Discovery of Potent Glycosidases Enables Quantification of Smoke-Derived Phenolic Glycosides through Enzymatic Hydrolysis.

When grapes are exposed to wildfire smoke, certain smoke-related volatile phenols (VPs) can be absorbed into the fruit, where they can be then converted into volatile-phenol (VP) glycosides through glycosylation. These volatile-phenol glycosides can be particularly problematic from a winemaking standpoint as they can be hydrolyzed, releasing volatile phenols, which can contribute to smoke-related off-flavors. Current methods for quantitating these volatile-phenol glycosides present several challenges, including the requirement of expensive capital equipment, limited accuracy due to the molecular complexity of the glycosides, and the utilization of harsh reagents. To address these challenges, we proposed an enzymatic hydrolysis method enabled by a tailored enzyme cocktail of novel glycosidases discovered through genome mining, and the generated VPs from VP glycosides can be quantitated by gas chromatography-mass spectrometry (GC-MS). The enzyme cocktails displayed high activities and a broad substrate scope when using commercially available VP glycosides as the substrates for testing. When evaluated in an industrially relevant matrix of Cabernet Sauvignon wine and grapes, this enzymatic cocktail consistently achieved a comparable efficacy of acid hydrolysis. The proposed method offers a simple, safe, and affordable option for smoke taint analysis.

Volatile phenols in wine: overview of origin, formation, analysis, and sensory expression.

Volatile phenols impart particular aromas to wine. Due to their distinctive aroma characteristics and low sensory thresholds, volatile phenols can easily influence and modify the aroma of wine. Since these compounds can be formed in wines in various ways, it is necessary to clarify the possible sources of each volatile phenol to achieve management during the winemaking process. The sources of volatile phenols in wine are divided into berry-derived, fermentation-derived, and oak-derived. The pathways and factors influencing the formation of volatile phenols from each source are then reviewed respectively. In addition, an overview of the sensory impact of volatile phenols is given, both in terms of the aroma these volatile phenols directly bring to the wine and their contribution through aroma interactions. Finally, as an essential basis for exploring the scientific problems of volatile phenols in wine, approaches to quantitation of volatile phenols and their precursors are discussed in detail. With the advancement of analytical techniques, more details on volatile phenols have been discovered. Further exploration is worthwhile to achieve more detailed monitoring and targeted management of volatile phenols in wine.

Smoky Characters in Wine: Distinctive Flavor or Taint?

The frequency of wildfires has significantly increased in recent years, posing concerns for many grapegrowers and winemakers. Exposure of grapes to smoke can result in wines with notable smoky notes, which in severe cases are described as "smoke tainted". However, smoky aromas in wine are not a priori quality defects but may be considered desirable in some styles of wines, as also widely found and appreciated in many spirits. In this perspective, we summarize recent research on sources and assessment of smoky sensory attributes in wine and provide an outlook on opportunities for managing excessive smoky characters.

Influence of saliva on the sensory properties of US commercial smoke affected wines: Preliminary findings.

Previous research has suggested that the impact of smoke affected wines require human evaluation due to in-mouth changes in perception, perhaps associated with saliva. Smoke affected wines (n = 36) from three major wine growing regions in the US were sourced from commercial wineries. A subset of these wines (n = 7) were evaluated by a consumer panel (n = 57) and electronic tongue (e-tongue) to determine the influence of saliva in the sensory profile. Consumers assessed the wines for aroma and other sensory attributes, before and after individual saliva addition. Pooled saliva from consumers was used to treat all wines obtained (n = 36) and then analyzed using the e-tongue. Results showed that saliva did not significantly alter the overall aroma, other than fruity or woody aroma liking by consumers (p > .05). However, the presence of saliva significantly lowered overall liking in both red and white wines that were affected by smoke (p ≤ .05). Consumers rated the subset of smoke affected wines below the "might purchase" category, indicating these wines were not considered acceptable by consumers. When individual pairs of smoke affected wines (before and after saliva additions) were assessed using the e-tongue, the device was able to differentiate the pairs, validating potential usefulness to discern wine changes, though the discrimination indices were moderate to low (68.8% to 11.9%). Based on these data, in human ratings of the aroma and appearance of smoked affected wines, saliva decreased overall liking, and this was somewhat distinguishable by e-tongue analysis.

