Application of Statistical Methods for Classification of Varietal and Regional Origin of White Wines

Abstract

White wines of four wine-growing regions of Krasnodar region were studied by the ICP-AES method. Nineteen elements were identified in 153 samples of white wines of the Chardonnay, Riesling, and Muscat brands. Methods of statistical analysis demonstrated that, despite the complicated technological cycle of wine production (soil–grapes–juice–wine), information about the grape variety and the region of its origin is preserved in the totality of trace elements. Probabilistic statistical models were constructed to identify the grape variety and the region where it was grown by the concentrations of a certain set of microelements contained in wine. The proposed models are based on the assumption that the soil origin is an important factor determining the microelement composition in wine.