Abstract
Vine performance was tested over five years (2005–2009) on Vitis vinifera L. cv. Barbera either manually spur-pruned (HP) or mechanically hedged with light (SMP-LF) or severe (SMP-SF) hand follow-up. Although mechanical treatments retained 2- to 2.5-fold higher count nodes per vine than did HP, yield per vine (~5 kg) was almost identical between treatments due to the strong offsetting effect of reduced budbreak. Weak compensation and no compensation were seen for cluster weight and bud fruitfulness, respectively. Except for a slight reduction in anthocyanin concentration, overall grape composition was similar among treatments throughout the trial. As minor differences in vine vigor and capacity were found and the leaf-to-fruit ratio (vine basis) was unaffected by treatments, the slightly lower anthocyanin berry content in the SMP vines may have derived from increased shoot density and, hence, more shade cast in the fruiting area. Winter pruning was performed in less than 25 hr/ha in the hedged vines, thereby cutting labor demand from 54 to 70% compared with HP. Thus, if all other vineyard operations are also mechanized, a single high-wire Barbera vineyard with a mostly erect canopy can be maintained in less than 70 worker hr/ha. Such a performance, coupled with overall unchanged yield and grape composition, represents a solid and reliable approach in a wine market that demands greater efficiency and competitiveness.
Winter pruning of grapevines is a time-consuming operation, as it requires ~60 to 120 worker hr/ha depending on vine vigor, trellis type and design, equipment, and skilled labor (an increasingly short commodity in many viticultural areas worldwide) (Brancadoro and Marmugi 1997, Corradi 2010, Intrieri and Poni 1995). In mechanically harvested, vertical shoot-positioned (VSP) training systems, hand pruning accounts for up to 75% of the yearly worker hours requirement, or 41% of total overhead (Corradi 2010). Such a high labor demand can be easily trimmed by 50 to 90% if a form of mechanical pruning is used, the time savings primarily dependent on the extent of hand clean-up (Intrieri and Poni 2000). Winter mechanical pruning appears to be facilitated when performed on single, high-wire trellises with free-growing vegetation concentrated in the upper 180° hemisphere of the cordon. This pattern renders stake and post-avoiding devices virtually unnecessary, thereby also reducing the need for hand clean-up, which is essentially limited here to removing ventral spurs. The most distinguishing trait of mechanically pruned vines is that their retained node number is usually largely increased compared with hand pruning. In effect, suitability to mechanical pruning is primarily a function of vine capacity for yield self-regulation, which is usually manifested as reduced budbreak rates (Poni et al. 2004, Studer et al. 1980), clusters per shoot (Clingeleffer 1993, Martinez de Toda and Sancha 1999), berry set (Jackson et al. 1984, Possingham 1994), cluster and berry weight (Freeman and Cullis 1981, Intrieri et al. 1988), and even self-pruning due to abscission of immature wood (Clingeleffer 1993, Morris 1993, Tassie and Freeman 1992). If sufficient yield compensation is also coupled in mechanically pruned vines to higher canopy efficiency (i.e., earlier canopy formation and larger leaf areas via increased shoot number), then this technique can deliver similar or even higher yields than hand pruning while maintaining unaltered grape composition (Clingeleffer 1993, Morris and Cawthon 1981, Keller et al. 2004, Zabadal et al. 2002).
However, some long-term field trials on mechanical pruning have shown that highly fruitful hybrids (Fisher et al. 1997, Pool et al. 1988) and some Vitis vinifera cultivars (Carbonneau and Zhang 1988, Di Collalto et al. 1988) are somewhat difficult to adapt to mechanical pruning in that they easily tend to overcrop, with a consequent loss of fruit quality (Andersen et al. 1996, Fisher et al. 1997, Pool et al. 1993, Possingham 1994, Zabadal et al. 2002). One such variety seems to be Sangiovese, which, given its high fruitfulness of basal and base buds, often shows insufficient crop self-regulation at high bud loads and requires some form of crop adjustment (Di Collalto et al. 1988). In Italy, the two most successful long-term applications of winter mechanical pruning were with cvs. Montuni (Intrieri et al. 1998) and Croatina (Poni et al. 2004), both sharing the low fruitfulness of the basal buds, an important inherent yield-regulating factor.
