Abstract
To determine the effects of timing and extent of regulated deficit irrigation (RDI) on grapevine (Vitis vinifera) canopies, whole-canopy transpiration (TrV) and canopy conductance to water vapor (gc) were calculated from whole-vine gas exchange near key stages of fruit development. The vines were managed under three approaches to RDI: 1) standard industry practice (RDIS), or weekly replacement of 60% to 70% of estimated evapotranspiration (ET) for well-watered grapevines; 2) early additional deficit (RDIE), or one-half of RDIS applied between fruit set and veraison; and 3) late additional deficit (RDIL), or one-half of RDIS applied between veraison and harvest. Compared with RDIS, the additional deficits (RDIE, RDIL) reduced daily cumulative Trv by about 45% (RDIE) and about 48% [RDIL (57% by unit leaf area)]. Diurnal patterns of gc indicated consistent moderate water stress in all RDI regimens (gc ≈50–150 mmol·m−2·s−1). Under RDIE and RDIL, there were transient occurrences of severe water stress, indicated by gc declining below 50 mmol·m−2·s−1. Across the day, vines under RDIE and RDIL had lower gc than RDIS. Under all deficit regimens, TrV exhibited opposing hysteretic loops with solar radiation [photosynthetic photon flux (PPF)] and vapor pressure deficit (VPD), with less sensitivity to VPD in RDIE and RDIL. For a given value of VPD, TrV was higher in the morning than in the afternoon. For a given value of PPF, TrV was higher in the afternoon than in the morning. Single-leaf measurements of transpiration overestimated TrV by an average of 45%. Instantaneous water use efficiency (WUE) declined during midday at the pre- and postveraison measurements for all RDI regimens. Whole-canopy daily integrated WUE (WUEd) did not differ among regimens during the additional deficits because daily cumulative values of whole-vine net carbon exchange (NCEV) and TrV changed proportionally: by about 43% to 46% in RDIE relative to RDIS. The case was less clear-cut for RDIL, where NCEv declined by 33% and TrV by 48% relative to RDIS. However, WUEd did not differ significantly between the two. More substantial water deficits than those are currently practiced in the industry through RDI could be used for potential water savings in semiarid climates.
Regulated deficit irrigation is used to manage soil water content to impose predetermined periods of plant water stress or soil water deficit that may elicit a desirable response in the plants (see reviews by Behboudian and Singh, 2001; Chaves et al., 2010; Lovisolo et al., 2010). Concomitant with this plant-based objective, RDI is a management tool that is relevant to concerns about water scarcity and the competition for water among agriculture, industry, and domestic use (Intergovernmental Panel on Climate Change, 2014). Regulated deficit irrigation has become widely adopted in the production of wine grapes in arid and semiarid areas. About two-thirds of the major grape-producing regions of the world may be subjected to moderate or severe drought in any given year (Flexas et al., 2010). Irrigation timing and frequency are used to control canopy size (Loveys et al., 2004) and to influence berry size and change berry composition (see reviews by Chaves et al., 2010; Lovisolo et al., 2010), which are particularly pertinent in red wine grapes. The effects of water deficits on vegetative growth, vine productivity, and fruit quality vary based on their timing, and results have not been entirely consistent among studies (e.g., Castellarin et al., 2007; Intrigliolo et al., 2012; Tarara et al., 2011).
Regulated deficit irrigation affects both net carbon exchange (NCE) and transpiration (Tr), which are coupled processes. The ratio between NCE and Tr is referred to as instantaneous WUE. Under RDI, the physiological regulation of WUE is not fully understood (Schultz and Stoll, 2010). For example, under moderate water stress, WUE was not different between irrigated and nonirrigated plants, but WUE decreased under more severe water stress (Schultz and Stoll, 2010). Whole-vine WUE (Poni et al., 2014) under moderate water stress (irrigated to 70% of full-vine ET) did not differ from that of well-watered vines, but whole-vine WUE was substantially lower for more severely water stressed vines (irrigated to 50% of full-vine ET). Variation in WUE under drought occurs across cultivars and within a cultivar depending upon the conditions in which the vines were grown (Lovisolo et al., 2010). There is a paucity of information on WUE at the whole-vine level, but recently WUE was investigated under various deficit irrigation regimens using concurrent single-leaf and whole-vine methods (Merli et al., 2015), the single-leaf measurements providing a link to the bulk of data in the literature.
We reported (Tarara et al., 2011) the consequences on NCE of different timing (pre- vs. postveraison) and degree of RDI [additional deficits of 50% of the industry standard RDI (RDIS)] in V. vinifera ‘Cabernet Sauvignon’. Under an early additional deficit [RDIE (fruit set to veraison)], vines fixed about 45% less carbon per day than did vines under RDIS. Under a late additional deficit [RDIL (veraison to harvest)], about 33% less carbon was fixed per day than under RDIS. As a function of cumulative irrigation applied, between fruit set and harvest a total water savings of about 40% accompanied RDIE, whereas about 20% less irrigation was applied to vines under RDIL. At harvest, there were no consistent differences among irrigation regimens in berry mass, fruit maturity indices, and quality indicators.
The objective of this report is to demonstrate how the timing and extent of water restrictions greater than that of standard industry RDI alter whole-vine transpiration, gc, and whole-vine WUE in field-grown, mature grapevines in a semiarid climate. The additional deficits were each imposed independently during one of two main periods of fruit development that have been implicated in affecting fruit quality: 1) shortly after fruit set to the onset of veraison and 2) veraison to commercial maturity. These two stages of berry development were selected to investigate the ultimate effect on yield and fruit quality—critical variables for the wine industry—and to contribute to the body of knowledge on physiological responses to deficit irrigation regimens. At key developmental stages, whole-vine measurements of gas exchange were made to minimize the scaling-up limitations of the single-leaf technique.
Materials and Methods
Site description.
The experiment was conducted during 2002 and 2003 in a commercial vineyard (4 ha) ≈15 km west of Paterson, WA (lat. 45°53′N, long. 119°45′W, 125 m above sea level). Long-term mean annual rainfall (1990–2013) is 206 mm and reference evapotranspiration (ETo) is ≈1075 mm [Washington State University AgWeatherNet (AWN), 2015]. The vineyard was located on a 14% south-facing slope on a uniformly deep (≈1 m) Burbank loamy fine sand (sandy-skeletal, mixed mesic Xeric Torriorthents) with an estimated field capacity (FC) of 14.6% v/v and permanent wilting point of 7.1% v/v (U.S. Department of Agriculture, 2015). The vineyard had been planted in 1992 to own-rooted grapevines (‘Cabernet Sauvignon’) in rows oriented north–south, with 2.7 m between rows and 1.8 m between vines for an average plant density of ≈2000 vines/ha. Vines were trained to a bilateral cordon (permanent, horizontal extension of the trunk) at a height of ≈1 m aboveground. Shoots were loosely trained vertically between two foliage wires spaced 0.25 m apart at 0.2 m above the cordon, allowing a “sprawl” architecture above that height. Vines were winter pruned annually to two-bud spurs. All other horticultural practices were according to commercial convention for red wine grapes grown in the area. Fertilizer and pest management interventions were applied uniformly across plots. Irrigation was delivered by drip through a single line per row using 1.8 L·h−1, pressure-compensated emitters spaced 1.2 m apart (three emitters for every two vines).
Meteorological variables were measured on-site (Tarara et al., 2011) or where applicable, reference data were obtained from the Alderdale AWN station [≈10 km west of site (2002, 2003)] and Paterson West [≈15 km east of site (long-term normals)]. Thermal time expressed as degree-days [DD (°C)] was computed in daily increments from the daily maximum and minimum temperatures using a lower threshold of 10 °C and no upper threshold. The daily values were summed arithmethically from 1 Apr. [day of year (DOY) 91] to 31 Oct. (DOY 304) according to local convention for grape production.
