CONICET | Buscador de Institutos y Recursos Humanos

Traceability of wines requires a deep knowledge of their characteristics, which are associated with the region of origin of grapes, its soil, water, climate as well as the winery techniques used. The regional name of the product plays an important role, especially for wines and other alcoholic drinks because it allows developing a particular standard of quality, which is appreciated by consumers (i.e. Scotch whisky, French champagne, etc.). The mineral profile present in a wine seems to be very sensitive to the type of grape used and its region of cultivation, winery techniques, etc. Therefore, this profile could be used as a chemical descriptor to classify wines in according to its region of origin. However, assessing the geographical origin of wine by this method requires verifying to which extent the mineral profile of a given wine is related to the soil where grapes were cultivated. The main goal of this work was to evaluate several minerals in both wines and soil from three producing areas of Argentina (Córdoba, La Rioja and San Juan) by Atomic Absorption Spectroscopy (AA), pointing out those minerals that enable differentiation among wines and soils from these regions, and also evaluating the correspondence between wines and soil profiles. Eleven minerals were determined (K, Na, Mg, Ca, Fe, Mn, Zn, Cu, Cr, Co, Ni) in 41 wine samples and 63 soil samples from 3 producing regions of Argentina: Cordoba (12 wines and 14 soils), La Rioja (18 wines and 12 soils) and San Juan (11 wines samples and 37 soils samples). We analyzed three different portions of each sample by triplicate. Separate data matrices were constructed for both soil and wine. These matrices were treated by multivariate analysis looking to classify wines and soils according to its geographical origin. The method is based on the assumption that the provenance soil is an important contributor to the element composition of wine. Stepwise discriminant analysis allows us to differentiate soils from these three regions with 92.9% accuracy using only 8 parameters (Ca, Ni, K, Na, Cr, Fe, Cu, Mg), while we obtained 100% right classification for wines using 9 variables (Ni, Cr, Mg, Fe, K, Ca, Zn, Cu, Na). Thus, it is possible to get good differentiation of both soils and wines from studied areas from their mineral profile. The next step was to establish if the characterizations of soil and wine presented significant relationship. For that we perform a modified canonical correlation analysis, showing up the origin wine as a classification variable. This analysis showed a canonical correlation coefficient of 0.85 (p-value<0.001) between both datasets. Thus, almost 85% of variability observed among wines from different areas could be attributed to the original soil, whereas the remainder 15% could be attributable to the climate, water quality, altitude, winery practices, etc. We conclude that the analysis of mineral profile in both wine and soil, along with the use of multivariate statistical techniques (chemometrics), provide with a powerful tool for geographical classification and characterization of wines, contributing to the traceability of this product.

enviar mensaje