimate research.

Airborne pollen concentration patterns reflect flowering phenology of a given species, and it may be a sensitive regional indicator in climate change studies. This paper presents the relationship between a strategic biological event, such as olive flowering, and the air temperature trend, registered over a large scale in the Umbria region. The aim of the study was to determine relationships between phenological behaviour (flowering) of olive trees and the air temperature trend in the Umbria region. The phenological data on flowering phase were registered indirectly through an aerobiological monitoring technique. The obtained results showed a strong relationship between phenology and thermal trend. This characteristic was confirmed from results of correlations between temperature (mean temperature from 1st March) and flowering dates, especially that of full flowering (r = -0.9297). Moreover, the results showed an advance trend of 6, 8 and 10 days, respectively of start, full and end of flowering dates. The advance of the recorded flowering time in this period is to ascribe mainly to the increase of mean temperature and above all to that registered in months of May and June.

Olives are one of the largest crops in the Mediterranean and in central and southern Italy. This work investigates the correlation of the Olea europaea L. pollen season in Perugia, the capital city of the region of Umbria in central Italy, with atmospheric parameters. The aim of the study is twofold. First, we study the correlation between the pollen season and the surface air temperature of the spring and late spring in Perugia. Second, the correlation between the pollen season and large-scale atmospheric patterns is investigated. The average surface temperature in the spring and late spring has a clear impact on the pollen season in Perugia. Years with higher average temperatures have an earlier onset of the pollen season. In particular, a 1 degrees C higher (lower) average surface temperature corresponds to an earlier (later) start of the pollen season of about 1 week. The correlation between the pollen season and large-scale atmospheric patterns of sea level pressure and 500-hPa geopotential height shows that the cyclonic activity in the Mediterranean is unequivocally tied to the pollen season in Perugia. A larger than average cyclonic activity in the Mediterranean Basin corresponds to a later than average pollen season. Larger than average cyclonic activity in Northern Europe and Siberia corresponds to an earlier than average pollen season. A possible explanation of this correlation, that needs further investigation to be proven, is given. These results can have a practical application by using the seasonal forecast of atmospheric general circulation models.