Volatile Organic Compound-Based Predictive Modeling of Smoke Taint in Wine.

Smoke taint in wine has become a critical issue in the wine industry due to its significant negative impact on wine quality. Data-driven approaches including univariate analysis and predictive modeling are applied to a data set containing concentrations of 20 VOCs in 48 grape samples and 56 corresponding wine samples with a taster-evaluated smoke taint index. The resulting models for predicting the smoke taint index of wines are highly predictive when using as inputs VOC concentrations after log conversion in both grapes and wines (Pearson Correlation Coefficient PCC = 0.82; R2 = 0.68) and less so when only grape VOCs are used (Pearson Correlation Coefficient PCC = 0.76; R2 = 0.56), and the classification models also show the capacity for detecting smoke-tainted wines using both wine and grape VOC concentrations (Recall = 0.76; Precision = 0.92; F1 = 0.82) or using only grape VOC concentrations (Recall = 0.74; Precision = 0.92; F1 = 0.80). The performance of the predictive model shows the possibility of predicting the smoke taint index of the wine and grape samples before fermentation. The corresponding code of data analysis and predictive modeling of smoke taint in wine is available in the Github repository (https://github.com/IBPA/smoke_taint_prediction).

Rapid headspace solid-phase microextraction sheets with direct analysis in real time mass spectrometry (SPMESH-DART-MS) of derivatized volatile phenols in grape juices and wines.

Volatile phenols possess "smoky, spicy" aromas and are routinely measured in grapes, wines and other foodstuffs for quality control. Routine analyses of volatile phenols rely on gas chromatography - mass spectrometry (GC-MS), but slow throughput of GC-MS can cause challenges during times of surge demand, i.e. following 'smoke taint' events involving forest fires near vineyards. Parallel extraction of headspace volatiles onto sorbent sheets (HS-SPMESH) followed by direct analysis in real time mass spectrometry (DART-MS) is a rapid alternative to conventional GC-MS approaches. However, HS-SPMESH extraction is poorly suited for lower volatility odorants, including volatile phenols. This work reports development and validation of an HS-SPMESH-DART-MS approach for five volatile phenols (4-ethylphenol, 4-ethylguiacol, guaiacol, 4-methylguaiacol, and cresols). Prior to HS-SPMESH extraction, volatile phenols were acetylated to facilitate their extraction. A unique feature of this work was the use of d6-Ac2O as a derivatizing agent to overcome issues with isobaric interferences inherent to chromatography-free MS techniques. The use of alkaline conditions during derivatization resulted in cumulative measurement of both free and bound forms of volatile phenols. The validated HS-SPMESH-DART-MS method achieved a throughput of 24 samples in ∼60 min (including derivatization and extraction time) with low limits of detection (<1 µg L-1) and good repeatability (3-6% RSD) in grape and wine matrices. Validation experiments with smoke-tainted grape samples indicated good correlation between total (free + bound) volatile phenols measured by HS-SPMESH-DART-MS and a gold standard GC-MS method.

Appraising California Zinfandel Exposure to Wildfire Smoke Using Natural Product Phenolic Diglycoside Biomarkers.

