Abstract
Effects of reducing irrigation from fruit set to veraison or harvest combined with crop-load adjustment by cluster or shoot thinning were determined for Merlot and Cabernet Sauvignon grapevines cultured on a coarse loamy sand. Geographic information system tools were used to develop maps of moisture distribution in the soil profile, which revealed reductions in total moisture levels and increased spatial variation in response to reduced emitter density. Stomatal conductance and leaf gas exchange decreased in response to reduced irrigation but also declined across all treatments during the lag phase of berry development and then increased postveraison. Pruning mass was affected little by treatments in Merlot but was reduced by either shoot or cluster thinning in Cabernet Sauvignon. Berry mass and anthocyanin and tannin contents were affected little and inconsistently by irrigation and crop-load adjustment and varied mostly among years, indicating a dominant influence of seasonal climate on berry development and composition.
Vineyard soils in the Okanagan Valley, British Columbia, consist mostly of coarse-textured glaciofluvial parent materials deposited during deglaciation in the late Pleistocene (Nasmith 1962, Bowen et al. 2005a). More than 60% of the vineyard blocks in the valley lie on loamy sands classified as Orthic Brown Chernozems belonging to the Osoyoos series (Bowen et al. 2005a). These soils are well drained and hold ~13% moisture by volume at field capacity. The objective to conserve water while maximizing fruit quality for winemaking has spurred interest in allowing mild water stress to occur in vines by reducing soil moisture availability such as prescribed in deficit-irrigation techniques (Hardie and Considine 1976, Behboudian and Singh 2001). The desired effects of water stress are (1) reduced shoot vigor leading to sparser canopies that transpire less water and increase fruit exposure and (2) smaller berries with higher relative amounts of skin and seed that contain phenolics and other extracts that contribute to wine quality (Matthews and Anderson 1989, Matthews et al. 1990, Kennedy et al. 2002, Roby et al. 2004). An additional benefit might be elevated abscisic acid in the fruit, which appears to increase production of phenolics including skin anthocyanins (Peppi and Fidelibus 2008, Owen et al. 2009).
Imposing beneficial levels of water stress to vines on extremely coarse-textured soils is challenging due to the physics of moisture retention within the soil matrix. Most plant-available water is lost to drainage within 1 to 2 days of irrigating (Hillel 2004), and most of that held can be readily taken up by the crop with little concomitant increase in matric tension (Kramer and Boyer 1995). However, depletion of soil moisture to below 10% (v/v) substantially increases matric tensions and the potential for extreme crop stress (Kramer and Boyer 1995). Therefore, without frequent irrigation there is high risk of sudden and intense stress. In the Okanagan Valley, this risk is exacerbated by the warm, sunny climate and low humidity during the growing season.
Reducing the volume of water applied via drip irrigation can be accomplished by reducing the application frequency, the density of emitters, or the volume delivered per emitter, all of which reduce the consistency of water supplied either temporally or spatially. Whereas infrequent irrigation can lead to periods of extreme stress, reducing the volume applied per emitter could shift moisture availability to the soil surface and promote development of a shallow root system vulnerable to surface drying and winter injury. In this study to help maintain the wetted depth in the profile and to avoid periods of whole-profile drying, the irrigation rate was decreased by reducing the emitter density while maintaining the frequency and volume delivered per emitter. To visualize the spatial effect of this treatment, the resulting distribution of moisture in the soil profile in one experiment was mapped through geographic information system (GIS) techniques.
Crop adjustment by thinning clusters or shoots can affect fruit composition including phenolic levels, particularly when the treatment advances fruit maturity (see Reynolds et al. 1994, 2005, Dokoozlian and Hirschfelt 1995, Guidoni et al. 2002). Crop adjustment could also affect grapevine response to soil water supply, as higher crop levels increase the requirements for sugar, leaf gas exchange, and soil moisture. Few studies have examined the interactive effects of water supply and vine crop load. Crop level exacerbated the effects of water stress on leaf gas exchange, but had few effects on berry ripening in a study of potted Pinot noir vines with high leaf areas relative to crop levels (Poni et al. 1993). A study with growing conditions similar to those of the Okanagan Valley found only minor and inconsistent main effects and very few interactive effects of irrigation deficits and cluster thinning on vine vigor, yield components including berry mass, and basic fruit composition in Cabernet Sauvignon (Keller et al. 2008). We examined the interactive effects of irrigation reduction and crop adjustment on soil moisture distribution, leaf gas exchange, yield components, and fruit composition including tannin and anthocyanin levels in Merlot and Cabernet Sauvignon.