RDI regimens.
A complete description of the experimental methods is given in Tarara et al. (2011) and is briefly summarized below. The RDI regimens were applied to plots consisting of five rows each within which the middle two rows were used for data collection. All were irrigated to FC just after budbreak (early April). Irrigation was then withheld until shoots were ≈1 m long to minimize the rate of growth in main shoots. Regulated deficit irrigation was imposed from shortly after fruit set until harvest. All RDI regimens were based on estimated crop evapotranspiration (ETc) derived from ETo (Allen et al., 1998) and a crop coefficient (Kc) for drip irrigated ‘Cabernet Sauvignon’ in eastern Washington (Evans et al., 1993). The RDI regimens had been imposed since 1999 as part of a larger experiment. The treatments were based on regional knowledge of water limitations and published data on the consequences of the timing of water stress on fruit quality in red wine cultivars.
To compute RDIS, Kc was multiplied by 0.7 or 0.6. Therefore, RDIS supplied 70% (1999–2002) and 60% (2003) of estimated ETc on a weekly basis from shortly after fruit set until harvest. The “early” and “late” deficits received 50% of the irrigation that was applied to RDIS vines during specific periods of fruit development. In RDIE, 35% (1999–2002) and 30% (2003) of ETc was applied weekly from shortly after fruit set until veraison, then the vines were returned to RDIS [70% (1999–2002) and 60% (2003) of ETc] until harvest. In RDIL, vines were supplied 70% (1999–2002) and 60% (2003) of ETc (i.e., RDIS) until veraison, then 35% (1999–2002) and 30% (2003) of ETc was applied weekly between veraison and commercial maturity.
Irrigation was delivered in two to four applications (sets) per week to achieve the weekly target. Actual water applied was estimated using the nominal flow rate of the drip emitters and the duration of water delivery as detected by pressure transducers in the drip line. For 4 to 5 weeks after harvest (2002) all plots were irrigated to replace 70% ETc, then were irrigated to FC in late October. In 2003, all plots were irrigated to FC immediately after harvest. Volumetric soil water content (θv) was measured using the neutron scattering method (HydroProbe 503 DR; CPN, Concord, CA) at 0.15-, 0.45-, and 0.75-m depths (n = 3 per plot; in-row and equidistant from emitters). The average of these values was used to represent a 0.9-m deep soil unit for which the grower-cooperator’s irrigation engineer adjusted upward or downward the scheduled irrigation amount in response to deviations from target θv (10% for RDIS, 8.3% for additional deficit).
Whole-vine transpiration.
In 2002 and 2003, whole-vine rates of transpiration were measured concurrently with whole-vine rates of NCE (NCEv; Tarara et al., 2011) during five periods corresponding to key developmental stages in grape berries: 1) fruit set (before initiation of the additional deficit); 2) preveraison (about the end of stage I of berry growth); 3) postveraison (early in stage III of berry growth); 4) preharvest (just as the fruit approached commercial maturity); and 5) postharvest (≈2 weeks after removal of the fruit). Instantaneous rates (millimoles H2O per second) and daily cumulative Tr (liters H2O per day) were calculated per vine (TrV) and per unit leaf area (TrV,LA). The TrV measurements were obtained using framed, open-top, flow-through chambers (≈8 m3 volume) that fully enclosed one vine each without modification of the canopy or trellis. Details of chamber design, operation, and calibration are provided in Perez Peña and Tarara (2004) and Tarara et al. (2011). Briefly, six chambers operated simultaneously, with an in-house-built gas multiplexer switching sample streams among chambers. The rate of airflow through the chambers was ≈16 m3·min−1 for an air exchange rate of about two chamber volumes per minute. Maximum differences in air temperature between ambient and that at canopy height in the chamber were 2.5 to 3.0 °C (Perez Peña and Tarara, 2004). All analog signals were recorded at 2.5-s intervals and averaged every 2 min by datalogger (CR7; Campbell Scientific, Logan, UT) so that a mean was recorded for each chamber every 12 min. Measurements were collected continuously for 36 to 48 h on six vines (two per regimen) after which the chambers were moved to a second set of replicate vines in each RDI regimen, the process was repeated, and again for a third set of replicate vines. Vines were paired across rows so that on any one measurement day (dm), the two vines within an RDI regimen were not along the same drip irrigation line. Replication was addressed by repeated measurements across days within a developmental stage (n = 6 vines). The same 18 experimental vines were retained for both years. Data were analyzed only for complete 24-h measurements.
where ue and uo are air flows (moles per second), we and wo are the mole fractions of H2Ov (moles per mole) entering and exiting the chamber, respectively, and LAV is the foliage area per vine (square meters). Because of distance limitations for drawing gas samples, all experimental vines were within 60 m of the mobile laboratory that housed the IRGA, gas handling units, and data acquisition system. Daily total TrV was calculated by integrating instantaneous rates of TrV over 24 h and is expressed as liters per day (per vine basis), and liters per square meter per day (per unit leaf area basis).
Complete measurement runs consumed 8 to 9 d, during which there were dm with clear skies and dm with cloudy skies, and during which two to four irrigation sets were applied, varying with season (i.e., ETo and Kc) and the RDI regimen. Therefore, two sets of analyses for TrV were conducted to address this variation in weather and in the days of the week on which irrigation was applied: 1) data were pooled across all dm in a developmental stage (n = 6 vines) and 2) data were extracted for the “optimal” measurement day in each run (n = 2 vines), which we defined as dm with clear skies and the lowest likelihood of an RDI regimen being confounded by irrigation scheduling (i.e., the timing of multiple irrigation sets to apply the required amount of water).
where es(Tin) is the saturation vapor pressure at the temperature of the air entering the chamber and ea is the actual vapor pressure.
A hypothetical maximum TrV was calculated using only the optimal dm from each measurement run to scale the values between developmental stages. Integration periods were defined by the duration of the interval between RDI treatments. Estimates were calculated using linear interpolation. This variable is not intended to represent an actual cumulative water use for the season but to indicate relative differences among irrigation regimens using the available data that were least confounded by irrigation sets and weather (i.e., optimal dm).
Single-leaf transpiration.
To provide context with the preponderance of gas exchange measurements in the literature, rates of single-leaf Tr (TrSL) were measured concurrently with TrV at three developmental stages in 2002 (preveraison, postveraison, preharvest) and at all stages at which TrV was measured in 2003. Repeated measurements (≈0800–1400 hr) of TrSL were recorded four times during the day in 2002 and six times per day in 2003, during each full day that TrV was recorded. Measurements were collected from fully expanded, sunlit leaves (n = 9 vines per RDI regimen at each developmental stage; Tarara et al., 2011). Gas exchange was measured under ambient irradiance with a portable photosynthesis system (CIRAS-2, PP Systems) using a 2.5-cm2 leaf cuvette [PLC6(U), PP Systems]. Air flow through the cuvette was 200 cm3·min−1. For the single-leaf measurements, VPD was calculated using leaf temperature and vapor pressure in the cuvette.
Leaf area.
Leaf area per vine was estimated twice during 2002 (veraison and preharvest) and four times during 2003 (fruit set, veraison, preharvest, and postharvest) using the indirect methods detailed in Tarara et al. (2011). Because LAV was measured less frequently in 2002 than in 2003, where data were pooled across years daily cumulative TrV was only assessed on a per vine basis and not per unit LA.
Water use efficiency.
Whole-canopy daily water use efficiency (WUEd), defined as the ratio of net CO2 fixed to water transpired, was estimated for the whole canopy and at the leaf level. Values of NCEV are reported in Tarara et al. (2011). The WUEd was calculated using daily total NCE (NCEd) and daily cumulative TrV (TrV,d). Leaf-level water use efficiency was calculated using daily averages of single-leaf NCE (NCESL) and TrSL. To avoid fluctuations in WUE due to VPD (Osmond et al., 1980), intrinsic WUE (WUEi,V and WUEi,SL) was calculated from the ratio between NCEV and gc (whole-canopy) or NCESL and stomatal conductance (gS).
