Summary
Goals: To evaluate how variations in the extent of fruit-zone leaf thinning at prebloom or post-fruit set would affect crop yield components, cluster compactness, primary juice composition, and Botrytis bunch rot of Cabernet Sauvignon grown under Mid-Atlantic conditions in the United States. We anticipated that our results would guide leaf removal practice to target crop yield, primary juice composition, and disease management goals.
Key Findings:
Prebloom leaf removal reduced crop yield compared with no leaf removal, and by a greater extent when eight (59%) rather than four (25%) basal leaves per shoot were removed.
Prebloom leaf removal treatments reduced crop yield by 20 to 55% compared with no leaf removal in the inaugural season, and by 50 to 77% in the third consecutive season of treatment implementation.
Post-fruit set leaf removal of six basal leaves per shoot did not affect crop yield when compared to no leaf removal.
Prebloom leaf removal reduced cluster compactness compared with no leaf removal, and by a greater extent when eight (47% reduction) rather than four (24% reduction) basal leaves per shoot were removed.
Regardless of timing or extent, leaf removal resulted in 19 to 22% lower Botrytis bunch rot incidence compared with no leaf removal.
Treatment effects on juice composition were generally unremarkable, besides lower total titratable acidities observed in leaf removal relative to no leaf removal plots.
Impact and Significance: Our results illustrate that exposed grape clusters have lower bunch rot incidence and produce musts with similar pH and with lower titratable acidity than musts produced from shaded grapes. Prebloom leaf removal should be judiciously used, however, as severe crop reduction may result, particularly if prebloom leaf removal is seasonally repeated. Our results complement previous findings by (1) reporting on the effects of multiple extents of prebloom leaf removal over several consecutive growing seasons in Cabernet Sauvignon, a cultivar that is relatively resistant to bunch rots, and (2) by illustrating generally positive primary (herein) and secondary metabolite (our unpublished data) responses in harvested grapes.
Overview
The recommendation for fruit-zone management in the eastern United States is to aim for an average of one to two leaf layers, or roughly 50 to 80% of fruit clusters receiving direct sunlight exposure for at least a portion of the day.1 That recommendation was intended to present a conservative balance between recommendations that suggested (1) there is no limit to the amount of fruit exposure required to maximize grape and wine quality potential,2 and (2) alternative views that quality potential of red wine can be decreased by excessive sunlight exposure.3,4,5 However, the humid climate of the eastern US, particularly that of the southeastern states, increases fungal disease pressure compared with more arid climates. Thus, practices that promote open fruit zones are recommended in humid climates, both from the standpoint of disease management and increased wine quality potential.6,7,8,9
The impacts of the fruit-zone microclimate and of leaf removal on grape quality and disease management has been extensively researched.2,6,7,8,9,10,11,12 The evaluation of leaf removal timing, particularly prebloom (PB) thinning, has recently received increased attention because of its impact on source-sink relationships, crop yield, disease management, and fruit composition. PB leaf removal usually reduces crop yield, depending on the magnitude of leaf removal, cultivar, and experimental conditions.13,14,15,16,17,18,19 Although crop yield reduction might be considered a negative response, PB leaf removal has documented benefits, including accelerated soluble solids accumulation,14,16 reduced cluster compactness and Botrytis bunch rot incidence,14,16,18,19,20,21 and reduced risk of fruit sunburn compared with midseason or later leaf thinning.15,22 The potential crop reduction due to reduced fruit set could be offset by reduced disease and more harvested crop where PB leaf removal is practiced. This might particularly be the case in rot-sensitive cultivars such as Pinot noir, Vignoles, and Chardonnay. PB leaf removal could thus be used as a cultivar-specific and production goal-dependent vineyard management tool to improve rot management and fruit composition in humid, subtropical winegrowing regions.
We hypothesized that fruit-zone leaf thinning could improve disease management without compromising fruit composition in Cabernet Sauvignon in a humid region. 20,21 We were also interested in evaluating the effect of seasonally repeated PB leaf thinning on vine capacity.
Major Observations and Interpretations
Experimental setting and treatments.
