Abstract
The response of grapevine carbohydrate reserves and seasonal growth and development to defruiting at the onset of ripening or complete defoliation at commercial harvest was examined at four sites in two hot, inland regions of New South Wales, Australia. Early defruiting over two consecutive seasons increased total nonstructural carbohydrates (TNC) in the roots, and to a lesser extent in the wood, and resulted in yield increases of up to 60% in the third season when fruit was allowed to remain on the vine until harvest. In contrast, defoliation at harvest caused a decline in TNC concentrations and reduced fruit production by up to 22% following one season of treatment and by 50% following two seasons. The higher yields developed after the two proceeding seasons of early defruiting were associated with depletion of the initially high carbohydrate reserves between budbreak and harvest, and reserve restoration did not take place until after harvest. However, in the lower yielding, previously defoliated vines, carbohydrate reserves were restored before fruit maturity. These findings show that carbon demands of ripening fruit, and photoassimilation capacity after harvest, can both limit the restoration of carbohydrate reserves to preseason levels. Marked differences between sites in the seasonal maxima of nonstructural carbohydrate concentrations in perennial tissues and aspects of floral and vegetative development are attributed to water deficits. The interaction between carbohydrate reserves and fruit and shoot growth suggests a feedback process whereby homeostasis, following environmental or cultural perturbations, is restored to a level determined by the capacity of the grapevine to assimilate and store carbohydrates.
In grapevine (Vitis vinifera L.), a deciduous, temperate climate species, spring growth requirements for carbohydrates are initially supported by reserves stored in the perennial organs over winter. Following the progressive reactivation of phloem after budbreak, mobilization of these reserves reaches a maximum at the 8 to 10 leaf stage (Yang and Hori 1979, 1980). As photosynthesis of the upper leaves becomes sufficient to support continuing shoot growth requirements, newly assimilated carbon is exported from the lower leaves back toward the perennial parts of the vine (Hale and Weaver 1962). This transition of shoots to carbon autotrophy is reflected in the dynamics of carbohydrates typically seen between budbreak and early summer, where starch and sugar concentrations in root and wood tissue show a rapid decline after budbreak and then begin to reaccumulate at bloom or shortly after (Winkler and Williams 1945, Williams 1996, Bates et al. 2002, Goffinet 2004, Zapata et al. 2004).
For grapevines in cooler climates, where conditions favor photosynthesis during ripening and fruit loads are lower, carbohydrate reserve accumulation may continue uninterrupted through fruit ripening and allow reserves to be completely restored by harvest (Bennett et al. 2005). However, weather conditions and cultural practices during the ripening period are not always optimal for photoassimilation in hot climates and thus attendant impacts on the distribution of carbohydrates for the vine development and restoration of reserves are likely. Highly productive vines in hot-climate regions of Australia can carry up to 35 kg of fruit to harvest, which would be expected to create a high demand for recent photoassimilates (Hunter and Visser 1988) and may induce further remobilization of reserves if photoassimilation during the ripening period is insufficient (Candolfi-Vasconcelos et al. 1994). Temperatures during the ripening period may frequently exceed 40°C, which is 10°C above the optimum for photosynthesis (Kriedemann and Smart 1971), and fruit weight to leaf area or pruning weight ratios may be double those recommended for vines in cooler areas (Smart and Robinson 1991, Kliewer and Dokoozlian 2005). Consequently, the postharvest period may be of critical importance for the replenishment of carbohydrate reserves when weather conditions or fruit load prior to harvest preclude it.
Previously we reported on the impact of postveraison fruit load and postharvest photoassimilation capacity on the fruit yield and composition in subsequent seasons for Semillon growing in a hot arid climates (Holzapfel et al. 2006). We found both factors strongly influenced the productivity of high-yielding vines and that conditions that favor postharvest photoassimilation are likely to be a key requisite for sustaining fruit production at high levels. Of the treatments evaluated, defruiting at the onset of ripening and complete defoliation at commercial harvest over two consecutive seasons had the greatest cumulative effect on yield and pruning weights. In this second paper, we report on the changes in the concentrations and dynamics of nonstructural carbohydrate reserves induced by these two treatments. We also report on the detailed response of shoot growth and bloom and discuss the role of fruit-load induced changes in carbohydrate reserve dynamics as a contributing factor to seasonal yield fluctuations.
Materials and Methods
Vineyards and treatments.