Statistical analyses.
Data were tested for normality using the Kolmogorov–Smirnov test and for homogeneity of variance using Brown–Forsythe test. A general linear model procedure was used for analysis of variance. Means were compared by Tukey or Tukey–Kramer (P < 0.05) as appropriate. Where data were normally distributed, correlations between plant response variables were assessed using the Pearson product moment correlation coefficient (R). Otherwise, the Spearman Rank correlation coefficient was used (rs). For linear regression analyses, data were transformed as needed to adjust for heterogeneous variances and for distributions that deviated from normal. All statistical analyses were performed using SAS (version 8.2; SAS Institute, Cary, NC). The data collected at fruit set in 2002 were excluded from the analysis because of technical difficulties with the whole-canopy system. Hereafter for conciseness, graphical presentation of TrV comprises the complete dataset from 2003, where gas exchange and LAV data were available at all developmental stages. Instances where TrV values differed significantly between years are noted in the text.
Results
Environmental summary.
Thermal time was below the long-term mean (1702 DD) in 2002 and above average in 2003 (Table 1; Tarara et al., 2011). Key developmental stages (budbreak, anthesis, veraison) occurred ≈1 week earlier in 2003 than in 2002. Annual rainfall during the study was 28% (2002) and 7% (2003) below the long-term mean, and was consistent with the seasonal pattern of rainfall in eastern Washington: 25% (2002) and 18% (2003) of the annual total fell between budbreak and leaf fall. No rain was detected between early June (prebloom) and early August (about veraison) in 2003.
Developmental stages, summary meteorological variables, and irrigation applied to mature grapevines under three regimens of regulated deficit irrigation (reprinted from Tarara et al., 2011).
Soil water content reflected the timing and intensity of the RDI regimens (Fig. 1; Schreiner et al., 2007). Per grower practice, θv declined into early summer until fruit set when RDI was initiated. Thereafter θv trended according to the RDI regimens. Obvious departures from target θv are indicative of the inherent limitations in the posthoc soil water balance approach, particularly under drip irrigation (Stevens and Douglas, 1994). Standard RDI resulted in consistently higher θv than the original target of 10% v/v. Between fruit set and harvest, 63% and 79% of the total water applied to RDIS plots was applied to RDIE and RDIL, respectively (Table 1). In both years, when RDIE plots were under the additional deficit, the vines received 55% of the cumulative amount of water that had been applied to RDIS, about the target value. Water application was somewhat more variable in RDIL: 45% (2002) and 31% (2003) of the cumulative RDIS application was delivered when the RDIL vines were under the additional deficit. The end-of-season variation in θv in 2003 reflects the cooperator’s attempt to immediately rewater the plots to FC (Fig. 1).
Volumetric soil water content (θv) in mature, field-grown grapevines during 2002 (A) and 2003 (B) for three regimens of regulated deficit irrigation (RDI): RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit of 50% of RDIS, RDIL = late additional deficit of 50% of RDIS. Arrows from left to right denote fruit set, veraison, and harvest, respectively. Redrawn by permission from Schreiner et al. (2007).
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Daily cumulative transpiration.
Across the experiment, dm were significantly different from one another (Table 2). There was a significant interaction between RDI regimen and dm during the periods of additional deficit, except during preverasion and preharvest if Tr was expressed as TrV,LA. Hence the computation of daily cumulative TrV and TrV,LA both as a mean across all dm within a developmental stage and for the optimal dm in each stage (i.e., clear skies and the least confounding effect of an irrigation set; Table 3). Treatment differences in daily cumulative TrV generally were conserved whether Tr was expressed as TrV or TrV,LA. Daily cumulative TrV differed between RDI regimens during the periods of additional deficit (Tables 2 and 3) but the differences were not always the same based on all dm or only the optimal dm.
Daily cumulative transpiration (Tr) by mature, field-grown grapevine canopies under three regimens of regulated deficit irrigation (RDI) near key developmental stages in 2003. Data are expressed per vine [TrV (top)] and per unit leaf area [TrV,LA (bottom)]. The Tr values are means of six vines averaged across all measurement days (dm) per developmental stage.
Daily cumulative transpiration (Tr) per day in mature, field-grown grapevines under regulated deficit irrigation (RDI) expressed per vine (TrV) and per unit leaf area (TrV,LA). Data are means of two vines from days with clear skies and the least confounding effect of irrigation application (the “optimal” measurement day) in 2003.
Daily cumulative TrV and TrV,LA did not differ among RDI regimens during fruit set. During preveraison, RDIE vines transpired nearly 60% less per day than RDIS and RDIL, whether expressed per vine or per unit leaf area (Tables 2 and 3). During the postveraison stage, across all dm, RDIE vines did not appear to have recovered to the TrV values of RDIS (Table 2) because of a missed irrigation set. However, on the optimal dm, TrV did not differ between RDIS and RDIE, and both exceeded RDIL. During postveraison, RDIL vines transpired nearly 50% less than RDIS and RDIE on a per vine basis, and nearly 60% less based on TrV,LA. Immediately before harvest, cumulative values of TrV in RDIS had declined from pre- and postveraison by about 40% (Tables 2 and 3). At both the pre- and postharvest measurements, daily cumulative TrV,LA did not differ among RDI regimens.
Transpiration dynamics.
Before the imposition of an additional deficit (i.e., fruit set), the diurnal course of TrV,LA was similar among RDI regimens (Fig. 2). The magnitude and diurnal course of TrV,LA between dm reflected a response to both irrigation application and rapid soil dry-down. Maximum TrV,LA values were about 3 to 3.5 mmol·m−2·s−1. Following 15 mm of irrigation (DOY 178), the diurnal course of TrV,LA largely followed that of solar radiation, with a slight lag in TrV,LA under decreasing solar radiation because of sensitivity to VPD (Fig. 2A). Between irrigation sets (e.g., DOY 181; Fig. 2B), daytime rates of TrV,LA were relatively flat between 0800 and 1300 hr following a maximum of about 2 mmol·m−2·s−1. Vines in RDIE (Fig. 2C) were erroneously irrigated at 40% of the scheduled water on one dm (DOY 183), causing lower rates of TrV,LA compared with RDIS and RDIL. Despite this error, across all dm, daily cumulative TrV,LA did not differ among RDI regimens (Table 2).
Whole-canopy transpiration rate per unit leaf area (TrV,LA) in mature, field-grown grapevines, photosynthetic photon flux (PPF), and air vapor pressure deficit (VPD) during the “fruit set” developmental stage in 2003. Symbols represent the mean of two vines; RDI = regulated deficit irrigation, RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit (preveraison) of 50% of RDIS, RDIL = late additional deficit (postveraison) of 50% of RDIS, ETo = reference evapotranspiration, DOY = day of the year. Data were collected before RDIE was imposed.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
After RDIE had been imposed for 5 weeks (preveraison stage), maximum TrV,LA values were about 4 to 4.5 mmol·m−2·s−1. The diurnal course of TrV,LA in RDIS and RDIL was generally similar to that at fruit set. There was lower TrV,LA in RDIE vines than in RDIS and RDIL vines after a morning maximum: for example, even following an irrigation (e.g., DOY 213; Fig. 3A), RDIE vines transpired 18% less than RDIS and RDIL vines during midday (1000–1400 hr). The TrV,LA in all RDI regimens converged during the latter part of the day when VPD was high. This convergence was not the case under cloudy skies and lower VPD (e.g., DOY 215; Fig. 3B). The largest differences among RDI regimens were apparent at the end of the irrigation cycle (DOY 218; Fig. 3C), where midday TrV,LA in RDIE was about 60% less than that in RDIS and RDIL. The lower TrV,LA in RDIE was reflected in a midday maximum gc of 65 mmol·m−2·s−1, about 60% lower than that of either RDIS or RDIL.