Two separate experiments were conducted on Cabernet Sauvignon clone #337. The first experiment compared three intensities of PB leaf thinning. The second experiment compared two intensities of leaf thinning at post-fruit set (PFS). The vines were planted in 2006 in 3.0-m-wide rows and spaced 1.5 m apart in rows oriented northeast/southwest and trained to low, bilateral cordons with vertical shoot positioning. Additional details are in the Experimental Design. Treatments were as follows:
PB leaf removal experiment. PB-NO (no fruit-zone leaves or lateral shoots removed), PB-4 (removal of leaves/lateral shoots from primary shoot basal nodes 1 to 4 at modified Eichhorn-Lorenz [E-L] stage23 17 [12 leaves separated, inflorescences well developed, and single flowers separated]), and PB-8 (removal of leaves/lateral shoots from primary shoot basal nodes 1 to 8 at modified E-L stage 17). The PB-4 and PB-8 treatments were annually applied to the same vines over three seasons to evaluate repeated treatment impact on crop yield components, cluster compactness, primary juice composition, and cane pruning weights. A depiction of seasonal treatment implementation and preharvest berry sampling are shown to help clarify crop yield responses described below.
PFS leaf removal experiment. PFS-NO (no fruit-zone leaves or lateral shoots removed) or PFS-6 (removal of leaves/lateral shoots from primary shoot nodes 1 to 6 at modified E-L stage23 29 [berries have peppercorn size and bunches tend downwards]). These treatments were applied to the same vines in 2014 and 2015 to evaluate treatment impact on crop yield components, primary juice composition, and cane pruning weight.
Meteorology.
Seasonal growing degree day (GDD) accumulation in 2015 was ∼8% greater than in 2013 and 2014 because of greater GDD accumulation in the postveraison period (Aug and Sept) of 2015 (Table 1). Monthly mean temperature was greatest in June 2013 and in Aug and Sept 2015 compared with the same months in the other two years. Monthly maximum temperatures were greatest in June and Oct 2013 and in Aug 2015 compared with the same months in the other years. Seasonal rainfall in 2013 was 25% higher than in the 2014 and 2015 seasons. Average relative humidity was 84, 75, and 75% in July, Aug, and Sept 2015, respectively (data not shown).
Seasonal and monthly growing degree day (GDD) and rainfall accumulation, and mean and maximum monthly temperatures for July, Aug, Sept, and Oct at the Alson H. Smith, Jr. AREC in Winchester, VA in 2013 to 2015.
Dormant cane pruning weight, shoot fruitfulness, and Botrytis bunch rot incidence at harvest.
Mean pruning weight during 2013 to 2015 was reduced by 30% by PB-8 and by 11% by PB-4 compared with PBNO (Table 2). Although PB-8 reduced pruning weights to the greatest extent, pruning weights from all treatments were above the range for the previously defined “optimal vine size” (0.3 to 0.6 kg/m/row),2 particularly considering the shoot-hedging required of all treatments. The treatment × year interaction was significant as follows: pruning weight was not affected in 2013, PB-8 reduced pruning weight by 0.3 kg/m/row (35%) in 2014 and by 0.7 kg/m/row (37%) in 2015 compared with PB-NO, while PB-4 depressed pruning weight by 0.3 kg/m/row (21%) compared with PB-NO, but only in 2015 (Figure 1). Our findings that PB leaf removal can reduce pruning weight was consistent with previous work,14 and that response was likely due to a seasonal net depletion of carbohydrate reserves from permanent vine organs.24 We predict that PB leaf removal can reduce vine size if repeated over years, whereas modest PFS leaf thinning is less likely to affect vine size. The 2015 season was marked by a general increase in vine pruning weights compared with those measured in the 2013 and 2014 seasons (Figure 1). An average of 40 mm greater rainfall was recorded between 1 April and 30 June in 2015 compared with the same period in 2013 and 2014 (data not shown); the greater moisture might have contributed to the general increase in cane pruning weights observed in that year. However, the cluster number data, below, suggested that more shoots per vine were retained in 2015 relative to earlier years; this too would have increased pruning weight.