Currently, we have no reported data on long-term adaptation trials carried out in Italy on highly fruitful basal node cultivars under optimal machine-training system configurations. The present contribution deals with a five-year data set on vine performance of cv. Barbera subjected to hand and mechanical pruning, the latter applied with a variable severity of hand finishing.
Materials and Methods
Plant material and experimental layout
The trial was carried out over five years from 2005–2009 in a vineyard of Vitis vinifera L. cv. Barbera (clone AT84 grafted onto 420A rootstock) established in 2002 in Piacenza (45.1°N, 9.6 °E), Italy, at a spacing of 0.9 m x 2.5 m (intra- and interrow) for a density of 4,444 plants/ha. Soil was classified as a mesic, fine clay, deep Hapludalf aquico with high water-holding capacity and low lime content. Barbera currently ranks fifth among Vitis vinifera cultivars grown in Italy, with ~28,000 ha concentrated primarily in the Piedmont and Emilia Romagna regions (2009; http://agri.istat.it). The high fruitfulness of its basal buds (Calò et al. 2006) makes it typically suited to short spur pruning. The vines were trained to a high cordon trellis set 1.7 m aboveground with no catch wires (sprawl vegetation), and canopy management was limited to light mechanical shoot trimming performed shortly after bloom to promote upright growth and optimal machine operating conditions for the follow-up winter pruning. Shoots were tipped to retain approximately 9 to 10 leaves on the main shoots.
Diurnal mean air temperature and rainfall for each season of the five years are reported for the 1 Apr to 30 Sept time span (Figure 1). Three pruning treatments—hand (HP), short mechanical followed by severe manual follow-up (SMP-SF), and short mechanical followed by light manual follow-up (SMP-LF)—were compared in a randomized complete block (RCB) design, where each block held three adjacent rows (~110 m length) to which treatments were randomly assigned. The trial thus comprised 12 rows and, within each central row of each block, five single-vine subreplicates were tagged and used for detailed growth, yield, and grape composition assessment. Hand pruning was applied to retain an average of six, 2-count node spurs per vine selected dorsally to the cordon. All wood with ventral insertion was removed.
The mechanical pruning was performed by a cutter bar unit side-mounted on a tractor featuring an over-row reverse-U cut profile applied as closely as possible to the cordon; hand clean-up was performed by two field workers with pneumatic shears working from the tractor-drawn platform; workers were instructed to primarily remove the ventrally inserted, long hangers that were missed by the pruning unit. “Severe” and “light” follow-up were defined as variable machine speed (second and fourth gear, reduced pace, respectively), thereby allowing more or less time for shortening and/or thinning of machine-pruned wood. Engine rounds per minute (rpm) were driver adjusted between 800 and 1500 rpm to negotiate situations of different wood loads on the vines. The time needed to prune either manually or mechanically was recorded each year for each test row.
Daily mean air temperature (°C) and rainfall (mm) recorded from 1 Apr to 30 Sept 2005–2009 at the experimental site (Piacenza, Italy). Arrow indicates harvest date.
Vegetative growth, yield, grape quality, and pruner performance
Vegetative growth components were recorded yearly in early May when vines had generally reached the stage of separate clusters (stage G; Baggiolini 1952). Number of shoots and their inflorescences were recorded per node of each retained spur; shoots bursting from base buds or latent buds and their inflorescences were also counted.
At the completion of canopy development (typically mid-August), 80 main and lateral leaves sampled from basal and apical positions along the shoot were taken per treatment and the area of each blade was determined by a LI-COR 3000 area meter (LI-COR Biosciences, Lincoln, NE). At leaf fall, the total number of nodes per primary shoot were counted, as were the number of nodes of each lateral shoot with respect to their position on the main shoot axis. Final total leaf area per vine was then estimated on the basis of node counts and leaf blade areas.
Yield and total cluster number per vine were recorded at harvest. Two 100-berry samples per vine were taken concurrently to ensure that the various positions within cluster (top, mid, bottom) and exposures (internal or external berries) were represented. While the first set of berry samples was weighed and stored at −20°C for subsequent color and phenolic analysis, the second set was immediately processed. The concentration of total soluble solids (Brix) was determined by a temperature-compensating Atago refractometer (model RX-5000; Atago Ltd., Tokyo, Japan). Titratable acidity (TA) was measured by a Crison Compact Titrator (Crison, Barcelona, Spain) with NaOH 0.1 N to the end point of pH 8.2 and expressed as g/L of tartaric acid equivalent.
