Modification of Cv. Merlot Berry Composition and Wine Sensory Characteristics by Different Leaf Area to Fruit Ratios

Abstract

This study aimed to evaluate the effects of cluster thinning and severe shoot trimming on berry and wine composition and wine sensory characteristics of Merlot variety, in the context of climate change challenges related to grapevine ripening and the corresponding high alcohol content in wine. In two seasons, two different crop sizes were obtained via cluster thinning and combined in a two-factorial design with severe shoot trimming (SST) and its respective high canopy control (HC). In both seasons, cluster thinning (CT) resulted in higher Brix in grape juice and higher alcohol in wine than full crop size (FC), whereas SST obtained lower values than HC. Total anthocyanins and phenolics in wine were increased by CT, whereas SST had no any significant effect on wine’s phenolic content. Several sensory characteristics of wine were positively affected by CT in both seasons, including aroma intensity, wine body and overall wine quality, whereas SST wines were in one season characterized by increased perception of vegetal aroma, acidity and bitterness, and decreased perception of body, persistency and taste balance. Our results demonstrate that practices which affect the leaf area to fruit ratio have a major impact on wine sensorial characteristics, concluding that their choice should be based on the desired wine style.

1. Introduction

Different crop and canopy management practices that alter crop size, source to sink ratio or fruit zone microclimate are used in viticulture production in order to achieve a targeted yield and composition of berries at harvest [1,2,3,4,5]. Some practices are used to regulate crop size, such as bud load management [6], shoot thinning [7,8] or cluster thinning [9,10]. They aim to achieve a lower yield and a greater leaf area to fruit ratio and, consequently, promote sugar accumulation in berries by increasing the assimilate availability per unit of grape mass [5,11].

For the production of premium quality, full-bodied red wines, cluster thinning at the beginning of veraison is a practice often used by winegrowers, especially if the yield in the current season is expected to be overly high for a targeted wine style [12]. The vine usually reacts to this practice with enhanced ripening dynamics, resulting in increased sugar concentration, pH, lower titratable acidity and, in most cases, increased concentration of anthocyanins and other phenolic compounds in grapes and their respective wines [13,14].

However, the practices of yield reduction are nowadays questioned due to the effects of climate changes; the earlier and more enhanced ripening of grapes, with higher sugar accumulation in berries, is a trend largely observed in recent decades in most wine growing regions worldwide [15,16,17,18,19]. This situation, in many cases, leads to excessively high alcohol concentration in wine, negatively affecting its sensory quality and leading to wines which are out of balance in term of the alcoholic notes or hotness sensation [20,21]. Furthermore, the desired sugar concentration for the production of red wines is often reached without a proper phenolic maturity in terms of the accumulation of secondary metabolites in berries [15,22,23], forcing the producers to harvest the grapes at supra-optimal sugar concentrations owing to the contribution of phenolic compounds to the formation of specific sensory characteristics of wine, such as color, aroma, mouth-feel and flavor [24].

In order to mitigate these rising challenges to viticulture production, several measures to delay ripening dynamics have been widely studied in recent years, such as the use of late ripening rootstocks, varieties or clones, delayed pruning, reduced leaf area to fruit ratio, late source limitation, anti-transpirants and shading nets, late irrigation or other methods [16,18,23,25,26,27]. Among these options, late source limitation is focused on the removal of the leaves in the apical part of the canopy around veraison as, at this stage, the apical leaves are the most functional, having reached full expansion while still short of senescence [28]. Late source limitation may be carried out as leaf removal in the apical part of the canopy [29,30] or severe shoot trimming [31,32,33]. Several studies obtained promising results with this technique as the sugar content was reduced without negatively affecting the concentration of anthocyanins in the berries [28,34,35,36].

However, despite the widely investigated impact of late source limitation on berry composition, only few studies have considered the effect of these techniques on wine sensory characteristics and its overall quality, leading to a knowledge gap in this regard. Caccavello et al. [22] showed that moderate post-veraison defoliation and trimming decreased soluble solids in berry and wine alcohol concentration and improved wine sensory score in Aglianico variety, whereas intense defoliation and trimming negatively affected this parameter. In other studies with severe shoot trimming a mild modification of wine sensory characteristics was obtained with this practice as the reduction in red fruit aroma attributes and color intensity [37] or the increase in the total wine score [33]. Quantitative descriptive sensory analysis of wines is widely used in studies if a detailed description of sensory attributes is sought or if sensory properties of different products have to be compared [38], such as the effects of specific viticultural or enological treatments on wine quality. Due to its greater sensitivity to differences between the samples, quantitative descriptive sensory analysis has a bigger discriminatory power for wine samples in comparison to the other methods, such as the 100-points OIV method, which is able to show only the general quality of wines [39].