Prebloom (PB) and post-fruit set (PFS) leaf removal effects on dormant cane pruning weight averaged over the 2013 to 2015 seasons and Botrytis incidence in 2015.
Prebloom (PB) leaf removal effect on dormant cane pruning weight over 2013 to 2015. PB-NO, PB-4, and PB-8 = PB removal of no, four, or eight leaves per shoot, respectively. Means within the same year not sharing the same letter were statistically significantly different, and means within the same year without letters were not statistically significantly different (α ≤ 0.05, Tukey’s honest significant difference). Error bars indicate the standard error; data points represent means of six replicates. Note: 2012 data indicate pre-experiment pruning weights, for comparison.
Count and basal shoot fruitfulness were not affected by the PB leaf removal treatment. Basal shoot fruitfulness was 124% greater in 2014 (0.66 clusters/shoot on average) than in 2015 (0.29 clusters/shoot on average) when averaged over the PB treatments (data not shown); this did not affect crop yield across years, as basal shoots were thinned out each season. Increased radiation can improve bud fruitfulness,25,26 potentially countering the negative impact that removal of source tissues can have on bud fruitfulness.24 Those opposing effects may partially explain why PB leaf removal reduces fruitfulness in some,19,21 but not all, cases.14,15,17,27 Botrytis bunch rot incidence was significantly reduced by all leaf removal treatments in both experiments (Table 2). This reduction was likely a function of the reduction in cluster drying time and improved canopy spray penetration associated with an open fruit zone.1,7,9,28
Crop yield components.
Yearly data were averaged by treatment (Table 3). We observed significant treatment × year interactions, and several of these responses are shown in Figure 2. When compared with PB-NO, PB-4 reduced crop yield by 0.9 kg/vine (25%), and PB-8 reduced crop yield by 2.1 kg/vine (59%) (Table 3). The fact that cluster number was not reduced by either PB-4 or PB-8 supported the notion that the aforementioned reduction in pruning weight was not simply a function of fewer shoots (and thus clusters) per vine. PB removal of four leaves reduced cluster weight by 35 g (34%), and PB-8 reduced cluster weight by 60 g (59%), compared with PB-NO. Berry number per cluster was reduced by PB-4 by 28 berries (36%), and reduced by PB-8 by 42 berries (55%) compared with PB-NO. Berry weight was reduced by PB-8 compared with both PB-NO and PB-4. There were treatment and year differences in crop loads; low crop loads reflected under-cropping and the vigorous vegetative growth that is characteristic of Cabernet Sauvignon in this region.
Prebloom (PB) and post-fruit set (PFS) leaf removal effects on crop yield components and crop load averaged over the 2013 to 2015 seasons (PB) or over the 2014 to 2015 seasons (PFS).
Prebloom (PB) leaf removal effect on crop yield (A), cluster weight (B), berry number per cluster (C), and berry weight (D) at harvest over 2013 to 2015. PB-NO, PB-4, and PB-8 = PB removal of no, four, and eight leaves per shoot, respectively; PB-NO est. = estimated responses if extra berry samples were not taken from PB-NO in 2014. Means within the same year not sharing the same letter were statistically significantly different, and means within the same year without letters were not statistically significantly different (α ≤ 0.05, Tukey’s honest significant difference). Note: PB-NO est. was not statistically analyzed. Error bars indicate the standard error; data points represent means of six replicates.
PB leaf removal’s influence on crop yield was through a reduction in berry number per cluster that equated to lighter clusters and lower crop yield. Previous work has indicated a positive relationship between the number of leaves removed at trace bloom and the extent of cluster weight reduction in different cultivars.21 Previous studies have also indicated that fruit set can be reduced when carbohydrates are limited by flowering,24,29,30 but fruit set was not reduced until eight (rather than four or six) leaves were removed at flowering.31 PB leaf removal can nonetheless reduce fruit set (8 to 37%), berry number per cluster (14 to 64%), cluster weight (20 to 69%), and crop yield (30 to 71%) across a wide range of conditions, cultivars, and application rates.13,14,15,16,17,18,19 Our results with Cabernet Sauvignon are consistent with results from those studies.
