Quantifying terroir with chemistry: still searching for the Holy Grail

It’s the Holy Grail of wine research: chemical proof of terroir. Though the quest has fallen out of favor with those who think that the Grail can never be found, or that it doesn’t exist, it should come as no surprise that many are still searching.

The most recent contribution: a team from France and Germany have used mass spectrometry to look for chemical differences among wines from four different vineyards — two in the Côte de Beaune, two in the Côte de Nuits — across the 2007, 2008, and 2009 vintages. Mass spectrometry* is in essence a way of separating out all of the chemicals in a sample by weight — each compound has a unique weight — and displaying what the sample contains by showing a blip on a graph for every unique compound. A sample containing only one compound will show only one blip, a sample containing two compounds will show two blips, and something as complex as wine will show — according to this study — up to 7016 unique blips. The researchers analyzed must, dried skins, and finished wine (fermented by the commercial winery) to come up with what is, essentially, a chemical fingerprint: a unique set of blips for each sample.

Think we know a lot about wine chemistry? These scientists could identify only about 5% of the compounds identified in their fingerprinting. If we ever wanted evidence that wine is more complex than current chemistry can explain, we have it here. We know about all of wine’s major constituents, and many of the minor constituents that affect flavor, color, and smell. But amongst all of those yet-unidentified minor components might be the explanations for any number of wine mysteries we can’t currently explain.

Even without knowing what all of the compounds are, though, we can still see their signature blips and make some generalizations about which wines are most similar based on the similarity of their fingerprints. No doubt the researchers hoped that the different wines would show nice, tidy similarities among wines from the same place and wines from the same year. They found similarities; they just weren’t tidy. The Côte de Beaunes wines weren’t as a group chemically distinguishable from the Côte de Nuits. And one of the Côte de Beaunes wines grouped more closely with the other vineyards’ 2007 vintages even in 2008 and 2009, though the other wines were roughly similar by vintage.

Maybe out of desperation — or maybe because it was part of the plan all along — the researchers fingerprinted the wines again in 2012, three, four, or five years after bottling for the 2009, 2008, and 2007 vintages, respectively. Encouragingly, the wines did group more closely after some age: both the regions and the vintages formed tighter clusters. The authors concluded that “this fundamental result highlights the fact that the terroir as a whole – i.e. that considers the vine-soil-climate-human ecosystem – definitely 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.” Diagenesis, by the way, is the process by which accumulating particles transform over time into sedimentary rock. It’s a geology term, though I can see the analogy here in how accumulating chemical changes can transform over time into something new with distinct properties.  This is taking things just a bit too far, I think. Their results from just-bottled wine aren’t convincing and, with their broad definition of terroir, they’re really saying nothing more than “different wines are different,” which isn’t very helpful. There’s also the problem that the vintages aged for different amounts of time, and that we only have two time points: some unspecified time very shortly after bottling, and three or four or five years later. In another five years, would the wines have grouped even more tightly? Or less so?

These scientists want their results to mean “that various and distinct chemical reaction pathways can actually exist in bottles, which provide after a few years of ageing an instantaneous picture of the chemical complexity associated with a specific terroir that could potentially be as small as the countless ‘climats’ of Burgundy.” I’ll grant them the point about chemical complexity, which is beautifully demonstrated in this paper even if it is something we already knew. The rest of their conclusions seem out of place. Disappointingly, they haven’t demonstrated that they can link that chemical uniqueness with individual vineyard plots. Not even close, I fear. First, their sample size is too small, looking only at four vineyards from two regions (with no mention of what the individual differences are between the vineyards in terms of soil, elevation, etc.). Second, the wines from each vineyard weren’t more similar to each other than to the other wines from the same year or the same region, which seems to quite clearly say that these chemical profiles can’t be used to distinguish local terroir. Again, all we can say with any confidence is that different wines are different.

The strength of this study lies in its being, with any luck a jumping-off point. Chemically fingerprinting terroir is a worthy goal, and if chemists analyze enough wines, they’ll eventually see useful patterns. I imagine these wine like stars in a crowded night sky. Amongst the billions of stars, if we isolate just four or six, we can say that some are closer together than others, or that they’re clearly different points in the sky. Ho hum. But if we look at enough stars, we can start to develop constellations — which, like terroir, are a story, a tool we use for understanding the world. And if we look at enough stars with the right kind of instruments, maybe we can start to develop even more useful tools for thinking about them.