Total anthocyanins and phenolics were determined after Iland (1988). The frozen 100-berry sample was left to thaw and then homogenized at high speed (20,000 rpm) for 1 min with an Ultra-Turrax homogenizer (Rose Scientific Ltd., Alberta, Canada). Two grams of the homogenate were transferred to a pre-tared centrifuge tube, enriched with 10 mL aqueous ethanol (50%, pH 5.0), capped and mixed periodically for one hour before centrifugation at 3500 rpm for 5 min. A portion of the extract (0.5 mL) was added to 10 mL 1 M HCL, mixed and let stand for 3 hr; the absorbance values were then registered at 520 nm and 280 nm as taken on a Kontron spectrophotometer (Tri-M Systems and Engineering, Toronto, Canada). Total anthocyanins and phenolics were expressed as mg per g of fresh weight.
Statistical treatment
A combined analysis of variance over years (Gomez and Gomez 1984) was performed using the GLM procedure of the SAS statistical package (SAS Institute, Cary, NC). Year was considered as a random variable and the error term for the pruning treatments was the year x pruning interaction mean square. Mean separation between pruning levels was performed with the Student-Newman-Keuls test. The year x treatment interaction was tested over the pooled error and discussed only in case of significance. Regression analysis was used to study relationships between continuous variables, whereas variation around means was given as standard error (SE).
Results
Growing degree days (GDD) from 1 Apr to 30 Sept at the trial site varied from 1886 GDD in 2009 to 2087 in 2007, which at 292 mm (cumulated April to September) was the year with the lowest rainfall (Figure 1). The 2008 growing season was marked by a heavy rainfall in June and very scant precipitation thereafter.
Five-year means of count nodes per vine varied from 13.4 in HP to 33.4 in SMP-LF, with SMP-SF at intermediate levels, yet more than doubling the value recorded for HP (Table 1). Treatment differences for total shoots per vine were considerably reduced as mechanically pruned vines, in addition to almost identical values, had 21% higher shoot number per vine than HP. Such compensation originated from notable differences in budbreak rates as, on average, 3.5 shoots were produced by each count node retained in HP versus 1.99 and 1.64 shoots per count node burst in SMP-SF and SMP-LF, respectively (Table 1). Regressing shoots per count node versus count nodes per vine by pooling data over treatments and seasons yielded an exponential relationship (r2 = 0.77), showing good separation between pruning regimes with lack of response beyond the 25 count node per vine threshold (Figure 2). Vine capacity given as total leaf area was slightly, albeit significantly, higher in mechanically pruned vines, which on the other hand displayed reduced vigor evaluated as leaf area per shoot and lateral leaf area development (Table 1). This latter was generally fairly weak, accounting for a maximum of 10% of the total leaf area scored in the HP treatment.
Partitioning of the significant year x pruning interaction shown by parameters of count nodes and shoots per vine and shoots per count node is reported (Figure 3). While count nodes retained in HP showed moderate variation over seasons, a steady increase up to 45 nodes/vine in 2009 was recorded in SMP-LF. SMP-SF leveled off around the 35 nodes/vine it reached in 2007 (Figure 3A). In both mechanical pruning treatments, shoots per vine increased over the seasons, whereas HP rose from ~30 shoots/vine in 2005 to 44 in 2006 and then remained essentially constant (Figure 3B). Budbreak given as shoots per count node was generally much higher in HP, although differences among treatments were less notable in the initial year of the experiment (2005) and in 2008 (Figure 3C).
Vegetative growth parameters of Barbera vines in response to three winter pruning systems: hand pruning (HP), short mechanical pruning plus light finishing (SMP-LF), and short mechanical pruning plus severe finishing (SMP-SF). Data are means of five years (2005–2009).
Expressing frequency (%), budbreak rate (shoots/node), and bud fruitfulness (clusters/shoot) versus node position on fruiting spurs in the different treatments shows that 70% of node positions in HP could be classified as base buds (position 0), while node positions 1 and 2 showed higher frequencies in the SMP treatments (Figure 4A). Sprouting from base buds was highest in HP (0.95 shoots/node) and decreased linearly with higher count nodes retained on the vine (Figure 4B, Table 1). Conversely, budbreak rates from positions 1 and 2 were similar among treatments. As expected, node fruitfulness considerably increased in each treatment, moving acropetally along the spur; HP showed slightly improved bud fruitfulness at positions 0 and 1 than the other treatments (Figure 4C).
