Investigation On Characteristic Sensory Descriptors Of Cabernet Sauvignon Red Wine By Multivariate Statistical Analysis

Sensory analysis is of the most importance in the evaluation of wine quality,and depends mainly on the reliable performance of tasters and the group,which greatly limits the accuracy and authority of sensory results.In this study,Cabernet Sauvignon wines were tasted and evaluated by a sensory panel.As a result,the reliable indexes of evaluating the performance of tasters and the panel was clarified by a Panel Check software;by using multivariate statistical analysis,a method of screening the sensory characteristic descriptors of wine was standardized,and the sensory characteristics of domestic bottled Cabernet Sauvignon wines were primarily determined.The main results obtained are as follows:(1)Selection of wine aromatic descriptors based on reliability evaluation of a sensory panelThe aroma attributes of nine domestic Cabernet Sauvignon wines were evaluated with quantitative descriptive analysis(6-point scale)by a trained sensory panel.Using variance analysis and multivariate statistical analysis for the sensory data,the proficiency of a sensory panel was first performed,and then a systematic method of screening characteristic aroma descriptors of a wine were proposed.The results showed that the reproducibility index(Ri),F value of the test statistic,MSE value of the mean square error,and such the plots as Manhattan,could effectively evaluate the reproducibility,distinguishing ability and repeatability of each assessor,as well as the consistency of the sensory panel.In accordance with the ranked M values calculated for each aromatic descriptor,principal component analysis(PCA)and cluster analysis(CA)were used to process the data obtained by the final sensory panel,and a standardized method of scientifically screening characteristic aroma descriptors of a wine was confirmed.(2)Primary identification of sensory characteristics of domestic bottled Cabernet Sauvignon winesQuantitative descriptive analysis(0-10 cm,unstructured scal)was used to evaluate the aroma,taste and mouthfeel sensations of 24 domestic bottled Cabernet Sauvignon wines by a reliable tasting panel.With the help of multivariate statistical analysis,the preliminary sensory characteristics were identified.The characteristic aroma were classified as:(1)black currant or jam,green pepper;(2)floral,mint or oxidized taste,boiled taste;(3)oak,licorice,roasted nuts.The overall taste is prominent with astringency and low sweetness.The established model for distinguishing origin and year of Cabernet Sauvignon achieves 95.8%and 100% judgment rates respectively,which further verifies the typicality of the selected sensory characteristic descriptors.In short,based on the sensory data obtained by the reliable tasting panel,using multivariate statistical analysis to screen the characteristic descriptors of wine is effective and objective.It provides a scientific method for further clarifying the quality and style of wines.