Abstract
Early defoliation is a novel cultural practice for crop management in grapevines. The effects of timing (prebloom or fruit set) and method (manual or mechanical) of early defoliation on yield and grape and wine composition of Vitis vinifera L. cv. Graciano and Carignan grapevines were investigated. Leaf removal induced a substantial increase in both cluster exposure and canopy porosity. Yield was significantly reduced by defoliation in both varieties. Yield per shoot was reduced between 30 and 70% by manual and mechanical leaf removal prebloom. In both varieties, postflowering leaf removal was ineffective at modifying fruit set, number of berries per cluster, or yield per shoot. Botrytis incidence was also reduced by leaf removal. Prebloom defoliation allowed full recovery of the leaf:fruit ratios to that seen in nondefoliated vines. Grape soluble solids and wine alcohol concentration were not affected by defoliation. Malic acid decreased with leaf removal at fruit set. Genotype-environment interactions on berry acidity variables such as titratable acidity and tartaric acid concentrations were found. Early leaf removal substantially increased anthocyanin and phenol concentration in grapes and wine of both varieties. In Carignan, early leaf removal resulted in heavier berries, which nevertheless had enhanced grape and wine color. These results support the hypothesis that in early-defoliated vines, the relative growth of various berry organs is affected independently from absolute berry mass. Mechanical early defoliation has the potential to be a cost-effective technique to control yield and to improve grape and wine composition.
Reducing wine surplus (IWSR 2006) and improving grape quality are important goals for the world wine industry, raising interest in crop regulation. Winter pruning and manual cluster thinning are the most widely used tools in viticulture for yield reduction. Cluster thinning can improve grape composition at harvest (Bertamini et al. 1991, Guidoni et al. 2002, Petrie and Clingeleffer 2006, Reynolds et al. 2007). However, other studies have failed to demonstrate that cluster thinning increases grape quality due to the compensatory growth of the retained clusters which end up thicker and with larger berries (Ridomi et al. 1995, Chapman et al. 2004, Keller et al. 2005). Furthermore, manual cluster thinning and winter pruning techniques are expensive due to their high labor requirements (Martinez de Toda and Tardaguila 2003).
Recently, two alternative techniques were tested for yield management: mechanical cluster thinning (Petrie and Clingeleffer 2006, Tardaguila et al. 2008b) and early leaf removal (Poni et al. 2006, Intrieri et al. 2008). Early defoliation is a novel and innovative viticultural practice to regulate yield and improve grape quality. It is carried out at bloom, unlike traditional leaf removal, typically conducted between fruit set and veraison on high-density canopies to improve fruit exposure and air circulation (Bledsoe et al. 1988, Percival et al. 1994, Tardaguila et al. 2008a). Crop regulation is achieved in early-defoliated vines through reduced fruit set, leading to smaller and looser clusters that are less susceptible to Botrytis rot (Poni et al. 2006, Intrieri et al. 2008). In these two studies, grape quality also improved in defoliated vines as soluble solids and anthocyanin concentrations increased. Further work showed that prebloom defoliation induced a consistent increase in the relative skin mass of Lambrusco and Barbera berries (Poni et al. 2009).
Most research conducted on early leaf removal has relied upon manual intervention (Poni et al. 2006, 2009), leaving the feasibility of a mechanical approach largely unexplored (Martinez de Toda and Tardaguila 2003). Only one study successfully used a leaf-plucking machine for early defoliation on the high-yielding cv. Sangiovese in Italy, on Combitrained vines (Intrieri et al. 2008). The machine effects were milder than those caused by manual leaf removal. No information is available on the performance of a leaf-plucking machine in hedgerow, vertically shoot-positioned (VSP) training systems.
For the wine industry, it will be of interest to assess the impact of early defoliation on wine composition and quality in different varieties under varying climate conditions. The purpose of this study was to evaluate the effects of timing (prebloom or fruit set) and method (manual or mechanical) of early defoliation on yield and grape and wine composition of Vitis vinifera L. cv. Graciano and Carignan VSP-trained grapevines grown under nonirrigated conditions.