Since the studies investigating the impact of late source limitation on wine sensory characteristics are limited, the aim of this study is to determine the effects of severe shoot trimming, performed at late veraison, on leaf area, yield components, berry and wine composition, and wine sensory characteristics of Merlot variety, which in favorable growing conditions has a tendency to accumulate high sugar content in a berry [40], resulting in overly high alcohol levels in wine [41]. Severe shoot trimming and its respective control were evaluated at two crop sizes obtained via cluster thinning at the onset of veraison, leading to four different combinations of treatments. The purpose of combining these two practices was to assess the effectiveness of late source limitation in reducing the sugar accumulation in berries at different crop sizes, evaluate the relative impact of these techniques on berry and wine composition, and study the potential interaction between these two treatments.

2. Materials and Methods

2.1. Vineyard Site

The experiment was conducted during seasons 2017 and 2018 in the vineyard of the Institute of Agriculture and Tourism in Poreč (latitude 45°13′20′′ N; longitude 13°36′00′′ E; 15 m above sea level), Istria winegrowing region, Croatia. Vitis vinifera L. Merlot vines (clone 347) were grafted on SO4 rootstock. The vines were planted in 2006 in terra rossa (chromic luvisol) soil, with a 2.5 m × 0.8 m spacing (row × vine), on a slope exposed to west, and with 5% inclination. The orientation of the vineyard rows was NNE-SSW, with a declination of 27° from N-S. Irrigation was not used. The training system was a bilateral spur cordon, vertically shoot-positioned, pruned to eight to nine spurs per vine, and with each spur containing two nodes. The removal of approximately two leaves per shoot around the clusters was carried out at berry set in order to avoid the formation of dense canopies. Shoot trimming was mechanically performed twice per year at a canopy height of 130 cm: the first time at berry set (in the first decade of June) and the second time three weeks thereafter.

The monthly mean temperature and average monthly rainfall values were obtained by the Croatian Meteorological and Hydrological Service from a weather station located 200 m from the vineyard. Grapevine phenological stages were recorded according to the modified E-L system [42].

2.2. Experimental Design

The treatments were applied as two crop sizes in combination with two different canopy heights. The two crop sizes were regulated via cluster thinning, which was manually performed at the beginning of the veraison (on 28 July 2017 and 23 July 2018) at grapevine growth stage 35 according to the modified E-L system [42], when approximately 3–5% of the berries started to change color and the concentration of soluble solids in the grape juice was approximately 9 Brix. Cluster thinning treatment (CT) was obtained via thinning 35% of the clusters, whereas in full crop treatment (FC) the clusters were not removed. Two canopy heights were obtained via trimming the vines to 130 cm of total canopy height (high canopy; HC) or removing the apical part of the shoots via severe shoot trimming (SST) to 70 cm of the total canopy height at late veraison (on 14 August 2017 and 10 August 2018) and grapevine growth stage 36, when approximately 80–90% of berries changed color and the concentration of soluble solids in grape juice was approximately 14–15 Brix. Treatments were set up as a 2 × 2 factorial design, leading to four combinations: FC-HC (full crop, high canopy), FC-SST (full crop, severe shoot trimming), CT-HC (cluster thinning, high canopy), and CT-SST (cluster thinning, severe shoot trimming). The experiment was set as a randomized complete block design in three adjacent rows, with each row as a block. In each row, four sections of two post spaces (containing 12 vines per plot) were randomly assigned to the four combinations of treatments. Two post spaces at the beginning and end of each row were not included in the experiment. In both seasons, treatments were applied on the same vines. The grapes were harvested manually on 20 September 2017 and 24 September 2018.

2.3. Leaf Area, Cluster Exposure, Yield Components and Berry Sampling

Leaf area, cluster exposure, yield components and berry sampling were assessed as previously described [32]. In summary, eight representative shoots per replicate were collected one week before harvest and leaf area was determined with a LI-3000 leaf area meter (LI-COR Bioscience, Lincoln, NE, USA). Photosynthetically active radiation (PAR) in the cluster zone was determined one day after the application of severe shoot trimming between 11:00 and 12:00 h under cloudless conditions using a portable QSO-S PAR Photon Flux sensor (Decagon Devices, Pullman, WA, USA), which was placed vertically and upward near clusters on both sides of the canopy. The yield and number of clusters per vine were recorded at harvest and the average cluster weight was calculated. The number of clusters per shoot was determined as clusters to shoot ratio. Single berry weight was calculated using a sample of 200 berries. In both years, berry sampling was performed at four stages of ripening in order to follow the accumulation of sugar content in berries. In both years, the first sampling also was performed on a day when severe shoot trimming was applied, while the fourth sampling was performed on a harvest day. The second and third samplings were performed in intervals of 11 to 13 days in 2017 and at intervals of 14 to 17 days in 2018. At harvest date, samples of 200 berries were used for total anthocyanins and total phenolic determinations. All sets of berries were weighed and immediately stored at −20 °C.

