Résumé
Fourteen Syrah red wines with different initial composition and antioxidant properties (polyphenols, antioxidant capacity, voltammetric behaviour, colour parameters and SO2) were selected. Three different accelerated ageing tests (AATs) were then performed on these wines: thermal test at 60 ◦C (60 ◦C-ATT), enzymatic test with laccase (Laccase-ATT) and chemical test with H2O2 (H2O2-ATT). The results showed high correlations between the initial phenolic composition and antioxidant properties of the samples. Partial least squares (PLS) regressions were used in order to establish some models that can predict the AATs test results based on their different initial composition and antioxidant properties. The PLS regression models had overall very good accuracy and involved different explaining variables for each test. The models taking into account all the measured parameters and the phenolic composition alone showed good predictive capacities with correlation coefficients (r2) > 0.89.