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?

Unlike most plant crops, especially annuals, wine grape cultivars are often grown best under sub-optimal conditions. Historically, and particularly so in the Old World, farmers preferred to use fertile land for grazing livestock and growing annual crops. Shallow soils, stony soils and steep slopes were dedicated to vineyards or olive tree plantations (van Leeuwen and Seguin, 2006). It is acknowledged that a range of environmental constraints may restrict vigour and yield and thereby enhance the winemaking potential of the grape (Coipel et al., 2006; van Leeuwen et al., 2007). Of all these constraints, it is the restriction of water supply that plays the most significant role in vine behaviour and berry composition. A limitation in vine water uptake reduces shoot growth, berry weight and yield and increases berry anthocyanin and tannin content (Hardie and Considine, 1976; Matthews and Anderson, 1988; Matthews and Anderson, 1989; van Leeuwen and Seguin, 1994; Koundouras et al., 2006). These effects are favourable to grape quality potential, particularly so in red wine production (Ribéreau- Gayon et al., 1998). The effect of water deficit stress on berry sugar content is yield-dependent; for low yields, vine water deficit enhances berry sugar content and for high yields, it depresses berry sugar content (Tregoat et al., 2002). In temperate climates, water deficit conditions are necessary to produce high quality red wine (Seguin, 1986). However, excessive water deficit stress may lead to yield and quality losses (Ojeda et al., 2002). Vine water status is dependant on soil and climate characteristics (van Leeuwen et al., 2004). Soil influences vine water status through its water-holding capacity. Climate acts on vine water status through rainfall and Reference Crop Evapotranspiration (ETo). Vines can be dry farmed or irrigated. Full irrigation increases yield but is detrimental for quality. Deficit irrigation strategies have been developed to produce high quality wines while limiting yield losses through severe water stress (Dry et al., 2001). Deficit irrigation also allows increasing water use efficiency, compared to full irrigation.

Vine water uptake conditions can be assessed through (i) soil water monitoring by means of tensiometers (Nadal and Arola, 1995), watermark probes (Hanson et al., 2000), neutron moisture probes (Seguin, 1986) or Time Domain Reflectometry (Koundouras et al., 1999), (ii) water balance modelling (Lebon et al., 2003; Pellegrino et al., 2006), or (iii) the use of physiological indicators (van Leeuwen et al., 2001a; Jones, 2004; Cifre et al., 2005). Among physiological indicators, two are particularly accurate and useful: stem water potential (Choné et al., 2001a) and the 13C/12Cratio (d13Cor carbon isotope discrimination) measured on grape sugar at ripeness (van Leeuwen et al., 2001b; Gaudillère et al., 2002).

In this article, precise data are presented about the impact of water status on vine development, yield parameters, grape ripening and vintage quality. Methods for assessing vine water uptake conditions and possible applications for water management in vineyards are discussed. Threshold values for water deficit stress levels are proposed. Vine water status has implications on yield and quality parameters and is therefore essential to the economics of vineyard management. Grape growers tend to focus on yields while grape buyers tend to focus on quality parameters. Hence, an accurate and easy-toimplement tool for assessing vine water status can clarify relations between grape growers and buyers in the wine industry.

1. Experimental plots

Three experimental plots planted with Vitis viniferacv. Merlot were chosen among existing dry-farmed blocks in a commercial vineyard in the Saint-Émilion region (44°56'N; 0°11'W). Soils and vine-rooting profiles were characterized by means of a soil pit study. Soil texture was gravelly (G), sandy (S) and clayey (C). The sandy soil had a water table within the reach of the roots. Soil water-holding capacities were highly variable among plots, so as to create a large range of vine water uptake conditions. The maximum distance between the experimental sites was 500 meters, on flat terrain, so that climate conditions can be considered identical. Rootstock was 3309C. Vines were Guyot pruned and trained with a vertical shoot positioned trellis, with two fixed and two movable wires. Vine density was 6000 vines/ha. The study covers the period 2004 – 2007.