I suppose, then, that the Holy Grail is still out there, amongst the stars.

*A more complicated explanation of the kind of mass spectrometry that these researchers used, along with a lovely explanatory video, is here, courtesy of the Johnson lab at Yale

Are wine drinkers healthier than beer drinkers? Not so fast…

Yet another study was published this month showing that habitual wine drinkers are, all things considered, healthier than habitual beer drinkers. This one was about Dutch people — part of a national dietary survey — but similar findings have come out of the United States and other parts of Europe. We can draw all kinds of conclusions from such data, some of them plausible, but many of them ill-advised. Let’s jump to some conclusions and see where they take us.

1. Wine is better for you than beer – Maybe. Let’s say that you’re a marathon runner who runs the eight miles each way to work every day, teaches yoga in the evenings, eats salads for breakfast and lunch and half a pizza for dinner most nights. Your next-door neighbor drives to his desk job, gets winded walking up the stairs, and eats stir-fried tofu and veggies for dinner most nights. A study could conclude that you, the pizza-eater were healthier than your neighbor, the stir-fry eater, but that wouldn’t mean that pizza was healthier than tofu. Clearly, other lifestyle choices are contributing to your difference in health. The same is true for studies of habitual wine or beer drinkers: wine drinkers may be likely to make healthier choices in general.

In fact, this study came to exactly that conclusion: wine drinkers ate more fruits and veggies and less red meat, soda, margarine, and snacks than beer drinkers. We clearly can’t ascribe the health difference between the two groups just to their choice of beverage. That said, the study found that when the effects of several other health behaviors were accounted for — that is, when we measure and remove the effect of total calorie consumption, or smoking — wine drinkers were still healthier. Moreover, this study looked at healthy lifestyle habits, not at  measures of the participants physical health other than body mass index, and measuring healthfulness by weight-to-height ratio alone is pretty simplistic. This study gives support, then, for the idea that wine might be healthier than beer, but doesn’t give us enough information to reach that conclusion.

2. People who want to be healthier should start drinking wine – Maybe but, again, not supported by this study. We’ve already concluded that beverage alone isn’t responsible for the health difference between wine and beer drinkers. There’s no reason to believe that starting to drink wine, or switching from beer to wine, will help you eat more veggies.

3. Wine drinkers are better educated than beer drinkers – Yes, actually. Among the four groups considered in this study — wine drinkers, beer drinkers, spirits drinkers, indiscriminate drinkers with no preference, and alcohol abstainers — wine drinkers were most likely to have university degrees.

4. Wine drinkers tend to have healthier lifestyles than beer drinkers – Yes. This is the study’s primary conclusion: “alcoholic beverage preference was associated with dietary habits in The Netherlands.” On average, wine drinkers eat more healthfully, consume fewer calories, are less likely to smoke, and tend to have a higher socioeconomic status (which is itself strongly correlated with healthful habits) than beer drinkers. But we don’t know why this is true, though we can say that socioeconomic status and education have a lot to do with it.

So, we can fairly say that regular wine drinkers tend to live healthier lifestyles than regular beer drinkers in the Netherlands. Generalizing these findings to other countries is dangerous. Similar studies in other northern European countries (Denmark, for example) and the U.S. have come to similar conclusions. No surprise: any American could probably guess that American wine drinkers in general tend to be better educated and wealthier than beer drinkers. But that’s a cultural phenomenon. In Italy and Spain, where regular wine drinking in volume is a habit of the old guard, wine consumption doesn’t necessarily go hand-in-hand with healthy eating or other markers of a healthy lifestyle.

The conclusion? Studies such as this are interesting. Maybe they help the medical profession get a better picture of who to target with healthy lifestyle interventions (though education and socioeconomic status do most of the work here). But they tell us nothing at all about whether wine is good for you.

What I’ve been reading

I’ve just added a page (linked at the right) entitled “What I’ve been reading” where I’ll keep short notes on books and articles that I’ve been reading of late, mostly about wine, science, and rhetoric (and sometimes about other things). I hope that some of these may spark your interest, too.


