Abstract: The aroma attributes of nine ‘Cabernet Sauvignon’ red wines from Chinese main wine producing areas were evaluated by quantitative descriptive analysis on a 5-point scale using a trained sensory panel. Through analysis of variance (ANOVA) and multivariate statistical analysis, the proficiency of the sensory panel was examined, and then a systematic method for screening for the characteristic aroma descriptors of red wine was proposed. The results showed that the reproducibility index (Ri), the F value of the test statistics, the mean-square error (MSE), and the Manhattan plots were effective in evaluating the reproducibility, distinguishing ability and repeatability of each assessor, as well as the consistency within the sensory panel. Using the reliable sensory panel data as raw variables, geometric mean (M) values for each aromatic descriptor were calculated and ranked, and principal component analysis (PCA) and cluster analysis (CA) were performed on the data from the final sensory panel to establish a normative and scientific approach to screening for the characteristic aroma descriptors of wine. Furthermore, it was determined that the major aromatic attributes of ‘Cabernet Sauvignon’ wine were black currant-like and red currant-like fruity notes as well as prominent roasted notes such as caramel and smoky, and spicy notes such as pepper-like.

Key words: wine; quantitative descriptive analysis; sensory panel; aromatic descriptor; multivariate statistical analysis