Which countries are similar to each other, in terms of their grape varieties?

A couple of weeks ago I noted that Regional diversity of grape varieties is important for climate change. The basic point is that with changing environmental conditions, we need a range of grape-vine varieties available, so that we can see which ones do best under the new conditions.

A related issue is, of course, what varieties are available at the national level. After all, the existence of varieties is not of much practical use to you if they are not actually in your country. One major concern that is sometimes expressed is about the diversity of wine-grapes narrowing down to a few so-called “international” (noble?) varieties — after all, 13 grape varieties cover one third of the global vineyard area. These are often presented as mono-varietal wines, which results from trying to maximize short-term economic returns, rather than planning for the future.

We can look at this issue with the same source of data used in my previous post, the Database of Regional, National and Global Winegrape Bearing Areas by Variety. This compendium has vineyard area data (in hectares) from 2010, for 1,446 named grape varieties in 48 grape-growing countries. You will never have heard of the vast majority of these varieties, as there have probably been at least 10,000 of them recorded at some time in history.

What I wish to do with these data is to examine just how similar the countries are to each other, in terms of the abundance (area) of these grape-vine varieties. This can be conveniently displayed as a network, as I have done in a number of previous posts. [Technical note: the relationships among the countries were calculated as Bray-Curtis distances, and the NeighborNet graph was drawn using the SplitsTree program.]

I have included the network at the bottom of the post, for those of you who would like to see the details. Here, I will simply summarize what the network tells us about the relationships among the cultivar varieties of the different countries.

1. There is no strong pattern in the network, as there has often been in my previous posts using networks. So, there are no clusters of distinctly different countries, as far as their grape diversity is concerned. That is, grape-vine varietal diversity is relatively evenly distributed among countries, which is a good thing for maintaining diversity.

2. The long edges around the outside of the network indicate that most countries are more different from each other than they are similar. This is interesting, because it means that each country is preserving grape-vine diversity in a somewhat different way. This is not necessarily unexpected, of course, mainly for historical reasons. Moreover, different countries often do have different climates, and therefore different varieties will do well. Also, many countries have chosen to specialize in different varieties, as a wine-making business strategy.

3. There are a few small clusters of countries, which I have highlighted with different colors in the network diagram. For example. clockwise from the bottom:

• there is a varietal connection between Bulgaria and Romania, which is not unexpected, but also between Bulgaria and China, which you may not have expected;
• there is quite a close grape-vine connection between the Ukraine and Moldova, and a lesser connection between the Ukraine and Russia;
• there is a reasonably close varietal connection between Croatia, Slovenia, Slovakia, Czechia, Austria, and Hungary, which presumably reflects the historical Austro-Hungarian Empire (dismantled after World War I), plus a somewhat separate similarity between Croatia and Serbia;
• there is a varietal association between Thailand and Myanmar, in South-east Asia, plus a more distance (intriguing) connection to Ethiopia;
• Taiwan and South Korea have a reasonable association of varieties;
• Armenia and Kazakhstan are closely connected, reflecting the cradle of wine-making, with Georgia being somewhat more distance;
• there is a small varietal connection between Luxembourg and the United Kingdom, which may reflect a focus on sparkling wines.

4. Then we come to two sets of large but somewhat more complex associations.

First, there are varieties shared among Tunisia, Algeria, and Morocco, reflecting their wine-making tradition as French colonies, supplying warm-climate blending wine to supplement the weaker red wines of France. These countries are then connected to Israel, Mexico, Cyprus, and Turkey, where the same varieties do well. Then there is an intriguing connection of Canada, Japan, Uruguay, and Israel, which bears some looking-into.

Finally, we come to the relationships that we might expect, connecting the varieties of the New World to their origins in Europe. That is, there is an obvious varietal connection between France, Italy, Argentina, South Africa, Chile, Australia, and the United States. However, there is also a separate association between Spain and France, presumably reflecting their shared varieties that have not (yet) become common in New World wine-making.

Note that the grape varieties that are popular in New Zealand are not similar to anywhere else, possibly because of the restricted range of popular varieties. Germany and Switzerland are also disconnected, presumably because most of their varieties have not emigrated.

Below is the network diagram (click to enlarge it). The technical interpretation of this network is described in my post: Summarizing multi-dimensional wine data as graphs, Part 2: networks. Basically, countries that are closely connected along the network edges are similar to each other in terms of their area of the different grape-vine varieties.