Effect of winter rainfall on yield components and fruit green aromas of Vitis vinifera L. cv. Merlot in California
Abstract
Background and Aims
A field trial during the 2009 and 2010 seasons evaluated the impact of winter rainfall on the main compounds responsible for green aromas in grapes and wines, 3-isobutyl-2-methoxypyrazine (IBMP) and C6 compounds. These compounds are considered undesirable in grapes and wines above the threshold concentration.
Methods and Results
One treatment subjected vines to average rainfall, while the other excluded winter rainfall by covering the ground with a plastic tarpaulin during the entire dormant season (November to mid-March). Irrigation for both treatments was maintained at a weekly rate of 70% of crop evapotranspiration until commercial harvest. Canopy growth, berry size and vine yield were significantly reduced by rainfall exclusion, and a significant increase in the fruit to pruning mass ratio was recorded from one season to another. Synthesis of IBMP was significantly greater in vines under normal rainfall, whereas C6 compounds were significantly different between treatments only at the end of the second season. Fruit and wine composition, mainly colour and mouthfeel compounds, were positively affected by the absence of rainfall in both years. Wine descriptive analysis showed that the lack of rainfall produced wines perceived as less green and of more intense fruit attributes in the first season. As a consequence of the reduction in vine growth, however, the same treatment produced wines less intense in fruit aromas and of inferior tannin quality in the following season.
Conclusions
These results show that the soil moisture level prior to budbreak affects both canopy growth and vine yield, even when irrigation is applied following budbreak. If the rainfall level is below normal, the positive effect on fruit and wine composition achieved through smaller berry size may be offset by a significant reduction in canopy growth, resulting in severely unbalanced vines, i.e. inadequate fruit to pruning mass ratio.
Significance of the Study
Growers aiming to minimise the level of IBMP at harvest would benefit from applying moderate deficit irrigation and nitrogen fertilisation rates and also might achieve an earlier harvest date for those vineyards where the absence of undesirable vegetal characters is considered a key harvest metric.
Introduction
Vineyards in most winegrape growing regions of California must be irrigated to achieve optimum fruit yield and quality. As the majority of California winegrape growing regions receive ample winter rainfall during the dormant season in most years, soils are usually at field capacity at budbreak, and early water deficits (i.e. between budbreak to fruitset) are rare. Vine water stress, however, is expected and even encouraged during mid and late season, i.e. fruitset to veraison and veraison to harvest. In most seasons, winter rainfall will be sufficient to achieve the desired canopy size, but there may still be years where the rainfall level is well below average, which may reduce the initial stages of canopy growth. Mild water stress reduces leaf area formation by decreasing terminal meristem activity, while carbon uptake is reduced because of leaf stomatal closure (Kliewer et al. 1983, During and Dry 1995). Stomatal closure reduces water loss, limiting the effect of water stress to a reduction in vegetative growth for mild stress. If water deficit remains, stomata are closed for a longer period of time, which leads to reduced photosynthesis (Wang et al. 2003). As shoot growth of grapevines is highly sensitive to water stress (Smart and Coombe 1983, Williams and Matthews 1990), a possible effect of lower rainfall will be increased fruit exposure. The effect of lower than normal winter rainfall on parameters, such as vine yield components, canopy size and fruit and wine composition, is not as well understood, but fruit green aroma compounds, such as 3-isobutyl-2-methoxypyrazine (IBMP), have already been shown to be highly influenced by fruit exposure and irrigation level, variables likely to be affected by a below-average rainfall season. 3-Isobutyl-2-methoxypyrazine and C6 compounds are the main contributors of green aromas in grapes and wine. While IBMP gives a characteristic herbaceous, bell pepper-like aroma (Allen et al. 1990, Hashizume and Umeda 1996, Marais and Swart 1999, Roujou de Boubee et al. 2000, Sala et al. 2002), C6 compounds exhibit leafy, grassy odours, and are usually formed when the cell structure is disrupted and exposed to oxidation (Cordonnier and Bayonove 1981, Lopez et al. 1999, Jorgensen et al. 2000, Klesk and Qian 2003, Qian and Wang 2005, Cullere et al. 2007).
Photo-degradation of IBMP in berries exposed to light has been reported (Heymann et al. 1986, Hashizume and Samuta 1999), as well as a reduced IBMP concentration in fruit because of increased fruit exposure (Allen and Lacey 1993, Marais and Swart 1999, Allen 2001, Ryona et al. 2008). The intensity of the characteristic bell pepper aroma of IBMP in wine is positively correlated to factors promoting vine growth, i.e. high vigour vines usually show lower fruit exposure and a higher IBMP concentration in the fruit compared to that of less vigorous vines (Allen 2001, Wilkinson et al. 2007). Seasonal irrigation level is also positively correlated to IBMP fruit concentration, and to the intensity of the resultant bell pepper wine aroma (Roujou de Boubee 2003, Chapman et al. 2005). The amount of IBMP at harvest may be the result of a complex interaction of factors affecting vine vigour, but nitrogen fertilisation has been shown to promote IBMP accumulation, probably because of increased fruit shading (Allen 2001).
