A Systematic Approach to Understanding the Sensory Character of South Australian Cabernet Sauvignon

C.G. Forde (1)*, D. C. Frank (1), P.J. O’Riordan (1), C.M. Delahunty (1), P. K. Boss (2), A.C. Cox (2), B. Loveys (2), E.R. Williams (3)

1) CSIRO Food Futures Flagship, Food Science Australia, Sydney, Australia
2) CSIRO Food Futures Flagship - Plant Industry, Glen Osmond, Adelaide, Australia
3Statistical Consulting Unit – Australian National University, Canberra, Australia
*Corresponding author email:  ciaran.fordecsiro.au

This is a comprehensive study of a best practice approach to understanding relationships between a product’s physico-chemical signature and its perceived sensory characteristics. Differences in Cabernet Sauvignon flavour arise from a wide range of complex reactions that occur during the oenological transition from grape to wine. The evolution of wine sensory character can be further influenced by both climatic and regional differences in the source material and these factors can change with each vintage. This study represents a systematic investigation of the flavour properties of Cabernet Sauvignon berries and wine and the underpinning volatile and non-volatile factors that influence flavour across vintages.

Cabernet berries were sourced from seven vineyard sites across South Australia and small scale wine making enabled comparison of the sensory properties of wines produced from the same berry lots. The use of paired sites within two of the vineyards sampled made it possible to correlate the sensory properties of grapes with different quality designations and value. The sensory properties were compared by ten trained descriptive panellists, and also by a group of eight wine making experts using descriptive analysis across three consecutive vintages (2004-06). The odour impact volatile compounds of the same Cabernet berries and wines were identified using Gas Chromatography Olfactometry with a panel of six trained assessors.

Descriptive analysis results demonstrated good agreement between the trained panel and wine making experts. The sensory results enabled sample differentiation based on climate, regional and quality designation and the key volatile differences between vineyards were identified. Multivariate modelling, including Partial Least Squares Regression (PLSR), established relationships between the sensory ratings from both panels and the volatile and non-volatile characteristics of the same samples. Relationships were established between Cabernet berries and wines within and across vintages. Although the current study focussed on Cabernet Sauvignon the same systematic approach could be applied to a wide range of complex products.