Abstract
An experiment was conducted in the Barossa Valley, South Australia, to examine the effect of rootstocks on reproductive performance of Shiraz (Vitis vinifera L.) under water stress. Vines were grown on own roots or grafted to 110R, 1103P, 99R, Ramsey, Schwarzmann, or 140Ru. Vines either were unirrigated or irrigation was applied at 56 to 128 mm/ha across three seasons. Water stress (Ψpd < 0.8 MPa) was apparent in the unirrigated vines from veraison onward. The absence of irrigation strongly influenced vine growth and performance. Pruning weight, cane weight, and cane number were all reduced as a consequence of zero irrigation. Yields were reduced in unirrigated treatments due to a reduction in cluster number, cluster weight, and berry weight rather than fruit set or berry number. Unirrigated Ramsey was the only rootstock able to maintain yield comparable with irrigated rootstocks. Unirrigated own roots performed well in the first season but not in the second and third seasons when water stress had a negative effect on yield. Millerandage, coulure, and seedless berry numbers were the main reproductive parameters found to have a negative impact on yield and both own-rooted and grafted vines were as susceptible to these parameters. Season had a greater influence than either rootstock type or irrigation. These findings have significant implications for regions facing future drought and declining water supplies.
Water use efficiency is critical for the sustainability of viticulture. The majority of the world’s grapegrowing regions are in Mediterranean-type climates, which are characterized by warm-hot temperatures and require irrigation throughout the growing season. However, water supplies are becoming increasingly limited and grapevines are more commonly experiencing water stress during the growing season (Green et al. 2008). Increased climate variability is likely to result in periods of water shortages, resulting in the management of grapegrowing regions with less water to optimize vine and fruit growth. Severe water stress may be detrimental to production and quality at various stages of grapevine growth (Hardie and Considine 1976). For many grapegrowing regions worldwide, there is an increased requirement to minimize irrigation-water application, limit the severity of water stress on the grapevine, and produce economically sustainable yields of desirable quality (Green et al. 2008).
The effect of water deficit on reproductive development has been well documented (Matthews et al. 1987, Matthews and Anderson 1989, Poni et al. 1993, McCarthy 1997). Early and late season water deficits can be detrimental to the development of both the current and the following season’s crop (Matthews and Anderson 1989, Petrie et al. 2004). Bud fruitfulness was shown to decline under deficit irrigation for the varieties Cabernet franc, Shiraz, and, in one season, Thompson Seedless through a reduction in shoot number or a low shoot internode number when treatments received 0 and 0.2 times the water used by vines grown in a weighing lysimeter (Matthews and Anderson 1989, Petrie et al. 2004, Williams et al. 2009). However, fruitfulness has been shown to overall increase with deficit irrigation through improved light interception to developing buds (Williams et al. 2009) and under minimal and mechanical pruning, as fruitful buds will burst in preference to less fruitful buds, increasing the number of inflorescence per shoot for Shiraz (Petrie et al. 2004). Furthermore, a mild water deficit may increase fruitfulness through improved light interception to the developing buds through a decrease in foliage (Keller 2005, Williams et al. 2009).
Early season water deficits can interfere with pollination and fertilization and can cause poor fruit set and/or abscission of inflorescences (Alexander 1965, Keller 2005) and result in fewer berries per cluster (Rogiers et al. 2004). Water stress during flowering can result in cluster shrivel and berry and branch abscission, so only the rachis remains (Alexander 1965). Although water defict is acknowledged to affect fruit set (Alexander 1965, Hardie and Considine 1976, Matthews and Anderson 1989, Rogiers et al. 2004), the fundamental causes of poor fruit set remain uncertain.
Drought tolerance of rootstocks under Australian conditions has been defined as the ability of the rootstock to maintain yield and produce dry matter under water-stressed environments (McCarthy et al. 1997, Iacono and Peterlunger 2000). Rootstocks Ramsey, 110 Richter (110R), 140 Ruggeri (140Ru), 1103 Paulsen (1103P), and 99 Richter (99R) are generally considered drought tolerant (Carbonneau 1985, Dry 2007, Walker and Clingeleffer 2009, Williams 2010). Rootstocks 5BB Kober, 5C Teleki, and SO4 are considered moderately susceptible to drought and K51-40, 101-14 Mgt, Schwarzmann, and 3309C are reportedly susceptible to drought (Nicholas 1997, Dry 2007). However, there are some exceptions, particularly when compared against own-rooted controls. Using yield as the main covariate, McCarthy et al. (1997) found an ~50% reduction in yield in unirrigated treatments of Shiraz grafted to a series of rootstocks. In that study, Shiraz vines grafted to Ramsey yielded more than other unirrigated rootstocks, but not significantly so when compared to own-rooted vines, 99R, or 140Ru. Conversely, both 110R and 1103P performed poorly in the absence of irrigation (McCarthy et al. 1997). In the absence of irrigation, own-rooted Shiraz vines were as drought tolerant as rootstocks.