Materials and Methods
The experiments were conducted in 2008 and 2009 in two commercial dry-farmed cvs. Graciano and Carignan (Vitis vinifera L.) vineyards located in Ollauri (lat. 42°31′N; long. 2°49′W; 527 m asl) and Hormilla (lat. 42°27′N; long. 2°47′W; 588 m asl), La Rioja, Spain. Graciano and Carignan vines were grafted onto 110R rootstock and planted in 1995 and 2000 in clay-loam and clay soils, respectively. Vines of both varieties had a between-row and within-row spacing of 2.70 m x 1.15 m, respectively, and row orientation was east-west. The vines were spur-pruned (12 buds per vine) in a bilateral cordon and trained to a VSP trellis system. The trellis featured a supporting wire at 0.70 m, two wires at 1.00 m aboveground for protection against wind damage, and a pair of movable shoot-positioned wires at 1.45 m. Vines were not irrigated during the growing season. Shoots were trimmed once at the end of July, before veraison.
In both vineyards, the experimental design compared four treatments: (1) control or nondefoliated; (2) hand removal of the first eight basal leaves prebloom, at stage 19 (Coombe 1995) (Man-PB); (3) hand removal of the first eight basal leaves at fruit set, at stage 27 (Coombe 1995) (Man-FS); and (4) mechanical defoliation prebloom (Mec-PB). Laterals, if present, were not removed in manual defoliation. Prebloom defoliation was performed on 13 June in 2008 and on 8 June in 2009. Fruit-set defoliation was carried out on 4 July in 2008 and on 28 June in 2009.
Mechanical leaf removal was conducted with a tractor-mounted, pulsed-air leaf remover (Collard, Bouzy, France), which operates by blowing compressed air with enough force to tear off a whole leaf or sections of leaf blades. The machine was driven at ~0.5 km/hr and removed leaves around the basal 60 cm of foliage in the fruiting zone. The leaf remover operated in two passes, one on each side of the canopy.
In each experiment, treatments were arranged in a completely randomized design of five replicates of 20-vine plots for each treatment. For each treatment, 25 vines (5 vines x 5 replicates) were tagged, and for each vine, a representative shoot was labeled before bloom. Additionally, the basal cluster of each labeled shoot was tagged.
In the 25 tagged basal clusters per treatment, fruit set was estimated as described (Poni et al. 2006). Each cluster was photographed against a dark background with a digital camera the day before defoliation. Estimation of initial flower number on tagged inflorescences was done using a linear regression between actual flower number and the number of flowers counted on photo prints, established for 30 inflorescences taken from extra vines (y = 1.7484x, R2 = 0.85 for Graciano; y = 1.9816x, R2 = 0.88 for Carignan).
Assessment of canopy porosity and cluster exposure was by digital image analysis (Dunn and Martin 2004). For each treatment, the 25 labeled vines were photographed between 07:00 and 09:00 one week before harvest. A 1.15 m × 0.70 m wooden frame was suspended in front of the fruiting zone of each vine, 0.70 m aboveground, and the frame canopy was photographed using a digital camera mounted on a tripod set normal to the canopy 2 m from row axis and 1.05 m aboveground. A white screen was placed behind the canopy to avoid confounding effects from background vegetation. The digital images were cut to include only the 1.15 m × 0.70 m portion of canopy using XnView version 1.96 (www.xnview.com), then analyzed using the image analysis software Envi 4.3 (ITT Visual Information Solutions, Boulder, CO). Red, green, and blue threshold values and tolerances were manually set to establish four different hue classes: clusters, green leaves, yellow-wilted leaves, and porosity. These settings were retained and applied to all images. The program was then used to automatically count total pixels in each class. To avoid the influence of yield on the percentage of cluster pixels in the image, the ratio pixels of cluster/yield per vine was calculated.
At harvest, the main, lateral, and total leaf area of each tagged shoot were assessed using a method based on the disc technique (Smart and Robinson 1991). All main and lateral leaves on each tagged shoot were separately removed and weighed and 100 discs (2 cm diam) were cut from these leaves and weighed. The weight of the discs was compared with the weight of main and lateral leaves, allowing main and lateral leaf area per shoot to be calculated.
