Smoke Taint: Effect of wildfires on fruit and wine composition

UCCE Sonoma County Grape Day February 7, 2019 Smoke Taint: Effect of wildfires on fruit and wine composition Thomas Collins, PhD Washington State University

Smoke taint Wines made from vineyards exposed to smoke from wildfires can develop characteristic aromas and flavors Smokey Ashy Cigar or cigarette smoke Strong ashy, cigar-butt aftertaste can also occur In many cases, the aftertaste is more offensive than the aroma Ashy, cigarette smoke notes linger long after the wine is gone.

First smoke taint research report--2007 Kennison, et al, demonstrated that fruit exposed to smoke produced wine which had smoky aromas and flavors Guaiacol, 4-methyl guaiacol, 4-ethyl guaiacol, 4-ethyl phenol, eugenol and furfural were identified in wines made with smoked grapes 3 guaiacol 4 methyl guaiacol syringol furfural

Timing of smoke exposure ug/l 100 90 80 70 60 50 40 30 20 10 0 Guaiacol (ug/l) All smoke exposure yielded some guaiacol Exposure seven days post veraison produced maximum Multiple exposures yielded additive amounts of guaiacol Kennison, et al, 2009 Days post veraison 4

B.C. smoke event, August 1 st 12 th 2017 Wildfires in B.C. resulted in poor air quality for much of Washington

Air quality monitors Particle counters were used to monitor smoke density PM 2.5, PM1.0 Particle size in microns Particle size generally decreases during smoke events Road dust PM10.0 Smoke particles generally PM1.0 and smaller

0.6 0.5 0.4 0.3 0.2 0.1 0 Smoke exposure at Roza vineyard, August 02-12, 2017 PM2.5 mg/m 3 16:47:09 20:57:09 1:07:09 5:17:09 9:27:09 13:51:50 18:01:50 22:27:12 2:37:12 6:47:12 11:06:27 15:16:27 12:41:08 16:51:08 21:01:08 14:39:35 18:49:35 22:59:35 3:09:35 7:19:35 11:37:59 15:49:59 19:59:59 0:09:59 8:41:30 12:51:30 17:01:30 21:16:04 1:26:04 5:36:04 9:54:20 14:04:20 18:14:20 22:31:20 2:41:20 6:51:20 11:06:31 15:16:31 19:33:48 23:43:48 3:53:48 8:03:48 12:20:17

Smoke exposure at Roza vineyard, August 02-12, 2017 0.6 0.5 0.4 0.3 0.2 TWA was less than 0.2 mg/m 3 0.1 0 16:47:09 20:57:09 1:07:09 5:17:09 9:27:09 13:51:50 18:01:50 22:27:12 2:37:12 6:47:12 11:06:27 15:16:27 12:41:08 16:51:08 21:01:08 14:39:35 18:49:35 22:59:35 3:09:35 7:19:35 11:37:59 15:49:59 19:59:59 0:09:59 8:41:30 12:51:30 17:01:30 21:16:04 1:26:04 5:36:04 9:54:20 14:04:20 18:14:20 22:31:20 2:41:20 6:51:20 11:06:31 15:16:31 19:33:48 23:43:48 3:53:48 8:03:48 12:20:17 PM2.5 mg/m 3

Smoke Exposure at Roza vineyard Berry samples analyzed for presence of smoke related volatile phenols and glycosides Samples collected from August 3 rd -12 th did not have observable levels of free compounds or glycosides Combination of relatively low smoke density and pre-veraison fruit limited the extent of smoke taint from this smoke event

Labor Day Smoke Event, September 5 8 th 2017 Smoke from Eagle Creek fire, Jolly Mountain and Norse Peak fires covered much of the Columbia Basin as well as the Columbia Gorge

Labor Day Smoke Event, September 5 8 th 2017 PM 2.5 was much higher than for the BC event ranged from 0.7 0.9 mg/m 3 at the Roza vineyard Smoke from the Eagle Creek fire seen from the WA side of the Columbia river Control wines from the smoking trials have detectable levels of smoke taint volatiles and glycosides

Smoke from fires in Siberia & British Columbia drifted down into eastern Washington, Oregon and Idaho. Smoke event 08/19/2018 PM 2.5 levels in excess of 2.0 mg/m 3 for several hours Sunday afternoon/evening

