This is an edited extract from a chapter in my book The New Viticulture. The brief message? Smoke exposure is a big problem for vineyards, especially later in the growing season. Fresh smoke is very bad, but the smoke isn’t so bad if it has travelled a long way. One of the key issues is testing: you want to know whether it is worth picking the grapes (a cost) and making the wine (another cost), and if you are buying grapes, you don’t want to be buying smoke tainted ones. But it’s very difficult to test for this.
Since I wrote this, I’ve come across some interesting work, including some posters from the last Australian Wine Industry Technical Conference in Adelaide last year:
[This is a picture of Mayacamas in the Napa Valley: the effect of the fires from 2020 can still be seen in this beautiful mountain landscape]
Here’s the article:
Consider the smell of smoke. It can be very comforting: think of sitting round a campfire, or logs burning in the grate of a country pub in winter. But in the wrong context it can be terrifying - the sign that something is badly wrong and danger is at hand. We have a similarly ambivalent relationship to the taste of smoke in our food and drink. Some people prize the distinctive bonfire and peat aromas of Islay whisky or German Rauchbier, while others dislike it. Rather more enjoy the taste of smoked salmon. And the flavour of barbecued beef brisket cooked for hours in a smoker is pretty much universally loved. But smoky wine? It has fewer followers.
Many wine regions are in areas that, from time to time, experience bushfires. These have been in the news a lot recently. In December and January 2020, Australia suffered devastating bush fires. ‘Australia is prone to fire,’ said Wine Australia’s Stuart Barclay, addressing journalists in London in a briefing about the bushfires as part of the Australia day tastings. ‘It is part of the natural landscape. But this year has been a monumental impact. 15 years of drought has made the country dry, and the fires have been almost impossible to manage.’ Barclay emphasizes that while 1500 hectares of vines are within fire-affected zones, this represents just 1% of Australia’s vineyard area. The Adelaide Hills, though, was particularly badly hit, with 648 hectares affected. ‘This ranges from complete devastation to partial devastation,’ says Barclay. ‘A major issue is the potential impact of smoke on grapes and wine.’
Australia isn’t the only country that has experienced severe bush fires in recent years. In October 2017 and October 2019 California had fires on a major scale that involved the destruction of some wineries and vineyards, although many of the grapes had already been harvested at this stage. Canada’s Okanagan Valley has also had issues with smoke tainted grapes in recent years. Forest fires in Portugal in June 2017 impacted vineyards. In light of the human and wildlife impact of these natural disasters it seems trivial to be thinking about the impact of bush fires on wine production, but the issue of smoke taint is something that is concerning many within the wine industry. And both those growing grapes and those buying grapes want a way of measuring whether or not grapes that have been exposed to smoke will produce tainted wines.
As a country where this is a recurring problem, Australia has led the way in this research. The Australian Wine Research Institute (AWRI) began looking at this issue in earnest in 2003 after a bad bush fire season. Of particular interest are the following questions. What are the components in smoke that can taint wines? Which are the vulnerable periods in the vine’s life cycle, where smoke exposure has the potential to cause problems? And how can you test grapes to see whether they are affected or not?
The smoke taint compounds
This first set of studies concluded that smoke taint was a real problem, with wines from smoke-affected vineyards displaying characters described in sensory work as “smoky,” “burnt,” “ash,” “ashtray,” “salami,” and “smoked salmon.” They also identified two compounds as key culprits: guaiacol and 4-methylguaiacol, although there probably were other chemicals involved, albeit at lower concentrations. They also showed that the taint compounds were found within the skins of the grapes, but not in the pulp or in the surface waxes on the skins.
Interestingly, it turns out that the chief chemical culprits in smoke taint are already present in many wines. Guaiacol and methylguiacol are formed by burning lignin, the structural component of wood. And when barrels are heated with fire to bend the staves, this results in the production of a number of flavour compounds from the wood that then find their way into wine. The concentrations typically found in barrel fermented or aged wines seem to be fine; It’s only when they are present at higher levels that they cause problems.
Further studies looked at the timing issue, and found that smoke exposure at any time between veraison (when the berries change colour and enter the final stage of ripening) and harvest can cause smoke taint, the most sensitive period being a week after veraison and afterwards. It is also worth noting that repeated smoke exposure builds up the level of smoke taint, as might be expected. The take-home message: the less smoke the better. Evidence suggests that the smoke taint compounds are absorbed by the skin of the berries (thus bigger, more developed berries are more at risk), where they bind to sugars in the plant tissue in a process known as glycosylation.