For yield assessment, all clusters on tagged shoots were separately weighed and yield per shoot was determined. In the laboratory, the following variables were appraised for each tagged basal cluster: cluster weight, berry number, berry weight, cluster compactness (following the OIV, 204 standard; OIV 2009b) and visual presence/absence of Botrytis bunch rot. If present, Botrytis incidence was then estimated as the percentage of affected clusters per treatment.
After yield estimation, all berries were frozen and stored at −18°C for fruit composition analyses. Berries were naturally defrosted overnight and a subsample of 50 berries was hand crushed and centrifuged at 4,000 rpm for 10 min. The total soluble solids concentration (Brix) was determined using a temperature-compensating digital refractometer (Atago, Tokyo, Japan) and titratable acidity (TA), pH, and tartaric and malic acid concentrations were determined according to standard methods (OIV 2009a). The remaining berry subsample was homogenized using an Ultra Turrax grind mixer (IKA, Staufen, Germany) at high speed (14,000 rpm for 1 min). Anthocyanin and phenolic concentrations were determined as described (Iland et al. 2004). Total anthocyanins were expressed as mg/berry and mg/g FW and total phenols were expressed as absorbance units (AU) at 280 nm: AU280/berry and AU280/g FW.
For each cultivar and treatment, the remaining grapes from the five labeled vines per replicate were harvested and blended. Grapes were transported to the winery of the University of Rioja and stored for 12 hr at 4.5°C. Microscale wine fermentations were conducted as described (Sampaio et al 2007). Grapes were destemmed and slightly crushed (Enomundi, Zaragoza, Spain). Sulfur dioxide was added at a rate of 60 mg/kg and musts inoculated with Saccharomyces cerevisiae (Uvaferm 71B; Lallemand, Montreal) at a rate of 20 g/hL. Fermentation temperature was maintained between 27 and 31°C. Alcoholic fermentations were completed after seven days, but extended maceration was allowed for eight more days. After fermentation, wines were manually racked and pressed. For each microfermentor, the free-run and pressed wine fractions were blended and bottled. For each wine, a 125 mL sample was taken for analysis of alcohol concentration, TA, pH, and malic acid concentration by standard methods (OIV 2009a). Color density was calculated by adding the absorbance at 420, 520, and 620 nm and hue was measured as the ratio of absorbance at 420 and 520 nm (OIV 2009a). Total polyphenol index was calculated from the absorbance at 280 nm as described (European Commission 1990). All analyses were run in triplicate.
A combined analysis of variance was performed using the Infostat Professional 2007 statistical package (Infostat, Cordoba, Argentina). The effect of defoliation, its timing, and its method were evaluated using a priori contrasts (p < 0.05). Dunnett’s t-test (Dunnett 1955) was used to demonstrate significant differences between each defoliation treatment from the control at p = 0.05.
Results
Canopy characteristics and leaf area.
Canopy porosity and cluster and leaf exposure were determined for Graciano (Table 1⇓) and Carignan (Table 2⇓). In both varieties, leaf removal induced a significant increase in cluster exposure, canopy porosity, and the pixel ratio of cluster/yield per vine. Additionally, defoliation decreased the proportion of green leaves but had no effect on the yellow/wilted leaf percentage. Leaf removal at fruit set resulted in improved cluster exposure compared with prebloom pulling in both varieties and also increased canopy porosity in Carignan. When cluster exposure was normalized by dividing yield, the differences from the control became more evident. Both fruit-set defoliation and mechanical leaf removal prebloom exhibited higher ratios than the control. An increase in yellow/wilted leaves was observed in Carignan as a result of the mechanical treatment.
Canopy porosity and cluster and leaf exposure of Graciano (Vitis vinifera L.) for different early defoliation treatments at harvest in 2008.
Canopy porosity and cluster and leaf exposure of Carignan (Vitis vinifera L.) for different early defoliation treatments at harvest in 2008.
In both cultivars, final total leaf area per shoot was significantly reduced by defoliation, essentially reflecting differences in primary leaf development with no evident compensation for lateral regrowth (Table 3⇓). Timing and method of defoliation had no impact on shoot vegetative development.
Influence of early defoliation on leaf area (LA) components of Graciano and Carignan grapevines at harvest in 2008 and 2009.
Yield components.