Smoke from fires in Siberia & British Columbia drifted down into eastern Washington, Oregon and Idaho. PM 2.5 levels in excess of 2.0 mg/m 3 for several hours Sunday afternoon/evening Smoke event 08/19/2018 In spite of the smoke density, this event did not result in appreciable smoke taint in affected areas

Smoke Exposure at Roza vineyard, 2017 Berry samples analyzed for presence of smoke related volatile phenols and glycosides Samples collected from August 3 rd -12 th did not have observable levels of free compounds or glycosides Combination of relatively low smoke density and pre-veraison fruit limited the extent of smoke taint from this smoke event Or so we thought at the time it may be more related to low smoke density and the distance the smoke was carried to reach our site

First smoke taint research report--2007 Kennison, et al, demonstrated that fruit exposed to smoke produced wine which had smoke taint aromas and flavors Guaiacol, 4-methyl guaiacol, 4-ethyl guaiacol, 4-ethyl phenol, eugenol and furfural were identified in wines made with smoked grapes 15 guaiacol 4 methyl guaiacol syringol furfural

Smoke from conifer bark GC/MS chromatogram of smoke collected from burning conifer bark phenol o cresol m cresol guaiacol 4 methylguaiacol 4 ethylguaiacol

Guaiacol Guaiacol is widely used as a marker for smoke exposure Guaiacol or 2 methoxy phenol

Guaiacol Guaiacol is widely used as a marker for smoke exposure Guaiacol is produced by thermal degradation of lignin when wood or other plant material is burned Guaiacol or 2 methoxy phenol

4-Methyl Guaiacol 4-methyl guaiacol is usually used along with guaiacol as a marker for exposure 4-MG is also produced during the thermal degradation of lignin

4-Methyl Guaiacol 4-methyl guaiacol is usually used along with guaiacol as a marker for exposure 4-MG is also produced during the thermal degradation of lignin Both guaiacol and 4-methyl guaiacol are found in barrels, again from the thermal degradation of lignin in the oak wood

4-Methyl Guaiacol 4-methyl guaiacol is usually used along with guaiacol as a marker for exposure 4-MG is also produced during the thermal degradation of lignin Guaiacol and 4-methyl guaiacol are good markers of smoke exposure but maybe not so good for the intensity of smoke perception

Lignin Lignin is an irregular highly cross-linked phenolic polymer It cross-links with carbohydrate polymers such as cellulose to provide rigidity and strength to wood

Lignin When lignin is heated, the bonds cross-linking the phenolic subunits can be cleaved, releasing guaiacol and other small volatile phenols Cleave here to produce guaiacol

Lignin When lignin is heated, the bonds cross-linking the phenolic subunits can be cleaved, creating guaiacol and other small volatile phenols Cleave here to produce vanillin

Volatile phenols Thermal degradation of lignin produces a range of small volatile phenols All of these compounds have been found in smoke affected wines Parker et al, 2012 showed cresols could contribute to smoky character in wines 4 ethyl guaiacol o cresol syringol 4 vinylguaiacol

Smoke taint glycosides In addition to the free volatile phenols, glycosides are also present Glycosides are less volatile than the free compounds and generally odorless Glycosidic bond is acid labile and can hydrolyze at wine ph Guaiacol glucoside

Glycosides of smoke compounds Guaiacol, arabinosyl glucoside Guaiacol, glucoside

Glycosides of smoke compounds Guaiacol, coumaroyl glucoside

Role of the glycosides in smoke taint Mayr, et al, 2014, reported that in a model system, the addition of both free volatile phenols and their glycosides produced the strongest off-aromas in a descriptive analysis experiment Enzymatic activity in saliva hydrolyzes both mono- and di-saccharides Variable hydrolysis dependent on glycosides present

Two questions on everyone s mind: Wildfires and Wine View from the WSU Roza vineyard on 05 September 2017

Two questions on everyone s mind: Wildfires and Wine How much smoke does it take to create the problem? View from the WSU Roza vineyard on 05 September 2017

Two questions on everyone s minds: Wildfires and Wine How much smoke does it take to create a problem? Can the wines be treated? View from the WSU Roza vineyard on 05 September 2017

Assessing Risk of Smoke Taint Project initiated in 2016 with funds from the Washington Wine Commission to address these questions Develop a system to smoke vines at our research center near Prosser 60 vines at a time, smoke and control 18-48 hour exposures Monitoring smoke intensity and composition Generate smoked fruit to make wines for mitigation studies

