Phenological Sensitivity of Cabernet Sauvignon to Water Stress: Vine Physiology and Berry Composition

Transcription

1 Phenological Sensitivity of Cabernet Sauvignon to Water Stress: Vine Physiology and Berry Composition Boris Basile, 1 * Jordi Marsal, 2 Mercè Mata, 2 Xavier Vallverdú, 2 Joaquim Bellvert, 2 and Joan Girona 2 Abstract: There is little information on the sensitivity of berry composition to early-season water stress and how it compares to the effects of late-season stress. This study aimed to quantify the effects of water stress on berry growth and composition of Cabernet Sauvignon grapevine at three phenological stages: anthesis to fruit set, fruit set to veraison, and veraison to harvest. Potted vines were used to facilitate imposing water stress early in the season. Four irrigation levels (0%, 25%, 50%, and 100% of calculated crop evapotranspiration, ET c ) were applied and midday leaf water potential and leaf gas exchange were measured. Berry composition was evaluated by measuring titratable acidity and concentrations of soluble solids, anthocyanins, and polyphenols. Water stress decreased net CO 2 exchange rate and vine green leaf area. Berry composition significantly correlated with the vine water status, but the nature of the relationship depended on the phenological stage and on the parameter measured. Berry composition (in terms of concentration of anthocyanins and polyphenols) was improved when no water stress occurred from anthesis to fruit set (irrigation replacing 100% of ET c ), with mild water stress between fruit set and veraison (irrigation replacing 25% and 50% of ET c ), and with moderate to severe water stress in postveraison (irrigation replacing 0% of ET c ). Key words: anthocyanins, polyphenols, soluble solids, leaf abscission The effect of water stress on berry growth and composition depends on the phenological stage at which irrigation is withheld (Matthews and Anderson 1989, Girona et al. 2009). Research has typically considered two broad phenological stages: preveraison and postveraison. That is the case for Cabernet Sauvignon, which is popular in dry areas such as California, Australia, and Spain. Yet the available information is not conclusive on how Cabernet Sauvignon berry composition is affected even considering only two broad phenological stages. Berry soluble solids concentration (SSC) and wine quality scores were reported as significantly higher for vines exposed 1 Dipartimento di Arboricoltura, Botanica e Patologia Vegetale, Università degli Studi di Napoli Federico II, Via Università, 100, Portici, Napoli, Italy; and 2 Irrigation Technology, Institut de Recerca i Tecnologia Agroalimentàries (IRTA), Centre UdL-IRTA, Lleida, Spain. *Corresponding author ( boris.basile@unina.it; tel: ; fax: ) Acknowledgments: This work was partly supported by the Spanish Ministry of Education and Science (PETRI OP and CONSOLIDER CSD ) and was carried out under a research agreement between Codorniu (Raïmat, Lleida, Spain) and the IRTA. The authors thank Jesús del Campo, Carles Paris, Iñigo Auzmendi, Núria Bonastre, Núria Civit, Gerard Piñol, Joan Ventura, Xavier Domingo, and Amadeu Arbonés for their important technical assistance in the field; Raïmat Wineries and Joan Esteve and Xavier Ferré for their support, valuable comments, and suggestions throughout the experiment; and M. Hossein Behboudian, Massey University, New Zealand, for his critical comments on the manuscript. Manuscript submitted Jan 2011, revised May 2011, accepted Jun Publication costs of this article defrayed in part by page fees. Copyright 2011 by the American Society for Enology and Viticulture. All rights reserved. doi: /ajev to deficit irrigation (DI) during postveraison (Bravdo et al. 1985), although plant water status was not measured. Severe water stress (midday stem water potential of to MPa), applied during both pre- and postveraison, reportedly increased wine color intensity and concentrations of phenols and anthocyanins compared to irrigated control (Ferreyra et al. 2004). In a different study, severe water stress (leaf water potential [Ψ leaf ] values of MPa) during early postveraison did not affect berry anthocyanin concentration at harvest (Castellarin et al. 2007a). One study reported that preveraison DI slightly decreased juice color compared to postveraison DI; there were no consistent significant effects of timing of DI on other berry composition parameters (SSC, titratable acidity, or ph) (Keller et al. 2008). The inconsistencies in the above results may be related to differences among studies in the timing of water stress, its severity, and duration. In addition, preveraison water stress is a long period that spans from flowering to fruit set and finishes at veraison. The sensitivity of berry composition to water stress might vary across the preveraison period. To the best of our knowledge, for Cabernet Sauvignon there is no comprehensive study that relates a broad range of water status values to berry composition for phenological stages from flowering to fruit set and fruit set to veraison as compared to postveraison. For other cultivars, there is also no information available about the effects of early-season water stress on berry composition at harvest. In one study where water stress was applied to Cabernet franc for 22 days starting from anthesis, yield was decreased but berry composition was not affected (Hardie and Considine 1976). Our research was conducted in Catalonia, Spain, which is a dry area (average annual rainfall of 377 mm for ) 452

