Agronomic Response of 13 Spanish Red Grapevine (Vitis vinifera L.) Cultivars under Drought Conditions in a Semi-Arid Mediterranean Climate

Abstract

Drought is perhaps the most important abiotic stressor affecting plants. Grapevine (Vitis vinifera L.) is a drought-tolerant species, and this feature makes it a traditional crop in semi-arid climate areas. However, not all cultivars respond to drought in the same way. Many studies on grapevine drought response have focused on physiological traits. This study mainly used agronomic indicators to assess the drought response of 13 red cultivars. Our results revealed high variability in must isotope ratios (δ¹³C and δ¹⁸O), yield components, and grape must quality. Bobal, Garnacha Peluda, Garnacha Tinta, Mazuela, and Moribel cultivars responded well to drought conditions, simultaneously maintaining high yields and must quality. By contrast, Garnacha Tintorera, Forcallat Tinta, and Tempranillo cultivars showed high water use efficiency but had low yield and must quality. Therefore, these cultivars can be considered poorly adapted to drought conditions. By knowing which cultivars perform well under drought conditions, viticulturists can reduce their reliance on water irrigation and continue to maintain vineyard sustainability in current and future semi-arid climatic conditions. This research also contributes novel information about the Castilla-La Mancha region, where there have been no previous similar assays.

1. Introduction

Drought stress is a critical factor in viticulture because it limits the yield and affects the grape quality [1,2,3] by altering grapevine physiology [4,5]. Currently, many wine-growing regions are vulnerable to drought, as vineyards are found mainly in dry climates. Many of these areas have semi-arid Mediterranean climates, characterized by warm and dry summers, in which vineyards are regularly exposed to long periods of seasonal drought. A 2016 report found that more than 90% of vineyards in Europe are rainfed [6]. This percentage is significantly lower in Spain, where it is 59.4% nationally and 49.3% in the Castilla-La Mancha region (2019 data obtained from the Spanish Ministry for Agriculture, Fisheries, and Food). Recent climate forecasts predict temperature increases and lower rainfall rates in the coming years, and consequently, drought events are likely to intensify [7]. Many traditional wine-growing regions will likely need to increase irrigation under these conditions. However, this is not a sustainable solution, particularly in semi-arid areas where water resources are increasingly limited [8,9]. In addition, the legislative regulation of water use by the European Union and individual countries is increasingly restrictive in these regions.

Although the grapevine genotypes currently cultivated in semi-arid environments are generally well adapted to water deficits and high temperatures, the suitability of some cultivars may soon become compromised [10,11,12]. As droughts intensify, they will negatively impact grapevine growth, physiology, and berry ripening, and vine yield and wines quality may suffer [13]. These effects are expected to be particularly severe for red cultivars, which may respond by producing grapes with lower concentrations of anthocyanins [14,15] and unbalanced in respect to sugar content [16,17].

Given these climate forecasts and their predicted effects on vineyards, identifying drought-tolerant cultivars is vital in developing strategies to maintain vineyard sustainability. However, drought tolerance is a complex concept [18], as it results from the combination of various types of stress, including low water availability, high evaporation demand, and heat stress. Consequently, no consensus has been reached as to which traits determine grapevine drought tolerance. Recently, a new quantitative measure named stress distance has been proposed. This parameter integrates four physiological traits related to the grapevines’ response to water stress: maximal transpiration rate, stomatal regulation, turgor loss point, and root volume [19].

To date, some studies on grapevine drought response have focused on agronomic indicators, such as yield and grapes quality [20,21,22,23,24,25,26]. Others have focused on physiological traits, such as stomatal regulation, carbon assimilation and hydraulic conductance, finding that both the rootstock and the scion genotypes confer drought adaptability traits, but in different ways [5,19,27,28,29,30]. While the rootstock influences the ability of the plant to extract water from the soil, the scion influences the sensitivity of the stomatal control [30]. Further studies are necessary to determine to what extent differences in yield or water use are attributable to innate genotypic differences between cultivars or environmental factors [24,31].