2.Climate conditions

Temperature and rainfall data were collected by an automatic weather station in Saint-Émilion (Bordeaux area), and compared to 1970-1999 averages.

3. Vine water status

Vine water status was assessed with the pressure chamber technique (Scholander et al., 1965) and by means of carbon isotope discrimination measured on grape sugar at ripeness (Gaudillère et al., 2002).

Stem water potential (Choné et al., 2001a) was measured with a pressure chamber several times during the season on leaves covered with an opaque plastic bag one hour prior to measurement. Each measurement was replicated 6 times, on 6 individual vines.

Stable isotope ratio in carbon (13C/12C) is modified by photosynthesis: 12Cis preferentially picked up by the enzymes involved in photosynthesis (Farquhar et al., 1989). This process is called « isotope discrimination », because 12C is preferred over 13C. In water stress conditions, this discrimination is less severe and sugars produced contain more 13Ccompared to those produced when plant water status is not limiting. Hence, an index called d13C, based upon 13C/12Cratio in grape sugar, can be used as an integrative indicator of water deficit experienced by vine during grape ripening. 13C/12Cratio is expressed compared to a standard and ranges from -27 p. 1000 (no water deficit) to -20 p. 1000 (severe water deficit stress, van Leeuwen et al., 2001b). For each plot, d13C measurements were carried out on 4 individual samples of grape juice at harvest by mass spectrometry.

4. Modelling of the water deficit of a vintage

In order to assess the water uptake conditions of a range of vintages in Bordeaux, average Water Deficit Stress Index (WDSI) between véraison and ripeness was calculated with the model published by Pieri and Gaudillère (2005). This model is based on a previous model of seasonal dynamics of soil water balance in vineyards (Lebon et al., 2003) to which a non-linear function is added to account for the stomatal regulation. ATotal Transpirable Soil Water (TTSW) content at field capacity of 200 mm was used, which corresponds to an average value for Bordeaux soils. The WDSI calculated by the model ranges between 0 (severe water deficit stress, i.e. stomata permanently closed) and 1 (no water deficit stress, i.e. no stomatal limitation).

5. Shoot growth slackening

Once every week, 50 apexes were sampled randomly on lateral shoots inside the canopy. Growth slackening was considered to have occurred when 50% or more of the apexes completely stopped their growth.

6.Yield parameters

Grape load was determined on 20 vines per plot so as to obtain the average yield per vine. Four replicates of eight hundred berries were sampled on each plot at harvest time and weighed to determine individual berry weight.

7.Leaf area

Acorrelation was established every year between the length of shoots and the leaf area they carried, for primary and secondary shoots, according to Mabrouk and Carbonneau (1996). Then, the length of all the shoots was measured for 10 vines on each plot. Leaf area was deduced from the previously established correlation.

8.Grape composition at ripeness

Grapes were considered ripe for the production of high quality red table wine when the technical staff of the commercial vineyard decided to harvest (generally at 13% potential alcohol). Harvest took place between 44 and 53 days after mid-véraison, depending on the year. Four replicates of eight hundred berries were sampled at ripeness and pressed at 0.5 MPa in a pneumatic laboratory press (BELLOT, 33170 Gradignan, France). Grape sugar was measured with a portable refractometer. Total acidity was measured by titrimetry. Malic acid was analyzed using an enzymatic kit (Bohringer, Mannheim, RFA; expressed in g.L-1).

9.Anthocyanin composition at ripeness

Four replicates of two hundred berries were sampled at ripeness and mixed during 2 minutes in a fruit mixer. Fifty mL of juice were centrifuged at 3500 rotations per minute during 10 minutes. Juice was diluted 50 times in a 1% HCl solution (A) and in purified water (B). Optic Density (OD) was measured at 520 nm through 1 cm. Anthocyanin content was obtained by the formula: Anthocyanin (mg/L) = OD520 (A-B) * 22.76 * 50 (Ribéreau-Gayon et al., 1998).

