Water use efficiency (WUE)
Regional estimates on changes in water availability under future climate conditions suggest that for most grape-growing regions, the propensity for water deficiency will increase during the growing season (IPCC 2008). Together with increasing competition for irrigation water, WUE in viticulture will be an important issue of the future. Therefore, to secure a sustainable and effective use of water, more information on the physiological and genetic basis of WUE is needed (Chaves and Oliveira 2004). There is evidence for variation in WUE among species and cultivars in grapevines (Düring and Scienza 1980, Eibach and Alleweldt 1984, Chaves et al. 1987, Bota et al. 2001, Schultz 2003a, Flexas et al. 2004, Souza et al. 2005b, Soar et al. 2006a, Flexas et al. 2008) and some of the aspects of genetic variation of WUE have been elucidated in Arabidopsis thaliana as a model plant (Masle et al. 2005, Nilson and Assmann 2007, Lake and Woodward 2008). Despite demonstrated interest in WUE, it can be defined in different ways and the physiological basis of its regulation is not fully understood. WUE depends on complex interactions between environmental factors and physiological mechanisms such as stomatal behaviour, photosynthetic capacity, and leaf and plant anatomy (for a review see Bacon 2004). Additionally, all mechanisms that tend to maintain plant survival or a certain productivity under conditions of limited water supply or high evaporative demand (or adverse conditions in general) come at certain ‘costs’, which, from an agricultural point of view, will either reduce dry matter production or competitive ability (Jones 1992). In that sense, high WUE genotypes may not be the ideal compromise between drought tolerance and economic performance.
WUE – challenges and limitations of measurements and data interpretation. WUE is often determined from single leaf gas-exchange measurements, relating net CO2 assimilation rate (A) either to stomatal conductance for water vapour (g), termed intrinsic water use efficiency (WUEi) (Osmond et al. 1980) (Eqn 1),
or A to leaf transpiration rate (E), termed instantaneous water use efficiency WUEinst (Eqn 2).
The former (Eqn 1) is a way to largely exclude the effects of changing evaporative demand on water flux out of the leaf (Bierhuizen and Slatyer 1965) and has been predominantly used in studies of water stress effects on grapevines (i.e. Düring 1987, Schultz 1996, Flexas et al. 1998, Escalona et al. 1999, Bota et al. 2001, Souza et al. 2005b, Chaves et al. 2007, Pou et al. 2008, Zsófi et al. 2009). In general, an increase in WUEi under drought or deficit irrigation strategies has been observed in these studies, but it is difficult to integrate WUEi over time. Diurnal changes in g and environmental conditions may preclude a close relationship to long-term WUE measurements such as, for instance, leaf carbon isotope composition described below.
Less frequently used as WUEi is WUEinst or the integrated WUE (Wp) (Farquhar et al. 1989) which includes losses of carbon beause of respiration at night or from non-photosynthetic organs such as roots and water loss at night because of incomplete stomatal closure or high cuticular conductance (Eqn 3). Data like these are most relevant but are confined to pot studies and are reflected in ratios such as dry matter produced per unit of water lost (i.e. Düring and Scienza 1980, Gibberd et al. 2001).
where ΦC = carbon loss via respiration
where ΦE = unproductive water loss
More recently, the analyses of 13C carbon isotope discrimination (Δ13C) has become popular and widespread for the evaluation of water use in plants and grapevines under field conditions (Gaudillère et al. 2002, Medrano et al. 2005, Souza et al. 2005b, Chaves et al. 2007). In contrast to gas-exchange techniques that provide measurements of photosynthesis rates at a single point in time, leaf carbon isotope signature (δ13C) integrates the ratio of intercellular (Ci) to atmospheric CO2 concentration (Ca) for longer periods of time (Condon et al. 2004). The basis of the biochemical discrimination (Δ) against 13C in C3 plants lies within the primary carboxylating enzyme, ribulose-1,5-bisphosphate carboxylase-oxygenase (Rubisco) which discriminates against 13C because of the intrinsically lower reactivity of 13C compared with 12C (Farquhar et al. 1982).
Variations in Δ13C are used to analyse temporal and spatial trends in plant–carbon water relations. In the most frequently used version of the Farquhar et al. (1982) model – the linear (or reduced) form relates Δ13C linearly to Ci/Ca (Eqn 4).
