Abstract
Many growers in the Niagara Peninsula of Ontario are reluctant to use canopy division to address high vigor in their vineyards. An alternative approach of delayed shoot thinning was explored. Pinot noir and Cabernet franc vines in Ontario were subjected to six different shoot-thinning timings, based on the Eichhorn and Lorenz phenological stages 9 to 31, during the 2001 and 2002 seasons. An additional double-prune treatment (retaining four canes at dormant pruning and removing two at fruit set) was imposed on Cabernet franc. Minor differences between treatments were observed in 2001 and 2002 with respect to yield, periderm formation, and vine size. However, the study was carried out primarily to assess the impact of these treatments on berry, must, and wine composition and canopy microclimate. Early shoot-thinning treatments on Pinot noir resulted in increased titratable acidity (TA) and soluble solids in berries and must. Early shoot-thinning treatments on Cabernet franc generally resulted in higher soluble solids and color intensity in berries, lower TA in musts, and higher color intensity, total phenolics, and total anthocyanins in wines. The double-prune treatment was characterized by higher soluble solids, hue/tint, color intensity, and total phenolics overall. Canopy assessment showed that treatments imposed after bloom produced lower leaf layer numbers and better leaf and cluster exposure than did the control and early shoot thinning, especially in Cabernet franc. Early shoot-thinning treatments induced higher leaf areas in both cultivars compared to later treatments. Double pruning resulted in lower leaf layer numbers and better leaf exposure and had no impact on mean leaf area or photosynthetic photon flux density. Despite reductions in canopy density, delayed shoot thinning appeared to lead to concomitant delays in fruit maturity in terms of soluble solids, anthocyanins, and phenolics.
There is general acceptance that high-quality wines are produced from vineyards where balance is maintained between yield and vegetative growth. Balance is defined variously as either a range of crop loads (yield:cane pruning weight ratio of 5 to 12) (Bravdo et al. 1985) or as a cane pruning weight of 20 to 30 g/cane (Smart and Smith 1988). Vine balance may be achieved in vigorous vineyards through increased canopy length; for example, canopy division or increased vine spacing (Reynolds et al. 2004). The ultimate effect of these cultural practices is a reduction in individual shoot vigor and concomitant improvement in wine quality (Reynolds et al. 2004). More conservative approaches for achieving vine balance might include delay of shoot-number adjustment (shoot thinning) until bloom or thereafter in order to allow shoots to compete. A more drastic form of this “vigor diversion” is the retention of extra canes (“double pruning”) that can be removed midseason after the competitive effects of the extra shoots have been felt. The ultimate goal is to achieve an appropriate shoot density and an optimal vine balance that may improve overall quality of the wines produced.
Achieving vine balance may ultimately lead to significant improvements in canopy microclimate over an overvegetative vine by reducing shoot vigor and leaf size (Smart and Smith 1988). Maintaining an appropriate shoot density by either pruning or shoot thinning can create an optimal canopy light microclimate that will result in high node fruitfulness, optimum vine hardiness, high soluble solids, low TA and pH, enhanced varietal character, minimized vegetative flavors, and improved color (Reynolds et al. 1994a,b,c, 1996, Smart 1988, Smart and Smith 1988). Excessive shoot densities may lead to shaded canopies and compromised fruit composition (Smart 1988, Smart and Smith 1988).
Studies investigating shoot-thinning effects are rare (Morris et al. 2004), and the use of timing of shoot thinning at various phenological growth stages to achieve an appropriate shoot density has not been studied thoroughly. From an intuitive standpoint, a low shoot density (10 to 15 shoots/m row) is believed to optimize canopy microclimate; however, excessive shoot thinning may increase the vegetative growth of remaining shoots, leading to diminished leaf and cluster exposure (Reynolds et al. 1994a). Suboptimal shoot densities may therefore occur when thinning is conducted early in the growing season when most vegetative growth occurs. However, shoot thinning performed later in the season may allow for more intershoot competition, shorter internodes, smaller leaves, fewer lateral shoots, and hence better sunlight penetration into the canopy (Reynolds et al. 1994a). The resulting increased fruit exposure to sunlight can lead to increased levels of anthocyanins, color, and phenolics in red wine-grape cultivars (Jackson and Lombard 1993, Price et al. 1995, Reynolds et al. 1994c, 1996, 2004).
