Seasonal vine nutrient dynamics and distribution of Shiraz grapevines

Aim: The nutrient reserves in the grapevine perennial structure perform a critical role in supplying the grapevine with nutrients when demand cannot be sustained by root uptake. The seasonal changes in these reserves largely depend on the developmental stage and the associated growth requirements. These stored reserves are, in turn, influenced by environmental conditions and vineyard management practices, such as production levels and water availability. The aim of this study was to assess the nutrient dynamics of a major wine grape variety grown in Australia to determine the key nutrient uptake periods and to understand the mobilization patterns throughout a season. Methods and results: The own-rooted 10-year-old Shiraz vines used for the trial were located in the Riverina, which is a hot inland grape-growing region in New South Wales, Australia. Uniformly sized vines, identified by trunk circumference, were selected for 11 excavation dates with four replicates, from a month before bud-burst to leaf-fall. The above-ground section of the vines were separated into the different perennial, vegetative and reproductive organs. The below-ground section of the vines were obtained in an allocated area (6 m2/vine) and were excavated to a depth of 1 m, and the roots were separated into rootstock and three root sizes. Sub-samples of each tissue were freeze-dried and the remaining tissues were oven-dried at 70 °C, and for both procedures the dry weight (DW) was recorded. For the nutrient analysis, the tissue sub-samples were ground up, and nutrients were determined with an nitrogen analyzer and an ICP-OES. The annual organs showed the highest nitrogen (N) concentrations in spring, with the leaves having 3 % and inflorescences 2.5 %, but stem N concentration was highest at the end of the season with 0.7 % DW. Root N concentrations are at least double the other perennial sections, with these reserves declining early in the season and being replenished by leaf-fall. The changes in concentrations for perennial sections are similar for the other macro nutrients, but differ for Ca and S in the annual tissues. The N content of the perennial structure declined considerably until flowering, with a sharp increase after harvest. The majority of the N uptake occurred four weeks before flowering and four weeks before veraison, and more than half the N of the vine was allocated to the annual organs at harvest. Other macro nutrients show a pattern of decline and replenishment in the roots and wood and most nutrients were taken up predominantly four weeks prior to flowering. Conclusions: An important finding from the study revealed that the amount of each nutrient allocated to the perennial structure and annual parts varied between the nutrients. This understanding of the nutrient dynamics will lead to an optimization of individual nutrient status and supply for grapevines. Significance and impact of the study: This is the first time that whole-vine nutrient levels were followed through the season under Australian conditions and on a wine grape variety for all macro nutrients. Such information is critical to allow precise prediction and modeling of grapevine nutrient requirements.


INTRODUCTION
The nutrient reserves in the grapevine perennial structure perform a critical role in supplying the annual growth with nutrients under conditions of insufficient nutrient uptake by the roots. The seasonal changes in these reserves largely depend on the developmental stage and the associated growth requirements of grapevines. These stored reserves are influenced by environmental conditions and certain vineyard management practices, such as production levels, water supply and canopy status. Thus, these stored resources likely contribute to the grapevine capacity within and between seasons.
The grapevine stores a substantial amount of nutrients in the perennial structure after leaf-fall. The nitrogen (N) amounts can vary between 10 and 75 g per vine (Löhnertz et al., 1989 ;Treeby and Wheatley, 2006), being influenced by environmental and management factors. Where reported, potassium (K) and calcium (Ca) are present in similar concentration ranges to N, at 30-65 g and 30-75 g per vine, respectively. Phosphorus (P) and magnesium (Mg) are present in smaller amounts, at 4-10 g for P and 10-13 g for Mg (Schreiner et al., 2006 ;Pradubsuk and Davenport, 2010). In the warmer climates, a substantial amount of N is acquired after harvest and provides 60 % of the stored N for the following season (Conradie, 1992). Other nutrients are taken also up in the post-harvest period, but any accumulated Mg and Ca is mostly lost with the leaves prior to dormancy (Conradie, 1980(Conradie, , 1981. However, the majority of nutrient uptake occurs from bloom to veraison, except for P for which uptake is predominantly prior to bloom (Schreiner et al., 2006). The importance of stored N, together with carbohydrate reserves, is required to support shoot growth in the following spring because demand at that time cannot be guaranteed by root uptake (Conradie, 1992, Conradie, 2005Bates et al., 2002 ;Cheng et al., 2004 ;Zapata et al., 2004). However, other nutrients could be equally important for the next season's growth (Tromp, 1983 ;Sánchez-Alonso and Lachica 1987), although the different mobility of each nutrient determines the accumulation and mobilization from the perennial structure. Root growth that is generally most pronounced between bloom and veraison and after harvest (Van Zyl, 1984), when carbohydrates are more available for the process (Candolfi-Vasconcelos et al., 1994), affects these nutrient dynamics of vines within a season. This project aimed to assess the nutrient uptake and partitioning of nutrition in a major red wine variety grown in Australia. For this purpose, whole vines were excavated monthly over the growing season to determine the key uptake periods of the nutrients and to understand some of the mobilization patterns during a season. Such a study has previously not being undertaken on wine grapes in Australia and similar work in warm grape-growing regions was mostly limited to N or not undertaken under field conditions. The work aimed to provide further information for the optimization of fertilizer application for vine productivity and grape composition.