Six treatments designed to induce variation in postharvest photoassimilation capacity were applied over two consecutive seasons to vines in four commercial Semillon vineyards as described elsewhere (Holzapfel et al. 2006). Two trial sites were located in high-yielding furrow-irrigated vineyards in the Riverina (34°S, 146°E) and two in lower-yielding vineyards in the neighboring Hilltops region (34°S, 148°E) of New South Wales. The Riverina vineyards (R1 and R2) were planted in 1998 and 1996, respectively, and the Hilltops vineyards (H1 and H2) in 1990 and 1997, respectively. All were grown on own roots, trained to a single bilateral cordon, and spur pruned. With nontreated vines used as controls, the response of carbohydrate reserve concentrations, shoot growth, and reproductive development to defruiting at the onset of ripening (crop removal treatment) and complete defoliation at harvest (leaf removal treatment) are reported here.
The study spanned the four southern hemisphere growing seasons from 2000/2001 to 2003/2004, but for clarity these are subsequently referred to as seasons 1 to 4 (Figure 1⇓). The defruiting and defoliation treatments were applied to four replicated panels of five vines in season 1 and season 2, with the middle three vines of each panel used for measurements and sampling. The defoliation treatment was repeated approximately six weeks after harvest in season 2 at R1 and R2 to remove regrowth. Treatments were not reapplied in season 3, and fruit on vines in the early defruited treatment was retained until maturity. The study concluded in season 4 with shoot length measurements and cluster counts at bloom to assess any carryover treatment effects.
Overview of treatment periods and measurement schedule in relation to the four seasons spanned during the study. Data for the carbohydrate (CHO) reserve sampling dates are presented in Figures 2⇓ and 3⇓, and data for flowering and fruitfulness in Table 4⇓. Shoot growth data are shown in Table 3⇓ and Figure 4⇓. For each season, the approximate duration of the budbreak to harvest and harvest to dormancy periods are indicated by dark and light grey shading, respectively.
In season 1, the defruiting treatment was applied 33 and 42 days before harvest at R1 and R2, and 30 and 40 days before harvest at H1 and H2, respectively. In season 2, the treatments were applied 41 and 53 days before harvest at R1 and R2 and 45 and 54 days before harvest at H1 and H2, respectively. Total soluble solids at the time of defruiting was only recorded in season 2 and was 6.5 Brix and 8.2 Brix for R1 and R2, respectively. At H1 and H2, veraison was earlier than expected because of low crop loads, and berries had already attained 12.8 Brix and 11.5 Brix, respectively, by the time the defruiting treatment was imposed. Exact irrigation and rainfall amounts were not available for all sites, but estimates of water inputs were compiled for season 2 and 3 from grower records and Australian Bureau of Meteorology climate data. Based on crop factors suggested for hot-climate vineyards (Nicholas 2004), rainfall and irrigation matched the estimated demand from crop evapotranspiration (ETcrop) at R1 and provided an average of 77% of requirements at R2. At both H1 and H2, rainfall and irrigation met 62% of estimated demand in season 2, but only 30% in season 3 (Table 1⇓). Water supply deficit was thus an important element of difference in growing conditions between the two regions, particularly in season 3.
Seasonal vineyard water supply in relation to estimated crop evapotranspiration and summary of yield data (reproduced in part from Holzapfel et al. 2006).
Single leaf gas exchange.
Single leaf gas exchange parameters from vines in the control and early defruited treatments were compared in season 2, following defruiting at the Riverina and Hilltops sites on 16 Jan and 1 Feb, respectively. A minimum of six single leaf gas exchange rate measurements were made on fully exposed leaves in each treatment panel using a LCA4 gas analyzer (ADC BioScientific, Hoddeson, UK). The measurements were made in full sunlight between 09:30 and 12:30 on 12 Feb at the Riverina sites, and between 10:00 and 01:00 at the Hilltops sites on 22 Feb. At harvest 10 leaves were randomly collected from each replicate in the control and defruiting treatment, and the starch and soluble concentrations were determined as described below.
Carbohydrate reserves.
Root and wood (combined trunk and cordon tissue) samples were collected from all treatments at harvest and pruning in season 2, and at budbreak, harvest, and leaf fall in season 3. Collections were undertaken earlier in the study, but because of a possible loss of sample integrity during the drying process, data for these dates are not reported here. The wood samples were collected with a 4.8-mm drill bit to a depth visually estimated as the center of the cordon or trunk. Approximately four holes per vine were required across each panel of three vines to ensure sufficient sample for analysis. Root samples ranging from 4 to 10 mm in diameter were collected from near the base of the vine and later washed and recut to remove all soil and any dead tissue. All samples were oven dried at 70°C and the ground to 0.12 mm using an ultracentrifugal mill (Retsch ZM100, Haan, Germany).