Most research on shoot thinning has involved determining optimal shoot densities in pursuit of optimal wine sensory attributes. Numerous studies (Kiefer and Crusius 1984, Murisier and Zeigler 1991, Reynolds et al. 1986, 1994a, Reynolds et al. b,c,d, 1996, Smart 1988) have indicated that improved canopy microclimate can be achieved at shoot densities in the range of 15 to 25 shoots/m row or less. The objectives of this study were to examine the phenological stages at which shoot thinning is most effective. It is of critical importance in cool-climate wine-producing regions such as Niagara where vigorous canopy growth is a common problem. We hypothesized that optimal phenological growth stages for shoot thinning would occur after fruit set and that late shoot thinning or double pruning may result in reduced vigor of remaining shoots, leading to enhanced fruit exposure and composition. We suspected that early shoot thinning might lead to increased vigor of remaining shoots and an enhancement of the growth of laterals, possibly resulting in a shaded canopy, producing fruit with lower quality in terms of soluble solids, pH, titratable acidity, color, phenolics, and anthocyanins.
Materials and Methods
Experimental design.
Two vineyards located ~1.0 km from each other on Niagara’s Beamsville Bench (Ontario) were used for the field trials. Six rows of 5-year-old Cabernet franc (3.0 m x 1.5 m spacing) at East Dell Estates Vineyard and three rows of 5-year-old Pinot noir (2.4 m x 1.2 m spacing) at the Novello Vineyard were used in the experiment. Vines in both blocks were trained to a double-Guyot system and pruned to two canes and at least two 2-node spurs, with the canes tied to fruit-bearing wire, such that about 20 nodes/m of row were retained. Canopy management consisted of vertical shoot-positioning and hedging at the top of the canopy at pea size of berry development. Experimental design was a randomized complete block with six blocks containing seven treatments in Cabernet franc and six treatments in Pinot noir. Each treatment consisted of two 5-vine treatment replicates (two post-length plots of five vines each) in each of the six blocks. One buffer plot was included at the start of each row where no treatments were imposed.
The timing of field shoot-thinning treatments was carried out according to five different phenological development stages (Eichhorn and Lorenz 1977): stage 09–12 (three to five expanded leaves), stage 15 (flower clusters visible with six to seven expanded leaves), stage 19–21 (full bloom), stage 25–27 (fruit set), and stage 29–31 (early stage I, berry cell division). Shoot thinning was carried out with the objective of retaining ~15 shoots/m of row. Those shoots removed were typically secondary and/or tertiary shoots originating from “count” positions (nodes retained at pruning) as well as any “noncount” shoots (base shoots) that originated from the base of renewal spurs. Control treatments were also imposed for both cultivars in which no thinning was done other than trunk suckering. A double-prune treatment in Cabernet franc involved the removal of a third cane that was retained on the vine during pruning. Shoot densities for the control treatment in Pinot noir were 18.6 shoots/m row in 2001 and 15.6 shoots/m row in 2002. These values were reduced by shoot thinning to 14.6 to 16.2 and 13.0 to 13.9 shoots/m row in 2001 and 2002, respectively. Values for Cabernet franc were: 21.7 and 26.0 shoots/m row (control, 2001 and 2002) and ranges of 16.6 to 17.1 and 14.3 to 16.0 shoots/m row (shoot thinned, 2001 and 2002, respectively). Double-prune treatments averaged 19.1 and 21.0 shoots/m row in the respective seasons. Each thinning treatment was done at intervals of approximately one to one and a half weeks based on the phenological development of the vines. The double-prune treatment was completed one week after the stage 29–31 treatment.
Yield components.
Each vine in both trials was harvested individually each season and the clusters were counted. Yield was measured using an electronic scale. Cluster weight data were calculated from the yield and clusters per vine data. Berries per cluster were calculated from cluster weight and berry weight data. Harvest dates were 2 to 3 Oct 2001 and 3 to 4 Oct 2002 (Pinot noir) and 10 to 11 Oct 2001 and 2002 (Cabernet franc).
Vines were pruned during the winter following each season of data collection, and weight of cane prunings was recorded for each vine. Periderm formation (a measurement of cane maturity) was monitored during vine dormancy after both seasons. The number of canes per vine in the following periderm classes was recorded (number of mature internodes per cane): <3; 4 to 9; >10.
Berry sampling and harvest.
Two 100-berry samples were collected from each of the treatment replicates in each of the two thinning trials. These samples were used for determination of berry weight and for analysis of Brix, titratable acidity (TA), pH, color, anthocyanins, and phenols. Sample collections occurred one day before harvest in a manner that was representative of the berries occurring naturally throughout the canopy. All samples were then stored at −25°C until analysis.
Winemaking.
Fruit from each of the Cabernet franc and Pinot noir blocks was separated by treatment replicate, crushed, destemmed, and fermented on the skins at 30°C for ~14 days. Fermentations were inoculated with Saccharomyces cerevisiae RC212 (Pinot noir) and D254 (Cabernet franc). Fermentations proceeded until dryness. Each treatment was then pressed using a bladder press to a pressure of 2.0 psi. Wines from each treatment were blended by block (blocks 1+2, 3+4, 5+6) into three replicates. The wines were then transferred to 20-L glass carboys and inoculated with a malolactic (ML) bacterial culture, and the ML conversion was allowed to proceed at 25°C over a two-week period. The young wines were sulfited to 50 mg/L, settled at 15°C, racked, resulfited, and cold stabilized at −2°C. In April 2002 and 2003, the wines were filtered using a 0.45-μm pad filter and bottled.