The use of rootstocks in Australian viticulture is relatively low in comparison to Europe and America, where phylloxera and nematodes necessitate their use. In South Australia and New South Wales, only 20% and 28%, respectively, of winegrape plantings are on rootstocks (Walker and Clingeleffer 2009). Drought tolerance is a key priority for rootstock choice in Australia (Walker and Clingeleffer 2009). Of the top five rootstocks planted in Australia in the last decade, three were drought tolerant: 1103P, Ramsey, and 140Ru (www.phylloxera.com.au/resources/rootstocks).
The vegetative and yield responses of rootstocks to deficit irrigation have been extensively studied in inland, irrigated regions of Southern Australia (Pech et al. 2008, Stevens et al. 2008, 2010). In these studies, irrigation volume was reduced by ~30% for Shiraz and Chardonnay, resulting in yield reductions of 31% and 9%, respectively. Yield was reduced for Shiraz under deficit irrigation due to a reduction in cluster number per vine for 1103P and 110R, compared to other rootstocks. There was also a reduction in berry number per cluster for 1103P and Schwarzmann (Stevens et al. 2010).
For Shiraz, deficit irrigation treatments exhibited lower midday leaf water potential (Ψl), which, in general, was not affected by rootstock genotype except on one occasion in which 1103P had a significantly lower Ψl than 110R and Ramsey (Stevens et al. 2010). In a recent deficit irrigation study of Chardonnay, Merlot, and Syrah, higher yields were reported for 3309C, 5C Teleki, and own roots than for 140Ru, 1103P, and 101CU (Keller et al. 2012). These higher yields were attributed to more clusters and berries per cluster. In that instance, scion cultivar had a greater effect on yield than rootstock and, similarly, the reported change to vigor between the treatments was a consequence of deficit irrigation rather than rootstock genotype.
The aim of this research was to ascertain whether rootstocks could mitigate the detrimental effects of water deficit on yield through improved reproductive performance. Although the effects of rootstock on yield under deficit irrigation has been examined (McCarthy et al. 1997, Stevens et al. 2008, 2010, Pech et al. 2008, Keller et al. 2012), limited attention has been paid to grapevine reproduction when vines are subject to a return to dryland conditions, as may occur under future climate conditions.
Materials and Methods
Experimental site.
In 2008, a three-year experiment was established at Nuriootpa, South Australia, Australia (34º48′S; 139.01º′E). The vineyard was planted in 2001 at 1481 vines per hectare, vine spacing and row spacing was 2.25 m × 3 m, and vines were trained to a bilateral cordon. The site was located within a phylloxera-free region that allows for the use of own-rooted Vitis vinifera vines. The vineyard was planted on a uniform site with little soil variation. The soil was a Light Pass fine sandy loam A horizon overlying a red brown mottled clay B horizon (Northcote 1954). The fine sandy loam A horizon (A1+A2) was 30 cm deep averaged across a 40 m2 sampling grid pattern with structural roots to one meter (Rudd 1975). All plant material (Shiraz clone BVRC30 and rootstocks) were certified by South Australian Vine Improvement nurseries as free from known virus and diseases. Prior to grafting, rootstocks were hot-water-treated for 30 min at 50ºC, and all vines were planted as potted one-year-old plants. Vines were spur-pruned to ~40 nodes per vine to match the commercial pruning level of the vineyard.
Irrigation and precipitation.
Vines were drip-irrigated, using irrigation sourced from either bore water (water sourced from below-ground aquifer) or water from the Murray River via the Barossa Infrastructure Limited scheme. The values used to schedule an irrigation event were based on soil water content measured with capacitance probes that recorded soil water hourly. Based on the capacitance sensors, the theoretical average total soil water content to 1.1 m depth was 236 mm for Sept 2008, 270 mm for Sept 2009, and 265 mm for Sept 2010.