Both PB leaf removal treatments tended to further depress crop yield over time compared with PB-NO; we attributed this to a decrease in crop yield in the PB-4 and PB-8 plots and to an increase in crop yield in the PB-NO plots (Figure 2). The PB leaf removal treatments reduced crop yield by 22 to 30% more, cluster weight by 21% more, and berry number per cluster by 17 to 20% more when compared with PB-NO in 2015 than when compared with PB-NO in 2013. Trends indicated increasingly depleted carbohydrates in the PB leaf removal treatments over time. It was thus possible that the greater reduction in berry number per cluster in 2015 relative to 2013 was due to carbohydrate-limited fruit set reduction.24,29,30 Although we were unable to verify this assumption, repeated PB leaf removal may have reduced the number of flowers per inflorescence over time. Although some PB leaf removal studies have observed no further reduction in crop yield over multiple years,16,27 others have reported trends in crop yield reduction, as those observed in our study.14,17,18,19,21 Data from the two experiments described here illustrated that PB leaf thinning reduces berry number per cluster, cluster weight, and crop yield, but that leaf removal after fruit set has negligible effects on crop yield components.
The significant leaf removal treatment × year interactions reflected the convergence of crop yield component means between treatments in 2014 (Figure 2). This was due to frequent preharvest berry sampling from PB-NO plots, but not from PB-4 or PB-8 plots, in 2014. The estimated crop yield components (see Crop yield components in Experimental Design) in the PB-NO plots in 2014 are shown in Figure 2 for visual comparison. These estimations revealed that: (1) 36% of the total number of berries per cluster (an average of 35 of 96 berries per cluster) were removed from PB-NO plots in 2014, (2) 33% of the total estimated crop weight at harvest (an average of an estimated 1.26 kg of an average of an estimated 3.81 kg of crop) was removed from PB-NO plots in 2014, and (3) the leaf removal treatment effects on crop yield components in 2014 would have more closely paralleled those in 2013 and 2015, had the extra berry samples not been taken from PB-NO plots throughout the 2014 season. When we added the estimated crop yield data from 2014 PB-NO plots into the mixed model, we detected no treatment × year interaction for crop yield, and PB-NO had greater crop yield than PB-8, and greater berry number per cluster and cluster weight than both PB-4 and PB-8 (data not shown).
Seasonal berry weight development.
Seasonal berry weight development was not affected by the PFS leaf removal. PB leaf removal treatment and date both significantly affected (Prob > F < 0.0001) seasonal berry weight development in 2013 and 2015 (Figure 3). Berries weighed less in PB-8 plots than in the PB-NO and PB-4 plots, and those differences were greater after than before veraison, which occurred on 21 Aug 2013 and 13 Aug 2015. There was a significant treatment × date interaction in 2015 (Prob > F = 0.0002) such that berry weight was not affected by treatment on 30 June 2015, but on every sample date thereafter (Figure 3). Our results confirmed that aggressive early defoliation can reduce berry size. The reduction in berry size may have been a function of decreased translocation of leafderived hormones to the grapes from early defoliated treatments,32 but may have also been due to a reduction in berry cell division, which occurs during the first two weeks after flowering.33
Prebloom (PB) leaf removal effect on seasonal berry weight development in 2013 (A) and 2015 (B). PB-NO, PB-4, and PB-8 = PB removal of no, four, and eight leaves per shoot, respectively. Means within the same date not sharing the same letter were statistically significantly different, and means within the same date without letters were not statistically significantly different (α ≤ 0.05, Tukey’s honest significant difference). Error bars indicate the standard error; data points represent means of six replicates. Note: berry samples were not evenly collected from treatments and, therefore, were not analyzed in 2014.
Our results further indicate that PB removal of more than four primary basal shoot leaves was necessary to depress Cabernet Sauvignon berry weight development under our experimental conditions. Previous research has indicated that PB removal of six leaves consistently reduces both berry number per cluster and berry weight development in Petit Verdot, but that these effects are less consistent in Cabernet franc.34 Some PB leaf removal studies have reported a reduction in berry weight,17,18 whereas others found that berry weight is unaffected as a compensatory response to a tandem reduction in berry number per cluster.27 If the reduced berry weight afforded by PB leaf removal increases the skin-to-pulp ratio, then PB leaf removal may be a practical tool to increase skin-derived sensory impact compounds in red wine musts. Our results and those of others show that the extent of PB leaf removal can differentially impact berry weight, but that responses may vary across cultivars.
