Combining Multi-Element Analysis with Statistical Modeling for Tracing the Origin of Green Coffee Beans from Amhara Region, Ethiopia

Abstract

Characterization of coffee terroirs is important to determine authenticity and provide confidence for consumers to select the right product. In this regard, Amhara Region, which is located at the northwestern part of Ethiopia, produces various local coffee types with distinct cup qualities. The coffees are, however, not yet registered with certification marks or trademarks for indications of their geographical origins. This study was aimed at developing analytical methodology useful to determine the geographical origin of green coffee beans produced in Amhara Region based on multi-element analysis combined with multivariate statistical techniques. For this, a total of 120 samples of green coffee beans, collected from four major cultivating zones (West Gojjam, East Gojjam, Awi, and Bahir Dar Especial Zones) were analyzed for K, Mg, Ca, Mn, Fe, Cu, Fe, Co, Ni, Zn, Si, Cr, Cd, and Pb using inductively coupled plasma-optical emission spectroscopy. The elemental analysis data were subjected to principal component analysis (PCA) and linear discriminant analysis (LDA). PCA was used to explore the natural groupings of samples and the discriminatory ability of elements. Accordingly, the elements K, Mg, Ca, and Na were found to be the main discriminators among samples. LDA provided a model to classify the coffee samples based on their production zones with an accuracy of 94.2% and prediction ability of 93.4%. Thus, the elemental composition of green coffee beans can be used as a chemical descriptor in the authentication of coffee produced in Amhara Region.

Keywords: Amhara Region; Geographical origin; Green coffee beans; Linear discriminate analysis; Metals; Principal component analysis.