Yield per vine (five-year basis) was very similar among treatments (~5 kg), corresponding to ~22 t/ha (Table 2). While no effects on node fertility (clusters per shoot), berries per cluster, and berry weight were seen, higher cluster number per vine in the mechanical pruning treatments was associated with a significantly lower cluster weight than data recorded in HP. In addition to the variation in absolute values among seasons, the nature of the significant year x treatment interaction for cluster weight showed that while scant within-season variation was found between treatments in 2005 and 2006, the remaining seasons consistently displayed smaller clusters in the mechanized plots (Figure 5). The ratio between final total leaf area per vine and yield in all treatments approached the threshold of one square meter per kilogram of fruit (Table 2).
Correlation between count nodes per vine and shoots per count node. Data represent single vine means and were pooled over years and pruning treatments: hand pruning (HP), short mechanical pruning plus light finishing (SMP-LF), and short mechanical pruning plus severe finishing (SMP-SF). Regression equation: y = 19.283 x−0.711, r2 = 0.77.
Grape composition differences between HP and either hedged treatment were minor, with between-treatment mean variation of 18.8–19.5 for soluble solids (Brix), 3.04–3.08 for pH, and 10.1–10.5 for total acidity. Conversely, a slight, yet statistically significant, lower anthocyanin concentration was found in mechanically pruned vines (0.75 mg/g versus 0.85 mg/g in HP). No significant pruning x year interactions were found for any of the quality parameters measured. Over the five-year run, time savings for winter pruning was 54% for SMP-SF as compared with HP and increased to 70% when a light manual follow-up was applied (Table 3).
Variation over years of count nodes per vine (A), shoots per vine (B), and shoots per count node (C) for three treatments (see Figure 2 for treatments). Vertical bars show SE for each year x treatment combination, n = 4.
Discussion
The hedged vines showed a more than two-fold increase in nodes per vine as compared with hand pruning without affecting yield per vine, thus indicating that full yield compensation occurred, which is in agreement with previous long-term mechanical pruning trials (Di Collalto et al. 1988, Intrieri et al. 1988). Barbera vines clearly tended to reach equilibrium at an environmentally driven yield threshold of ~5 kilos, which was wholly independent of the bud load retained at winter pruning and which proved to be an unreliable controller of the final cropping level. In this respect our findings do not match those reported in a four-year mechanical pruning trial on Sangiovese, a cultivar that shares with Barbera a high fruitfulness of the basal nodes but where yield significantly increased in machine-pruned vines (30% higher than HP control) and must soluble solids concentration consequently decreased (Di Collalto et al. 1988).
Frequency distribution (%) (A), shoots per node (B), and inflorescences per shoot (C) of base buds and count nodes along spurs in each pruning treatment. Numbers atop bars indicate absolute values (see Figure 2 for treatments). Data pooled over years. Vertical bars show SE for each pruning treatment, n = 16.
Interestingly, our data evinced no “peak yield” in the mechanical treatments at the end of the first trial year, as crop per vine (mean ± SE) was 2.48 ± 0.24 in HP, 3.10 ± 0.28 in SMP-SF, and 3.63 ± 0.35 in SMP-LF (data not shown). A steep yield rise is usually expected in year 1, since an abrupt increase in the node number per vine is linked to the typical nonselective machine action and is unlikely to be offset by other counteracting yield factors, which usually require several years to become established. In a seven-year trial with cv. Montuni, a similar yield between hand pruning and mechanical pruning was reported, although the year 1 yield per vine with mechanical pruning peaked at 13.0 kg versus 8.0 kg for hand pruning, with a remarkable decrease in must soluble solids concentration (Intrieri et al. 1988). Generally, such a response might have two major drawbacks depending on site, genotype, and grower management skills. The first is that an “on” year might trigger alternate bearing while having a negative impact on grower expectations from the mechanical approach. In response to an overly high year 1 yield, which is unavoidably accompanied by poorer grape composition, a grower may be tempted to resume hand pruning the next season or apply overly severe hand finishing that would result in vines similar to those that are hand pruned. In the present study, the response was mild because the vines in the first season of mechanical pruning (2005, fourth after planting) were in a transition phase between long canes and spur selection. Therefore, the maximum number of retainable spurs at pruning was entirely a function of internode length and, eventually, canes per node produced the previous growing season. Count nodes per vine were lowest for the SMP treatments the first season and steadily increased over the years due to a progressive increase in shoots (canes) per vine (Figure 3A). Overall, the decision to begin mechanical pruning in year 4 after planting may well be viable for smooth vine adaptation to the pruning regime.