For nonstructural carbohydrate analysis, soluble sugars were first extracted from a 20 mg subsample of each tissue using 3 × 1 mL × 10 min washes of 80% aqueous ethanol. The first two volumes were at 80°C and the third at room temperature. After centrifuging between each wash, the three aliquots were combined, diluted to 10 mL, and the concentration of sucrose, d-fructose and d-glucose determined with commercial enzyme assays (Megazyme International, Bray, Ireland). For starch analysis, the remaining wood sample was resuspended in 200 μL dimethylsulfoxide and heated at 98°C for 10 min. The remainder of the analysis was then performed using commercial enzymes and glucose assay kits (Megazyme International). Briefly, 300 μL thermostable α-amylase in MOPS buffer was added, mixed, and incubated for 15 min in a 98°C water bath. After cooling, 400 μL amyloglucosidase in sodium acetate buffer was added and incubated at 50°C for 60 min. The samples were mixed at 20-min intervals, and then centrifuged at 10,000 rpm for 2 min. Supernatant from root samples was diluted 1:11, and trunk and shoot samples 1:6 in Ultra-pure water. Glucose concentration of the diluted samples was then determined colorimetrically and the amount of starch in the original 20 mg sample calculated.
Shoot, flower, and fruit development.
At budbreak in season 3, 10 buds across the three vines in each replicate of the control, crop removal, and leaf removal treatments were randomly selected and the length of the shoot that subsequently developed from each bud recorded at weekly intervals for the following two months. Shoot length was also measured at bloom in seasons 2 and 4 on 10 and 20 random shoots per panel, respectively. Prior to the start of bloom in season 3, 10 inflorescences from the same vines were randomly selected, labeled, and covered with mesh bags (1.4 × 0.5 mm mesh). At the end of bloom, the bags were removed and the number of flowers per inflorescence determined by counting flower caps. The number of fully developed berries was counted on each cluster at harvest to determine percentage fruit set. Shoot fruitfulness in season 2 was determined from counts of total shoot and cluster numbers before bloom and in season 4 was determined from cluster counts at bloom on 20 random shoots per treatment panel.
Statistical analysis.
Statistical analysis was conducted using GenStat release 10.0 (VSN, Hertfordshire, UK). An analysis of variance (ANOVA) was used for each site, with means separated using the least significant difference (LSD) where the overall analysis was significant (p = 0.05). For comparison of treatment effects on percentage fruit set, flower number was used as a covariate and with data reported as adjusted means.
Results
Single leaf photosynthesis and carbohydrates after early defruiting.
On the single dates that gas exchange measurements were made at the Riverina and Hilltops vineyards, no significant difference in photosynthesis rates were found between the control and early defruiting treatments (Table 2⇓). At R1, R2, and H1 vineyards, average CO2 assimilation rates were 12.7, 15.7, and 12.7 μmol m−2 s−1, respectively, which is consistent with maximum photosynthesis rates reported for well-watered vines at these leaf temperatures (Williams 1996). Photosynthesis rates for vines at the drought-affected H2 vineyard were lower and averaged 9.3 μmol m−2 s−1 for the two treatments. At harvest, starch and soluble sugar concentrations were higher on the defruited vines than the control vines at all sites. However, these results were only significant for soluble sugars at R1 and R2.
Influence of defruiting at the onset of ripening (crop removal) on leaf gas exchange parameters before harvest and leaf nonstructural carbohydrates at harvest in season 2.
Carbohydrate reserve dynamics.
The concentration dynamics of total nonstructural carbohydrates (TNC) in wood and root tissue over the five sampling dates from harvest in season 2 to pruning in season 3 are shown (Figure 2⇓). The first two dates show the changes in TNC concentrations between harvest and dormancy induced by the second season of early defruiting or complete defoliation. The next two dates show the response of TNC concentrations between budbreak and harvest, when fruit was retained on all treatments until commercial harvest. The final date shows the postharvest change in TNC concentrations after cessation of treatments and return to normal vineyard management regimes.