2 Responses of Cabernet Sauvignon to Water Stress 453 and drought can occur from May until harvest. Since dry winters are common in this area, water stress may occur between anthesis and fruit set in some years and pass unnoticed by growers. Container-grown plants were used to ensure that water stress was introduced in plants during fruit set. There is little information on how berry composition is affected by water stress during early phenological stages. Our objective was to determine how Cabernet Sauvignon would respond to water stress, in terms of fruit growth and composition at harvest, depending on the severity of water stress and phenological stage. These findings should be useful in devising a suitable irrigation strategy for Cabernet Sauvignon. Materials and Methods Experimental site and plant material. The experiment was carried out in Raïmat ( N; E), Catalonia, Spain. In March 2007, 400 one-year-old Cabernet Sauvignon grapevines grafted onto rootstock SO4 were planted in 50 L plastic containers (50 cm top diameter, 41 cm depth) with drainage holes. The growing media was a substrate mix of 50% peat and 50% silty-loam soil. Vines were fully irrigated throughout 2007 and then dormant-pruned to two, 2-bud spurs per vine. In 2008, 360 uniform vines were selected and aligned in 10 rows of 36 containers, with rows 3 m apart. The pots were placed next to each other, allowing 0.5 m spacing between vines. Growing shoots were positioned vertically using a trellis system made of four horizontal galvanized steel wires located along each row of containers at four different heights (50, 100, 150, and 200 cm aboveground). The vines were managed (disease control, nutrition) according to the winegrape production protocol defined by the Costers del Segre Denomination of Origin (Catalonia, Spain). Crop evapotranspiration. Crop evapotranspiration (ET c ) was calculated according to the following equation: where ET o is the reference evapotranspiration calculated by the Penman-Monteith model (Allen et al. 1998) using the meteorological data recorded by a weather station located within 1 km of the experimental site. Crop coefficient (K c ) ranged from 0.15 to 0.80 depending on the phenological stage calculated for young vines (Williams et al. 2003). With a vine spacing of 1.5 m 2 (3.0 m x 0.5 m) and a shaded area of m 2 per vine (corresponding to the top surface of the container), the percentage of shaded area was 13%. This produced a reduction coefficient K r of 0.32 (Fereres et al. 1981). A correction factor of 1.2 was applied, which considered that evapotranspiration is often higher for potted plants, perhaps because of substrate heating (Burger et al. 1987). Irrigation treatments. Ten irrigation treatments were applied (one control and nine DI treatments). To assign these, the growing season was divided into three phenological stages: stage I (anthesis to full fruit set), stage II (full fruit set to 60% of veraison), and stage III (end of stage II to harvest). These stages did not correspond to the classical three stages describing the double sigmoid growth curve of berries. Four irrigation levels were applied as percentages of calculated ET c : 0%, 25%, 50%, and 100% of ET c. Vines of the control treatment were irrigated to replace 100% ET c during stages I, II, and III (100%-SI-II-III). The nine DI treatments were 0%, 25%, and 50% of ET c each applied to one group of 16 vines in each of the three phenological stages. These vines were fully irrigated during the other two stages. The treatments were therefore: 0%, 25%, and 50% ET c applied during stage I and 100% ET c during stages II and III (hereafter called 0%-SI, 25%-SI, and 50%-SI, respectively); 0%, 25%, and 50% ET c during stage II and 100% ET c during stages I and III (0%-SII, 25%-SII, and 50%-SII, respectively); 0%, 25%, and 50% ET c during stage III and 100% ET c during stages I and II (0%- SIII, 25%-SIII, and 50%-SIII, respectively). Deficit irrigation treatments applied during stage I started on 4 June 2008 and ended on 17 June (full fruit set); those during stage II ended on 24 July (60% veraison completed); and those during stage III were applied until harvest (29 August). Supplementary irrigation was given to vines of the 0% ET c treatment to prevent midday Ψ leaf dropping below the threshold of -1.8 MPa. The total amounts of supplementary water given to 0%-SI, 0%-SII, and 0%-SIII vines were 5.1%, 12.1%, and 14.1% of ET c during stages I, II, and III, respectively. From budbreak to the beginning of stage I and from harvest to leaf fall, all vines received 100% of ET c. Each container had two emitters (3 L/hr each) and vines were irrigated daily. The amount of water applied to each experimental unit was measured with digital water meters (CZ2000-3M, Contazara, Zaragoza, Spain). The electrical conductivity of irrigation water was 0.4 ds/m. Experimental design. The experiment had a complete randomized block design with 10 treatments and four replications. Each experimental unit consisted of four vines, for a total of 160 vines (4 vines x 10 treatments x 4 replications). The potted vines were aligned in eight rows (each replication consisted of two rows), and two additional rows of similar potted vines were considered as borders. Two border vines were also placed at both the beginning and end of each row. Meteorological data. Air temperature, relative humidity, solar radiation, rainfall, wind speed, and wind direction were measured hourly in a weather station located within 1 km of the experimental site. These data were also used to calculate ET o. Vapor pressure deficit (VPD) was calculated following an established method (Buck 1981). Effective rainfall was estimated as half of the rainfall for a single event-day with more than 10 mm precipitation. Vine water status. Midday Ψ leaf was measured weekly, at solar noon from 29 May to 27 Aug, selecting one fully mature and sun-exposed leaf per experimental unit. These leaves were normally found in the median part of the shoot. A pressure chamber was used (3005 series portable plant water status console; Soil Moisture Equipment Corp., Santa Barbara, CA) following the recommendations of Turner and Long (1980). Additional measurements were carried out twice weekly on the most extreme treatments (control and 0% ET c ) to identify the need for supplementary irrigation. To consider both severity and duration of water stress, midday Ψ leaf was integrated over each phenological stage.