Materials and Methods
The experiments were conducted in Merlot and Cabernet Sauvignon blocks planted in 1999, 300 m apart, in Vincor Canada’s Sunrock vineyard, 3 km northeast of Osoyoos, British Columbia. The vineyard soil is a loamy sand of the Osoyoos series with a water-holding capacity of ~13%. Slope, aspect, and row direction were, respectively, 4%, 296°, and east-west for the Merlot block and 1%, 221°, and north-south for the Cabernet Sauvignon block. Row × vine spacing was 2.4 × 1.2 m in both blocks. The scion clones and rootstocks were Merlot 347 on 3309 and Cabernet Sauvignon 191 on Vitis riparia. The vines were trained to have bilateral cordons with shoots positioned vertically each year. The experiments were initiated in 2004 in the Merlot block and in 2005 in the Cabernet Sauvignon block and were conducted through 2006. Ambient air temperature was measured hourly in the vineyard at 2 m from the soil surface using a shielded thermistor attached to a datalogger (models 107 and 21X, respectively, Campbell Scientific, Logan, UT).
Treatments applied were a factorial combination of three irrigation and three crop adjustment regimes. The full irrigation treatment was the standard industry practice in the region. It maintained the volumetric soil moisture content in the root zone (to 60 cm depth) at above 10% and had been applied in the vineyard through the previous two years. Two reduced irrigation treatments provided two-thirds of the full rate beginning at fruit set, ~2 weeks postbloom, and until either veraison (RV) or harvest (RH), after which full irrigation was restored. The grower’s existing irrigation systems were used. In the Merlot planting there were three emitters per vine spaced 40 cm apart with every third emitter midway between vines. In the Cabernet Sauvignon the emitters were in-line and spaced 45 cm apart. Emitter delivery rate in both blocks was 2 L/hr. Irrigation was applied when the soil moisture, measured twice weekly to 60 cm depth using a capacitance probe (C-Probe, Adcon Telemetry, Boca Raton, FL) installed between vines under full irrigation in each variety block, was depleted to below 10%. Volumes applied per irrigation had been determined previously to wet the profile to between 45 and 60 cm depth. The full irrigation volumes applied pre- and postveraison, respectively, were 21 and 24 L/vine to the Merlot, and 24 and 26.7 L/vine to the Cabernet Sauvignon. The irrigation frequency was the same for the reduced irrigation treatments, but the volume was reduced over the treatment period by plugging every third emitter. Plugging was accomplished in the Merlot block by replacing the external emitters with plugs and in the Cabernet Sauvignon block by clamping rubber strips over the emitter outlets. Three to four irrigations were applied before late June each year, after which the irrigations were regular and frequent. The number of irrigations applied before and after veraison were: for Merlot, in 2004, 13 before and 9 after; in 2005, 12 before and 10 after; and in 2006, 12 before and 11 after; and for Cabernet Sauvignon, in 2005, 13 before and 15 after; and in 2006, 16 before and 15 after.
The crop adjustment treatments were combinations of shoot and cluster thinning that retained the same higher number of shoots on full-crop (FC) and cluster-thinned (CT) vines and the same lower number of clusters on CT and shoot-thinned (ST) vines. The adjustments were made over two days during the third week of May when the clusters were ~2.5 weeks prebloom. The numbers of clusters and shoots retained differed among years and varieties depending on general vine vigor and grower practice at the site (Table 1⇓).
In the first year (2004) of the Merlot experiment, the design was a randomized complete block with four replicates each consisting of three vines flanked by two guard vines. In 2005 and 2006 the design for both the Merlot and Cabernet Sauvignon experiments was a split plot with four replicates and with irrigation regimes applied to main plots and crop adjustments to subplots. The irrigation main plots in the Merlot experiment had three vines, flanked by two guards (two guard plants separating each main plot), and with crop adjustments applied randomly to single vines. In the Cabernet Sauvignon experiment the irrigation main plots had eight vines, flanked by two guards, and with crop adjustments each applied to two randomly selected vines (six of the eight).