Total nonstructural carbohydrate (TNC) concentrations in the wood (combined trunk and cordon, a–d) and root tissue (e–h) at the Riverina and Hilltops sites, from harvest in season 2 until dormancy in season 3, as influenced by leaf removal and crop removal treatments applied in season 1 (not shown) and season 2. H, harvest; D, dormancy (sampled at date of winter pruning); BB, budbreak. Error bars ± SE mean. ★ indicates a significant difference from ANOVA (p < 0.05).
Wood TNC concentration dynamics and absolute concentrations were similar at all four sites, and there were only small variation in concentrations between sampling dates within each site (Figure 2a–d⇑). In both postharvest periods there was a slight increase in TNC, but it appears that aboveground reserves, which would have been mobilized to support canopy reestablishment in spring, were largely restored by harvest. Generally, early crop removal did cause a slight increase in wood TNC concentrations and early defoliation caused a slight decrease. These changes were sufficient to produce small but statistically significant differences between the defruited and defoliated vines. Water supply deficit at the H1 and H2 sites did not appear to substantially reduce TNC concentrations in the wood or alter the magnitude of treatment responses.
Root TNC concentrations showed greater responsiveness to the defruiting and defoliation treatments than wood TNC concentrations and also appeared to be reduced by the water supply deficit at the H1 and H2 (Figure 2e–h⇑). The changes in reserve dynamics induced by the treatments were greater in magnitude in Riverina vineyards than in Hilltops vineyards, but the general responses at R1, R2, and H1 were similar: e.g., the amount of TNC stored in the roots by leaf fall, as indicated by concentrations at dormancy in season 2, was lowered by defoliation and increased by defruiting. Root TNC concentrations were also lower after defoliation at H2, but the effect was not statistically significant. There was one site-specific difference: at R2 the differences between the defoliation and crop removal treatments were not evident until dormancy (Figure 2f⇑). At R1 and H1, root TNC reserves increased between harvest and leaf fall, but for the defoliated and defruited vines, TNC concentrations remained constant. At R2, TNC concentrations at harvest were similar for all treatments, whereas the differences seen at dormancy were a result of a postharvest decline in reserves in the defoliated vines and a more rapid increase in the defruited vines than the control treatment.
In season 3, when all fruit was retained on the vines until commercial harvest, there was a consistent interaction between yield and the change in root TNC concentrations between budbreak and harvest. The previously defoliated vines, which began the season with low reserves and low yields, were able to increase or at least maintain TNC concentrations in the roots between budbreak and harvest (Figure 2e–h⇑). In contrast, vines in the high-yielding previously defruited treatment, which began the season with high TNC concentrations, had the highest rate of reserve depletion between budbreak and harvest. Consequently, higher yielding vines were more dependent on the postharvest period for the replenishment of root reserves than low-yielding vines, indicating that fruit load was an important factor in determining the relative importance of the pre- and postharvest periods for reserve replenishment at these sites.
For the separate starch and soluble sugar fractions, there was a distinct difference between the treatment responses in the wood and roots. At dormancy in season 2, starch concentrations in wood and root tissue were lower than control vines at all sites following postharvest defoliation (Figure 3i–p⇓). However, for the wood tissue, the decrease in starch concentrations following defoliation was nearly offset by a similar increase in soluble sugar concentrations (Figure 3a–h⇓). Similarly, the slightly higher wood starch concentrations following early defruiting at R1, R2, and H1 were accompanied by lower soluble sugar concentrations at dormancy in season 2. Consequently, there was only a small net effect of the defruiting and defoliation treatments on wood TNC (Figure 2a–d⇑), but a significant change in the relative proportions of starch and sugars that was still evident at budbreak in season 3. In the roots, starch was the predominant nonstructural carbohydrate, and soluble sugar concentrations were largely unresponsive to defruiting or defoliation. Therefore, treatment effects on starch concentrations largely determined the net effect on total root carbohydrates (Figure 2e–h⇑).
Soluble sugar concentrations (sucrose + glucose + fructose; a–h) and starch (i–p) in the wood (combined trunk and cordon) and root tissue at the Riverina and Hilltops sites, from harvest in season 2 until dormancy in season 3, as influenced by leaf removal and crop removal treatments applied in season 1 (not shown) and season 2. H, harvest; D, dormancy (sampled at the date of winter pruning); BB, budbreak, with days from budbreak in each season shown on the secondary axis. Error bars ± SE mean. ★ indicates a significant difference from ANOVA (p < 0.05).
Shoot growth.
