Accurate discrimination of tea from multiple geographical regions by combining multi-elements with multivariate statistical analysis

Minglu Zhang,Congwei Huang,Jiangyang Zhang,Haoran Qin,Guicen Ma,Xin Liu,Jie Yin
DOI: https://doi.org/10.1007/s11694-020-00575-1
2020-01-01
Abstract:This study focused on building accurate and robust statistical models for geographical discrimination of teas, especially for the application on classification of the core geographical region teas from the non-core regions in tea. In order to achieve this goal, the content of mineral elements of tea samples was investigated. To be specific, taking Duyun maojian (DYMJ) tea as a sample, we applied inductively coupled plasma mass spectrometry (ICP-MS) combined with multivariate statistical analysis to profile the mineral elements of DYMJ collected from five adjacent production areas. Results demonstrated that a total of 39 elements were significantly different among the 5 regions, and in the stepwise linear discriminant analysis model (S-LDA), a satisfying prediction ability (88.3%) was obtained by cross-validation for the whole 5 regions, in which the correct classification rates for the geographical region (Duyun county) teas could reach 96.0%. The overall results suggested that the combination of ICP-MS with S-LDA could be successfully used as a rapid and reliable method for geographical origin identification of teas. At the end, the benefit of the essential elements supplement and also the risk of the toxic elements excess have been assessed during people’s daily consumption of tea.
What problem does this paper attempt to address?