Berry, must, and wine analysis for Brix, TA, and pH.
Samples of 100 berries from each field-treatment replicate from 2001 and 2002 harvests were stored in the −25°C freezer until analysis. Approximately 50 berries from each sample were placed into 250-mL beakers, thawed, and heated to 80°C in a water bath to redissolve any precipitated tartrates. Samples were then cooled to room temperature and juiced using an Omega 500 fruit juicer. The collected juice was analyzed and the soluble solids were measured using an American Optical Abbé refractometer (AO Corp., Buffalo, NY) with temperature compensation. A 5.0-mL portion of juice was used to measure TA using a PC Titrate automated titration system (Mann-Tech, Guelph, ON) with standardized 0.1 N NaOH. pH was measured with an Accumet pH meter (Fisher Scientific, Mississauga, ON).
Must samples of 250 mL from each treatment replicate from the 2001 and 2002 harvests were also stored in the −25°C freezer analysis. Samples were heated to 80°C and analyzed for soluble solids, pH, and TA in the same manner as the berry samples. The remainder of the must was placed back into the −25°C freezer. Wine samples from each of the treatment replicates were used directly for measurement of pH and TA, also using the previously described methods.
Berry, must, and wine analysis for color, phenols, and anthocyanins.
Berry and must samples were heated and juiced as described for soluble solids, pH, and TA analysis. The juice was then cooled to 4°C in a refrigerator. Approximately 15 mL of each sample was centrifuged at 8500 rpm for 10 min at 4°C with an IEC Centra CL2 centrifuge (International Equipment Co., Needham Heights, MA). Clarified samples were then filtered through 0.45-μm HV Durapore syringe membrane filters (Millipore Corp., Bedford, MA). Wine samples were filtered to 0.45 μm during bottling and were therefore not subjected to the centrifugation and filtration process prior to analysis.
Color analysis was based on a modified version of the method provided by Mazza et al. (1999). Using 1.0-mm glass cuvettes, absorbance measurements were completed on all berry, must, and wine samples at 420 and 520 nm wavelengths against a blank using a Biotech Ultraspec 1000E UV/Vis spectrophotometer (Pharmacia, Peapack, NJ). Solutions of 20% glucose + 10 g/L tartaric acid and 10% ethanol + 10 g/L tartaric acid were prepared for use as blanks for all berry/must and wine samples, respectively. Color intensity and hue/tint were calculated using the following formulas: hue/tint = A420/A520; color intensity = A520 + A420. Values were expressed as 10-mm equivalents (that is, their equivalent values if measured in 10-mm cuvettes).
Anthocyanin measurements were completed on all berry, must, and wine samples using the pH shift method described by (Metivier et al. 1980). Two buffer solutions were prepared as follows: pH 1.0 buffer using 0.2 M KCl + 0.2 M HCl; pH 4.5 buffer using 1.0 M sodium acetate + 1.0 M HCl. Samples were diluted 10-fold with aqueous pH 1.0 and pH 4.5 buffer into test tubes and placed in the dark at room temperature for a one-hour reaction time. The samples were then placed in 10-mm glass cuvettes, and the absorbance was measured at 520 nm using the Biotech Ultraspec 1000E UV/Vis spectrophotometer. All sample measurements were blanked using the buffers. Total anthocyanin concentration was calculated as using the following formula:
Phenol analysis was completed on all samples by the phenolic reduction method using Folin-Ciocalteu reagent (Zoecklein et al. 1995). Phenol standard stock solution (5000 mg/L gallic acid; GA) was prepared by dissolving 500 mg of GA in distilled water to a final volume of 100 mL at room temperature. Calibration standards were prepared from the phenol stock solution to concentrations of 0, 50, 100, 150, 250, and 500 mg/L in 100-mL volumetric flasks. A 1.0-mL aliquot of each sample was diluted 10-fold and added to a 100-mL volumetric flask, along with 60 mL of distilled water, 5.0 mL of Folin-Ciocalteu reagent, and 15 mL of 20% sodium carbonate. The reactants were mixed well, brought to volume with distilled water, and allowed to incubate for two hours at room temperature. Using 10-mm glass cuvettes, the absorbance at 765 nm was measured with the Biotech Ultraspec 1000E UV/Vis spectrophotometer for each standard and sample against the 0 mg/L GA standard blank. A calibration curve was constructed from the absorbance values of the standards. Concentrations for the juice and wine samples were read from the calibration curve and multiplied by the dilution factor (10) to obtain the total phenol concentration (mg/L GA equivalent) in the original sample.