This is important, because it means that if you taste grapes that have just been picked, you can’t taste these bound smoke taint compounds. The sugars have to be cleaved off, either by enzyme activity or hydrolysis. So the only way to know whether grapes are smoke affected is to do a microvinification and then taste the results. But if there was a way to know just after the smoke exposure whether the grapes are tainted or not, this could save growers a lot of money: the work that is needed to farm the vines if the crop is going to be picked is a lot more expensive than just keeping the vines alive, and if the resulting grapes are tainted, this is money badly spent. And for insurance purposes, or a dispute between a grower and a winery, a chemical analysis would be much easier than arguing over a sensory analysis.
Smoke composition
Smoke is chemically complex, with an estimated 500 aroma compounds present in its composition. Smoking food has been a traditional means of preservation, dehydrating meats and fish while at the same time inhibiting bacterial growth. The main chemicals associated with the smoke flavour in affected wines are guaiacol, 4-methylguaiacol, syringol, p-cresol, o-cresol and m-cresol. Initially, one thought was that it might be possible to wash the grapes while they were still on the vine to get rid of the smoke compounds so they didn’t end up in the wine. In a study from the Australian Wine Research Institute (AWRI) in 2003, attempts were made to wash fruit in various ways. The treatments tested were cold water, cold water plus wetting agent, warm water, cold water plus 5 percent ethanol, and milk. None of these worked.
This is because these flavour compounds enter the grapes soon after exposure via the waxy cuticle around the berry. They enter the berry skin cells under the cuticle, and this is where things get interesting. These cells respond to smoke compounds by using a detoxifyication process called glycosylation, which involves binding a sugar molecule to them. ‘Glycosylation is a defence mechanism,’ says Eric Hervé of California’s ETS Laboratories, who have been at the forefront of looking at the science of smoke impact on grapes. ‘Smoke compounds have some toxicity, and so the defence is to bind them to sugars and make them water soluble so they can be translocated into the cell and inter them in the vacuoles where they won’t interfere with metabolism.’
This creates an immediate problem. As well as making the smoke compounds less toxic, glycosylation also takes away their aromatic potential: they stop smelling.
So if winemakers want to see whether their grapes are affected, then they can’t just sample them; something needs to happen to cleave the sugars off so that their aromatic potential can be revealed. And this is exactly what yeasts do during fermentation, via a process called hydrolysis. ‘It looks like it is yeast driven and mostly in the first days of fermentation,’ says Hervé. For this reason, the AWRI recommend testing smoke-affected grapes for a panel of smoke-related compounds, and then doing a sensory assessment of a small-scale ferment. This is more accurate the closer to the actual harvest date it takes place: they recommend a couple of weeks before harvest, which gives enough time for the results to be available so that the right action can be taken.
What about analysing for the glycosylated compounds? It turns out this is not so straightforward. ‘Our current investment in measuring these glycosylated compounds exceeds a million dollars in hardware,’ says Gordon Burns of ETS. ‘It is entirely a speculative investment that as we go down the road we will be able to serve the industry better. We are building that very necessary background information about native levels of all these compounds.’ In the meantime, they believe that looking at the levels of two of the free volatiles, guaiacol and 4-methylguaiacol, is the best way to go. Even though most of these compounds will be glycosylated, looking at the small amount that is free gives a good indicator of whether the grapes are smoke-impacted. ‘We have been doing this for 10 years now,’ says Hervé. ‘Each year the fire events are different. The timing of exposure to smoke is different. The fuel is different. But over the years we haven’t seen much difference in terms of the numbers. Typically, guaiacol is three to five times higher than 4-methylguaiacol. This is a good signature that it is coming from smoke.’
‘We can always add more compounds,’ says Hervé, ‘but the problem is that the throughput can suffer, when we need a fast screening tool. And I’m not sure it will improve the prediction that much. When we analyse grapes exposed to smoke, based on the results we can’t say that there is no risk, but we can say that the risk is low to minimal, or significant but still moderate, or high, or very high. That’s the best that we can do.’
How to treat affected grapes
So what can be done if the grapes look to be affected? Winemakers have a few options.
The first is to try to minimize any transfer of smoke compounds from the grape skins into the must. This is easier for whites and rosés than it is for reds, which rely on extended skin contact for colour and structure. So hand harvesting is a good first step, because it minimizes skin maceration. It’s also important to make sure that there are no leaves in the harvest bins, because these also have smoke compounds in them. Fruit should be kept cool, because this results in less extraction, and it’s ideal to press the intact whole bunches. The free run (the first 400 litres per ton from the press) should be kept separate because this will have fewer smoke compounds. But this is just applicable to whites and rosés.
For reds, the received wisdom is that shorter extraction time is better than longer, but ETS dispute this. ‘Smoke compounds are mainly in skins so if you prevent skin contact you minimize the problem,’ says Hervé. ‘This works beautifully for whites or rosés, but for reds this doesn’t work in our experience. The smoke compounds migrate into fermenting juice extremely quickly.’ They have studied this. ‘In all the red ferments we have followed, the maximum number of smoke markers are already extracted after three days. Trying to shorten the fermentation with reds is actually making the situation worse because it is a lighter wine that shows up smoke impact more clearly.’