With a and b' being isotopic fractionation coefficients (a = 4.4‰, fractionation during CO2 diffusion through stomata, O'Leary 1981; b' = 27‰, fractionation associated with reactions by Rubisco and phosphoenol pyruvate (PEP)-carboxylase, Farquhar and Richards 1984). The linear relationship between Ci/Ca and Δ13C can be used to calculate WUEi (A/g) because Ci/Ca reflects the balance between A and stomatal conductance for CO2 (gc) and because gc and stomatal conductance for water vapour (g) are related by a constant factor (g = 1.6 gc) (Farquhar et al. 1980) (Eqn 5)
Inherit in Equations 4 and 5 are, that if Ci/Ca decreases, then Δ13C decreases and WUEi will increase. If we relate the linear model of carbon isotope discrimination in an analogous manner to WUEinst, where leaf transpiration rate, E, is a function of g and the leaf to air vapour pressure deficit (LAVPD) (kPa), (Eqn 6) (Farquhar and Richards 1984)
which implies that WUEinst will vary with Ci (affecting Δ13C) and LAVPD because Ca is essentially constant. It also implies that because evaporative demand is considered in WUEinst, but not in WUEi, both can show different trends (opposite directions) while these can still be consistent with the same trend in Δ13C (Seibt et al. 2008).
VPD response and signalling mechanisms
Research has shown that stomatal conductance of grapevines is sensitive to the vapour pressure deficit of the air (VPD) (kPa), (i.e. Düring 1987, Soar et al. 2006a,b, Poni et al. 2009) closely related to LAVPD, which will directly affect WUEinst (g, E) and WUEi (g). Figure 1 shows the relationships between WUEinst and WUEi with LAVPD for four grapevine varieties grown in pots (30L, 3-year old vines) under a wide range of climatic conditions (photosynthetic photon flux density, PPFD 750–1750 µmol/m2/s; air temperature, Tair, 19°C–37°C) during different drought and recovery experiments (pre-dawn water potential ψPD, −0.15 to −1.45 MPa) over a period of several weeks (Chouzouri and Schultz, unpublished data). When all data are pooled, there is a clear relationship of WUEinst to LAVPD, with WUEinst decreasing with increasing LAVPD, while there is none for WUEi. This is consistent with the optimisation theory (Cowan 1977) and is important because (i) it indicates that LAVPD negatively affects WUEinst over a wide range of plant water status; and (ii) because this is not apparent in WUEi, measurements conducted under high LAVPD and water deficit may show opposite tendencies in these two parameters. This becomes clear from diurnal gas-exchange data of a field experiment with the variety Syrah (Shiraz), where LAVPD was high during a continuous dry down over several months (Figure 2). As expected, WUEi was higher for stressed than for irrigated plants throughout most of the day at any level of water deficit (Figure 2(a),(b),(c)). However, WUEinst was either not different between irrigated and non-irrigated plants (Figure 2(d)) or lower in the stressed plants when water deficit became more severe (Figure 2(e),(f)) and stomatal closure caused LAVPD to rise substantially (Figure 2(j)–(l)) because of increases in leaf temperature (Figure 2(g),(i)).
Figure 1. Response of instantaneous water use efficiency (WUEinst) (a) and intrinsic water use efficiency (WUEi) (b) to leaf to air vapour pressure deficit (LAVPD) of four grapevine varieties (Silvaner, Syrah, Grenache, Airen) over a wide range of climate conditions (photosynthetic photon flux density (PPFD) 750–1750 µmol/m2/s; air temperature (Tair), 19°C–37°C) during different drought and recovery experiments (pre-dawn water potential (ψPD), −0.15 to −1.45 MPa) over a period of several weeks. Plants were grown outdoors in 30 L pots in 2003 (Chouzouri and Schultz unpublished data).
Download figure to PowerPoint
Figure 2. Diurnal time courses in intrinsic water use efficiency (WUEi) ((a)–(c)), instantaneous water use efficiency (WUEinst) ((d)–(f)), leaf temperature ((g)–(i)) and leaf to air vapour pressure deficit (LAVPD) ((j)–(l)) of irrigated (open symbols) and non-irrigated (stressed) (closed symbols) field-grown Syrah grapevines in a commercial vineyard near Montpellier, France on 30 June (2 weeks post bloom), 15 July and 11 August (veraison). The development in ψPD is indicated in (d)–(f). For details on experimental set-up and materials and methods see Schultz (2003a).
