Abstract
The analysis of the strontium isotope (87Sr/86Sr) ratio is recognized as a valid and promising tool to determine the origin of horticultural products, as its value is related to the local geo lithological features of the growing area. With this thesis, in-depth studies and investigations were conducted to describe unexplored aspects of the 87Sr/86Sr ratio in horticultural products, particularly apples, in relation to agricultural practices and soil features, with the aim to provide a valuable support to its use as geographical tracer. In the first study, the impact of multiple Sr inputs on the 87Sr/86Sr ratio of the soil-plant system was evaluated through two pot-experiments (oat seedlings and apple tree) (Chapter 3). From the overall results, it emerged that plants tend to progressively approximate the isotope composition of the soil where they were growing. In the oat experiment, the 87Sr/86Sr ratio of oat stems was significantly altered by exogenous Sr additions to the soil, proportionally to their Sr concentration and isotope fingerprint (Chapter 3.4.2). Even though it was not possible to detect a significant effect of irrigation water and fertilizers on the 87Sr/86Sr ratio of apple trees (Chapter 3.4.3), it is likely that prolonged applications alter the tree 87Sr/86Sr ratio and may loosen their correlation with the soil of the growing area. Apple trees in pots gradually adapted their 87Sr/86Sr ratio to that of the soil where they were transplanted and approximated the soil 87Sr/86Sr ratio after two vegetative seasons. This is due to a low uptake rate and low Sr mobility to the shoots. As a practical implication, this study suggests not to use young transplanted trees for traceability studies. A field investigation in two apple orchards allowed studying the variability of the 87Sr/86Sr ratio intra-part, intra-tree, and inter-tree (Chapter 4). The intra-part variability was evaluated through the standard deviations of the 87Sr/86Sr ratios of multiple sub-samples for each tree part (shoot axes, leaves, apple peels, and pulps). Interestingly, a lack of uniformity in the 87Sr/86Sr ratio among sub-samples, and hence among different tree branches, was observed. At the same time, in both orchards, the 87Sr/86Sr ratio was homogeneous intra-tree (comparison of the mean ratio of each tree part within the same tree) but significantly different among trees (comparison of the mean ratio of each tree part among different trees). The 87Sr/86Sr ratio in soil samples showed horizontal and vertical (10-80 cm of depth) 2 differences that might explain both the inter-tree and intra-part variability. In the last case, these results suggest a sectorial Sr transport from the roots to the shoots in apple trees. Differences in the soil 87Sr/86Sr ratios were more evident in one of the two orchards and highlighted the presence there of a greater geo-lithological complexity. To improve the results of traceability studies, preliminary considerations should precede samplings in such complex areas. Additionally, attention should be paid on data handling before generalizing the results. It could occur that plant species growing in areas with high 87Sr/86Sr ratio heterogeneity along the soil profile show differences in their Sr isotope composition, especially if these plants have fine lateral roots localized in different soil layers (e.g. arboreal vs herbaceous plants). Finally, due to the high homogeneity of the 87Sr/86Sr ratios among apple tree parts, the sampling can be performed indifferently on one of these parts without significantly affecting the results. In Chapter 5, the 87Sr/86Sr ratio was used to characterize the Golden Delicious apples collected from five regions in Northern Italy, and specifically to distinguish the cultivation districts of PDO and PGI apples from the others. Moreover, apple samples were used for a preliminary inter-laboratory comparison that showed a high reproducibility of the 87Sr/86Sr ratio analysis. The 87Sr/86Sr ratios of apples showed a good correlation with those of the respective soils and were consistent with the local geo-lithological data. Comparing the ratios of the different cultivation districts, PGI apples from Val Venosta had higher 87Sr/86Sr ratios than apples from some of the other production districts, but not always significant and a clear distinction among all the districts was not possible. The 87Sr/86Sr ratios were combined with the results of the multielement analysis to develop a classification model through the linear discriminant analysis (LDA). The best model was obtained grouping apples in four groups (PDO-Val di Non, PGI-South Tyrol, PGI-Valtellina and non-GI apples), reaching satisfactory preliminary accuracy (88% in calibration, 78% in prediction). The LDA model could be further improved increasing the sample size, especially for areas with high data variability, or including further variables. Numerous interesting aspects emerged from the results of this thesis and could serve as the basis for further investigations. Moreover, the results of the 87Sr/86Sr ratio analysis of apples provided for the first time information about several cultivation districts located in Northern Italy. The high potential of the 87Sr/86Sr ratio as geographical tracer was confirmed and could be enhanced by preliminary investigations that take into account the critical points discussed by this study (multiple Sr inputs, 87Sr/86Sr ratio variability, local geological features).