Abstract
A multichamber whole-canopy gas exchange system coupled with an automatic pot weighing device was tested for continuous 24 hr recording over 50 days in a trial comparing cv. Sangiovese vines subjected to progressive reduction of total transpiration water supply to well-watered vines. The system ran smoothly under regular maintenance for the entire period and gravimetric vine water loss was highly correlated with chamber-derived vine transpiration (r = 0.95) for data pooled over treatments. Seasonal and diurnal whole-canopy net CO2 exchange rate (NCER) and transpiration (Tc) showed that supplying 50% and 30% of daily gravimetric vine water loss (Tg) consistently corresponded to a NCER more than proportionally limited as compared to Tc, hence leading to lower canopy water-use efficiency (WUE) expressed as NCER/Tc ratio. Conversely, canopy WUE did not differ between treatments at 70% Tg restitution and rewatering. Similarly, during the most limiting water supply periods, the WUE difference between treatments was greatly reduced during cloudy days with lower vapor pressure deficit and higher diffuse-to-direct light intensity ratio. Data sets taken over different time frames on whole-canopy WUE provide a scenario different from that which might derive from traditional single-leaf assessment, reporting in almost all cases that intrinsic WUE increases under stress and suggests that the methodology used can mask or alter conclusions about adaptive response of grapevine cultivars to water stress.
Since the pioneering application of an assimilation chamber to study gas exchange of a whole apple tree in the field (Heinicke and Childers 1936), several studies have used custom-built, whole-plant enclosure systems to monitor gas exchange in fruit trees for variable time lengths (Buwalda et al. 1992, Corelli-Grappadelli and Magnanini 1993, Poni et al. 1997, Perez Peña and Tarara 2004, Petrie et al. 2009, Tarara et al. 2011). Overall, while the flow-through systems have almost entirely replaced the closed (i.e., transient) systems, equipment configuration still varies notably in chamber design (e.g., rigid versus flexible, closed versus open top), ease of assembly, dismantling and transportability, number of canopies that can be simultaneously monitored, chamber material, method for measurement of air flow through the chamber, degree of automation in data checking, logging, and transfer, need of surveillance and periodic maintenance, and cost of single components of the whole apparatus.
As outlined elsewhere, regardless of specific technical features any “enclosure” system unavoidably causes some degree of alteration of the microclimate surrounding the canopy and such change needs to be minimized (Garcia et al. 1990). Major considerations are chamber shapes that allow uniform air mixing while avoiding formation of pockets of still air, the most stable inlet (reference) CO2 concentration, which requires air intake to be placed a few meters aboveground and the use of a large buffering container before air enters the system, and, most importantly, calibrated air flow. The air flow entering and leaving the chamber impacts directly on the number of chamber volume exchange per unit of time, which, according to Perez Peña and Tarara (2004), should not be less than one exchange per 30 sec to keep overheating within ~3°C above ambient. On the other hand, depending on vine size and type of study to be conducted, air flow cannot be raised ad libitum, as the inlet-outlet (CO2) would progressively decrease to nil detection. In previous work, raising air flow from 30 to 47 L/sec in field-chambered vines with a leaf area of 14 m2 reduced the outlet-inlet temperature differential from 5.0 to 2.1°C (Poni et al. 1999). Thus, a good compromise between thermal regulation and detection of an adequate gas-exchange differential seems to be reached when ~3 to 4 L/sec · m2 of leaf area is fed to the chambers.
In grapevine, the use of “whole-vine” enclosure systems have led to novel findings that involve effects of training system and pruning regime (Petrie et al. 2009, Poni et al. 2000), leaf removal and shoot thinning (Intrieri et al. 1997, Petrie et al. 2003, Poni et al. 2008, Bernizzoni et al. 2012), row orientation (Intrieri et al. 1998), regulated deficit irrigation (Poni et al. 2009, Tarara et al. 2011), response to light intensity and interception (Petrie et al. 2009, Poni et al. 2003), and model calibration (Poni et al. 2006). When available, these systems can also continuously monitor the concentration of inlet and outlet water-vapor pressure, thereby allowing measurements of whole-canopy transpiration (Tc) and, hence, whole-canopy water-use efficiency (WUE) given as net CO2 exchange rate (NCER)/Tc. The issue of calibration of whole-canopy transpiration rates versus other methods has primarily relied upon comparisons with heat-balance sap flow (Dragoni et al. 2006) and heat-pulse velocity (Petrie et al. 2009) readings, which yielded very close correlations (R2 = 0.88 to 0.98 across the two studies). In trunk sap flow velocity measurements with and without chambers, the physical presence of the chamber did not affect sap flow rates, suggesting its minimal effect on vine microclimate and resulting transpiration (Petrie et al. 2009).