Canopy characterization.
Several variables were measured in order to assess the impact of shoot-thinning time on the vine canopy. This included point quadrat analysis, mean leaf area per shoot (main shoot and lateral shoot leaves), and cluster photosynthetic photon flux density. These measurements were all completed 3 to 4 weeks before the 2002 harvest.
Point quadrat analysis was completed using a 1.0-m rod by horizontal insertion into the canopy fruit zone (Smart and Robinson 1992). All leaf and cluster contacts on the rod per insertion were recorded and characterized with respect to whether they were a leaf or cluster. Twenty insertions were done at regular intervals for each treatment replicate. These data were used to determine mean leaf layer number, percent exposed and shaded leaves, and percent exposed, partially exposed, and shaded clusters using standard techniques (Smart and Robinson 1992).
To determine mean leaf area, one representative shoot per treatment replicate was collected for the control, double-prune (Cabernet franc only), and the two extreme shoot-thinning treatments (stages 09–12 and 29–31). Both primary and lateral leaves were removed from the shoots and the total primary and lateral leaf areas per shoot were measured using a LI-3100 leaf area meter (LI-COR, Lincoln, NE).
Cluster photosynthetic photon flux density (PPFD) was measured using a LI-COR LI-250 light meter. Twenty representative clusters in the canopy were selected from each treatment. The PPFD was measured by placing the light meter sensor parallel to the face of each cluster to obtain a light reading. Measurements were taken every hour for six hours on a cloudless day.
Wine sensory analysis.
Sensory descriptive analysis was performed on the 2001 wines. There were three preliminary taster-training sessions before the formal data collection. The first session consisted of a subset of three wines from each of the Pinot noir and Cabernet franc trials, during which aroma and retronasal (flavor) descriptors were generated. In the second session, the prospective panelists were provided with a prototype score sheet based upon the descriptors agreed upon during the first session, as well as a series of corresponding sensory aroma standards (Table 1⇓) that they were requested to evaluate for appropriateness (accuracy and intensity) together with a subset of three wines from each trial. In the third session, the sensory standards were confirmed and mutually agreed upon, and the ability of each panelist to detect each standard at a similar magnitude of intensity was verified.
Sensory data collection was carried out using Compusense software (Guelph, ON) with the assistance of 10 panelists. Both tasting replicates were evaluated twice by each panelist in flights of wines consisting of all treatments for each cultivar (six samples for Pinot noir and seven samples for Cabernet franc), for a grand total of eight formal tasting sessions overall. The scorecard was a 100-mm line score sheet. Preliminary bench tasting suggested that little or no differences were present among the 2002 wines, and therefore sensory analysis was not performed on that vintage.
Statistical analysis.
Statistical analysis was performed using the SAS statistical software package (SAS Institute, Cary, NC). All yield data, berry, must, and wine compositional data, canopy data, and sensory data were analyzed by analysis of variance for each of the field treatments. Single degree-of-freedom polynomial contrasts were tested on the equally spaced shoot-thinning treatments (09–12, 15, 19–21, 25–27, 29–31) in order to elucidate linear, quadratic, cubic, and quartic trends (Bowley 1999). The model excluded both the unthinned control and double-prune treatments. Multiple comparisons were also performed on all data using Duncan’s multiple range test at p ≤ 0.05. Dunnett’s t test (Dunnett 1955) was used to determine whether or not the individual means significantly differed from the control at p ≤ 0.05.
Results
Yield components.
In general, shoot thinning had relatively minor impacts on yield components over the two years of study. In Pinot noir, yield in 2001 was indeed reduced by all shoot-thinning treatments collectively (that is, the control versus thinning contrast was significant) and the two latest timings were lower than the control based upon Dunnett’s test (Table 2⇓). In 2002, a quadratic (parabolic) trend was observed in the five timings, whereby the 15 and 19–21 treatments were highest yielding, while the 29–31 treatment was significantly lower than the control (Table 2⇓). These responses were presumably driven by reductions in cluster number, since thinning treatments collectively and individually had fewer clusters than the control in 2001, and the treatments displayed a decreasing linear and quadratic trend with delayed timing of thinning, with fewer clusters than the control in the 29–31 treatment (Table 2⇓). Yield compensation effects were not apparent, since no differences were observed in cluster weights in 2001 and only two treatments differed from each other in 2002: 15 (lowest) and 25–27 (highest). Berries per cluster were also minimally affected; a slight decreasing linear trend with increased delay in thinning was observed in 2001, but no differences were found in 2002 (Table 2⇓). Berry weight, however, was substantially affected in both seasons. In both seasons, increasing linear and quadratic trends were elucidated with increased delay in thinning, and all treatments differed from the control; treatments were higher than the control in 2001 but lower in 2002 (Table 2⇓).