In 2005, soil moisture distribution was monitored in the Cabernet Sauvignon experiment. Volumetric moisture content was measured in the vine row in two replicate plots of the full and RH irrigation treatments using time domain reflectometry (TDR) probes (Moisturepoint, ESI, Victoria, BC, Canada) with measurement depth increments of 0–15, 15–30, 30–60, 60–90, and 90–120 cm. The slender (1.3 × 1.8 cm) cross section of the probes enabled insertion with minimal disruption of soil structure. Two probes were installed in each of the full irrigation plots, at 10 and 22.5 cm from different emitters, the latter halfway between emitters. Three probes were installed in each RH plot at 10, 22.5, and 35 cm from different emitters, the latter two measured toward emitters to be plugged. Measurements were taken 2 to 3 times per week (29 times total) during the season and averaged for the three developmental periods: preset, set to veraison, and set to harvest. A map of soil moisture distribution was made for each treatment and developmental period using ESRI ArcMap software and the Spatial Analyst extension package (ESRI Canada Inc., Toronto, ON, Canada). A surface raster representing the soil moisture distribution was created using the Spatial Analyst Tension Spline function with the moisture levels measured at each TDR probe segment as the point data source. Contours were then generated from the soil moisture raster layer using the Spatial Analyst Surface Contour function.
In 2006, the dynamics of moisture addition and depletion in the soil profile were monitored using a data-logging TDR system installed in the Cabernet Sauvignon experiment. The probes were 60 cm long with measurement increments of 15 cm. Installation was in the vine row 10 cm from an emitter in each of two replicate plots receiving full and RH irrigation. Measurements were taken at each probe segment every 15 minutes.
In the Merlot experiment the lengths of eight (2004) or 10 (2005 and 2006) randomly chosen fruit-bearing shoots per vine were measured each year after peak growth in early July. Each year in the Merlot, and in 2006 in the Cabernet Sauvignon, cluster exposure (visible surface area) and canopy gaps (open area between the cordon and top catch wires) were estimated when viewed from eye level (1.8 m height) at midrow. Dormant vine prunings were weighed each year.
Leaf gas exchange was measured on several days during the reduced irrigation periods using a portable leaf gas exchange system (6400, LI-COR, Lincoln, NE). Two sets of measurements were taken on most measurement days, the first between 1000 and 1130 hr, and the second between 1300 and 1430 hr. On days when only one set of measurements was taken it was between 1100 and 1330 hr. For each set, two fully exposed mature leaves at midheight of the canopy were measured on each vine. On 19 Aug 2004, 9 days postveraison and two days after irrigating in the Merlot experiment, leaf water potential (Ψl) was measured directly after and on the same leaves as those used for gas exchange measurements, using a pressure bomb (PMS, Portland, OR) as described by Turner (1988).
Basic composition of the berry juice was determined from samples of 20 to 40 berries per vine (1 to 2 berries per cluster), collected biweekly beginning ~1 week postveraison. The samples were weighed to determine average berry mass then crushed with a mortar and pestle. The released juice was pressed through cheesecloth and analyzed for soluble solids by a digital refractometer (Pallette, Atago, Tokyo, Japan), pH by electrode (London Scientific, London, ON, Canada), and titratable acidity (TA) by titrating 2 to 5 mL juice with 0.1 M NaOH to a pH endpoint of 8.1 using an automatic titrator (Metrohm Canada, Mississauga, ON, Canada).
All remaining fruit was harvested when the juice soluble solids averaged ~25 Brix for Merlot in 2004 (23 Sep) and Cabernet Sauvignon in 2005 and 2006 (12 and 24 Oct, respectively), and ~26 Brix for Merlot in 2005 and 2006 (29 and 26 Sep, respectively). Average cluster mass was calculated by dividing total yield mass by the number of clusters. A 200-berry sample was collected at random from the fruit harvested from each vine and stored at −20°C before analyses for phenolics. While frozen, the skins and seeds were removed, weighed, and lyophilized prior to phenolics extraction with acidified methanol (MeOH:HCl, 99:1, v/v). Anthocyanins were determined using the Glories method (Glories 1984). Tannins were determined using a vanillin assay (Burns 1971).
Treatment main and interactive effects and relationships between response variables were determined by analysis of variance (ANOVA), appropriate for the experimental designs (randomized block and split plot), and regression analysis conducted using the SAS GLM procedure (SAS Institute Inc., Cary, NC). Growth-phase effects on leaf gas exchange parameters within varieties and years were detected using phase × block as the error term in the analysis of variance. Year-to-year effects on berry mass and phenolics within varieties were determined using year × block as the error term in the analysis of variance.
Results
Air temperatures and vine phenology.
Annual growing degree day (GDD) accumulations (base 10°C, from 1 Apr to 31 Oct) in the vineyard were 1576 (2004), 1498 (2005), and 1549 (2006). Daily maximum temperatures varied more than minimum temperatures (Figure 1⇓). Despite this variation, the temperature differences among years appeared to be randomly distributed through the seasons such that the dates of budbreak, bloom, veraison, and harvest were similar among years. A notable difference among years occurred during the first week of bloom when daily maximum temperatures in 2005 averaged 30.5°C, ~8°C higher than in 2004 and 2006. Each year, fruit harvest maturity was ~3 weeks later for Cabernet Sauvignon than for Merlot.