The above results indicate that WUE, a key parameter for assessing adaptation of different genotypes to water deficit, can be reliably measured on a whole-canopy basis and compared to other techniques, including single-leaf intrinsic and extrinsic WUE (Schultz and Stoll 2010), isotopic composition (13C/12C), or biomass accumulation per unit of water used (Gibberd et al. 2001). Indeed, an unresolved issue seems to be how cultivars classified as iso- or anisohydric modify their WUE under either well-watered or water deficit conditions. For instance, Lovisolo et al. (2010), pooling data of 27 references reporting midday leaf water potential, stomatal conductance (gs), and intrinsic WUE (A/gs) for isohydric or anisohydric grapevine species or cultivars fully irrigated or subjected to some degree of water stress, showed that the two categories present very similar A/gs under both irrigation and water stress and concluded there was no clear pattern of correlation between WUE and iso- or anisohydric character.
The goals here were to describe improvements and novel technologies used in a multichamber gas exchange system tested on grapevine canopies, to assess the degree of correlation between water loss determined gravimetrically versus whole vine transpiration, to determine if chamber presence significantly affects actual vine transpiration rates, and to provide data for seasonal and diurnal canopy WUE during a dry-down experiment and a follow-up rewatering.
Materials and Methods
Gas exchange apparatus.
The main components of the system (a configuration similar to the version reported in Poni et al. 1997) features alternating current, centrifugal blowers (Vorticent C25/2M Vortice, Milan, Italy) delivering a maximum air flow of 950 m3/hr, flexible plastic polyethylene chambers allowing 88% light transmission, 6% diffuse light enrichment and no alteration of the light spectrum, a CIRAS-SC single-channel absolute CO2/H2O infrared gas analyzer (PP-Systems, Amesbury, MA), and a CR1000 datalogger wired to an AM16/32B Multiplexer (Campbell Scientific, Logan, UT) (Figure 1). Air sampling is switched from one chamber to another at programmed time intervals using a set of solenoid valves (model 177 2/2, normally closed type; Sirai, Bussero, Italy).
Air is drawn ~3 m from the soil surface and is directed by the blowers to the vines through 100 mm diameter rigid, white plastic pipes (straight segments, T-pieces, curves). To facilitate air mixing and ensure higher stability in inlet (CO2), air is forced through a buffer tank (500 L capacity) before it is directed to the chambers. The air-flow rate to each chamber is controlled by a butterfly valve and measured, at least 1 m downstream from the valve itself, with a Testo 510 digital manometer (Farnell, Lainate, Italy) according to a flow-restriction method (Osborne 1977). Flow restriction was achieved by inserting a 44 mm inner-diameter ring in the 100 mm pipe so that a pressure drop of 10 mm water column measured upstream and downstream from the ring corresponded to 12.01 L/sec.
A major improvement is that the system can now simultaneously monitor up to 12 chambers, thereby solving the issue of an adequate number of replicates, which former versions could not provide (Corelli-Grappadelli and Magnanini 2003). A major limitation of the increased number of chambers, in addition to cost, design, and assembly complexity, was the increase in both tube length and the time required to flush the tubes upon switching to reach a steady state. This can result in long intervals before resampling in the same chamber and, hence, spoiling the necessary consistent diurnal trends. Thus, an additional rotary vane pump (model G-01-K-EB; Gardner Denver Thomas, Pucheim, Germany) with an air flow rate of 1.4 L/min was mounted to speed tube flushing while resolving the issue of water condensation within the tubing during the first hours in the morning or during rainy days. The 90-sec switching interval of the 12-chamber system allows for excellent stability of the recorded values.