1.Climate conditions

Climatic conditions varied considerably during the four vintages studied (Figure 1). The year 2004 was slightly warmer and slightly drier than average, while 2005 was warm and exceptionally dry. The year 2006 was dry, except for September which was very rainy, and warm, except for August, which was cool. The 2007 vintage was cool and wet, after a dry and warm month of April.

2. Effect of vine water status on vine growth, yield and grape ripening

Water deficit stress may occur more or less rapidly, depending on the climatic conditions of the year, as shown by the development of stem water potential in the 2004, 2005 and 2007 vintages in the Saint-Émilion region (Figure 2). On the three soils and during the four vintages, shoot growth slackening was related to the intensity of water deficit stress (Figure 3A) and so was berry enlargement. Berry weight was reduced by as much as 50% in water stress conditions (Figure 3B). Grapes contained less malic acid (Figure 3C) and more anthocyanin (Figure 3D) when vines faced water deficit. The highest sugar level was reached when water deficit was mild (Figure 3E).

The occurrence of water deficit stress during the season has a profound effect on the physiological functioning of the vine. Water deficit will cause the stomata to close for a part of the day (Hsiao, 1973). This adjustment by the stomata restricts photosynthesis. Hence, dry matter production is reduced and yield is negatively affected. When roots are present in drying soils, they produce abscisic acid (Stoll et al., 2000). This hormone promotes grape ripening. Hence, vine water restriction has a negative effect on the maturation of the grape (it restricts photosynthesis) and several positive effects (production of abscisic acid, restriction of competition for carbon substances by the shoot tips, smaller berry size). When the water deficit is mild, the positive effects outweigh the negative; grapes contain less malic acid and are higher in sugar, anthocyanin and tannin contents (van Leeuwen and Seguin, 1994; Trégoat et al., 2002). When water stress is excessive, photosynthesis is overly restricted and fruit ripening may be delayed, particularly when the yield is high.

When the water supply is moderately restricted, it improves the quality of red wines by causing the grapes to achieve optimal sugar levels. When the water supply is not restricted at all, grape sugar levels are lower for two reasons: (i) due to competition for carbon between berry ripening and shoot growth and (ii) increased berry size (dilution of sugar in a greater berry volume). When water deficit is severe, grape sugar content is depressed due to limited photosynthesis. Anthocyanin content increases in a linear manner with water deficit stress (Figure 3D). The quality of a red wine depends more on its level of phenolic compounds (anthocyanins and tannins) than on sugar levels in the grape at maturity. Therefore, the winemaking potential of red grapes can still be excellent even if a severe water deficit slightly limits the amount of sugar in the must.

3. Effects of water deficit and soil characteristics on the timing of grape ripening

The moment at which grapes attain ripeness depends both on the precociousness of the phenological cycle (which may be assessed by the date of mid-véraison), and the ripening speed, which can be calculated according to the method published by Duteau (1990). The precociousness of the budding depends on air and soil temperatures (Morlat, 1989). Soil temperature is directly related to soil water content. Hence, budding, as well as the occurence of the following phenological stages, is delayed on wet soils. The grape ripening speed is largely determined by the water status of the vine (van Leeuwen and Seguin, 1994). Moderate water deficit encourages rapid ripening while it restricts the size of the berries (and thus the volume to be filled with sugars). It also reduces competition for carbohydrates between the berries and the shoots, allowing a greater percentage of the sugars produced by photosynthesis to be available for grape ripening (Lebon et al., 2006).

Figure 4 shows an example of the effect of water uptake conditions on the ripening speed (as a function of temperature sum base of 10 °C, also referred to as « Growing Degree Days » by Winkler et al., 1974). Vines were low yielding (

