The relationship between grape and wine phenolics is of key interest for the wine industry to predict wine quality from analysis of grapes. Research suggests that the physical properties of skin tissue are related to the composition and extractability of compounds with importance to wine quality. What is lacking is an understanding of the relationship between grape texture, extractability of grape components, and wine quality under real winemaking conditions. Although analytical texture methodologies are in place, routine analytical procedures to predict wine quality are still unavailable. To make such assessment more applicable in a winery setting, prediction of the anthocyanin composition and color of red wines from the detailed anthocyanin composition and skin texture was investigated using a multiple regression model. Grape extracts and wines were produced from two Vitis vinifera L. varieties, Merlot and Cabernet Sauvignon, grown in different regions in California, from warm interior valley sites to cool north coast sites. A linear combination of three independent variables could accurately predict the total concentration of anthocyanins in wines. Total skin anthocyanins, skin hardness, and skin weight were significant predictors of wine color, and the model developed was independent of region and cultivar. With this model, winemakers could predict anthocyanin extractability and adapt the process accordingly.

Funding Support: The California State University Agricultural Research Institute (ARI), E & J Gallo Winery