In the current basic configuration, ambient (inlet) air temperature and the air temperature at each chamber’s outlet were measured by shielded external/internal (1/0.2 mm) diameter PFA–Teflon insulated type-T thermocouples (Omega Engineering, Stamford, CT), whereas direct and diffuse radiation were measured with a BF2 sunshine sensor (Delta-T Devices, Cambridge, UK) placed horizontally on top of a support stake next to the chambers that enclosed the canopies.
An automatic pot weighing system was embedded into the system (Figure 1), with each pot placed on a single-cell platform scale (ABC Bilance, Campogalliano, Italy), 400 × 400 × 140 mm (length × width × height), having a maximum load of 150 kg and 50 g accuracy. Each scale was wired to the CR1000 datalogger and instantaneous pot weight was measured concurrently with environmental and gas-exchange parameters. Pots placed on each scale were equipped with underpots to avoid free drainage and minimize errors in calculating actual amount of water loss. The system was connected to an external laptop computer configured with Team Viewer software (TeamViewer GmbH, Uhingen, Germany) to allow remote access for data checking and downloading from any location using the datalogger support software package LoggerNet (Campbell Scientific), which diminished the need for attendance.
Plant material and treatment layout.
Data examples for system testing were derived from a broader preveraison water-stress experiment conducted in 2013 on three-year-old cane-pruned Montepulciano and Sangiovese (Vitis vinifera L.) grafted on SO4. Canes were 1 m long, each containing 8 to 10 dormant buds each. Twelve vines, six per cultivar, were arranged along a single, vertical shoot-positioned, 35° northeast-southwest-oriented row and randomly assigned to a well-watered (WW) and a water-stressed (WS) treatment. For the purpose of this article, only data taken on the six Sangiovese vines were used.
All six Sangiovese vines were kept well watered until DOY 170 (19 June, pea-size stage) by supplying a daily amount of 4 L/vine, which was slightly in excess of actual vine transpiration (Tg) detected by the weighing system over the five days prior to the start of reduction estimated at 3.806 ± 0.133 L/vine. Starting on DOY 171, a progressive water deficit was imposed on half of the vines by reducing water supply to 70% Tg (2.8 L/day delivered for WS and 4 L day for WW) until DOY 179. On DOY 180 (29 June), supply was reduced to 50% Tg (2 L/day for WS and 4 L/day for WW) until DOY 190 and then raised to 2.5 L/day for WW and 5 L/day on DOY 191 and 192. The increase for 100% Tg from 4 to 5 L/vine was due to canopy growth and increasing air vapor pressure deficit (VPD) as midsummer approached. Beginning on DOY 193, maximum stress as 30% Tg was applied (1.5 L/day WS and 5 L/day for WW) until rewatering on DOY 198 (17 July) to 100% Tg. No sign of berry pigmentation was seen at rewatering. To ensure maximum accuracy, fractional water restitution was manually applied twice a day. During water stress, pot surface in both treatments was covered with a plastic sheet to minimize losses due to soil evaporation.
Progression of water stress was monitored by measuring predawn leaf water potential (ψpd) on DOY 169, 179, 186, 189, 197, and 206. Measurements were taken before sunrise on three leaves per vine using a Scholander pressure chamber (model 3500; Soil Moisture Equipment, Santa Barbara, CA).
Gas exchange measurements.
Whole-canopy net CO2 exchange rate (NCER) and transpiration (Tc) measurements were taken using the new multichamber system. The chambers were set up on each vine and continuously operated, 24 hr/day, from DOY 157 (6 June, 14 days before stress) until DOY 205 (24 July, 7 days after rewatering). The flow rate fed to the chambers was set at 10.4 L/sec and kept constant throughout the entire measuring season. The polyethylene chambers had a volume of 0.602 m3 ± 0.0579, so a complete volume air change occurred at an interval of ~60 sec. Whole-canopy NCER and Tc (μmol CO2/sec and mmolH2O/sec, respectively) were calculated from flow rates and CO2 and water vapor differentials after Long and Hallgren (1985). Whole-canopy water-use efficiency (WUE) was calculated as NCER/Tc and given as mmolCO2/molH2O. The impossibility of automatically decreasing air flow in the evening prevented accurate measurements of nighttime gas exchange. To gain some estimate of probable effects of the chamber presence on canopy water loss, gravimetric daily vine water use over the five days before dismantling (DOY 202–205) was compared to that recorded over the five days following dismantling (DOY 207–210).
