WINETECH Technical Yearbook 2020

wines (spiking level zero) allowed for some degree of differentiation. Cabernet Sauvignon had pronounced characteristics of ‘berry jam’, ‘savoury’, ‘soy’, ‘caramel’ and ‘balsamic’. Merlot and Shiraz were quite similar, both sharing ’dark berries’ and ‘prune’ as descriptors; but while the former had ‘caramel’, the latter had wood- related characteristics expressed as ‘oak’ and ‘vanilla’. Lastly, Pinotage in addition to ‘dark berries’ and ‘oak’ (same as Shiraz) had some reductive notes labelled as ‘cooked vegetables’. Once the levels of each thiol were increased, the cultivars lost their individuality and moved to more fruity notes (at medium level) described as berries-like and sweet- associated, to ‘earthy’, ‘green’ and most importantly ‘blackcurrant’ at the highest concentration. Looking at the large overall number of attributes generated when describing the spiked wines, it becomes clear that the contribution of the matrix plays a significant role. The total number of descriptors generated across sets (thiols x FIGURE 2. CA loading plot for Experiment 1. In figure 2a (3MH – top); 2b (3MHA – middle), the samples are colour-coded according to cultivar; 2c (4MMP – bottom), the samples are colour-coded according to thiol level.

0,8

M1_3MH_M

0,6

M1_3MH_0

M1_3MH_H

CS1a_3MH_H_Duplicate

0,4

SH1_3MH_0

SH1_3MH_M

CS1a_3MH_0

FIGURE 1. Experimental layout showing the project design. Individual thiols at three different concentrations (0 = zero, M = medium and H = high) spiked into five matrices. The wine samples codes are CS1a – Cabernet Sauvignon 1a, CS1b – Cabernet Sauvignon 1b (or blind duplicate), M1 – Merlot, SH1 – Shiraz and PT1 – Pinotage.

0,2

CS1b_3MH_H_Duplicate

0,0

CS1a_3MH_H

SH1_3MH_H

CS1a_3MH_M

-0,2

PT1_3MH_M

CS1b_3MH_H

Dimension 2 (13.09%)

PT1_3MH_H

-0,4

CS1b_3MH_0

PT1_3MH_0

CS1b_3MH_M

-0,6

2a

-0,8

-0,8

-0,6

-0,4

-0,2

0,0

0,2

0,4

0,6

0,8

Dimension 1 (32.58%)

0,8

SH1_3MHA_M

CS1a_3MHA_H_Duplicate

0,6

CS1b_3MHA_H

0,4

CS1b_3MHA_M

SH1_3MHA_0

M1_3MHA_H

0,2

SH1_3MHA_H

0,0

PT1_3MHA_M

CS1b_3MHA_0

-0,2

CS1b_3MHA_H_Duplicate

PT1_3MHA_0

Dimension 2 (20.75%)

CS1a_3MHA_0

M1_3MHA_0

-0,4

PT1_3MHA_H

CS1a_3MHA_H

CS1a_3MHA_M

M1_3MHA_M

-0,6

2b

-0,8

-0,8

-0,6

-0,4

-0,2

0,0

0,2

0,4

0,6

0,8

Dimension 1 (33.36%)

wines, Cabernet samples grouped together and were profiled as ‘raisin’, ‘prune’, ‘animal’ and ‘berry jam’; Shiraz as ‘planky’, ‘coffee’, ‘wood’ and ‘smoke’; and Merlot and Pinotage were clustered, and described as ‘cooked vegetables’, ‘savoury’, ‘herbaceous’ and ‘plum’. In contrast, spiking of de-aromatised red wine with 4MMP brought a separation of the samples according to increasing concentration, with the cultivars playing a lesser role (figure 2c). Only the base

the wine samples was mainly according to cultivar (figure 2a and 2b). In this case, the increasing thiol concentrations did not play a role. Nevertheless, with each thiol, the judges used different descriptors for each cultivar. In the case of 3MH spiked samples, Cabernet Sauvignon was described as ‘berry jam’, ‘earthy’, ‘plum’ and ‘soy’; Merlot was associated with ‘toast’ and ‘caramel’; Shiraz with ‘oak’ and ‘smoke’; and Pinotage was characterised by ‘smoke’, ‘spices’ and ‘planky’. When 3MHA was added to the de-aromatised

NONE

0,8

CS1b_4MMP_0

0,6

CS1a_4MMP_0

CS1b_4MMP_M

CS1a_4MMP_M

M1_4MMP_0

0,4

M1_4MMP_M

HIGH

0,2

SH1_4MMP_0

CS1a_4MMP_H

0,0

CS1a_4MMP_H_Duplicate CS1b_4MMP_H_Duplicate

PT1_4MMP_0

-0,2

CS1b_4MMP_H

Dimension 2 (17.77%)

SH1_4MMP_M

SH1_4MMP_H

-0,4

M1_4MMP_H

PT1_4MMP_M

PT1_4MMP_H

-0,6

2c

-1,2 -1,0 -0,8 -0,6 -0,4 -0,2 0,0 0,2 0,4 0,6 0,8 1,0 1,2 Dimension 1 (20.84%) -0,8

FIGURE 2. CA loading plot for Experiment 1. In Figure 2a (3MH – top); 2b (3MHA – middle), the samples are colour-coded according to cultivar; 2c (4MMP – bottom), the samples are colour-coded according to thiol level.

WINETECH TECHNICAL YEARBOOK 2020 87