This article contains scientific facts, as well as the researcher’s own personal opinion, about internal and external factors in terroir expression.
Eucalyptus and terroir
I tasted a Jacob’s Creek 2006 Cabernet Sauvignon the other day, definitely a rare occurrence for me.
Despite the wine being 15 years old, the eucalyptus note leapt enthusiastically out of the glass and evoked Australian vineyards baking in the afternoon sun, with kookaburras (very original, I know) calling in the hot gum trees nearby.
Australian red wine is known for a minty green attribute that may present anywhere on the spectrum from full-on crushed gum-leaf to subtle garden spearmint.1
The compound responsible for the minty/eucalyptus flavour in the Jacob’s Creek is 1,8-cineole, (commonly known as eucalyptol). Higher concentrations in wine are related to the presence of eucalyptus trees growing near the grapevines. For me, this is a powerful expression of terroir.
Methoxypyrazines and terpenes
The terroir concept is complex and encompasses a huge number of factors including the grapevine’s own responses to climate, water, canopy management practices and soil composition.
The methoxypyrazine concentrations found in grape berries are a good example of an endogenous (internal) grapevine response to environmental conditions around it.
Australian researchers have been able to differentiate international terroir/origin of wines based on their 3-isobutyl-2-methoxypyrazine (IBMP: ‘green pepper’ aroma) concentrations.
Bordeaux wines were shown to have significantly lower levels of IBMP than their Australian counterparts. Climate, grape maturity, light exposure, various viticultural management practices and crop yield can all affect IBMP concentration, so it makes sense that the terroir will have a huge influence on these compounds.
Another set of compounds that reflect grapevine’s endogenous responses are terpenes (spicy, floral smelling compounds in Muscat and Riesling grapes). It is well known that terpene concentrations increase in these cultivars in response to sunlight exposure and higher temperatures.
Australia vs Bordeaux
Coming back to the Jacob’s Creek example, we see quite another set of factors in the terroir effect. External or exogenous factors may not cause a grapevine response, but they are certainly present in wine, and are arguably every bit as important as an expression of terroir as the endogenous factors.
Analysis of Aussie Cabernets has revealed varying concentrations of eucalyptol in wines from different regions within Australia, prompting the researchers to suggest that this compound could help characterise terroir.
At the International Terroir Congress in 2020, Capone2 demonstrated that Aussie wines also have far higher levels of eucalyptol than Bordeaux Cabernets, so it’s definitely an “Australian thing”.
Another important exogenous factor in terroir expression is the microbes found on grape berries when they are harvested. These microorganisms originate from the vineyard and are present in the crush.
The initial steps in fermentation are carried out by many non-Saccharomyces yeast genera, which produce a wide range of higher alcohols and esters.
This may explain why Coonawarra wines were found by Capone and the team to have ‘liquorice’, ‘dark fruit’ and ‘spice’, traits associated with hexanol, ethyl propanoate and ethyl butyrate, whereas Yarra Valley wines are described as ‘red fruit’ and ‘jammy’, due to the presence of 1-octanol, 2-octanone and dimethyl sulphide.
A cluster of wines from Margaret River were also characterised by ‘grassy’, ‘savoury’ and ‘chocolate’ aromas, which were probably due to the presence of ethyl decanoate, butan-1-ol, (Z)-3-hexen-1-ol and ethyl-3-methyl butanol.
Each Australian region seemed to have some unique aroma attributes, as well as the common cultivar characteristics. The unique microbiome in each terroir may well have a lot to do with it.
Zooming in on volatile phenols
The volatile phenols (VPs) are yet another exogenous factor that could be considered an important aspect of terroir expression.
The two VPs, 4-ethyl phenol (4-EP) and 4-ethylguaiacol (4-EG), are part of a large group of compounds that are associated with smoke taint in wine, but 4-EP and 4-EG can also come from Brettanomyces spoilage (‘Brett’).
They are found in significantly higher concentrations in Bordeaux wines than in Australian wines. The characteristic ratio (4-EP:4-EG = 8:1) that is symptomatic of Brett appeared to account for their presence in the French wines tested by Capone and her colleagues, giving a range of ‘earthy’ and ‘yeasty’ aromas, which can be typical of Bordeaux blends.
The point of all this is that different regions do give us different wine chemistry, not all of them are related directly to the grape itself.
Some of the effects that we see that give wines its regional typicality are related to external factors (eucalyptus leaves and Brettanomyces for starters). So why don’t we use this terroir-chemistry relationship to nail down the uniqueness and authenticity of our own wine region?