To determine the vine water status, the carbon isotope ratio of the grape must (δ¹³C) can be used. This is because the δ¹³C of the grape must at harvest is considered a physiological indicator of water stress that differs between rootstocks and cultivars [32]. It is an integrated marker of vine water status during berry growth, particularly during the veraison to harvest periods [33,34,35,36]. Carbon in plant tissues is fixed through photosynthesis and comes from existing in the atmosphere. In nature, there are two stable carbon isotopes: ¹²C is the lightest and more abundant form (98.93%), whereas ¹³C is heavier and represents 1.07% [37]. When C3 plants, such as grapevines, experience water deficit, stomatal closure is enhanced, and as a consequence, the ¹³C/¹²C ratio (δ¹³C) increases. This is because under these conditions there is less discrimination of the ¹³C form, due to the joint action of two processes. On the one hand, its diffusion through the stomata is prevented [38] and, on the other hand, its lower reactivity against the enzyme ribulose-1.5-biphosphate carboxylase-oxygenase (Rubisco), which prefers the ¹²C form, favors the relative proportion of ¹³C in the leaf. Therefore, it is common to observe that tissues of C3 plants growing under water deficit show greater δ¹³C than those from non-water-stressed plants [39,40].

In the present study, 13 Spanish red cultivars authorized for cultivation in Castilla-La Mancha region were monitored for three years under drought conditions in a multivarietal vineyard. The plants responses were assessed based on agronomic indicators, including the yield, grape quality, and vegetative development. The carbon isotope ratio of the grape must was measured at harvest as an indicator of vine water status. The study aimed to identify cultivars maintaining good yield and high grape quality under rainfed conditions, regardless of the genotypic traits or tolerance mechanisms leading this adaptation. This is the first comparative study of the drought behavior of grapevine cultivars performed in the La Mancha region.

2. Materials and Methods

2.1. Location and Plant Material

The present study was conducted from 2018 to 2020 in a multivarietal vineyard in Tomelloso, Spain (39°10′34” N; 3°00’01” W; 663 m.a.s.l.), belonging to the Regional Institute of Agri-Food and Forestry Research and Development of Castilla-La Mancha (IRIAF). Thirteen traditional red cultivars were selected for the present study: Bobal, Forcallat Tinta, Garnacha Peluda, Garnacha Tinta, Garnacha Tintorera, Graciano, Mazuela (syn. Carignan), Monastrell, Moravia Agria, Moribel, Tempranillo, Tinto de la Pámpana Blanca, and Tinto Velasco. The cultivars were arranged in parallel rows of 140 plants each. The plants were 15 years old and grafted onto 110-Richter rootstock. They were trained on a bilateral Royat cordon system and planted with a row spacing of 3 m and a vine spacing of 1.5 m. Twenty-two contiguous plants per cultivar were included in the study.

2.2. Climate, Soil, and Vine Water Regime

Tomelloso’s climate is Mediterranean continental semi-arid, with hot and dry summers and cold and moderately rainy winters. The annual thermal amplitude is high (21.5 °C). The mean annual rainfall is about 380 mm, of which only about 40% falls during the growing season, and drought periods are long (4.5 months). Figure 1 shows the monthly rainfall and mean air temperature at the study site during the three agronomic years of the experiment alongside the 20-year averages (2001–2020). The weather data used in this study were recorded at the Argamasilla de Alba weather station 6 km away from the experimental site. The station belongs to the SIAR network of the Spanish Ministry for Agriculture, Fisheries, and Food. The vapor pressure deficit (VPD) values were calculated using the FAO 56 methodology [41]. Each VPD value represents the average of the daily mean VPD values during the 7 days prior to the harvest date.

The soil at the study site is classified as Petric Calcisol (FAO soil classification) or Petrocalcic Calcixerept (USDA soil classification), typical of the La Mancha region. The main characteristic of this soil is the presence of a dense C_km horizon 30 cm below the surface, impenetrable to grapevine roots. Subscripts k and m indicate carbonate accumulation and strong cementation or hardening, respectively. The pedoclimatic soil conditions are characterized by the typical xeric moisture regime of Mediterranean climates. The vineyard soil was managed mechanically by mowing the permanent natural plant cover throughout the year. Water was supplied to the vineyard by rainfall. However, at specific times, it was necessary to rehydrate the vines to ensure their survival. For this purpose, three 10 mm irrigations with local water from an aquifer well were carried out throughout the season each year (after fruit set, when the branches stopped growing, and when the berries reached the size of a pea).