Whole-canopy transpiration rate per unit leaf area (TrV,LA) in mature, field-grown grapevines, photosynthetic photon flux (PPF), and air vapor pressure deficit (VPD) during the “preveraison” developmental stage in 2003. Symbols represent the mean of two vines; RDI = regulated deficit irrigation, RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit (preveraison) of 50% of RDIS, RDIL = late additional deficit (postveraison) of 50% of RDIS, ETo = reference evapotranspiration, DOY = day of the year. Data were collected 3 weeks after RDIE had been imposed.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Three weeks after RDIE vines had been returned to RDIS and RDIL was started (postveraison measurements), maximum rates of TrV,LA in RDIS were about 4 to 4.5 mmol·m−2·s−1 (Fig. 4). The general course of TrV,LA was generally similar to that observed earlier in the season except for lower TrV,LA in RDIL vines compared with RDIE and RDIS after about 0800 hr. Under cloudy skies, RDI regimen appeared to have had no effect on TrV,LA (DOY 234; Fig. 4A). Under nearly clear skies (DOY 238; Fig. 4B) and at the end of the irrigation cycle, midday TrV,LA in RDIE vines was 22% less than that of RDIS vines. However, cumulatively there was no difference in Trv,LA between RDIS and RDIE on this optimal dm, indicating that the RDIE vines had essentially recovered from the 5 weeks of additional deficit. The RDIL vines transpired about 77% less than RDIS during the midday hours on the optimal dm (DOY 238; Fig. 4B). Concurrent midday gc in RDIL ranged from 21 to 37 mmol·m−2·s−1, indicative of severe water stress according to proposed ranges for gS in grapevine (Lovisolo et al., 2010). By contrast, gc in RDIS and RDIE was between 73 and 135 mmol·m−2·s−1, indicating moderate water stress. An irrigation scheduling error (DOY 239) that omitted irrigation from RDIE plots induced enough cumulative water stress that on the subsequent day (DOY 240; Fig. 4C) both the instantaneous rates of TrV,LA and the daily cumulative TrV (data not shown) fell below those of RDIL plots.
Whole-canopy transpiration rate per unit leaf area (TrV,LA) in mature, field-grown grapevines, photosynthetic photon flux (PPF), and air vapor pressure deficit (VPD) during the “postveraison” developmental stage in 2003. Symbols represent the mean of two vines; RDI = regulated deficit irrigation, RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit (preveraison) of 50% of RDIS, RDIL = late additional deficit (postveraison) of 50% of RDIS, ETo = reference evapotranspiration, DOY = day of the year, PP = precipitation. Data were collected 3 weeks after RDIL had been imposed and RDIE had been returned to RDIS.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
At the preharvest developmental stage, maximum TrV,LA in RDIS was about 2.5 mmol·m−2·s−1 (Fig. 5). The TrV,LA in RDIL was below that of RDIE and RDIS by as much as 40% in midafternoon, but this difference was not reflected in daily cumulative TrV,LA (Tables 2 and 3). Midday values of gc indicated continued water stress of varying degrees, where gc in RDIL ranged between 40 and 90 mmol·m−2·s−1, and that in RDIS and RDIE between 60 and 150 mmol·m−2·s−1 (e.g., DOY 256). Rates of TrV,LA at the end of the measurement run were uniformly low and corresponding midday gc indicated moderate to severe water stress [24–76 mmol·m−2·s−1 (RDIL), 40–106 mmol·m−2·s−1 (RDIE and RDIS)]. In the postharvest measurement run the daily pattern of TrV,LA was similar to that during preharvest (data not shown). During both pre- and postharvest measurements, canopies showed some senescence.
Whole-canopy transpiration rate per unit leaf area (TrV,LA) in mature, field-grown grapevines, photosynthetic photon flux (PPF), and air vapor pressure deficit (VPD) during the “preharvest” developmental stage in 2003. Symbols represent the mean of two vines; RDI = regulated deficit irrigation, RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit (preveraison) of 50% of RDIS, RDIL = late additional deficit (postveraison) of 50% of RDIS, ETo = reference evapotranspiration, DOY = day of the year.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Across the entire duration of measurements (fruit set to postharvest), a hypothetical cumulative TrV differed by about 20% between RDIS and RDIE (Table 4). Between RDIS and RDIL, the difference was only 2.6%. During the early additional deficit period, the hypothetical maximum difference between RDIS and RDIE was estimated at about 28%. About half of the measurement season’s TrV (computed from the hypothetical maximum values) was between fruit set and veraison in RDIS and RDIE, whereas in RDIL about three-quarters of the hypothetical maximum water was transpired. Between veraison and harvest, about the same percent of the hypothetical seasonal maximum was transpired in RDIS and RDIE; absolute values were within 11%, supporting the notion that vines under RDIE recovered well from the additional deficit. The hypothetical maximum was low between preharvest and postharvest regardless of RDI regimen.
Estimated hypothetical maximum transpiration per vine (TrV) by mature, field-grown grapevine canopies under three regimens of regulated deficit irrigation (RDI) near key developmental stages in 2003.
We were not able to conduct hourly single-leaf measurements of gas exchange late enough in the day to sufficiently capture the diurnal pattern of Tr. Measurements did reflect the treatment differences that were observed in the whole-vine measurements (data not shown). On the whole, concurrent measurements of TrSL overestimated TrV,LA (Fig. 6). There was a moderate linear relationship between TrSL and TrV,LA (r2 = 0.34, P < 0.001), consistent with our findings for NCE, albeit the NCE relationship was stronger (r2 = 0.61; Tarara et al., 2011).
Linear association between rate of transpiration in mature, field-grown grapevines at the single leaf level (TrSL) and at the canopy level expressed per unit leaf area (TrV,LA), for vines under an industry standard practice of regulated deficit irrigation of weekly replenishing of 60% of crop evapotranspiration, or under an additional deficit that reduced the standard application by half. Symbols represent means over 1 h of simultaneous measurement (n = 6 for TrSL, n = 10 for TrV,LA). Data are from all developmental stages in 2003 except harvest, when there were no coincident measurements.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Henceforth for brevity, only fruit set, preveraison, and postveraison data are shown, and for the optimal dm. The NCEV was strongly related to gc at fruit set and at the pre- and postveraison measurements, whether the vines were under RDIS or an additional deficit (Fig. 7). Under RDIS, TrV,LA was a curvilinear function of solar radiation (expressed as PPF) and of VPD, exhibiting hysteresis between morning and afternoon (Fig. 8). The hysteretic loop was smaller as a function of solar radiation, indicative of the tighter coupling between TrV and solar radiation. Under the additional deficit, both relationships (TrV,LA vs. PPF, TrV,LA vs. VPD) were weak. The hysteresis was clockwise as a function of VPD (Fig. 8A–C) and counter clockwise as a function of PPF (Fig. 8D–F). In other words, under RDIS more water was transpired per unit VPD in the morning than in the afternoon; the contrary was the case as a function of PPF. At a given VPD or a given PPF, TrV,LA under RDIS was higher than that of vines under the additional water deficit (except at either end of the day), consistent with the diurnal curves presented above.