4. Water deficit stress and vintage quality in Bordeaux

In order to assess vineyard water availability of 32 vintages in Bordeaux (1974 – 2005), average Water Deficit Stress Index (WDSI) between véraison and ripeness was calculated by the model published by Pieri and Gaudillère (2005). The WDSI calculated by the model ranges between 0 (severe water deficit stress, i.e. stomata permanently closed) and 1 (no water deficit stress, i.e. no stomatal limitation). A correlation between WDSI values and vintage quality ratings by Bordeaux wine brokers Tastet and Lawton (33000 Bordeaux, France) shows that 45% (significant at a = 0.001 ; n = 32) of the quality of red Bordeaux vintages is explained by the intensity of the water deficit stress (Figure 5). Thus, vine water status can be considered an important factor for red wine vintage quality in Bordeaux. Poor quality vintages are without exception wet vintages: 1974, WDSI = 0.85; 1977, WDSI =0.97; 1984, WDSI = 0.91; 1992, WDSI = 0.99. Vintages where the vines were subjected to moderate to severe water deficit stress correspond without exception to very good or great vintages: 1990, WDSI = 0.42; 1995, WDSI = 0.35; 1998: WDSI = 0.18; 2000, WDSI = 0.31; 2001, WDSI = 0.19; 2003, WDSI = 0.39; 2005, WDSI = 0.05. In not one vintage during these 32 years has the overall quality of wine suffered due to excessive water deficit stress in Bordeaux. Even if during a dry summer excessive water deficit stress might have occurred on some blocks with negative consequences on the wine quality (particularly in blocks with young vines with superficial root systems), the overall quality of the vintage did not suffer.

Temperature is a less decisive factor on vintage quality in Bordeaux than water deficit stress. During the period 1974 - 2005, vintage ratings are neither well-correlated to the sum of active temperatures (R2 = 0.20, data not shown), nor with the average growing season temperature (April – October, R2 = 0.26, data not shown). Although some very warm vintages produced high quality red Bordeaux wines (1989, 1990, 2003, 2005), other warm vintages were not outstanding (1994, 1997). Some cool vintages (1978, 1985, 1988, 1996) were among the finest over the last three decades. However, temperature and water deficit stress are not completely independent factors: high net radiation generates high temperatures as well as high Reference Evapotranspiration rates (ETo).

That said, vintage quality in Bordeaux cannot be explained by vine water status alone: some very good vintages, such as 1982, were vintages without significant water deficit.

5. Managing vine water deficits to optimize berry quality

The ideal water uptake conditions for producing highquality grapes for making red wine corresponds to a moderate water deficit early in the season (before véraison). Early water deficit stress has a greater effect on the reduction of berry size than late water deficit stress (Becker and Zimmermann, 1984). The winemaking potential of the grape will suffer in the cases of both no water deficit and excessive water stress.

The loss of quality through excessive water deficit stress occurs much less frequently than the loss of quality due to insufficient water deficit, even though the latter consequence is often overlooked. If vines face excessive water deficit stress, berry shrinkage and leaf drop may occur, a consequence that is particularly visible and striking to growers. In the case of unlimited water uptake conditions, berries tend to be large and constituents diluted and though no damage to the vines is visible, the negative impact on grape quality potential can be significant. When summer rainfall and water reserves in the soil prevent a regular occurence of water deficit, it is necessary to increase the leaf area per hectare in order to increase evapotranspiration. This can be achieved in short term by increasing the trimming height and in long term by changing the trellis system (higher posts) or to increase vine density (number vines.ha-1). Another solution is to select a rootstock, which only partially uses water reserves of the soil (Riparia Gloire de Montpellier for example). Insufficient water deficit can also be addressed by choosing to plant specific grape varieties (i.e. white varieties or early ripening red varieties, van Leeuwen, 2001).

In situations where a loss of quality occurs in some years due to excessive water deficit stress (i.e. in very dry climates, in soils with low water reserves), it is possible to limit the negative effects on the vine by adapting the training system (i.e. to train the vines as spur-pruned bush vines) and by using appropriate plant material (Choné et al., 2001b). The most effective protection against harmful effects of water stress is yield restriction. A moderate yield per hectare allows a moderately low leaf surface area per hectare without altering the leaf / fruit ratio. Mediterranean bush vines are adapted to arid regions when irrigation is not an option due to water restrictions. Rootstocks mediate vine responses to soil water restrictions through chemical signaling (Soar et al., 2006). Choosing drought-resistant rootstocks (for example 110 Richter) is an efficient and cost-effective way of adaptation to dry conditions. Resistance to drought can also be achieved by choosing varieties adapted to arid regions. Drought resistance is highly variable among cultivars. For example, Vitis vinifera cv. Merlot is sensitive to drought, while Vitis vinifera cv. Grenache and Carignane are more resistant. Fertilization practices impact drought resistance. Low vine nitrogen status reduces water requirements by limiting vigour and leaf area.