Zinfandel grapes are ubiquitous in California and its wine quality could be negatively impacted from wildfire smoke. Thus, the occurrence of fires prior to grape harvest presents a persistent problem to both viticulture and enology processes. This is the first broad study on Zinfandel to investigate wine quality defects produced by natural wildfires. The project, guided by UHPLC separations and MS2 multiple reaction monitoring, involved measuring natural product phenolic diglycosides (PDs) bioaccumulated in grapes, and expands outcomes published in 2022 by our team (called the Santa Cruz Campaign, SCC). The plan was implemented by exploiting a panel of six marker PDs 1-6 and their deuterated analogues. Examined in the study were 24 different Zinfandel wines obtained from 2016 to 2021 vintages of nine different American Viticulture Areas (AVAs) that were also within five of the eight California Zinfandel viticulture zones. The goal was to extend understanding on PD variations using patterns that possibly change as a function of appellation and fire intensity. Preliminary data was obtained to examine the relative amounts of PDs localized in berry skin versus pulp. The baseline of <15 ppb was proposed by surveying 18 distinct unsmoked Zinfandel wines. It was proposed to estimate the smoke impact on other Zinfandel wines by using seven PD ppb concentrations categories. A pilot study was also launched to assess conclusions by comparing ppb-based ratings versus sensory evaluation quality estimates. General findings presented herein should provide an important foundation to build understanding of using PD patterns to forecast possible Zinfandel wine wildfire damage.

Correlating Sensory Assessment of Smoke-Tainted Wines with Inter-Laboratory Study Consensus Values for Volatile Phenols.

Vineyard exposure to wildfire smoke can taint grapes and wine. To understand the impact of this taint, it is imperative that the analytical methods used are accurate and precise. This study compared the variance across nine commercial and research laboratories following quantitative analysis of the same set of smoke-tainted wines. In parallel, correlations between the interlaboratory consensus values for smoke-taint markers and sensory analyses of the same smoke-tainted wines were evaluated. For free guaiacol, the mean accuracy was 94 ± 11% in model wine, while the free cresols and 4-methylguaiacol showed a negative bias and/or decreased precision relative to guaiacol. Similar trends were observed in smoke-tainted wines, with the cresols and glycosidically bound markers demonstrating high variance. Collectively, the interlaboratory results show that data from a single laboratory can be used quantitatively to understand smoke-taint. Results from different laboratories, however, should not be directly compared due to the high variance between study participants. Correlations between consensus compositional data and sensory evaluations suggest the risk of perceivable smoke-taint can be predicted from free cresol concentrations, overcoming limitations associated with the occurrence of some volatile phenols, guaiacol in particular, as natural constituents of some grape cultivars and of the oak used for barrel maturation.

Evaluating the Potential for Smoke from Stubble Burning to Taint Grapes and Wine.

It has been well established that bushfire/wildfire smoke can taint grapes (and therefore wine), depending on the timing and duration of exposure, but the risk of smoke contamination from stubble burning (a practice employed by some grain growers to prepare farmland for sowing) has not yet been established. This study exposed excised bunches of grapes to smoke from combustion of barley straw and pea stubble windrows to investigate the potential for stubble burning to elicit smoke taint. Increased levels of volatile phenols (i.e., chemical markers of smoke taint) were detected in grapes exposed to barley straw smoke (relative to control grapes), with smoke density and the duration of smoke exposure influencing grape volatile phenols. However, the sensory panel did not perceive wine made from grapes exposed to low-density smoke to be tainted, despite the presence of low levels of syringol providing compositional evidence of smoke exposure. During the pea stubble burn, grapes positioned amongst the burning windrows or on the edge of the pea paddock were exposed to smoke for ~15-20 and 30-45 min, respectively, but this only resulted in 1 µg/kg differences in the cresol and/or syringol concentrations of smoke-affected grapes (and 1 µg/L differences for wine), relative to controls. A small, but significant increase in the intensity of smoke aroma and burnt rubber flavor of wine made from the grapes positioned amongst the burning pea stubble windrows provided the only sensory evidence of any smoke taint. As such, had vineyards been located immediately downwind from the pea stubble burn, it is unlikely that there would have been any smoke contamination of unharvested grapes.