Periderm data indicated that thinning led to reductions in canes with “unripe” internodes. Canes with <3 mature internodes were reduced linearly with increased delay in thinning in both seasons, and those with 4 to 9 mature internodes were likewise reduced in 2002 (Table 2⇑). All thinning treatments collectively reduced cane numbers relative to the control in these categories during both seasons. Also, all individual treatments were lower than the control in 2001, while all but the 29–31 vines (<3 category) and the three latest (4 to 9 category) were lower than the control in 2002 (Table 2⇑). The category with >10 mature internodes contained the most desirable canes, but these were also reduced linearly with delayed thinning in 2001, and the latest treatment was lower than the control. In 2002, thinning treatments collectively reduced canes in this category, and two of five individual thinning treatments were less than the control (Table 2⇑). Weight of cane prunings (vine size) data were not available following the 2001 season, and minimal differences were observed in 2002, with only the 29–31 treatment being lower than the control in 2002 (Table 2⇑).
Shoot-thinning effects on Cabernet franc yield components were similar in magnitude to those on Pinot noir. Thinning treatments collectively had no effect on yield components in 2001 and no trend in the data was apparent; however, yields of the double-prune and 25–27 treatments were less than the control (Table 3⇓). However, in 2002, the greater magnitude of difference between the control and the various treatments in terms of number of shoots retained likely contributed to greater differences between treatments with respect to yield components. For instance, yield in 2002 decreased linearly with increasing delay in shoot thinning; collectively, thinning treatments decreased yield, and four treatments (double prune, 15, 25–27, 29–31) were different from the control (Table 3⇓). This pattern was similar in the clusters per vine data; no trends were evident in 2001 except for decreased cluster numbers in two treatments (09–12 and 19–21). However, in 2002, all treatments had fewer clusters than the control (Table 3⇓). Cluster weight showed some indications of yield compensation in both seasons. Thinning treatments collectively increased cluster weight over the control each year, and several individual treatments (09–12 and 19–21 in 2001; 09–12 to 25–27 in 2002) exceeded the control. Increasing delay in thinning decreased cluster weight (Table 3⇓). Shoot thinning impacted berries per cluster minimally in 2001 except for an increase in the 19–21 treatment. In 2002, increased delay in thinning decreased berries per cluster, although like cluster weight, all thinning treatments collectively exceeded the control and all treatments from 09–12 to 25–27 exceeded the control (Table 3⇓). Berry weight displayed a linear trend with increased delay in thinning in 2001, and the thinning treatments collectively as well as the double-prune treatment had lower berry weights than the control (specifically, the 25–27 and the 29–31 treatments). Only the double-prune and 19–21 treatments differed from the control in 2002; double prune had lower berry weights, whereas 19–21 had larger berries (Table 3⇓).
The number of canes with insufficient periderm formation (either the <3 or 4 to 9 internode categories) was reduced by thinning collectively as well as by double pruning (<3 in 2001 excepted)(Table 3⇑). Delaying thinning increased number of canes in the <3 category in 2002. No treatment effects were apparent in the >10 category in 2001, but thinning and double pruning reduced cane number in this category in 2002, and all but the 09–12 treatment decreased cane number relative to the control. Delaying thinning also linearly decreased cane number in the >10 category in 2002 (Table 3⇑). Weight of cane prunings decreased quadratically (2001) and linearly (2002) with increasing delay in shoot thinning, and both thinning and double pruning reduced vine size relative to the control in 2002 (Table 3⇑). The 09–12 treatment increased vine size slightly in 2001, whereas four treatments decreased vine size in 2002 (Table 3⇑).
Berry composition.
Soluble solids in Pinot noir were increased by thinning collectively in both seasons and were much higher than the control in the 09–12 and 15 treatments (2001) and all but the 09–12 (2002) treatments (Table 4⇓). A linear and quadratic trend was observed in 2002 whereby Brix increased relative to increased delay in shoot thinning. No thinning effects were observed for TA in 2001; however, in 2002, thinning treatments collectively reduced TA relative to the control, and all individual treatments were different from the control (09–12 and 15 were higher than the control and the three latest treatments were lower). A linear and quadratic trend indicated that delayed thinning decreased TA (Table 4⇓). Thinning treatments collectively increased pH in both seasons, and four treatments (09–12 to 25–27) exceeded the control in 2001; although four treatments were different from the control in 2002, only 19–21 and the 29–31 exceeded the control. Berry pH also increased in a linear and quadratic fashion with increased delay in thinning (Table 4⇓). No changes in hue/tint were observed across thinning treatments in 2001; in 2002, the 19–21 treatment increased hue (Table 4⇓). Color intensity, total phenols, and anthocyanins were also unaffected in 2001, but two treatments differed from the control in intensity in 2002, and intensity decreased linearly with delayed thinning (Table 4⇓). Thinning collectively increased phenols over the control in 2002, with the two earliest treatments exceeding the control, and a decreasing quadratic trend evident as thinning was delayed. For anthocyanins, the 19–21 treatment exceeded the control in 2002 (Table 4⇓).