Soil moisture.
The high-frequency measurements of soil moisture taken in the Cabernet Sauvignon experiment provide insight into the dynamics of the soil wetting and drying over irrigation cycles and in response to reduced irrigation. In the full-irrigation plots, at 10 cm from directly beneath emitters, moisture in the top 30 cm of the profile reached ~22% within 2 hr of irrigation commencement (Figure 2⇓). That moisture level was not exceeded but was retained for more than 2 hr during some irrigations (i.e., after veraison), indicating it was near saturation. Between set and veraison moisture penetrated to below 30 cm beginning ~3 hr after irrigations commenced in the full-irrigation plots, then accumulated in the 30–45 cm depth interval to ~15% which is just slightly above field capacity (13%). In reduced-irrigation plots, irrigation between set and veraison saturated only the top 15 cm of the profile, and moisture accumulation in the 30–45 cm interval was less than in the full-irrigation plots.
Soil moisture depletion was rapid for ~10 hr after an irrigation was applied, bringing levels to near field capacity in the top 30 cm of the profile, presumably because of simultaneous drainage and crop uptake (Figure 2⇑). Slower depletion followed until moisture levels in the top 45 cm of the profile averaged below 10% in full-irrigation plots, triggering the next irrigation. In the RH plots, moisture in the top 45 cm of the profile declined to 7% before the next irrigation was applied. Postveraison wetting and depletion patterns were similar to those preveraison, except more moisture reached the 30–45 cm depth interval, resulting in significant drainage to below 45 cm where depletion between irrigations was apparent.
Maps of soil moisture distribution in the Cabernet Sauvignon experiment revealed that under both full and RH irrigation lowest moisture levels were found in the 30–60 cm depth interval, indicating it was where water uptake by the vines was most active (Figure 3⇓). Reducing irrigation through emitter elimination reduced soil moisture through most of the profile, but especially in the 30–60 cm depth interval beneath where emitters had been eliminated. There, average moisture levels were depleted to 5% in RH plots compared with 8% preveraison and 11% postveraison in full-irrigation plots. Although the spatial pattern of soil moisture in the profile changed between preset and set-to-veraison in the full-irrigation plots, reflecting the combination of emitter locations and the likely zone of maximum moisture uptake, the range of soil moisture levels changed little (i.e., remained between 8% and 13%). In contrast, under reduced irrigation over the same period, moisture was depleted through the whole profile but especially below where emitters were eliminated (i.e., to 5%). Between set to veraison and veraison to harvest, soil moisture levels increased through the entire profile in full-irrigation plots but mainly under emitters in the reduced-irrigation plots.
Water relations and leaf gas exchange.
Midday Ψl at 9 days postveraison and 2 days after irrigating (midway between irrigations) in the Merlot experiment in 2004 was higher (p < 0.01) under full (−0.97 MPa) than under RH (−1.33 MPa) irrigation. At the time of measurement, full irrigation had been resumed for 9 days in the RV treatment but Ψl had not fully recovered (−1.24 MPa).
In both varieties each year, the gas exchange parameters, stomatal conductance (gs), and photosynthesis (as net CO2 assimilation, A) declined between the morning and afternoon from 16% to 54% for gs and from 8% to 39% for A (p < 0.01 in all cases, data not tabulated). Water stress may have educed these changes over the day but reduced irrigation enhanced the decline only in 2006 for both parameters in Merlot (p < 0.05) and for gs in Cabernet Sauvignon (p < 0.01).
There was a seasonal pattern in leaf gas exchange parameters which was similar between varieties and among years. Within three weeks before veraison, presumably coincident with the onset of the berry growth lag phase, gs, A, and transpiration (T) declined and then recovered at veraison or later (Figure 4⇓, Figure 5⇓). The reduction in T was in some cases several-fold (i.e., in Merlot in 2005 and in Cabernet Sauvignon in 2005 and 2006). The leaf gas exchange parameters also declined in response to reduced irrigation, and then mostly recovered after veraison in the RV treatment when full irrigation was restored. That recovery effect is expressed in the phase × irrigation interaction. In both varieties and all years, the relative decline in A was less than in gs and T, revealing an increase in water use efficiency under reduced irrigation.
