Structural Equation Modelling (SEM) applied to sensory profile of Vinho Verde monovarietal wines

Vinho Verde is the biggest DOC of Portugal, and it produces wines distinguished by their high acidity and flavor. In this Portuguese wine-region, three grape varieties, Alvarinho, Loureiro, and Arinto, are used to produce monovarietal wines of great quality. Sensory characterization of wines has the aim of detecting unique qualities in wines, and the most common measurement for the sensory aspects of wines sensory profile is descriptive sensory analysis, typically performed by trained sensory panels. To a list of attributes that can characterize a wine, we call a "sensory lexicon" that can be used by marketers allowing the articulation of flavor perceptions to consumers. So, the aim of this work was to recruit and train a sensory panel able to describe the sensory aspects associated with the monovarietal wines from Loureiro, Arinto and Alvarinho grape-varieties. Furthermore, a second order factor analysis statistical model was developed for each grape-variety wine, based on the data-set of sensory attributes, allowing a new perspective in the sensory characterization of these wines. We discover that there are some similarities between the sensory lexicons that can be used to better describe the wines. However, there are descriptors that specify each monovarietal wine. Loureiro wines present "Loureiro and mineral aromas"; Alvarinho wines are described by having "citrus" aroma and freshness in the mouth and Arinto wines possess a distinctive fruity aroma (peach and pineapple) with an alcoholic, acid and fresh flavor. Additionally, it was also shown that Structural Equation Modelling (SEM) is a worthy statistical tool to be used in sensory analysis data treatment.