During season 3 shoot growth rates were higher in the Riverina vineyards than in the Hilltops vineyards, and within regions shoots were longer by bloom at R1 than R2, and at H1 than H2 (Figure 4a–d⇓). The differences between sites reflected the water supply at each vineyard, and were most pronounced at the severely drought affected H2 vineyard where average shoot length was only 25 cm at bloom. There were no treatment effects on percentage budbreak at any site (not shown), but there were significant differences in shoot growth and leaf appearance rates. At R2 and H1, shoot growth rates were increased by early defruiting in the two preceding seasons and reduced by postharvest leaf removal at R1, R2, and H1 (Figure 4e–h⇓). At R1, which was the highest yielding vineyard (Holzapfel et al. 2006), significant differences in growth rate were observed within the second week after budbreak when an average 5.6 leaves had separated from the shoot tip. The relative differences between treatments reached a maximum four weeks after budbreak, at which point the control shoots had 12 separated leaves and the crop removal and leaf removal treatments had 13 and 12, respectively. At R2 and H1, treatment effects on shoot growth rates developed more slowly, with significant differences first seen in the third and fifth week after budbreak, respectively. At R1 and R2, the growth rate of the leaf removal treatment reached the lowest point relative to control vines at approximately the 10-leaf stage.
Shoot length (a–d) and growth rates (e–h) between budbreak and bloom in season 3 and following two preceding seasons of early crop removal or complete leaf removal at harvest. Dotted lines in a–d indicate daily maximum and minimum air temperatures. Error bars ± SE mean. ★ indicates a significant difference from ANOVA (p < 0.05).
Shoot length measurements were also made at bloom in season 2 and season 4 to assess the effect of one year of postharvest treatments on shoot growth or the recovery of the vines following a full season of normal management (Table 3⇓). At bloom in season 2 there was no significant difference between treatments, but in most cases there was a trend for leaf removal to decrease shoot lengths and crop removal to increase shoot lengths in a manner consistent with the larger treatment response seen in season 3. At bloom in season 4 significant treatment effects on shoot growth were still evident at the Riverina vineyards, with a similar trend at the H1 vineyard. However, the relative difference between treatments was less than in the previous season.
Shoot length of Semillon as influenced by complete postharvest defoliation (leaf removal) or early defruiting at the onset of ripening (crop removal) in season 1 and season 2.
Reproductive development.
Although no attempt was made to assess possible treatment effects on bud necrosis during dormancy, or to distinguish between shoots arising from primary or less fruitful secondary buds at the time of shoot and cluster counts, the effect of the early crop removal and postharvest defoliation on cluster numbers per shoot provides an indication of how inflorescence initiation was influenced by these two treatments (Table 4⇓). In season 3, there was a trend at the R1, R2, and H1 vineyards for leaf removal to decrease fruitfulness, and at R2 and H1 for defruiting to increase fruitfulness. These effects were significant for leaf removal at R1, crop removal at R2, and all treatments at H1, indicating that one season of postharvest treatment was sufficient to influence inflorescence initiation or the maintenance of these inflorescences over winter. In contrast, for shoot growth, where residual treatment effects were still evident at bloom, there was no significant difference in fruitfulness in season 4.
Fruitfulness and reproductive development of Semillon as influenced by complete postharvest defoliation (leaf removal) or early defruiting at the onset of ripening (crop removal) in season 1 and season 2.
Flower numbers per inflorescence in season 3 were significantly influenced by the leaf removal and crop removal treatments at the R1, R2, and H1 vineyards (Table 4⇑). Specific effects varied between vineyards, but the general response was for vines from the leaf removal treatment to have the same or a reduced number of flowers as the control, and the crop removal treatment to have the same or greater number of flowers. There was a negative correlation between inflorescence flower number and percentage fruit set at R1, R2, and H2, and a significant covariate effect for flower number (p < 0.05) when that parameter was included in the analysis of treatment effects on fruit set (Table 4⇑). At all three vineyards, fruit set was significantly lower than the controls on the previously defoliated vines and significantly higher for the defruiting treatment at H1. At H2 the bagged inflorescences were exposed on the outside of the canopy because of short shoot lengths and almost 50% had broken off by the time bags were removed. A further loss of tagged clusters had occurred by harvest. Consequently, berry counts were made on additional random clusters so an estimate of fruit set could still be made for H2. Statistical analysis was conducted on the replicate means within treatments, as there were insufficient paired flower and berry number data for covariate analysis. However, the results for both Hilltops sites, but particularly H2, indicate a strong depressing effect of water stress on percentage fruit set.