Rapid Analysis of Volatile Phenols from Grape Juice by Immersive Sorbent Sheet Extraction Prior to Direct Analysis in Real-Time Mass Spectrometry (DART-MS).

Poly(dimethylsiloxane)-based thin-film sorbent sheets (SPMESH) have previously been used for parallel headspace (HS) extraction prior to direct analysis in real-time mass spectrometry (DART-MS) for rapid quantitation of odorants in complex matrices. However, HS-SPMESH extraction is poorly suited for less volatile odorants, e.g., volatile phenols. This report describes modifications to the previous SPMESH extraction device, which make it amenable to parallel extraction of low-volatility analytes from multiwell plates under direct immersion (DI) conditions. Optimization and validation of the DI-SPMESH-DART-MS approach were performed on four volatile phenols (4-ethylphenol, 4-ethylguaiacol, 4-methylguaiacol, and guaiacol) of relevance to the quality of grape juices. Negative-ion mode DART-MS spectra showed a series of oxygenated adducts [M + nO - H]- for all analytes, but isobaric interferences could be limited for three of the four analytes by selecting an appropriate MS/MS transition. Signal suppression from nonvolatiles (sugars, acids) could be overcome by a rinse step. DI-SPMESH-DART-MS analysis of 24 samples could be performed in ∼45 min (30 min extraction, 16 min DART analysis) with 0.5-3 µg/L detection limits in aqueous and model juice solutions. In real grape juices (n = 5 cultivars), good accuracy (72-137%) could be achieved for two of the four volatile phenols initially investigated, 4-ethylphenol and 4-ethylguaiacol. However, poor accuracy was observed for guaiacol in some cultivars, and 4-methylguaiacol could not be quantitated due to interferences with other volatile phenols. Despite these limitations, DI-SPMESH-DART-MS/MS may be useful for prescreening a large number of samples prior to more selective conventional analyses.

Evaluation of Different Interstimulus Rinse Protocols on Smoke Attribute Perception in Wildfire-Affected Wines.

Wildfires produce smoke that can carry organic compounds to a vineyard, which are then absorbed by the grape berry and result in wines with elevated levels of smoke-related phenols. These phenols have been found to have a large impact on the flavor of wines, being the cause of a smokey flavor with a lasting ashy aftertaste. When evaluating the sensory profile of these wines, there is an observed problem due to the lasting nature of these undesirable attributes and potential flavor carryover between samples. Through the use of standard and temporal attribute check-all-that-apply, this research desires to better understand the impact of smoke on the sensorial profiles of wines with various levels of smoke phenols (high, moderate, and low). Additionally, through the employment of different interstimulus protocols, the effectiveness of rinses on diminishing the smoke flavor in wines and optimal time separation were investigated. It was determined that a 1 g/L pectin rinse in between samples with a 120 s separation is optimal to ensure the removal of smoke attribute perception. This work also indicated the need to look deeper at the effects of the in-mouth hydrolysis of glyconjugate phenols that impact overall smoke flavor.

Development and Utilization of a Model System to Evaluate the Potential of Surface Coatings for Protecting Grapes from Volatile Phenols Implicated in Smoke Taint.

Due to the increasing frequency of wildfires in recent years, there is a strong need for developing mitigation strategies to manage the impact of smoke exposure of vines and occurrence of 'smoke taint' in wine. One plausible approach would be to prevent or inhibit the uptake of volatile phenols from smoke into grape berries in the vineyard. In this study we describe a model system we developed for evaluating under controlled conditions the effectiveness of a range of surface coatings (including existing horticultural sprays) for reducing/preventing the uptake of volatile phenols and their subsequent conversion to phenolic glycosides. Grapes were coated with the materials to be tested and then exposed to gaseous phenols, via evaporation from an aqueous solution, in a semi-closed glass container. Analysis of volatile phenols and their glycosidic grape metabolites demonstrated that the treatments typically did not provide any significant protection; in fact, some resulted in higher concentrations of these compounds in the grapes. The highest concentrations of volatile phenols and their glycosides were observed after application of oily, hydrophobic materials, suggesting that these materials may enhance the adsorption or transfer of volatile phenols into grape berries. Therefore, it is important to consider the types of sprays that are being applied in the vineyard before and during smoke events to prevent the potential of exacerbating the uptake of smoke compounds by grape berries.