Components of cluster compactness.
Cluster compactness was reduced by PB-4 (24%) and by PB-8 (47%) compared with PB-NO; this was a function of PB leaf removal’s reduction in berry number per cluster (Table 4), as rachis lengths were unaffected (data not shown). The leaf removal treatment × year interaction showed very similar trends in berry number per cluster (Figure 4) to those shown in Figure 2C. Cluster compactness was thus unaffected by the leaf removal treatments in 2014, but was reduced in PB-8 when compared with PB-NO in both 2013 and 2015. These observations were, again, due to the preharvest berry sampling that occurred in PB-NO, but not in PB-4 or PB-8, plots in 2014.
Prebloom (PB) leaf removal effect on berry number and cluster compactness of the same clusters at harvest averaged over the 2013 to 2015 seasons.
Prebloom (PB) leaf removal effect on berry number per cluster (A) and cluster compactness (B) of the same randomly selected clusters at harvest, and on the relationship between cluster compactness and Botrytis bunch rot incidence at harvest in 2015 (C). PB-NO, PB-4, and PB-8 = PB removal of no, four, and eight leaves per shoot, respectively. Means not sharing the same letter within the same year were statistically significantly different, and means within same year without letters were not statistically significantly different (α ≤ 0.05, Tukey’s honest significant difference). Error bars indicate the standard error; data points represent means of six replicates.
We found a positive relationship between cluster compactness and Botrytis bunch r ot i ncidence at harvest (Figure 4), as has been observed in previous studies.14,16,18,19,20,21 Although PB-8 had significantly less compact clusters than PB-4, neither treatment had measurable Botrytis bunch rot incidence. The data distribution in Figure 4 suggests that the zero leaf layers created by both PB-4 and PB-8 (unpublished data) was of primary importance for reducing Botrytis, as documented in previous studies.7,9,24 That argument was strengthened by the fact that PFS-6 also reduced Botrytis incidence (Table 2), but unlikely reduced cluster compactness (Table 3) compared with no leaf removal. We hypothesize that less compact clusters were of secondary importance for reducing Botrytis given the greater absolute and percentage-based reduction in Botrytis in PB-4 and PB-8 plots relative to PFS-6 plots (Table 2). The earlier implementation of PB versus PFS leaf thinning might have further reduced Botrytis incidence via improved fungicide canopy penetration at both flowering and fruit set. Bunch rot in Pinot noir was not controlled in a humid region until clusters were loosened by removal of eight and 10 (rather than four or six) leaves at flowering.31 Our results may differ from those conducted with other cultivars because Cabernet Sauvignon is less susceptible to fruit rots than more tightly clustered cultivars.
Primary fruit composition.
Aside from a measurable impact on fruit titratable acidity (TA), the impact of leaf removal on primary fruit composition was modest (Table 5). PB-8 reduced Brix values on average by 3% compared with PB-4 and PB-NO, but PFS leaf removal did not affect Brix. Although PB-8 had numerically lower Brix in each season, within-year analyses revealed that PB-8 reduced Brix only when compared with PB-4 in 2013 (data not shown). PB leaf removal has been reported to increase Brix by several mechanisms, including increased leaf area-to-fruit weight ratios, improved canopy and water use efficiency, crop load regulation, restored and more efficient leaf area due to lateral shoot regrowth, and hastened translocation of assimilates to the fruit.14,15,16,17,19,27 It is possible that our results differ from those of previous studies because (1) removal of eight leaves coupled with aggressively managed lateral shoot growth throughout the season restricted leaf area to a point that fruit ripening was delayed, or (2) the postveraison weather patterns in humid regions may result in comparably slower ripening than in regions where PB leaf removal increases Brix.14,15,16,17,19,27
TA was reduced by PB-4 by 8% and by PB-8 by 11% compared with the 7.96 g/L TA recorded in PB-NO (Table 5). Within-year analyses revealed that TA was reduced by both PB-4 and PB-8 only in 2015, when TA was 6.60 g/L in PB-8, 6.73 g/L in PB-4, and 7.81 g/L in PBNO (data not shown). PFS removal of six leaves (PFS-6) had a TA of 6.16 g/L, a 17% reduction compared with the TA observed in PFS-NO. Exposed grapes often have lower acidity as a function of radiant heat-driven increases in malic acid respiration rates.10,35 The 2015 season had a relatively warm postveraison period (Table 1), which might explain the lower TA values observed in 2015 than in 2013 in the PB experiment (Prob > F < 0.0001), and when compared with 2014 in the PFS experiment (Prob > F < 0.0001) (data not shown). Juice pH was unaffected by PB leaf removal, was increased by PFS-6 compared with PFS-NO (Table 5), and was greater in 2014 than in 2013 (in the PB leaf removal experiment) and 2015 (in both experiments) (data not shown).