Sensory speculations on the Riedel Coca-Cola glass

The wine news is making hay this week with specialty glassware maker Riedel’s newest custom glass shape designed for Coca-Cola. While the process for selecting the glass sounds pretty empirical — a panel tried Coke out of a bunch of prototype glasses and chose the one they liked best — I can’t help but wonder about the sensory chemical logic behind the design. There’s no sense in being pretentious about this: even if Coke is a mass-produced beverage and a cultural and health nightmare, it’s still very sensorily complex (and unquestionably popular).

The glass recapitulates the wide-shouldered hourglass shape of the old-fashioned glass Coke bottle. That’s simple brand congruity: only the shape of the glass opening and the upper bowl affect the glass’s sensory properties in terms of directing aromatics to the nose and affecting how the liquid hits the palate. The dynamics of how Coke behaves in a glass will be wildly different than wine, with the possible exception of a sweet sparkling. Coke has bubbles, which actively convey volatile aromatics into the head space above the glass. It’s extremely high in both sugar and acid (phosphoric, as opposed to tartaric, malic, and lactic in wine) and contains caffeine, none of which should have a significant affect on aroma save insofar as the sugar increases viscosity. How that compares with the viscosity of wine, where alcohol and glycerol (and sometimes sugar) are responsible for viscosity, I’m not sure.

A spokesperson from Riedel says that the top of the glass is the same shape as the Riedel O-series Sauvignon Blanc glass, which immediately sent me on a search for methoxypyrazines and thiols (two prominent characterizers of Sauv Blanc) in Coke.

No joy, and no surprise: methoxypyrazines are responsible for green bell peppery notes, thiols for various tropical fruit and grapefruit-y aromas. It’s been a while since I had Coke, but I’m pretty confident that neither bell pepper nor passionfruit feature prominently in its flavor profile, even if its citrusy notes are easily agreed-upon. The Open Cola Project recipe, which we might reasonably expect to be in the right ballpark, calls for orange, lemon, and lime oils along with cassia (Chinese cinnamon), nutmeg, coriander, and lavender, and a lot of sugar and acid and caramel color.

From a theoretical perspective, then, I’m going to guess that the glass emphasizes Coke’s spritely and refreshing citrus aromatics first and foremost, leaving the sweet caramel/vanilla and spicy notes to bring up the rear. That testers would prefer that effect is congruent with the famous 1980’s and ’90’s Coke vs. Pepsi trials, which showed that Pepsi tended to win out in sip tests — both because it was sweeter and because it has a heavier initial citrus impression — but that Coke had more lasting fans — because it was less sweet, because Pepsi’s citrus tends to fade after the first few sips, and because Coke has a more robust caramel backbone.

On that basis, the glass should either be really good for a rum-and-coke — if you’re using cheap, sweet rum and want to maintain the refreshing balance of the drink against the extra sugar and body — or really bad for a rum-and-coke — if you’re using decent rum and want to play up the sweet/vanilla/barrel aromas.

If I get hold of a glass, I’ll test the theory — this is worth one small exception to my long-standing boycott of Coca-Cola (as well as Pepsi and a number of other food mega-companies) as a response to the company’s massive funding of campaigns against mandatory labeling of GMO-containing foods (and I can’t stand the stuff in any case). But I’d love to know what characteristics the glass brings out, and to play with fresh vs. flat, ice vs. no ice.

What’s next? A gin and tonic glass? A raw milk glass? An orange juice glass? A Pepsi glass? Tea glasses are apparently in the works, but I can only hope that they’ll differentiate oolong from lapsang souchong from pu ehr.

A Humanist Rationale for Wine Science

Wine writers have all manner of reasons for writing about wine: the people, the moments of beauty, deciphering complexity for the hapless consumer, a passion for free samples. Why I write about wine is a question I perennially reconsider – the answer seems to keep changing – but I can always take the easy (and true, if not exclusively true) cop-out: wine is a fantastic vehicle for helping people learn about science.