Site details
The own-rooted Shiraz grapevines (clone PT23, South Australia) used for the trial were located at the Charles Sturt University vineyard (35°03'38"S, 147°21'50"E). Wagga Wagga is a hot and semi-arid area with a mean January temperature (MJT) of 24°C and average annual rainfall of 572 mm. The vineyard was located within the Riverina wine region, which is classified as a hot climate with a mean growing season temperature of 21.5°C (Hall and Jones, 2010). The vineyard soil was a red Kandosol, with an A horizon of dark reddish-brown sandy clay loam overlying, at ca. 40 cm a porous B horizon of massive granitic saprolite. The growing season 2007/08 was warmer and drier than the average, with 378 mm of rain and a MJT of 26°C (Figure 1), and the data were obtained from a weather station nearby (Wagga Wagga Agriculture Institute). The vines were planted in 1997 and trained to a single bilateral cordon 1.2 m from ground level to an approximately total length of 2 m. The vines were first mechanically trimmed and then hand pruned to 80 to 90 buds per vine. The grapes were harvested mechanically. Vine and row spacing was 2 m and 3 mm respectively, with a planting density of 1,667 vines/ha. The vineyard was drip irrigated, and in the 2007/2008 season received approximately 3.5 ML/ha of irrigation. The vineyard was harvested at 24.1°Brix and carried an average yield of 14.6 t/ha with a pruning weight of 2.2 t/ha.

Vine excavations
The trunk circumference of all vines in the 1.1ha trial area were measured during winter in Bruno Peter Holzapfel et al. 2007, and 44 uniformly sized vines were selected to allow for 11 excavation dates with four replicates. These were selected by assessing the trunk circumference of each Shiraz vine at 30 cm above ground, with circumference ranging from 12.3 to 14.5 cm. Vines were grouped into four size-based replicates of 5 mm intervals (12.3-12.8, 12.9-13.4, 13.5-14.0 and 14.1-14.5). The excavation began prior to budburst on August 21, 2007 and the last excavation was undertaken after leaf-fall on May 29, 2008. The nine dates in between were September 20, October 17, November 14, December 13, January 9, February 6, March 3, April 2 and April 30, in the 2007/2008 growing season. The timing of excavation dates and key E-L developmental stages (Coombe, 1995) are indicated in Figure 1. The procedure at each excavation was as follows: the above-ground section of the vine was cut off and separated into different parts, with the number of tissues varying with the destructive harvest dates. Prior to bud-burst the sections were was trunk, cordon, and spurs sites, while during the season the sections were stems, leaves (including petioles) and inflorescences/bunches. The excavation date was March 3, which was a few days before the commercial harvest on March 7. The grapes from the vines for the last three excavation dates after March 3 (post-harvest period) were harvested before the commercial harvest on March 7 ( Figure 1). All of the individual tissues were cut into smaller pieces, and representative sub-samples were collected, washed in phosphate-free detergent (Deconex, Borer Chemie AG, Zuchwil, Switzerland) and triple rinsed with deionized water. The sub-samples were then freeze-dried for later analysis, and the dry weight (DW) was recorded. The remaining wood and annual tissue was oven-dried at 70°C, and the DW was also recorded.
The below-ground sections of the vines were obtained on the same day the wood samples were collected. The area allocated to each vine (6 m 2 ) was marked out and then excavated with a backhoe to a depth of 80-100 cm ( Figure 2). Deeper excavation was prevented due to high density sub-soils and the size of machinery, meaning that deeper structural roots occasionally observed as cut off at the base of the excavation pit could not be followed. After each excavation, a large majority of the root system was immediately collected. This was separated into the rootstock and three root size categories (described below), and were sub-sampled and dried as described for the wood.
Over the next three days, the soil from the four vine excavation sites was manually sieved through 10×10 mm mesh to remove all FIGURE 1. Climate data from the 2007/08 growing season (left), monthly rainfall and average temperatures. Excavation dates with E-L numbers and key stages (right).
The key developmental stages are indicated for bud-break (BB), flowering (F), veraison (V), harvest (H) and leaf-fall (LF). Excavation of Shiraz vines at the first sampling date in winter (right). *key development stage only.

