Geographical Authentication of Aquilaria sinensis Using Integrated C and O Stable Isotope Analysis Coupled with Chemometric Profiling

Multivariate carbon and oxygen stable isotope analyses combined with chemometric methods were employed to investigate Aquilaria sinensis samples collected from six major regions in China (Honghe Hani and Yi Autonomous Prefecture and Xishuangbanna Dai Autonomous Prefecture in Yunnan Province; Zhongshan City and Maoming City in Guangdong Province; and Danzhou City and Chengmai County in Hainan Province). Isotopic δ-values were analyzed across different wood parts (longitudinal and north–south orientations), chemical fractions (de-extracted wood and α-cellulose), and geographical origins. Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), Decision Tree, and Random Forest were applied to screen and classify the samples. Four discriminant models were successfully established, achieving a maximum accuracy of 85.7% for distinguishing Aquilaria sinensis from different regions, and 88.1% for discrimination at the provincial level. These results demonstrate that stable isotope signatures, when combined with chemometrics, provide a reliable technical approach for the traceability of incense wood and offer a reference framework for verifying the authenticity of Agarwood and related plant-derived materials.