Discussion
Root and wood carbohydrate reserves.
We have previously shown that postveraison fruit load influences yield in succeeding seasons and that postharvest assimilation is required to maintain the productivity of high-yielding Semillon grapevines in hot, irrigated regions of inland Australia. As suggested in an earlier study, these findings could be explained by an inability of heavily cropped vines to sufficiently restore carbohydrate reserves before harvest (Holzapfel et al. 2006). Subsequent analysis of perennial tissue samples collected during the same study confirms that for the roots, and to a lesser extent the wood, accumulation of carbohydrate reserves did occur after harvest under normal commercial management conditions. Some regional and site specific differences were observed, notably that root starch concentrations were higher at the Riverina sites and the magnitude of treatment effects on root starch concentrations were greater at Riverina than at Hilltops. At R2, the response of root reserves to defruiting was also delayed relative to R1 and H1. However, the general treatment responses and seasonal dynamics of TNC reserves (Figure 2⇑), and soluble sugar and starch fractions (Figure 3⇑), were otherwise remarkably similar across the four vineyards.
The high root starch concentrations in the Riverina may reflect the higher carbon assimilation capacity that would be expected with large canopies in the absence of water stress and are consistent with other reported values for vines grown in warmer climates (Winkler and Williams 1945, Hunter 1998). The greater responsiveness of root TNC reserves to the defruiting and defoliation treatments, and to the subsequent variations induced in fruit yield, suggests that root TNC reserves are more sensitive to factors that influence internal carbon availability than wood TNC reserves. Therefore, the combination of a long season and nonlimiting water supply in the Riverina may allow vines to accumulate high concentrations of carbohydrate reserves on a season to season basis. However, within the season, the high fruit yields carried by these vines may temporarily prevent storage of reserves in the roots or lead to further mobilization (Candolfi-Vasconcelos et al. 1994). Consequently, while higher maxima of TNC reserves may be attained, the magnitude of fluctuations between developmental stages within a season may be greater.
At R2, where defruiting did not cause an increase in root starch concentrations until after harvest in season 2, it is possible that the concentrations were lower than the control when the treatment was applied or that defruiting resulted in a increase in root or shoot growth. The possibility that renewed shoot growth did occur is supported by visual observations that some new green leaves remained at R2 at pruning. This retention of leaves into early winter was not observed at R1 and suggests that renewed growth may have provided an alternate sink for carbohydrates after the fruit was removed, but that the canopy subsequently maintained photosynthetic capacity for a longer period after harvest.
Water supply deficits comparable to those seen in season 3 at the Hilltops vineyards have been shown to halve assimilation in other circumstances (Perez Peña and Tarara 2004), and it is likely that reduced carbon assimilation capacity may have impacted on carbohydrate reserve concentrations in the present study. At both Hilltops vineyards moderate water supply deficits were also experienced during season 2, suggesting that lower starch concentrations in the roots may have been a consequence of reduced photosynthesis over several seasons. Fruit yields in season 2 may also have been sufficiently low to allow earlier replenishment of reserves, therefore reducing any potential impacts of defruiting and defoliation. The more advanced berry maturity when the defruiting treatments were applied in season 2 would also have meant that the fruit requirement for carbohydrates was already slowing at the time of removal. Consequently, despite an overall reduction in carbohydrate reserve storage under water stress, attempting to manipulate vine carbohydrate accumulation by reducing fruit loads or defoliating vines after harvest would have little impact if the fruit was already a weak sink or if postharvest photosynthetic capacity was already too low to contribute significantly to the restoration of reserves.
Reproductive development and shoot growth.
Yield reductions following one or more seasons of partial or complete defoliation have been well documented (Scholefield et al. 1977, Mansfield and Howell 1981, Candolfi-Vasconcelos and Koblet 1990, Bennett et al. 2005) and are consistent with the yield response to partial or complete postharvest leaf removal in the present study (Holzapfel et al. 2006). Likewise, defruiting 30 days after bloom has been shown to increase yield potential in the following season (Goffinet 2004). The importance of photosynthesis and carbohydrate supply for the processes of floral initiation has long been implied by studies showing a decrease in fruitfulness following defoliation (May et al. 1969) or by an improvement under high light intensity and temperature (May and Antcliff 1963, Buttrose 1969). More recent work has attributed this fruitfulness response to the carbon assimilation capacity at the whole shoot level, rather than to light exposure of individual buds, and to the net carbohydrate availability at the time of initiation (Sánchez and Dokoozlian 2005). However, there is little information on the role of reserve carbohydrates if the supply of carbon from current photoassimilation is limiting.