Still, that answer begs the question: why is it good or valuable or worthwhile to help people learn about science? I’m currently studying in a department called the Centre for Science Communication, a magical place where everyone cares about bringing Science to the Everyman in accessible and entertaining ways. Some reasons for teaching people about science are pragmatic: better-educated people should make better decisions. Show people the beauty of the penguin or the wild orchid and empower them to care more about how their lives affect the natural world. It’s the democratic argument for why we have compulsory education: educate everyone enough to make them good (enough) citizens. Educate them more and make them better citizens. Or, if you prefer, the capitalist argument: people who know more science can do jobs that involve more science and generate more revenue than the ignorant. All good.

But I chose a small liberal arts college for my undergrad years, where I studied music and philosophy along with my molecular biology. I’m not all pragmatist; I believe in my heart of hearts in a liberal education. Learning makes us better people because it increases our appreciation and enjoyment of the world, because men were made to live and living is more than mere work and production. Moreover, learning makes us able to see. We only see what we can conceptualize, things for which we’ve created mental boxes. The Inuit can see twenty different kinds of snow; the educated oenophile can see twenty different kinds of pinot noir. The more we see, the more we are able to observe and to strive to understand, the more able we are to seek truth in all of its forms, and this is the highest calling of man. I don’t really write about wine science because it helps people learn about science. I write about wine science because I want to help people see and seek to understand.

I’ve been sipping on the taMt. Beautiful 2011 North Canterbury Pinot Noiril-end of a bottle of Mt. Beautiful 2011 North Canterbury Pinot Noir as I’ve been writing this. The wine is pretty (see below). The website gives me more than the usual information about how the wine was made, and that makes me happy. Can I taste the seven days of cold soaking or smell the twice-daily punch downs? Heck no. By knowing these details, am I thereby prompted to see and seek to understand more about the wine? Yes.  

Mt. Beautiful 2011 North Canterbury Pinot Noir (NZ $29.90) – smells like flowers and raspberries; very fresh, very nice. Fruity, but enough acidity and astringency to save it from being just fruity, with helpful if unexpected tannins on the finish. If not wildly complex, a pleasant iteration of a light, cool-climate style Pinot Noir. 

On Palate Press: Playing with yeasts isn’t quite science (yet)

My January column at Palate Press gives some thought to how new, non-Saccharomyces winemaking yeasts (and new combinations of those yeasts) are like faux “antique” furniture: achieve that natural look, but without all the time and risk of a spontaneous fermentation (or actually letting your farm table be beaten up by generations of rambunctious youngsters). While some very nice microbiology is exploring what minor-player yeasts can do, actually using them — and predicting the results — is still as much (or more) art as it is science.  Find it here.

Arsenic, BeverageGrades, and the Power of Withholding Information

UPDATE: Kevin Hicks brought a class action lawsuit against multiple California wineries for selling wine containing undisclosed high concentrations of arsenic in March 2015. I’ve posted about that here.

I received an unusual “story” by way of Palate Press this week. It looked like this:


Alcoholic beverage testing company, BeverageGrades has discovered lead
and arsenic present in wine at levels that exceed the Environmental
Protection Agency (EPA) standard Maximum Contaminant Levels for
drinking water.
Independent laboratory testing has been conducted on the top selling
white wines in the United States. Approximately one in every three
bottles tested was found to contain either arsenic or lead levels that
exceed the Maximum Contaminant Level.
Regular consumption of these elements, even at low levels, can present
a serious health risk.

Topic (wine, food, wine pairing, travel, etc.)
Wine Science

Suggested Title
High Levels of Arsenic and Lead Detected in Wine

Approximately one third of the top selling wines in the United States
found to contain lead and arsenic.

Your Website or Other Writing Sample

Preferred payment method:

Mailing Address (Check)
7000 Broadway
Suite 307
Denver, CO

PayPal email

BeverageGrades, it seems, is a young company creating a private database of beverage health information, summarized in their trademarked “grades.” They say that they run a bottle – of wine at the moment, of beer, spirits, and coolers in the future – through a pile of lab tests to create a profile for each product. Users can see calories, sugar, and carbohydrates per serving, plus a set of “grades” that are supposed to indicate how healthy a wine is overall, how “skinny” it is, how “pure” it is, how likely it is to trouble your allergies, and how likely it is to give you a headache. The problem with HealthyPour™, SkinnyGrade™, and the rest of these ratings is that if you’re looking for more information on how they’re calculated, you’re out of luck. New users have to click “I do” to having read BeverageGrade’s methodologies page, but those “methodologies” say nothing more detailed than “we take lots of measurements of lots of stuff.” And, rather than showing numbers for any of these measurements, wines are rated as average or better or worse than average.

