The data from the Napping test were digitized by writing in a table, for each product, its X-coordinate and its Y-coordinate on the sheet. The origin was placed on the left bottom corner of the sheet.The chemical and sensory data of the wines were analyzed by multivariate analysis of variance , with the factors as the wines and replicates, and frequency distributions were analyzed by the Chi-square test; all statistical analyses were completed using Statgraphics Centurion . Principal component analysis was performed using XLSTAT v. 2018.3 ; partial least square regression analysis was performed using Unscrambler . A multiple factor analysis in which each subject of the Napping® panel constitutes a group of two un-standardized variables was performed using XLSTAT. The typicality scores of the second table and the Color scores of the third table were considered as two sets of 11 + 11 supplementary variables: They do not intervene in the axes construction, but their correlation coefficients with the factors of MFA are calculated and represented as in a usual PCA. Descriptive Analysis data were analyzed by MANOVA, to check overall differences among the products for aroma, taste, and mouthfeel terms. Following a three-way ANOVA with the factors wine, judge, and replicate as well as their two-way interactions, Fisher’s LSD test was used to detect differences among wines for the separate attributes. In those cases, where the effect of the wine was significant, but one of the interaction terms included wine as a factor, a pseudo-mixed model was applied. Here, a new F-value was calculated with the mean sum of squares from the significant interaction as an error term for the factor wine. The significance level for all statistical tests was set to p < 0.05. The chemical eligibility profile of the wines was represented by the standard chemical parameters , color indices and polyphenol composition.
The chemical data were elaborated using the PCA, growing blueberries and the results were reported in Figure 1a,b. Figure 1a shows the observations on a two-dimensional map. Figure 1b shows the correlationcircle and the projections of the initial variables in the factors space. The first two dimensions accounted for 57.90% of the total variance. The first dimension separated the wines between Italy and California based largely on the polyphenol composition. According to the squared cosines of the variables for the two dimensions, it was possible to determine that on the first dimension , quercetin , gallic acid , myricetin , total phenols index , color intensity , polymeric phenols , hue , residual sugar , volatile acidity , and pH were the variables well linked to this axis. On the second dimension , samples were separated by anthocyanin , delfinidin-3-O-glucoside , peonidin-3-O-glucoside , malvidin-3-O-glucoside, and quercetin-3-O-galactoside composition . According to the importance of the above mentioned variables, the Italian wines on the right side of the plot were characterized by the polymeric phenols, monomeric anthocyanins , color intensity, and total phenols index. The Californian wines on the left side of the plot were instead characterized by hue, residual sugar, pH, volatile acidity, malvidin-3-O-glucoside, and quercetin. The two ellipses defined the interval of confidence and helped to evidence a better separation between wines in the two regions. The chemical identity profile of the Sangiovese wines was represented by the volatile compounds originating in the grape and by the alcoholic and malolactic fermentations . The data were elaborated using the PCA, and the results were reported in Figure 2a,b. Figure 2a shows the observations on a two-dimensional map. According to the squared cosines for the variables on the two axes , the volatiles that were well linked to the first dimension were ethyl octanoate , octanoic acid , ethyl butanoate , ethyl decanoate , isoamyl acetate , 4-terpineol , and β-phenylethanol , while compounds on the second dimension were β-phenethyl acetate , TDN , 3-methylbutan-1-ol , isoamylbutanoate , and β-damascenone . Wines were separated according to the region of origin along the first dimension. In particular, most of the Californian wines were located on the right side of the plot described by volatile compounds such as esters , acetates and fatty acids .
On the left side of the plot, the Italian wines were characterized mostly by varietal volatile compounds such as terpenes β-citronellol, β-linalool, 4-terpineol, α-terpineol and norisoprenoids . However, the wines from the two different origins were not completely separated by the identity profile as evidenced by the overlapping confidence intervals of the two ellipses . Using Descriptive Analysis, the panel of trained judges described the sensory attributes of the Italian and Californian wines. Figure 3a,b shows the distribution of the wines according to the sensory eligibility and identity descriptors . Only significant descriptors were used for the PCA, and 79.30% of the total variance was explained by the first two factors/dimensions. One of the main objectives of this study was to examine intrinsic quality of the samples, evaluating how chemical differences in wines from Italy and California, in terms of eligibility and identity profiles , could reflect on the sensory perception of wines. In this context, the chemical eligibility profile of the wines was represented by the standard chemical parameters , color indices and polyphenols composition, while the chemical identity profile was represented by the volatile compounds originating in the grape and by the alcoholic and malolactic fermentations . The experimental data showed that the Sangiovese wines from Italy and California resulted in differences mostly for chemical eligibility profile. In particular, it was very evident that the Italian and Californian wines differed in their color indices and polyphenol composition . In fact, the Italian wines were higher in polyphenols compounds and in color intensity. These results were in agreement with the chemical characterization of Sangiovese wines from Italy and California for the 2016 harvest where the Italian wines resulted in higher color intensity and total phenols index compared to the Californian ones, that showed instead a higher hue. The values for these indices were consistent with other findings for Sangiovese wines. The differences in color indices were better explained by the polyphenols compounds of the wines that resulted in higher amounts for the Italian wines. They showed higher amounts of pigmented polymers and monomer anthocyanins . Sangiovese red grape is considered a variety with a neutral aroma since the total amount of terpenes is lower than 1 mg/L, and this variety is not dependent upon monoterpenes for its varietal flavor. In the Sangiovese pulp and skin some norisoprenoids precursors such as TDN, riesling acetale, damascenone, and vitispirane are detected. The above varietal volatile compounds were determined in both Italian and Californian wines indicating that the varietal aspects of the Sangiovese grape were maintained in both regions. Important differences were instead evidenced in wines from both regions according to the fermentative volatile compounds. The Californian wines were richer in the composition of these fermentative volatiles than the Italian wines confirming the trend observed for the same regions for the 2016 vintage. Significant differences for volatiles between the two regions were reported in Table S2 as supplementary material.
