Picchi M1, Recchia A1, Bertuccioli M1, Tuorila H2, Naes T3, Monteleone E1.
1 Dipartimento di Biotecnologie Agrarie, University of Florence, Italy
2 Department of Food Technology, University of Helsinki, Finland
3 Nofima, Matforsk, Oslo, Norway
Sensory evaluation conducted with experts (wine professionals, enologists) is largely applied in wine industry. The projective map method (also known as “napping”) has recently been proposed as a suitable method for wine experts because of its spontaneous aspect and its flexibility.
In the present study, the method was used to establish to what extend aromatic properties of Sangiovese wine can be discriminated from those of other Italian wines.
The aim of the paper was to evaluate the suitability of the method for expert evaluation of wines and compare it with regular sensory profiling and rating techniques for typicality and quality. Attention was given both to aspects related to conducting the test and to the type and amount of information that the napping method gives.
Eleven wines (6 Sangiovese and 5 other wine styles) were selected and evaluated by both a trained panel and an expert panel. The trained panel was composed of 9 subjects. Wines were described in three replicates in relation to 8 odour attributes, using a 9 point category scale. Thirteen experts performed the “mapping” based on positioning 12 samples (11 plus one replicate Sangiovese sample) on a 60 x 40 cm white paper sheet. In two further sessions, experts were also asked to rate wine quality and typicality using a 9 point category scale. All the tests were based on the ortho-nasal odour evaluation of samples.
Data from conventional profile were submitted to a PCA and resulting score and loading plots used as a benchmark for validating the projective map method. The sensory space appears to be basically two-dimensional (as determined by cross-validation) with a clear separation of the 11 wines.
The data from the napping technique were first analysed by generalised procrustes analysis (GPA) in order to reduce the individual effects related to using the directions/dimensions on the white paper sheet differently. Quite large individual differences were found even after rotation. Then the consensus configuration was analysed by PCA and the two dimensions plotted. The first dimension had strong resemblance with the second axis from the PCA of the sensory data. Relating the two GPA dimensions to the sensory data was attempted. Strong relations were found for the first axis, but no significant relation was found for the second, indicating that the experts focus primarily on one dimension when using the napping method. The average typicality and quality scores were very comparable except for one of the samples. The relation between quality and the first component from the GPA was strong.