Whole-vine net carbon exchange (NCEV) in mature, field-grown grapevines as a function of bulk canopy conductance (gc) before noon (AM) and after noon (PM), for vines under an industry standard practice of regulated deficit irrigation or under an additional deficit that reduced the standard application by half in 2003. Symbols represent the mean of two vines; RDI = regulated deficit irrigation, RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit (preveraison) of 50% of RDIS, RDIL = late additional deficit (postveraison) of 50% of RDIS.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Rate of whole-vine transpiration per unit leaf area (TrV,LA) in mature, field-grown grapevines as a function of vapor pressure deficit [VPD (A–C)] and photosynthetic photon flux [PPF (D–F)] before noon (AM) and after noon (PM), for vines under an industry standard practice of regulated deficit irrigation [RDIS (A–F)] or under an additional deficit that reduced the standard application by half [RDIE (B,E)], RDIL (C,F)] in 2003. Symbols represent the mean of two vines; RDI = regulated deficit irrigation, RDIS = industry standard practice of weekly replenishing of 60% of crop evapotranspiration, RDIE = early additional deficit (preveraison) of 50% of RDIS, RDIL = late additional deficit (postveraison) of 50% of RDIS.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Water use efficiency.
Daily water use efficiency did not differ among RDI regimens during the periods of additional deficit when considered either as a mean across all dm or under the optimal dm per measurement run (Table 5). The diurnal curves of WUE were synchronized among treatments during fruit set (Fig. 9A). The diurnal pattern of WUE at preveraison did not differ between RDIS and RDIE (Fig. 9B). However, under the late deficit, midday WUE in RDIL dropped below that of RDIS (Fig. 9C). Despite variation in diurnal patterns among RDI regimens and between NCEV and TrV, on a daily basis under RDIE, NCEV and TrV declined about proportionally from RDIS (44% less carbon fixed; 45% less water transpired). On average during the postveraison measurements, 48% less water was transpired and 33% less carbon was fixed under RDIL than under RDIS. When only the optimal dm was considered, the values of both TrV and NCEV for RDIE (preveraison) and RDIL (postveraison) were about 57% less than RDIS. Relative differences in WUEd from single-leaf estimates were consistent with treatment differences in whole-vine data (data not shown). Absolute values of single-leaf WUEd tended to be lower than those from the whole-vine estimates (data not shown). At all measurement periods, gc reached a daily maximum value before noon.
Daily water use efficiency (WUEd) for mature, field-grown grapevine canopies under three regimens of regulated deficit irrigation (RDI) near key developmental stages in 2003. Data are a mean of six vines across all measurement days [dm (top)] and means of two vines on days with clear skies and no confounding irrigation sets [the “optimal” dm (bottom)].
Whole-vine instantaneous water use efficiency (WUE) and canopy conductance (gc) of mature, field-grown grapevines during three measurement periods: (A) “fruit set,” during which all vines were under standard regulated deficit irrigation (RDIS) of weekly replenishing of 60% of crop evapotranspiration; (B) “preveraison,” during which an early additional deficit was imposed (RDIE) that reduced RDIS by half; and (C) “postveraison,” during which a late additional deficit was imposed (RDIL) that reduced RDIS by half in 2003.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 140, 5; 10.21273/JASHS.140.5.413
Using the optimal dm per measurement run, WUEi (NCEV/gc) at fruit set and preveraison was negatively correlated with gc (data not shown); i.e., WUEi generally rose as gc declined. This relationship is in contrast to WUE. Before the imposition of the additional deficits, the correlation between WUEi and gc was nearly identical among regimens (rs = −0.631 to −0.663, P < 0.0001). During the early additional deficit, the relationship between WUEi and gc was stronger for RDIS (rs = −0.767, P < 0.0001) than for RDIE (rs = −0.553, P < 0.0001). The case differed during the postveraison run, where WUEi was significantly correlated with gc in neither RDIS nor RDIL [rs = −0.173, P = 0.182 (RDIS); rs = 0.053, P = 0.697 (RDIL)]. During the early additional deficit, the average midday WUEi did not differ between RDIE and RDIS (data not shown). Around midday in RDIE, both NCEV and gc were 59% less than in RDIS, resulting in a constant proportionality. Both regimens exhibited a small decline in WUEi between midmorning and midday. However, the same outcome was not observed during the postveraison measurement run, where RDIL exhibited a large decline in WUEi during midday (data not shown). The average midday WUEi differed between RDIL and RDIS because average NCEV was 85% less in RDIL than in RDIS, whereas midday gc was 75% less in RDIL than in RDIS. Thus the ratio between NCEV and gc changed in response to the additional deficit.
Discussion
Vines under additional water deficit (RDIE, RDIL) transpired less than those under RDIs during the specific periods of additional deficit, as we observed in carbon fixation (Tarara et al., 2011). Because of the duration of the respective additional deficits, cumulative irrigation applied over the course of the season was about 40% and 20% less than RDIS in RDIE and RDIL, respectively. Therefore, under increased demand for water resources, growers may consider the interaction between the effect of various RDI regimens on fruit quality goals and total irrigation water applied. As with the response that we observed in NCEV (Tarara et al., 2011), all vines responded rapidly to irrigation, increasing TrV within 24 h or less. This response is similar to response rates of three cultivars evaluated by Pou et al. (2012). In the present study, the response to soil drying occurred over 2 to 3 d. Sinclair (2005) suggested for a wide range of soils that Tr and NCE, mediated by gc, change little until about two-thirds of the available soil water is transpired, after which NCE and Tr decrease linearly with decreasing θv. From calculations reported in Tarara et al. (2011), during midsummer more than 50% of available water could have been transpired in 2 d.
Under the additional water deficits, the diurnal patterns of TrV differed somewhat from those of NCEv in that rates of TrV approached relatively steady values across midday, whereas NCEV declined steadily from a morning maximum (Tarara et al., 2011). The highest rates of NCEV and the highest rates of change in NCEV occurred earlier in the morning than for TrV. Despite these different diurnal patterns, during the additional deficit periods there were similar relative cumulative differences from RDIS in both TrV,d and NCEV,d. This outcome substantiates the notion that stomatal and not photochemical limitation dominated the response in NCE (Tarara et al., 2011). Flexas et al. (2002) found that under slowly induced water stress in field-grown vines, as gS decreased to values of about 50 mmol·m−2·s−1, NCE decreased progressively with almost no change in photosynthetic efficiency. Below this approximate threshold, nonstomatal limitations may be dominant (Medrano et al., 2002), but in our case this situation would not have occurred extensively. Ranges of gS in Vitis sp. have been associated in a general sense with mild water stress (decreasing from about 200–500 mmol·m−2·s−1 down to 150 mmol·m−2·s−1), moderate water stress (about 50–150 mmol·m−2·s−1), and severe water stress (<50 mmol·m−2·s−1; Lovisolo et al., 2010). Analogous ranges for gc have not been proposed so we take available values of gS as a starting reference. Based on gc we observed at least transient occurrences of severe water stress in RDIE and RDIL; RDIS consistently would have been classified as under moderate water stress. All vines were under at least moderate water stress for the entire period of fruit set to harvest.
The daily maximum TrV,LA values presented here are comparable with others in the literature that evaluated water-stressed vines using whole-canopy gas exchange [e.g., ≈5 mmol·m−2·s−1, field-grown (Katerji et al., 1994); ≈3 mmol·m−2·s−1, field-grown (Ollat and Tandonnet, 1999); ≈3 mmol·m−2·s−1, potted (Palliotti et al., 2014); ≈2–3 mmol·m−2·s−1, potted (Poni et al., 2009)]. The lower values of gc that we observed under the additional deficits were in the general range reported for single-leaf measurements (gS) of water-stressed potted vines by Palliotti et al. (2014), but much lower than those reported for field-grown, deficit-irrigated vines elsewhere (de Souza et al., 2003). In context, we found gS to overestimate gc. Under progressive soil water deficit in other field-grown vines, midday gS declined to values similar to those that we observed (de Souza et al., 2005). Chaves et al. (2007) measured midday gS under water deficit as high as 230 mmol·m−2·s−1 in 2 years, but reported a value as low as 80 mmol·m−2·s−1 in the third year of the study, the driest year.
