The most common descriptor associated with terroir is “undefinable”—it’s attributed to a mix of influences including soil, exposure, and climate that create distinctive character in the wine coming from different vineyard sites. At its simplest extreme, it is scarcely rocket science to accept that grapes (or for that matter any other crop) will be quite different if grown in a sunny, well drained spot at the top of a hill from those grown in a shady, waterlogged spot at the bottom of the same hill. That reductio ad absurdum disguises the fact that there can be vineyard sites which to all appearances are identical in all the parameters you might think distinguish them, but which nonetheless consistently produce different wines. No one to date has been able to explain this mystery in any rational way.
A large part of the mystique of wine would disappear if we could explain terroir on a scientific basis, but there is actually little threat from the latest attempt, a report in the Proceedings of the National Academy of Sciences from a group at the University of California, Davis headed by Dr. David Mills. The basic finding is that microorganisms on the skins of the grapes are different in various vineyards in California. These observations were made possible by the ability of rapid DNA sequencing to identify large numbers of microorganisms without the need actually to culture them.
So vineyards in different places have different sets of bacteria and fungi, very likely related to both region and climate. Wind, temperature, and humidity were identified as pertinent factors. This is scarcely a major surprise. But that’s a far cry from supporting the conclusion that “these differences may help explain regional patterns in wine chemicosensory properties,” which is quasi-scientific speak for arguing they might be involved in terroir.
For these microorganisms to be involved in determining wine character, they would most likely need to play some role during fermentation (their mass could scarcely be sufficient to provide any significant level of flavor components in themselves). Well, here is a problem. A significant proportion of the microorganisms that were detected are so-called spoilage organisms, which if allowed to act on the grapes in fact spoil the wine. Precautions are taken to prevent this (most typically the addition of sulfur to block bacterial or fungal action before fermentation).
In fact, wine producers divide into two groups on the issue of the role of yeasts in fermentation. Probably most producers in California are in the group who kill off all the indigenous yeasts before fermentation; cultured yeasts are then added. This provides control of the process and prevents spoilage organisms from acting, making it quite unlikely that any yeasts or bacteria on grape skins are involved in determining wine character.
Some producers allow fermentation to be catalyzed by indigenous yeasts. But here is an interesting point. In spite of a longstanding general belief that this is part of the properties of the vineyard (which makes the present article appear somewhat less than novel), it appears that the yeasts that actually catalyze fermentation are different every year. Eminent oenologist Ribereau-Gayon said, “In a given vineyard, spontaneous fermentation is not systematically carried out by the same strains each year; strain specificity does not exist and therefore does not participate in vineyard characteristics.” So it is hard for this to explain terroir. Curiously, the present paper finds that there is greater vintage variation between small vineyards than between wider areas (but this is based on a comparison of only two years). To make a case that microorganisms are involved in terroir, wouldn’t you need to show that there is some consistency in them over substantial periods of time?
The microorganisms also differ from regards to grape variety, again not a surprise, as grapes with thin skins (more easily damaged) are likely to attract different microorganisms from tougher grapes with thicker skins. This goes back to the point that many of the microorganisms are in the spoilage class, and the last thing you want is for them to affect wine character.
My bet would be that if microorganisms are involved in terroir, it would be more likely to be those in the soil that act on the roots of the grapevines, and which might therefore indirectly affect the properties of berries as they develop. Maybe the answer lies in the soil.
Reference: http://www.pnas.org/cgi/doi/10.1073/pnas.1317377110
Lewin on Wine 
I’m not so sure, Ben. First, much of what we term terroir is determined pre-fermentation. Otherwise, all wines made with, e.g., Pasteur Red yeast would have a similar flavor pattern. That clearly isn’t the case as even an amateur winemaker like myself knows. Secondly, wines do in fact change from year to year, as an amateur taster like myself knows, and an MW like yourself *really* knows.
Allow me to raise the flag for the neglected indigenous microbes on the skin. Those compounds that we associate with taste and color are secondary metabolites. Whose secondary metabolites? is the question. If Dekkeria (Brettanomyces) can contribute flavor (good or bad, your choice) post-fermentation to old-line Burgundy or Chateaueuf, why couldn’t some other (as yet unknown) microorganism contribute a (good) flavor to the grapes pre-fermentation? We know that they exist on the skins, and that they (1) cause the synthesis of secondary metabolites in the plant, presumably for defense. (2) They have the genetic capacity for secondary metabolism, themselves. Some of the compounds they make undoubtedly have aroma.
To rule this possibility out would require determining the ability of V. vinifera to make *all* the 3000+ known compounds in wine. I’ll hold out for the microbes.
Well I certainly agree that most, if not virtually everything, that we associate with terroir depends on events pre-fermentation. I think most people would agree that (assuming terroir exists) it is caused by differences in the grapes at time of harvest. The cause of those differences is of course subject to much discussion. Yes, there is a view that microorganisms associated with grapes are part of terroir, and to that extent, using cultured rather than indigenous yeasts reduces (but does not necessarily eliminate) expression of terroir. There is no doubt that the major part of the aromas of wine are created during fermentation — but they are created from the basis of what’s in the grape. Yeast certainly affect the character of the wine, but current data suggest they are not consistent from year to year: so if terroir is a long-term effect, it’s hard to associate it with the yeasts that undertake fermentation. This is not to minimize the importance of the yeast, but just to argue that whatever is responsible for terroir effects, e.g. the association of minerality with some sites, may have a different basis. I am not at all sure that microbes on the skin of the grape have a sufficient influence on events inside the grape to support an argument that they contribute significantly to aroma or flavor. Going back to the defining example of Burgundy, consistent differences over the years between, say, Gevrey Chambertin Cazetiers and Combe aux Moines, or between Chambertin and Clos de Beze, seem to transcend vintage and producer. The mystery of terroir is that it’s so difficult to envisage just what could be responsible for such a long-term, consistent effect.
Excellent piece of writing.