All yield components, with the exception of initial flower number, were significantly reduced by defoliation in Graciano (Table 4⇓). The same applied to Carignan, except for fresh berry weight (Table 5⇓). Yield per shoot was drastically reduced by defoliation (−27% and −31% in Graciano and Carignan, respectively, averaging all treatments). Botrytis incidence was also reduced by leaf removal in both cultivars. In both varieties, postflowering leaf removal was ineffective at modifying fruit set, berries per cluster, or yield per shoot, but prebloom pulling severely reduced these variables. The per berry fresh weight responded differently to leaf removal in the two cultivars: it was generally reduced in Graciano but increased in Carignan. In general, mechanical defoliation was more effective in reducing yield than manual leaf removal.
Influence of early defoliation on yield of Graciano grapevines at harvest in 2008 and 2009.
Influence of early defoliation on yield of Carignan grapevines at harvest in 2008 and 2009.
The total leaf area:yield ratio at harvest per shoot showed a substantial reduction for Man-FS in Carignan and increased for Mec-PB in both cultivars. In general, prebloom defoliation allowed full recovery of leaf:fruit ratios to that as seen in nondefoliated vines regardless of intervention.
Grape and wine composition.
Grape maturity, assessed as soluble solids concentration and TA, was affected differently by defoliation in Graciano (Table 6⇓) and Carignan (Table 7⇓). In Graciano, early leaf removal increased soluble solids concentration only in 2009, especially for the prebloom treatments, whereas in Carignan no effect on soluble solids was recorded.
Influence of early defoliation on grape composition of Graciano at harvest in 2008 and 2009.
Influence of early defoliation on grape composition of Carignan at harvest in 2008 and 2009.
Defoliation at fruit set decreased TA in Graciano, but in Carignan, manual early defoliation increased TA regardless of the timing of intervention in 2008, and had no effect in 2009. In both cultivars, differences in TA between manual and mechanical operations were also encountered. Leaf removal diminished tartaric acid in Graciano in 2009, especially in the Man-FS and Mec-PB treatments, while manual defoliation at either time increased tartaric acid concentrations in Carignan in both years. Fruit-set leaf removal in Graciano and both manual treatments in Carignan decreased must pH in 2008, but in 2009 must pH was enhanced by early defoliation for all treatments in Graciano and by mechanical leaf removal prebloom in Carignan. Malic acid was differentially affected by the timing of leaf removal: in both cultivars, malate decreased upon fruit-set leaf removal compared to the earlier timing.
Early leaf removal enhanced anthocyanin and phenol concentrations in Graciano in both seasons. In Carignan, regardless of the unit chosen to express the data, anthocyanins and phenols were markedly enhanced in fruits from leaf-removed shoots, with a tendency for a sharper effect with earlier defoliation.
Despite the lack of effect on acidity, wine composition closely resembled grape composition in Graciano. There were differences in alcohol concentration, color density, and total phenols among treatments (Table 8⇓). Carignan wines from leaf removal treatments did not differ in alcohol, TA, or malic acid, but had a lower pH for the Man-FS treatment in both seasons (Table 9⇓). Wine color density and total phenols were also improved in defoliated vines. However, while postbloom defoliation had the highest color and phenols in 2008, there were no differences between timings in 2009. Mechanical leaf removal produced wines with the most color and phenols in this season. A trend toward a more bluish tint for wine hue (lower values) was observed in Carignan wines from defoliated vines, and differences were observed between the control and prebloom manual treatment in 2008 and for all treatments in 2009, with no differences between timings.
Influence of early defoliation on wine composition of Graciano in 2008 and 2009.
Influence of early defoliation on wine composition of Carignan in 2008 and 2009.
Discussion
Canopy characteristics and leaf area.
The increase in cluster exposure and canopy porosity in both varieties was an expected outcome, since the viticultural practice of defoliation causes a direct and usually long-lasting effect on canopy leaf density in the fruiting zone. However, especially if the leaf pulling is executed early in the season, partial cluster reshading can occur as a function of intrinsic vine vigor, postdefoliation weather pattern, compensating growth of retained primary leaves, or maintenance or removal of growing lateral tips at the time of defoliation.