In Cabernet franc, linear decreases in berry soluble solids relative to delayed thinning were observed in both seasons, while Dunnett’s test indicated higher Brix than the control in double prune in both seasons, the three earliest thinning treatments (2001), and the 15, 25–27, and 29–31 (2002) (Table 5⇓). Titratable acidity showed no treatment effects in 2001 but displayed a linear/quadratic decrease relative to delays in thinning in 2002, with the two latest treatments being less than the control (Table 4⇑). No significant trends were observed in pH in 2001, except for an increase in stage 25–27 compared to the control (Table 5⇓). In 2002, all but one treatment exceeded the control in pH, and a linear/quadratic trend was present whereby pH decreased relative to delayed thinning (Table 5⇓). Double-prune and later thinning treatments had higher hue/tint values compared to the control in 2001, but no effects were noted in 2002 (Table 5⇓). Delays in shoot-thinning time resulted in linear decreases in color intensity both seasons, with the highest intensity in the double-prune treatment. Thinning collectively increased intensity in 2002 and four thinning treatments exceeded the control (Table 5⇓). Phenolics were not strongly affected by thinning, although the double-prune treatment exceeded the control in 2001, while the 15 treatment was greater than the control in both seasons (Table 5⇓). Anthocyanins were minimally affected in both seasons except for a slight increase over the control by the double-prune treatment in 2001 (Table 5⇓).
Must composition.
For Pinot noir must composition, thinning collectively increased Brix in 2001, and the control was significantly different from all treatments except for stage 29–31 (Table 6⇓). Linear relationships were observed both seasons; in 2001, Brix decreased with delayed shoot thinning, but in 2002 the trend was reversed (Table 6⇓). Delayed shoot-thinning time led to a linear decrease in TA in 2001 and two treatments exceeded the control; however, in 2002, the trend was quadratic and one treatment (25–27) was less than the control (Table 6⇓). pH was unaffected in both seasons (Table 6⇓). No first- or second-degree polynomial trends were observed (cubic trends only in 2001) for both hue/tint and color intensity in both seasons, although one treatment exceeded the control in intensity each season (Table 6⇓). Values for total phenolics and total anthocyanins had no apparent noteworthy differences among treatments, with the exception of quadratic trends for both variables in 2002 (Table 6⇓).
In Cabernet franc must, no trends were observed for soluble solids in either season except for a slight decrease in the 19–21 treatment relative to the control (Table 7⇓). Titratable acidity was reduced by shoot thinning collectively in 2001, all individual treatments were less than the control, and the data displayed an increasing linear trend relative to delayed thinning that season; however, minimal treatment effects were observed in 2002 (Table 7⇓). Shoot thinning collectively increased pH over the control both seasons, and two treatments exceeded the control in 2001, while three were greater than the control in 2002 (Table 7⇓). An increasing linear/quadratic trend was apparent in 2002 relative to delayed thinning. Hue/tint was affected minimally by thinning (Table 7⇓). There were no trends observed for color intensity, total phenolics, or total anthocyanins (Table 7⇓).
Wine composition.
Pinot noir wine TA differed slightly among shoot-thinning treatments in 2001, but no clear patterns were apparent (Table 8⇓). Values for pH, hue/tint, and color intensity showed no significant differences in either season. Total phenolics only showed a difference between the control and the stage 25–27 treatment in 2001 and no differences in 2002 (Table 8⇓). Total anthocyanins varied somewhat between treatments in 2001, with the 09–12, 15, and 25–27 treatments exceeding the control (Table 8⇓).
In Cabernet franc, wine TA varied slightly across the shoot-thinning treatments in 2001, and the 29–31 treatment exceeded the control (Table 9⇓). Values for pH showed no significance in 2001, but three treatments exceeded the control in 2002 (Table 9⇓). All treatments differed from the control with regard to hue/tint, color intensity, and total phenolics in 2001, as did two treatments for total anthocyanins (Table 9⇓). In 2002, only one treatment exceeded the control in hue and intensity, and both total phenolics and anthocyanins were unaffected (Table 9⇓).
Canopy measurements.
Point quadrat analysis.