2.4. Vinification

Microvinifications were performed separately for each experimental plot. At the date of harvest, 20 kg of grapes per treatment replication were de-stemmed, crushed, treated with 50 mg/L SO₂, supplied with the yeast nutrient Go-Ferm Protect (Lallemand, Montreal, Canada; 0.3 g/L) and inoculated with Saccharomyces cerevisiae (Enoferm BDX; Lallemand, Montreal, Canada; 0.3 g/L). Fermentations were carried out in 10 L glass vessels at a temperature of 25 °C. Three days after the beginning of fermentation, 0.2 g/L of yeast nutrient Fermaid E (Lallemand, Montreal, Canada) was added. The pomace was mixed two times per day. After 10 days of fermentation and maceration, the pomace was pressed in a small pneumatic press to 1.5 Bar. The wines were racked and sulfited with 50 mg/L SO₂ and stored in 5 L glass vessels at 15 °C. During ageing of the wines, racking and sulfiting to 30 mg/L of free SO₂ was performed every three months. Wine samples were taken for the chemical analyses four months after the end of fermentation, whereas quantitative descriptive sensory analysis of the wines was performed six months later.

2.5. Chemical Analyses of Grape Juice, Berries and Wine

Soluble solids (Brix) in the grape juice were determined using a HR200 digital refractometer (APT Instruments, Litchfield, IL, USA), pH was determined using a MP220 pH-meter (Mettler Toledo, Giessen, Germany), and titratable acidity (expressed as g/L tartaric acid equivalents) was measured via titration with 0.1 N NaOH (BDH Prolabo, VWR Chemicals, Leicester, UK). Standard physico-chemical wine parameters were determined according to the methods of the International Organization of Vine and Wine [43]. Sugar content per berry was approximated from berry volume (obtained by immersing 200 berries in a graduated cylinder) and Brix, as reported by Previtali et al. [44]. Total anthocyanins and phenolic substances in berries were determined via spectrophotometry using a 0.1 M HCl (Carlo Erba Reagents, Milan, Italy) and 96% EtOH p.a. (Lab Expert, KEFO d.o.o, Sisak, Croatia) according to Iland et al. [45] and expressed either as mg/g of berry fresh weight and mg/berry. Wine color intensity (optical density 420, 520 and 620 nm) and the concentrations of total anthocyanins and phenolics in wine samples were determined using UV/VIS spectrophotometer (Cary 50, Varian, Palo Alto, CA, USA). Total phenolics in wine were determined via spectrophotometry using a Folin–Ciocalteu reagent of analytical grade (BDH Prolabo, VWR, Leicester, UK) and quantified according to Singleton and Rossi [46], whereas total anthocyanins were analyzed via spectrophotometry using concentrated HCl (Carlo Erba Reagents, Milan, Italy), 50% solution of sodium hydrogensulfite (BDH Prolabo, VWR, Leicester, UK) and 96% EtOH p.a. (Lab Expert, KEFO d.o.o, Sisak, Croatia), as reported by Ribérau-Gayon and Stonestreet [47]. All analyses were carried out in duplicate.

2.6. Quantitative Descriptive Sensory Analysis of Wines

Quantitative descriptive sensory analysis of wines was performed by a panel of eight trained certified panelists, who were highly experienced in cv. Merlot wine sensory analysis. Tasters were seated in separate booths and the environment was free of interference in terms of noise, visual stimulation and ambient odor. A constant volume of 40 mL for each wine was evaluated under white light in wine tasting glasses at 18 °C as described by the International Organization for Standardization ISO standard [48]. Qualitative (selection of main descriptors and standardization of vocabulary) and quantitative criteria (intensity of perception) of the panelists were attuned using tasting representative samples of Merlot wine through several preliminary training sessions and at the beginning of the sensory analysis. A detailed list of sensory descriptors and definitions of the attributes used in the quantitative descriptive sensory analysis of wines are indicated in Table S1. Samples were served to tasters in random order according to a randomized three-digit number for identification. The tasters used a 10-point structured scale to rate the color, aroma or taste intensity of each descriptor (0 = descriptor not perceptible; 10 = descriptor strongly perceptible). Sensory analysis performed in this study was in accordance with the standards of the Institutional Ethical Committee and the ‘Guidance Note—Ethics and Food-Related Research’ defined by the European Union.

2.7. Statistical Analysis

For the determination of the effects of the two investigated factors (cluster thinning and severe shoot trimming), data were processed using GenStat (VSN, Hemel Hempstead, UK; Version 10.2) separately by year, using two-way ANOVA in randomized blocks design. For the assessment of sugar content in berries during ripening, sensory descriptors of wine, and cases where other variables obtained significant interaction between the two factors, data were processed for the four combinations of treatments (FC-HC, FC -SST, CT-HC and CT-SST) using one-way ANOVA in randomized blocks design.

