“Does bottle age reveal terroir ?” A discussion.

When  I received the notification from Erika Szymanski’s (also of Palate Press) blog Wine-o-Scope that a new post was up entitled “Quantifying terroir with chemistry: still searching for the Holy Grail”, I knew immediately that I’d find myself responding.  The destiny became reality when I saw the article she was citing – the most recent release from the laboratory where I had started my internship just days before.

In the final phase of my Master Vintage program, I am currently doing a 6-month internship in the laboratory of Régis Gougeon, at the Institut Universitaire de la Vigne et du Vin, (Institut Jules Guyot) of the University of Burgundy.  His current PhD student, Chloé Roullier-Gall, has taken me under her wing, and is the first author on the concerned paper discussed on Wine-o-Scope.  Thus I had no choice but to seize the opportunity to respond.

The paper, published in Food Chemistry, presents findings using an ultra-high resolution method of mass spectrometry, known as Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FTICR-MS), which is capable of determining the exact masses, and thus the exact chemical formulas, of thousands of compounds in a sample. Gougeon and his collaborator, Phillippe Schmitt-Kopplin of Helmholtz Zentrum Muenchen in Germany, use this methodology to look at the metabolomic profiles of grape, must and wine samples, effectively ‘fingerprinting’ them to look for particularities of individual or groups of samples.  In the present paper, the team looks at a series over 3 vintages of grapes, musts, and wines from four different vineyards, two in the Côte de Nuits region and two in the Côte de Beaune region of Burgundy.

Vougeot 1View from Château Vougeot in the Côte de Nuits

In the grapes, musts and just-fermented wines, no major differences were seen based on the vineyard locations.  Each of the three vintages gives a distinct profile in the freshly-made wine, but wines from the Côte de Nuits and Côte de Beaune region cannot be distinguished based on statistical analysis of their FTICR-MS profiles.

However, the team re-analyzed the wines in 2013, thus when the 2007s had been in bottle for 5 years, the 2008s for 4, and the 2009s for 3, and this is where the story gets interesting.  In her post, Szymanski wonders whether this follow-up analysis was out of desperation or always part of the plan, but frankly, it doesn’t matter.  In science, it is certainly true that many experimental decisions are made ‘out of desperation’, and honestly I don’t know the motivation in this case, but what is key is that they did the analysis, as this is where the most interesting result of the study was revealed.

Unlike in the wines analyzed immediately post-fermentation, the bottle aged wines do appear different based on the vineyard zone (the vintage effect remains important, as well).  Thus there is an evolution that occurs in bottle, that actually takes us back to a point where we can see differences based on the origins of the grapes – how do we explain that?! In the article, Roullier-Gall et al. propose that the terroir “impacts the initial chemical complexity of a wine, but time – i.e. bottle ageing – might be required to fully reveal it through the in-bottle diagenesis of complex chemical signatures.”  In other words, the terroir imprints its characteristics on the wine, but these signature elements are not revealed until the wine has aged in the bottle.  Could this correspond with some of the language that we employ to talk about bottle aging?  That wines need time to “open up,” “assimilate,” “integrate,” “mature” ?  Maybe there is a deeper truth to these statements than we know – something that goes beyond tannin softening and integration.  Maybe there are certain elements of a wines terroir ‘trademark’ that are only revealed after a bit of time in bottle.

Szymanski isn’t happy with this conclusion, as she says they haven’t gone this far in the study, that they have simply shown that “different wines are different.”  This is true, in effect, but dismissing it as such overlooks the real importance of the study, which is exactly this :  some kind of change is occurring in the bottle that allows us to characterize wine metabolomic fingerprints differently than immediately after fermentation.  Key changes are thus occurring to the chemical makeup during bottle aging, and we thus need to be keenly aware of this fact when choosing at which moment to analyze a wine, as this choice of time point may have an enormous impact on the results obtained.  And from a more romantic standpoint, yes, perhaps there are important changes occurring during bottle aging that reveal a wine’s identity.  But here Szymanski is right to be hesitant – it is a key finding of this study that opens up new questions to be explored, but for the moment, these questions remain wide open (but likely tied to the 95% of compounds found that are as-yet unidentifiable).

st aubin 2014 st vincentSaint Aubin, in the Côte de Beaune, during the Saint Vincent Festival 2014

In defending her statement that this paper’s sole valid conclusion is that “different wines are different,” Szymanski characterizes the researchers’ definition of terroir as broad.  They define it as the “vine-soil-climate-human ecosystem”, which I find to be an accurate summary of the most widely-accepted definitions of terroir, such as that of the OIV (Resolution OIV/VITI 333/2010) which states :

 Vitivinicultural “terroir” is a concept which refers to an area in which collective knowledge of the interactions between the identifiable physical and biological environment and applied vitivinicultural practices develops, providing distinctive characteristics for the products originating from this area. “Terroir” includes specific soil, topography, climate, landscape characteristics and biodiversity features.