So, what’s unique about terroir?
Easier to ask what isn’t unique! In the Western Cape, we have fynbos, the world’s richest and smallest floral kingdom. Other than the fact that fynbos contains a rich and pungent variety of flora, it also needs to burn to survive.
Fire is essential to maintain the healthy vegetation dynamics promoting synchronous germination and sprouting of bulbs, rhizomes and woody species. Periodic wildfires (veldfires to us South Africans) are a natural phenomenon in fynbos areas with a frequency of one to 40 years, although fire occurrence has increased a lot over the decades due to human activity.3
Fynbos fires are rapid and fairly cool, moving very fast over mountainous regions with the assistance of often gale-force South Easterly winds in summer. Vineyards are located all over the province and are frequently in the path of these bushfires during their peak-ripening season in February and March when the veld is at its driest, and fire risk is high.
These veldfires are accompanied by smoke that can cover hundreds of square kilometres.
Anyone who believes that fires (even under Covid lockdown conditions that restrict human activity) are not a problem in the Cape, only needs to look at the smoke plume covering the winelands in the NASA image (Figure 1) taken on 26 February last year.4
This fire was in the Jonkershoek mountains, a fair distance from human habitation, and yet the smoke plume choked Franschhoek, Stellenbosch, the Helderberg and even extended as far as Durbanville, all prime viticultural areas where grapes were ripening on the vines. The fire burned 33 000 hectares, the smoke drifted over vineyards for around a week, and a number of prominent producers were affected. Although large in area, this Western Cape fynbos fire event was definitely not unusual.
Veldfires emit large amounts of volatile compounds and particles and these emissions can significantly influence the chemical composition of the atmosphere in the region.
The impact of wildland fires depends on meteorology, fire plume dynamics, the amount and chemical composition of the emissions, as well as land use history and environmental conditions, so it is difficult to predict what will happen in any given veld fire event.
We also know very little about the specific emission characteristics (smoke quality and character) of fynbos. What we do know is that VPs like guaiacol, 4-methyl guaiacol, 4-ethyl phenol and the cresols are well known as smoke components in most fire events, and have been shown to be absorbed by grapes, and carried through to wine5 where they manifest as smoky, burnt aromas and an ashy taste (Figure 2).
Not only is the effect of the smoke difficult to predict, it’s also very difficult to manage smoke compounds once they are present in the wine. The VPs can be volatile, but they can also be locked up as glycosides (non-volatile compounds) that are released during winemaking and ageing.
The wine matrix also influences how easily the VPs can be perceived. Add to this that VPs are persistent little beggars that can’t easily be removed or chemically reacted away. This was underlined yet again in a recent comprehensive review of mitigating factors for smoke taint.
“Although the issue has prompted a surge in research on the subject in recent years, no singular solution has yet been identified that is capable of maintaining the quality of wine made from some-affected grapes.”6
Thus, one is yet again made aware of how much we know as a global industry about smoke taint (the ‘what’ and the ‘why’ are very well covered, specifically by the Australians), but the ‘what now’ seems to have been left out in the cold. Isn’t it time we cut our winemaking industry a little slack and took a slightly different view of this issue?
“The subtle, herbal notes imparted by fynbos along with discreet smokiness may give our Western Cape red wines a highly desirable edge in a competitive market.”
What have we found out about our own backyard braai?
We have been working on this issue consistently at the Department of Viticulture and Oenology, since around 2012 when we graduated our first Master’s student specialising in volatile phenol effects on wine.
Since then, at least five Master’s and one PhD have beavered away looking at how we can work with what we have, helping our industry to solve the problem as pragmatically as possible.
For example, during 2018, 12 wines were submitted to the Department of Viticulture and Oenology by South African wine industry members concerned that the wines might be smoke-tainted. These wines were from regions that had experienced fire events prior to harvest, some bad, some not so bad.
The wines were screened for a broad range of VPs, and the results were published in SAJEV. Historical data concerning fire events in the regions from which the wines originated, and during the relevant vintages were also investigated. The wines were evaluated by our “smoke expert” sensory panel and analysed chemically for VPs.
The wines contained a wide range of VPs in different concentrations. On the sensory side, out of 12 wines, four were described with negative attributes (‘smoky’, ‘ashy’ and ‘burnt rubber’), at significantly higher levels than the others.