Shoot-thinning time in Pinot noir displayed a linear decrease in leaf layer number, with the two latest thinning treatments differing from the control (Table 10⇓). A cubic effect was observed for percentages of shaded and exposed leaves across thinning treatment times. Cluster exposure (full and partial exposure) increased linearly with thinning time with the two latest treatments differing from the control, while no shaded clusters occurred in any treatments (Table 10⇓). Leaf layer number in Cabernet franc displayed a linear decrease across thinning treatments (Table 11⇓). All thinning treatments, except for stage 09–12, were different from the control. The percentage of leaves exposed to sunlight also increased linearly, and shaded leaves decreased concomitantly, with all treatments differing from the control (Table 11⇓). Cluster exposure also increased linearly by thinning treatment, while partially exposed and shaded clusters displayed linear decreases (Table 11⇓). The double-prune treatment showed no difference in cluster exposure from the control (Table 11⇓).
Mean leaf area per primary shoot and lateral shoot.
In Pinot noir, shoot thinning at stage 09–12 resulted in the highest primary and lateral leaf areas per shoot (Figure 1A⇓). In Cabernet franc, none of the treatments differed from the control for primary leaf area per shoot (Figure 1B⇓). Differences were observed at the two extreme thinning times, whereby thinning at stage 09–12 resulted in the highest primary and lateral leaf area per shoot (Figure 1B⇓).
Photosynthetic photon flux density.
No significant trends were observed for Pinot noir at any time during the day across treatments (data not shown). However, the control PPFD value at 1300 hr was significantly higher than any of the shoot-thinning treatments (data not shown). Cabernet franc clusters displayed linear increases in PPFD across thinning times at 1000 hr and 1300 hr with no differences between double prune and control. No major differences were observed at any other time during the day (data not shown).
Sensory analysis.
Effects of treatments on Pinot noir were minimal and included quadratic trends for cherry aroma and significant differences among treatments for cherry, vegetal and black pepper aromas, prune and vegetal flavors, and finish (Figure 2⇓). Cherry aroma was highest in the 19–21 and 29–31 treatments, and these also had correspondingly lowest vegetal aromas and flavors (Figure 2⇓). Black pepper aroma was highest for 09–12 wines. The 29–31 treatment also had lowest prune flavor and longest finish (Figure 2⇓). In Cabernet franc, vegetal aroma and color displayed quadratic trends, and significant differences were found for cherry, dark berry and vegetal aroma, bell pepper flavor, and acidity (Figure 3⇓) as well as color (data not shown). The 19–21 treatment had highest dark berry aroma and the control and double-prune treatments had lowest vegetal aromas (Figure 3⇓). Among flavor attributes, the 09–12 treatment displayed highest bell pepper flavor (Figure 3⇓). Color was decidedly lowest for the control (data not shown); non-shoot-thinned wines had a mean of 40 on a 100-point scale, whereas 09–12, 15, and double-prune treatments ranged from 93 to 99, and other treatments ranged from 60 to 88.
Discussion
The objectives of this study were to examine the phenological stages at which shoot thinning might be most effective for improvement of canopy microclimate in dense vertical canopies. We hypothesized that the optimal phenological growth stages for this would occur after fruit set and that late shoot thinning or double pruning may result in reduced vigor of remaining shoots, leading to enhanced fruit exposure and composition. We suspected, moreover, that early shoot thinning might lead to increased vigor of remaining shoots and an enhancement of the growth of laterals, possibly resulting in a shaded canopy, producing fruit with lower quality in terms of its composition, and possibly producing wines of inferior quality.
Shoot thinning impacted canopy microclimate variables positively, although the magnitude of effect was not large. This is attributable to two reasons. First, vine size was considered moderate (Smart and Robinson 1992), with control vines in the two cultivars ranging from 10 to 22 g/cane, or 0.35 to 0.48 kg/m row. Second, leaf layer number (LLN) was not as high as those measured in extremely vigorous canopies; some LLN of vertical shoot-positioned canopies reported in the literature have exceeded 4.5 (Reynolds et al. 1996). Nonetheless, canopy measurements in both cultivars suggested that later shoot-thinning treatments and double pruning allowed for more sunlight penetration to the interior of the canopy than early season treatments. The stage 29–31 treatment resulted in the lowest leaf layer numbers at about 1.8 to 1.9 leaf layers. The leaf and cluster exposure results for Cabernet franc supported this LLN trend, as both percentage of exposed leaves and clusters also increased with thinning treatment time. Pinot noir cluster exposure showed the same linear trend as the Cabernet franc. Prior literature suggested that shoot thinning regardless of timing can reduce LLN (Reynolds et al. 1994a,c, Smart 1988). Our data supports the hypothesis that later shoot-thinning treatments would result in greatest control of shoot vigor. Leaf area data further support this hypothesis; the Pinot noir data showed an increase in mean leaf area on both main and lateral shoots when thinning was done at stage 09–12 in comparison with the control and stage 29–31 treatment, and Cabernet franc showed a similar trend with lateral leaf area per shoot. Reynolds et al. (1994a) showed that severe shoot thinning of Riesling at early stages in the growing season led to increased lateral shoot growth. Double pruning did not appear to control overall vegetative vigor sufficiently enough to impact the canopy variables measured in this trial. Cluster PPFD measurements on Cabernet franc indicated a better fruit zone light environment with delayed thinning time, with highest PPFD values seen at stage 29–31. This is consistent with shoot density studies whereby treatments imposed at veraison had a substantial impact on canopy microclimate (Smart 1988).