Prebloom (PB) and post-fruit set (PFS) leaf removal effects on juice soluble solids, pH, and titratable acidity at harvest for the 2013 to 2015 seasons (PB) or for the 2014 to 2015 seasons (PFS).
Broader Impact
Fruit-zone leaf thinning has wide application, yet industry questions persist as to “How much fruit exposure is desirable?” and “What is the optimal timing to achieve the desired effects?” Answers to these questions depend upon whether grapes are grown in hot, arid climates or cooler, cloudier wine regions. When leaves were removed before bloom, Botrytis was reduced to a greater extent than when leaves were removed after fruit set, which suggested that both exposure and cluster architecture affect Botrytis incidence. The questionable robustness of one year of Botrytis data may be negated by the fact that treatments reduced rot in Cabernet Sauvignon, which generally has greater bunch rot resistance than other vinifera cultivars. Crop yield was not affected when leaves were removed after fruit set, but was reduced when leaves were removed before bloom.
PB leaf removal reduced crop yield by a greater extent in the third consecutive year than in the first year of treatment implementation. Considering treatment impact on crop yield and Botrytis incidence at 5% severity in 2015, an informal economic assessment in that year estimated that PB leaf removal reduced total crop revenue by $4589 to $7325 per acre compared with no leaf removal, but the crop value between PFS leaf thinning and no leaf removal plots was essentially equal. Perennial PB leaf removal reduced vine capacity. Here we evaluated a high-capacity cultivar in a nonlimiting environment, and similar treatments might therefore cause substantial reductions in vine capacity if practiced in low-capacity cultivars grown in resource-limited conditions. PB leaf removal should be judiciously practiced to avoid undesirable reductions in vine capacity and crop yield. The most consistent response of juice composition to leaf removal was a decrease in TA. Grape sunburn was not observed, and this was perhaps a function of skin acclimation to ambient radiation and temperature from a relatively early stage of berry development (peppercorn berry size or earlier), or due to the relatively cloudy climate in which this experiment was conducted.
Taken together with unpublished (phenolics and anthocyanins) data, our results illustrate that PFS leaf removal offers benefits for disease management and fruit composition that are similar to those of PB leaf removal, without the potential for crop reduction. We can therefore recommend PB leaf thinning only if the earlier thinning was logistically superior to PFS thinning, or if the anticipated crop yield reductions of modest PB leaf thinning were desired as a crop regulation tool alternative to cluster thinning.
Experimental Design
Experimental vineyard and treatments.
Experiments used Cabernet Sauvignon ENTAV-INRA clone 337 vines, grafted onto 420-A rootstock and grown at Virginia Tech’s Alson H. Smith, Jr. Agricultural Research and Extension Center near Winchester, VA (39°11′N; 78°28′W). Vines were planted in May 2006 in rows running northeast/ southwest (39°/219°) at a 3.0 m (row) × 1.5 m (vine) spacing, and were trained to bilateral cordons with vertically positioned shoots. The soil was a Poplimento Hagerstown sandy loam (A. Blackburn, personal communication, 2013). The interrow groundcover, established in 2001, initially comprised a mixture of orchard grass (Dactylis glomerata) and tall fescue (Festuca arundinacea) cv. ‘Shenandoah,’ with the fescue dominating after about six years. Basal and count shoot thinning was implemented at E-L stage 13 (six leaves separated) to standardize the count shoot number across all treatment plots to ∼12 to 13 shoots/m of cordon. Shoot topping occurred before shoots fell over the top catch wire to avoid canopy selfshading. Treatment integrity was maintained from initiation through harvest by weekly revisits to experimental plots to remove vegetation infringing on the fruit zone. Pest and nutrient managements were standard for the region and were uniformly applied to all treatments and experimental replicates.
