Authenticating vintage in white tea: Appearance-taste-aroma-based three-in-one non-invasive anticipation

To safeguard the legal rights of tea enterprises and promote sustainable development in the tea industry, this study proposes a rapid, non-destructive method for authenticating white tea vintages based on the hypothesis that the appearance, taste and aroma cannot be simultaneously replicated in counterfeit teas. Using visible-near infrared hyperspectral imaging, this three-in-one appearance-taste-aroma method was applied to Bai Mudan white tea, produced from the Jinggu Dabai Tea cultivar harvested in 2020, 2021 and 2022. Hyperspectral imaging captured appearance data from dry samples of different vintages, with preprocessing using multiplicative scatter correction (MSC) and standard normal variate (SNV). Partial least squares regression (PLSR) and support vector regression (SVR) models were used to explore correlations between appearance data, electronic tongue-measured taste and electronic nose-measured aroma. The results showed that appearance data can predict tea infusion taste (0.6540 < R_p < 0.8873) and aroma (0.8880 < R_p < 0.9703) across vintages. Further integration of high-performance liquid chromatography (HPLC), high-performance liquid chromatography (GC-IMS) and regression models revealed that appearance-based spectral data predict taste through gallic acid (GA), catechin (C) and gallocatechin gallate (GCG), and predict aroma via styrene, 2,5-dimethylpyrazine and 2-octanone. This non-invasive method, leveraging visible-near infrared spectroscopy, provides a standardized approach for white tea vintage authentication by integrating appearance, taste and aroma assessments.