Concerns about proprietary details aside, that’s not enough information. What in tarnation does “pure” mean? What counts as a “contaminant” or an “additive” in BeverageGrade-speak? How do they summarize how healthful a wine is when a whole pile of scientists are still trying to figure out how and why and whether wine is healthy? They list sodium and “vitamins and minerals,” but should I care about sodium in wine, and is wine a good source of “vitamins and minerals” in the first place?

Instead of seeking to educate consumers, BeverageGrades is patronizing them with overly simplistic branded products while hiding information. And that brings us back to arsenic.

Arsenic is found in wine. Recent studies in New Hampshire and France have shown higher arsenic concentrations (in toenails and urine) in wine drinkers versus non-wine drinkers (fish, beer, and Brussels sprouts were similarly implicated, as has been rice in studies based in Asian countries). Arsenic is found in water and soil; drinking water remains, in fact, our most significant source of arsenic exposure. Grapevines and some other plants seem to take up and concentrate it and other heavy metals. The FDA has set maximum limits on how much arsenic can be present in apple and pear juices in the US and is working on similar limits for rice; wine very well may be next on the list. If we stop here, things look bad. 

But stopping there means withholding important information. Saying that arsenic is found in wine is, crucially, different from saying that arsenic in wine poses a significant health risk. Arsenic in wine and arsenic as a health risk in general aren’t things we understand well yet. We don’t yet have a clear picture of how much arsenic is too much arsenic (the same is true for a lot of environmental toxins, including lead).

The several recent studies I found all indicate that arsenic in wine probably doesn’t pose a significant risk to most drinkers. The “probably,” “significant,” and “most” hedges in the previous sentence are all a nod to that yet-unsolved “how much arsenic is too much” problem. But an analysis of arsenic in wine, sake, and beer for sale in Central Europe concluded that it’s not a significant risk to consumers. A comparative risk assessment of fifteen “known and suspected human carcinogens” judged that ethanol was the only really significant carcinogen in alcoholic beverages; the rest, including arsenic, “may pose risks below thresholds normally tolerated for food contaminants.” In other words, these fish are too small to bother frying.

I suspect that BeverageGrades isn’t counting alcohol itself against a wine’s HealthyPour™ rating – though I don’t know for sure, since they’re not sharing that information. In any case, I appreciated a comment made by one reader of a popular press story about arsenic in wine, who suggested that not just drinking, but eating and breathing and sleeping were hazardous to our health, too, and that we should probably stop doing all of those things. Wine, unlike breathing, isn’t a necessity. But if we try to eliminate every risk from our way of life, we’ll end up not living at all. 

Wine’s environmental impact: the bottle hits hard (but so does the fertilizer)

The book The 100-Mile Diet by Alisa Smith and J. B. MacKinnon opens with a glorious British Columbia meal made entirely from ingredients grown outside the door of an isolated cabin in the woods. Except the wine. The wine is from Australia.

Locavorism and the sustainable food culture has long since turned its gaze — and its mouth — to drinkables, with the Drink Local movement and a plethora of sustainable wine associations and alliances and the wine selection at Whole Foods attesting to its success. If you don’t care about the earth-friendliness of your wine, you at least know that there are people who do care.

A lot of environmentalist action seems intuitive. It took more energy to convey a Sangiovese from Umbria to my table than this bottle of Washington Carménère that I picked up from the winery on my way home from grocery shopping last Saturday. And I’m better off buying this bunch of organic radishes from the lady who grew them at the community farmers’ market than industrial radishes from a massive farm in Mexico. Duh. But decisions about wine are rarely so obvious, mostly because the vast majority of wine lovers are unsatisfied with the prospect of only drinking local wines. Like the 100- or 50- or 10- mile diet — a regimen of eating only foods produced within a limited radius of their kitchens — it can be done, but not without missing things like coffee and chocolate and pepper and olive oil and rice…and Australian Shiraz, evidently. It’s telling that so many farm-to-table restaurants still serve primarily non-local wines.