The goal of this study was to evaluate the effect of rainfall exclusion on fruit and wine green aromas, both IBMP and C6 compounds and other variables, including canopy growth, vine yield components and fruit and wine composition indices, such as colour and mouthfeel compounds. The results of this study will improve current understanding of irrigation practices for vineyards located in regions with little winter rainfall in order to better achieve desired yield and composition targets.
Materials and methods
Vineyard location
Field trials were established in a commercial vineyard of Vitis vinifera L. Merlot located in the Central Valley of California during the 2009 and 2010 seasons. The vineyard was located 19.8 km southwest of the city of Madera (36°49′42.85′N, 120°12′55.06′W). Vines were planted on a Cajon loamy sand soil (82% sand, 15% silt, 3% clay, 5% plant available water), with a rooting depth of approximately 1.8 m. Twelve-year-old Merlot (FPS clone 3) vines grafted onto Teleki 5C rootstock (Vitis berlandieri x Vitis riparia) were planted in north–south oriented rows, spaced 2.1 m × 3.4 m (vine x row spacing – approx. 1400 vines per ha), trained to a bilateral cordon and pruned to 24 two-bud spurs per vine, on a two-wire vertical trellis system (cordon wire at 1.2 m above-ground with a cross arm at 1.5 m above-ground). The vines were not crop or shoot thinned, or subjected to leaf removal at any time during either of the two seasons.
Experimental design and treatments
Experimental treatments were established in a randomised complete block design; with each treatment replicated four times using 25-vine plots located in separate rows with four buffer rows in between. The experiment was designed to exclude rainfall during the dormant season for one set of vines (i.e. rainfall exclusion), while another set was maintained at field capacity (i.e. normal rainfall). Rainfall exclusion was achieved by covering the soil surface with a tarpaulin to prevent rainfall from reaching the soil throughout the dormant period (1 November to 15 March). Sprinklers were installed for the normal rainfall vines, and a total of 50 mm of water applied during the dormant season in order to maintain the soil at field capacity (i.e. normal conditions for the region during the dormant season). Irrigation was initiated when vines reached a midday leaf water potential of −1 MPa, and was maintained at a weekly rate of 70% of crop evapotranspiration (ETc), as is standard commercial practice. Daily water requirement was estimated with the following equation, ETc = Kc × ETo (Williams 2001), where Kc is the seasonal basal crop coefficient (Kcb), and ETo is the reference evapotranspiration obtained from a nearby weather station (within 1 km, Western Weather Group, Chico, CA, USA) using the Penman Monteith equation (Allen et al. 1998). Crop coefficients were estimated by measuring the shaded area beneath the trunk of vines, and using the equation Kc = SA% × 0.017–0.008 (Williams and Ayers 2005), where SA% is the proportion of shaded area. The shaded area was estimated weekly using a solar panel located beneath the foliage at midday. Electrical current produced by the solar panel is proportional to the amount of direct sunlight striking the panel; thus, as the solar panel surface area is increasingly shaded by the vine leaf canopy, it produces proportionally less current (Battany 2008). Since ETo data from the weather station were available weekly, irrigations were scheduled on a weekly basis to replenish the ETc value estimated for the previous week on each site. The calculated amount of irrigation was applied in equal amounts over a 5-day period each week. Overall, irrigation applied during the season corresponded to 1360 and 1990 L/vine for the vines receiving the rainfall exclusion and normal rainfall treatments, respectively, in 2009, and 1920 and 2690 L/vine in 2010 (1.9 and 2.8 ML/ha in 2009; 2.7 and 3.8 ML/ha in 2010).
Yield components
Experimental vines were harvested at the same time the vineyard was commercially harvested. Six vines per replicate and treatment (total of 24 per treatment) were manually harvested, bunch number and mass per vine were recorded, and average bunch mass calculated. One hundred berries were randomly sampled from each replicate and mean berry mass calculated. Three berries per bunch, one from the top, one from the middle and one from the bottom, were collected in zip lock bags, placed in ice coolers and transported immediately to the laboratory for counting and weighing.
Vine physiology
Measurements were taken bi-weekly around midday (between 12:30 and 1:30 pm, PST) unless specified, beginning at irrigation initiation. Leaf water potential and stomatal conductance were measured on two vines per replicate at the time treatments were imposed, at veraison, and at harvest (for a total of eight leaves per treatment on each date). Leaf water potential was measured with a 610 PMS pressure chamber (PMS Instrument Co., Albany, OR, USA) using fully exposed, mature leaves (Williams 2001). Pre-dawn leaf water potential was measured at veraison only. Leaf stomatal conductance was measured also using fully exposed, mature leaves, with a SC-1 leaf porometer (Decagon Inc., Pullman, WA, USA).
