Inquiring Winemaker


Do You Know What's Inside Your Barrel?

March 2014
by Tim Patterson

Split up the harvest of a small vineyard among five boutique wineries, do the winemaking in all cases along similar lines, and no one will be surprised when the results in the bottle all taste and smell noticeably different. But open up five barrels from the same cooperage, with the same toast level, and your jaw may drop at the variation on display.


  • Research has shown a high degree of variability in the toast levels and aromatic compound content of oak barrels.
  • Because barrels are hand-made products, some degree of variation—even along a single stave—is inevitable.
  • Wineries can benefit from testing barrel aromatics, using the analysis ?to augment sensory evaluation.

The further you dig into barrel variation—between cooperages, between barrels from the same cooperage, between the staves in a single barrel, between the ends and the bellies of the staves—the more of it you find. Tanks are pretty much tanks, plus or minus a few bells and whistles; barrels are more like snowflakes. The next barrel you buy from Fils et Frères may not be like the last one, and the order from Quercus Expensivus may not really echo the one you used in your barrel trial.

This is not the warm-up for a flaming exposé of the shocking mendacity of the world’s great cooperages, a sordid tale of fraud in the woodlot. It’s a story of how hand-made, artisan products can’t and don’t come out identical. In fact, we probably don’t want barrels to be utterly uniform widgets. But the breadth of variation is something winemakers have to understand and keep an eye on, and every now and then, it might be worth the hassle to open the hood on one and see what’s inside.

Tom Collins, barrel detective
This column isn’t exactly ripped from the headlines, but it was inspired by a presentation given by Tom Collins, director of the Food Safety and Measurement Facility at the University of California, Davis, during an all-day session held at Davis in December to focus on winemaking measurements. Before coming to Davis, Collins worked for Treasury Wine Estates, where he launched a project to understand how barrel production and toasting affected oak aromatics, in order to better match cooperage practices with specific wine styles. His talk drew from both the results of the Treasury study and his own later work.

Barrel wood contains certain aromatic compounds from the start, just by virtue of being oak, like oak lactones (fresh oak and coconut) and others that get produced or amplified by toasting, like guaiacol (smoky, spicy). The duration and temperature of the toasting and barrel-assembly process are the main drivers of the aromatic compounds and properties of the final barrels, and so Collins and his crew inserted thermocouple sensors at various points along the length of beaucoup de staves and tracked them through the toasting process. Later, they took apart a number of barrels and tested wood scrapings—again, from different parts of different staves—and analyzed the aromatic compound content.

Since Collins is still working up his data for formal publication, I will spare you some of the charts, tables and statistical numerology. Let’s just say he found a ton of variation—a heap, a bunch, a bushel and a peck—at every stage from tree to finished barrel.

It all starts with the wood. Different trees—even ones raised right next to each other—go under the axe with slightly different chemical composition and tissue structure. Staves cut from different parts of the tree can differ, and how staves are cut and planed can leave them rougher or smoother, which affects how the wood later responds to heat.

And then there’s the fire part. Because the fire pot is down on the floor, and the staves are sticking up three feet into the air, it is guaranteed that the ends of the staves will get more heat treatment than the belly, which virtually guarantees every stave in every barrel will bridge at least two official toast levels and bring two sets of aromatic goodies to the party.

Different staves headed for the same barrel may also get slightly differently toasted. Even if the overall fire pot temperature is well-controlled, adding a piece of wood to the fire makes that side of the circle hotter (at least for a while) than the other side, generating more of compound X and less of compound Y.

Color, it turns out, is not a very reliable indicator of actual toast level and compound content. But the pictures Collins showed of all the staves from a single barrel lined up side by side, with the light and dark areas splattered around like a Jackson Pollack, were indeed worth a thousand words.

My favorite from Collins’ list of factors that can affect stave and barrel toast levels is what kind of ventilation system is in place in a particular cooperage and how well it is working that day.

If these are some of the factors that influence the supply of barrels, the demand for barrels can play a role, too. Mel Knox, barrel broker extraordinaire, distributor of both Francois Frères and Tarransaud, pointed out that if a major wine region like California has a short crop in a particular year, the supply of barrel wood will back up, meaning some staves will dry another year, changing their composition and toastability.

Are so-called oak alternatives more reliable? According to Collins, if they come from fire-toasted wood, then all of the above issues apply. If they come oven-roasted, there may be more uniformity—though ovens can have hot and cold spots, just like fermentation tanks.

Finally, in an era when heavy toast has considerable cachet, Collins points out that attaining toast precision becomes more difficult as the toast level gets higher. Lighter toasts can be made with more consistency; the line between medium-plus and heavy toast—both involving lots of heat and time—is particularly iffy, even for the output of a single cooperage.

