Wine aroma is arguably the most important intrinsic factor used to judge wine quality. The perception of wine aroma is the result of a number of different factors including the composition of volatile compounds in the wine, the perceptual interactions between the volatiles, the physical and chemical effects of non-volatile components of the wine matrix, and the ability and experience of the person perceiving the aroma.
Compounds, some of them malodorous, can enter wine from different sources during the production process. For example, methoxypyrazines are part of primary aroma, meaning that they are naturally present in some grape varieties, and are extracted from the grapes during winemaking. On the other hand, esters are part of the secondary aroma, which means that they are released by yeast during fermentation as metabolites from certain amino acids. Wine can also have tertiary aromas, for example, those deriving from oxidation compounds that form during long-term storage.
With climate change, heatwaves, and more frequent forest fires across the world, the effect of smoke on crops is viewed as increasingly important. When wine is made from smoke-affected grapes, smoke volatiles are transferred from the grapes during the winemaking process, and manifest in the wine as an unpleasant burnt, ashy aroma called “smoke taint.” Some of the main culprits in smoke tainted wines are volatile phenols (VPs). This group of compounds can also be derived from a number of other sources including toasted oak barrels, microbial fermentations (specifically Brettanomyces spoilage), but they have become mostly associated with off-flavors in wine made from smoke-affected grapes.
The usual way to deal with problematic aroma compounds is to quantify them by chemical/analytical means. The possible impact on human perception is then evaluated using Odour Threshold (ODT, the lowest concentration of a compound that is perceivable by smell) values and Odour Activity values (OAVs, the ratio between concentration and ODT). If the measured compounds are present at concentrations above their ODTs, their OAV is higher than 1 and they are considered to have an impact on the smell. In other words, we can smell a compound present at a concentration higher than its ODT.
Like any other aroma compound, VPs can be quantified by analytical means. Previous research has established that, at higher levels, certain VPs contribute to a continuum of smoke taint related off-flavors including “burnt,” “bretty,” “smoky,” and “ashy.” At subthreshold levels, VPs are generally accepted as being benign to wine aroma.
Increasing importance is being placed on understanding the effects of smoke on agricultural crops, and there are numerous articles available on the effects of smoke on grapes and wine. Most of these reports are the result of experimental conditions looking into replicating field conditions in a reproducible manner in small batches, including micro-vinifications. In contrast, research into the composition of commercial wines affected by naturally-occurring (i.e. not experimentally induced) smoke events is still rare. As such, we have decided to investigate the impact of bushfires on the chemical (VP concentrations) and sensory profile of commercially-produced red wines.
Twelve wines submitted by the industry as potentially smoke-tainted since the grapes have been exposed to bush fires close to harvest time. The wines were screened for a broad range of VPs using a specially developed GC-MS method. The wines were also characterized using Descriptive Analysis by a sensory panel highly trained in smoke-taint evaluation. In the second stage of the project, the results were compared statistically to see if the chemistry results did actually support the findings from the sensory evaluation. If so, simply submitting samples for chemical analysis of VPs would be enough to establish if a wine is smoke tainted or not. We also investigated historical data concerning fire events in the regions from which the wines originated, and during the relevant vintages.
Chemically, the wines submitted for evaluation contained different VPs and a wide range of VP concentrations. This was promising for the sensory work, as now we knew that the set had good chemical variability. On the sensory side, it was notable that out of twelve wines, the four that were described with the most negative attributes, at significantly higher levels than the others, were all from regions that had experienced severe fire events prior to harvest. Certain attributes (“smoky,” “ashtray”) in some of the wines could be attributed simply to levels of specific or combinations of VPs at concentrations higher than their respective ODT.
Looking more into detail, there were some unexpected findings. Guaiacol (one of the compounds linked to smoke taint) was present in the majority of samples at or above ODT; as the wines had been submitted by industry for suspected smoke taint, this result was not surprising. It was surprising though that guaiacol did not cause the perception of “smoke” in any of the wines unless it was in combination with other phenols. Even more, combinations of compounds (for example, cresols and xylenols) at subthreshold levels led to unexpected sensory effects (“earthy/dusty,” “chemical,” and “tar/burnt rubber”).
Using the combined dataset (chemistry and sensory), we started to clearly see that concentration and composition of VPs correlated with certain sensory attributes in the wines. There were (sometimes unexpected) links between VPs and specific off-odors. Some of these links were previously demonstrated, others we could not predict based on existing knowledge. Wines with very low (sub-threshold) levels of VPs showed fruity and sweet-associated characteristics, and those with supra-threshold levels showed negative attributes. In some cases, sensory effects (“earthy/dusty/potato skin,” “moldy/musty,” and “cooked veg”) could not be attributed to a certain concentration of VPs based on their ODT, but may have been due to combinations of volatile phenols at subthreshold levels, possibly influenced by the presence of other compounds.
The idea of interactions with masking and synergistic effects is not conceptually new. We knew that perceptual interaction phenomena between aroma compounds in red wines represent an important source of complexity. It has also been already demonstrated in studies on other compounds (thiols, esters, methoxypyrazines, terpenes, etc.) that sensory results do not always correlate well with predictions that are based on the chemistry of the solution.
However, our findings did highlight a number of issues. Starting with the simplest, ODT values found in the literature are usually determined in model solutions and are not comparable to those determined in wine. Odor thresholds in any matrix other than the study matrix may be irrelevant. A review of the literature revealed that studies in wine can use inappropriate ODTs for work carried out in a new wine matrix, and matrix effects are often ignored. More to the point, many VPs (as well as other compounds) have not yet been characterized in any wine matrix or assigned threshold values in any formal sensory study. The aspect of “matrix effect” is also significant in the context of our work, as the commercial wines were diverse, i.e. different chemical matrices as dictated by cultivar-related grape genetics, ripeness levels, winemaking procedures, and bottle storage. A significant gap exists in the literature with regards to matrix effects on VPs, specifically the effect of the VP composition on different cultivar aroma profiles.
Our study also highlighted the fact that that aroma compounds interact at threshold and subthreshold levels and cause aromatic changes to wine, unpredictable from the effects of a single compound, and certainly not predictable from the many peri-threshold studies on aroma compounds. So even if the ODT is accurately determined in the wine matrix of choice, it is still possible to find sensory effects when the compound is present at concentrations below threshold, especially in the presence of certain other compounds (see the example of guaiacol above).
For us, this study emphasizes the importance of understanding the effects of problematic compounds like VPs on wine and of increasing awareness of the interactions and synergies between them even at levels where they are usually considered benign. Crucially, low levels of VPs may cause or exaggerate negative odor attributes in red wines, which has implications for winemaking with smoke-affected grapes. This new knowledge should prove useful to wine industries world-wide that are increasingly being affected by smoke taint. Ultimately, the aim is to produce wines that satisfy the rigorous demands of the global market despite the escalation of wildfires globally.
These findings are described in the article entitled Profiling Potentially Smoke Tainted Red Wines: Volatile Phenols and Aroma Attributes, recently published in the South African Journal of Enology and Viticulture.