Light interception and leaf area index were recorded on six vines per replicate and treatment at veraison only (i.e. total of 24 vines per treatment). Light interception in the fruiting zone was estimated using a LP-80 ceptometer (Decagon Inc.) placed inside the fruiting zone, and facing outwards. Full sun readings (ambient) were taken simultaneously using the external PAR sensor provided by the manufacturer, and results are presented as light interception in the fruiting zone as a proportion of ambient. Leaf area index was estimated using an LP-80 ceptometer.
Soil moisture measurement
Changes in soil moisture over time were evaluated with 200SS Watermark sensors (Irrometer Inc., Riverside, CA, USA) for each replicate at three locations: under the drip hose, halfway in between rows, and halfway between the two previous locations, at 30, 60 and 120 cm below the soil surface.
Fruit analysis
For green aroma analysis, 60 berries per replicate and treatment were sampled bi-weekly starting 10 days after fruit set, and ending at harvest. Berries were placed in plastic vials and ground (fresh) using a GENO grinder model 2000 (SPEX Certiprep Ltd, Metuchen, NJ, USA), at 1400 oscillations/min for 4 min. The concentration of IBMP was determined after Chapman et al. (2004a). The concentration of C6 compounds was assessed by solid-phase microextraction coupled with gas chromatography mass spectrometry after Sanchez-Palomo et al. (2005).
Randomly selected 20-bunch samples were also collected from each replicate and treatment combination at commercial harvest. Fruit was placed in zip lock bags, and then taken in ice coolers to the laboratory for analysis. Bunch samples were analysed for soluble solids (°Brix), pH, malic acid, titratable acidity (TA). Berry moisture content was determined by near-infrared spectroscopy (Cozzolino et al. 2008). Anthocyanins and phenol-free glucose glycosides were measured as described by Iland et al. (2004). Polymeric tannins and quercetin glycosides were assessed using the methodology described by Waterhouse et al. (1999).
Weather characterisation
Weather was characterised with data from a weather station less than 1 km away (Penman Monteith equation used for reference evapotranspiration). Monthly and cumulative growing degree days were calculated from daily average temperature data obtained from the weather station, and a base temperature of 10°C. Cumulative growing degree days were calculated from 1 March to 31 October. Rainfall was determined as annual and winter rainfall.
Winemaking
Small-scale vinifications were carried out at commercial harvest and replicated four times by collecting a single composite 50-kg fruit sample from each replicate. The grapes were chilled overnight and crushed the following day using a destemmer crusher (model A15DC, Magitec, Paarl, South Africa). Sulfur dioxide (40 mg/L) was added to the grapes at the time of crushing and TA adjusted to between 0.6 and 0.7 g/100 mL without allowing pH to drop below 3.5. Commercial yeast (N96, Anchor Yeast, Industria, South Africa) was inoculated at a rate of 0.18 g/L, and the cap was plunged daily. The must was fermented at 25°C and pressed at 0°Brix in a membrane press. The wines were cold settled at 2°C for 2 weeks prior to bottling. Wines were bottled 3 months after fermentation was completed and stored at 13°C. Free SO2 was adjusted to 30 mg/L prior to bottling. Wine lots were analysed for alcohol (gas chromatograph model 6980, Agilent Technologies, Santa Clara, CA, USA), TA (autotitrator model DL22, Mettler Toledo, Columbus, OH, USA), malic acid and volatile acidity (VA) (enzymatic), pH (pH meter model 700, Oakton, Vernon Hills, IL, USA), absorbance at 420 nm and 520 nm (Iland et al. 2004), and polymeric tannins and quercetin glycosides (Waterhouse et al. 1999).
Wine sensory analysis
The wines were subjected to descriptive analysis (DA) 3 months after bottling to study the effects of the experimental treatments on not only green aromas but also on basic sensory attributes. Wines were evaluated at The University of Adelaide's sensory laboratory by a panel of 12 judges in 2009 (eight females and four males, six panellists with previous experience) and 11 judges (seven females and four males, seven panellists with previous experience) in 2010, respectively. Panellists were The University of Adelaide staff and students enrolled in postgraduate coursework aged between 20 and 60 years. At a minimum, prior to the commencement of the DA panel, panellists received training in aroma, taste and mouthfeel evaluation, ranking and aroma identification over a 7-week period, with 2-h sessions weekly. The wines were subjected to DA during weekly, 2-h sessions over 2 weeks each year. Each panellist was given 30 mL of the 12 research wines in coded, covered XL5 (ISO Standard), 215 mL tasting glasses, and instructed to individually generate descriptors for each wine that differentiated the samples, then asked to reach consensus on descriptive terms. An unstructured 15-cm line scale with indented anchor points of ‘low’ and ‘high’ intensity placed at 10 and 90% of the scale and a midline anchor was used for the training and subsequent formal evaluation sessions. The descriptive terms ultimately agreed upon included colour, aroma, flavour (defined as aroma by mouth) and mouthfeel (Tables S1 and S2). Standard aroma solutions diluted in 15 mL of a Shiraz cask wine from South Australia (Yalumba) in covered black tasting glasses were provided to the panellists during each session as aroma references. Panel performance was evaluated in the last two training sessions by having each person evaluate a subsample of the wines in triplicate. Data were analysed using SENPAQ version 4.82. (Qi Statistics Ltd, Reading, England). When no significant panellist–sample interactions were found, the panel commenced final evaluation of the samples.