3 454 Basile et al. These calculations were performed using only the data from the dates when midday Ψ leaf was measured in all the treatments (one measurement per week throughout the experiment). The additional measurements carried out weekly only on the control, and the 0% treatments were not used for this purpose. Integrated Ψ leaf was calculated according to Myers (1988) with some modifications. The values between two successive measuring dates were first calculated according to the following equation: where IntΨ i, i+1 is the integrated Ψ leaf calculated between date t i and date t i+1 ; t i and t i+1 are two consecutive measuring dates expressed as days of year; and Ψ i and Ψ i+1 are midday Ψ leaf measured on dates t i and t i+1, respectively. These values were used to calculate integrated Ψ leaf for each phenological stage: weight (SLW) values, which were calculated by collecting a random leaf sample in each experimental unit and measuring its area and dry weight. Leaf areas were measured with a leaf area meter (model 3200; LI-COR, Lincoln, NE). At the end of stages I and II, these samples consisted of 12 leaves per experimental unit, whereas at the end of stage III, 50 leaves were collected per experimental unit. Leaf samples collected at the end of the first two stages were smaller to avoid inducing artificially a significant reduction in vine leaf area. Total area of healthy leaves at the end of stages I and II was estimated by counting the number of healthy leaves per experimental unit and multiplying this value by mean leaf size calculated with the 12-leaf samples. To estimate total area of healthy leaves at the end of stage III, all the leaves were collected from each experimental unit and their dry weights were converted into leaf areas using the SLW values previously calculated with the leaf samples. Percent abscised or chlorotic leaves were calculated as: Here, IntΨSx is the total integrated Ψ leaf accumulated during phenological stage x and i represents the number (n) of days when Ψ leaf was measured during the same phenological stage. The units of both IntΨ i, i+1 and IntΨSx are MPa day. As previously described (Girona et al. 2009), the weighted average of Ψ leaf was calculated for each phenological stage: where ΨSx is the weighted average of Ψ leaf accumulated during phenological stage x and l represents the length (in days) of the same phenological phase. The unit of ΨSx is MPa. Net CO 2 exchange rate. Net CO 2 exchange rate was measured at midday on three leaves per experimental unit on 16 July (stage II) and 27 Aug (stage III). Gas exchange measurements were conducted on the most recently matured leaves and were taken at midday to coincide with measurements of Ψ leaf. Measurements of 16 July were on control and DI vines in stage II (0%-SII, 25%-SII, and 50%-SII). Measurements of 27 Aug were on control and DI vines in stage III (0%-SIII, 25%-SIII, and 50%-SIII). Net CO 2 exchange rate was measured with a portable gas exchange analyzer (model ADC-LCi; Analytical Development Co., Hoddesdon, UK). Ψ leaf was also measured at the same time. All measurements were carried out in less than one hour. Midday air temperature and VPD were 29.3 C and 2.7 kpa, respectively, on 16 July and were 30.8 C and 2.6 kpa, respectively, on 27 July. Leaf area, leaf senescence, and abscission. Total areas of healthy leaves and of abscised or chlorotic leaves were estimated at the end of the three phenological stages. Leaves abscised from each experimental unit (four vines) were collected every 3 to 4 days and their dry weight measured. At the end of the three phenological stages, chlorotic leaves were collected from each experimental unit and their dry weight was measured. Dry weights of abscised and chlorotic leaves were converted into leaf areas by means of specific leaf where %LA a+c is the percent of abscised or chlorotic leaf area; LA a+c is the area of abscised or chlorotic leaves per experimental unit; and LA h is the area of healthy leaves per experimental unit. LA h and LA a+c at the end of stage I and II were estimated only in control vines and in the treatments that were exposed to DI in that specific phenological stage (i.e., 0%-SI, 25%-SI, and 50%-SI at the end of stage I, and 0%-SII, 25%-SII, and 50%-SII at the end of stage II). These estimates were carried out on all the experimental units at the end of stage III. Fruit growth. Samples of eight berries per experimental unit (2 berries/vine) were collected weekly from 18 June to 27 Aug 2008, and fresh and dry weights measured. The latter were measured after berries were dried to constant weight in a ventilated oven (set at 60 C to avoid sugar caramelization). Percentage of berry dry matter (DM%) was calculated as the ratio of dry to fresh weight ( 100). Harvest. Soluble solids concentration (SSC) in berry was used as the basis for harvest. Fruit were harvested on 29 Aug when all treatments were ~23 Brix. By this time, it appeared that SSC had leveled off in all treatments. To measure SSC, 8-berry samples were collected on four dates (2 to 4 days apart) starting on 21 Aug until harvest from each experimental unit (2 berries/vine). Berry juice was extracted from each sample and SSC was measured with a refractometer (Palette PR-32α, ATAGO Co., Tokyo, Japan). At harvest, fruit yield per vine was measured, weighing separately the clusters from each vine. The number of berries per vine (crop load) was estimated by dividing fruit yield per vine by the mean berry fresh weight (measured separately for each treatment as previously described). Physiological crop load was defined for each experimental unit as the ratio of number of berries to total leaf area. Fruit from the four vines of each experimental unit were combined and then split into two subsamples. The first subsample was immediately frozen