One indicator of vine water stress that would have been valuable to include, for example, are measurements of midday stem water potential. Unfortunately, stem water potential measurements that had been recorded by the commercial grower-cooperator could not be included in this data set because they had not been recorded at a consistent time of day and thus were not comparable over time. Future work would clearly benefit from such measurements, with the caveat that they are single-leaf measurements and subject to the same sampling bias of single-leaf measurements of gas exchange. Furthermore, a limited number of samples can be collected during the short midday window. Nonetheless, an independent indicator of vine water stress is useful in context.
There may be other physiological implications from the deficit regimens. Under water stress that was indicated by both photosynthesis and transpiration behavior, the additional water deficits could have generated more extensive hydraulic and chemical signals than RDIS, for example root water transport mediated by aquaporins; reduced root water potential; increased xylem sap pH; and/or increased abscisic acid (ABA) concentrations in the leaf, xylem, or root, all implicated in the control of stomatal aperture. However, there is yet to be consensus on the relative importance of each factor (see reviews by Chaves et al., 2010; Lovisolo et al., 2010; Schultz and Stoll, 2010), not least because of observed differences in responses among cultivars, for example to ABA (e.g., Beis and Patakas, 2010; Rogiers et al., 2011a; Tramontini et al., 2014). In field-grown ‘Cabernet Sauvignon’ (Speirs et al., 2013), gene expression for ABA synthesis was highest under their more severe deficit irrigation regimen. In potted ‘Cabernet Sauvignon’ (Tramontini et al., 2014), gS was strongly related to leaf ABA.
Under very high gc (little if any water stress), TrV is proportional to gc and the coupling of the leaf or canopy to the atmosphere is strong (Jones, 1990). For example, in well-watered grapevines in the field, gc was mostly a function of VPD once stomates fully opened in response to solar radiation (in their case >200 W·m−2; Lu et al., 2003). In a sparse crop like grapevine, one would expect the plants to be well coupled to the atmosphere, but this coupling can be modified by water stress. For example, Chaves et al. (2002) proposed that under progressive water stress in field-grown vines, gS declines beginning in midmorning and for increasingly longer portions of the day as stress intensifies. We observed that gc reached a daily maximum value in the morning, generally earlier for the vines under the additional deficit than for those under RDIS, indicating that under more severe water deficits TrV may be less coupled to the atmosphere (i.e., VPD). Rogiers et al. (2011b) found that the magnitude of the response of Tr to VPD was smaller under dry soils (lower gS) than under well-watered conditions. Responses are not necessarily cultivar-specific. In ‘Cabernet Sauvignon’ (Collins and Loveys, 2010), which is thought of as displaying anisohydric behavior (delayed stomatal closure), longer term and increasing water deficits resulted in higher sensitivity of gc to VPD, eliciting what are considered isohydric responses (earlier stomatal closure). In other water stressed vines, Tr became limited as VPD increased (Patakas et al., 2005), leading to the proposition that under water stress, stomates are more sensitive to VPD. Our data from the additional water deficits appear to support these observations.
We found contrasting (clockwise vs. counterclockwise) hysteretic loops in the relationship between TrV and VPD, and between TrV and PPF. Zhang et al. (2014) proposed that the phase angle difference, or kinetic lag, between radiation and VPD prompts asymmetric responses of Tr to VPD and of Tr to net radiation (solar radiation as proxy). Because Tr is more in phase with solar radiation than with VPD, the hysteretic loop between Tr and radiation is smaller than that between Tr and VPD. The same was the case under both RDIS and the additional deficits. The hysteretic responses can be modulated by abiotic factors like soil water potentials and by biotic factors like leaf water potential, which are particularly relevant under water deficit. That Tr can be less correlated with VPD under low θv is evident in the compressed hysteretic loops of the additional deficits. As in the grapevines in this study, in trees growing under water stress (Zeppel et al., 2004) there was a clockwise hysteretic pattern between Tr and VPD, meaning that for a given VPD, more water was transpired in the morning than in the afternoon. They reported a counterclockwise hysteretic pattern between Tr and solar radiation, meaning that for a given value of solar radiation, less water was transpired in the morning than in the afternoon, as we also observed. The same patterns in Tr were reported for Ziziphus jujuba (Chen et al., 2014), a shrubland meadow (Zheng et al., 2014), and for prairie forbs and Zea mays (Matteos-Remigio, 2015). Postulated abiotic and biotic causes of the phenomenon include an increase in stomatal sensitivity to VPD in the afternoon, thereby prompting lower gc; decreasing soil water potential, leading to decreased hydraulic conductance; and reduced leaf water potential in the afternoon, also eliciting reduced gc (Zeppel et al., 2004).
That our values of gS and TrSL were higher than gc and TrV,LA is substantiated elsewhere by experiments in which the grapevine canopy was divided into sections: interior, exterior, and aspect of the vine row. Both Tr and NCE varied with leaf position in the canopy and cardinal direction of the canopy section, driven by radiation interception (Escalona et al., 2003). Drought progressively affected leaves from the sunlit exterior to the shaded interior. They found that severe water stress reduced NCE and Tr in all locations of the canopy except the innermost shaded leaves. Medrano et al. (2012) confirmed these observations for instantaneous rates of Tr and NCE (expressed as WUE and WUEi) as functions of incident solar radiation at a specific section of the canopy. One known limitation of leaf-level measurements is the propensity for selecting sunlit, fully-exposed leaves on the canopy exterior. Difficulties in “scaling up” leaf level measurements to canopy or larger scales are ubiquitous both empirically (Medrano et al., 2012; Merli et al., 2015; Poni et al., 2009; Tomás et al., 2014) and in the modeling community (e.g., Launiainen et al., 2011).
An optimization theory of stomatal aperture (e.g., Cowan, 1982) predicts that daytime variation in gS minimizes Tr while maximizing NCE; changes in gS theoretically maintain a constant ratio of ΔTr:ΔNCE, “conserving” so to speak, WUE (Bacon, 2004). We found that the relative decline in TrV under the additional deficits was similar to the relative decline in NCEV, resulting in no differences in WUEd among RDI regimens during the respective periods of additional deficit. This result is in contrast to other findings where the decline in net photosynthesis under water deficit was less than that of Tr [leaf-level (Bowen et al., 2011)], indicating an increase in WUE. In other leaf-level measurements (Pou et al., 2008), WUE increased with increasing water stress but became more independent of VPD. In whole-vine measurements (Poni et al., 2014), deficit irrigation limited NCEV more than TrV and thus WUE was dramatically reduced during the second half of a deficit period.
There is a general hypothesis that TrV is more sensitive to changes in gc than is NCEV (e.g., Jones, 2014) because of differences in [H2Ov] and [CO2] gradients, respectively. However, water deficits potentially alter this relationship. For example, if leaf temperature diverges from air temperature (leaf > air), which is plausible under water deficit, TrV may not be proportional to gc, resulting in a change in gc not necessarily leading to a change in WUE (Jones, 2004). For fully sunlit leaves in field-grown vines (Medrano et al., 2012), WUE did not differ between moderate and severe water stress for instantaneous midday WUE or WUEd. They postulated that stomatal closure during water stress resulted in increased leaf temperature which in turn increased VPD, counteracting the effect of reduced gS on Tr. Similar to our observations, WUE decreased during the day (Merli et al., 2015) with a maximum limitation around solar noon, under a caveat that their vines received a second daily irrigation application in midafternoon. Apparent inconsistencies in the literature on WUE under water deficits undoubtedly are compounded by the dependence of WUE on cultivar and on the environmental conditions under which the vines are grown (Tomás et al., 2014).