In general, shoot thinning had relatively minor impacts on yield components over the two years of study. Most of these differences are likely attributable to the fact that some fruit clusters were removed in the process of shoot thinning, which provided the potential either for yield reductions because of lowered cluster numbers or for yield compensation in the form of increased berry and cluster weights.
Soluble solids concentration in both Pinot noir and Cabernet franc berries and musts showed significantly higher levels at early shoot-thinning times (stages 09–12 and 15) compared with the control. The later thinning treatments were not beneficial, despite speculation that increased leaf and cluster exposure would favorably impact soluble solids levels (Bergqvist et al. 2001, Crippen and Morrison 1986a,b, Haselgrove et al. 2000, Morrison and Noble 1990). This phenomenon may have been due to a reduction in photosynthetic capability of the vines. As Miller et al. (1996a,b) indicated, leaving more shoots per vine later in the season reduces the shoot leaf area and length and results in fewer laterals. Therefore, removal of fully grown shoots along with fewer lateral shoots significantly decreases the vine’s total photosynthetic leaf area, resulting in less carbohydrate accumulation in berries (Vasconcelos and Castagnoli 2000). The higher soluble solids concentration in the double-prune treatment may have been due to the fact that a portion of the fruit was removed from the vine as a result of this treatment.
The linear decrease in TA relative to delayed thinning in berries and musts of both cultivars suggests an implication of cluster exposure (Bergqvist et al. 2001, Crippen and Morrison 1986a, Haselgrove et al. 2000, Morrison and Noble 1990), perhaps explained by an increase in malic enzyme activity in the more exposed fruit in the late-thinned treatments (Ruffner 1982). The response of pH is consistent with previous findings (Reynolds et al. 1994c), where reduced shoot densities resulted in higher pH values. However, Bergqvist et al. (2001) suggested that temperature may be a more influential factor on pH than actual sunlight exposure.
Hue/tint and total phenolics also showed no substantial differences in berries, musts, and wines of both grape cultivars, although prior studies have found higher phenols with increased fruit exposure (Crippen and Morrison 1986b, Mabrouk and Sinoquet 1998, Mazza et al. 1999, Price et al. 1995, Zoecklein et al. 1996). Color intensity in Cabernet franc showed a linear decrease with later thinning times. This may be in direct correlation with the decreased soluble solids levels, as there is a glucoside component in anthocyanin pigments which are responsible for color in grape skins (Haselgrove et al. 2000).
Conclusions
Shoot thinning improved canopy microclimate of Pinot noir and Cabernet franc grapevines, and thinning after bloom appeared to be superior to early season thinning in terms of leaf layer number, leaf exposure, cluster exposure, and reduction of lateral shoot growth. Shoot thinning and its timing had limited and inconsistent effects on yield components of both cultivars. Berry, must, and wine composition of Pinot noir and Cabernet franc were impacted, particularly Brix and TA, whereby both decreased relative to increased delay in thinning. Double pruning in Cabernet franc increased soluble solids and phenolics in berries as well as color variables and phenolics in wine. Despite differences in berry, must, and wine composition, impact of the thinning treatments on sensory attributes were relatively minor. Delaying shoot thinning or use of double pruning may therefore have positive impacts on canopy microclimate, but these effects are not of a great enough magnitude to influence wine quality.
Footnotes
Acknowledgments: The authors thank Sebastian Novello, Novello Vineyard, and both Michael East and Philip Clarke, East Dell Estate Winery, for their cooperation. Financial assistance from the Natural Sciences and Engineering Research Council and the National Research Council of Canada, and the Niagara Peninsula Fruit and Vegetable Growers Association is hereby acknowledged. Efforts of the many sensory panellists are also acknowledged.
This paper represents work included in the undergraduate thesis of TM. Presented at the 28th Annual Meeting, ASEV–Eastern Section, Corning, NY, July 2003.
- Received March 2005.
- Revision received June 2005.
- Copyright © 2005 by the American Society for Enology and Viticulture