!"#$ %&'()*+,$-.$/(0#"1$
1= IEF&(60 1: M*EN&>E?? IEF&'G remaining roots. All the roots from each vine were separated into three size classes taht corresponded approximately to the diameter range of the main structural roots (large >7 mm), secondary roots (medium 3-7 mm), and tertiary roots (small <3 mm) and washed. On the basis of random measurements made after the roots from each vine were separated, the average diameter in each class was 10.7, 4.0 and 1.1 mm, respectively. The method of excavation did not allow the inclusion of any of the new season's fine root growth, as this was either too fine to be caught in the sieve or too delicate to survive the excavation process. The tissues were then ovendried as described for the wood.

Nutrient analysis
For the nutrient analysis, the sub-sampled tissue samples were ground to 0.12 mm using a heavy duty cutting mill (Retsch ZM2000, Haan, Germany) and then an ultra-centrifugal mill (Retsch ZM100). Total N was determined on a 50 mg subsample with a VarioMAX combustion analyzer (Elementar, Hanau, Germany), and other nutrients (P, K, Mg, S and Ca) with an inductively coupled plasma-optical emission spectrometer (ICP-OES) (ARL 3580B, Applied Research Laboratories, Ecublens, Switzerland) on a 300-400 mg subsample.

Data analysis
Data were compiled using Microsoft Excel 2016 (Microsoft Corporation 2016) and further manipulated and analyzed using SigmaPlot 14.0 (Systat Software Inc., San Jose, USA). The concentrations reported are on the medium-sized roots and cordon, while the amounts included are all combined root fractions (all root sizes and rootstock), and the cordon and spurs sites were also combined. Individual data values and points in the tables and graphs are shown with standard errors of the mean (n=4).

Results and discussion
The nutrient dynamics of Shiraz grapevines, which during this study were in full production, showed considerable changes over the season in both annual and perennial organs. Similarly, the dry matter changed through the season, with some decline in the perennial structure in early spring and significant annual accumulation due to shoot growth and berry development. This study was undertaken on an irrigated major wine grape variety in a region with a long post-harvest period, providing a contrasting situation to earlier work on grapevines that were considerably older and were either rain-fed or furrow-irrigated with either much lower or higher yield levels (Pradubsuk and Davenport, 2010 ;Schreiner et al., 2006).
The seasonal concentration and content changes of the six macronutrients were more pronounced in the annual parts of vine than in perennial tissues (Figure 3 and 4). declined during the season for four nutrients, but Mg and Ca increased until leaf-fall. Therefore, the concentrations of Mg and Ca were highest after harvest, and for the other nutrients this occurred pre-harvest. Almost all nutrient concentrations in the stems declined rapidly until flowering or veraison, but for Ca the changes in the stems were minor. The concentration changes of the perennial structure were relatively small, with a decline in spring and replenishment during the season. However, the total content of macronutrients varied substantially during the growing season, with 50 % or more allocated to the below-ground reserve sections of the vine both early and late in the season. These results indicate the buffer capacity of the perennial structure, providing the annual organs with sufficient nutrients when the supply is limited. The vines total nutrient content reached a seasonal maximum prior to harvest, and was highest for N at 60 g/vine and lowest for S at about 4 g/vine, which corresponds to 100 kg N and 6.5 kg S per hectare, respectively. For most nutrients, more than half is located in the annual organs, with the majority in the leaves ; however, the majority of K was present in the bunches, and therefore the removal of this nutrient is the highest from the vineyard system.