Glycosylation of Volatile Phenols in Grapes following Pre-Harvest (On-Vine) vs. Post-Harvest (Off-Vine) Exposure to Smoke.

Taint in grapes and wine following vineyard exposure to bushfire smoke continues to challenge the financial viability of grape and wine producers worldwide. In response, researchers are studying the chemical, sensory and physiological consequences of grapevine smoke exposure. However, studies involving winemaking trials are often limited by the availability of suitable quantities of smoke-affected grapes, either from vineyards exposed to smoke or from field trials involving the application of smoke to grapevines. This study compared the accumulation of volatile phenol glycosides (as compositional markers of smoke taint) in Viognier and Cabernet Sauvignon grapes exposed to smoke pre- vs. post-harvest, and found post-harvest smoke exposure of fruit gave similar levels of volatile phenol glycosides to fruit exposed to smoke pre-harvest. Furthermore, wines made from smoke-affected fruit contained similar levels of smoke-derived volatile phenols and their glycosides, irrespective of whether smoke exposure occurred pre- vs. post-harvest. Post-harvest smoke exposure therefore provides a valid approach to generating smoke-affected grapes in the quantities needed for winemaking trials and/or trials that employ both chemical and sensory analysis of wine.

Glycosidically-Bound Volatile Phenols Linked to Smoke Taint: Stability during Fermentation with Different Yeasts and in Finished Wine.

When wine grapes are exposed to smoke, there is a risk that the resulting wines may possess smoky, ashy, or burnt aromas, a wine flaw known as smoke taint. Smoke taint occurs when the volatile phenols (VPs) largely responsible for the aroma of smoke are transformed in grape into a range of glycosides that are imperceptible by smell. The majority of VP-glycosides described to date are disaccharides possessing a reducing ß-d-glucopyranosyl moiety. Here, a two-part experiment was performed to (1) assess the stability of 11 synthesized VP-glycosides towards general acid-catalyzed hydrolysis during aging, and (2) to examine whether yeast strains differed in their capacity to produce free VPs both from these model glycosides as well as from grapes that had been deliberately exposed to smoke. When fortified into both model and real wine matrices at 200 ng/g, all VP-disaccharides were stable over 12 weeks, while (42-50 ng/g) increases in free 4-ethylphenol and p-cresol were detected when these were added to wine as their monoglucosides. Guaiacol and phenol were the most abundantly produced VPs during fermentation, whether originating from natural VP-precursors in smoked-exposed Pinot Noir must, or due to fortification with synthetic VP-glycosides. Significant yeast strain-specific differences in glycolytic activities were observed for phenyl-ß-d-glycopyranoside, with two strains (RC212 and BM45) being unable to hydrolyze this model VP, albeit both were active on the guaiacyl analogue. Thus, differences in Saccharomyces cerevisiae ß-glucosidase activity appear to be influenced by the VP moiety.

Compositional Changes in Smoke-Affected Grape Juice as a Consequence of Activated Carbon Treatment and the Impact on Phenolic Compounds and Smoke Flavor in Wine.

