Geochemical and spectroscopic fingerprinting for authentication and geographical traceability of high-quality lemon fruits.

Geochemical (mineral element and Sr isotope ratio) and spectroscopical fingerprinting (Near Infrared Spectroscopy) were proposed to authenticate and track the two high-quality lemon fruits from the Campania region (Limone di Sorrento PGI and Limone Costa d'Amalfi PGI) to protect them from frauds. Considering the geochemical indicators, we built different chemometric discriminant models based on mineral profile and 87Sr/86Sr isotope ratio. These two techniques were applied to discriminate fruits from different territorial scales, small territorial scales (region scale), and large territorial scales. The results of different discriminant models applied on mineral profiles of lemon juices, both on a small and large territorially scale, showed good discrimination according to provenance, especially for non-essential elements as Rb, Ba, Sr, Ti, and Co. These same elements have shown a good correlation with cultivation soils and stability between the two production years. It is worth noting that although, the performance of the whole elemental profile gave a better result than the profile of the non-essential elements, the reliability of the two models, calculated as the ratio between the percentage of correctly validated and classification samples, was similar. In addition, the Sr isotope ratio had shown a clear differentiation among the fruits from the Campania region and extra-regional samples, and by analysis of 86Sr/87Sr of soils, it was clear that the strontium isotope ratio of lemon juices was closely related to that of the bioavailable fractions of the soil. Furthermore, combining both isotopic and mineral profiles in lemon juices by a low-level data fusion approach, the results showed a better clustering according to geographical origins than the two-determination taken separately, although on an explorative level. In addition, the spectroscopical data (NIR) on intact lemon fruits showed the strong influence of environmental growing conditions on the samples. For this, the application of Linear Discriminant Analysis (LDA) models suggested building the discrimination models according to origins (PGI and not PGI productions) based on one production year. In the same way, the application of MLR models, that showed a strong relationship between quality properties of lemon fruits and NIR spectra, suggested the applicability of this technique to build predictive models for the quality properties. In addition, on a part of the total samples collected only in 2019 (intact lemons and juices), have been successfully applied two different chemometrics models i.e., LDA and Partial Least Square Discriminant Analysis (PLS-DA). The results showed better provenance discrimination using the lemon juices than the intact lemons. Comparing the results obtained, of the two approaches used, the results of geochemical fingerprinting have shown more stability for discriminate lemon fruits derived from two different production years, especially for not essential elements. However, considering the various vantages of the application of NIR spectroscopy (non-destructive, rapid, and cheap) and the results obtained, this technique can be used for rapid screening of samples in order to verify the quality and origins of lemon fruits during the year. The study of the pedoclimatic features was fundamental to understand the nature of discriminating variables, in both approaches. Additional research should be conducted to include a greater number of lemon farms (or sampling points) in the PGI area and to enlarge the existing database including lemon samples from other regions and validate the models built. These discriminant models based on geochemical and spectroscopical profiles of lemon fruits could substantially contribute to implementing a blockchain system for Campanian lemon traceability, providing real-time information not only to the final consumers but also to manufacturers, distributors, and retailers