4 Responses of Cabernet Sauvignon to Water Stress 455 and used for anthocyanin and polyphenol analyses. The second subsample was used to measure titratable acidity (TA). To measure TA, 10 ml filtered juice was diluted with 10 ml distilled water and titrated with 0.1 N NaOH to an endpoint ph of 8.2. Titratable acidity was expressed as grams tartaric acid per milliliter juice. To measure anthocyanin and polyphenol concentrations in the juice, two 30-berry samples were collected and separately crushed. A blender (T25 Basic Ultra-Turrax; IKA Werke GmbH & Co. KG, Staufen, Germany) was used to homogenize the sample at 24,000 rpm. A homogenized sample (1 g) was then added to 10 ml ethanol (ph 2) and shaken for 1 hr with a rotator set at 10 rpm. The samples were then centrifuged for 20 min at 1450 g and a 0.5 ml sample of supernatant was added to 10 ml 1 M HCl. After 3 hr, the absorbances of the samples at 520 nm (anthocyanins) and at 280 nm (polyphenols) were measured with a spectrophotometer (T60; PG Instruments Ltd., Wibtoft, UK) using 1 M HCl as blank. Total anthocyanin and polyphenol concentrations (expressed as mg/g sample fresh weight) were calculated (Iland et al. 2004). Statistical analyses. All relationships between the different berry composition parameters and weighted average of Ψ leaf, between percent abscised or chlorotic leaf area and weighted average of Ψ leaf, and between net CO 2 exchange rate and midday Ψ leaf were assessed by regression analyses, considering the experimental unit values (each consisting of four vines) as elemental data. This yielded a maximum of four points per treatment. Curve fitting was by the simplest model that allowed a random distribution of residuals. The significance of the differences between treatments in mean Ψ leaf was calculated for each phenological stage; in leaf area, number of clusters, and number of berries per vine; and in physiological crop load was assessed by one-way ANOVA using the Duncan test (p 0.05) as a post-hoc test for separation of means. All regression analyses and ANOVA analyses were performed with SAS statistical software (ver. 9.2; SAS Institute, Cary, NC). Results Weather conditions. During the experiment (4 June to 29 Aug 2008), ET o and total rainfall values were 405 mm and 71 mm, respectively (Figure 1). Total ET o was 58, 186, and 161 mm during stages I, II, and III, respectively. Only two events of effective rainfall occurred during the experiment. The first (7.9 mm) occurred on 16 June (stage I) and the second (12.4 mm) occurred on 14 Aug (stage III). Both rainfall events were effectively forecasted by the local weather service, which allowed for timely installation of plastic covers on the containers to keep out the rain. During the three phenological stages, midday air temperature ranged from 21 to 37 C and midday VPD ranged from 1.1 to 4.7 kpa. Water applied and vine water status. Control vines received 29, 108, and 121 L water/vine during stages I, II, and III, respectively, a total of 258 L/vine (Table 1). Vines in the 0%-SII and 0%-SIII treatments received ~160 L/vine and those in the 0%-SI treatment were irrigated with 93% of the water given to control. Irrigation treatments significantly affected midday Ψ leaf (Figure 2, Table 1). For each of the three stages, deficit irrigated vines had significantly lower Ψ leaf than control vines. The 0% treatment had the lowest Ψ leaf values and those for 25% and 50% treatments were intermediate between control and 0% treatment. Weighted averages of midday Ψ leaf (MPa) ranged from to in stage I, from to in stage II, and from to in stage III. Net CO 2 exchange rate and leaf senescence and abscission. As midday Ψ leaf decreased, so did net CO 2 exchange Figure 1 Seasonal pattern of (A) rainfall and (B) midday air temperature, reference evapotranspiration (ET o ), and midday vapor pressure deficit (VPD) measured daily from 4 June to 29 Aug Figure 2 Seasonal pattern of average midday leaf water potential (Ψ leaf ) in Cabernet Sauvignon vines exposed to 100%-SI-II-III, 0%-SI, 0%-SII, and 0%-SIII irrigation treatments in SI, SII, and SIII indicate phenological stage I (anthesis to fruit set), stage II (fruit set to veraison), and stage III (veraison to fruit harvest), respectively. For greater clarity, other treatments were not included. Vertical bars represent least significant differences (LSD) (p 0.05) between irrigation treatments for the midday Ψ leaf data.

5 456 Basile et al. Treatment Table 1 Volume of applied irrigation water and average midday Ψ leaf values for stages I, II, and III in Cabernet Sauvignon vines exposed to 10 different irrigation treatments. Water applied (L/vine) Midday Ψ leaf (MPa) Stage I Stage II Stage III Stage I Stage II Stage III 100%-SI-II-III a a a a 0%-SI c b a 25%-SI b ab a 50%-SI a ab a 0%-SII a d a 25%-SII a c a 50%-SII a b a 0%-SIII a ab c 25%-SIII a ab b 50%-SIII a ab a a Within columns, means followed by different letters are significantly different according to the Duncan test (p 0.05). Figure 3 Relationship between abscised or chlorotic leaf area as the percentage of total vine leaf area at the end of each phenological stage and weighted average of Ψ leaf calculated for the three stages. rate (data not shown). These relationships were linear during stage II and exponential during stage III. Irrespective of the phenological stage, the percentage of total vine leaf area that abscised or became chlorotic significantly increased in an exponential relation with decreasing weighted average of midday Ψ leaf (Figure 3). The relationships between these two parameters calculated for stage I and stage III were very similar for values of weighted average of Ψ leaf greater than MPa, whereas below this value the relationship appeared to be steeper for stage I than stage III. The curve calculated for stage II appeared to be shifted down compared to the other two phenological stages. Leaf abscission induced by the irrigation treatments caused significant differences between treatments in total leaf area at harvest (Table 2). Leaf area at harvest of vines exposed to water stress during stage I was not significantly different from control vines. Similar results were obtained with 50%-SIII vines. Vines of the 0%- SII, 25%-SII, and 0%-SIII had the lowest leaf area at harvest. Crop load and physiological crop load. The number of berries per vine was not affected significantly by deficit irrigation compared to the control (Table 2). Physiological crop Treatment Table 2 Leaf area, number of clusters per vine, number of berries per vine, physiological crop load (number of berries per unit leaf area) at harvest, and fruit yield per vine for Cabernet Sauvignon vines exposed to 10 different irrigation treatments. Leaf area/ vine (m 2 /vine) Clusters/vine (n) Berries/vine (n) Crop load (berries/m 2 leaf area) Fruit yield (g/vine) 100%-SI-II-III 0.81 a a 8.0 a ab de ab 0%-SI 0.73 a 7.7 a b e bc 25%-SI 0.72 a 7.4 a ab e bc 50%-SI 0.74 a 7.2 a ab de abc 0%-SII 0.24 c 6.6 a ab a abc 25%-SII 0.35 c 7.1 a b b bc 50%-SII 0.53 b 8.2 a ab bcd a 0%-SIII 0.32 c 7.3 a b b 91.4 c 25%-SIII 0.57 b 8.1 a a bc ab 50%-SIII 0.86 a 7.9 a a cde ab a Within columns, means followed by different letters are significantly different according to the Duncan test (p 0.05).