How much smoke does it take to create a problem? Wildfires and Wine View from the WSU Roza vineyard on 05 September 2017

How much smoke does it take to create a problem? Wildfires and Wine It depends View from the WSU Roza vineyard on 05 September 2017

How much smoke does it take to create a problem? Wildfires and Wine It depends: On the smoke View from the WSU Roza vineyard on 05 September 2017

How much smoke does it take to create a problem? Wildfires and Wine It depends: On the smoke fuel source View from the WSU Roza vineyard on 05 September 2017

How much smoke does it take to create a problem? Wildfires and Wine It depends: On the smoke fuel source proximity of the fire View from the WSU Roza vineyard on 05 September 2017

How much smoke does it take to create a problem? Wildfires and Wine It depends: On the smoke fuel source proximity of the fire On timing of the fire relative to berry development (?) View from the WSU Roza vineyard on 05 September 2017

The role of fuel source GC/MS chromatogram of smoke from softwood bark mulch Sampled smoke from 15 plant species commonly found in rangeland of eastern Washington Controlled conditions, smoke collected on 47 mm filter papers

Filters for Big Sagebrush 47 mm filter paper; three 9mm discs from each filter put in separate GC vials for analysis

Rabbit brush and sage GC/MS Chromatograms Rabbit brush Artemisia tridentate: Big Sagebrush

Smoke from Conifer bark Phenol m-cresol o-cresol Guaiacol 4-Ethylphenol 4-methylguaiacol 4-ethylguaiacol Eugenol

Smoke from Rabbitbrush Phenol o-cresol Guaiacol m-cresol 3,5 xylenol 4-methyl guaiacol 4-ethyl guaiacol Eugenol Cadinadiene Trimethylnapthalene Benzophenone Tetramethylnapthalene Phenanthrene 2-methylanthracene

Rabbit brush tar Gray rabbit brush (Ericameria nauseosa) was evaluated during WWII as a source for natural rubber and other resins. The species name nauseosa refers to the foul aromas that arise when the leaves or flowers are crushed (or burned, as it turns out!)

2017 Smoking Trials Impact of fuel source Chardonnay and Merlot smoked with proprietary blend of rangeland plant materials Cabernet smoked with mix of conifer back mulch 26 48 hour exposures Increase from 2016 Wines made from exposure to smoke from rangeland fires express different aromas than wine from bark mulch smoke

How much smoke does it take to create a problem? Wildfires and Wine It depends: On the smoke fuel source flaming vs smoldering fires proximity of the fire On timing of the fire relative to berry development View from the WSU Roza vineyard on 05 September 2017

2018 Smoking trials Timing of smoke exposure relative to berry development Pre véraison At véraison 3 weeks post véraison 6 weeks post véraison

2018 Smoking trials Timing of smoke exposure relative to berry development Pre véraison At véraison 3 weeks post véraison 6 weeks post véraison All of the wines from this trial have smoke taint, including the wines made from fruit smoked 4 weeks pre veraison Wine from pre veraision unsmoked control are not tainted, in spite of the smoke from the August 2018 event

2018 Smoking trials Timing of smoke exposure relative to berry development Pre véraison At véraison 3 weeks post véraison 6 weeks post véraison Australian work had suggested the risk prevéraison was lower than post véraison, but not zero We plan to conduct an even earlier smoking in 2019, plus confirmation of results from 4 weeks prevéraison smoking in 2018

Key Take-aways Smoke related compounds are present both as free compounds and as glycosides Free compounds responsible for wine aromas Glycosides play a role in the flavor/aftertaste Glycosides hydrolyze over time to release more free compounds Smoke composition varies with plant source; differences in smoke composition affect the aromas and flavors in the resulting wines Smoke exposures both pre- and post-veraison can result in smoke affected fruit and wine

Thank you!

R.O. treatment of six 2008 red wines 60 50 40 (ug/l) 30 20 Guaicol Pre Guaicol Post 10 0 A B C D E F 56

Guaiacol content of R.O. treated wines over time 18 16 14 µg/l 12 10 8 6 4 June August October December February April June 2 0 Wine 3 Wine 20 Wine 25 Wine 32 Wine 40 57

Guaiacol released by strong acid hydrolysis one year after R.O. treatment 70 60 (ug/l) 50 40 30 20 10 June August October December February April June June hydrolyzed 0 58