Yield compensation is notably dependent on the degree of adjustment to increased node number as shown by budbreak (shoots per node), bud fruitfulness (inflorescences per shoot), and cluster weight (Reynolds 2010, Tassie and Freeman 1992). Our results show that the regulation of budbreak, that is, fewer shoots per count node in the mechanically pruned vines, was the primary yield determinant. Yet, results suggest that, despite the fairly low count node level left on the HP vines (13.4) (Table 1, Figure 4), yield here was stimulated by budbreak from noncount nodes, which in HP represented 70% of the total bud population (Figure 4A). This response primarily relates to the inherent lower node number retained at pruning. However, the absolute value for base buds (position 0) per vine in HP was 30.4 on a five-year basis (Figure 4A), which was ~4- to 5-fold higher than the count expected from spur number per vine (6 to 7 on average, hence an equivalent amount of “noncount” nodes). Such a value is reasonable considering that hand pruning was rigorously applied as clear-cut removal of the canes on the ventral portion of the cordon, whereas removal of side or dorsally located canes was performed with maximum care to preserve the noncount buds to facilitate year-after-year spur renewal along the cordon and vegetative viability of the cordon itself.
Results show no consistent reduction in bud fruitfulness of hedged vines, which has been reported as one of the most consistent yield compensation factors (Clingeleffer 1993, Reynolds 2010). However, there are at least three explanations for this finding. The degree of reduction in bud fruitfulness in hedged vines is often a primary consequence of the actual yield increase caused by retaining more buds, which was nil in our trial (Table 2). Changes in bud fruitfulness in hedged versus hand-pruned vines also depend on the variation of frequency distribution of bearer length caused by pruning strategies. Data reported in a previous study indicated that all hedged treatments caused a strong reduction in pruning length (i.e., fewer nodes on retained spurs) as compared with hand pruning by retaining higher numbers of less fruitful buds (Intrieri et al. 1988). This effect is ambiguous since the measured reduction in bud fruitfulness may simply be the outcome of the highest frequencies of less fruitful nodes, without bud fruitfulness itself directly affected. That does not seem the case in the present study, however, since the frequency distribution of the type of bearing units did not substantially differ among treatments, although HP vines showed a higher fraction of base buds than the other pruning treatments (Figure 4A). It is noteworthy that average spur length did not increase in our machine pruning treatments as compared with hand pruning, given that field workers were instructed to retain spurs not exceeding 2-count nodes in length on them. The incidence of 3- and 4-count node spurs was negligible. In addition to driver ability, this outcome results from an efficient integration of trellis design, canopy configuration, and machine settings, as successfully performing box-hedge pruning as close as possible to the cordon depends upon the absence of any obstacle in the upper 180° above canopy and on a prevailing upright shoot growth. On the other hand, the extent in severity of hand finishing was chosen to leave enough time for the removal of any ventrally inserted canes that the cutter bars may have missed.
Decreased bud fruitfulness is generally expected when canopy or shoot physiology imposes some limitation on the induction and differentiation processes. A good case in point is with minimal pruning (Possingham 1994), where weak shoots produced in the spring exert a source limitation on bud differentiation that provides some preliminary crop control for the next season. A negative effect on bud fruitfulness can also occur when excessive shading is cast on the basal part of the shoots (May et al. 1976, Sanchez and Dokoozlian 2005). Although the dynamics of vine-leaf area formation were not monitored in the present study, it is unlikely that both effects occurred. Shoot development appeared similar among treatments until shoot trimming was performed (usually first week of June), a practice that helped to maintain a more erect canopy with better light penetration even in the inner parts. Then, too, lateral development subsequent to trimming was quite moderate in all treatments and, in addition to concentration in the most apical nodes, it is highly unlikely that significant shade was cast on the subtending dormant buds.