6 Responses of Cabernet Sauvignon to Water Stress 457 load was significantly increased by the 0% and 25% DI treatments applied in stages II and III because leaf area decreased with increasing water stress. Berry growth, yield, and composition. Berry dry weight was affected by deficit irrigation (Figure 4). It was significantly higher in control than in 0%-SII and 0%-SIII vines, whereas the 0%-SI treatment had an intermediate fruit dry weight. Because there is more berry dry weight accumulation in stage III than the previous stages, as seen for the control treatment, 0%-SIII in stage III reduced dry matter accumulation to a greater degree than did 0%-SII in stage II. Berry dry weight increased linearly with weighted average of midday Ψ leaf in stages II and III (Figure 5), but not in stage I (data not shown). The greatest decrease in yield, compared to control, was for the 0% treatments, and significantly so for the 0%- SIII treatment (Table 2). All berry composition parameters measured at harvest were significantly affected by the irrigation treatments. For each of the three phenological stages, there was a significant curvilinear relationship between berry SSC and weighted average Ψ leaf (Figure 6A). The maximum values of SSC, as read on the fitted curve, occurred at a weighted average Ψ leaf (MPa) of for stage I, for stage II, and for stage III. Figure 4 Seasonal patterns of berry dry weight for Cabernet Sauvignon vines exposed to 100%-SI-II-III, 0%-SI, 0%-SII, and 0%-SIII irrigation treatments in For greater clarity, the other treatments were not included. The vertical bars represent the least significant differences (LSD) (p 0.05) between irrigation treatments. Figure 5 Relationship between weighted average of Ψ leaf for phenological stages II and III and berry dry weight at harvest. Circles (, ) represent 100% S-II and 100% S-III vines (100% of ET c during stages II and III), triangles (, ), diamonds (, ), and squares (, ) represent vines irrigated at 0%, 25%, and 50% of ET c during stage II (empty symbols) and stage III (solid symbols), respectively. Each point represents an experimental unit with four vines. The relationship for stage I was not significant. Figure 6 Relationship between weighted average of Ψ leaf and berry soluble solids concentration for phenological stages I, II, and III (A) and between weighted average of Ψ leaf and titratable acidity for stage II (B). Circles (,, ) represent 100%-SI-II-III vines (100% of ET c during stages I, II, and III), triangles (,, ), diamonds (,, ), and squares (,, ) represent vines irrigated at 0%, 25%, and 50% of ET c during stage I (grey solid symbols), stage II (empty symbols), and stage III (black solid symbols), respectively. Each point represents an experimental unit, with four vines. For stages I and III the relationships were not significant for inclusion in panel B.

7 458 Basile et al. Titratable acidity decreased in an inverse relationship with weighted average of Ψ leaf during stage II (Figure 6B). There were no significant relationships between the two parameters during stages I and III (data not shown). Berry polyphenol and anthocyanin concentrations decreased curvilinearly with decreasing weighted average of Ψ leaf during stage I, whereas they both increased linearly with decreasing weighted average of Ψ leaf during stage III (Figure 7). Anthocyanins increased with increasing water stress during stage II until a weighted average Ψ leaf value of MPa was reached (Figure 7B). The response leveled off at more negative Ψ leaf values. The relationship between polyphenols and weighted average of Ψ leaf during stage II was not significant (data not shown). Figure 7 Relationship between berry polyphenols concentration and weighted average of Ψ leaf for stages I and III (A) and between berry anthocyanin concentration and weighted average of Ψ leaf for stages I, II, and III (B). Circles (,, ) represent 100%-SI-II-III vines (100% of ET c during stage I, II, and III), triangles (,, ), diamonds (,, ), and squares (,, ) represent vines irrigated at 0%, 25%, and 50% of ET c during stage I (grey solid symbols), stage II (empty symbols), and stage III (black solid symbols), respectively. Each point represents an experimental unit, with four vines. For stage II the relationship was not significant for inclusion in panel A. Discussion The response of berry size (in terms of dry weight) and composition of Cabernet Sauvignon to water stress was strongly dependent on phenological stage and stress severity (Figures 4, 5, 6, and 7). In addition, the nature and the sensitivity of these responses strongly changed with the measured parameter. Different parameters will therefore be discussed separately according to the nature of their responses. For the three phenological stages, SSC and water status were related in a bell-shaped quadratic relationship (Figure 6A). Water stress up to certain thresholds (weighted average of Ψ leaf of MPa for stage I and for stages II and III) induced an increase in SSC at harvest, whereas more severe water stress (weighted average of Ψ leaf < MPa for stage I and Ψ leaf < MPa for stages II and III) caused an opposite effect. For Tempranillo berries, a similar curvilinear response of SSC to postveraison water status was reported (Girona et al. 2009) and the authors defined a threshold (weighted average of Ψ leaf = MPa) that was similar to the one found here for Cabernet Sauvignon. For Merlot grapevines, there was a relationship between berry