Treatments were applied in a randomized complete block design and replicated in six blocks. Experimental units were one vine for the PB leaf removal trial and two vines for the PFS leaf removal trial. The experiments were conducted separately because of the limited number of similarly cultivated experimental units available in the research vineyard. The intrarow groundcover in the PB leaf/lateral shoot removal experiment consisted of perennial creeping red fescue (Festuca rubra), established in Sept 2008, and an 85-cm-wide herbicide-treated strip in the PFS leaf/lateral shoot removal experiment. Treatments were as follows:
PB leaf removal experiment. PB-NO (no fruit-zone leaves or lateral shoots removed), PB-4 (removal of leaves/lateral shoots from primary shoot basal nodes 1 to 4 at modified E-L stage23 17), and PB-8 (removal of leaves/lateral shoots from primary shoot basal nodes 1 to 8 at modified E-L stage 17). The PB-4 and PB-8 treatments were annually applied to the same vines from the 2013 through 2015 seasons to evaluate repeated treatment impact on crop yield components, cluster compactness, primary juice composition, and cane pruning weight.
PFS leaf removal experiment. PFS-NO (no fruit-zone leaves or lateral shoots removed), PFS-6 (removal of leaves/lateral shoots from primary shoot nodes 1 to 6 at modified E-L stage23 29). These treatments were applied to the same vines in 2014 and 2015 to evaluate repeated treatment impact on crop yield components, primary juice composition, and pruning weight.
Meteorology.
Seasonal rainfall, temperature, and relative humidity were recorded over 2013 to 2015 with research weather stations located within 300 m of the research site. GDD were summed from 1 April using a base temperature of 10°C.
Dormant cane pruning weight, shoot fruitfulness, and Botrytis bunch rot incidence.
The weight of pruned canes was measured on a per-vine basis with a field scale during the dormant periods that followed the 2013 to 2015 growing seasons. Shoot fruitfulness was indexed as inflorescence number per shoot at the modified E-L stage 15/16 in 2014 and 2015. The number of inflorescences were divided by the number of shoots originating from both the cordon (basal) and one-year-old spurs (count). Botrytis bunch rot incidence was estimated by visual inspection of every harvested cluster in 2015 only. A ≥5% estimated severity per cluster resulted in an “incident” for that cluster, but severity was anecdotally observed to be greater than 5% in most cases. Percent incidence was calculated as the number of clusters with Botrytis divided by the total number of harvested clusters in each experimental unit.
Crop yield components.
Crop was harvested at commercially acceptable composition and integrity. Crop yield per vine was measured with a field scale at harvest on 9 Oct 2013, 20 Oct 2014, and 5 Oct 2015 to obtain yield components. Cluster weight was estimated from the quotient of crop yield weight and cluster number. Berry weight was determined from 120 berry samples, half of which were randomly collected from each of east and west canopy sides immediately prior to harvest. Berry number per cluster was derived from the quotient of average cluster weight and average berry weight. In 2014, the second year of the PB leaf removal experiment, the PB-4 and PB-8 treatments were not berry-sampled until harvest, while the PB-NO plots were sampled five times throughout the season. Thus, because of the inconsistency of berry sampling across treatments, the crop yield relationship among treatments was unique in 2014 compared with 2013 and 2015. To “correct” for the extra number of berries that were sampled before harvest, crop yield components were estimated in PB-NO plots (PB-NO est.) as follows: the product of the average berry weight at harvest and total number of berries sampled before harvest was summed with the total harvested crop weight; estimated crop yield components were then calculated as described above. Crop load was expressed as the ratio of crop yield weight to dormant cane pruning weight, and calculated on a per-vine basis.