3. Results and Discussion

3.1. Meteorological Conditions

The values of average air temperature and monthly rainfall during the 2017 and 2018 growing seasons are presented in Figure 1. The 2017 season was characterized with generally lower temperatures than 2018, especially in April, May and September, as well as with higher temperatures in June. During the period covering June to August, less rainfall was recorded in 2017 compared to 2018 and the difference in this regard was especially marked in July and August (with 15 and 32 mm of rain, respectively), leading to drought conditions in 2017 as the vineyard was not irrigated. Following this event, abundant rainfall occurred in September 2017.

3.2. Leaf Area and Yield Components

As a direct consequence of the removal of the upper part of the canopy, the primary, lateral and total leaf area of the vine was reduced in both years via severe shoot trimming (SST) as compared to the high canopy (HC) treatment, whereas cluster thinning (CT) did not have any impact on vine leaf area (

Table 1. Leaf area, photosynthetically active radiation in fruit zone and yield components of Merlot vines subjected to cluster thinning (FC, full crop and CT, cluster thinning) and shoot trimming (HC, high canopy and SST, severe shoot trimming).

	Season 2017							Season 2018
	Cluster Thinning			Shoot Trimming			Int. b	Cluster Thinning			Shoot Trimming			Int.
	FC	CT	Sign. a	HC	SST	Sign.	Sign.	FC	CT	Sign.	HC	SST	Sign.	Sign.
Primary leaf area (m2/vine)	1.89	1.91	ns	2.43	1.37	***	ns	2.05	1.94	ns	2.52	1.46	***	ns
Lateral leaf area (m2/vine)	0.90	0.79	ns	1.16	0.54	***	ns	0.89	0.88	ns	1.16	0.61	***	ns
Total leaf area (m2/vine)	2.79	2.70	ns	3.58	1.91	***	ns	2.94	2.82	ns	3.68	2.08	***	ns
PAR c (% ambient)	11	11	ns	7	15	***	ns	10	9	ns	5	14	**	ns
Yield/vine (kg)	2.52	1.77	***	2.18	2.12	ns	ns	2.98	1.85	***	2.46	2.37	ns	ns
Clusters/vine	24.6	16.0	***	20.2	20.4	ns	ns	26.6	16.8	***	22.2	21.2	ns	ns
Cluster weight (g)	103	111	ns	109	105	ns	ns	112	110	ns	111	111	ns	ns
Shoots/vine	17.7	17.0	ns	17.4	17.3	ns	ns	16.8	16.4	ns	16.7	16.4	ns	ns
Clusters/shoot	1.39	0.94	***	1.15	1.18	ns	ns	1.59	1.02	***	1.32	1.29	ns	ns
Berry weight (g)	1.18	1.30	**	1.28	1.20	*	ns	1.43	1.52	*	1.46	1.49	ns	ns
Leaf area/yield (m2/kg)	1.12	1.52	**	1.72	0.93	***	ns	0.99	1.53	***	1.58	0.94	***	ns

^a Data were analyzed via two-way ANOVA in randomized blocks design (ns, not significant; *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001). ^b Interaction between cluster thinning and shoot trimming. ^c Photosynthetically active radiation.

). The photosynthetically active radiation (PAR) in the fruit zone was, in both years, higher in SST vines because more sunlight reached the clusters from the upper part of the canopy; this finding is also corroborated in some other studies with late severe shoot trimming [30,32,37]. Yield per vine, clusters per vine and clusters per shoot were affected only with CT treatment as a direct consequence of the removal of clusters, while late source limitation had no effect on these variables, as observed in other studies where severe shoot trimming was applied after the onset veraison [9,28,37]. Berry weight was higher in CT in both years and was reduced via SST in 2017, presumably because of the differences in assimilates’ availability among treatments [11] as the source/sink (leaf area/yield) ratio was significantly increased by CT and decreased via SST in both years. The absence of difference in berry weight between HC and SST in 2018 was possibly due to more favorable conditions for berry ripening in this season because of higher water availability during the summer months.

3.3. Berry Composition

As a consequence of the difference in leaf area/yield ratio between the treatments (Table 1), sugar content (mg/berry) was significantly affected by the combination of treatments in most sampling dates (Figure 2). This outcome resulted from the differences among treatments in the leaf area/yield ratio, which determine the availability of assimilate to the berries [11]. In both seasons, the highest values of sugar content (mg/berry) were obtained via CT-HC as it had the largest leaf area and the lowest crop size, thus enabling a high availability of assimilates for the berries throughout the ripening period. In 2017, FC-SST obtained the lowest values for sugar content per berry, whereas the other two treatments were characterized with intermediate values during the latter two sampling dates. On the other hand, in 2018 both full crop treatments obtained similar and low values of sugar content per berry in all sampling dates, regardless of the canopy size. This difference between 2017 and 2018 possibly indicates a more pronounced reduction in assimilate availability for the ripening berries if a large crop is combined with reduced canopy under conditions of restricted water availability [49], whereas if water is not a limiting factor satisfactory sugar accumulation is also achieved with reduced source/sink ratio [12].