sugar concentration at harvest and minimum seasonal stem water potential with a shape similar to that described in the present study (Van Leeuwen et al. 2009). In Cabernet Sauvignon, there were slight decreases in SSC under severe water stress conditions (mean postveraison Ψ leaf = MPa) (Chapman et al. 2005). On the other hand, the curvilinear relationship describing the effect of water stress during stage II on SSC (Figure 6A) is different from the linear negative relationship reported elsewhere for Tempranillo in stage II (Girona et al. 2009). To compare our results with those of Girona et al. (2009) for the same values of plant water status, it is necessary to convert from weighted Ψ leaf to integrated Ψ leaf. To do so, Equation 4 can be used by multiplying weighted Ψ leaf by the length of the time in stage II. Thus, the Tempranillo cultivar was exposed during stage II to similar water stress (integrated Ψ leaf values between 25.0 and 46.6 MPa day) (Girona et al. 2009) as Cabernet Sauvignon in our experiment (integrated Ψ leaf values between 28.9 and 49.9 MPa day). We therefore suggest that the different responses of SSC to water stress could have been due to cultivar differences. The part of the curvilinear response where SSC is positively affected by postveraison water stress supports the widely accepted hypothesis that water deficit applied between veraison and harvest can have a dehydration-concentration effect on berry solutes (Williams and Matthews 1990). However, this cannot explain the positive effect of mild-to-medium water stress applied from anthesis to fruit set (weighted average of Ψ leaf equal to MPa) or between fruit set and veraison (weighted average of Ψ leaf equal to -1.07) on SSC, because berries of vines exposed to deficit irrigation only during these two phenological stages were fully rehydrated by harvest. Water stress between anthesis and early stages of berry growth was reported to cause a decrease in berry number in Cabernet franc (Hardie and Considine 1976). The decrease in crop load may lower berry-to-berry competition for carbohydrates, which may have positive effect on soluble solids accumulation in

8 Responses of Cabernet Sauvignon to Water Stress 459 fruit. However, our mild-to-medium DI treatments applied in stages I and II either did not affect the physiological crop load or increased it in one case (25%-SII; Table 2). The low-to-mild water stress applied during stages I and II might have caused better exposure of leaves in the proximity of berry clusters to light, which could have resulted in the observed increase in berry SSC (Figure 6A). However, we do not have data on bunch microclimate to substantiate this hypothesis. The part of the curvilinear response where SSC was progressively decreased by increasing water stress (Figure 6A) may be related to the occurrence of carbon limitation. Water stress decreasing weighted average of Ψ leaf to less than MPa for stages II and III induced large decreases in net CO 2 exchange rate (data not shown) and green leaf area (Figure 3). These decreases in leaf area by harvest caused significant increases in physiological crop load (Table 2), and high physiological crop load has been associated with source-sink relationship conditions that are unfavorable for berry growth and composition (Jackson and Lombard 1993, Kliewer and Dokoozlian 2005). Interestingly, even though leaf abscission appeared to be less sensitive to water stress applied in stage II than in stages I and III (Figure 3), total leaf areas at harvest of 0%-SII and 25%-SII vines were very small and similar to that of 0%-SIII, suggesting that vegetative growth of vines exposed to water stress during stage II did not recover during stage III. During postveraison, shoot growth is strongly inhibited by shoot-to-berry competition for carbohydrates (Williams and Matthews 1990). On the other hand, vines exposed to different levels of water stress during stage I (0%- SI, 25%-SI, and 50%-SI) fully recovered their leaf area by harvest (Table 2), which may be due to the vegetative growth occurring during stage II. The lower berry SSC in 0%-SI than 0%-SII and 0%-SIII (Figure 6) could be due to the dilution effect arising from the higher water content in 0%-SI. Values of berry water content (%) for 0%-SI, 0%-SII, and 0%-SIII were 74.3, 73.4, and 72.5, respectively. Juice polyphenol and anthocyanin concentrations progressively decreased in a negative quadratic relationship with increasing water stress applied between anthesis and fruit set (Figure 7). These decreases could be a result of a dilution effect, as explained above for the decrease in berry SSC in stage I. But the decreases were linearly and positively correlated with postveraison water stress, in agreement with previous reports indicating that these attributes improve with water stress (Ginestar et al. 1998, Ojeda et al. 2002, Castellarin et al. 2007a, Romero et al. 2010). The positive effect of deficit irrigation on berry anthocyanin concentration may be due to both indirect and/or direct effects of water stress. The indirect effects may be related to the decrease in berry size, leading to a high skin-topulp ratio that is thought to improve anthocyanin extraction from berry skin (Williams and Matthews 1990). In the present study, postveraison DI significantly decreased dry matter accumulation in the berry compared to the control (Figure 4, Figure 5). However, other results have supported the hypothesis that water stress induces