Irrigation began when the soil water content to 1.1 m depth declined to a predetermined value in each season. Irrigation of control plants was intended to mimic high-quality production in the Barossa region, and therefore control vines were subject to a degree of water stress throughout the season. Due to irrigation supply-flow constraints, all rootstock combinations received the same volume of irrigation on a weekly basis as a single irrigation. Applied water was 1.1 ML/ha (110 mm) in 2008/2009, 1.3 ML/ha (128 mm) in 2009/2010, and 0.6 ML/ha (56 mm) in 2010/2011. All unirrigated treatments had no irrigation applied for the duration of the experiment and only received water through precipitation.
A bureau of meteorology weather station, located ~500 m west of the trial site, was used for climate data. Total rainfall for the site for the years 2008 to 2011 was 434, 502, 589, and 640 mm, respectively. Growing season rainfall (September to April) was 227 mm in 2008/2009, 267 mm in 2009/2010, and 516 mm in 2010/2011. Long-term annual rainfall for the site was 500 mm (http://www.bom.gov.au/climate/averages/tables/cw_023373.shtml).
Experimental design.
Vitis vinifera L. Shiraz (clone BVRC30) was grafted to six American Vitis rootstocks: Ramsey (V. champinii), Schwarzmann (Schwarz) (V. riparia × V. rupestris), 1103 Paulsen (1103P) (V. berlandieri × V. rupestris), 140 Ruggeri (140Ru) (V. berlandieri × V. rupestris), 99 Richter (99R) (V. berlandieri × V. rupestris), and 110 Richter (110R) (V. berlandieri × V. rupestris). These rootstock were compared to ungrafted, own-rooted Shiraz (V. vinifera). The experiment was performed across 10 rows for each rootstock (whole plot). Within each plot, there were three replicate blocks of irrigated and three replicate blocks of unirrigated vines (subplots), consisting of seven treatment vines (split-plot design). The unirrigated treatment was established in August 2008 before the start of the 2008/2009 growing season. Treatments undergoing zero irrigation and adjacent buffer rows had their drip-irrigation lines bypassed.
Vegetative and reproductive measurements.
During winter dormancy, cane number and pruning weights were recorded and average cane weight determined from these measures (presented on a per meter of cordon basis). Twenty canes from each replicate were also collected at this time to assess bud fertility. Compound (latent) buds (formed after the prompt bud, N+1, which contains the main shoot; May 2004) at node positions one to four on each cane were dissected and scored for the number of inflorescence primordia (IP) and the presence of primary bud necrosis (PBN) using a binocular microscope (Leica, model MS5; Wetzlar, Germany) at 10 to 40× magnification. The number of IP per compound bud was recorded in the primary bud (N+2); however, if the primary bud was necrotic, then the largest secondary bud (N+3) was scored for the number of IP. An average of the number of IP in nodes one to four was determined to give an indication of the potential fruitfulness per node. Primary bud necrosis was also assessed at each node and the incidence expressed as a percentage. Actual fruitfulness per shoot (shoots with ≥5 nodes) was determined by the number of inflorescences per meter of cordon divided by the number of shoots per meter of cordon at pruning time in the following winter to give a mean number of inflorescences per shoot. To assess fruit set, three inflorescences per vine from five vines per treatment and replicate (45 samples per irrigation treatment) were randomly selected and enclosed in a fine mesh bag before flowering. After flowering, bags were removed and the dehisced caps of the flowers in each bag were counted to determine flower number per inflorescence (Collins and Dry 2009). At harvest, the corresponding clusters were counted and weighed for inclusion in final cluster numbers and yield. Reference clusters were then collected to determine berry number per cluster. Berries within each cluster were assessed for the proportion of seeded berries, seedless berries, and live green ovaries (LGO), while average berry weight was determined through the weight of the cluster (g) divided by the sum of seeded berries and seedless berries on the cluster. At harvest, all clusters were counted, weighed, and recorded (kg) per meter of cordon. Briefly, calculations included total berry number per cluster, fruit set (%), berry weight (g), fruit yield (kg) per meter cordon, coulure index (CI), and millerandage index (MI). Millerandage and coulure are abnormal conditions of fruit set that can result in an increase in the amount of LGOs and seedless berries (millerandage) or increase in excessive flower shedding (coulure) (May 2004, Collins and Dry 2009).