Results
Predawn leaf water potential (ψpd) progressively decreased together with a gradual reduction of the Tg fraction supplied to the vines, reaching the lowest negative value (~−1.0 MPa) toward the end of the 50% Tg restitution period (Figure 2). Further reduction in delivered Tg did not affect ψpd. One week after the rewatering on DOY 198, ψpd of previously stressed vines recovered to the same level of WW pots (−0.2 MPa). Visual observations on DOY 192 (11 July, final day of 50% Tg regime) indicated that during the warmest part of the day leaves tended to display a vertical orientation and some basal leaves showed initial yellowing.
The multichamber system ran smoothly and without major technical glitches for the entire recording period (DOY 157–205), thereby ensuring regular 24-hr data recording (Figure 3). Most of the 49 measuring days were marked by clear skies (i.e., PAR ≥800 μmol/m2 · s) and ambient VPD fluctuating between 1.5 and 2.2 kPa. Lower VPD (≤1.5 kPa) was recorded only a few days before stress and before imposition of the 50% Tg stress level (Figure 3A). Mean seasonal daily averaged values of air temperature recorded at chamber inlet, outlets WW, and outlets WS were 26.9, 28.7 (+1.8), and 29.3 (+2.4)°C, respectively (Figure 4).
Despite day-to-day fluctuations due to climate variability, prestress NCER/canopy rates were very similar among vines allocated to treatments, thus attesting to similar vine size and overall system integrity (Figure 3B). Upon imposition of the different fractions of supplied water, averaged NCER rates per canopy were 70, 46, and 23% of rates measured in WW vines for 70, 50, and 30% Tg, respectively. After rewatering, mean NCER/canopy in WS was 74.8% of rates recorded in WW. Canopy transpiration averaged over the respective length of any water supply level was 72, 55, and 39% of rates recorded in WW at 70, 50, and 30% restitution levels, respectively (Figure 3C). Whole-canopy WUE expressed as NCER/Tc ratio (mmolCO2/molH2O) did not differ significantly between treatments at prestress, 70% Tg restitution, and upon rewatering. Conversely, WUE markedly decreased in WS vines during most of the 50% and 30% Tg restitution periods (Figure 3D).
A close correlation was found between gravimetric daily vine water loss and canopy transpiration calculated via gas exchange for data pooled over treatments (r = 0.95) (Figure 5) after the nonsignificance of regression slopes between treatment was assessed. Equation forced through the origin indicated that the degree of over- or underestimation of the chamber system was negligible. Maximum daily water use per vine reached a threshold of 4.5 L/day for WW, whereas peak values were ~4 L/day for WS. In both treatments, the lowest water use (0.9 L/day) was reached on 9 June (DOY 160), a very cloudy day. Mean Tg and air VPD over the five days before system dismantling were 4.34 ± 0.22 L and 1.98 kPa, respectively, whereas values recorded over the five days after the chambers were removed were 4.84 ± 0.33 and 2.42 kPa, respectively. When evaluated over a shorter integration period (i.e., the days before and after dismantling were both cloudless), water loss per vine was 4.21 and 5.78 L, respectively, against a VPD of 1.98 and 2.68 kPa; thus, calculated vine Tg/VPD was almost identical (2.12 and 2.15 L/kPa, respectively).
Data examples refer to diurnal trends of environmental, gravimetric water loss, and gas exchange parameters of one clear day per trial period (Figure 6). Diurnal NCER trends in WW were fairly smooth and consistently had a peak at ~10:00 hr and a mild, yet constant decline until ~18:00 hr, when NCER was severely limited by incoming radiation (Figure 6B, F, J, N, R). Data taken on DOY 176 (fifth day at 70% water supply) showed that NCER and Tc had mild limitation in the morning hours (Figure 6F, G) and stronger limitation later in the day, despite some recovery after the second water supply at 14:30 hr. Overall, canopy WUE during the day was similar for the two treatments, although WS vines had occasional lower values (Figure 6H).