Cooperage constraints
Even with the best of cooperage intentions, barrels are, in the original Middle-Frenc h sense of the term, manufactured: that is, hand made. They are not extruded from computer-controlled machines, which means every one is a little different from the one before and the one after.

You can, of course, see this as a good thing. Gordon Burns at ETS labs, which has tested oak aromatics for years (see more below), notes that elsewhere in winemaking, we applaud the fact that the raw materials that go into wine are varied and different, coming from distinct vineyards and vintages. Knox remembers a time when there was a certain fondness for barrel individuality, as though each one had a cooper’s signature on it (as some do); but as wineries have gotten bigger and more competitive, pressure for consistency has mounted.

Cooperages don’t produce barrel variation on purpose, or because they are too cheap or too incompetent to implement quality-control programs. Knox explained that his two major cooperages have detailed protocols for aging, air exposure, grain width and toasting time and temperature; barrels are, however, he says, “made by human beings.”

What’s a winemaker to do?
Even if we think of barrel variation as charming, or admit that it’s inevitable, it still poses practical problems for wineries. In particular, variation makes small-scale barrel trials, which nearly all wineries do nearly every year, a somewhat perilous exercise. If the same wine is put in four different medium-plus barrels, and a taste-off leads to ordering 50 of the winning wood, how would you know that your trial barrel wasn’t an outlier? How confident can you be that you won’t have more vanilla or less spice in 5,000 cases of wine two years from now?

The fact of variability leaves wineries with four basic coping strategies to choose from. The first and easiest to implement is to embrace variety, to celebrate the surprises that show up in the barrel room. Second is to emphasize the tried and true: If you have been getting good wood from a particular cooper or combination of coopers, stick with the program. A third approach, common in the industry, is to rely on safety in numbers, using barrels from several cooperages and multiple toast levels. If this year’s barrels from Quercus Expensivus don’t deliver enough caramel flavor, maybe the ones from Fils et Frères will pick up the slack.

And as Gordon Burns points out, numbers matter in another way. Most final wines are blended from multiple barrels, often dozens or hundreds, which can mean that the differences settle out and that winemakers can pick and choose from among lots of options. With multiple coopers or a single, monopoly cooperage, the aromatic impact of barrel aging, barrel by distinctive barrel, is always going to be critical for making final blends.

A fourth option, highly recommended by Collins and, of course, by ETS, is to do some testing of your barrel stock, thereby obtaining chemical profiles to go along with sensory evaluations. ETS does this by taking samples of wine—or, less often, wood—extracting volatiles from the headspace and analyzing their presence and concentration through GC/MS (gas chromatography/mass spectroscopy). The analysis focuses on nine “indicator” compounds that have the greatest sensory impact and are also good proxies for the dozens of other similar compounds that may show up. Tom Collins uses the same methodology and compound list; he originally had ETS do his testing, but now he has his own set of expensive equipment.

The levels of almost all of these compounds can be traced to toast level. Cribbing from the ETS website, here are a couple of examples:

Eugenol and Isoeugenol: Eugenol is the main aroma compound found in cloves. Present in raw oak, eugenol is reported to increase during open-air wood seasoning. Eugenol and isoeugenol possess a very similar spicy, clove-like aroma. Release into wine is reported to increase with toasting level.

Furfural and 5-Methylfurfural: Furfural and 5-methylfurfural result from caramelization of cellulose and hemicellulose during barrel toasting. They possess sweet, butterscotch, light caramel and faint almond-like aromas. They are also markers for the whole family of caramelization compounds.

Results of the tests are displayed in spider plots (see illustration), a graphical form that makes the relative strength of multiple components quite visually evident, facilitating the comparison of different samples. Systematic testing practices also make it possible to do year-to-year comparisons.

Burns says the number of wineries that are working with ETS on oak aromatics on a regular basis has increased in recent years, and he now has several dozen clients with regularized testing programs. Collins says it won’t hurt any winery to take a barrel or two apart every now and then and see where its wine has been living, sending some scrapings off for testing. He claims it isn’t that hard to put the barrels back together again, once you get the hang of it.

So, lots of options for handling the barrel challenge. The only thing winemakers cannot do is blithely assume a barrel is a barrel.

Tim Patterson is the author of “Home Winemaking for Dummies.” He writes about wine and makes his own in Berkeley, Calif. Years of experience as a journalist, combined with a contrarian streak, make him interested in getting to the bottom of wine stories, casting a critical eye on conventional wisdom in the process.

Currently no comments posted for this article.