Measures of vine water status.
In 2009, vine water status was determined by measures of midday leaf water potential (Ψl) at the commencement of the first irrigation in December. For the 2010 and 2011 seasons, vine water status was determined by measures of predawn leaf water potential (Ψpd) at the commencement of the first irrigation in December and then again at veraison (E-L stage 35). For all seasons, a total of five vines per replicate were sampled and assessed for leaf water potential. One leaf from each vine was excised using a single-edged razor blade through the petiole. The leaf blade was then inserted into the chamber to measure leaf water potential. Leaves were not enclosed in a plastic bag prior to the analysis of Ψpd but were placed immediately into the chamber after removal from the vine. Water potential was measured using a 3000 series leaf pressure chamber (Soil Moisture Equipment Corp., Santa Barbara, CA).
Statistical analysis.
A repeated measures ANOVA was performed on all variables for the three years of data collection using GenStat statistical package (ver. 13.1; VSN, Hemel Hempstead, UK). As the season effect was always highly significant (p < 0.001) and the rootstock x irrigation interaction was always significant (Supplemental Table 1), data were further subjected to a split-plot analysis (rootstock [whole plots] x irrigation [subplots]). This analysis was performed for each season to test for consistency of irrigation effects on rootstocks.
Results
Plant water stress indicators.
Applied irrigation varied depending on season and climate. Irrigation of control plants was intended to mimic high-quality production in the Barossa and therefore control vines were subject to a degree of water stress throughout the season. Zero irrigation did not affect midday leaf water potential (Ψl) or predawn leaf water potential (Ψpd) at the commencement of the first irrigation (December) across the three seasons (Table 1). Mean Ψl values in December ranged from −1.21 to −1.65 MPa while Ψpd values in December ranged from −0.6 to −0.9 MPa for 2009/2010 and −0.2 MPa to −0.3 MPa for 2010/2011. Flowering commenced on 17 Nov 2008, 7 Nov 2009, and 26 Nov 2011. For the 2008/2009, 2009/2010, and 2010/2011 seasons, the average flowering temperatures were 21.6ºC/9.5ºC, 36.4ºC/16.7ºC, and 23.4ºC/12ºC, respectively, and the rainfall was 12.6, 0.4, and 120 mm, respectively (Figure 1). The 2009/2010 growing season (Sept–April) was hotter and drier than the 2008/2009 and 2009/2011 growing seasons. There was 516 mm annual rainfall in the 2010/2011 growing season, which contributed to the higher Ψpd values of 2010/2011. The majority of rainfall in 2010/2011 occurred in November and December (134 mm) and February (92 mm), which coincided with flowering-fruit set and harvest time points for that season. At veraison, all unirrigated treatments had lower Ψpd values than irrigated treatments (Table 1), with the exception of Schwarz in 2009/2010. Mean values at veraison were similar between seasons (−0.89 MPa) and ranged from −0.55 to −0.92 MPa for irrigated and −0.85 to −1.2 MPa for unirrigated. Unirrigated rootstocks 1103P, 99R, and 140Ru had the lowest Ψpd values of all treatments and irrigated own-rooted Shiraz had the highest in the two seasons of analysis. At each phenological stage measured, Ψpd values differed between seasons and were always lower for 2009/2010 than 2010/2011.
2008/2009 season.
Yield response was driven by a combination of cluster weight, cluster number, and seeded berry number and was negatively correlated with MI and seedless berries (Table 2, Table 3). Actual fruitfulness was lower when fruit set and cane number were high. Fresh weight to pruning weight ratio (FW:PW), flower number, and LGO were positively correlated with each other but were negatively correlated with cluster number, fruit set, and cane number.
Unirrigated rootstocks 1103P, Ramsey, and 99R had a higher number of seedless berries and MI (Table 2). Irrigated and unirrigated own-rooted Shiraz and unirrigated 110R had higher yields as a consequence of higher cluster weight and seeded berry numbers. Irrigated own-rooted Shiraz and 140Ru also had higher seeded berry numbers, actual fruitfulness, and cluster numbers (Table 2, Table 3). Both own-rooted Shiraz treatments differed from the other rootstocks as a consequence of higher yields due to higher cluster weights and seeded berry numbers (Table 2).