The use of WUEi (NCE/g) allows WUE (NCE/Tr) to be normalized to g, allowing the near exclusion of the confounding effect of VPD on Tr (Schultz and Stoll, 2010). Most studies of grapevine NCE report WUEi, undoubtedly a consequence of an emphasis on leaf-level measurements of photosynthesis. Chaves et al. (2010) mention the importance of tracking WUEi throughout the day to adequately estimate an integrated WUEd, although this integration is difficult (Schultz and Stoll, 2010). However, physiologically it can be argued that the ratio of interest is NCE:Tr, on short- and long-term bases (e.g., Merli et al., 2015). Furthermore, measurements conducted under high VPD and water deficit, as in this study, may show opposite tendencies in WUE and WUEi: VPD negatively affects WUE over wide range of plant water status but there is no obvious relationship between VPD and WUEi (Schultz and Stoll, 2010). This study was conducted in a semiarid climate with high evaporative demand: in midseason, we calculated maximum daily VPD up to about 5.5 kPa.
Conclusion
The daily cumulative relative decline in TrV under additional water deficits was similar to that of NCEV, resulting in no differences in WUEd among RDI regimens during the respective periods of additional deficit. In the case of single-leaf measurements, relative differences among treatments may support that conclusion. However, whole-canopy measurements provide a powerful tool for identifying integrated absolute differences in NCE and Tr that are free from the scaling-up uncertainties of leaf-level measurements. In terms of irrigation management, there was greater water savings in RDIE than in RDIL because the majority of water that was transpired during the season occurred between fruit set and veraison. The previous report of this study showed that nonetheless there was no apparent loss of yield or compromise in fruit quality.
Literature Cited
Allen, R.G., Pereira, L.S., Raes, D. & Smith, M. 1998 Crop evapotranspiration. FAO Irr. Drainage Paper No. 56. United Nations Food and Agriculture Organization, Rome, Italy
Bacon, M.A. 2004 Water use efficiency in plant biology, p. 1–26. In: M.A. Bacon (ed.). Water use efficiency in plant biology. CRC Press, Boca Raton, FL
Behboudian, M.H. & Singh, Z. 2001 Water relations and irrigation scheduling in grapevine Hort. Rev. 27 189 224
Beis, A. & Patakas, A. 2010 Differences in stomatal responses and root to shoot signalling between two grapevine varieties subjected to drought Funct. Plant Biol. 37 139 146
Bowen, P., Boganoff, C., Usher, K., Estergaard, B. & Watson, M. 2011 Effects of irrigation and crop load on leaf gas exchange and fruit composition in red wine grapes grown on a loamy sand Amer. J. Enol. Viticult. 62 9 22
Campbell, G.S. & Norman, J.M. 1998 An introduction to environmental biophysics. 2nd ed. Springer-Verlag, New York, NY
Castellarin, S.D., Matthews, M.A., Di Gaspero, G. & Gambetta, G.A. 2007 Water deficits accelerate ripening and induce changes in gene expression regulating flavonoid biosynthesis in grape berries Planta 227 101 112
Chaves, M.M., Pereira, J.S., Maroco, J., Rodrigues, M.L., Ricardo, C.P.P., Osório, M.L., Carvalho, I., Faria, T. & Pinheiro, C. 2002 How plants cope with water stress in the field. Photosynthesis and growth Ann. Bot. (Lond.) 89 907 916
Chaves, M.M., Santos, T.P., Souza, C.R., Ortuño, M.F., Rodrigues, M.L., Lopes, C.M., Maroco, J.P. & Pereira, J.S. 2007 Deficit irrigation in grapevine improves water-use efficiency while controlling vigor and production quality Ann. Appl. Biol. 150 237 252
Chaves, M.M., Zarrouk, O., Francisco, R., Costa, J.M., Santos, T., Regalado, A.P., Rodrigues, M.L. & Lopes, C.M. 2010 Grapevine under deficit irrigation: Hints from physiological and molecular data Ann. Bot. (Lond.) 105 661 676
Chen, D., Wang, Y., Liu, S., Wei, X. & Wang, X. 2014 Response of relative sap flow to meteorological factors under different soil moisture conditions in rainfed jujube (Ziziphus jujuba Mill.) plantations in semiarid northwest China Agr. Water Mgt. 136 23 33
Collins, M. & Loveys, B. 2010 Optimizing irrigation for different cultivars. 2010. Final research report to Grape and Wine Research & Development Corporation. GWRDC Project no. CSP 05/02. Grape and Wine Research & Development Corp., Adelaide, Australia
Cowan, I.R. 1982 Regulation of water use in relation to carbon gain in higher plants, p. 589–613. In: O.L. Lange and J.D. Bewley (eds.). Encyclopedia of plant physiology. Vol. 12. Springer, New York, NY
de Souza, C.R., Maroco, J.P., dos Santos, T.P., Rodrigues, M.L., Lopes, C.M., Pereira, J.S. & Chaves, M.M. 2003 Partial rootzone drying: Regulation of stomatal aperture and carbon assimilation in field-grown grapevines (Vitis vinifera cv. Moscatel) Funct. Plant Biol. 30 653 662
de Souza, C.R., Maroco, J.P., dos Santos, T.P., Rodrigues, M.L., Lopes, C.M., Pereira, J.S. & Chaves, M.M. 2005 Control of stomatal aperture and carbon uptake by deficit irrigation in two grapevine cultivars Agr. Ecosyst. Environ. 106 261 274
Escalona, J.M., Flexas, J., Bota, J. & Medrano, H. 2003 Distribution of leaf photosynthesis and transpiration within grapevine canopies under different drought conditions Vitis 42 57 64
Evans, R.G., Spayd, S.E., Wample, R.L., Kroeger, M.W. & Mahan, M.O. 1993 Water use of Vitis vinifera grapes in Washington Agr. Water Mgt. 23 109 124
Flexas, J., Bota, J., Escalona, J.M., Sampol, B. & Medrano, H. 2002 Effects of drought on photosynthesis in grapevines under field conditions: An evaluation of stomatal and mesophyll limitations Funct. Plant Biol. 29 461 471
Flexas, J., Galmés, J., Gallé, A., Gulías, J., Pou, A., Ribas-Carbo, M., Tomàs, M. & Medrano, H. 2010 Improving water use efficiency in grapevines: Potential physiological targets for biotechnological improvement Austral. J. Grape Wine Res. 16 106 121
Intrigliolo, D.S., Pérez, D., Risco, D., Yeves, A. & Castel, J.R. 2012 Yield components and grape composition responses to seasonal water deficits in Tempranillo grapevines Irrig. Sci. 30 339 349
Intergovernmental Panel on Climate Change 2014 Working Group II contribution to the fifth assessment report of the intergovernmental panel on climate change. Cambridge Univ. Press, Cambridge, UK
Jones, H.G. 1990 Physiological aspects of control of water status in horticultural crops HortScience 25 19 26
Jones, H.G. 2004 What is water use efficiency? p. 27–41. In: M.A. Bacon (ed.). Water use efficiency in plant biology. CRC Press, Boca Raton, FL
Jones, H.G. 2014 Plants and microclimate. 3rd ed. Cambridge Univ. Press, Cambridge, UK
Katerji, N., Daudet, F.A., Carbonneau, A. & Ollat, N. 1994 Etude à l’échelle de la plante entière du fonctionnement hydrique et photosynthétique de la vigne: Comparaison des systèmes de conduite traditionnel et en Lyre Vitis 33 197 203
Launiainen, S., Katul, G.G., Kolari, P., Vesala, T. & Hari, P. 2011 Empirical and optimal stomatal controls on leaf and ecosystem level CO2 and H2O exchange rates Agr. For. Meteorol. 151 1672 1689