The level of stress at six days after lowering water supply to 50% Tg was severe: mean daily NCER was only 16.8% of WW, dropped to nil around solar noon, with mild recovery after the afternoon irrigation (Figure 6J). Conversely, canopy transpiration on the same day was less limited (37.8% Tg measured in WW), with a more regular, bell-shaped pattern during the day (Figure 6K). As a result, WUE was much lower in WS during most of the day, with a tendency to reach WW levels when direct PAR was <600 μmol/m2 · sec. Data taken on DOY 195 (second day at 30% restitution) confirmed the overall trends observed on DOY 186, although limitation of both NCER and Tc was milder (Figure 6 N, O) despite a lower water supply to the vines (1.5 L on DOY 195 and 2 L on DOY 186 in 50% fraction). Diurnal trends of canopy WUE (Figure 6P) confirmed that WUE drastically decreased in WS throughout the day, with maximum limitation around solar noon. Rewatering of the WS vines restored NCER and Tc patterns similar to the prestress phase (Figure 6R, S), although mean daily values recorded on DOY 202 were ~75% of WW vines. Noticeably, canopy WUE calculated at rewatering (Figure 6T) no longer showed differences between the two treatments. Overall, diurnal Tg on any day was in good accordance with chamber-derived transpiration, with no major shift or lag between these two parameters (Figure 6C, G, K, O, S).
Discussion
The system ran smoothly under regular maintenance and without major mishaps from DOY 157 to 205 when the chambers were dismantled. Seasonal and diurnal gas exchange parameters trends were fairly clean and consistent, making it possible to detect subtle differences between treatments and to study dynamic and long-term responses to the imposed treatments.
Interfacing an automated weighing tool to the system has two major advantages. First, knowing exact vine water use before applying treatments allows users to conduct detailed water-stress studies. Second, the calibration of chamber-derived transpiration rates is performed against an objective and independent parameter. Previously, little attention was given to calibrate whole-canopy “balloon” transpiration versus other methods. To our knowledge, the only data available correlating gravimetric vine water loss and whole-canopy T were from a study that reported a close linear relationship (r = 0.96) for canopies with ~14 m2 leaf area and fed with an air flow of 40 L/sec (Poni et al. 1999). However, while others have successfully validated chamber-derived vine transpiration rates versus sap flow techniques (Dragoni et al. 2006), one study ascertained that physical chamber presence did not significantly modify absolute amounts of vine water loss (Petrie et al. 2009), in agreement with findings reported here.
Although vine performance data pertaining to this study are not reported here, knowing that final leaf area of the Sangiovese WW vines was 2.87 m2 (data not shown) and that gravimetric vine water loss averaged over a week post-rewatering was 3.88 L, leads to a calculated water use of 1.35 L/m2 · day. For comparison, water use per vine of 0.79 L/m2 · day was reported for cv. Lambrusco grown in the same environment and conditions (Poni et al. 2009); however, whole-canopy gas exchange data were taken only over the first week of September under lower air T and VPD, hence lower evaporative demand, than in the current study. Total daily transpiration per unit of leaf area ranged from 1.0 to 2.3 L/m2 · day for measurements taken on mature Concord vines grown in upstate New York (Dragoni et al. 2006), whereas in the cool, humid climate of Geiseheim (Germany) peak water use of mature Riesling field vines measured with Granier-type sap flow gauges (Braun and Schmid 1999) was ~1 L/m2 · day.
Seasonal trends of canopy NCER, Tc, and WUE (Figure 3) underscore a major issue: under the conditions of this trial, 50% and 30% Tg supply was more limiting to NCER than to Tc and, hence, WUE was drastically reduced in WS vines during the second half of the 50% stage and during the entire 30% Tg water supply level. Establishing an adequate comparison for the long-term (seasonal) response of WUE is not an easy task; several studies containing both whole-canopy-derived NCER and Tc data extended measurements to a maximum of 24 to 48 hr at key phenological stages and, even on a diurnal basis, WUE trends were either not presented or not discussed (Perez Peña and Tarara 2004, Petrie et al. 2009, Tarara et al. 2011).