The other wines, despite the grapes having been exposed to smoke, and having higher volatile phenol levels, were not overwhelmingly smoky, and showed attributes which included ‘berries’, ‘floral/ violets’, ‘prunes/raisins’, ‘vanilla/caramel’, ‘tobacco’ and ‘pencil shavings’.
High fruit intensity seemed to mask any sensory contribution by the VPs present in these wines, as has been indicated by previous work. Our own South African research7 showed that guaiacol could contribute ‘sweet, woody’ notes to wine and French workers8 found that VPs could enhance red wine esters (‘sweet, fruity’ notes).
In fact, in the wines we investigated, guaiacol (even at higher levels) did not seem to be correlated with a perception of ‘smoke’ in any of the wines unless it was in combination with other phenols.9
Oakwood also contributes a range of volatile phenols including guaiacol, 4-methyl guaiacol, the furfurals and eugenol, and nobody complains too much about the olfactory impacts of those. In fact, the judicious use of a bit of wood might work to help mask some of the VPs by adding sweet, coconut and spice flavours.
Taint? What taint?
Coming back to my point about chemical composition. So, what if the grapes have been exposed to smoke? It’s not always the case that the wine is an ashy mess, even if the smokiness is there.
In fact, it may be that a combination of the ‘subtle, herbal’ notes imparted by fynbos, along with a little smokiness may give our red wines an unmistakable identity in a highly competitive market.
It might be argued that a wine containing some smokiness is more accurately reflecting the landscape of its origin than one that has been stripped of these vintage and terroir effects.
Why are we trying to pretend that our indigenous, ecologically necessary fires never happen? The olfactory space in a red wine is complex, and if grapes have been exposed to smoke, they don’t necessarily need to be written off or banished to an entry-level blend.
And it’s not only me that is thinking this way… See Part 2 for the emerging thinking around smoke ‘taint’ in wine.
Discussing the exogenous or external factors that form part of world-famous wine terroir in Part 1, this opinion piece lays out the rationale for well-managed volatile phenols derived from inevitable wildfire events to form a part of our own Western Cape terroir expression.
Part 2 shares some insights into consumer reactions to smoke events around the world, demonstrating that wine lovers are not always predictable in their reactions, and quality is in the mind (and on the palate) of the beholder.
- Hooke, H., 2012. Huon Hooke in Wine News. https://www.therealreview.com/2012/10/30/eucalyptus/
- Capone, D., Boss, P., Souza Gonzaga, L., Bastian, S. & Jeffery, D., 2020. Characterising the chemical typicity of regional Cabernet Sauvignon wines. XIIIth International Terroir Congress, 17 – 18 November, Adelaide, Australia.
- Soós, V., Badics, E., Incze, N. & Balázs, E., 2019. Fire-borne life: A brief review of smoke-induced germination. Amaryllidaceae alkaloids and related natural products. Review Natural Product Communications 1 – 5. Doi: 10.1177/1934578X19872925.
- Gabbert, B., 2021. South Africa wildfire grows to more than 33 000 acres. Posted on February 27. CategoriesWildfireTagsSouthAfrica. https://wildfiretoday.com/2021/02/27/two-firefighters-injured-in-south-africa-wildfire/
- Ristic, R., Danner, L., Johnson, T.E., Meiselman, H.L., Hoek, A.C., Jiranek, V. & Bastian, S., 2019. Wine-related aromas for different seasons and occasions: Hedonic and emotional responses of wine consumers from Australia, UK and USA. Food Quality and Preference 71: 250 – 260.
- Mirabelli-Montan,Y., Marangon, M., Graça, A., Mayr, C. & Wilkinson, K., 2021. Techniques for mitigating the effects of smoke taint while maintaining quality in wine production: A review. Molecules 26(6), 1672. https://doi.org/10.3390/molecules26061672
- De Vries, C.J., Brand, J., Buica, A. & McKay, M., 2016. Effect of smoke taint on sensory characteristics of Cabernet Sauvignon wine. South African Journal of Enology and Viticulture 37(1): 22 – 30.
- Lorrain et al., 2013. Influence of phenolic compounds on the sensorial perception and volatility of red wine esters in model solution: An insight at the molecular level. Food Chemistry 140: 76 – 82.
- McKay, M., Bauer, F.F., Panzeri, V., Mokwena, L. & Buica, A., 2019. Profiling potentially smoke-tainted red wines: Volatile phenols and aroma attributes. South African Journal of Enology and Viticulture 40(2): 1 – 16. https://doi.org/10.21548/42-2-3270
– For more information, contact Marianne McKay at firstname.lastname@example.org.