An increase in bushfires and wildfires globally and consequent smoke exposure of grapevines has seen an elevated need for remediation options to manage the impact of smoke taint in the wine industry. Two commercially available activated carbons (PS1300 and CASPF) were evaluated at 1, 2, and 4 g/L with juice from smoke-affected Pinot Noir and Chardonnay grapes. PS1300 and CASPF treatments removed up to 75 and 92% of the phenolic glycosides in the smoke-affected Pinot Noir rosé juice, respectively, and both carbons removed virtually all (i.e., 98-99%) of the phenolic glycosides in the smoke-affected Chardonnay juice at the highest dose rate (4 g/L). The free volatile phenols in the wines were similarly lower in concentration following treatment. Sensory analysis confirmed that the wines made from carbon fined juice had reduced smoke aroma and flavor compared to those from the nontreated controls. However, desirable sensory properties such as color and fruity attributes were also negatively affected by the treatment. The dose rate should be optimized in industry practice to find a balance between reducing the intensity of smoke-related sensory attributes while maintaining or enhancing positive attributes.

Large-Scale Reassessment of In-Vineyard Smoke-Taint Grapevine Protection Strategies and the Development of Predictive Off-Vine Models.

Smoke taint in wine is thought to be caused by smoke-derived volatile phenols (VPs) that are absorbed into grape tissues, trapped as conjugates that are imperceptible by smell, and subsequently released into wines as their free odor-active forms via metabolism by yeasts during fermentation. Blocking VP uptake into grapes would, therefore, be an effective way for vineyards to protect ripening grape crops exposed to smoke. Here, we re-evaluated a biofilm that had previously shown promise in pilot studies in reducing levels of smoke-derived VPs. A suite of nine free and acid-labile VPs were quantitated in Pinot Noir grapes that had been exposed to smoke after being coated with the biofilm one, seven or 14 days earlier. In contrast with earlier studies, our results demonstrated that in all cases, the biofilm treatments led to increased concentrations of both free and total VPs in smoke-exposed grapes, with earlier applications elevating concentrations of some VPs more than the later time points. Tracking VP concentrations through the grape ripening process demonstrated that some (phenol, p/m-cresol, and guaiacol) were not entirely sequestered in grapes as acid-labile conjugates, suggesting the presence of VP storage forms beyond simple glycosides. Free VPs in grapes, though a minor portion of the total, most clearly correlated with concentrations present in the resulting wines. Finally, red table grapes, available year round, were observed to replicate the effects of the biofilm treatments and were capable of transforming most VPs into acid-labile conjugates in under 24 h, indicating that they might be an effective model for rapidly assessing smoke-taint prophylactic products in the laboratory.

Compositional Changes in Grapes and Leaves as a Consequence of Smoke Exposure of Vineyards from Multiple Bushfires across a Ripening Season.

The negative effects of smoke exposure of grapes in vineyards that are close to harvest are well documented. Volatile phenols in smoke from forest and grass fires can contaminate berries and, upon uptake, are readily converted into a range of glycosylated grape metabolites. These phenolic glycosides and corresponding volatile phenols are extracted into the must and carried through the winemaking process, leading to wines with overtly smoky aromas and flavours. As a result, smoke exposure of grapes can cause significant quality defects in wine, and may render grapes and wine unfit for sale, with substantial negative economic impacts. Until now, however, very little has been known about the impact on grape composition of smoke exposure very early in the season, when grapes are small, hard and green, as occurred with many fires in the 2019-20 Australian grapegrowing season. This research summarises the compositional consequences of cumulative bushfire smoke exposure of grapes and leaves, it establishes detailed profiles of volatile phenols and phenolic glycosides in samples from six commercial Chardonnay and Shiraz blocks throughout berry ripening and examines the observed effects in the context of vineyard location and timing of smoke exposure. In addition, we demonstrate the potential of some phenolic glycosides in leaves to serve as additional biomarkers for smoke exposure of vineyards.

Potential Mitigation of Smoke Taint in Wines by Post-Harvest Ozone Treatment of Grapes.

