Development of tools to verify origin and varietal nature of wines
Origin verification of wines relies on the use of parameters that not only reflect the geology and water source of the location where grapes are grown, but which are also not altered during vinification. Building on a recently completed AWRI feasibility study, several additional elemental isotope ratios such as boron, lithium and lead are now being measured as part of the Wine Australia project AWR1505, in order to provide information additional to the strontium isotope ratio, on the crustal composition of source vineyard location.
To supplement this dataset of trace elements and their isotopes, and to achieve true independence of multiple data sources, oxygen-18 will also be analysed as an indication of the vineyard water source. These parameters will be measured in a large number of Australian wines, and wines from North and South America, Europe, and mainland Asia and Africa.
In order to verify that chosen parameters are not affected by variety (as has been highlighted in several recent papers in the literature), laboratory-scale wines will be made to assess numerous varieties grown in the same soil across several different locations. In addition, the temporal variation of these parameters will be assessed through the analysis of several vertical series of wines made from single vineyards over the last ten years. This information will be combined with the continued survey of the parameters determined to be most robust over a five-year period, to provide a series of reliable analytical protocols and metrics for the determination of wine origin.
In the later years of the project, a review of chemical and genetic techniques for the determination of varietal origin of wine will be undertaken, with the aim of targeting appropriate tools for this aspect of wine authentication.
Is it Australian?
The focus of the most recent work has been on expanding the data set of Australian wines to include as many regions as possible and increasing the range of isotopic ratios for each wine to include those for boron, oxygen, lithium and lead. These extra isotopic ratios were chosen as literature had suggested that they had some potential to be related to the provenance of wine. The final data set was expanded to include 292 wines from around Australia and more than 90 wines from other countries.
Figure 28 shows that the boron, oxygen and strontium isotopic ratios have an inherent ability to separate Australian wines from overseas wines.
A principal component analysis using all available isotopic ratios, however, led to differentiation of red and white wines (Figure 27). This was undesirable as it suggested that the differences in winemaking process between red and white wines were introducing differences in the isotopic ratios, masking the signals from the grape origins. Further study of these effects showed that the main driver of the differentiation was the lead isotopic ratios, which could be traced back to lead introduced into white wines by bentonite fining for protein stabilisation. This result ruled out further use of lead isotopic ratios in the study.
In a similar vein it was decided to investigate if there was any possible contribution of boron from the glass in the wine bottles to the isotopic variation for this element found in wine. Flint and green glass bottles were sourced from different countries and wine was stored in them to elucidate any impacts. No significant differences were seen in the wine stored in bottles from different continents, suggesting that the bottle source had little impact.
Using the strontium, lithium, boron and oxygen isotopic ratios with a statistical technique called orthogonal projection of latent structures-discriminant analysis (OPLS-DA) it was possible to differentiate Australian wines from those produced overseas without any contribution from wine processing factors (Figure 28).
Using similar approaches, it was also possible to classify Australian wine (regardless of grape colour) based on subsoil types, although detailed discrimination of regionality would likely require detailed studies of more wines from each region and perhaps the inclusion of additional data such as trace metal concentrations.