Seasonal berry weight development.
Seasonal berry weight development was evaluated in 2013 and 2015 in the PB leaf removal experiment, and in 2014 in the PFS leaf removal experiment. Berry weight development was determined from randomly collected, composite 120 (30 Aug and 9 Oct 2013 and 16 July, 2 Aug, 8 Sept, and 5 Oct 2015) and 200 (9 and 24 July and 9 Aug 2013 and 30 June 2015) berry samples from PB leaf removal experimental units, and from randomly collected, composite 120 berry samples from PFS leaf removal experimental units on 14 Aug, 12 Sept, 30 Sept, and 20 Oct in 2014. Average berry weight in each sample was determined by dividing the total composite sample weight by the number of berries contained in the sample.
Components of cluster compactness.
Cluster compactness was measured at harvest on 10 clusters per experimental unit in the PB leaf removal experiment only over 2013 to 2015. Clusters were randomly selected from harvest lugs and weighed until 10 clusters were obtained, each of which weighed within ±25% of the mean harvested cluster weight for each respective experimental unit. Compactness was indexed on each of the 10 clusters by dividing the total berry number by the summed length of the main rachis and top two branches from the main rachis. Rachis “wings” or “shoulders” were not included in the compactness assessment.
Primary fruit composition.
Primary composition was measured on juice obtained from combined, composite samples of 30 east and 30 west canopy side-positioned berries that were randomly collected from each experimental unit immediately prior to harvest over 2013 to 2015. Juice samples were obtained from fresh (nonfrozen) berries by equally hand-pressing samples in a plastic bag for about two minutes, and then immediately centrifuging them for five minutes at ∼3500 rpm. Soluble solids were measured with a digital refractometer (Pocket PAL-1, ATAGO USA, Inc.). Juice pH was measured and TA was determined by titration to an endpoint of pH 8.2 with an 848 Titrino Plus autotitrator (Metrohm USA) and 0.1 N NaOH.
Statistical analysis.
JMP Pro v. 13 was used for statistical computation. Mixed models were used to evaluate the random effects of block, block × year, and block × date; the fixed effects of treatment, year, and date; and treatment × year and treatment × date interactions. It was unnecessary to include all model effects and interactions to evaluate every response (e.g., Botrytis, berry weight development). Significant mean separation (α ≤ 0.05) was determined with Tukey’s honest significant difference (HSD) for the PB leaf removal experiment. Student’s t test was used to determine mean separation (α ≤ 0.05) for the PFS leaf removal experiment.
Tukey’s HSD was used to determine significant within-year mean separation (α ≤ 0.05) when analyzing the leaf removal treatment × year interactions. Tukey’s HSD was also used to determine significant within-date separation when analyzing the leaf removal treatment × date interactions for seasonal berry weight development, which was analyzed separately within each of 2013 and 2015. The estimated crop yield responses from PB-NO in 2014 were not statistically analyzed, as these were simply estimations. The bivariate fit function in JMP Pro v. 13 was used to run simple linear regression (α ≤ 0.05) to determine the relationship between cluster compactness and Botrytis bunch rot incidence. When data were not presented, “(α ≤ __)” was used in the text to indicate a significant separation between treatment means, and “(Prob > F = __)” was used in the text to indicate a significant fixed-treatment effect within the model.
Acknowledgments
The authors gratefully acknowledge funding for this work from the Virginia Wine Board; Virginia Agricultural Council; and the National Institute of Food and Agriculture, U.S. Department of Agriculture, under agreement no. 2010-51181-21599. Appreciation is also extended to Rachael White, Trevor Wolf, Danielle Bunce, Dana Melby, Brycen Hill, and Hannah Kasabian for their collective vineyard and laboratory assistance, to Jerry Davis for the statistical assistance, and to Dr. Matthew Chappell for his review efforts.
- Received April 2018.
- Revision received July 2018.
- Accepted July 2018.
- Copyright © 2018 by the American Society for Enology and Viticulture. All rights reserved.