The reduction in sugar accumulation per berry with severe shoot trimming was more expressed in both years for vines with lower yields, whereas in higher-yield conditions this reduction was modest or absent. These results indicate that late source limitation has the largest effect if performed on vines with high assimilate availability for developing clusters due to high leaf area/yield ratio.

The differences among treatments in sugar content per berry were reflected in total soluble solids content in the grape juice (

Table 2. Berry composition of Merlot vines subjected to cluster thinning (FC—full crop; and CT—cluster thinning) and shoot trimming (HC—high canopy; and SST—severe shoot trimming).

	Season 2017							Season 2018
	Cluster Thinning			Shoot Trimming			Int. b	Cluster Thinning			Shoot Trimming			Int.
	FC	CT	Sign. a	HC	SST	Sign.	Sign.	FC	CT	Sign.	HC	SST	Sign.	Sign.
Soluble solids (Brix)	21.7	22.8	***	22.9	21.6	***	*	22.3	23.5	**	23.5	22.3	**	*
Titratable acidity (g/L)	7.0	6.4	**	6.6	6.8	ns	ns	6.8	6.4	*	6.7	6.5	ns	ns
pH	3.40	3.49	**	3.46	3.42	ns	ns	3.42	3.56	**	3.49	3.49	ns	ns
Total anthocyanins (mg/g)	0.75	0.83	ns	0.80	0.79	ns	*	0.73	0.79	ns	0.78	0.75	ns	ns
Total anthocyanins (mg/berry)	0.91	1.07	*	1.03	0.95	ns	**	1.04	1.20	ns	1.14	1.11	ns	ns
Total phenolics (mg/g)	1.72	1.76	ns	1.72	1.76	ns	ns	1.58	1.61	ns	1.60	1.58	ns	ns
Total phenolics (mg/berry)	2.09	2.23	ns	2.22	2.11	ns	**	2.25	2.43	ns	2.33	2.36	ns	ns

^a Data were analyzed via two-way ANOVA in randomized blocks design (ns, not significant; *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001). ^b Interaction between cluster thinning and shoot trimming.

). In both years, CT obtained higher Brix than FC, whereas SST obtained lower values than HC. The increase in soluble solids via cluster thinning is a common reaction in over-cropped vines [12,13,14], as is the decrease in soluble solids via late source limitation due to a lower assimilate availability for the berries during ripening [31,34,35,49]. Besides the differences in Brix, SST did not significantly affect other berry composition variables compared to HC, confirming the previously reported findings from several authors that, if the severe shoot trimming or apical leaf removal is performed at late veraison, it does not negatively affect the phenolic composition of berries [28,31,32,35,37,50]. On the other hand, due to more advanced ripening CT decreased titratable acidity and increased pH of grape juice in both years; this effect is often obtained with this practice [1,12,13,14].

Concerning the berry phenolic composition, cluster thinning only significantly increased the total anthocyanins expressed as mg per berry in 2017, though the concentration of total anthocyanins (expressed as mg/g) had a tendency to increase with reduced crop size in both years. This effect presumably stems from the increased sugar concentration in berries of CT treatment as sugars, together with abscisic acid, are thought to trigger the biosynthesis of anthocyanins in grape berries [51,52]. In other studies, the effect of cluster thinning on the concentration of total anthocyanins in berries differs according to the growing conditions and ranges from no impact to a major positive impact. The final outcome is mostly affected by the assimilate availability for the developing berries before the execution of cluster thinning; vines which experience greater limitation in assimilate availability generally obtain more benefits from cluster thinning in terms of improved berry ripening, greater accumulation of assimilates in the berries and higher berry and wine quality [11,12,53,54].

For some berry composition variables, the interaction between the two investigated factors was obtained (Table 2) and, in these cases, the results of all specific combinations of treatments were presented in order to elucidate the effects of the particular combinations (Figure 3). Concerning the soluble solids in grape juice, the interaction was a consequence of the particularly great positive effect of the combination of cluster thinning and high canopy (CT-HC), which was obtained in both years of study. On the other hand, for total anthocyanins expressed as mg/g and mg/berry in 2017 and total phenolics expressed as mg/berry in 2017, the effect of severe shoot trimming differed depending on the crop size. In this year, severe shoot trimming had a negative effect on low crop size vines; we speculate that the reduction in assimilate production in this case also had a negative impact on the secondary metabolism in the berry—a reaction observed in situations with low assimilate availability [55]. On the other hand, at full crop size the effect of severe shoot trimming on total anthocyanins and total phenolics was positive, possibly because, in these conditions, the vines profited from better light infiltration in the fruit zone during the ripening period as a result of the removal of the upper part of canopy [32,36,37].