What’s a savvy, environmentally-conscious oenophile to do? Data to the rescue.

A number of studies — more than a number, really — have examined the environmental impact of pieces of the winemaking process or elements that contribute to shipping a final bottle, mostly in terms of carbon dioxide emissions. A few have looked at the entire winemaking process, but generalized for an imaginary average wine from a particular region. Fusi, Guidetti, and Benedetto aimed for a more specific and precise picture by performing a complete life cycle analysis — from “cradle” with planting the winery to “grave” with bottle disposal — of a single wine: a Sardinian Vermentino made by Sella and Mosca, the area’s largest producer. A life cycle analysis (LCA), considered the gold standard for evaluating a product’s impact, tracks a product from creation through to distribution (cradle to gate, a partial LCA) or disposal (cradle to grave, a complete LCA) and sums the impact of everything that goes into making that product. It also considers multiple forms of environmental impact beyond just CO2 emissions to potential acid rain production, eutrophication (dumping nutrients into water, causing algae blooms and disrupting the ecosystem), ozone depletion, water use, and so forth.

Smith and MacKinnon found that their simple-seeming 100-mile diet became a lot more complicated when they considered everything that went into food production. If the cow down the street in Illinois is being raised on grain grown in Washington, is the cow’s milk still part of a 100-mile diet on the street in Illinois? The number of fiddly little inputs — the lubricating oil used on the tractor used to help plant the vines, for example — that contribute to Fusi and company’s analysis is impressive. The list of different types of emissions they considered, to soil, to water, and to air, at every stage of vine planting and growing and winemaking, is impressive and moderately unpronouncable. The authors had to make some assumptions when precise data just wasn’t available — how much of the electricity used by the entire winery could be attributed specifically to the production of the Vermentino, for example — but they do seem to have been thorough.

Results confirmed what we already know from other studies, but they serve as both a good confirmation and a good reminder.

First: glass bottles are the most significant source of a wine’s environmental impact in every regard save ozone depletion. Machinery that goes into maintaining the vineyard wins in causing ozone depletion.

Second: What the authors call “the agricultural phase” — the work of planting and maintaining the vineyard before grapes are harvested and winemaking begins — is a major contributor to use of fossil fuels, water pollution, global warming potential, and ozone depletion. Maintaining the vines caused more harm in most categories than planting them, but vine planting still accounted for a third or more of the overall impact of the agricultural phase almost entirely because of diesel fuel used to prepare the land and lay down trellises. More ozone depletion was caused by the production of fertilizers used in the vineyard than by any other component of the winemaking process.

And this is at a winery that has won awards for its environmentally friendly practices. A similar LCA of a Portuguese vinho verde (made at a winery with presumably less scrupulous vineyard practices) and an imaginary average white from Ribeiro showed that agricultural inputs — essentially fertilizers, pesticides, and machinery — were the number one cause of environmental impact in every category.

Third: What about distribution? Shipping the wine from Sardinia to the US was a major source of potential acid rain production, and a smaller but still significant source of other types of pollution. Shipping the wine from Sardinia to anywhere else in Italy never accounted for more than 5% of any impact measure. Shipping to the rest of Europe landed somewhere in the middle.

The most important ways in which the average wine is hurting the environment? Making the glass bottle, diesel used to prep and maintain the vineyard, fertilizer production, and shipping the bottle overseas if it needs to go that far to get to you.

The ramifications in practical terms for that savvy, environmentally-conscious oenophile: favor wineries practicing low-input viticulture, especially those eschewing synthetic fertilizers and using horses instead of diesel. Take advantage of wine on tap when it’s available. And yes, drink local.

In real terms, I can’t imagine that oenophile — let’s say she’s living in San Francisco — not drinking the occasional Sardinian Vermentino, or any other interesting wine from Europe or South Africa or New Zealand. But maybe her standbys, if she has them, can come from closer to home. We have good data to say that Drink Local and low-input viticulture are more than just marketing schemes.

**For the record, in light of articles I’ve written on the environmental impact of wine closures (here and here), the wine in this study was closed with a cork  which, in the grand scheme of things, wasn’t important enough to warrant discussion.