In the present study, fruitfulness in season 3 was generally increased by early defruiting and decreased by postharvest defoliation (Table 4⇑). Shoot length at bloom and shoot number (not shown) in season 2 did not differ greatly between treatments, providing some circumstantial evidence that these differences in bud fruitfulness were not related to light exposure of the shoots within the canopy. While we do not have carbohydrate reserve data for budbreak in season 2, cane weights at the end of season 2 were reduced by postharvest defoliation and increased by early defruiting (Holzapfel et al. 2006). These findings are consistent with the response of Chardonnay to one season of defoliation treatments (Bennett et al. 2005) and suggest the reduced fruitfulness following postharvest defoliation in season 1 may be related to reduced carbohydrate availability in the following spring. It is also possible that the differences in fruitfulness were related to the number of buds that remained viable following initiation in season 2 rather than to treatment effects on initiation per se, as localized carbohydrate deficiency may also influence the extent of bud necrosis in grapevine (Vasudevan et al. 1998). However, without information on the extent of any bud necrosis during season 2, we cannot determine the relative contributions of initiation and bud viability to the differences in fruitfulness seen between treatments in the following season.
In season 4, where the initiation and maintenance of inflorescence primordia were subject to the cumulative effects of two seasons of early defruiting or postharvest defoliation, there was no longer any significant difference in fruitfulness between treatments (Table 4⇑). In season 3 significant differences in shoot length were evident at bloom, with shoot growth rates (Figure 4⇑), leaf size, and lateral development (not shown) reduced by postharvest defoliation and increased by early defruiting. Although direct light measurements were not made, it is likely that the exposure of individual shoots within the canopy would have been increased in the previously defoliated vines. Consequently, these differences in shoot growth and canopy development during season 3 may have been sufficient to offset any direct effects of carbohydrate reserves on initiation or bud survival by altering the light environment around individual shoots. This was particularly evident with the postharvest defoliation treatments, where pruning weights were reduced by up to 50% relative to the controls or the previously defruited vines (Holzapfel et al. 2006), but fruitfulness did not differ significantly between treatments at any vineyard. These findings suggest that carbohydrate reserves may play a role in the process of inflorescence initiation, but ensuring adequate shoot exposure and photosynthetic capacity may be of greater importance for promoting bud fruitfulness.
Flower numbers per inflorescence can be reduced by defoliation-induced carbohydrate reserve depletion in the previous season (Scholefield et al. 1977, Bennett et al. 2005), suggesting a possible role for reserves in regulating continued inflorescence development at budbreak. A positive relationship was found between flowers per inflorescence and trunk or root starch concentrations at budbreak (Bennett et al. 2005), and a similar, although weaker, relationship was evident in the present study (not shown). However, given that treatments in season 1 and season 2 could have altered the extent of inflorescence development after initiation in season 2, it is difficult to attribute any differences in flower numbers directly to starch concentrations at budbreak in season 3. Further, for wood tissue the increase in soluble sugar concentrations after defoliation corresponded closely with the decrease in starch, so there was only a minor effect on TNC reserve concentrations (Figure 2⇑). In the roots, sugar concentrations were low and largely unresponsive to the defruiting and defoliation treatments. It is possible that flower development may have been influenced by root starch reserves as the inflorescence resumed development before budbreak. However, buds are initially isolated from the vasculature of the parent vine (Jones et al. 2000) and may be more dependent on local reserves prior to the reactivation of phloem at budburst (Esau 1948, Koussa et al. 2005). Therefore, further information regarding the capacity of more remote carbohydrate reserve pools to supply developing buds may be required before predictive relationships between reserve status and reproductive development can be interpreted with more confidence.
The postharvest period and seasonal yield variations.