2009/2010 season.
Cluster weight, yield, and berry weight were positively correlated with each other, while fruit set was negatively correlated with CI, berry weight, and LGO (Table 4). Irrigated Schwarz had a higher yield than unirrigated Schwarz (Table 4) and both Ramsey treatments were positively correlated with measures of fruitfulness and FW:PW (Table 5). Lower yields due to lower cluster and berry weights were observed for unirrigated 1103P, 140Ru, 110R, Schwarz, and 99R (Table 4) compared with the other treatments. Similarly, pruning weights were lower in unirrigated Schwarz, 110R, and 99R than in irrigated treatments, while both Ramsey treatments had low pruning weights. Lower pruning weights were due to lower cane number per vine (Table 5). Irrigated and unirrigated own-rooted Shiraz were positively correlated with high MI and pruning weights (Table 4, Table 5).
2010/2011 season.
Fruit set was positively correlated with number of seeded berries, cane number, and berry number and negatively correlated with CI and berry weight (Table 6). Irrigated 140Ru had higher yield, FW:PW, and fruitfulness than unirrigated 140Ru. Unirrigated own-rooted Shiraz had higher CI and lower seeded number, total berry number, and fruit set. On the other hand, irrigated own-rooted Shiraz was positively correlated with measures of fruitfulness, yield, and FW:PW (Table 6, Table 7). Both own-rooted Shiraz treatments had higher cane and berry weights than other treatments. Unirrigated 1103P and irrigated Schwarz had higher seeded number, berry number, and fruit set and lower CI and berry weights (Table 6). Unirrigated Schwarz, 110R, and 140Ru were positively correlated with cane number and negatively correlated with yield and fruitfulness (Table 6, Table 7).
Reproductive performance over three seasons.
For the three seasons of measurements, the main drivers of yield in each season were cluster weight, cluster number, and seeded berries per cluster. In each of the three seasons, yield was highest for the own-rooted controls (Tables 2, 4, 6). Millerandage, coulure, and seedless berries were the main reproductive parameters with a negative impact on yield and were also high when fruit set was low. The incidence of PBN in 2010 was greatest when there were higher cane numbers, and for that year there was a higher incidence of MI when there were higher pruning weights (Tables 3, 5, 7). High flower numbers correlated with low seeded berry number and total berry number per cluster.
Rootstock performance over three seasons.
Over the three seasons, own-rooted Shiraz grafted onto rootstocks had lower yield and yield parameters than own-rooted Shiraz, with the exception of irrigated 140Ru and Ramsey in the 2010/2011 season (Tables 2, 4, 6). Own-rooted Shiraz had higher yield than 1103P in all three years as a consequence of higher cluster weights, berry weights, and berry number per cluster. Own-rooted Shiraz also had higher pruning weight and cane number than rootstocks (Tables 3, 5, 7). PBN was higher in own-rooted Shiraz than in rootstocks in 2009/2010, with the exception of 110R (Table 5).
The absence of irrigation reduced yield by 25% in 2008/2009, 22% in 2009/2010, and 23% in 2010/2011 (Tables 2, 4, 6). Over three seasons, irrigated own-rooted Shiraz had the highest yield. Irrigated 140Ru and 110R and unirrigated own-rooted Shiraz had similar yield to irrigated own-rooted Shiraz. Rootstocks with the lowest yield were unirrigated 1103P, Schwarz, 99R, and 140Ru. Lower yields were largely due to a combination of lower berry and cluster weights. Over three seasons, MI was lowest for unirrigated 140Ru, Schwarz, and 110R and irrigated 110R and highest for unirrigated 99R and own-rooted Shiraz. CI was highest for irrigated Ramsey and 140Ru and unirrigated own-rooted Shiraz and 140Ru. Unirrigated 1103P had, on average, the highest fruit set, but this was only significantly greater than irrigated Ramsey because all other rootstocks had similar mean fruit set values.
Water stress and reproductive performance.
In each of the three seasons, zero irrigation reduced pruning weights, cane weights, and cane numbers (Tables 3, 5, 7). Yield was lower with zero irrigation due to fewer clusters per vine and lower cluster and berry weights (Tables 2, 4, 6). Zero irrigation decreased the number of LGO and seedless berries and increased seeded berries. There was no significant effect of zero irrigation on fruit set in any of the three years. However, in the second season CI was lower for unirrigated treatments and in the third season MI was lower for unirrigated treatments (Table 4).