Each panellist was presented with four wines in each rating session. Each wine was randomly evaluated in triplicate over the course of the DA (i.e. two sessions of 2 h each per season). Thirty-millilitre wine samples were presented in coded, covered XL5, 215 mL tasting glasses covered with Petri dishes. Distilled water and unsalted crackers were available for palate cleaning, and panellists had a 1-min break between samples and a 5-min break between each flight of four wines. Aroma reference and intensity standards were available at the beginning of each session, and panellists had free access to these outside their booths during the rating period as required.
Statistical analysis
Data from DA were subjected to statistical analysis using a mixed model two-way analysis of variance with assessors as random and samples as fixed factors, Fisher's least significant difference post-hoc test and P < 0.1 as significance threshold using SENPAQ (version 4.82, Qi Statistics). Principal component analysis was performed using XLSTAT version 3.01 (Addinsoft SARL, Paris, France).
Data other than sensory data were subjected to analysis of variance and t-tests using SPSS (Version 19, IBM, New York, NY, USA). Polynomial regression lines were used to fit the curves showing the evolution of fractional ground coverage and green aroma compounds.
Results
Regional weather
Weather was significantly different between the two seasons under study. The first year (2009) was a warmer and drier season than average, whereas 2010 was colder and wetter (Table 1). Annual and winter rainfall was lower in 2009 compared with that in 2010. Cumulative growing degree days were greater in 2010, mainly because of the temperature during spring being higher (April, May) than that in 2009.
| Variable | 2009 | 2010 | Historical average (1999–2010) |
|---|---|---|---|
| Annual rainfall (mm) | 200 | 359 | 229 |
| Winter rainfall (mm) | 144 | 190 | 147 |
| GDD April (°C) | 137 | 72 | 124 |
| GDD May (°C) | 361 | 193 | 273 |
| GDD June (°C) | 358 | 389 | 371 |
| GDD July (°C) | 487 | 467 | 464 |
| GDD August (°C) | 444 | 409 | 427 |
| GDD September (°C) | 396 | 346 | 343 |
| GDD October (°C) | 161 | 229 | 184 |
| Cumulative GDD (°C) | 2345 | 2104 | 2186 |
| Irrigation start date† | 22 May | 16 May | n/a |
| Harvest date | 24 Sep | 14 Sep | n/a |
- †Irrigation was initiated when vines reached a midday leaf water potential of −1 Mpa. GDD, growing degree days.
Vine physiology and yield components
As expected, soil moisture level at budbreak was much lower when rainfall was excluded, although this trend was reversed by the time of treatment initiation, such that vines which received winter rainfall maintained a lower soil matric potential level during the remainder of the season, especially for those soil areas located further away from the drip hose (i.e. towards the middle of the row, Figure 1).
Soil matric potential of Merlot vines with normal rainfall over winter (●) and of vines where rainfall was excluded during the winter (□) during the (a, b, c) 2009 and (d,e,f) 2010 growing seasons. The letters TI, F, V and H denote the timing of treatment initiation, flowering, start of veraison and commercial harvest, respectively. Values represent average soil matric potential for sensors installed at 60 and 120 cm deep and located either (a, d), under the drip hose, (b, e), half the distance between the drip hose and the middle of the rows and (c, f) in the middle of two vine rows. Bars correspond to +/− one standard error.
No difference was observed for midday leaf water potential at treatment initiation, although lower midday water potential, lower pre-dawn leaf water potential and lower stomatal conductance were recorded at veraison for vines that did not receive any rainfall during the winter. Reduced canopy growth resulted in greater fruit exposure in the rainfall exclusion treatment, which also had lower midday water potential at harvest, compared to the normal rainfall treatment (Table 2). The exclusion of rainfall had a significant effect on vine growth in particular, reducing shoot length, fraction of ground coverage and basal crop coefficients (Tables 2 and 3 and Figure 2). Vines stopped growing 2 weeks after flowering in 2009 but continued to grow until 5 weeks after flowering in 2010. In contrast, shoot growth continued later into the season for vines under normal rainfall conditions (Table 3).