Severe water deficits may compel viticulturists to resort to irrigation, although irrigation is restricted in most high-quality wine producing areas in Europe. Due to increasingly limited water resources and the harmful effects of irrigation on soil quality (salinity), vineyard design for dry farming is recommended over irrigation when possible (i.e. when economically sustainable). However, cultivating vines without irrigation while maintaining economically viable yields is difficult when the annual rainfall is less than 400 mm. This value however needs to be adjusted depending on the distribution of rainfall throughout the year and the soil water-holding capacity. In very dry climates, wellmanaged irrigation can enhance quality, but poorly conducted irrigation may cause significant losses in grape quality potential. When irrigation is necessary, it should ideally lead the vine progressively into a situation of moderate water deficit while avoiding a situation of excessive water stress (Choné et al., 2001b).

Historically, grapevines are cultivated amidst challenging environmental conditions. The idea of the necessity of a limiting factor in the production of high quality wines is now making its way into irrigation management which strategically reduce water supply to the vines. In the case of Regulated Deficit Irrigation (RDI), water deficit is deliberately provoked in the early part of the season by withholding irrigation with the specific intent of reducing berry size (McCarthy, 1997; Dry et al., 2001). The second method, Partial Rootzone Drying (PRD), draws on an alternating system of irrigation. Approximately every two weeks, each side of a row is alternately irrigated. In this way, half of the root system is always in a zone of the soil which is in the process of drying out. This method produces a positive effect in the winemaking potential of the grape, which may be partially explained by a greater synthesis of abscisic acid compared to vines irrigated without any water deficit stress (Stoll et al., 2000). Another method of deficit irrigation, called « Sustained Deficit Irrigation » (SDI) has also proven to be effective for reducing water use (Fereres and Soriano, 2007).

6. Methods for assessing the water requirements of the vine

Given the importance of water relations in vines on yield and quality parameters, the measurement of vine water status is very important for vineyard management and research in both irrigated and non-irrigated vineyards. Many methods have been developped over the last five decades. Most of them were first implemented on annual crops and then adapted to vineyards. The vine is a perennial crop, which is often cultivated in stony soils, and which may have deep roots (up to several meters). Because of these particularities, some methods that give satisfactory results on annual crops or fruit orchards do not really work well on vines. Hence, one has to be very prudent when choosing a method for monitoring water supply to the vines. Measurements of vine water uptake can be grouped according to three different approaches: (i) measurements of soil water, (ii) water balance modelling or (iii) physiological indicators. Some methods are more suitable for research purposes, others for practical vineyard management.

a. (i) Assessment of soil water

Soil matric potential indicates the availability of soil water. It can be monitored with tensiometers (Nadal and Arola, 1995). Tensiometers are succesfully implemented for irrgation management in annual crops and fruit tree crops. However, this technique has major drawbacks when used in vineyards. A first limitation of this technique is the rooting depth of vines, generally over one meter. Vines explore several soil layers, each of which has its own matric potential. Tensiometers have thus to be installed at several depths. A second technical limitation is the narrow range of soil matric potential that can be measured: from 0 to –70kPa for classic tensiometers. Watermark probes (Hanson et al., 2000) can assess a greater range of soil matric potentials: 0 to –0.2MPa. However, even at –0.2MPa there is almost no limitation in vine water uptake. Tensiometers and watermark probes can thus only be used for vine water status control when water deficit is weak or when there isn’t any water deficit stress. Hence, deficit irrigation cannot be managed with these tools.

