Abstract
This study determined the content in 39 trace and ultratrace elements, among them 16 rare earth elements, in 153 wines from the Canary Islands (Spain). Analyses were performed by ICP-MS. Rare earth elements were strongly correlated among one another, red wines mean contents in these elements being lower than those of white and rosé wines, and they are not suitable to classify wines according to the island of origin. Contrarily, good recognition and prediction abilities were achieved with the other 23 elements in both linear discriminant analysis (96.7 and 95.4%, respectively) and back-propagation artificial neural network (100 and 99.3%, respectively) to classify wines according to the island of origin. The most discriminant elements were established.
Acknowledgments
This project was financially supported by the French Ministry of Economy and FEDER grant No. 13 02 01 Z0098. Dr. Pérez-Trujillo was supported by the Cabildo Insular of Tenerife.