There are some other winemaking techniques that can be used to mitigate smoke impact, but they all have their downsides. One is to fine with a product called activated carbon, which is already used in some winemaking settings. Carbon grabs hold of organic materials and removes them: for this reason it’s widely used during filtration in water treatment. The problem is, it isn’t particularly selective. It can be used for treating juice or wine, but it has a negative effect on other aroma compounds. Another approach has been to use reverse osmosis, which is a specialized filtration process, coupled with a technique called solid-phase adsorption. Reverse osmosis takes out a permeate fraction from the wine that contains water, acids, alcohols, and chemicals below a certain molecular point cutoff. This fraction is then treated alone with solid-phase resorption to take out the smoke compounds before returning it to the wine. It removes some beneficial aroma compounds too, but the hope is that because just this small fraction of the wine is treated, its negative impacts will be limited. One of the problems with this approach is that while treatment has been effective in some cases, the smoke impact came back after a while, presumably because of the ongoing presents of some glycosidases that later hydrolysed to reveal smoke compounds that weren’t removed, so wines treated this way have to be drunk quickly.
Chris Carpenter of Jackson Family Wines, who makes the wines of Lokoya, Cardinale, La Jota and Mt Brave in Napa, has had some experience with smoke impact. In his experience, different varieties show the smoke differently. ‘Merlot, Petit Verdot and Malbec were crushed by it,’ he says. ‘Cabernet Sauvignon and Cabernet Franc fared relatively better.’ He adds that Syrah is a problem for analysis, ‘because many of the compounds you are testing for in smoke occur naturally in Syrah.’ After crushing affected grapes, Carpenter added inactivated yeast hulls, which are supposed to soak up some of the smoke residue, but he’s not sure whether this helped. ‘At the time any of the scrubbing technologies that were available were not able to scrub bound compounds,’ he says. ‘Some of the companies that are working on that have now purported to be able to deal with more of the bound complexes but there is no independent science that is backing those technologies.’ One of the difficulties he finds in assessing smoke-impacted grapes is how quickly these compounds affect your sense of taste and smell. ‘After two or three wines that are tainted your pallet is dead and you need to reset or you lose the ability to taste anything,’ he says. This complicates assessing whether wines have a problem or not. ‘We still do not understand the dynamics of smoke taint, the compounds and their evolution over time,’ says Carpenter. ‘There is no solution that we have yet to be presented. If you bottle wines that you suspect have smoke in them you are taking a huge risk.’ He also thinks it’s a huge risk even to bottle wines from grapes that were on the vine during smoke events, even if they don’t show any effect, because they could in the future. ‘We ended up scrapping two-thirds of my inventory from 2017 because the risk of bottling those wines was too great.’ So no Lokoya 2017 was produced.
Dan Petroski of Larkmead has experience processing potentially smoke-affected grapes. ‘In 2017, after our own harvest was complete, we crushed 40 tons of grapes for a vineyard client whose winery was in a quarantined zone and the grapes sat through a week of the fires and smoke exposure on the valley floor of Napa,’ he recalls. As part of the process, in order to support the client, they did sensory evaluation and chemical analysis along the way. ‘Thankfully, the client’s grapes did not have perceptible smoke issue from a chemistry or sensory perspective. We don’t know why in this particular instance this is true for us and not for others. But this is all I know.’ He also had a look at some wines that were in barrel to see whether any taint was seeping into the wine from the smoky air. ‘The new French oak barrels of the aging wine (2016 vintage) had barrel char similar to what we would normally see in a barrel aged wine with no perceptible smoke impact.’
One winery that did end up bottling some wine from smoke-affected grapes was Anthill Farms in Healdsburg. This single-vineyard Pinot specialist had smoke in Anderson Valley in 2008. ‘We declassified all the wines, carbon filtered the good lots and blended into one wine of about 200 cases that we sold to our customers with this explanation, I believe for $25,’ says co-owner Anthony Filiberti. ‘It was good wine, not amazing. Most people were happy and enjoyed the wine.’ However, some people customers complained and posted their thoughts on various wine bulletin boards, so it ended up being bad publicity for the winery. Filiberti says that he still drinks the wine from time to time, opening bottles at customer pick-up events. ‘Some taste pure and delicious, some taste like barbecue chips.’
Thanks for sharing. A lot of great research and work is being done on this. It's unfortunate tho that we have to do it.
Thanks for spreading the word on this important subject. The West Coast Smoke Exposure Task Force has compiled additional resources for grape growers and winemakers to utilize during smoke events that may also be useful to readers: https://www.wcsetf.org.