Crop adjustment did not influence the degree of decline or recovery of gas exchanged with growth phase (i.e., no growth phase × crop adjustment interactions were detected) and did not influence the irrigation effects on gas exchange parameters (i.e., no irrigation × crop interactions were detected). Except in Cabernet Sauvignon in 2005, stomatal conductance was higher in shoot-thinned than in full crop and/or cluster-thinned vines and was associated with increased A and/or T (Table 2⇓).
Canopy characteristics.
Crop adjustment had more influence than did irrigation on vine canopy characteristics. In Merlot, longer shoots and an increased canopy gap area, together indicative of taller and more open canopies, were produced by shoot thinning compared with cluster thinning in 2004 and 2005 (Table 3⇓). In 2004, when the crop adjustment treatments were most severe, cluster thinning reduced Merlot cluster exposure. There were no main effects of crop adjustment on Merlot pruning mass, but in 2005 the pruning mass of cluster-thinned compared with shoot-thinned vines was higher under full irrigation and lower under reduced (RV or RH) irrigation. In Cabernet Sauvignon, canopy gap area was not affected by crop adjustment but cluster exposure was higher in shoot-thinned than in cluster-thinned vines (Table 4⇓). Pruning mass was reduced by shoot or cluster thinning in Cabernet Sauvignon in both years. The suppressive effect of cluster or shoot thinning on pruning mass was consistent across irrigation treatments in 2006, but in 2005 the effect was found only in vines under full or RV irrigation.
Irrigation treatments had few and inconsistent effects on measured canopy characteristics. In Merlot, shoot length was unaffected by irrigation in any year and pruning mass was affected only in 2005 when it was enhanced by reduced irrigation in shoot-thinned vines (Table 3⇑). Canopy gap area and cluster exposure in Merlot were higher under RV than full irrigation but only in one year (2004). In Cabernet Sauvignon, irrigation did not affect canopy gap area or cluster exposure (data collected only in 2006) and had no main effect on pruning mass in either year (Table 4⇑).
Yield and yield components.
Fruit yield per vine, cluster mass, and seed number per berry were relatively low in 2005 in both varieties (Table 5⇓). Low yield resulting from small clusters was an industrywide phenomenon that followed unusually hot bloomtime weather. Maximum temperatures were near 35°C for several days in late April 2005 (see weekly average maximum temperatures, Figure 1⇑), which may have reduced fruit set.
Crop yield was unaffected by irrigation in any year, but was reduced equally by cluster thinning and shoot thinning (Table 5⇑). In Merlot, shoot thinning in 2004 to retain half the number of clusters of full-crop vines resulted in larger (higher mass) clusters across all irrigation treatments. This effect did not occur in Merlot and Cabernet Sauvignon in the other years when crop adjustment was less severe. In 2006, when Merlot yields were highest, the response of cluster mass to crop adjustment was influenced by irrigation such that the highest cluster mass was produced by cluster thinning under full irrigation and by shoot thinning under reduced irrigation.
Berry mass was unaffected by crop adjustment except in Merlot in 2004 when cluster thinning caused a transitory increase in berry mass not evident at harvest (Figure 6⇓, Figure 7⇓). Reduced irrigation decreased the mass of berries during their development in both varieties in 2005 and 2006, but the effect diminished by maturity and was then detected only in Cabernet Sauvignon in 2005 and in Merlot in 2006. In both varieties, across years there was a negative relationship between the berry number per vine and berry mass, but this relationship did not exist within years (Figure 8⇓).
Berry composition.
Treatment effects on juice soluble solids were mostly transitory. Soluble solids accumulation was advanced by thinning clusters or shoots in Merlot each year and by thinning shoots in Cabernet Sauvignon in 2005 and was delayed by reduced irrigation (RV and RH) in Cabernet Sauvignon in 2006 (Figure 6⇑, Figure 7⇑). In mature berries, soluble solids was unaffected by treatments except in Merlot in 2004 when it was enhanced slightly (<1 Brix) by thinning clusters or shoots. Reduced irrigation led to lower juice TA levels, detectable at harvest in Merlot in 2005 and in Cabernet Sauvignon in both years, and to increased juice pH, detectable at harvest in Merlot in 2004 and 2005.
Seed number per berry and the skin and seed mass components of berries were not affected by irrigation treatments in either variety and were affected by crop adjustment only in Merlot. Shoot thinning in Merlot increased skin relative mass in 2004 (Table 6⇓) and reduced seed number per berry (Table 5⇑) and seed relative mass in 2005 (Table 7⇓). In 2006, cluster thinning in Merlot increased seed relative mass (Table 7⇓).