3.4. Wine Composition

As expected, the alcohol content in wine corresponded to the results of soluble solids in grape juice. In both years of study, the practice of cluster thinning increased the alcohol content in wine, whereas a reduction was obtained via severe shoot trimming (

Table 3. Wine composition obtained from grapes of Merlot vines subjected to cluster thinning (FC—full crop; and CT—cluster thinning) and shoot trimming (HC—high canopy; and SST—severe shoot trimming).

	Season 2017							Season 2018
	Cluster Thinning			Shoot Trimming			Int. b	Cluster Thinning			Shoot Trimming			Int.
	FC	CT	Sign. a	HC	SST	Sign.	Sign.	FC	CT	Sign.	HC	SST	Sign.	Sign.
Alcohol (vol.%)	12.6	13.3	***	13.4	12.6	***	**	13.0	13.8	***	13.8	13.0	***	*
Titratable acidity (g/L)	6.5	6.3	ns	6.4	6.4	ns	ns	6.4	6.2	ns	6.3	6.2	ns	ns
pH	3.33	3.43	*	3.41	3.36	ns	ns	3.38	3.46	*	3.41	3.43	ns	ns
Color intensity	0.77	0.87	ns	0.84	0.80	ns	ns	0.81	0.90	ns	0.88	0.83	ns	ns
Total anthocyanins (mg/L)	248	295	**	281	262	ns	ns	353	426	***	398	381	ns	*
Total phenolics (mg/L)	1074	1171	ns	1136	1110	ns	ns	1277	1454	**	1359	1372	ns	ns

^a Data were analyzed via two-way ANOVA in randomized blocks design (ns, not significant; *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001). ^b Interaction between cluster thinning and shoot trimming.

). The same corresponding results occurred with pH, which was increased via CT in both years. On the other hand, insignificant differences were observed between CT and FC for titratable acidity, though it tended to be lower in CT. Interestingly, though the effect of cluster thinning on berry phenolic composition was in most cases insignificant, the phenolic composition of wine was largely affected by this practice; that is, CT wines had significantly higher concentrations of total anthocyanins in both years of study than FC, whereas the differences in berries were not significant. A similar outcome was reported in some other studies [56,57], where the viticultural treatments did not change the concentration of anthocyanins in the berries, followed by their significant increase in the respective wines. Given that the concentration of extractable anthocyanins increases in the final stages of berry ripening [58] and higher anthocyanin concentration in wines can be obtained from grapes which are more ripe [59], we hypothesize that the observed discrepancies in the concentration of total anthocyanins in berries and wines of CT-treated vines are related to changes in skin cell wall material, which occur during the ripening period [60]. Similarly, although no significant differences were observed in the concentration of total phenolics in berries, CT obtained significantly higher concentrations of total phenolics in wine in 2018. This variability was most likely caused by the ethanol concentration in the wine, which results from the permeability of the cell walls and membranes that are mostly affected by factors such as the presence of ethanol, which causes a more significant release of phenols located in the inner-thick cells of the skin’s hypodermis [61]. According to several studies, higher ethanol concentration leads to higher concentrations of total phenolics, tannins and non-tannin phenolics in wine [59,62], which was the case for the CT treatment compared to FC. The color intensity of the wines did not differ significantly between treatments, though a tendency toward more intense coloration was observed in CT compared to FC in both years; this trend was probably due to higher concentration of total anthocyanins in CT and the fact that higher ethanol concentration favors the formation of polymeric pigments, which result in darker wines [59]. Anthocyanins are responsible for the color of red wines and their interactions with other phenolic compounds are of crucial importance in determining wine color [63].

Apart from the reduced alcohol content in wine, severe shoot trimming had no other significant impact on wine composition, confirming the findings obtained by some previous studies, which reported that late source limitation does not negatively impact the concentration of anthocyanins or phenolic compounds in wine [28,30,33,34,35].

Following the response observed for the sugar content in grape juice, the interaction between the two investigated factors was also obtained for the alcohol content in wine (Table 3). The cause of this interaction was the particularly great positive effect of the combination of CT and HC on the increase in alcohol content, which was observed in both years of this study (Figure 4). Moreover, the interaction also occurred for the concentration of total anthocyanins in 2018 as, in full crop conditions, severe shoot trimming obtained a mild increase in the concentration of total anthocyanins, whereas in the reduced crop conditions obtained via cluster thinning, severe shoot trimming strongly decreased the concentration of total anthocyanins in wine.