Vine water consumption can be determined and soil water balance established through regular monitoring of soil volumetric water content. Soil volumetric water content can be measured by means of a neutron moisture probe (Seguin, 1986) or Time Domaine Reflectometry (TDR, Koundouras et al., 1999). Studies using a neutron moisture probe have provided details of vine water uptake in gravely soils (e.g. Haut-Medoc: Seguin, 1975), in clayey soils in Pomerol (Duteau et al., 1981), as well as in the calcareous soils of the limestone plateau at Saint-Émilion (Duteau, 1987). However, the neutron moisture probe has some drawbacks when implemented in vineyards. Water assessments conducted in this way do not take into account water coming in laterally when a water table is present in the soil; run-offs from sloping sites are not taken into account either. Eventually, root development can reach the access tube, which would alter the results (van Leeuwen et al., 2001a). Vine rooting can be very deep and neutron moisture probe readings can only be interpreted when the access tube covers the whole root zone, which is rarely the case. Moreover, the types of soils typically found in vineyards (gravelly, rocky) can make it very difficult to insert the access tube. Neutron moisture probes measure soil water content in a sphere with a diameter of less than 50 cm. Many access tubes per plot are necessary to take spatial soil moisture variations into account. Even if the neutron moisture probe is used in certain countries to control vine irrigation, the unwieldiness of this tool will be an obstacle to its widespread use. The use of TDR probes (Time Domain Reflectometry) for establishing vine water uptake assessments presents similar drawbacks.

b.(ii) Water balance modeling

Water balance modelling is a way to estimate soil water availability (Riou and Lebon, 2000; Lebon et al., 2003; Pieri and Gaudillère, 2005). The idea is to simulate the amount of water remaining in the soil throughout the summer by using data regarding the amount of available water in the soil at the beginning of the season (total transpirable soil water at field capacity). The inputs (rainfall, irrigation) are added to this number and the outputs (crop transpiration, evaporation at soil surface) are subtracted (I):

Soil water content is most precisely modelled at daily time step, but water balance models can also be run with a time step of a decade or a month. Rainfall can easily be determined by using a weather station. Crop transpiration and soil evaporation can be accurately estimated by using ETo (reference crop evapotranspiration) and rainfall data from a weather station and also from taking canopy size into account. The main problem with the modeling approach is in estimating the total transpirable soil water at field capacity which is extremely difficult due to varying conditions in vine cultivation (rooting depth, amount of stones in the ground, variations in soil texture, variations in soil bulk density, etc.). Hence, application of these models at plot level provides inconsistent results (Pellegrino et al., 2006). However, these models are useful for describing the climatic and plant-related aspects of vine water deficits at a regional scale and they are powerful tools for classifying vintages according to their dryness.

c. (iii) Physiological indicators

Vine water status can also be monitored by the use of physiological indicators (Cifre et al., 2005). These indicators are based on the principle that vine physiology is modified by water deficit. Many physiological indicators have been developed over the last four decades: (1) transpiration (Hsiao, 1973), (2) water potentials (Begg and Turner, 1970), (3) microvariations in stem or berry diameter (Garnier and Berger, 1986; Greenspan et al., 1994; van Leeuwen et al., 2001a), (4) differences between leaf and air temperatures (Jones, 1999), (5) carbon isotope discrimination measured on grape sugars (Farquhar et al., 1989; van Leeuwen et al., 2001b; Gaudillère et al., 2002), (6) sap flow measurement (Escalona et al., 2002) and (7) growth parameters (Pellegrino et al., 2005). Among these accurate physiological indicators, two are practical tools for determining vine water status: stem water potential and carbon isotope discrimination measured on grape sugars.

Predawn, leaf and stem water potentials

Water potentials in vascular plants can be measured by means of a pressure chamber (Scholander et al., 1965). These potentials are generally measured on leaves. The pressure chamber technique is applied to: (i) leaf water potential, (ii) pre-dawn leaf water potential and (iii) stem water potential. The range of midday leaf water potential values, pre-dawn leaf water potential values and midday stem water potential values with regard to vine water deficit are presented in table 1. However, these values are average thresholds which may vary from block to block depending on root distribution, vine vigour and yield.