Canopy development (ground coverage) of Merlot grapevines with normal rainfall over winter (●) and of vines where rainfall was excluded during the winter (□) in (a) 2009 and (b) 2010. *Indicates a significant difference among treatments at a given date (P ≤ 0.05). TI denotes the time when treatments were imposed.
| Variable | Timing | 2009 | 2010 | ||
|---|---|---|---|---|---|
| Winter rainfall exclusion | Normal rainfall | Winter rainfall exclusion | Normal rainfall | ||
| Light interception in the fruiting zone (% of ambient) | Veraison | 9.6 | 4.6† | 5.3 | 2.0† |
| Midday leaf water potential (MPa) | First irrigation | −0.98 | −1.00 | −1.10 | −1.10‡ |
| Veraison | −1.25 | −1.15† | −1.08 | −1.07‡ | |
| Harvest | −1.40 | −12.5† | −14.1 | −12.5† | |
| Pre-dawn leaf water potential (MPa) | Veraison | −0.23 | −0.21† | −0.21 | −0.20‡ |
| Stomatal conductance [mol/(m2●s)] | Veraison | 150 | 240† | 199 | 201‡ |
| Harvest | 200 | 311† | 268 | 402† | |
| Basal crop coefficient (kc) | Veraison | 0.42 | 0.63† | 0.53 | 0.73† |
| Leaf area/crop mass ratio (m2/kg) | Veraison | 0.65 | 0.73‡ | 0.68 | 1.08† |
- †Indicates a significant difference among treatments (P ≤ 0.05). ‡No significant difference between treatments (P ≤ 0.05).
| Shoot length (cm) | ||
|---|---|---|
| Weeks from flowering | Rainfall exclusion | Normal rainfall |
| −5 | 6 | 8‡ |
| −4 | 11 | 16† |
| −3 | 15 | 27† |
| −2 | 22 | 41† |
| −1/TI | 25 | 53† |
| 0 | 28 | 63† |
| 1 | 34 | 73† |
| 2 | 40 | 81† |
| 3 | 41 | 86† |
| 4 | 41 | 89† |
| 5 | 42 | 93† |
| 6 | 43 | 95† |
| 7 | 43 | 95† |
- †Indicates a significant difference among treatments (P ≤ 0.05). ‡No significant difference between treatments (P ≤ 0.05). TI denotes the timing of treatment initiation.
Berry growth was significantly greater when vines received normal rainfall prior to budbreak, with differences between treatments recorded as early as 2 weeks after flowering (Table 4). Vines in the rainfall exclusion treatment produced significantly smaller berries and fewer bunches in both seasons (Table 5). This effect on the number of bunches per vine during the first season has not been reported previously; we can only hypothesise that the lower level of soil moisture during the dormant season affected the final stages of inflorescence primordial differentiation. While both vine total yield and pruning mass were negatively affected by the lack of rainfall, the decrease observed for the latter was greater than for the former. As a consequence of this, the fruit to pruning mass ratio was significantly higher for the rainfall exclusion treatment in both seasons; almost twice as high as that of the normal rainfall treatment in 2010 (Table 5). Also, the leaf area to crop mass ratio decreased over time for vines receiving no winter rainfall, being significantly lower than the ratio recorded for vines receiving normal rainfall conditions in the second season (Table 2).
| 2009 | 2010 | ||||
|---|---|---|---|---|---|
| Weeks after flowering | Berry mass (g) | Weeks after flowering | Berry mass (g) | ||
| Winter rainfall exclusion | Normal rainfall | Winter rainfall exclusion | Normal rainfall | ||
| 2 | 0.15 | 0.23† | 2 | 0.11 | 0.12† |
| 4 | 0.46 | 0.62† | 4 | 0.40 | 0.44† |
| 6 | 0.60 | 0.76† | 6 | 0.57 | 0.66† |
| 8 – V | 0.75 | 0.96† | 8 – V | 0.73 | 0.90† |
| 10 | 0.97 | 1.16† | 10 | 0.95 | 1.09† |
| 12 | 1.14 | 1.42† | 12 | 1.24 | 1.38† |
| 14 | 1.22 | 1.47† | 14 | 1.22 | 1.39† |
| 16 | 1.16 | 1.45† | 16 – H | 1.19 | 1.36† |
| 18 – H | 1.15 | 1.38† | |||
- †Indicates a significant difference among treatments at a given date (P ≤ 0.05). Letters V and H denote the timing of veraison and commercial harvest, respectively.
| Variables | 2009 | 2010 | ||
|---|---|---|---|---|
| Winter rainfall exclusion | Normal rainfall | Winter rainfall exclusion | Normal rainfall | |
| Tonnes/ha | 14.6 | 23.5† | 17.3 | 23.8† |
| Vine yield (kg/vine) | 10.4 | 16.8† | 12.4 | 17.0† |
| Bunch mass (g) | 103 | 126† | 104 | 154† |
| Bunch/vine | 101 | 133† | 99 | 109† |
| Berries/bunch | 87 | 88‡ | 100 | 109 |
| Berry mass (g) | 1.15 | 1.38† | 1.19 | 1.36† |
| Rachis mass (%) | 3.5 | 4.1‡ | 5.1 | 5.3 |
| Pruning mass (kg) | 0.8 | 1.5† | 0.8 | 2.0† |
| Fruit to pruning mass ratio | 13.9 | 11.5† | 15.0 | 8.6† |
- †Indicates a significant difference among treatments (P ≤ 0.05). ‡No significant difference between treatments (P ≤ 0.05).