For me, then, the authors’ definition of terroir is correct and incorporates all of the key elements, not broadening it by any sense.  In any case, even if they are only saying that different wines are different, that is not a conclusion to be dismissed.  It is, in fact, precisely the point that is interesting about this paper, which is that said differences only seem to appear after a certain time in bottle.

The criticism of merely two time points not being enough is a valid one, but this is only one article, and this work will surely be followed up, by this group or another, with a longer-term study that attempts to verify and further characterize these changes over time.  The present experiment is a proof-of-principle, successfully showing that this methodology can be applied to show differences from wines from different places, and further experimentation will need to be done to understand these differences, including additional sampling points, and certainly increasing the sample size as well.

Yes it is true that the vintage effect remained more prominent than the effect of vineyard location, but this does not, as Szymanski suggests, nullify the methods ability to distinguish terroir.  The vintage effect is stronger, yes, but the profiles of the wines can also be statistically grouped, within these vintage years, by location.  This is not surprising, as year-to-year climactic variation can be much more significant than the variation that occurs between vineyards located only a few kilometers apart (the entire zone of Côte de Beaune and Côte de Nuits spans a length of around 60 km), though this is complicated by the fact that terroir includes not only climate, but also geographical, pedological, and vitivinicultural factors.

In sum, the paper presents some interesting findings, not on the existence of measurable terroir effects so much as a methodology that allows us to elucidate distinguishing characteristics of wines only after a certain amount of time spent aging in bottle.  With further research this may eventually be traced to terroir fingerprints that are only revealed over time, but for now this remains an interesting theory, one to be questioned, debated, and explored, as Szymanski has already begun to do.

Personal note to Erika Szymanski : I must say that despite a bit of healthy debate here, I adore the concept of your work and of your PhD.  You hit precisely upon the theme that launched me into the world of wine with my Watson Fellowship – that it is an extraordinary context in which to understand relationships between scientific disciplines, between scientific research and those who put it into practice, and between empirical research and artistic expression. Wine is a fascinating subject matter precisely because the boundaries between all of these are constantly being pushed, prodded, dissolved, and redefined in ways that we don’t always see in other fields.  I’d love to read your thesis once its finished, and if ever you run across material that is particularly interesting in this regard, I’d be thrilled if you’d pass it along, either directly or via your blog!

6 thoughts on ““Does bottle age reveal terroir ?” A discussion.

  1. This is a fascinating discussion. Thanks for sharing it. It seems to me that if two self-contained systems are indistinguishable at time T1, but distinguishable at time T2, there must be SOME physical difference between the two which is eluding detection. I am a naïf in wine science (and so much else), and I wonder if differences in micro-organisms and/or their enzymes might escape detection by your assays and be a cause of the differences at T2.

    • Thanks Eric – yes I think that there are most certainly physical differences, but they remain, for the moment, uninterpretable, buried in the 95% of compounds seen on FTICR-MS that are, as yet, unidentified in the wine/grape/must model. Your proposal of micro-organisms is an interesting one, though I think that their enzymes are more likely to be responsible, as wine *should* be microbiologically stable (ie more or less sterile) at the moment of bottling.. but yes – very possible that terroir differences, however they’re revealed, are more of a function of microbiological differences between sites than differences in grape composition. This would certainly explain why in the grape and must samples, no location-based differences were found in this study..

      • In commenting here, I fear that I may appear like those people who feel they know more about cancer than do the oncologists, and post “cures” for cancer based on apricot pits and moonbeams. Alissa, you have spent more time considering the science of wine in your dreams than I have spent in all my wakeful years. Nonetheless, I’ll share a thought:

        There is a detectible chemical difference at T2 and T1. Either, I speculate, chemicals present at T1 have (cleaved or bonded and thereby) achieved a lower energy state by T2 spontaneously, or via enzymatic action. I know that for animal enzymes, there is an ideal temperature for their activity. For humans and other mammals, it is near 37 degrees C., of course. But the higher the temperature (ex vivo, where the enzyme cannot be replaced) the more rapid the degradation of the enzyme. Thus, as one raises the temperature from below ideal to ideal for reaction rate, there is an inevitable trade-off between the amount of reaction per time and its duration.