an increase in skin anthocyanin concentration independent of berry size (Ojeda et al. 2002, Roby and Matthews 2004, Roby et al. 2004), which occurs by stimulating biosynthesis (Castellarin et al. 2007a, 2007b). Irrespective of the mechanism, the response of juice anthocyanin concentration of Cabernet Sauvignon to postveraison water stress was different from that previously described for Tempranillo (Girona et al. 2009). For Tempranillo, juice anthocyanin concentration was positively affected only for mild-to-medium water stress conditions during postveraison, whereas severe water stress caused a decrease in this parameter. In the present study, juice anthocyanin concentration increased over a wide range of water stress conditions (Figure 7B). In this respect, Cabernet Sauvignon seems better adapted to water stress than Tempranillo. Water stress during stage II also positively affected juice anthocyanin concentration (Figure 7B), confirming previous reports (Matthews and Anderson 1988, Ojeda et al. 2002). However, in the present study the response of juice anthocyanin concentration to water stress during stage II was different than the response during postveraison (Figure 7B). In particular, water stress from fruit set to veraison positively affected juice anthocyanin only until average values of weighted Ψ leaf of approximately -1.0 MPa were reached. Below this threshold, anthocyanin concentration appeared to level off. For this reason, water stress inducing a weighted average of Ψ leaf less than -1.0 MPa during stage II was less efficient in increasing juice anthocyanin concentration than similar levels of water stress applied during postveraison. The higher concentration of berry anthocyanins for 0%-SIII than for 0%-SII could be explained by a dilution effect. Average berry fresh weight was 0.55 g for 0%-SIII and 0.68 g for 0%-SII. Titratable acidity was curvilinearly and negatively correlated with water stress during stage II, with lower TA values at lower weighted average of midday Ψ leaf. Water stress can induce a decrease in TA for Cabernet Sauvignon (Bindon et al. 2008), Tempranillo (Esteban et al. 1999), and Merlot (Van Leeuwen et al. 2009). This decrease may be explained by the fact that berry clusters of water-stressed vines are generally more exposed to sunlight (Romero et al. 2010), as leaves in the productive layer of the canopy either become burnt or drop (Ginestar et al. 1998), thus causing an increase in berry temperature (Santos et al. 2007) and therefore in berry respiration (Williams et al. 1994). There would be higher losses of malic acid at higher respiration rates (Williams et al. 1994), leading to lower TA. In our experiment, most of the burnt and/or abscised leaves were located in the bottom part of the canopy, where berry clusters were located. The responses of plants to water stress are generally very complex because different physiological processes may be affected depending on the intensity and the duration of water shortage. Mild water stress affects mainly cell turgor, whereas as water stress increases, other physiological processes such as whole-plant photosynthesis and solute transport are negatively affected (Kramer and Boyer 1995). Our results support the hypothesis that mild-to-medium water stress conditions may affect grape berry composition at harvest by mainly decreasing fruit size and concentrating solutes in

9 460 Basile et al. the pulp. However, when water stress becomes severe, the occurrence of carbon limitations may play a major role in the response of berry composition to water stress. This experiment was carried out in potted vines, so extrapolations to field-grown vines must be made with due caution. In addition, comparison of the results obtained here for Cabernet Sauvignon with those reported by other researchers for different cultivars suggests that extrapolations to other cultivars should be tempered with caution. Conclusion Deficit irrigation treatments led to development of moderate water stress in stage I and moderate to severe stress in stages II and III. Harvest was conducted based on berry soluble solids concentrations and all treatments had values higher than the required minimum of 23 Brix. Berry quality was therefore mainly determined by anthocyanin and polyphenol concentrations. Applying water stress in stage I was disadvantageous. Yield decreased as did berry soluble solids, anthocyanins, and polyphenols. There is therefore no justification for application of deficit irrigation in stage I. Anthocyanin and polyphenol concentrations increased with postveraison water stress (stage III). Similarly, anthocyanin concentration increased with water stress applied between fruit set and veraison (stage II). However, this positive effect leveled off for weighted average of Ψ leaf less than -1.0 MPa. Responses of Cabernet Sauvignon to water stress, in terms of fruit growth and composition, therefore depended on the severity of stress and the phenological stage at which it occurred. Literature Cited Allen, R.G., L.S. Pereira, D. Raes, and M. Smith Crop Evapotranspiration. Guidelines for Computing Crop Water Requirements. FAO irrigation and drainage paper no. 56. FAO, Rome. Bindon, K., P. Dry, and B. Loveys Influence of partial rootzone drying on the composition and accumulation of anthocyanins in grape berries (Vitis vinifera cv. Cabernet Sauvignon). Aust. J. Grape Wine Res. 14: Bravdo, B., Y. Hepner, C. Loinger, S. Cohen, and H. Tabacman Effect of irrigation and crop level on growth, yield and wine quality of Cabernet Sauvignon. Am. J. Enol. Vitic. 