2.3. Phenology

The phenology of the cultivars was monitored, noting the dates on which the main season phenological stages took place, according to the BBCH scale [42]: budbreak (07), flowering (65), veraison (81), and maturity (89). The phenological stage dates were assigned when 50% of the buds/bunches of monitored vines reached that stage, with the exception of maturity. This date was assigned when 100% of the berries were mature. Figure 2 graphically represents the mean phenology stages of each variety.

2.4. Vine Water Status and Must Carbon Isotope Analysis

The carbon isotope composition of grape must was measured by on-line analysis using a ThermoQuest Flash 1112 Elemental Analyzer equipped with an autosampler and coupled to a Delta-Plus IRMS (ThermoQuest, Bremen, Germany) through a ConFlo III interface (ThermoQuest). One microliter of must was placed in a tin capsule and sealed. All the carbon in the sample is oxidized to CO₂ by the reactors of the elemental analyzer. The analyzer passes the gas through a gas chromatography (GC) column to separate the CO₂ from other gases and then brings the CO₂ into the mass spectrometer by a helium flow [43]. Carbon isotope composition was expressed as:

δ¹³C_sample = [(R_s/R_std) − 1] × 1000

(1)

where R_s is the ¹³C/¹²C ratio of the sample and R_std is the international reference standard Vienna Pee Dee Belemnite. Five vines of each cultivar were sampled, with two must samples per vine.

2.5. Must Oxygen Isotope Analysis

An on-line gas equilibration and headspace introduction system model, GasBench II (ThermoQuest, Bremen, Germany), was used to analyze must oxygen isotopes. It is equipped with a GC column (PoraPlot Q, 25 m, 0.25 mm; Varian, Palo Alto, Santa Clara, CA, USA) operating at 70 °C and adapted to an autosampler CombiPAL (CTC-Analytics, Zwingen, Switzerland). For each must sample, 500 µL of must and a spatula tip of benzoic acid (to avoid possible fermentation) were transferred to a 10 mL vial with silicone septa. The vials were placed in the GasBench II, flushed with 0.3% CO₂ in He for 10 min, and left for 48 h at 22 °C before analysis. During this equilibration time, an exchange reaction took place between the oxygen in CO₂ and H₂O:

¹²C¹⁶O₂ + H₂¹⁸O ↔ ¹²C¹⁶O¹⁸O + H₂¹⁶O

(2)

The CO₂ was then isolated from the vial headspace and introduced in the IRMS system. The GasBench II was coupled to a Delta-Plus IRMS (ThermoQuest) equipped with three Faraday cup detectors that simultaneously and continuously monitor the [CO₂]⁺ signals for the three major ions at m/z 44 (¹²CO₂), m/z 45 (¹³CO₂ and ¹²C¹⁷O¹⁶O) and m/z 46 (¹²C¹⁸O¹⁶O). Oxygen isotope composition was expressed as:

δ¹⁸O_sample = [(R_s/R_std) − 1] × 1000

(3)

where R_s is the ¹⁸O/¹⁶O ratio of the sample and R_std is the international reference standard Vienna Standard Mean Ocean Water. Five vines of each cultivar were sampled, with two must samples per vine.

2.6. Yield Components, Pruning Weight, and Grape Quality Components

To determine the harvest date, regular sampling was performed until the grapes reached 22.5–24.5 °Brix. However, in some varieties, the harvest date had to be brought forward because the yield was too high and the grapes were unable to reach a higher sugar concentration. In these cases, and in view of the possibility that the quality of the grapes might deteriorate, it was decided to bring forward the harvest date. By contrast, Graciano reached 26.3 °Brix while maintaining a high level of acidity (6.52 g L⁻¹).

Five vines were harvested per cultivar. The vine yield was measured in terms of weight (kg vine⁻¹) and bunches per vine. The mean bunch weight (g) of each vine was calculated by dividing the vine yield by the number of bunches. To determine the mean berry weight (g), 100 berries from each vine were randomly selected and weighed. During the winter, pruning weight (kg vine⁻¹) was measured by pruning the same five vines and weighing the pruned wood. To evaluate the balance between production and vigor in each sample, the Ravaz index was calculated. The grapes of each harvested vine were pressed with a manual screw press to extract the must. Following the official methods of the International Organization of Vine and Wine [43], the total soluble solids (°Brix) of the must was measured by electronic refractometry (RX-5000α-Bev, Atago, Tokyo, Japan), and the total acidity (g L⁻¹ tartaric acid) and pH were measured by potentiometry (HI 902, Hanna, Eibar, Spain). The must was sampled (12 mL per vine) and frozen in polycarbonate test tubes to allow later isotope ratio analysis (δ¹³C and δ¹⁸O).