The data for cluster exposure suggest that defoliation in both varieties led to higher cluster exposure that was not predicted simply on the basis of the lateral formation by the whole canopy level (Table 3⇑). Cluster exposure per yield was significantly higher for the manual postbloom and mechanical prebloom treatments. Although manual defoliation postbloom did not reduce yield, the removal of eight leaves of each shoot created a more porous and clear canopy, especially in the fruiting zone, that resulted in more exposed clusters. For mechanical defoliation prebloom, some lateral removal due to the blowing machine action might have contributed to the higher cluster exposure. In general, there were not strong differences between the two methods of defoliation. These defoliation treatments impose a more permanent change in cluster exposure, with the risk of having overexposed clusters during the hottest time of the season. In Pinot gris, the incidence and severity of skin burning increases as defoliation of the basal shoot zone is delayed throughout the season (Mescalchin et al. 2008).
Yield components and vine balance.
Carbohydrate supply at flowering is a primary determinant of fruit set (Coombe 1962, Caspari and Lang 1996, Poni et al. 2006). Yield per shoot was greatly reduced in prebloom defoliated vines in both cultivars. As an expected yet desirable side effect, since the yield component which was most diminished by defoliation was berries/cluster, clusters borne on Man-PB and Mec-PB vines were looser, and thus less susceptible to rot (Table 4⇑, Table 5⇑).
A key result of this trial was that crop control was achieved only if prebloom defoliation was used. Postbloom defoliation was ineffective in significantly lowering cluster weight and berry number per cluster. In principle, this finding contrasts with work showing that postflowering manual leaf removal considerably restrained berry number in Trebbiano over three years (Poni et al. 2006). These authors used the same degree of defoliation (eight primary basal leaves removed) but laterals were also removed.
In general, mechanical prebloom defoliation was as effective as manual leaf removal in regulating yield. These findings confirm the performance obtained with a leaf-plucking machine either pre- or postbloom on Sangiovese vines (Intrieri et al. 2008), which removed about 50% of the leaf area pulled from the six basal nodes in the manual treatments. In our study, the fraction of leaves blown away by the machine was particularly high (>90%, data not shown) because of the different working principle (air pressure instead of a sucking machine) and to more accurate action due to two passes per row. Any lateral shoot that already had developed within the basal shoot zone targeted by the machine was either destroyed or removed, while in the hand treatments they were preserved.
The present study also confirms that effects of early leaf removal on final berry size are erratic. The weight of Graciano berries was generally reduced by defoliation, but in Carignan, there was a tendency toward larger berries in defoliated shoots. Thus, despite the severe source limitation brought about by removal of fully functional primary leaves at flowering, factors acting thereafter allowed full growth compensation of the berries.
Furthermore, defoliated vines of both cultivars showed less cluster rot. These results could be explained by the looser clusters and greater cluster exposure that provided better ventilation and more effective spray treatments from flowering to harvest.
The final leaf area:fruit ratios at harvest reinforce the conclusion that, in defoliated shoots, the curtailment in leaf area due to removing leaves might be fully counterbalanced by the corresponding reduction in yield, leaving the source:sink ratio essentially unchanged. However, in this study Man-FS suffered from a somewhat decreased final leaf:fruit ratio, which primarily reflects the poor performance of this treatment in reducing fruit set, and hence crop level per shoot.
Grape composition.
The small changes in must soluble solids among treatments in both cultivars generally reflected minor changes in the final leaf:fruit ratio. Leaf removal led to few changes in TA in either cultivar, except for manual defoliation at fruit set in Graciano, where it was diminished, and a significant increase in 2008 in Carignan. The enhancement of TA in the cool, rainy 2008 season together with the lack of effect observed in 2009, which was a very hot, dry year, is a positive outcome and of special interest due to the ripening advancement induced by global warming, despite Carignan being a late-ripening variety. Titratable acidity and must pH were differently affected by early defoliation depending on weather and cultivar. In Graciano and Carignan, early defoliation in cool and rainy 2008 induced no changes or opposite effects than in 2009, which was warm and dry. Genotype-environment interactions on berry ripening and acidity variables, such as TA and pH, have been reported in Barbera, Croatina, and Malvasia cultivars during a classical (veraison) defoliation trial over four years (Bavaresco et al. 2008).