Fruit and wine composition
Synthesis of IBMP in the fruit was substantially reduced by the rainfall exclusion in both seasons, being two to three times lower than that of the normal rainfall. The concentration of IBMP peaked in both treatments about 1 week before veraison (Figure 3). During fruit maturation C6 compounds also declined, but a difference was not apparent between treatments in 2009. A significant difference was recorded on the last three sampling points in 2010 when fruit from vines that had not received any rainfall contained a concentration of C6 compounds significantly lower than that of fruit from vines under normal rainfall conditions (Figure 4).
Concentration of 3-isobutyl-2-methoxypyrazine in fruit of Merlot vines with normal rainfall over winter (●) and of vines where rainfall was excluded during the winter (□) during the (b,d) 2009 and (a, c) 2010 seasons where irrigation was initiated (a, b) pre-veraison and (c, d) post-veraison. Letters V and H denote the timing of veraison and of commercial harvest, respectively. *Indicates significant differences among treatments at a given date (P ≤ 0.05).
Concentration of C6 compounds in fruit of Merlot vines with normal rainfall over winter (●) and of vines where rainfall was excluded during the winter (□) from the (a) 2009 and the (b) 2010 seasons. Letters V and H denote the timing of veraison and commercial harvest, respectively. *Indicates a significant difference among treatments at a given date (P ≤ 0.05). ns = No significant difference between treatments (P ≤ 0.05).
A significant difference in fruit colour, polymeric tannins, quercetin glycosides, and malic acid was recorded at harvest in both seasons, although the difference between the treatments was much smaller in the second season (2010). Fruit from vines contained a lower concentration of both polymeric tannins and quercetin glycosides, and higher malic acid concentration compared to that of fruit from the rainfall exclusion treatment (Table 6). Fruit moisture content, fruit colour and aroma precursors were significantly different between treatments in 2010 but not in 2009. Fruit moisture content was significantly lower for the rainfall exclusion treatment in 2010, and fruit colour and aroma precursors were significantly higher in this treatment in 2009. Differences in fruit composition translated to the subsequent wines in both seasons; wines corresponding to the normal rainfall treatment had less colour and mouthfeel, but a higher concentration of IBMP and malic acid compared to that of wines made from the rainfall exclusion treatment vines (Table 7).
| Variable | 2009 | 2010 | ||
|---|---|---|---|---|
| Rainfall exclusion | Normal rainfall | Rainfall exclusion | Normal rainfall | |
| Soluble solids (°Brix) | 25.7 | 26.3‡ | 24.1 | 23.7‡ |
| Moisture content (%) | 70.7 | 70.8‡ | 71.7 | 72.9† |
| pH | 3.83 | 3.82‡ | 3.66 | 3.68‡ |
| Malic acid (mg/L) | 810 | 1078† | 1178 | 1489† |
| Titratable acidity (g/L) | 5.8 | 5.8‡ | 6.1 | 6.0‡ |
| Phenol free glucose glycosides (mg/L) | 49.1 | 42.8† | 39.8 | 37.8‡ |
| 3-Isobutyl-2-methoxypyrazine (ng/L) | 1 | 3† | 1 | 2† |
| C6 compounds (mg/L) | 16.9 | 15.8‡ | 8.0 | 9.2† |
| Anthocyanins (mg/g) | 0.81 | 0.57† | 0.85 | 0.82‡ |
| Polymeric tannins (mg/L) | 1.6 | 1.0† | 0.9 | 0.7† |
| Quercetin glycosides (μg/L) | 173 | 98† | 155 | 103† |
- †Indicates a significant difference among treatments (P ≤ 0.05). ‡No significant difference between treatments (P ≤ 0.05).