The anthocyanin concentration in berry skins of both varieties was not affected by irrigation treatments and was affected by crop adjustment only in Merlot in 2006 when it was higher in full crop than in cluster-thinned vines (Table 6⇑). The anthocyanin relative content of berries (mg/g fresh mass) was not affected by treatments in either variety.
Treatment effects on skin tannins were few and inconsistent. Merlot berry skin tannin concentration was affected by treatments only in 2005 when high concentrations were produced by full crop vines only under full irrigation (Table 8⇓). The skin tannin relative content of Merlot berries was affected only in 2004 when it was lower in full-crop than shoot-thinned vines. Cabernet Sauvignon skin tannins were affected only by irrigation treatments in 2006 when skin tannin concentration and skin tannin relative content were lower in response to reduced irrigation (RV and RH).
As with the skin phenolics, treatment effects on seed tannins were inconsistent and detected only in one year in each variety (Table 7⇑). In Merlot in 2004, seed tannin concentration was higher in response to RV than to full irrigation, whereas in Cabernet Sauvignon in 2005, full rather than reduced (RV or RH) irrigation produced higher seed tannin concentrations. Irrigation did not affect the seed tannin relative content of berries of either variety. Crop adjustment by shoot thinning increased seed tannin concentration in Merlot in 2004 and in Cabernet Sauvignon in 2006, but enhanced the seed tannin relative content only in Merlot berries in 2004.
Berry mass and the relative contents of berry phenolics varied more among years than in response to treatments (Figure 9⇓). In both varieties, the analysis of variance mean squares for year compared with the irrigation and crop adjustment mean squares was at least five-fold higher for berry mass and at least 100-fold higher for each of the berry phenolics relative contents. Berry anthocyanin relative content was unrelated to berry mass in Merlot, whereas in Cabernet Sauvignon, negative relationships were detected across years and within 2005. Skin tannin relative content was unrelated to berry mass in Cabernet Sauvignon and within years in Merlot but positively related across years in Merlot. Seed tannin relative content was highest in 2005 in both varieties (Table 7⇑), despite fewer seeds per berry that year (Table 5⇑). However, the seeds were so massive that seed relative mass was highest in 2005 (Table 7⇑). No relationship between seed tannin relative content and berry mass was found in Merlot, but in Cabernet Sauvignon there was a positive relationship across years (Figure 9⇓).
Discussion
Frequent monitoring of soil moisture through the profile in the Cabernet Sauvignon block revealed the dynamic spatial pattern of moisture distribution resulting from vine uptake punctuated by irrigation events and altered by the elimination of emitters. The maps of moisture in the soil profile revealed that lowering the emitter density in the planted row reduced the amount of available moisture through the profile while maintaining the wetted soil depth. However, increasing the range of moisture levels in the root zone has been found to influence vine development (Dry et al. 2000a, 2000b), physiological processes such as stomatal conductance (Dry and Loveys 1999, Dry et al. 2000a), and the production or translocation of abscisic acid (Stoll et al. 2000). It may therefore be difficult to elucidate fully all interacting mechanisms involved in responses to irrigation reduction.
The full irrigation treatment, which is the current industry practice based on a low soil moisture threshold, replenished crop uptake through to veraison and then slightly exceeded uptake through to harvest. Under reduced irrigation, uptake exceeded that applied until after veraison when moisture levels increased slightly under emitters in the RH plots. The increased application of water relative to consumption late in the season, despite using a consistent soil moisture threshold for irrigation, might be the consequence of applying slightly more water per irrigation, or monitoring soil moisture midway between emitters where it was most depleted and would underestimate average moisture in the profile, especially later in the season.
The response of leaf water potential to the reduced irrigation treatment was similar to reported responses to deficit irrigation (Kennedy et al. 2002, Shellie 2006, Keller et al. 2008). Stress caused by the irrigation reductions was sufficient to reduce stomatal conductance and leaf gas exchange, as found for gs and A (Escalona et al. 1999), but had few and inconsistent effects on vine vigor, as indicated by pruning mass which averaged ~400 and 650 g/m of canopy for Merlot and Cabernet Sauvignon, respectively. Under comparable growing conditions but in own-rooted vines, Keller et al. (2008) reported similar pruning masses which were also influenced little by irrigation deficits. These results may indicate that vigor of vines cultured on coarse soil in a dry climate is not easily manipulated by irrigation, as vigor was not excessive under full irrigation or reduced substantially in response to deficit irrigation.