The fractional reduction in NCER was either very similar to or slightly higher than the fractional constraint in water supply at any stage of stress in the current study. In other words, 70% Tg corresponded to 70% NCER measured prestress, a result that suggests this outcome confers validity to our experimental protocol, which appears to have the capability of predetermining actual vine water use gravimetrically prior to water stress and allows a tighter control of the water stress than can typically be achieved under field conditions. If the initial amount of water delivered to the vines corresponds to their actual water consumption, then it can be assumed that prestress water supply is “optimal”: it allows full leaf function but does not reach excessive water supply, a common field occurrence. Paradoxically, in terms of being able to control stress levels, working in pots has another advantage as basically all the supplied water is made available for root uptake.
A general feature of the overall data set, which also affected WUE response during the most severe water deficit periods of 50% and 30% Tg restitution, is that NCER was proportionally more limited than Tc. This response seems to be linked to both genotype characteristics and specific environmental conditions. Field-grown Sangiovese subjected to multiple summer stresses like water shortage, high temperature, and radiation load typically shows chronic photoinhibition followed by yellowing and necrosis of the basal leaves (Palliotti et al. 2009). Although chlorophyll fluorescence was not measured here, visual canopy assessment performed on 11 July (end of the 50% Tg supply period) confirmed that basal leaf yellowing was already present. Such nonstomatal limitation, together with partial stomatal closure and temperature-driven increase in photorespiration, can account for NCER having been more severely limited than Tc (Palliotti et al. 2009). It also has to be considered that the present study was conducted in the middle of the Po Valley where maximum daily T peaks between 35 and 40°C in summer.
Canopy WUE decreased in WS only under severe stress while it remained unaffected under the mild stress induced by withdrawing 30% of the needed water. Yet closer examination of daily WUE variation during the 50% Tg restitution reveals that WUE, being the constant daily amount of delivered water, had fairly large fluctuations that were not due to random variability. Rather, relatively high WUE of WS vines recorded on DOY 184 and 191 corresponded with partial cloudiness (average daily PAR 577 and 686 μmol/m2 · sec, respectively, and >800 μmol/m2 · sec on the other days), lower VPD, and a higher diffuse-to-direct PAR ratio than the remaining days. It thus seems that, the level of soil water stress being equal, WUE is sensitive to environmental conditions and, especially, to evaporative demand. When daily diffuse-to-direct PAR ratio was plotted against canopy WUE in WW vines, there was a significant positive linear relationship (y = 0.0368x + 2.2004; R2 = 0.55), thereby confirming that higher diffuse light conditions can significantly contribute to raise WUE under severe water stress. This outcome agrees with Petrie et al. (2009), who reported that under diffuse light (cloudy) conditions chambered vines appeared to be more efficient as they photosynthesized at a higher rate per unit of light intercepted. It is possible that, as hazy days have a higher fraction of diffuse light than clear days, they would still saturate photosynthesis at the exterior of the canopy while increasing the photosynthetic contribution of internal, mostly shaded leaves; transpiration is concurrently lowered by the lower evaporative demand and WUE benefits from the balance.
Diurnal trends of gas exchange parameters reported for a representative day within each water regime (Figure 6) add more to the central issue of WUE. Analyses of NCER and Tc trends recorded on DOY 186 and 195 show that while NCER was nil during the central part of the day and had a mild recovery upon early afternoon water supply, Tc proceeded at a more regular pace during the day and was overall less limited. It cannot be ruled out that high temperature and radiation may have sustained transpiration by increasing cuticular conductance of vine organs (Boyer et al. 1997), while having a negative effect on photosynthesis through photoinhibition and increased photorespiration (Palliotti et al. 2009). It might appear odd that the gas exchange limitation recorded on DOY 195 was slightly milder than the reduction observed on DOY 186, despite less water delivered to the vines (1.5 L versus 2 L). However, previous studies have shown that evaporative demand adds to soil water stress in stomatal conditioning (Chaves et al. 2010, Soar et al. 2006). In fact, mean air VPD was markedly lower on DOY 195 (1.65 kPa) than on DOY 186 (2.18 kPa) as was mean air temperature (27.2 and 28.9°C, respectively).