36: Buck, A.L New equations for computing vapor pressure and enhancement factor. J. Appl. Meteorol. 20: Burger, D.W., J.S. Hartin, D.R. Hodel, T.A. Lukaszewski, S.A. Tjosvold, and S.A. Wagner Water use in California s ornamental nurseries. Calif. Agric. 41:7-8. Castellarin, S.D., M.A. Matthews, G. Di Gaspero, and G.A. Gambetta. 2007a. Water deficits accelerate ripening and induce changes in gene expression regulating flavonoid biosynthesis in grape berries. Planta 227: Castellarin, S.D., A. Pfeiffer, P. Sivilotti, M. Degan, E. Peterlunger, and G. Di Gaspero. 2007b. Transcriptional regulation of anthocyanin biosynthesis in ripening fruits of grapevine under seasonal water deficit. Plant Cell Environ. 30: Chapman, D.M., G. Roby, S.E. Ebeler, J.X. Guinard, and M.A. Matthews Sensory attributes of Cabernet Sauvignon wines made from vines with different water status. Aust. J. Grape Wine Res. 11: Esteban, M.A., M.J. Villanueva, and J.R. Lissarrague Effect of irrigation on changes in berry composition of Tempranillo during maturation. Sugars, organic acids, and mineral elements. Am. J. Enol. Vitic. 50: Fereres, E., W.O. Pruitt, J.A. Beutel, D.W. Hemerson, E. Holzapfel, H. Schulbach, and K. Uriu Evapotranspiration and drip irrigation scheduling. In Drip Irrigation Management. E. Fereres (ed.), pp Division of Agricultural Sciences, University of California, Davis. Ferreyra, R.E., G.V. Selles, J.A. Peralta, and J.B. Valenzuela Effect of water stress applied at different development periods of Cabernet Sauvignon grapevine on production and wine quality. Acta Hortic. 646: Ginestar, C., J. Eastham, S. Gray, and P. Iland Use of sap-flow sensors to schedule vineyard irrigation. II. Effects of post-veraison water deficits on composition of Shiraz grapes. Am. J. Enol. Vitic. 49: Girona, J., J. Marsal, M. Mata, J. del Campo, and B. Basile Phenological sensitivity of berry growth and quality of Tempranillo grapevines (Vitis vinifera L.) to water stress. Aust. J. Grape Wine Res. 15: Hardie, W.J., and J.A. Considine Response of grapes to waterdeficit stress in particular stages of development. Am. J. Enol. Vitic. 27: Iland, P., N. Bruer, G. Edward, S. Weeks, and E. Wilkes Anthocyanins (colour) and total phenolics of grape berries. In Chemical Analysis of Grapes and Wine: Techniques and Concepts, pp Winetitles, Goodwood, Australia. Jackson, D.I., and P.B. Lombard Environmental and management practices affecting grape composition and wine quality A review. Am. J. Enol. Vitic. 44: Keller, M., R.P. Smithyman, and L.J. Mills Interactive effects of deficit irrigation and crop load on Cabernet Sauvignon in an arid climate. Am. J. Enol. Vitic. 59: Kliewer, W.M., and N.K. Dokoozlian Leaf area/crop weight ratios of grapevines: Influence on fruit composition and wine quality. Am. J. Enol. Vitic. 56: Kramer, P.J., and J.S. Boyer Water Relations of Plants and Soils. Academic Press, San Diego. Matthews, M.A., and M.M. Anderson Fruit ripening in Vitis vinifera L.: Responses to seasonal water deficits. Am. J. Enol. Vitic. 39: Matthews, M.A., and M.M. Anderson Reproductive development in grape (Vitis vinifera L.): Responses to seasonal water deficits. Am. J. Enol. Vitic. 40: Myers, B.J Water stress integral A link between short-term stress and long-term growth. Tree Physiol. 4: Ojeda, H., C. Andary, E. Kraeva, A. Carbonneau, and A. Deloire Influence of pre- and postveraison water deficit on synthesis and concentration of skin phenolic compounds during berry growth of Vitis vinifera cv. Shiraz. Am. J. Enol. Vitic. 53: Roby, G., and M.A. Matthews Relative proportions of seed, skin and flesh, in ripe berries from Cabernet Sauvignon grapevines grown in a vineyard either well irrigated or under water deficit. Aust. J. Grape Wine Res.10: Roby, G., J.F. Harbertson, D.A. Adams, and M.A. Matthews Berry size and vine water deficits as factors in winegrape composition: Anthocyanins and tannins. Aust. J. Grape Wine Res. 10: Romero, P., J.I. Fernández-Fernández, and A. Martinez-Cutillas Physiological thresholds for efficient regulated deficit-irrigation management in winegrapes grown under semiarid conditions. Am. J. Enol. Vitic. 61:

10 Responses of Cabernet Sauvignon to Water Stress 461 Santos, T.P., C.M. Lopes, M.L. Rodrigues, C.R. Souza, J.R. Silva, J.P. Maroco, J.S. Pereira, and M.M. Chaves Effects of deficit irrigation strategies on cluster microclimate for improving fruit composition of Moscatel field-grown grapevines. Sci. Hortic. 112: Turner, N.C., and M.J. Long Errors arising from rapid water loss in the measurement of leaf water potential by the pressure chamber technique. Aust. J. Plant Physiol. 7: Van Leeuwen, C., O. Tregoat, X. Choné, B. Bois, D. Pernet, and J.P. Gaudillère Vine water status is a key factor in grape ripening and vintage quality for red Bordeaux wine. How can it be assessed for vineyard management purposes? J. Int. Sci. Vigne Vin. 43: Williams, L.E., and M.A. Matthews Grapevine. In Irrigation of Agricultural Crops. B.A. Steward and D.R. Nielsen (ed.), pp Agronomy monograph no. 30. ASA-CSSA-SSSA, Madison, WI. Williams, L.E., N.K. Dokoozlian, and R. Wample Grape. In Handbook of Environmental Physiology of Fruit Crops. Vol. 1. Temperate Crops. B. Schafer and P.C. Andersen (eds.), pp CRC Press, Boca Raton, FL. Williams, L.E., C.J. Phene, D.W. Grimes, and T.J. Trout Water use of young Thompson seedless grapevines in California. Irri. Sci. 22:1-9.