Variation over years of cluster weight for the different pruning treatments (see Figure 2 for treatments). Vertical bars show SE for each year x treatment combination, n = 4.
Yield components and leaf-to-fruit ratio of Barbera vines in response to three winter pruning systems: hand pruning (HP), short mechanical pruning plus light finishing (SMP-LF), and short mechanical pruning plus severe finishing (SMP-SF). Data are means of five years (2005–2009).
Field worker hours/ha (hr, min) needed to complete manual and mechanical pruning, yearly and averaged data, 2005–2009: hand pruning (HP), short mechanical pruning plus light finishing (SMP-LF), and short mechanical pruning plus severe finishing (SMP-SF).
Yield compensation data indicates that cluster weight had significant reduction in hedged vines compared with HP beginning from 2007 (third trial year) and thereafter. Thus, cluster weight may prove a sensitive factor at the pruning level once it has overcome the threshold of about 30 count nodes per vine (Figure 3A). Conversely, no long-term effects of pruning regime on berry size were seen. Other researchers reporting a ratio of retained nodes per vine for hedged versus manual pruning in the range of 2:1 or 3:1 have shown that the yield component berry size is not very sensitive to increased node number per vine (Possingham 1994, Martinez de Toda and Sancha 1999).
Results here allow for a discussion of the efficiency of the compared treatments with respect to yield level, grape composition, and labor. No change in yield, minor modifications in grape composition, and a time savings in winter pruning from 54 to 70%, as compared with HP, underscore that mechanical pruning followed by hand finishing is a rewarding practice. This is particularly true for a highly mechanized, single high-wire trellis where moving from hand pruning to SMP plus hand finishing would currently result in an estimated cost for winter pruning of approximately $820/ha (€570/ha), down from $2170/ha (€1500/ha) for HP (Corradi 2010). The light manual follow-up (faster machine speed) enabled only an 8-hr time savings per hectare as compared to severe follow-up (two runs per row). Yet, this trend is not so surprising considering that the removal of wood having a ventral insertion on the cordon (not negotiable by the cutter bars) accounted for the majority of the manual follow-up and was indicated as an absolute priority to the field workers.
Overall, a slight change in grape composition among treatments corresponds with very similar leaf-to-fruit ratio. Yet, it is unclear why, with yield and vine balance being almost equal, anthocyanin concentration was slightly, albeit significantly, lower in SMP treatments (Table 3). Although no microclimate fruiting area assessment was performed in this study, there is a general consensus that heavily shaded clusters have poorer berry pigmentation than more exposed clusters (Downey et al. 2006). In the single high-wire trellis with a decidedly upright shoot growth, as in this study, the prevalent cluster regime was inherently dim light broken by occasional sunflecks penetrating the canopy from different sun angles. Canopy density given as shoot number per vine was highest in the SMP (Table 2), which may have lessened light availability at the cluster level, thereby slightly reducing berry color accumulation.
Conclusions
Our data supports the hypothesis that a basal-node fruiting cultivar like Barbera responds successfully to mechanical pruning followed by hand finishing. Full yield compensation was achieved despite a more than doubling in the bud load retained at pruning, with only slight changes in grape composition and a time-savings ranging from 54 to 70% compared with hand pruning. A fully mechanized single high-wire Barbera trellis can thus be realistically managed in less than 70 worker hours per hectare. Self-regulation at increased node number occurred mostly in terms of reduced budbreak, and vines showed no further budbreak rate response beyond the threshold of about 25 count nodes per vine. Regardless of pruning regime, the nearly 22 t/ha yield level was quite high, suggesting that improved grape composition would require using additional crop-adjustment techniques.
Acknowledgments
Acknowledgments: This investigation was partially supported by a grant from CRPV, Filiera Viti-Vinicola, Regione Emilia-Romagna, Italy.
Footnotes
-
The authors thank David Verzoni for editing and restyling the English text and Giovanni Fugazza for lending the vineyard plots.
- Received September 1, 2010.
- Revision received December 1, 2010.
- Accepted January 1, 2011.
- Published online December 1969
- © 2011 by the American Society for Enology and Viticulture
Literature Cited
Vol 62 Issue 2
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