Download figure to PowerPoint
There is uncertainty whether a pronounced stomatal response to VPD during water deficit is beneficial for improving WUE, but there are clear differences in this response between varieties (Düring 1987, Soar et al. 2006b) and rootstock/scion combinations (Soar et al. 2006a, Koundouras et al. 2008). Water stress intensified the VPD response in some rootstock/scion combinations but had the opposite effect on others (Soar et al. 2006a). Irrespective of the irrigation strategy applied, g decreased in response to rising VPD under partial root zone drying (PRD) (Loveys et al. 2004) and regulated deficit irrigation (RDI) (Patakas et al. 2005, Pou et al. 2008) which would allow the plant to better control E. However, depending on the intensity of the closure response and the extent of the subsequent rise in temperature and LAVPD, WUEinst may be reduced (as in Figure 2), which has been observed in other studies on grapevines (Naor et al. 1994, Naor and Bravdo 2000, Koundouras et al. 2008) and trees (Jifon and Syvertsen 2003, Baldocchi and Bowling 2003). In this context, it is important to realise, that a reduction in WUEinst may occur despite Δ13C data suggesting the opposite (Koundouras et al. 2008, Seibt et al. 2008). Figure 3 illustrates this using different scenarios and combining Equations 4, 5 and 7. At the same carbon isotope signature indicating an increase in WUEi with decreasing Δ13C, WUEinst may increase, stay constant or even decrease (Figure 3(b)) depending on the development in LAVPD (Figure 3(a)). It is recognised that this model ignores the contribution of changes in mesophyll conductance with drought and/or temperature increase and photorespiration (Farquhar et al. 1982, Flexas et al. 2002, 2007) or even decreased activities of photosynthetic enzymes under severe stress (Flexas et al. 2006). Incorporating these aspects into a more comprehensive model can yield similar results (Baldocchi and Bowling 2003, Seibt et al. 2008). Added in Figure 3 are some data from Gibberd et al. (2001) obtained in a pot study of 19 different grapevine genotypes, where different groups were separated depending on obtained relationships between WUE on the whole plant level (dry matter produced/amount of water transpired, e.g. close to Eqn 3) and Δ13C under well-watered conditions. Functional groups which exhibited a greater decrease in g had lower Δ13C values but gained little in WUE (Figure 3B), probably because increasing leaf temperature and LAVPD increased E and did offset decreases in g. This effect will strengthen with decreasing wind-speed, decreasing water availability, increasing leaf size, increasing ambient temperature and denser canopies, where the latter four could be expected to be more important in the future because of warming and rising ambient CO2 concentrations. All the listed factors and their impact on leaf and canopy temperature are also playing a role in the use of new proxy and remote sensing techniques for stress monitoring and irrigation management (Jones et al. 2002, Möller et al. 2007, see section Canopy (leaf) temperature and imaging technologies).
Figure 3. Modelled changes in intrinsic water use efficiency (WUEi) and instantaneous water use efficiency (WUEinst) at a given Ci/Ca ratio and three cases of possible changes in leaf to air vapour pressure deficit (LAVPD) (a) as a function of Δ13C (b). The model is based on Equations 4, 5 and 7 and the conceptual analysis presented by Seibt et al. (2008). Some data from Gibberd et al. (2001) obtained in a pot study with 19 different grapevine genotypes are added. In this study, different groups were separated depending on obtained relationships between water use efficiency (WUE) on the whole plant level (dry matter produced/amount of water transpired, e.g. close to Eqn 3) and Δ13C under well-watered conditions.
Download figure to PowerPoint
Differences in the responsiveness of stomata to environmental factors, such as VPD and/or water deficit, seem to be modulated by abscisic acid (ABA), either through the increased ability of the xylem to supply ABA (Loveys 1984a, Loveys et al. 2004), increases in xylem pH affecting ABA ionization and general metabolism (Stoll et al. 2000) or through differential gene expression in the ABA biosynthetic pathway involved in regulating varietal responses to VPD (Soar et al. 2006a). However, absolute control of g by ABA is disputed, and hydraulic signalling may also play a role (Schultz 2003a, Rodrigues et al. 2008) and both factors may actually partly depend on each other in the regulation and recovery of g during and after a water stress (Lovisolo et al. 2008a). Depending on the degree of stress, the rate and degree of recovery may be very different. Flexas et al. (2004, 2006) and Galmés et al. (2007) found that grapevines subjected to a mild deficit (i.e. maximum g above 100 mmol/m2/s) fully recovered within a day after rewatering, whereas more severely stressed plants recovered slowly during the following week and did not reach pre-stress levels in A and g.
