It is difficult enough attempting to define the key influences underlying any flavour (see, for example, TasteMatters, July 2012: Driving a better tomato), but wine seems to be in a category all its own. Wine is chemically complex, but so are many foods and beverages we consume each day. What really sets wine apart (although some products such as cheese do come close) is its immersion in an artisanal mystique. The craft of winemaking resists scientific analysis, and this is particularly true of Old-World wines. After all, if winemaking was ‘just’ science, then any wine course graduate could plant vines in a paddock somewhere and 3 years later – hey presto! – Chateau Cheval Blanc ’47!
What really sets wine making apart is, of course, the land. In attempting to understand what makes a good cornflake, debates about the role of the soil in which the corn was grown are relatively uncommon. But the idea of terroir is now so entrenched in our appreciation of wine that its influence is taken for granted. The characteristic qualities of the land – the amount of clay or minerals in the soil or the local micro-climate – are seen as crucial to understanding not just the ripening and health of the grapes used in the wine, but the flavour of the wine itself. In essence, the wine flavor becomes an expression of the soil. This is sometimes taken to extremes, with the flinty character of a white wine being derived from the flinty soil, for example.
Beyond the romance of the idea of terroir, there are some practical reasons for linking the land to the flavour of the wines it produces. Thus, accepting this link means accepting too that wines from one region will in most cases taste quite different from those of another region, even if all other factors – grape variety, wine maker, storage type and so on – are kept constant. This is important, both for marketing purposes and for establishing that certain flavour characteristics are typical of a region and hence deserving of protected status. In turn, the idea of terroir underpins the rationale for using organic, or even the slightly whacky biodynamic, practices. It means too that the final say in what a wine tastes like cannot be left to consumer demands or market forces since the winemaker operates under constraints of the soil, and the soil determines what the wine ought to taste like.
Some of the attempts to study terroir have provided little support that this is an especially large contributor to flavour. In a descriptive sensory analysis of German Riesling wines from a number of wine estates, Fischer et al.  noted the huge variation in sensory properties among wines from the same vineyard, suggesting that terroir was a relatively minor influence on flavour, compared to the major influence caused by vintage and wine estate. They cite another study published in German by Wahl & Patzwald (1997) in which the researchers went to the effort of transplanting seven different soil types to the same vineyard to study the impact of soil type on wine composition and sensory quality of Silvaner wines. They reported no significant impact on wine ̄flavour of the soil type, beyond different grape yields.
In fact, a scientific analysis of the impact of terroir has proven difficult. Not only do soil types and climate vary with geography, but of course so do other factors including such things as the location of the vines relative to drainage and sunshine. Even when two locations have a winemaker in common, the entrenched belief by the winemaker in the influence of terroir may be an explicit or implicit source of handling the grapes or the wine in different ways.
Recently, Cadot et al.  surveyed wine producers from the Anjou region of the Loire in France to determine their concept of wine (flavour) typicality and how they thought that it related to terroir. Not surprisingly, for these producers, the main characteristic that explained both typicality of the wine and its flavour was the terroir. Soil and climate characteristics were important for 93% of the wine producers (compared to 65% for wine-making practices and 5% for harvest quality). The winemakers’ judgments of what constituted a typical wine of this region (and hence what was the main influence of terroir) were sensory attributes such as colour intensity, red fruits, and soft tannins.
In contrast, a descriptive sensory evaluation of the region’s wine produced a profile of the ‘perceptual typicality’ of the wines. Here, visual descriptors, spiciness and astringency, but not red fruits or soft tannins, were important. Moreover, when it came to those factors that distinguished the more prestigious style (Anjou-Villages Brissac) from a more quaffing variety (Anjou Rouge), only those technical factors under direct control of the winemaker - maturation time, vatting time, harvest date, and proportion of the Cabernet Franc grape – were influential.
Similar conclusions were drawn from a study undertaken some years ago by one of my students at the University of Otago, Sara Springhall . Sara asked 27 experts/semi-expert wine tasters (all either teaching enology, undertaking enology courses, or members of wine clubs) to taste 13 Chardonnay wines sourced from three distinct regions in New Zealand - Hawkes Bay, Marlborough and Central Otago. According to http://www.winesofnz.com, these regions are characterized respectively as having (a) high sunshine hours and variety of soil types; (b) Lots of sun, cool nights, low autumn rains and free draining alluvial soils; and (c) hot, dry summers, snowy winters, and soil structures that are very different to those of New Zealand's other regions, with heavy mineral deposits in silt loams.
The tasters were asked to sort and group the wines based on similarity of flavour and then provide descriptors for the most prominent sensory characteristics. The sorting data were analyzed by using the number of times that wines were grouped together as a measure of their “distance” from one another. This allowed the data to be represented as a multidimensional map, which showed both substantial overlap between regions (Central Otago and Marlborough) as well as clear separation (Central Otago and Hawkes Bay). The axes of the map (essentially, North-South vs. East-West) were shown to be related strongly to the common descriptors for the wines. The major axis, the one along which the wines differed most, was found to be associated positively with the woody and caramel attributes of the wines, and negatively with the wines’ sourness. Variations in wine citrus flavours were associated with the secondary axis.
These axes were also strongly associated with the chemical characteristics of the wines. In particular, alcohol and sugar content varied positively with the main axis of the map, while pH correlated negatively and volatile acidity positively with the minor axis.
What all these data mean is that both sensory attributes and chemical characteristics underpinned the ways in which the wines were sorted. However, since the wines grouped according to geographical regions were not strongly aligned to either the sensory or chemical dimensions on which these wine experts sorted the wines, the data suggest that the most important influences on the flavor of these wines occurred during winemaking. Woody and caramel qualities, for example, generally originate as a result of contact with oak during fermentation and maturation, while variations in citrus flavours, alcohol and sugar content are linked closely to when the grapes are harvested.
A failure to reveal the impact of terroir does not mean that there aren’t better or worse soils or microclimates in which to grow grapes that make good wines. But it does mean that we ought – for the moment – to be sceptical about wine producers’ claims that wine flavours are to any great extent a product of the soil. Alternatively, you can accept the claims, because of the romance of the idea …. but to be consistent, it’s probably a good idea to start asking your bartender about the soil in which the hops were grown next time you order a beer.
1. Fischer, U., D. Roth, and M. Christmann, The impact of geographic origin, vintage and wine estate on sensory properties of Vitis vinifera cv. Riesling wines. Food Qual Pref, 1999. 10: p. 281-288.
2. Cadot, Y., et al., Characterisation of typicality for wines related to terroir by conceptual and by perceptual representations. An application to red wines from the Loire Valley. Food Qual Pref, 2012. 24: p. 48-58.
3. Springhall, S., et al. Multidimensional sorting applied to understanding flavour variations in Chardonnay wines in 5th Australasian Association of Chemosensory Science Annual Scientific Meeting. 2002. Heron Island.