2.7. Statistical Analysis

Statistical treatment of the data was performed using Statgraphics Centurion XVIII software (Statgraphics Technologies, The Plains, VA, USA). First of all, the three-year data set (n = 15) was analyzed for outliers (data exceeding three standard deviations). No outliers were found, so all data for each variable came from the same distribution. For each of the observed variables, the mean values of the cultivars were compared by applying the Student–Newman–Keuls (S–N–K) multiple range test with a 95% confidence level using a one-factor analysis of variance (ANOVA). The method used for simple regression was least squares with a 95% (α = 0.05) confidence interval (bilateral). Principal component analysis (PCA) was also performed to study the possible grouping of varieties to drought response. The PCA groups were identified by cluster analysis. Box plots were generated using IBM SPSS Statistics 25 Core System software (SPSS, Armonk, New York, NY, USA).

3. Results

3.1. Must Carbon Isotope Ratio

The carbon isotope ratio of grape must was highly variable between cultivars (Figure 3). The δ¹³C of individual vines ranged from −25.93 ‰ to −21.41‰, and the mean δ¹³C of cultivars ranged from −23.94‰ to −22.46‰. These wide ranges reveal substantial differences in the water status of the vines during ripening. According to the Student–Newman–Keuls test, the δ¹³C of grape must at harvest was significantly affected by genotype (p < 0.001). Garnacha Tintorera exhibited the highest mean δ¹³C, indicating that this cultivar is very efficient in water use [44,45]. By this measure, Garnacha Tinta is moderately efficient, whereas Mazuela and Graciano displayed low δ¹³C and thus are less efficient. The intra-cultivar variability of δ¹³C was also high across the three years of the study. Tinto de la Pámpana Blanca was the most highly variable cultivar in addition to having the lowest minimum δ¹³C value.

3.2. Must Oxygen Isotope Ratio

The oxygen isotope ratio of the must water was closely related to the mean VPD in the week before harvest, as indicated by a double inverse model (Figure 4). The VPD values ranged from 0.5 kPa to 2.3 kPa, and Tempranillo was the cultivar with the greatest VPD.

The δ¹⁸O in grape must water ranged widely (Figure 5), and differences between cultivars were highly significant (p < 0.001). Tinto Velasco had the smallest mean δ¹⁸O (8.12‰), indicating that this cultivar had the lowest transpiration rates during the seven days before harvest [46,47,48], while Garnacha Tinta exhibited the largest mean (11.33‰; Figure 5). As was observed for δ¹³C, intra-cultivar variability in δ¹⁸O was high. Tempranillo and Bobal presented the highest and lowest variability, respectively. Tempranillo also had the largest δ¹⁸O value.

3.3. Yield Components and Pruning Weight

Considering all cultivars, when comparing 2019 to 2018, the mean yield per vine fell by 46.4%, and pruning weight fell by 39.7%. The subsequent declines from 2019 to 2020 were somewhat smaller: 32.0% for yield and 36.2% for pruning weight. All the yield components were highly variable between cultivars.

Table 1. Yield components and pruning weight (mean value ± standard deviation, n = 15) of 13 red cultivars.

