**1. Introduction**

Olive oil is an ancient food widely spread throughout the Roman Empire. Olive trees are now mainly diffused around the Mediterranean Sea and Italy—located in the center of this region—is one of the most important olive producers. The extra virgin olive oil (EVOO) produced in various Italian regions has an excellent quality and is greatly appreciated worldwide for its attributes, which depend, among other factors, on the geographical origin. The importance of the geographical origin for EVOO has been documented by the European Union since 1992, when the Protected Designation of Origin (PDO) and Protected Geographical Indication (PGI) arrangements were created to protect and support foodstuffs of particular quality [1]. Bianchi et al. (1993) and Angerosa et al. (1999) used Isotopic Ratios Spectrometry (IRMS) to find olive oil components and to characterize the geographical origin [2,3]. The traceability and authenticity of olive oils have been widely investigated by Nuclear

Magnetic Resonance (NMR) spectroscopy [4–6], Fourier transform infrared (FT-IR) [7] and mass spectrometry (MS) [6]. More recently a new regulation (EU 29/2012) [8] has affirmed that "*Extra virgin olive oil and virgin olive oil shall bear a designation of origin on the labelling*" and that "*the designation of origin shall appear on the packaging or on the label attached to the packaging*". This EU regulation is very important, because it recognizes that the characteristics of virgin olive oils are closely related to geographical origin and also with the agricultural practices and techniques used during the EVOO extraction procedure. Since the authenticity of foodstuff origin represents a very important issue for the clear definition of the concept of food quality [9,10], in recent years, various analytical techniques, in combination with multivariate statistical analysis (MVA) methods, have been applied for this purpose. Resonance intensities of triacylglycerols obtained by 13C NMR spectroscopy were used to classify olive oils from three production areas of the Puglia region. By applying linear discriminant analysis (LDA) and the leave-one-out cross-validation procedure, most of the oils used for this purpose were correctly assigned to their groups [11]. Different techniques, such as IRMS and 1H NMR, in combination with MVA, have successfully been used to discriminate Italian olive oils from Tunisian ones [12]. The authors declared that the high quality of Italian olive oils was essentially due to the high content of squalene and unsaturated fatty acids, when compared to the Tunisian samples. NMR spectroscopy is one of the most widely used analytical techniques, together with MS, and has shown grea<sup>t</sup> success in food analysis [10,13–15], as it is able to supply a complete view of the olive oil metabolic profile, giving qualitative and quantitative information on its compounds, particularly on minor ones. Among the minor components, phenolic compounds are very important, because they have been linked to the healthy properties of EVOO. Spanish EVOOs were analyzed by a 1H NMR standard pulse and by an experiment suppressing the main lipid signals, allowing the detection of minor component resonances. Usually, standard 1H NMR is used to determine high concentrations of oil components, whereas the multisuppression approach is useful for increasing the sensitivity of NMR, allowing the detection of minor components [16–18]. Nowadays, the whole 1H NMR metabolic profile of EVOO is considered strongly related to the geographic origin of the olive oil, as well as to its cultivar, maturation index, and/or technological factors [9,19]. In recent works on metabolomics, the combination of 1H NMR fingerprinting and multivariate analysis has been successfully applied to predict the geographic origin of olive oils from different Mediterranean regions [18,20,21]. In our work, the 1H NMR-based MVA approach has been used also to examine the historical meteorological parameters in order to obtain a model for olive oil geographical origin prediction. Most of the studies carried out in the past regarding the provenance of olive oils were performed with oils obtained from olive trees belonging to a single cultivar and produced in the same year. In the present study all oil samples were blends and were collected in different years. Therefore, the aim of this research is to apply a new approach able to better discriminate commercial olive oils from different geographic areas and to reduce the influence of the year [22]. The originality of this research is that we also considered climate parameters alongside the NMR and MVA techniques, with the final aim of developing a powerful tool for easily tracing Italian olive oil production.

#### **2. Materials and Methods**