Nitric oxide (NO) may be another signalling molecule (aside of or together with ABA) involved in stomatal responses to drought. NO has been shown to be involved in many biological events (Quan et al. 2008), and the synthesis of NO by nitric oxide synthase and nitrate reductase in response to drought may elicit the increase of anti-oxidant enzymes which can scavenge water stress-induced hydrogen peroxide (Sang et al. 2008) or may mediate ABA-induced closure of stomata (Neil et al. 2003). Little is known about the relationships between drought and NO accumulation in grapevines, but Patakas and Zotos (2005) found a rapid increase in NO-induced fluorescence close to the stomates in leaves of stressed plants.
Isotopic signatures in leaves and fruits
For grapevine leaves from plants exposed to different irrigation regimes, even the relationships between WUEi and Δ13C may be weak (Souza et al. 2005b, Chaves et al. 2007), but there are reasonable relationships of WUEi with Δ13C of the fruit (Medrano et al. 2005, Souza et al. 2005b, Chaves et al. 2007). These differences have been interpreted as being related to the fact that sugar accumulation starts late in the season and results from current leaf photosynthates, which are produced during the water deficit as compared with the isotopic signature of leaves which are formed earlier during the season (Gaudillère et al. 2002). There could also be more post-photosynthetic fractionation processes (namely respiration) in berries, which might result in differences in the carbon isotope composition of the two organs (Badeck et al. 2005). Despite good correlations to plant water status during berry ripening (i.e. van Leeuwen et al. 2001, Gaudillère et al. 2002), so far the final evidence is lacking that fruit Δ13C is an indicator of crop scale water use efficiency (i.e biomass produced per unit water used, Jones (1992), because correlations to WUEi may not reflect relationships on the whole-canopy level.
Aside from the isotopic discrimination of carbon during photosynthesis and respiration, it may also be useful in future studies to measure the isotopic enrichment of oxygen (18O‰) occurring in leaves during the evaporation of water. The enrichment will increase with a decrease in g, an increase in leaf temperature, a decrease in relative humidity (or an increase in VPD and LAVPD) (Figure 4), a decrease in E, and as such may allow the distinction between effects of soil water deficit and evaporative demand on WUE (Farquhar et al. 2007). This response to relative humidity on the leaf level can also be retraced in grape juice (Tardaguila et al. 1997) and wines coming from areas with different evaporative demand where isotopic enrichment can be related to the prevailing conditions 30 days before harvest (Figure 4) (Hermann and Voerkelius 2008).
Figure 4. Enrichment of leaf water versus relative humidity for five C3 grass species. Original data from Helliker and Ehleringer (2000a, closed symbols, 2000b, open symbols) redrawn in Farquhar et al. (2007). Copyright American Society of Plant Biologists used with permission (http://www.plantphysiol.org). The line represents the published modelled correlation between the isotopic oxygen value in wine as a function of relative humidity of the region of origin from Hermann and Voerkelius (2008). More details can be found in these references. V-SMOW, Vienna Standard Mean Ocean Water (the standard isotopic reference).
Download figure to PowerPoint
Varietal differences in stress responses (isohydric versus anisohydric)
There is clear evidence that varieties differ in their physiological response to stress in their response velocity (Düring and Scienza 1980), adaptive mechanisms (i.e. Düring and Loveys 1982, Loveys 1984b, Schultz 1996, Bota et al. 2001), and effectiveness in regulating gas-exchange (i.e. Chaves et al. 1987, Escalona et al. 1999, Flexas et al. 2002, Schultz 2003a, Souza et al. 2005a,b). A fortunate choice of varieties to study has led to the illustration of these differences in a field experiment with the varieties Grenache and Syrah, whose stomatal behaviour under water deficit led to a classification termed ‘near-isohydric’ and ‘anisohydric’ in reference to Stocker (1956) (Schultz 2003a). Subsequent studies have shown, that these differences in g were also triggered in response to VPD (Soar et al. 2006a), and correlated with differential expression of key genes in the ABA biosynthetic pathway in leaves but not roots (Soar et al. 2006a). The isohydric – anisohydric classification leaves sufficient space to test other varieties, whether they can fit these categories or can enlarge the existing spectrum in order to quantify a general type of WUE, which may be important in irrigated viticulture in the future. However, recent findings in an outdoor ‘heating experiment’ (40°C for three consecutive days) showed that anisohydric Syrah vines up-regulated g, E, and A in response to temperature (Soar et al. 2009) at a common VPD. This may be a positive adaptation in terms of heat tolerance through increased evaporative cooling at the expense of long-term transpiration efficiency (Soar et al. 2009).