The reproductive development and vegetative growth responses associated with the change in carbohydrate reserve status, where fruiting and shoot growth were promoted by early defruiting but reduced by postharvest defoliation, highlight the underlying role of the postharvest period for maintaining the productivity of high-yielding grapevines. That is, the depletion of reserves between budbreak and harvest in heavily cropped vines necessitates a postharvest recovery, but the amount of carbohydrates that can be accumulated once the fruit is removed may in turn ensure the development of a high yield in the following season. It is possible that a long postharvest period may also mean that vines can carry crop loads that would traditionally be considered excessive in relation to the amount of leaf area without causing a depletion of reserves, weakening of vine growth, or yield fluctuations known to be associated with overcropping in some varieties (Weaver and McCune 1960, Goffinet 2004). Therefore, if the postharvest restoration of the reserves of high-yielding vines is prevented for any reason, then vine growth and yield may be reduced in the following season. Previous work with Sultana grown under similar conditions to the present study has shown vines in hot climates can tolerate up to 60% defoliation after harvest without a significant impact on yield or carbohydrate reserve status (Scholefield et al. 1978). However, when vines are subjected to 100% defoliation, we have found that carbohydrate reserve accumulation is reduced, particularly in the roots, and that there is a cumulative reduction in yield and vegetative growth when the treatment is repeated over successive seasons.
The apparent relationship between carbohydrate reserves and reproductive development, where winter reserves were positively correlated with yield level in the following season, also provides insight into the role of reserves as a possible factor underlying seasonal yield variations. Yield can be reduced if carbohydrate reserve accumulation is reduced in the previous season (Bennett et al. 2005), and the results of the present study confirm that deliberate manipulation of carbohydrate reserve status can be associated with yield responses in the following season. However, when the vines were allowed a full season of growth to recover from these treatments, the subsequent fruiting response tended to allow carbohydrate reserves to return to similar concentrations to the control vines by the end of the season. Again, this was most evident in the roots, although the relative amounts of starch and sugars in the wood were also similar to the control vines by the end of the study. One interpretation of these findings is that the interaction between carbohydrate reserves and reproductive development may tend to stabilize grapevine productivity in the longer term if sudden changes in yield or assimilation capacity occur due to weather conditions or a change in management practice in the shorter term.
However, as previously reported (Holzapfel et al. 2006), there was a trend for increased trunk growth following defruiting and reduced trunk growth relative to control vines following defoliation. Consequently, while reserve concentrations may have been similar, total reserves in the previously defruited vines may still have been higher because of the larger perennial structure and lower on the previously defoliated vines. Differences in shoot length were still evident in season 4 (Table 3⇑), suggesting some residual treatment effects could be expressed over future seasons because of differences in vine size. Concentration measurements of carbohydrate reserves may be sufficient for shorter term studies where established vines have started at the same size, but in the longer term, or where comparing different vineyards, some assessment of vine size is required to relate productive capacity to the total amount of carbohydrates stored in the vine.
Conclusions
In hot-climate vineyards, where productivity is not otherwise limited by water stress, postharvest photoassimilation appears to be critical for maintaining the productivity of high-yielding grapevines. This period appears particularly important for root reserves, which are more sensitive to fruit yield than wood reserves and may not be replenished before harvest under high crop loads. If the postharvest recovery is prevented by defoliation at harvest, then carbohydrate reserve restoration is reduced, yields decline as a result of decreased cluster weight and number per vine, and vegetative growth is reduced. In contrast, early harvest may favor carbohydrate reserve accumulation and leads to increased vigor and higher yields through increased cluster weight and number per vine.
The study was not designed to determine if there is a direct influence of carbohydrate reserves on fruiting or vegetative development, but did demonstrate that shorter term perturbations in carbohydrate reserve status were associated with yield and shoot growth responses in following seasons. To further our understanding of the potential role of carbohydrates reserves in determining yield in the following season, more information is needed on the importance of various reserve pools for the early stages of bud development. Further information on carbohydrate reserve dynamics under normal commercial management practices, as opposed to the more severe treatments imposed in the present study, is needed to establish whether reserves actually fluctuate sufficiently from season to season to justify any proposed link with seasonal yield variations.
Footnotes
Acknowledgments: The work was supported by Australian grapegrowers and winemakers through their investment body, the Grape and Wine Research and Development Corporation, with matching funds from the Australian Government.
The authors thank Robert Lamont, David Foster, and Kirstina Lamont for past and ongoing technical support, Markus Keller, Roger Mandel, Shayne Hackett, and Kerry DeGaris for their input at various stages of the study, and Jim Hardie for constructive comments on the manuscript. We greatly appreciate the assistance of growers involved in the study, and thank them for allowing us access to their vineyards.
- Received December 2008.
- Revision received May 2009.
- Accepted May 2009.
- Copyright © 2009 by the American Society for Enology and Viticulture
Literature Cited
Vol 60 Issue 4