When bushfires occur near grape growing regions, vineyards can be exposed to smoke, and depending on the timing and duration of grapevine smoke exposure, fruit can become tainted. Smoke-derived volatile compounds, including volatile phenols, can impart unpleasant smoky, ashy characters to wines made from smoke-affected grapes, leading to substantial revenue losses where wines are perceivably tainted. This study investigated the potential for post-harvest ozone treatment of smoke-affected grapes to mitigate the intensity of smoke taint in wine. Merlot grapevines were exposed to smoke at ~7 days post-veraison and at harvest grapes were treated with 1 or 3 ppm of gaseous ozone (for 24 or 12 h, respectively), prior to winemaking. The concentrations of smoke taint marker compounds (i.e., free and glycosylated volatile phenols) were measured in grapes and wines to determine to what extent ozonation could mitigate the effects of grapevine exposure to smoke. The 24 h 1 ppm ozone treatment not only gave significantly lower volatile phenol and volatile phenol glycoside concentrations but also diminished the sensory perception of smoke taint in wine. Post-harvest smoke and ozone treatment of grapes suggests that ozone works more effectively when smoke-derived volatile phenols are in their free (aglycone) form, rather than glycosylated forms. Nevertheless, the collective results demonstrate the efficacy of post-harvest ozone treatment as a strategy for mitigation of smoke taint in wine.

Techniques for Mitigating the Effects of Smoke Taint While Maintaining Quality in Wine Production: A Review.

Smoke taint has become a prominent issue for the global wine industry as climate change continues to impact the length and extremity of fire seasons around the world. Although the issue has prompted a surge in research on the subject in recent years, no singular solution has yet been identified that is capable of maintaining the quality of wine made from smoke-affected grapes. In this review, we summarize the main research on smoke taint, the key discoveries, as well as the prevailing uncertainties. We also examine methods for mitigating smoke taint in the vineyard, in the winery, and post production. We assess the effectiveness of remediation methods (proposed and actual) based on available research. Our findings are in agreement with previous studies, suggesting that the most viable remedies for smoke taint are still the commercially available activated carbon fining and reverse osmosis treatments, but that the quality of the final treated wines is fundamentally dependent on the initial severity of the taint. In this review, suggestions for future studies are introduced for improving our understanding of methods that have thus far only been preliminarily investigated. We select regions that have already been subjected to severe wildfires, and therefore subjected to smoke taint (particularly Australia and California) as a case study to inform other wine-producing countries that will likely be impacted in the future and suggest specific data collection and policy implementation actions that should be taken, even in countries that have not yet been impacted by smoke taint. Ultimately, we streamline the available information on the topic of smoke taint, apply it to a global perspective that considers the various stakeholders involved, and provide a launching point for further research on the topic.

Assessment of Smoke Contamination in Grapevine Berries and Taint in Wines Due to Bushfires Using a Low-Cost E-Nose and an Artificial Intelligence Approach.

Bushfires are increasing in number and intensity due to climate change. A newly developed low-cost electronic nose (e-nose) was tested on wines made from grapevines exposed to smoke in field trials. E-nose readings were obtained from wines from five experimental treatments: (i) low-density smoke exposure (LS), (ii) high-density smoke exposure (HS), (iii) high-density smoke exposure with in-canopy misting (HSM), and two controls: (iv) control (C; no smoke treatment) and (v) control with in-canopy misting (CM; no smoke treatment). These e-nose readings were used as inputs for machine learning algorithms to obtain a classification model, with treatments as targets and seven neurons, with 97% accuracy in the classification of 300 samples into treatments as targets (Model 1). Models 2 to 4 used 10 neurons, with 20 glycoconjugates and 10 volatile phenols as targets, measured: in berries one hour after smoke (Model 2; R = 0.98; R2 = 0.95; b = 0.97); in berries at harvest (Model 3; R = 0.99; R2 = 0.97; b = 0.96); in wines (Model 4; R = 0.99; R2 = 0.98; b = 0.98). Model 5 was based on the intensity of 12 wine descriptors determined via a consumer sensory test (Model 5; R = 0.98; R2 = 0.96; b = 0.97). These models could be used by winemakers to assess near real-time smoke contamination levels and to implement amelioration strategies to minimize smoke taint in wines following bushfires.