The malic acid concentration in both cultivars diminished more after fruit-set leaf removal than prebloom. These results accord with highly improved cluster exposure in fruit-set defoliated vines (Table 1⇑, Table 2⇑). While the malic acid decrease in Carignan berries might be explained by temperature-driven enhanced degradation (Kliewer 1971, Rüffner 1982), the increase in tartaric acid in Carignan berries from the defoliated shoots confirms reported findings on Trebbiano (Poni et al. 2006) and strengthens the hypothesis that the sudden surge in light exposure and temperature caused by early leaf removal might lead to greater synthesis of tartaric acid. Previous work reported more 14CO2 incorporated into tartaric acid by berries in full sun than in shaded treatments (Kliewer and Schultz 1964). However, genotype and variations in canopy porosity play an important role in this physiological response and could explain the opposite changes in tartaric acid concentrations observed for Graciano and Carignan defoliated vines. In fact, Graciano clusters were less exposed than Carignan clusters.
Early leaf removal substantially increased color and phenols in berries. Indeed, the enhanced anthocyanin and total phenol concentrations in Graciano are not directly and uniquely related to smaller berry size induced by early defoliation (Table 4⇑), which would presumably allow a higher skin:pulp ratio (Johnstone et al. 1996, Clingeleffer et al. 2000). In Carignan, early leaf removal induced heavier berries which nevertheless showed enhanced pigmentation. This finding matches with the response seen in early-defoliated Barbera shoots which, despite larger berries than nondefoliated shoots, also had increased relative skin mass (Poni et al. 2009). Therefore, our results add more evidence to the hypothesis that in early-defoliated vines, the relative growth of various berry organs is affected independently from absolute berry mass. Also, skin growth can be promoted locally by improved cluster microclimate which, in turn, enhances cell division three to four weeks after flowering, when skin deposition is almost completed (Roby et al. 2004, Caporali et al. 2005, Poni et al. 2009).
Wine composition.
The minimal effect of early defoliation on grape soluble solid concentration is reflected in the alcohol concentration of the resulting wines. Only in 2009 was increased alcohol concentration induced by leaf removal in Graciano. It is widely accepted that anthocyanin concentration in grape is closely related to anthocyanin concentration in wine (Iland 1987) and the final quality (Francis et al. 1998) of wine. This is consistent with the results described here in which increased anthocyanin concentration of berries enhanced the color density of the wines. Moreover, the relationship between wine anthocyanin and phenolic concentrations and wine quality is of utmost importance, especially in premium wines intended for aging.
In the two studied varieties, early leaf removal significantly improved wine color and phenolic concentration, with little change in alcohol or pH. An outstanding quality wine is seldom described by its alcohol and acidity, but more typically by its color and phenolic attributes, as grape-derived phenolics are critical components of wine quality, contributing to color, bitterness, and mouthfeel.
Conclusions
This study confirms the effectiveness of early leaf removal, either manual or mechanical, as a tool to control crop load in Graciano and Carignan grown in the Rioja region. This is accomplished primarily via a reduction in the number of set berries, leading in turn to looser clusters that are less susceptible to rot. Soluble solids concentration and total acidity were not markedly affected by defoliation, likely because of slight changes in the leaf:fruit ratio of defoliated and control vines. Color and phenolic accumulation improved in the defoliated vines, suggesting that this intervention may uncouple the patterns of phenolic ripening from those classically bound to sugar accumulation and degradation of acids. This may be due to the local influence of altered light exposure and temperature on inflorescences and developing clusters which, regardless of final berry mass, promote skin growth through either adaptive responses to long-term exposure to high light or a microclimate more favorable to cell division and skin deposition.
Footnotes
Acknowledgments: The authors thank the Agencia de Desarrollo Económico de La Rioja (ADER-2006-I-ID-00157) and the Ministerio de Ciencia e Innovación for financial support AGL2007-60378) and the Agrupación de Bodegas Centenarias y Tradicionales de Rioja and New Holland for their assistance and help.
- Received August 2009.
- Revision received December 2009.
- Revision received March 2010.
- Accepted April 2010.
- Published online September 2010
- Copyright © 2010 by the American Society for Enology and Viticulture
Literature Cited
Vol 61 Issue 3
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