        Is there an ideal temperature for the enzymes in grapes and their micro-organisms? I suspect so, but don’t know. I suspect that that ideal temperature for them would be well below the ideal temperature of mammalian enzymes. I also suspect that the temperature-degradation curves for those enzymes would be no different from those of mammalian enzymes. (After all, the chemical bonds of between amino acids in a mammalian enzyme are no different from those in plant or micro-organism enzymes. A protein is a protein.) Therefore, I speculate that if there are enzymes present in wine, they could remain active longer at their ideal temperature than mammalian enzymes could remain active at their ideal temperature.

        Lotta speculation there, I see.

        I wonder if adding an irreversible broad-spectrum enzyme inhibitor (or perhaps a protease) to the wine would block the changes occurring between T1 and T2. If doing that did not prevent the change, you’d have good evidence against the role of enzymes.

        But for all I know, this experiment has been done decades ago.

        Best wishes for ongoing success!

      • Eric – thank you for your thoughts!! Contrary to your concern about appearing like you think you know more than the enologists, I think you bring up a very interesting point here, and one on which very little is known. I had to spend some time confirming today, because I had never even heard anyone talk about enzymes in finished wine today, and it turns out that this is precisely because very little is known. Grape enzymes and their role in grape juice are fairly well understood, and pectolytic and other enzymes are often added during the winemaking process to help with settling of solid matter, color extraction, etc.

        The work on finished wine, however, is relatively limited. It is primarily limited to the proteins that are of enological importance – notably those that cause protein haze (literally causing haziness in the – usually white – wine which, while harmless, tends to be quite disconcerting to consumers), and thus winemakers generally aim to eliminate them at a maximum. We know that the primary proteins involved in haze come from grapes – notably pathogenesis related proteins formed as a stress response by the vine – and are mostly eliminated by fining with Bentonite, a clay that, at wine pH, electrostatically interacts with proteins to pull them out of solution.

        Thus, in finished wine, not a lot of protein is found, and what is present is relatively poorly understood. There are some grape proteins, some yeast proteins, but I couldn’t find any research that talked about the activity of any enzymes that are in there. I don’t know how active they’d be in the wine environment (high [EtOH], low pH), but it seems feasible to me that there could be some enzymatic reactions occurring in the bottle.. but no – it appears that these types of experiments have not been done, and that people seem to assume (whether their reasons for doing so are valid or not, I am not yet entirely sure) that reactions in finished wine are spontaneous. But yes – by all means it seems like a good idea to test out some protein-inhibitors in wine (though it appears that working with a protease might not work so well, as given the breakdown of grape and yeast cells that occurs in winemaking, a lot of proteases are present, and so the proteins that do remain in finished wine are those that are most resistant to proteolytic attack…

  2. Alissa, thanks very much for the thoughtful discussion. I agree with you that the main import of this study is in demonstrating that something is happening in the bottle; that’s interesting, though I’d say it isn’t very new nor specific enough to add (I’ll hedge here and say “much”) new knowledge to the field. As a starting point, it’s great. In speaking of this operating definition of terroir, by the way, I’m not so much criticising the definition (I’ve written about this a bit at Palate Press, and my February article for them coming out next week addresses it head-on) as I am pointing out how it limits the scope of what this study can say: “something is happening in the bottle (but we don’t know what) related to something that happens with the wine (but we don’t know what). Again, as a starting point, nice; just, from my outsider’s perspective, no more than that. You’ll get me to agree any day that the knowledge shared in a published article is (for all sorts of very interesting reasons) a far cry from the knowledge held by people in the lab out of which the article is published.

    You adore my PhD concept, and I adore the concept of your Masters. What a fantastic idea to study wine — both as science and as art — in these different cultural and institutional contexts! I’m interested to know about the direction in which you next plan to take this experience.

    • Thanks for your response, Erika. I think that it is *potentially* new and different, as here the changes seen in the bottle may be revealing some differences present since the ‘conception’ of the wine – in the vineyard – that are not seen earlier. And, of course, this paper, as so many that turn out to become key, is primarily a methodological one, demonstrating the power of FTICR-MS and corresponding statistical analyses to study changes/differences in wine samples. But yes – it is all definitely a starting point, and a long road lies still ahead to unlock the potential of what all can be learned from analyzing wine in this way!

      As for the direction in which I go next, also a great question, and one which I am in the full throes of debating. A good chance that I am headed toward an oenological chemistry PhD, but I want to find a way to integrate these questions that I raise on the blog – questioning science, its relationship to art, etc etc., so that I remain capable of seeing the the forest for the trees, if you will, and not becoming so caught up in the minutiae of research that I lose touch with some of these bigger, meta- questions… in short, TBD.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s