The decrease in photosynthesis because of reduced irrigation appeared insufficient to affect carbon supplies to fruit, as yield and fruit soluble solids were unaffected. However, growth of roots and permanent (structural) vine wood, not measured in this study, can be impacted by reduced irrigation (Shellie 2006).
It appears that gas exchange was regulated diurnally and seasonally by the vines independently of water supply. The afternoon reductions in gs and A were associated with maintenance of a near-constant leaf internal CO2 concentration (Ci) (data not shown), similar to our previous findings (Bowen et al. 2004), and would serve to maintain water use efficiency. The seasonal pattern in gas exchange, marked by the lag-phase reduction, has been observed in other studies (Zufferey et al. 2000) and attributed to enhancement of photosynthesis by the increasing fruit-sink activity beginning at veraison (Candolfi-Vasconcelos et al. 1994). This finding indicates that total transpiration and irrigation requirements are overestimated during the lag phase when calculated from ambient humidity and a crop coefficient based on exposed leaf area (as in Keller et al. 2008). Carbon demand by the fruit (Loveys and Kriedemann 1974) may explain the seasonal pattern but photosynthesis was little influenced by crop adjustment, crop load, or yield, indicating that gas exchange was not significantly controlled by the source-sink balance (Poni et al. 1993, 1994). Only in Cabernet Sauvignon in 2006 were photosynthesis rates higher in full-crop than in cluster-thinned vines (Table 2⇑). Shoot thinning compared with cluster thinning enhanced photosynthesis rates in Merlot in 2004 and 2005 and Cabernet Sauvignon in 2006, despite the similar yields and crop loads resulting from these treatments. Shoot-thinned vines had fewer leaves and lower total leaf areas early in the season, which perhaps increased the supplies of nutrients (especially N) and water to each developing leaf and enhanced its photosynthetic capacity. The lack of consistent effects of shoot thinning on pruning mass indicates that shoot vigor was enhanced by an increased resource supply.
Crop adjustment through shoot or cluster thinning had mostly predictable effects on yield components, i.e., cluster numbers and yield were reduced accordingly. The interactive effect of crop adjustment and irrigation in Merlot in 2006 that enhanced cluster mass by cluster thinning under full irrigation and by shoot thinning under reduced irrigation might have involved soil moisture conservation by the lower amount of transpiring leaf area on shoot-thinned vines. The lack of major main and interactive effects of crop adjustment and irrigation on berry mass and basic juice composition (i.e., soluble solids, titratable acidity, and pH) is similar to findings elsewhere (Keller et al. 2008). While other studies have observed significant reductions in berry mass in response to irrigation deficits (Hardie and Considine 1976, Kennedy et al. 2002, Roby and Matthews 2004, Shellie 2006), we have found no reports of large reductions in berry mass in response to commercially acceptable irrigation reductions (i.e., not causing severe yield reductions). In our study, although yield and berry mass were little influenced by irrigation, during midseason there often were visible leaf symptoms of mild water stress (mild wilting and flagging) in vines under the reduced-irrigation treatment, and in one season some scorching of leaf margins was evident. A larger reduction in irrigation would likely have increased the stress severity and possibility of permanent stress damage to leaves. Water deficits sufficient to reduce berry mass by 10 to 30% also reduced fruit yields by 20 to 46% (Hardie and Considine 1976). Shellie (2006) attained 20 to 34% reductions in berry mass through water deficits that reduced yields by 41 to 48%. In the present study, there were minor and mostly transitory effects of reduced irrigation on berry mass and soluble solids, perhaps indicative of developmental delays, but there was also higher pH and lower TA in the berry juice, as reported elsewhere (Shellie 2006). The reduction in acidity was likely due to higher microclimate temperatures associated with the reduced transpiration rates, as reported from another irrigation study conducted nearby (Bowen et al. 2005b).
Our finding that crop adjustment had little influence on berry phenolics has also been observed for cluster thinning (Reynolds et al. 1994, Keller et al. 2005) and for shoot thinning (Reynolds et al. 2005). However, we also found little influence of irrigation on berry phenolics, which is inconsistent with reports of increased anthocyanin (Matthews and Anderson 1988) and phenolic levels (Roby et al. 2004) in response to irrigation deficits. The lack of major irrigation effects on berry characteristics and vine vigor at our study site might be due to the moisture release physics of the coarse-textured soil, which allowed for only short periods of vine stress before irrigation was necessary. Enhancement of berry phenolics by water stress may require more severe or prolonged moisture deficits, which are perhaps commercially practical only on finer textured soils or under higher humidity conditions, which would allow vines to endure prolonged stress. On the other hand, the lack of irrigation effects under our conditions indicates that a substantial amount of water (33%) can be conserved, without negative impacts on fruit yield or quality, by reducing emitter density while maintaining high irrigation frequency.