Discussion
The absence of irrigation reduced both vegetative and reproductive parameters in Shiraz grapevine grafted to several different rootstocks. Yields were reduced due to a reduction in either cluster number, cluster weight, or berry weight. A cumulative effect of prolonged zero irrigation was most notable in own-rooted Shiraz than in rootstock treatments. The cumulative effect of drought over two seasons severely affected unirrigated own-rooted Shiraz, despite a wetter than average third season. In season 1, unirrigated own-rooted Shiraz was able to mitigate the effects of zero irrigation by maintaining yields comparable with irrigated own-rooted Shiraz. Thereafter, the prolonged absence of irrigation resulted in a cumulative yield decline due to a reduction in cluster number and weight. In contrast, unirrigated Ramsey was able to maintain comparable yields with irrigated Ramsey in seasons 1 and 2.
Vine water status.
In 2009/2010 and 2010/2011, measures of vine water status were changed from midday (Ψl) to predawn (Ψpd). Measures of Ψpd have previously been considered a surrogate for soil matric potential as leaf water potentials become equilibrated with soil water potential (Loveys et al. 2005). Further to this, sampling was changed to capture water potential at the commencement of the first irrigation and at veraison.
The warmer growing season for 2009/2010 contributed to the lower Ψpd values in December while the higher rainfall in 2010/2011 contributed to the higher Ψpd values for all treatments (Figure 1).
In a study conducted on Shiraz (Stevens et al. 2010), measures of Ψl under deficit irrigation lowered Ψl values by 0.11 MPa. The present study reduced Ψpd by an average of 0.29 MPa at verasion for the two seasons, and although the measures of vine water status differed, a reduction in vine water status was observed.
Different scion varieties have anisohydric and isohydric behavior driven primarily by the stomata (Schultz 2003). Shiraz, regarded as an anisohydric variety in the absence of or under deficit irrigation, has limited control of stomatal conductance and, therefore, the inability to maintain a constant Ψl as the soil dries and with increased evaporative demand (Schultz 2003). Although the present study did not consistently measure Ψl, an observed lower Ψpd value was reflective of the water potential of the soil and the increased difficulty in water extraction from the soil by the vine (Iland et al. 2011).
Effect of irrigation on reproductive performance.
In the absence of irrigation, fruitfulness was lower in each season, which is in accordance with Matthews and Anderson (1989). Bud fruitfulness has been shown to decline under deficit irrigation for the varieties Cabernet franc, Shiraz (Matthews and Anderson 1989, Petrie et al. 2004), and, in one season, Thompson Seedless (Williams et al. 2009) through a reduction in shoot number or a low shoot internode number. Accordingly, when fruitfulness was reduced, cane weights and pruning weights were also reduced in the unirrigated treatments and fruitfulness was almost always positively correlated with yield, cluster number, and FW:PW and negatively correlated with PBN when PBN values were significantly different.
Higher incidences of PBN and cane number were associated with reductions in yield, cluster number, and actual fruitfulness. A decrease in fruitfulness per node and a reduction in cluster weight have been shown to occur with the loss of the primary bud, as the secondary bud is generally considered to be less fruitful with fewer and smaller inflorescence primordia (Dry and Coombe 1994, Dry 2000).
The proportion of seeded and seedless berries and LGO within a cluster is indicative of the success of reproductive performance. A linear regression analysis was performed for fruit set and both MI (increase in the proportion of seedless berries and LGO relative to seeded berries) and CI (increase in excessive flower shedding) for each season. No relationship between MI and fruit set was observed across all seasons. In contrast, there was a strong linear regression correlation for CI and fruit set (average r2 = 0.936) in all seasons. The inverse relationship between fruit set and CI appears to be a good indicator of disruption to reproductive development for Shiraz, in accordance with the finding that some varieties such as Shiraz express high CI and low MI compared with other varieties such as Merlot or Cabernet Sauvignon (Dry et al. 2010).
Yield was significantly affected by season and irrigation. In all seasons, yield and cluster weight were reduced for the unirrigated treatments, with the exception of unirrigated own-rooted Shiraz in the first season and unirrigated Ramsey in the second and third seasons. A 30% reduction in irrigation to Shiraz has been shown to result in reduced yields due to reduced cluster numbers and berry weights (Stevens et al. 2010).