Leaf water potential can be measured during the course of the day. The more negative the water potential in the leaf, the greater the water deficit in the vine. Leaf water potential reaches a minimum in the early afternoon and it is this time of day that is generally chosen for comparing measurements. Although it has been shown that leaf water potential varies with vine water status, it is also highly variable depending on the microclimatic environment of each particular leaf (Jones, 2004). Moreover, vines may have an isohydric behaviour (Schultz, 2003) and they limit variations in water potential of their leaves by stomatal regulation. For these two reasons, midday leaf water potential is not the most accurate indicator of vine water status.

When water potential is measured on leaves at the end of the night (so-called pre-dawn leaf water potential), microclimatic conditions are homogeneous among leaves and vines are not transpiring. At this time of day, each single leaf of a vine has a similar water potential. This water potential is in equilibrium with the most humid soil layer explored by the root system. Although pre-dawn leaf water potential better represents vine water status in relation to soil water availability than midday leaf water potential does, it underestimates water deficits when soil water is heterogeneous (Améglio et al., 1999). For a given level of pre-dawn leaf water potential, vines will be more subject to water deficit stress during the day when soil water content is heterogeneous. A small humid soil layer might be able to rehydrate a vine overnight (with predawn leaf water potential values close to zero), but might not be able to provide enough water to the vine during the day to meet evaporative demand, particularly when the canopy is large. This is typically the case with regard to irrigated vines and therefore, pre-dawn leaf water potential is not an accurate indicator of vine water status for irrigation management (Améglio et al., 1999).

Stem water potential is measured during the day on a leaf that is bagged with an opaque plastic bag at least one hour prior to measurement (Begg and Turner, 1970). The opaque bag prevents the leaf from transpiring. During the hour between bagging and taking the water potential measurement, the water potential in the leaf balances with the water potential in the stem xylem. Although the measurement is carried out on a single leaf, the obtained value represents whole vine water potential. Hence, when measurements are carried out on several leaves of the same vine, the coefficient of variation (%) of stem water potential is consistently lower compared to pre-dawn leaf water potentials or leaf water potentials (van Leeuwen et al., 2007). Stem water potential values reach a minimum in the early afternoon. This moment is generally chosen for comparing measurements among sites. Stem water potential values reflect soil water availability, but they also depend on climatic parameters. Any comparisons of soil water availability through stem water potential readings should therefore be carried out in similar climatic conditions, for example on sunny days without extreme temperatures. Stem water potential is consistently well- correlated to vine transpiration which is not always the case for midday leaf water potential (Choné et al., 2001a). Stem water potential is closely related to other physiological parameters of the plant (McCutchan and Shackel, 1992).

Because stem water potential represents whole vine water status during the day, it is a particularly useful tool for irrigation management. It accurately represents vine water status, even if soil water content is heterogeneous, which is the case in irrigated vineyards (Shackel, 2006). It can also be used in dry-farmed vineyards for measuring residual water deficits after rainfall (Choné et al., 2000). However, the specific threshold of stem water potential that gives way to irreversible damage on canopy and grapes varies with the vigor of the vines. Water deficit damage is caused by vascular embolism (Schultz and Matthews, 1988). Xylem vessels are small in vines that have been exposed to early water deficit stress and large in vines that developed under unlimited water uptake conditions (Lovisolo and Schubert, 1998). On vigorous vines with large xylem vessels that are suddenly exposed to excessive water deficit, irreversable embolism might occur at stem water potential readings of -1.2 MPa. Low vigor vines that are progressively exposed to water deficits might resist to stem water potential levels of -1.6 MPa (no excessive leaf drop of, no berry shrivel). In dry-farmed, Bordeaux vineyards, it is a current observation that vines are remarkably resistant to drought conditions when water deficits develop progressively during the season. This happened during the 2005 vintage. Although it was the dryest ever recorded in Bordeaux, no major water stress damage was recorded on vines.

