Similarities and Concentrations in the Mix of Winegrape Cultivars

This study empirically summarizes the extent of similarities and concentrations in the mixes of winegrape cultivars across countries and seeks to determine by how much these mixes are becoming more or less similar and more or less concentrated. We use a database that provides the area by cultivar for 53 countries that account for 99% of the world’s winegrape area. For analysing similarities in the mixes of winegrape cultivars we use an index of similarities between countries, and we develop a hierarchical clustering method based on this index. This hierarchical clustering reveals a great degree of heterogeneity in the extent of similarities across countries. We also introduce a novel index for quantifying the degree of concentration in the mix of winegrape cultivars. This concentration index suggests that if two different winegrape blocks are randomly chosen anywhere in the world, the probability of those winegrape blocks having the same cultivar is just 2.2%. Using both indexes, we use a k-means clustering method to classify countries into three clusters: (i) a cultivar mix that is similar to the world’s and not very concentrated (79% of the world’s winegrape area), (ii) a mix that is not very similar nor very concentrated (18% of the area), and (iii) a mix that is not very similar but is quite concentrated (just 3% of the area). Further, we use these indexes to compare changes between 2000 and 2016, and conclude that the mix of winegrape cultivars has become more similar across countries and more concentrated globally over that period.