The final weight of grape berries is a consequence of the potential for cell division, determined at flowering along with cell expansion (Alexander 1965, Hardie and Considine 1976, Matthews et al. 1987). Limited water supply during this critical period of cell division and expansion can result in a reduction in berry size and berry weight (Matthews et al. 1987, Matthews and Anderson 1989, Poni et al. 1993, Ussahatanonta et al. 1996, McCarthy 1997, McCarthy et al. 1997, Rogiers et al. 2004, Keller 2005). Lower yields for both irrigation treatments were observed in the first season, due to a combination of lower cluster and berry weights and a consequence of higher number of seedless berries, lower fruit set, higher MI, and higher CI. Similar reductions in yield for the 2009/2010 and 2010/2011 seasons were observed (−22% and −23%, respectively) despite differences in leaf water potentials between the two years (Tables 4 and 6). In 2009/2010, reductions in cluster and berry weight, due to the combination of a warmer ripening period and a lack of water in the unirrigated treatments, were the main drivers of lower yield. The unusually wet growing season for 2010/2011 resulted in significantly heavier cluster weights due to heavier berry weights, contributing to an overall higher yield. Unirrigated cluster weights for 2010/2011 were heavier than most irrigated cluster weights for 2008/2009 and 2009/2010. McCarthy (1997) observed a reduction in berry weight between 9% and 45% for unirrigated treatments over a four-year study when compared with irrigated treatments. McCarthy et al. (1997) also found an approximate 50% reduction in yield in unirrigated Shiraz grafted to a series of rootstocks growing in the same region.
The highest yields obtained from two out of the three seasons were from irrigated own-rooted Shiraz vines (Tables 2 and 4). However, in the absence of irrigation, Ramsey was the best performing rootstock and maintained values similar to irrigated treatments, with the exception of season 3. Elsewhere, Shiraz vines grafted to Ramsey yielded higher than other unirrigated rootstocks, but not significantly so when compared to 110R or 140Ru (McCarthy et al. 1997). Overall, 1103P unirrigated in every season was associated with the lowest yields, and 1103P has previously performed poorly in the absence of irrigation (McCarthy et al. 1997). Despite this poor performance, it continues to be ranked as drought tolerant for Australian conditions (Nicholas 1997, Dry 2007). In 2009, unirrigated own-rooted Shiraz vines performed equally to irrigated own-rooted Shiraz vines, and although yield was reduced in the absence of irrigation, the yield of own-rooted vines regardless of irrigation treatment was higher than grafted vines. Similarly, deficit Shiraz treatments, either through witholding irrigation for approximately one month at a defined phenological time point (McCarthy et al. 1997) or by reducing irrigation applied based on soil water content for the period between fruit set and veraison (Keller et al. 2012), could outperform deficit rootstock treatments. Keller et al. (2012) attributed the geographic origin of V. vinifera as having an influence on drought tolerance, inferring that V. vinifera was likely to be more drought tolerant than some American Vitis hybrids. If yields were averaged across three seasons, then that would support the above findings (McCarthy et al. 1997, Keller et al. 2012) or, more specifically, that unirrigated own-rooted Shiraz was as drought tolerant as irrigated 140Ru and 110R (Figure 1). However, as mentioned previously, the seasonal effect on yield outweighed many of the rootstock effects, which is consistent with findings for Chardonnay, Shiraz, and Merlot grafted to a selection of rootstocks (Keller et al. 2012). Unirrigated own-rooted Shiraz was as drought tolerant as irrigated own-rooted Shiraz for only the 2008/2009 season. Thereafter, a cumulative decline in yield was observed in subsequent seasons due to a reduction in cluster number and weight. Researchers have identified that water stress may have a negative carryover effect on productivity in the following season (Matthews and Anderson 1989, Petrie et al. 2004), further supported through analysis of FW:PW, which declined in each season. This suggests that unirrigated vines became progressively unbalanced in the absence of irrigation, producing more vegetative growth relative to reproductive growth due to a combination of reduced fruitfulness and cluster weight.
Effect of irrigation on vegetative growth.
