西北地区干红葡萄酒质量相关理化指标的判别功能解析

为解析西北地区干红葡萄酒的质量相关理化指标的溯源判别功能,以新疆、宁夏和内蒙古乌海产区共27款干红葡萄酒为材料,基于分光光度法分析检测与色泽、味感和香气质量相关的花色苷、单宁、酒石酸酯等指标,且进行香气特征的感官量化分析。统计分析结果表明,CIELab色空间和花色苷等色泽指标中前两个主成分占总体方差71.08%,判别干红葡萄酒年份的效果明显,但对不同葡萄品种和产地的判别效果不佳。单宁、多酚等味感相关理化指标的前两个主成分占总体方差的65.93%,对干红葡萄酒产地和年份的判别效果较好,但不能有效判别品种。酒石酸酯、总酸和香气特征等指标中前两个主成分占总体方差的83.10%,表现出较好的产地和品种的判别效果,而对年份的判别效果一般。色泽、味感和香气相关的总计20个理化指标数据在主成分分析中表现出更好的产地、品种和年份的判别效果,前两个主成分占总体方差的67.32%。聚类分析能将葡萄酒产地准确地分为3类,准确率超过90 %。基于分光光度法开发的色泽-风味理化指标的大数据具有判别我国西北地区干红葡萄酒产地、品种和年份的潜在应用前景。

Discriminant Analysis of Physicochemical Indexes Related to Quality of Dry Red Wines from Northwest China

The discriminant functions of physiochemical indexes related to the quality of dry red wines from Northwest China were evaluated. In this study, 27 dry red wines originated in Xinjiang, Ningxia and Wuhai regions of Inner Mongolia were selected. The physiochemical indexes related to color, taste and aroma quality such as the anthocyanins, tannins, and tartaric acid esters were analyzed using spectrophotometric method, and aroma attributes were also quantified by quantitative sensory analysis. Statistical analysis of the data showed that the first two principal components (PC) of the color indexes, such as CIELab parameters and anthocyanins, accounted for 71.08% of the total variance in principal component analysis (PCA), which had good effect on identifying wine vintage, while had no obvious effect on identifying variety and producing area. The first two PCs of the taste indexes, such as tannin and polyphenols, accounted for 65.93% of the total variance, which had good effect on identifying producing area and vintage, while had no effect on variety. The first two PCs of aroma indices, such as tartaric esters, acid and aroma attributes, accounted for 83.10% of the total variance, showing a good effect on identifying producing area and variety, while had no effect on vintage. A total of 20 physicochemical indexes related to color, taste and aroma showed better effect on identifying wine region, variety and vintage in PCA, the first two PCs accounted for 67.32% of the total variance. Cluster analysis could classify wine regions into 3 kinds with an accuracy rate over 90%. The big data of color-flavor physicochemical indexes based on spectrophotometry had potential application prospects for identifying the region, variety and vintage of dry red wine in Northwest China.