Carbon isotope discrimination

Ambient CO2 contains 98.9% of 12C isotope and 1.1% of 13C isotope. 12C is more easily used by the enzymes of photosynthesis in their production of hexoses. Therefore, the sugar produced by photosynthesis contains a higher rate of the 12C isotope than ambient CO2. This process is called « isotope discrimination ». When plants face water deficit conditions, isotope discrimination is reduced because of stomatal closure (Farquhar et al., 1989). Therefore, the 12C/13C ratio in products of photosynthesis forms a signature of plant water uptake conditions over the period in which they were synthesised. When measured on grape sugar at ripeness, the 12C/13C ratio (so-called d13C) indicates average vine water status during grape ripening (van Leeuwen et al., 2001b, Gaudillère et al., 2002).

d13C can easily be measured by mass spectrometry in specialized laboratories. The 12C/13C ratio in the sample is compared to that in an international standard, the so-called PDB standard which is a rock in which this ratio is particularly stable. The results vary from –20 p. 1000 (severe water deficit stress) to –27 p. 1000 (no water deficit stress, table 1). d13C is well-correlated to stem water potential (Figure 6).

The advantage of the carbon isotope discrimination method lies in the fact that it does not require any field measurements other than sampling the grapes at maturity (van Leeuwen et al., 2001b; Gaudillère et al., 2002). Hence, many measurements can easily be carried out. This is not the case with a pressure chamber, which is a more time consuming and labour intensive tool. In dry farmed vineyards, d13C is a valuable tool for determining water deficits at block scale or even at intra-block scale (van Leeuwen et al., 2006). On the Bordeaux estate presented in figure 7A, dry blocks (high d13C values) correspond to gravelly soils with low water-holding capacity. On these blocks, soil resistivity is high (Figure 7B). Blocks with higher water-holding capacities (due to a higher clay content in the soil) show more negative d13C values (Figure 7A). On these blocks, soil resistivity is low (Figure 7B). However, the fact that the result of carbon isotope discrimination is only available after harvest date makes this tool unsuitable to assess dayto- day vineyard irrigation requirements.

In many countries, the wine industry buys grapes from independent growers. Buyers of red grapes generally ask growers to deficit irrigate their vines in order to enhance grape quality potential (reduced berry size, increased anthocyanin and total phenolic content). Growers might be reluctant to reduce irrigation however, because deficit irrigation tends to reduce yields. It is overly labour intensive for grape buyers to control irrigation in the field. However, the d13C measurement provides a tool with which grape buyers can, at a reasonable cost, control the level of irrigation performed on the purchased grapes. In cooperatives, the d13C measurement can be a useful tool for block selection or for differential payment of grapes depending on vine water status.

Vine phenology and grape ripening are highly dependent on water uptake conditions. Moderate water deficits reduce shoot growth, berry size and yield and enhance fruit ripening and phenolic compound synthesis in the berries. These factors generally enhance grape quality for the production of red table wines. The quality of red Bordeaux wine can be better correlated to the dryness of the vintage (calculated by means of a water balance model) than to the sum of active temperatures.

Although the vine is a very drought-resistant species, severe water stress can sometimes alter grape quality potential. In many situations, however, quality losses through severe water stress can be avoided if growers use drought-adapted plant material, practice accurate canopy management, reduce nitrogen fertilization to avoid excessive vigour and reduce yield. If rainfall is particularly low (

A large range of methods for assessing vine water uptake conditions have been developed over the years. Some of them are more accurate and more easily performed than others. Vine water status is most accurately assessed by using physiological indicators. In our experimental conditions, stem water potential is one of the best ways to assess plant water status. Stem water potential may be the most accurate tool for irrigation management.

Carbon isotope discrimination (d13C), measured on grape sugar at ripeness, indicates the average water deficit stress experienced by the vines during grape ripening. Because carbon isotope discrimination facilitates the execution of many measurements, it is a useful tool for block characterisation (for example in terroir studies). While grape quality potential is closely related to the occurence of water deficits during grape ripening, d13C can be used as a quality predictor in viticulture. Because the carbon isotope discrimination measurement takes place at ripeness, it cannot be used for irrigation management during the growing season. However, it can be used at harvest to validate irrigation strategies.