Cultivar	Yield (kg vine−1)	Bunch Weight (g)	Berry Weight (g)	Pruning Weight (kg vine−1)	Ravaz Index
Monastrell	2.06 ± 1.00 a	120.01 ± 40.36 a	1.05 ± 0.20 a	0.28 ± 0.09 ab	7.27 ± 2.62 abcd
Forcallat Tinta	2.20 ± 1.13 ab	172.05 ± 47.84 ab	1.45 ± 0.13 b	0.33 ± 0.13 ab	6.80 ± 2.83 abc
Graciano	2.47 ± 1.81 abc	138.94 ± 55.47 ab	0.97 ± 0.18 a	0.27 ± 0.14 a	9.39 ± 4.83 bcd
Tempranillo	2.74 ± 1.85 abc	156.83 ± 79.98 ab	1.34 ± 0.20 b	0.47 ± 0.25 bc	5.64 ± 3.00 a
Garnacha Tintorera	2.89 ± 2.21 abc	114.20 ± 51.42 a	1.49 ± 0.27 b	0.36 ± 0.21 abc	7.30 ± 3.05 abcd
Garnacha Peluda	3.00 ± 1.74 abc	126.95 ± 44.46 ab	1.36 ± 0.17 b	0.42 ± 0.20 abc	7.26 ± 2.76 abcd
Garnacha Tinta	3.11 ± 2.17 abcd	128.10 ± 60.41 ab	1.34 ± 0.20 b	0.40 ± 0.20 abc	7.15 ± 2.49 abcd
Bobal	3.18 ± 0.91 abcd	329.63 ± 98.70 d	2.01 ± 0.21 c	0.56 ± 0.18 c	6.13 ± 2.08 ab
Mazuela	3.56 ± 1.20 abcd	150.90 ± 38.53 ab	1.39 ± 0.29 b	0.47 ± 0.14 bc	7.56 ± 1.63 abcd
Moribel	4.04 ± 2.08 bcd	185.77 ± 47.03 bc	1.56 ± 0.18 b	0.41 ± 0.16 abc	9.82 ± 2.58 cd
Tinto Velasco	4.22 ± 1.28 cd	222.29 ± 53.00 c	1.98 ± 0.35 c	0.52 ± 0.14 c	8.52 ± 3.31 abcd
Tinto de la Pámpana Blanca	4.38 ± 1.48 cd	236.97 ± 54.73 c	2.00 ± 0.19 c	0.45 ± 0.15 abc	10.22 ± 3.31 d
Moravia Agria	4.88 ± 2.62 d	230.01 ± 76.64 c	1.34 ± 0.18 b	0.39 ± 0.18 abc	12.56 ± 3.41 e

Different letters in the same column denote statistically significant differences between cultivars (ANOVA, S–N–K test, p < 0.001).

presents the mean yield per vine of the cultivars in ascending order. Moravia Agria was the cultivar with the highest mean yield per vine (above 4 kg vine⁻¹). However, when the cultivars were ranked by their berry and bunch weight, Bobal stood out above the other cultivars. Bobal and Tinto Velasco exhibited the highest pruning weight (above 0.50 kg vine⁻¹), a measure of vigor. By contrast, Graciano had the smallst pruning weight (0.27 kg vine⁻¹). Regarding the Ravaz index (yield/pruning weight), Tempranillo and Bobal stood out with the smallest values (5.64 and 6.13, respectively), while Tinto de la Pámpana Blanca and Moravia Agria exhibited the largest (values above 10).

3.4. Must Quality Components

The must quality components also displayed high variability between cultivars.

Table 2. Must quality components (mean value ± standard deviation, n = 15) of 13 red cultivars.

Cultivar	Total Soluble Solids (°Brix)	Total Acidity (g L−1)	pH
Moravia Agria	20.27 ± 1.59 a	6.58 ± 1.05 f	3.12 ± 0.12 ab
Garnacha Tintorera	20.83 ± 2.39 ab	4.29 ± 0.65 b	3.34 ± 0.21 cd
Tinto de la Pámpana Blanca	21.09 ± 1.97 abc	4.88 ± 0.44 bc	3.26 ± 0.13 bcd
Tinto Velasco	21.65 ± 1.43 bcd	4.85 ± 0.45 bc	3.37 ± 0.22 cd
Forcallat Tinta	22.11 ± 0.52 cde	3.57 ± 0.35 a	3.43 ± 0.18 cd
Tempranillo	22.39 ± 1.37 cdef	4.21 ± 0.78 b	3.45 ± 0.20 d
Mazuela	22.71 ± 1.11 defg	6.09 ± 0.77 ef	3.26 ± 0.19 bcd
Monastrell	22.92 ± 1.00 defg	4.81 ± 0.54 bc	3.37 ± 0.14 cd
Garnacha Tinta	23.53 ± 0.90 efg	4.83 ± 0.59 bc	3.26 ± 0.19 bcd
Moribel	23.55 ± 1.51 efg	5.54 ± 0.84 cde	3.31 ± 0.18 cd
Bobal	23.74 ± 1.20 fg	5.80 ± 1.34 de	3.36 ± 0.16 cd
Garnacha Peluda	24.12 ± 1.08 g	5.30 ± 0.71 cd	3.23 ± 0.20 bc
Graciano	26.32 ± 1.87 h	6.52 ± 1.06 f	3.06 ± 0.15 a

Different letters in the same column denote statistically significant differences between cultivars (ANOVA, S–N–K test, p < 0.001).

presents the mean total soluble solids of the cultivars in ascending order. Moravia Agria and Graciano exhibited the lowest and highest total soluble solids content (20.27 °Brix and 26.32 °Brix, respectively); the two cultivars also stood out for their high total acidity (6.58 g L⁻¹ and 6.52 g L⁻¹, respectively). Tempranillo was the cultivar with the highest must pH (3.45).

