Origin Traceability And Year Identification Of White Tea

White tea is one of the six traditional tea categories in China,dry tea is full of pekoe,pure and persistent aroma,fresh and sweet taste,with the efficacy of clear the heat,loved by consumers.In recent years,the “white tea fever” has continued,making non-traditional white tea production areas begin to experiment with white tea,driving the development of the white tea industry.The flavor quality of white tea is easily affected by the difference of origin or aging,and the information on origin and aging is an important basis for consumers to evaluate white tea.In this study,the feasibility of mineral elements and near-infrared spectroscopy(NIRS)technology for the origin traceability of white peony tea was explored by using the first-grade white peony tea produced in different provinces and different producing districts of Fujian Province as materials;the sensory quality and physicochemical characteristics of first-grade white peony tea in different storage years(current year,3 years,6 years,9 years,12 years and 15 years)in Fuding district were analyzed,the feasibility of NIRS,quality-related component content and parameters for the identification of white tea years was explored.It aims to provide a reference for the construction of high-quality origin traceability models and aging year identification models for white tea,promote market normalization and standardization,and promote high-quality development of the industry.The main research results were as follows.1.Traceability models of white tea origin based on mineral element analysis technology were constructed.White tea in different producing regions has different characteristics of mineral elements,the content of some mineral elements in different producing areas has certain commonalities,but the content and abundance was often different.The recognition accuracy of the recognition models of different provinces of white tea and different producing districts in Fujian Province based on the content of mineral elements was higher than 85.00 %,and the area under the curve(AUC)was higher than 0.80,models had good prediction and generalization ability.It was confirmed that the mineral element analysis technology was suitable for the identification of the differences between the geographical large-scale producing areas(different provinces)and the small-scale producing areas(different producing districts in Fujian Province).2.Traceability models of white tea origin based on NIRS technology were constructed.Among the pre-processing algorithms,standard normal variate(SNV)performed the best;among the feature extraction algorithms,linear discriminant analysis(LDA)had obvious advantages;among the machine learning algorithms,random forest(RF)had the strongest generalization,and k-nearest neighbor(KNN)was more suitable for modeling the origin of white tea based on NIRS data;The best models constructed for identifying white tea producing provinces,white tea producing districts in Fujian Province and the authenticity of Fuding white tea all had recognition accuracy close to or greater than 90.00%,with AUCs close to or greater than 0.90,the models had excellent overall performance.It has been systematically and comprehensively confirmed that NIRS could be an effective tool for tracing the geographical origin of white tea.3.The pattern of the effect of storage time on the flavor quality of white tea was revealed.The white tea stored for 0,3,6,9,12 and 15 years was evaluated by quantitative descriptive analysis(QDA),and the quality parameters,biochemical indexes and volatile components of tea infusion were detected.The results showed that during storage process,its taste changed from clear,fresh and sweet to sweet,thick,smooth and sour;the aroma changed from sweet,floral and fragrance to stale,plum and medicinal aroma;in the quality parameters of tea brew,L(brightness)and p H value decreased with the increase of storage time,while a(red and green),b(yellow and blue)and conductivity showed the opposite trend;EGCG,flavonoids,CAF,GC,EGC,ECG and water extract were considered as the key taste substances affecting the taste in aged white tea;linalool,geraniol,hexahydropseudoionone,phenethyl alcohol,methyl salicylate,phenylacetaldehyde,palmitic acid ethyl ester,benzyl alcohol,palmitic acid,α-terpineol,3,5-octadien-2-one,cedrol,9-hexadecenoic acid ethyl ester,nonanoic acid,2-bornene and ethyl salicylate were considered as the key volatile components affecting the aroma presentation of aged white tea.4.White tea vintage identification models were constructed based on NIRS,tea liquor quality parameters,flavoring substances and volatile components data.All the models constructed for the identification of white tea storage years had recognition accuracy greater than 80.00%,with AUC≥ 0.90,the models had good prediction ability and strong generalization ability,which indicated that the different types of data used were suitable for the identification of white tea storage years,and could achieve rapid and accurate identification of white tea storage years.