| Variable | 2009 | 2010 | ||
|---|---|---|---|---|
| Rainfall exclusion | Normal rainfall | Rainfall exclusion | Normal rainfall | |
| Alcohol (%) | 14.1 | 14.3‡ | 13.4 | 13.5‡ |
| Residual sugar (g/100 mL) | bdl | bdl‡ | bdl | bdl‡ |
| Titratable acidity (g/L) | 0.54 | 0.54‡ | 0.55 | 0.54‡ |
| Malic acid (mg/L) | 923 | 1081† | 1183 | 1554† |
| pH | 3.7 | 3.8‡ | 3.5 | 3.5‡ |
| Volatile acidity (g/L) | 0.015 | 0.015‡ | 0.014 | 0.015‡ |
| Absorbance at 420 nm | 1.9 | 1.9‡ | 1.56 | 1.53‡ |
| Abssorbance at 520 nm | 3.6 | 3.10† | 3.36 | 3.07† |
| Polymeric tannins (μg/L) | 212 | 215‡ | 168 | 170‡ |
| Quercetin glycosides (μg/L) | 11 | 7† | 17 | 11† |
| 3-Isobutyl-2-methoxypyrazine (ng/L) | 2 | 5† | 3 | 6† |
| Hexanol (μg/L) | 1569 | 1503‡ | 1961 | 1945‡ |
- †Indicates a significant difference among treatments (P ≤ 0.05). ‡No significant difference between treatments (P ≤ 0.05).
Wine descriptive analysis showed significant differences in some of the attributes judged by panellists (Tables S1 and S2). In the first year, wines from the rainfall exclusion treatment were considered to exhibit less intense asparagus, peas/bean, and capsicum aromas, but higher red berry and red confectionary aromas, and more intense red confectionary and red berry flavour than wines made from the normal rainfall treatment. The quality of tannin was scored higher on wines from the normal rainfall treatment in 2009. In 2010, wines from the normal rainfall treatment were judged as having higher red fruit and confectionary aromas, and higher red fruit and confectionary taste (Table 8) compared to those from the rainfall exclusion treatment.
| 2009 | Winter rainfall exclusion | Normal rainfall | P-value |
|---|---|---|---|
| Quality of tannin | 6.8 | 7.7 | 0.03 |
| Aroma intensity | 9.0 | 8.7 | 0.10 |
| Red berry_N | 6.9 | 6.0 | 0.03 |
| Red confectionary_N | 6.4 | 5.3 | 0.06 |
| Peas/bean_N | 2.8 | 3.7 | 0.06 |
| Capsicum_N | 1.5 | 2.1 | 0.09 |
| Asparagus_N | 1.3 | 1.9 | 0.002 |
| Red berry_P | 6.1 | 5.3 | 0.01 |
| Red confectionary_P | 4.6 | 3.7 | 0.002 |
| 2010 | Winter rainfall exclusion | Normal rainfall | P-value |
| Earthy | 2.5 | 2.2 | 0.09 |
| Acidity | 6.2 | 5.9 | 0.01 |
| Confectionary_N | 4.8 | 6.0 | 0.06 |
| Red fruit_N | 5.5 | 6.4 | 0.05 |
| Confectionary_P | 4.5 | 5.8 | 0.05 |
| Red fruit_P | 5.7 | 6.8 | 0.02 |
- Letters N and P denote attributes judged by nose and ‘palate.
Discussion
Absence of winter rainfall significantly reduced canopy growth, and, thereby, increased fruit exposure. Rainfall exclusion severely restricted shoot growth in both seasons, which might be explained by increased synthesis of abscisic acid (ABA), a known inhibitor of shoot growth, whose production in the roots is stimulated in drying soils (Zhang and Davies 1990, Khalil and Grace 1993). Vines from the normal rainfall treatment did not experience this growth restriction and were able to develop larger canopies over time. Rapid canopy growth in spring, however, significantly depleted soil moisture, and soils reached lower moisture content in the normal rainfall treatment compared to that of the rainfall exclusion treatment. Soil moisture further away from the vine was more rapidly depleted prior to the initiation of irrigation, which also would stimulate the synthesis of ABA, and therefore affect plant water status. This may explain why both treatments reached the midday leaf water potential threshold of −1 MPa at a similar time. Changes in leaf physiology (i.e. stomatal conductance) are related more to soil water status than to leaf water status (Davies and Zhang 1991). When irrigation was commenced, the larger canopies continued to deplete soil moisture at a rate faster in those areas more distant from the trunk (i.e. middle of the row), especially in the warmer 2009 season without exhibiting a high level of water stress. This is in agreement with studies where partial drying of the root system was used to obtain vigour control without the effect of severe water stress such as yield reduction, a technique now used commercially and known as Partial Rootzone Drying (Dry 1997, Loveys et al. 1997).
A change in canopy growth also influenced fruit exposure, such that a higher rate of light interception occurred in the fruiting zone for vines with a smaller canopy (i.e. rainfall exclusion treatment). This is likely to explain the difference in the IBMP concentration seen in both years, where vines with more open canopies and higher fruit exposure had a lower IBMP concentration in fruit, also reported previously (Allen and Lacey 1993, Hashizume and Samuta 1999, Ryona et al. 2008). Vine vigour has previously been linked to fruit IBMP concentration (Allen 2001, Chapman et al. 2004a,b, Wilkinson et al. 2007), and the effect of fruit exposure and vigour on IBMP concentration is difficult to separate. A direct effect of vine vigour cannot be ruled out as we also found that a higher fruit to pruning mass ratio was associated with a lower concentration of IBMP in fruit and wine. It is also possible that synthesis of IBMP was reduced by the higher level of water stress recorded on vines not receiving winter rainfall, as it has been shown that water deficits are usually associated with a low concentration of IBMP in grapes and wines (Roujou de Boubee 2003, Chapman et al. 2005, Sala et al. 2005).