Loveys, B.R., Stoll, M. & Davies, W.J. 2004 Physiological approaches to enhance water use efficiency in agriculture: Exploiting plant signalling in novel irrigation practice, p. 113–141. In: M.A. Bacon (ed.). Water use in plant biology. CRC Press, Boca Raton, FL
Lovisolo, C., Perrone, I., Carra, A., Ferrandino, A., Flexas, J., Medrano, H. & Schubert, A. 2010 Drought-induced changes in development and function of grapevine (Vitis spp.) organs and in their hydraulic and non-hydraulic interactions at the whole-plant level: A physiological and molecular update Funct. Plant Biol. 37 98 116
Lu, P., Yunusa, I.A.M., Walker, R.R. & Müller, W.J. 2003 Regulation of canopy conductance and transpiration and their modelling in irrigated grapevines Funct. Plant Biol. 30 689 698
Matteos-Remigio, V.S. 2015 Assessing the ecohydrological impact of incorporating perennial vegetation into an agricultural watershed in central Iowa, USA. PhD Diss. Iowa State Univ., Ames
Medrano, H., Escalona, J.M., Bota, J., Gulías, J. & Flexas, J. 2002 Regulation of photosynthesis of C3 plants in response to progressive drought: Stomatal conductance as a reference parameter Ann. Bot. (Lond.) 89 895 905
Medrano, H., Pou, A., Tomás, M., Martorell, S., Gulias, J., Flexas, J. & Escalona, J.M. 2012 Average daily light interception determines leaf water use efficiency among different canopy locations in grapevine Agr. Water Mgt. 114 4 10
Merli, M.C., Gatti, M., Galbignani, M., Bernizzoni, F., Magnanini, E. & Poni, S. 2015 Water use efficiency in Sangiovese grapes (Vitis vinifera L.) subjected to water stress before veraison: Different levels of assessment lead to different conclusions Funct. Plant Biol. 42 198 208
Ollat, N. & Tandonnet, J.P. 1999 Effects of two levels of water supply on the photosynthesis and the transpiration of a whole grapevine (Vitis vinifera, cv. Cabernet Sauvignon) Acta Hort. 493 197 204
Osmond, C.B., Björkman, O. & Anderson, D.J. 1980 Physiological processes in plant ecology. Springer-Verlag, New York, NY
Palliotti, A., Tombesi, S., Frioni, T., Famiani, F., Sivestroni, O., Zamboni, M. & Poni, S. 2014 Morpho-structural and physiological response of container-grown Sangiovese and Montepulciano cvv. (Vitis vinifera) to re-watering after a pre-veraison limiting water deficit Funct. Plant Biol. 41 634 647
Patakas, A., Noitsakis, B. & Chouzouri, A. 2005 Optimization of irrigation water use in grapevines using the relationship between transpiration and plant water status Agr. Ecosyst. Environ. 106 253 259
Perez Peña, J. & Tarara, J. 2004 A portable whole canopy gas exchange system for several mature field-grown grapevines Vitis 43 7 14
Poni, S., Bernizzoni, F., Civardi, S., Gatti, M., Porro, D. & Camin, F. 2009 Performance and water-use efficiency (single-leaf vs. whole-canopy) of well-watered and half-stressed split-root Lambrusco grapevines grown in Po valley (Italy) Agr. Ecosyst. Environ. 129 97 106
Poni, S., Merli, M.C., Magnanini, E., Galbignani, M., Bernizzoni, F., Vercesi, A. & Gatti, M. 2014 An improved multichamber gas exchange system for determining whole-canopy water-use efficiency in grapevine Amer. J. Enol. Viticult. 65 268 276
Pou, A., Flexas, J., del March Alsina, M., Bota, J., Carambula, C., de Herralde, F., Galmés, J., Lovisolo, C., Jiménez, M., Ribas-Carbó, M., Rusjan, D., Secchi, F., Tomás, M., Zsófi, Z. & Medrano, H. 2008 Adjustments of water use efficiency by stomatal regulation during drought and recovery in the drought-adapted Vitis hybrid Richter-110 (V. berliandi × V. rupestris) Physiol. Plant. 134 313 323
Pou, A., Medrano, H., Tomás, M., Martorell, S., Ribas-Carbó, M. & Fexas, J. 2012 Anisohydric behaviour in grapevines results in better performance under moderate water stress and recovery than isohydric behaviour Plant Soil 359 335 349
Rogiers, S.Y., Greer, D.H., Hatfield, J.M., Hutton, R.J., Clarke, S.J., Hutchinson, P.A. & Somers, A. 2011a Stomatal response of an anisohydric grapevine cultivar to evaporative demand, available soil moisture, and abscisic acid Tree Physiol. 32 249 261
Rogiers, S.Y., Greer, D.H., Hutton, R.J. & Clarke, S.J. 2011b Transpiration efficiency of the grapevine cv. Semillon is tied to VPD in warm climates Ann. Appl. Bot. 158 106 114
Schreiner, R.P., Tarara, J.M. & Smithyman, R.P. 2007 Deficit irrigation promotes arbuscular colonization of fine roots by mycorrhizal fungi in grapevines (Vitis vinifera L.) in an arid climate Mycorrhiza 17 551 562
Schultz, H.R. & Stoll, M. 2010 Some critical issues in environmental physiology of grapevines: Future challenges and current limitations Austral. J. Grape Wine Res. 16 4 24
Sinclair, T.R. 2005 Theoretical analysis of soil and plant traits influencing daily plant water flux on drying soils Agron. J. 97 1148 1152
Speirs, J., Binney, A., Collins, M., Edwards, E. & Loveys, B. 2013 Expression of ABA synthesis and metabolism under different irrigation strategies and atmospheric VPDs is associated with stomatal conductance in grapevine (Vitis vinifera L. cv. Cabernet Sauvignon) J. Expt. Bot. 64 1907 1916
Stevens, R.M. & Douglas, T. 1994 Distribution of grapevine roots and salt under drip and full-ground cover microjet irrigation systems Irrig. Sci. 15 147 152
Tarara, J.M., Perez Peña, J.E., Keller, M., Schreiner, R.P. & Smithyman, R.P. 2011 Net carbon exchange in grapevine canopies responds rapidly to timing and extent of regulated deficit irrigation Funct. Plant Biol. 38 386 400
Tomás, M., Medrano, H., Escalona, J.M., Martorell, S., Pou, A., Ribas-Carbó, M. & Flexas, J. 2014 Variability of water use efficiency in grapevines Environ. Expt. Bot. 103 148 157
Tramontini, S., Döring, J., Vitali, M., Ferrandino, A., Stoll, M. & Lovisolo, C. 2014 Soil water-holding capacity mediates hydraulic and hormonal signals of near-isohydric and near-anisohydric Vitis cultivars in potted grapevines Funct. Plant Biol. 41 1119 1128
U.S. Department of Agriculture 2015 Web Soil Survey. 6 June 2015. <http://websoilsurvey.sc.egov.usda.gov/App/WebSoilSurvey.aspx>
von Caemmerer, S. & Farquhar, G.D. 1981 Some relationships between the biochemistry of photosynthesis and the gas exchange of leaves Planta 153 376 387
Washington State University 2015 AgWeatherNet. 6 June 2015. <http://weather.wsu.edu/awn.php?page=dailydata>
Zeppel, M.J.B., Murray, B.R., Barton, C. & Eamus, D. 2004 Seasonal responses of xylem sap velocity to VPD and solar radiation during drought in a stand of native trees in temperate Australia Funct. Plant Biol. 31 461 470
Zhang, Q., Manzoni, S., Katul, G., Pororato, A. & Yang, D. 2014 The hysteretic evapotranspiration—vapor pressure deficit relation J. Geophys. Res. Biogeosci. 119 125 140
Zheng, H., Wang, Q., Zhu, X., Li, Y. & Yu, G. 2014 Hysteresis responses of evapotranspiration to meteorological factors at diel timescale: Patterns and causes. PLoS One. 9:1–10 (e98857)