Most of the observed variation in vine growth and fruit characteristics was from year-to-year rather than treatment effects, especially in Merlot. Average pruning mass increased with progressive years in both varieties. Average crop loads were similar between years in Cabernet Sauvignon (successively, 4.0 and 4.1 fruit/pruning mass), but varied substantially among years in Merlot (successively, 7.6, 4.6, and 5.3 fruit/pruning mass) due more to variation in yield than to pruning mass. Interestingly, the low fruit yields in 2005 were associated with lower per berry seed numbers but enhanced development of individual seeds rich in tannins, which led to the lowest ratio of skin to seed tannins among years (data not shown). These results demonstrate that coincidence of an unusual weather event (hot temperatures) with a vulnerable developmental stage (pollination) can substantially affect fruit development and quality.
The year-to-year differences in anthocyanin levels emerged from different developmental sources. Highest berry anthocyanins were found in 2005 in Merlot and in 2006 in Cabernet Sauvignon when skin relative mass and anthocyanin concentration were both relatively high. In Merlot, lower levels were due in 2004 to lower skin relative mass and in 2006 to lower skin anthocyanin concentration. These year-to-year differences in anthocyanins were not associated with harvest maturity, as indicated by basic juice components (soluble solids, pH, and TA), or with vine vigor, which can affect fruit shading and anthocyanin levels (Tarara et al. 2008). Large year-to-year differences in skin anthocyanin levels, of similar magnitude to the effects of fruit shading, were also reported (Tarara et al. 2008). Seasonal growing conditions that correlated with the differences in skin mass and anthocyanin concentration were not evident from our data.
The year-to-year pattern in skin tannins followed neither that of seed tannins nor skin anthocyanins. In both varieties the skin tannin content of berries decreased from year to year, as did average berry mass, but the decline was mainly due to decreasing concentrations in the skin rather than to relative skin mass. The corresponding year-to-year increase in pruning mass suggests that vine vigor effects on canopy light levels or berry temperature might have influenced skin tannin production. However, within years the skin tannin content of berries did not correlate with pruning mass (data not shown).
As mature berry mass was unaffected by crop adjustment and exhibited only a weak response to irrigation in only one year in each variety, there appears to be little scope to manipulate berry mass through irrigation or crop adjustment under the study site conditions. Both varieties exhibited an inverse relationship between berry number per vine and berry mass when data from all years were considered, as found elsewhere (Keller et al. 2008), but no correlations were found within years or in response to the early manipulations to berry number by crop adjustment. Berry mass varied mostly among years and did not correlate well with phenolic contents. Only in Cabernet Sauvignon were anthocyanin levels (as % berry fresh mass) higher in berries with lower mass. Skin tannins decreased with the year-to-year decline in berry mass in Merlot, but were unrelated to berry mass in Cabernet Sauvignon. Seed tannins followed the decreases in berry mass from 2005 to 2006 in both varieties, but were about average for Merlot in 2004 when berry mass was highest. The seasonal influences on berry phenolics appear to have been unrelated to effects on berry size. Likely they were due to climatic differences, since edaphic conditions and management practices changed little from year to year.
Conclusions
Our results suggest that grapevines respond differently to deficit irrigation when cultured on coarse rather than finer textured soils. Berry mass and phenolic levels were influenced little in Merlot and Cabernet Sauvignon vines cultured on sand, when subjected to soil moisture deficits leading to stress. The contrast of this result with reports of smaller, more phenolic-rich berries resulting from deficit irrigation might be due to the effects of soil texture on moisture release physics and the resulting degree and duration of water stress. We found that the compositional development of Merlot and Cabernet Sauvignon berries was influenced more by seasonal differences, likely climatic, than by irrigation or crop adjustment, emphasizing the importance of climate in determining wine grape quality.
Footnotes
-
Acknowledgments: The authors thank the British Columbia Wine Grape Council and Agriculture and Agri-Food Canada’s Matching Investment Initiative for grants in support of the research. We are also grateful to Vincor Canada for provision and maintenance of the vineyard sites.
-
Excellent technical support was provided by Steve Marsh, Emmanuelle Jean, and Tom Kopp.
- Received May 2010.
- Revision received September 2010.
- Accepted October 2010.
- Published online December 1969
- Copyright © 2011 by the American Society for Enology and Viticulture