4. Discussion

Previous water stress assays have reported variation in drought response between different grapevine cultivars [28,30,44,45,49,50], clones [45,51], and rootstocks [27,30,50,52,53,54,55,56,57,58]. However, the tolerance mechanisms driving this variation in the adaptability of grapevines to drought conditions have not yet been completely elucidated. Many studies have explored the role of physiological traits in controlling differences between genotypes in drought response [5,59,60,61,62], but fewer studies have investigated the role of agronomic indicators. The present work is the first to assess the response of grapevine cultivars to drought in the Castilla-La Mancha wine region.

4.1. Cultivar Water Status and Isotope Ratios

This study found significant variation between cultivars in the δ¹³C of their grape must. The observed δ¹³C values suggest severe drought stress in all cultivars, with mean values above −24‰ [36]. These findings are consistent with previously reported values from viticultural areas including Bordeaux [32,63], Montpellier [64], Tavel [65], Navarra [66], and La Mancha [67]. Conversely, in assays performed in other geographical regions, including the Balearic Islands, the mean δ¹³C was substantially lower [44]. These differences could be due to the environmental conditions, rootstock traits, or vine organ sampled (fruit or leaf). For example, it is well established that the δ¹³C values obtained from the grape pulp are higher than values obtained from other grapevine organs and tissues such as leaves, whole berries, skins, or seeds [1,61,68]. In this study, the mean δ¹³C values of the grape must did not vary much between cultivars over the three years (2‰). The values observed are similar to those of previous studies obtained from leaves [44,61,69,70,71,72], but they are substantially lower than values reported from grape must [32,63,64,67].

Although the cultivars were all grown under the same conditions, they varied in the water use efficiency. According to results obtained, Garnacha Tintorera was the most efficient cultivar (high δ¹³C values), whereas Graciano and Mazuela (low δ¹³C values) behaved as the least efficient cultivars. The remaining cultivars were categorized into four groups. Some of the most widely cultivated varieties in Spain were among these remaining cultivars, including Garnacha Tinta and Tempranillo. A previous study performed in rainfed vineyards in the Bordeaux region has found that Mazuela is not a very water-use-efficient cultivar, whereas Garnacha Tinta is [32]. The δ¹³C values they obtained and their conclusions regarding water use efficiency are consistent with the present study. By contrast, in the La Mancha region, the δ¹³C of non-irrigated Garnacha Tinta vines was much higher than what was observed in the present study [67]. These differences could be due to differing environmental conditions during the growing season in which the assays were performed, sampling methods, or rootstocks traits. For the Tempranillo cultivar, the carbon isotope ratios of different organs and tissues have been widely evaluated [60,66,67,73], but few studies have evaluated its grape must. In this study, Tempranillo’s high δ¹³C suggested that it is more efficient in terms of water use than Garnacha Tinta. In contrast with these results, other researchers have considered Tempranillo a less efficient cultivar than Garnacha Tinta [67].

Previous studies have shown that δ¹⁸O can complement δ¹³C in estimating the accumulated water deficit under drought conditions since the grapevine’s transpiration rate in the days preceding harvest determines the oxygen isotope composition in must water [36,38]. Conversely, assays performed on the wood of forests trees have found that the correlation between the two isotope ratios is generally poor, suggesting that although both ratios are somewhat related to transpirative demand, they are relatively independent [74]. These previous findings are consistent with those obtained in the present study, in which no significant correlation was found between the two isotope ratios. By contrast, δ¹⁸O and the average of the daily mean VPD values during the 7 days prior to the date of harvest were closely correlated (Figure 4). Previous reports have found that δ¹⁸O isotope ratios are affected by the presence of vine irrigation and the climatic conditions in which the assays are performed. The δ¹⁸O values observed in this study are comparable to those obtained for non-irrigated cultivars in the same geographic region [67]. By contrast, in other areas such as Korea [75], the Crimean Peninsula, the Krasnodar region, the Rostov region, and the Republic of Daguestan [76], the δ¹⁸O is much lower, with some areas even presenting negative values. These differences may be because transpiration through the leaves and skin of the grape during the ripening season is high in semi-arid climatic conditions. Consequently, the δ¹⁸O of the grape water is elevated under these dry conditions [47]. In addition, the oxygen isotope composition of water differs depending on the geographic location. For example, the δ¹⁸O of atmospheric precipitation water and groundwater are higher in southern Europe than northern Europe [77].