3.5. Wine Sensory Characteristics

Corresponding to the overall effect of the investigated treatments on wine composition, the sensory characteristics of wines were more strikingly affected through the application of cluster thinning than severe shoot trimming. In both years of study, CT significantly increased the perception of aroma intensity, aroma complexity, dark berry aroma, wine body, velvety astringency, persistency, taste balance and overall wine quality, whereas it decreased the perception of acidity compared to FC (

Table 4. Sensory characteristics of wines obtained from grapes of Merlot vines subjected to cluster thinning (FC—full crop; and CT—cluster thinning) and shoot trimming (HC—high canopy; and SST—severe shoot trimming).

	Season 2017							Season 2018
	Cluster Thinning			Shoot Trimming			Int. b	Cluster Thinning			Shoot Trimming			Int.
	FC	CT	Sign. a	HC	SST	Sign.	Sign.	FC	CT	Sign.	HC	SST	Sign.	Sign.
Color intensity	7.5	8.2	ns	7.9	7.9	ns	ns	6.8	8.2	*	7.7	7.3	ns	ns
Aroma intensity	7.1	7.8	*	7.7	7.2	ns	ns	6.2	7.0	*	6.9	6.4	ns	*
Aroma complexity	6.5	7.3	*	7.2	6.6	ns	ns	5.8	6.9	*	6.7	6.0	ns	*
Vegetal aroma	2.7	2.3	ns	2.3	2.7	ns	ns	3.2	2.5	**	2.5	3.2	**	*
Red berry aroma	4.3	4.7	ns	4.8	4.1	ns	ns	4.5	4.4	ns	4.6	4.2	ns	ns
Dark berry aroma	5.8	7.2	*	6.6	6.4	ns	ns	4.5	6.0	*	5.7	4.8	ns	*
Acidity	6.5	5.9	*	5.9	6.5	*	ns	7.0	6.5	*	6.6	6.8	ns	ns
Body	6.6	7.4	*	7.2	6.8	ns	ns	6.4	7.4	**	7.2	6.5	*	*
Bitterness	4.2	4.1	ns	3.9	4.3	ns	ns	5.1	4.7	ns	4.9	4.8	ns	ns
Coarse astringency	5.9	5.7	ns	5.4	6.3	*	ns	6.4	6.3	ns	6.4	6.3	ns	ns
Velvety astringency	6.5	7.6	*	7.4	6.8	ns	ns	5.9	7.1	*	6.9	6.1	ns	*
Tannin perception intensity	6.6	7.1	ns	6.8	6.9	ns	ns	6.7	7.5	*	7.4	6.9	ns	ns
Persistency	6.6	7.5	*	7.5	6.6	*	ns	6.0	7.2	*	6.7	6.4	ns	*
Taste balance	7.4	8.0	*	8.1	7.3	*	ns	6.7	7.7	*	7.5	7.0	ns	ns
Overall wine quality	7.5	8.1	*	8.0	7.6	ns	ns	6.8	7.8	*	7.6	7.0	ns	ns

^a Data were analyzed via two-way ANOVA in randomized blocks design (ns, not significant; *, p ≤ 0.05; **, p ≤ 0.01). ^b Interaction between cluster thinning and shoot trimming.

). Furthermore, in the second year of the study the same treatment significantly increased the perception of color intensity and tannin perception intensity, whereas it decreased the perception of vegetal aroma. On the other hand, the application of severe shoot trimming had a milder effect on wine sensory characteristics than cluster thinning. This practice significantly increased the perception of vegetal aroma, acidity and bitterness compared to HC in one year of study and decreased the perception of wine body, persistency and taste balance, also in one year of study.

Although it could be expected that the application of cluster thinning would result in wines that are out of balance in term of the alcoholic notes or the hotness sensation due to their higher alcohol content [20,21], under the conditions of our study the sensory quality of wines benefitted from this practice. The taste balance of CT wines even increased, suggesting that the alcohol content in these wines was not above the threshold that can negatively affect the sensorial perception of the wine, while the wine composition was adequate to support the given alcohol content. The higher alcohol content in CT wines resulted in enhanced perception of wine body, which is in accordance with the results previously reported by other authors [64,65]. The same wines were also characterized by improved persistency (length in mouth), which is a wine sensorial characteristic that may be positively affected by the ethanol content [21,66]. On the other hand, SST treatment had less enhanced persistency and lower taste balance than HC in 2017, as well as lower wine body in 2018.

Interestingly, although no differences in titratable acidity among the investigated treatments were observed (Table 3), CT decreased the sensorial perception of acidity in both years and SST increased it in 2017 (Table 4). This outcome is presumably due to differences in perceived wine body and its ethanol concentration, which is capable of suppressing the perception of acidity in wines [20,67]. Zamora et al. [67] reported that the reduction in perceived acidity may be due to the increase in the sensation of sweetness caused by the increased concentration of ethanol.