The continuum between classical isohydry and anisohydry was well illustrated in a study by Turner et al. (1984) with nine woody and herbaceous species, where the response of E to VPD varied largely. Despite of the correlation to ABA physiology (Soar et al. 2006a), the underlying mechanisms for isohydric or anisohydric behaviour are little understood, and there is evidence that plant hydraulic conductance may play a decisive role (Franks et al. 2007). This conclusion was based on the fact that even in plants capable of strongly down-regulating g as a reaction to water deficit (Franks et al. 2007) it could not be prevented that midday leaf water potential fell to levels where xylem air embolisms may have become the dominant factor in determining water flow and where a functional correlation between embolism formation and leaf gas-exchange existed (Domec et al. 2006, Maherali et al. 2006). Nevertheless, Alsina et al. (2007) did not find a correlation between drought tolerance and vulnerability to xylem embolisms of different grapevine varieties, whereas other data from pot and field experiments showed a correlation (Schultz 2003a, Chouzouri and Schultz 2005, Lovisolo et al. 2008b). The solution to the dilemma in terms of experimental set up may be to test varieties in their combined response of gas-exchange to environmental stresses using the Ball-Woodrow-Berry model (Ball et al. 1987) (Figure 5(a),(b)) and to accompany its application with both ABA and hydraulic conductance measurements to gain clarity. The advantage of the model is that it incorporates the composite sensitivity of stomata to assimilation rate, CO2 concentration, humidity and temperature (Ball et al. 1987), is applicable to grapevines (Schultz et al. 1999, Schultz 2003b), and gives an indication of overall stomatal sensitivity to these factors (Figure 5(a)) (Eqn 8):
Figure 5. (a) Relationship of stomatal conductance to the Ball–Woodrow–Berry ‘index’ (A [rh/Ca]) for plants subjected to progressive soil drying in the field. Lines represent linear regressions from diurnal data at a ψPD of −0.2 MPa (continuous line), −0.50 MPa (long dashed line), and 0.71 MPa (short dashed line) (data from Schultz et al. 2001). K denotes the slope of the lines and is termed stomatal sensitivity factor (Tenhunen et al. 1990). (B) Relationship of k to ψPD for the varieties Grenache (●) and Syrah (○) during water stress in the field then re-watered (recovery) (▵, ▴) and measured 1 day after re-watering. ψPD data are means ± SE (n = 5–6 leaves) (for experimental details see Schultz 2003a).
Download figure to PowerPoint
where g0 is a residual stomatal conductance to H2O vapour (A0 when light intensity approaches 0), and k is a constant representing stomatal sensitivity (Tenhunen et al. 1990, Harley and Tenhunen 1991) (Figure 5(a)). Measurements of relative humidity, rh, and CO2 partial pressure (Ca) outside the leaf boundary layer can be used to drive the model. Because this relationship is purely empirical, k and (A [rh/Ca]) are dimensionless (Ball et al. 1987). While the use of rh has been criticised as being non-mechanistic (Aphalo and Jarvis 1993), the model has been successfully coupled to assimilation models of individual perennial crops (Katul et al. 2000), including grapevines (Schultz 2003b) or entire terrestrial ecosystems (Lloyd et al. 1995, Wang et al. 2007, Mo et al. 2008). The constant representing stomatal sensitivity, k, i.e. the slope of the relationship between g and (A [rh/Ca]) is sensitive to water deficit (Figure 5(a)) and can be used to distinguish the sensitivities of g of different varieties (Figure 5(b)). Note, the smaller k, the higher is the stomatal sensitivity (Harley and Tenhunen 1991), thus Figure 5(b) shows that stomates of Grenache exhibit a higher sensitivity (smaller k) than Syrah at similar water potentials.