Treatments had little effect on the concentration of C6 compounds in the berries at any time. Differences between treatments were found only in the last two sampling points during 2010, when C6 compounds were significantly lower in fruit from vines that had not received rainfall. Their absolute concentration was close to their sensory threshold (Guth 1997), and was therefore not considered important. The lack of response of these compounds to vineyard management practices reported here is similar to that reported in other studies (Webster and Edwards 1993, Bureau et al. 2000). Both IBMP and C6 compounds peaked before veraison, and decreased afterwards, a pattern described before (Augustyn and Rapp 1982, Allen and Lacey 1993, Lamorte et al. 2007, Ryona et al. 2008, Kalua and Boss 2009).
Along with the reduction in leaf area, a significant effect on berry size has been documented in the literature as a consequence of grapevines being subject to water stress, especially early in the season (Belmonte and Williams 2002, Salon et al. 2005, Shellie 2006). Vines that did not receive winter rainfall had lower midday and pre-dawn leaf water potential over the season compared to that of vines under normal rainfall conditions. Berry and bunch size decreased, and fruit and wine composition was favourably affected by the resulting higher skin to flesh ratio, especially in the first season. It is worth noticing that yield components were already affected by the first season, which suggests a negative effect of low soil moisture prior to budbreak on the latter stages of inflorescence differentiation.
The increase in anthocyanins in berries of water-stressed vines (Kliewer et al. 1983, McCarthy 1997, Ginestar et al. 1998) and a positive effect on the concentration of phenolic substances in response to water stress have been reported in the literature (Ojeda et al. 2002). Decreased malic acid and increased mouthfeel compounds, mainly quercetin glycosides, are also cited as key effects of water stress and fruit exposure on fruit composition (Matthews and Anderson 1988, Haselgrove et al. 2000, Salon et al. 2005, Shellie 2006, Joscelyne et al. 2007). Wine DA followed this trend, with the wines made from the rainfall exclusion treatment judged as fruitier and less vegetal compared to that of wines from the normal rainfall treatment in 2009. Although these results agree with the fruit chemistry data, and with previous findings (Chapman et al. 2005), the reduction in berry size observed for vines subject to the rainfall exclusion treatment in the first year was not enough to offset the reduction in canopy growth observed in the second season. The 2010 wines were judged quite differently from those of the year before, and wines made with fruit from the normal rainfall treatment were rated fruitier by nose and palate. This is probably related to the differences seen between treatments in both the leaf area to crop mass and fruit to pruning mass ratios. By the second season (2010), vines that had not received any rainfall had a lower leaf area to crop mass ratio and a higher fruit to pruning mass ratio compared to that of their counterparts under normal rainfall, and were out of the optimal ranges for both ratios (0.8 to 1.2 m2/kg and 4 to 10 m2/kg, respectively (Kliewer and Dokoozlian 2005).
These results suggest that vines not receiving a level of rainfall will tend to overcrop over time, recording a significant increase on the fruit to pruning mass ratio by the second season, and likely resulting in canopies that are not large enough to adequately ripen their crop load. Shoot growth has previously been shown to be sensitive to water stress (Smart and Coombe 1983, Williams and Matthews 1990), but our findings show that regardless of in-season irrigation, soil moisture at budbreak plays an extremely important role in determining the potential canopy size, also affecting the grape and wine secondary metabolites measured, as well as some yield components.
Conclusions
The effect of reduced availability of soil moisture on fruit green aromas was consistently positive with a lower concentration of IBMP observed in the absence of winter rainfall; this was, however, associated with a substantial decrease in yield. Moreover, the decrease in canopy size was much larger than the reduction in crop load, which caused a severe vine imbalance by the second season, offsetting the positive effect of reduced berry size on grape and wine composition. As growers face new challenges including rising temperature and reduced water availability, vineyards may need to be irrigated earlier in the season in order to develop canopies capable of adequately ripening a profitable crop. Otherwise, severe crop reduction such as bunch removal will be needed to achieve a balance between the amount of fruit and the size of the canopy.
Acknowledgements
We are grateful for the assistance of Mike Cleary, Bruce Pan, Hui Chong and Steve Tallman with the grape and wine chemical analyses. The advice of Dr Paul Boss from CSIRO made a significant contribution to this research. The field assistance of Don Katayama and Nona Ebisuda, and the research winemaking support of Cyd Yonker and David Santino are also gratefully acknowledged.