Based on the chemical differences in the composition of the wines from the two regions, we further explore how intrinsic quality, in terms of chemical differences, could be reflected on eligibility and identity sensory profiles of the wines. Moving to the perceived quality, square plant pots the second target of the study was to see how Tuscan wine experts perceived the peculiarity/typicality of the Sangiovese wines from Italy and California and to link the sensory descriptors that might be associated with the wines’ typicality. The sensory profiles of Californian and Italian samples were separated from each other, and the analysis of the correlation with the descriptive attributes allows interpreting this separation in terms of differences of both eligibility and identity profiles. The Californian samples were correlated to the identity attributes Bell Pepper, Cherry, Red Berries, Citrus, Honey. The attributes Barnyard, Earthy and Rubber were correlated to 4I and 5I wines, while there was not any correlation with the Californian wines. Given that all the samples were checked for the defects before all the sensory tests, this contrast seemed to describe a freshness range, for which varietal aromas were perceivable in some samples , while in other they were hidden by some typical aromas of a full-developed wine . The eligibility profile underlines this separation, with the samples on the left side of Figure 3a correlated to Sweetness that, even if associated to Alcohol and Burning/Hot, elicits a softer sensation respect to the Astringency and Sour of the Italian ones on the right side. The perceived quality was studied by the Napping and typicality test. The Napping test results showed that the Californian and Italian wines were clearly separated, evidencing that the two kind of wines were perceived differently for the gustative and olfactive characteristics. Despite of that, the results of the typicality evaluation did not show the same clear discrimination between the wines from the two regions: The average score of all the wines were very similar with the Californian wines slightly higher but not significantly different . Figure 4 showed the correlation between the distribution of the wine samples according to Napping X- and Y-coordinates: the subjects were positioned in the second, third and fourth quadrants of the graphic with a concentration of the higher scores in correspondence of the position of two Californian samples and one Italian . This result can be explained by the fact that, even if the expert subjects perceived differences among the wines, they did not associate them uniquely to typicality. Given the extensive training and experience of the experts, the lack of agreement among them can be interpreted not only as a variability of their opinion but also as an indication that from the point of view of perceived quality in terms of typicality of the Sangiovese wine, the experts viewed all of the wines as falling within the identity profile. At the same time, the distribution of the higher average scores denotes that typicality has been correlated to fruity and floral attributes, in opposition to Bell pepper, Barnyard, Rubber, and Earthy descriptors. In other words, the typicality of Sangiovese has been connected to the perception of the varietal characteristics that in this wine were related overall to fruity and secondly to floral.These findings were evidenced by the PLS prediction of the typicality by the identity sensory attributes such as Cherry and Red Berries. In the case of color evaluation, the experts more clearly separated the wines and overall the Italian wines had significantly higher scores. In fact, these samples, reflecting the chemical parameters, had a more intense color and overall a lower hue compared to the Californian ones. These results showed that the Sangiovese variety is recognizable even if grown abroad, very far from the original terroir of Italy and in particular in Tuscany. This is supported by the fact that the varietal volatiles were found in both wines from both countries, even if the Californian wines were more intense in fermentative volatiles than Italian wines were. Despite this, the main differences seemed related more to the intrinsic quality in terms of eligibility chemical and sensory profiles. Important and significant differences were found in wines for the polyphenol composition since Italian wines were higher in color intensity, tannins, monomeric anthocyanins, and pigmented polymers content. Consequently, they were perceived more intense in color and astringency. On the other hand, Californian wines were higher in alcohol content and pH and lower in titratable acidity compared to the Italian wines. These results reflected the eligibility sensorial perception of the wines in which the Italian wines tend to be more acidic, less sweet, and more astringent than their Californian counterparts. These results evidenced that the terroir seemed to influence the eligibility characteristics of the Sangiovese grape variety, in particular for the polyphenol composition. In Italy, wines with a designation of origin are subject to production requirements that dictate many aspects of wine production such as the maximum grape yields, alcohol level, irrigation, and other quality factors, before an appellation name may legally appear on a wine bottle label. In general, the US has the highest national average yields, at 6.5 tons/acre , and the only requirement to use the AVA name on the wine label is that 85% of the wine must have come from grapes grown within the geographical AVA boundaries.