Positive correlations among pruning weight, cluster number, and yield, indicating that larger vines had a greater yield potential, have been reported (Keller et al. 2012). In all seasons, irrigated rootstocks have higher pruning weights, cane weights, and yields. The balance between vegetative and reproductive growth is indicated by FW:PW. The ideal range of FW:PW has been suggested as between 5 and 10 (Kliewer and Dookoozlian 2005), with values >10 indicating overcropping and values <5 indicating undercropping. The majority of vegetative growth occurs early in the season, and cool to warm, humid environments coupled with high soil moisture are conducive to high shoot vigor (Iland et al. 2011). There were significant differences in two out of the three seasons, whereby FW:PW was higher for the irrigated than unirrigated treatments. Mean values for FW:PW for each season were 5.1, 5.9, and 3.3, respectively, for the irrigated treatments and 4.6, 4.7, and 2.7, respectively, for the unirrigated treatments. The 2010/2011 season had the lowest FW:PW. The unusually humid and wet 2010/2011 growing season contributed to the higher shoot vigor that resulted in a dense canopy and significantly higher pruning weights. This, coupled with low fruitfulness resulted in vines that had a high FW:PW ratio regardless of irrigation treatment.
Effect of climate on reproductive performance.
The 2009/2010 growing season (Sept–April) was hotter and drier than the 2008/2009 and 2010/2011 seasons (Figure 2). Maximum flowering temperatures at Nuriootpa in 2009/2010 averaged 35ºC and ranged from 25 to 41ºC for the duration of flowering. For grapevines, an effect of high temperature (>32ºC) on reproductive development has previously been reported (Alexander 1965, Kliewer 1977, Greer and Weston 2010). Temperature regimes between 32 and 40ºC for 12-hr days between flowering and fruit set markedly reduced fruit set by 42% at 40ºC and by 23 % at 35ºC and increased the number of seedless berries within the cluster compared with treatments at 25ºC in Pinot noir (Kliewer 1977). Similar reductions were observed for Carignane in the same study. For the experiment reported here, fruit set was higher in 2009/2010 than in 2008/2009 and 2010/2011. In addition, increased millerandage was not observed in 2009/2010. The extreme temperature regimes of 40ºC and 35ºC for 12 hours (Kliewer 1977) were more severe than conditions experienced in the vineyard at Nuriootpa, where average temperature was 26ºC. Moreover, from early work by Alexander (1965), water stress rather than high temperature may reduce fruit set when temperatures are high for a prolonged period at flowering. There was no significant effect of zero irrigation on fruit set in any of the three seasons. Matthews and Anderson (1989) also did not find an effect of irrigation on fruit set, which they attributed to an absence of water deficit at this phenological time point. Our results support a similar finding.
Conclusion
The absence of irrigation strongly influenced vine growth and performance. Pruning weights and cane weights were reduced as a consequence of zero irrigation. Although yields were reduced in all seasons due to zero irrigation, this was mainly due to a reduction in cluster number, cluster weight, and berry weight rather than fruit set.
The absence of irrigation did have an effect on the performance of rootstock through a combination of effects on both vegetative and yield parameters. This is one of few studies where an own-rooted V. vinifera was included for comparison with grafted vines in assessing the impacts of water stress on reproductive performance, using quantifiable measures of fruit set. In the absence of irrigation, Ramsey was the best performing rootstock and maintained values similar to irrigated treatments. In contrast, unirrigated own-rooted Shiraz had a cumulative decline in yield while unirrigated 1103P in every season was associated with the lowest yields. These findings may have significant consequences for rootstock choice in grapegrowing regions faced with future drought and water allocation issues.
Acknowledgments
Acknowledgments: This project was funded by Phylloxera and Grape Industry Board of South Australia and by Australia’s grapegrowers and winemakers through their investment body, the Grape and Wine Research and Development Corporation. The authors thank Anne Hasted from Qi Statistics for statistical advice, Treva Hebberman and Tony Gerlach for their assistance throughout the trial, and the South Australian Research and Development Institute for use of their vineyard at Nuriootpa. Paul Petrie and John Innes are gratefully acknowledged for their comments on an early version of the manuscript.
Footnotes
-
Supplemental data is freely available with the online version of this article.
-
Publication costs of this article defrayed in part by page fees.
- Received May 2013.
- Revision received October 2013.
- Accepted October 2013.
- Published online February 2014
- ©2014 by the American Society for Enology and Viticulture