Spatial variability in various indices of winegrape quality was studied over several vintages in blocks planted to Cabernet Sauvignon and Ruby Cabernet in the Coonawarra (1999-2002) and Sunraysia (2000-2002) regions of Australia. At both sites, inter-annual variation was marked whilst intra-annual variation was much greater for some indices (e.g. concentration of total phenolics) than others (e.g. Baume). The magnitude of intra-annual variation was readily identified in terms of the 'spread, defined as the difference between the maximum and minimum values, expressed as a % of the median value. Typical values of the spread were 20% for Baume, but could be as high as 117% for phenolics, and better indicated the extent of variation facing the winemaker than the coefficient of variation (CV; typically 3% for Baume and 14% for phenolics). For all attributes, variation in any given year showed marked spatial structure, with the patterns of variation being broadly consistent for each attribute in each year of the study, and with many attributes following similar patterns. The results therefore strongly support the idea of zonal vineyard management. However, fruit quality zone identification is dependent on a large sampling effort. Therefore, given the current availability of yield monitors, the finding that between-zone differences in quality indices were generally significant (P < 0.05) for zones identified on the basis of yield alone, and, in the absence of an on-the-go sensing capability, it is suggested that zonal management should proceed on the basis of zones of characteristic yield productivity. Based on the present work, it is suggested that development of an on-the-go fruit quality sensing technology would enable the wine industry to maximise its opportunity to gain benefit from differential vineyard management such as selective harvesting. indeed, the results of this work suggest that in the absence of zonal management, preferably supported by on-the-go quality sensing, winemaker demands for delivery of uniform parcels of fruit are unlikely to be satisfied.