Even though SST did not negatively affect the concentrations of total anthocyanins and phenolics in wine as compared to HC (Table 3), the sensory characteristics observed in some SST wines, such as the more pronounced vegetal aroma and acidity or the decreased perception of wine body and persistency, are usually obtained using unripe grapes [20,59,62,65,66]; these characteristics are commonly associated with lower wine quality by professional wine judges and consumers [21,66]. From these results, it can be deduced that the lower accumulation of sugars in berries with the application of late source limitation and the lower content of alcohol in SST wines resulted in a negative impact of this treatment on the sensory quality of wine.

In 2017, wines from SST had a more intense perception of coarse astringency; this trend was most probably associated with the less ripened grapes and the resultant higher amount of seed proanthocyanidins, which contribute to coarse astringency [60], and also possibly because of the increased acidity perception in the wines as the coarse astringency sensation perceived in wines may result from higher acidity [68,69]. The same wines also had lower taste balance because of the mentioned stronger perception of acidity and coarse astringency.

Higher grades of aroma intensity, aroma complexity and dark berry aroma observed in CT treatment are usually found in wines obtained with more ripe grapes and/or higher sugar content [21,59,65,66,70]. On the other hand, the more pronounced vegetal aroma perception in SST wines from 2018 is a typical sensory characteristic of grapes which are less ripened or early harvested [59,62,65,66].

The higher intensity of tannin perception in CT wines from 2018 (Table 4) was most probably due to their higher concentration of total phenolics (Table 3). On the other hand, the higher perception of velvety astringency, which was observed in CT wines from both years and considered as having smoother/softer tannins, is usually related to a higher alcohol content [20]; better fruit maturity [65]; higher content of extractable skin flavonoids, which usually increase with berry ripening and contribute to pleasant astringency; lower amount of seed pro-anthocyanidins, which decrease with ripening and contribute to coarse astringency [60]; and the higher concentration of anthocyanins in wine, which contribute to the velvety astringency [60,71].

Interestingly, in a study by Lu et al. [33] the practice of severe shoot trimming managed to increase the wine sensory score, though the concentration of alcohol in wine was not affected by this practice. In other studies where late source limitation reduced the wine alcohol content, a reduction in red fruit aroma attributes and color intensity was observed [37]; the wine overall sensory score was increased only with moderate post-veraison defoliation or trimming, whereas the more intensive removal of leaves negatively affected this parameter [22].

From the results of the specific combination of all treatments (Figure 5), it is clear that, in both years of study, the CT-HC combination obtained the most favorable sensory quality of wines, having a more intense perception of color intensity, aroma intensity, aroma complexity, dark berry aroma, body, velvety astringency, tannin perception intensity, persistency, taste balance and overall wine quality; the same treatment obtained lower intensity of vegetal aroma in wines. On the other hand, with some small differences between the two years, the other three combinations of treatments obtained values comparable to most sensory descriptors and similar overall wine quality.

Although we expected that the higher alcohol concentration obtained via cluster thinning, and especially via the CT-HC combination, would negatively affect taste balance and overall wine quality, this was not the case under the conditions of our study. Indeed, the alcohol content was the main driver of overall wine quality and its sensory perception, confirming the findings of some studies that used the addition of sugar or water to regulate the initial soluble solids content in grape must of several red varieties [59,62] or that had different harvest dates [65]. On the other hand, the increase in alcohol content via cluster thinning is a potential drawback if the sugar content in grapes from untreated vines is high, leading to excessive levels via cluster thinning which, as a practice, enhances fruit ripening. In fact, this is the case in several vineyards in the last years due to the changing climatic conditions [16,17,18,19,27]; according to the Intergovernmental Panel on Climate Change [72], global temperatures will further rise in the following decades, which will further increase this challenge in viticulture production.

4. Conclusions

Even though severe shoot trimming at late veraison (at 14–15 Brix) is an effective vineyard management practice to reduce the sugar content in berries, it can negatively affect the sensory quality of wines and, therefore, it is reasonable to apply it only in conditions that lead to unacceptably high alcohol content. On the other hand, cluster thinning promotes ripening dynamics and increases the wine sensory quality of Merlot variety, though it should be practiced with caution in conditions which are favorable for ripening because it may be responsible for the potentially excessive alcohol content in the wine. Therefore, the decision to adopt a practice to reduce or increase wine alcohol content should be based on the historical experience of a given vineyard and its resultant wines, while also taking into account the meteorological and growing conditions of the current season.

Based on the obtained results, we consider that severe shoot trimming could be used in viticulture for the production of lighter, fresher and simpler red wines with medium-to-light body which are not used for long maturation, instead being pre-determined for a relatively fast consummation and usually marketed at a lower price. On the other hand, if the aim of production is a full-bodied red wine that is rich in aromas and flavor and destined for a longer maturation before being released on the market, usually at a higher price, cluster thinning may be generally considered an appropriate vineyard management practice.