The variation observed in δ¹⁸O was mainly observed for the VPD in the seven days immediately before harvest and not by differences between the cultivars. The δ¹⁸O was higher in moderately early harvest cultivars such as Garnacha Tinta, Garnacha Peluda, and Garnacha Tintorera (Figure 2). Their maturation took place in atmospheric conditions with high daily mean VPD values (low relative humidity and high air temperature), and consequently, these vines had a high transpiration rate. Conversely, in late harvest cultivars such as Tinto Velasco, Graciano, Forcallat Tinta, and Tinto de la Pámpana Blanca, the grapes ripened on days with low VPD, therefore, these varieties exhibited lower δ¹⁸O values. Although Tempranillo was harvested before Garnacha Tinta, its δ¹⁸O values were slightly lower than Garnacha Tinta’s, which agrees with the results obtained by other authors for the same cultivars under rainfed conditions [67].

4.2. Agronomic Response

Cultivars differed significantly in their yield components and pruning weight. The highest yields were observed for cultivars such as Tinto Velasco, Tinto de la Pámpana Blanca, and particularly for Moravia Agria, which often exceeded 4 kg vine⁻¹. These cultivars also exhibited heavy grape bunches, but Bobal had the highest bunch weight of all cultivars. Bobal and Tinto de la Pámpana Blanca had the heaviest berries. Conversely, Monastrell was the cultivar with the lowest yield per vine and low bunch and berry weight. This finding is consistent with a previous study comparing rootstocks, which found that Monastrell has a low yield when grafted onto 110-R [58]. In the present study, Bobal and Tinto Velasco were the most vigorous cultivars, as measured by pruning weight. Garnacha Tinta exhibited moderate yield per vine and vigor, as well as low bunch and berry weight. These findings partially contradict the results obtained by other authors concerning the effects of water stress in Mediterranean climates [78]. The different results may be due to differences in soil, crop management, or rootstock.

In the higher yielding cultivars, such as Moravia Agria, Tinto de la Pámpana Blanca and Tinto Velasco, the grapes could not ripen fully, because they had a high Ravaz index. As a consequence, their final total soluble solids content was low. However, other cultivars such as Moribel and Graciano had similar ratios but matured properly. Moravia Agria and Graciano were the cultivars with the highest total must acidity. Conversely, the two cultivars differed in total soluble solids content: while Graciano had the highest content of all cultivars, Moravia Agria had the lowest. Garnacha Tinta matured adequately, reaching acceptable soluble solids content and acidity level, consistent with the results obtained from other regions with Mediterranean climates [78].

4.3. Categorizing Cultivars Based on Their Agronomic Response to Drought

A multivariate data analysis using PCA was performed to determine how the agronomic traits varied between cultivars. The five variables least correlated with each other were selected for the PCA: yield, pruning weight, total soluble solids, total acidity, and δ¹³C. Other variables were excluded to simplify the interpretation of the PCA. Figure 6 shows the biplot defined by the first two Principal Components (PCs), explaining 75.1% (41.07% and 34.03% for PC 1 and PC 2, respectively) of all variance for the 13 cultivars. To simplify the multivariate analysis, only these first two PCs were considered. The cluster analysis enabled the cultivars to be separated into four groups. A group formed by Mazuela, Moribel, Bobal, Garnacha Peluda, and Garnacha Tinta is located in the center of the graph, and a group composed of Tempranillo, Monastrell, Garnacha Tintorera, and Forcallat Tinta is located toward the bottom of the graph. Graciano is isolated from the other cultivars, and is located on the upper left side of the biplot. The last group, formed by Moravia Agria, Tinto Velasco, and Tinto de la Pámpana Blanca, is located on the upper right side.

