Harvest date and crop level influence sensory and chemical profiles of Ontario Vidal blanc and Riesling icewines.

There is likelihood that periods of freezing and thawing endured by icewine grapes change their chemical and sensory profiles. The experimental objective was to determine the influence of harvest date and crop level on icewine sensory profiles and their relationships with chemical variables. Riesling and Vidal blanc (hereinafter "Vidal") icewines were made from four harvest dates in 2004-2005; Harvest 1 (H1): 19 December; H2: 29 December; H3: 18 January; H4: 11 February (Vidal only). Riesling and Vidal icewines were additionally made from three crop level treatments [control (fully cropped), fruit set cluster thinning (i.e. partial fruit removal) to one (basal) cluster per shoot, veraison cluster thinning] and were evaluated over two seasons (2003-2004; 2004-2005). Triangle tests showed differences between harvest date and crop level treatments (both cultivars). Ten and 11 aroma/flavor attributes differed for Vidal and Riesling harvest date icewines, respectively, based upon descriptive analysis by 14 trained tasters. For Vidal, later harvest dates had higher aroma/flavor intensities than H1. Riesling H1 wines had highest fresh fruit descriptor intensities whereas H3 wines were highest for dried fruit and nutty descriptors. Partial least squares regression (PLS) found Vidal icewines described by dried fruit/raisin and honey flavors and viscosity; these attributes were correlated to several aroma compounds and associated with later harvest dates. Sensory differences existed between crop level treatments; thinned treatments had higher fruity, honey, sherry and nut aroma/flavor intensities (both cultivars). PLS showed that sherry flavor was the most important explanatory variable in 2003, and correlated 4-vinylguaiacol and banana flavor in 2004. Overall, harvest date and crop level both affected sensory profiles and chemical composition of Niagara Peninsula icewines.