Diurnal trends of canopy WUE strongly support the responses and mechanisms of the seasonal data discussed above. When water shortage became very severe (≤50% of daily vine requirement) canopy WUE in WS vines fell below WW values, with the greatest gap recorded during the central hours of the day. That canopy WUE was closer to WW values in midmorning and late afternoon confirms that the process is sensitive to the diffuse-to-direct PAR ratio. Data examples (Figure 6I, P) together with trends (Figure 3D) suggest that evaluation of the response of a given cultivar to water stress might greatly vary simply depending on the methodology used to assess WUE. This crucial issue has been critically reviewed (Schultz and Stoll 2010), mostly as discrepancies between intrinsic water-use efficiency (WUEi), defined as the ratio between assimilation (A) and leaf stomatal conductance for water vapor (gs), and instantaneous water-use efficiency (WUEinst), calculated as A to leaf transpiration rate (E). Most studies conducted on the effects of water deficits on grapevine leaf gas exchange give preference to the WUEi expression since it largely rules out the effects of changing evaporative demand on water flux out of the leaves. Almost every study has shown that, under moderate-to-heavy water deficit, WUEi increases compared to well-watered conditions (Chaves et al. 2010, Düring 1987, Flexas et al. 2002, Palliotti et al. 2009). These results raise two major issues. First, when both WUEi and WUEinst were measured on the same plants in the same trial, quite often the outcome was inconclusive. For example, in a continuous dry-down field experiment on Shiraz where water deficit reached the maximum level of ψpd = −0.87 MPa after veraison, WUEi was higher for stressed than for irrigated plants throughout most of the day at any level of stress, whereas WUEinst was similar to well-watered vines under moderate stress and actually decreased under more severe water deficit (Schultz 2003). Second, when WUEi and canopy WUE were both measured in the same trial (Poni et al. 2009), the results were again contradictory, as WUEi increased in split-root water-stressed vines whereas canopy WUE was slightly higher in the well-watered vines, especially during the morning hours. These findings match those reported in a detailed study on eight varieties grown in pots to compare WUEi and whole-plant WUE given as annual increment of dry weight per unit of water loss; the authors concluded that WUEi was not a reliable parameter to predict whole-plant WUE (Tomás et al. 2012).
Data here strongly support that canopy WUE markedly decreased under water stress on both a seasonal and a diurnal basis and confirm that opposite conclusions regarding adaptation mechanisms to water stress of a given genotype can be drawn depending on the method used for WUE assessment (Tomás et al. 2014). Indeed, part of the discrepancy may be that single-leaf WUE assessment is typically made on a fairly limited number of well-exposed, healthy leaves, while a whole-canopy approach integrates the contribution of the entire leaf population without modifying their natural position within the canopy.
Conclusions
An improved multichamber system for determining whole-canopy gas-exchange coupled with a new device for automated and concurrent recording of container and plant weight operated successfully in a fully unattended and automated manner for several weeks with no major technical inconveniences or damage to vines. Canopy transpiration obtained under an array of environmental conditions and soil water levels closely correlated with gravimetric determinations and chamber presence did not seem to significantly affect the rates of water exchange between leaf and surrounding air. Most importantly, seasonal and diurnal canopy WUE (NCER/Tc) showed that, especially when conditions of vine water supply of ≤50% occurred together with high evaporative demand, WUE was drastically reduced in stressed vines. This consistent outcome seems to contrast with conclusions that may be driven by measuring single-leaf WUEi and makes a case for further work to develop reliable and consistent methods that objectively assess the response of a given genotype to water deficit. For example, it would be quite informative to run both single-leaf and whole-canopy WUE assessment on the same vines. The pot model used here is certainly biased for not being able to account for effects that can easily occur in the field, such as roots exploring deeper soil layers under water stress. However, coupling an automated weighing system within the apparatus makes it possible to perform targeted water-deficit experiments since prestress, actual transpiration, and any fraction thereof can be accurately measured.
- Received October 2013.
- Revision received December 2013.
- Accepted January 2014.
- Published online December 1969
- ©2014 by the American Society for Enology and Viticulture