Finally, to synthesize the results obtained by this study, the cultivars were classified again according to three arbitrary tolerance categories (high, moderate, and low) using the same traits as those chosen for the PCA (

Table 3. A synthesis of the study results categorizes the cultivar’s drought tolerance according to five traits: yield, pruning weight, total soluble solids, total acidity, and δ¹³C. Category cut-offs are arbitrarily chosen (adapted from [44]).

Trait	Category 1 (High)	Category 2 (Moderate)	Category 3 (Low)
Yield	(>4 kg vine−1) Moribel, Tinto Velasco, Tinto de la Pámpana Blanca, Moravia Agria	(3–4 kg vine−1) Garnacha Peluda, Garnacha Tinta, Bobal, Mazuela	(<3 kg vine−1) Monastrell, Forcallat Tinta, Graciano, Tempranillo, Garnacha Tintorera
Pruning weight	(>0.50 kg vine−1) Tinto Velasco, Bobal	(0.35–0.50 kg vine−1) Garnacha Tintorera, Moravia Agria, Garnacha Tinta, Moribel, Garnacha Peluda, Tinto de la Pámpana Blanca, Tempranillo, Mazuela	(<0.35 kg vine−1) Graciano, Monastrell, Forcallat Tinta
Total soluble solids	(>24.5 °Brix) Graciano	(22.5–24.5 °Brix) Mazuela, Monastrell, Garnacha Tinta, Moribel, Bobal, Garnacha Peluda	(<22.5 °Brix) Moravia Agria, Garnacha Tintorera, Tinto de la Pámpana Blanca, Tinto Velasco, Forcallat Tinta, Tempranillo
Total acidity	(>6.0 g L−1) Mazuela, Graciano, Moravia Agria	(4.5–6.0 g L−1) Bobal, Monastrell, Garnacha Tinta, Tinto Velasco, Tinto de la Pámpana Blanca, Garnacha Peluda, Moribel	(<4.5 g L−1) Forcallat Tinta, Tempranillo, Garnacha Tintorera
δ13C	(>−22.5‰) Garnacha Tintorera	(−22.5 to −23.5‰) Forcallat Tinta, Garnacha Peluda, Tinto de la Pámpana Blanca, Garnacha Tinta, Tinto Velasco, Monastrell, Moravia Agria, Tempranillo, Bobal	(<−23.5‰) Moribel, Mazuela, Graciano

). As expected, there were no cultivars with all traits classified in the high category. Bobal, Garnacha Peluda, Garnacha Tinta, Mazuela, and Moribel had high to moderate drought tolerance according to this categorization system. All five cultivars were classified in the high or moderate categories for yield, pruning weight, and must quality, therefore, they can be considered appropriate cultivars for semi-arid conditions. However, there were some differences between the cultivars. As indicated by δ¹³C, Bobal, Garnacha Peluda, and Garnacha Tinta were more efficient in their water use than Mazuela or Moribel. By contrast, Garnacha Tintorera, Forcallat Tinta, and Tempranillo were highly water efficient, but had low yield and must quality. Therefore, they can be considered the least adapted to drought conditions of the 13 studied cultivars. These findings reveal that a cultivar’s water-regulation strategy, such as isohydric or anisohydric behavior, is not critical in determining its agronomic response to drought. Graciano and Monastrell exhibited low yields and high must quality. Conversely, Moravia Agria, Tinto de la Pámpana Blanca, and Tinto Velasco had high yields, but an imbalance between the yield and leaf surface prevented the grapes from fully ripening, leading to must with high acidity, but low total soluble solids content. This categorization of the cultivars using arbitrarily chosen characteristics is consistent with the groupings formed by the PCA, confirming the similarities and differences between cultivars in their agronomic responses to drought.

5. Conclusions

For viticulture to remain sustainable in the coming years, it is essential to understand how grapevine cultivars yield and grape quality respond to drought. This study classified 13 red grapevine cultivars based on their agronomic drought response. The results demonstrate that not all cultivars respond equally well to drought. The δ¹³C was not correlated to any agronomic variables, and consequently, no water use efficiency patterns could be established. However, cultivars with lower water use efficiency generally had a better agronomic response to drought. Bobal, Garnacha Peluda, Garnacha Tinta, Mazuela, and Moribel showed high to moderate drought tolerance, simultaneously maintaining high yield and must quality. This behavior suggests that these cultivars may be suitable for cultivation in current and future semi-arid climatic regions, where water resources are scarce.