Dry matter accumulation
A decline of perennial structure biomass was observed in roots, trunk and cordon in spring ; this starts to increase again around flowering, but only starts to significantly recover during berry maturation, with further gains after harvest ( Table 1). The roots made up around 50 % of the perennial structure and, even at harvest, roots and wood had more than half of the total biomass. The dry mass of the shoots (stems and leaves) particularly increased from a month after bud-burst to a week after flowering, while the bunches had the largest increase from pea-size to veraison. These periods have considerable nutrient requirements (Schreiner et al., 2006), which need to be matched firstly from the perennial reserves and then from accessing more nutrients from the soil, with further requirements after harvest for replenishing reserves. The dry matter declined during the post-harvest period for the stems and leaves, as parts of shoots would have broken off (non-mature wood) and leaf-fall occurred.

Nutrient concentrations
The concentrations of N are generally the highest compared to the other nutrients. The annual organs started with high concentrations in spring: the leaves had 3 %, inflorescences had 2.5 %, and only the emerging buds were higher at nearly 4 % DW (Figure 3). Bunches at harvest and leaves prior to leaf-fall declined to about 1 % DW, and the stems regain N concentration at the end of the season and finish with an N concentration of 0.7 % DW. Both below-and above-ground parts declined during the season, until the post-harvest period when the concentrations increase. The concentrations were at least double in the roots compared to the other sections of the perennial structure. At the end of the season, the concentration largely returned to the levels observed prior to but-burst. The changes for both perennial and annual sections were similar for P and K, however only the reserve tissues for K, Ca and Mg had similar results. The concentrations varied: K and Ca were slightly below 0.5 % DW ; Mg concentrations were below 0.1 %; and sulfur (S) concentrations were below 0.5 % DW. The K concentrations were highest shortly after bloom for both bunches and leaves, at 2.2 % and 1.5 %, respectively. At harvest, the concentrations declined to 1 % in bunches and just over 0.5 % in the leaves prior to leaf-fall. The concentration changes of P and S declined from early spring toward the end of the season, in a similar way to N but at much lower levels. In contrast, the nutrients Ca and Mg increased substantially in the leaves: Ca from 1 to 3.7 % and Mg from 0.18 to 0.46 %. While the Ca concentration was highest in the inflorescence at 1 %, Mg had a concentration peak of 0.25 % in bunches after flowering. The decline in spring of nutrient concentration in the perennial structure indicates the mobilization from these tissues for the support of annual growth and development, while the accumulation of Ca and Mg in the leaves indicates the poor redistribution of the nutrients into the perennial structure during senescence (Conradie, 1981). Both observations are demonstrated in the dynamics by nutrient content on a whole-vine basis.

Nutrient accumulation
During the growing season the grapevine accumulates considerable nutrients in the annual growth, which is supplied mostly from new nutrient acquisition but also from reserves located in the roots and the wood (Figure 4). The values presented here are additive contributions of each organ, for instance the perennial reserves at bud-burst consist of 27 g root, 3 g trunk and 6 g cordon N. The N reserves in this study were 36 g/vine prior to bud-burst, which is about half that reported in other work from Australia for Sultana (Wheatly and Treeby, 2006) but similar to research undertaken on Pinot Noir in Oregon (Schreiner et al., 2006) or on Concord in Washington (Pradubsuk and Devonport, 2010). The N content of the perennial structure declined considerably until around flowering, where accumulation was only seen after veraison with a sharp increase after harvest. In this study, the decline of whole-vine N one month after budbreak indicates that the N in the root and wood tissues appears to have been at least half mobilized for growing new fine root growth. Sifting soil through 10×10 mm mesh did not allow new fine growth to be accounted for during the destructive harvests. Fine roots are known to be the major nutrient absorbing structures for vines (Keller, 2015) and considerable fine root growth occurs prior flowering (Comas et al., 2005). With the majority of N uptake occurring four weeks before flowering in this study, it supports the view that a significant amount of N had been used for fine root growth during the period after bud-break. Other nutrients analyzed showed this same trend of decline after bud-break, indicating the use of all perennial nutrient reserves for an early root growth flush to support nutrient uptake prior to flowering.
The four weeks before veraison and then four weeks after harvest were two additional uptake periods for N. The decline in N directly after harvest could again be the result of a fine root growth flush that we could not account for with the recovered root biomass (Comas et al., 2010). At harvest the annual growth contains more than half the N of the vine, with grapes being 10 g/vine. Most of the annual N is in the leaves and the least was in the stems ; however, this will not be lost from the system as both amounts are available to the vine at decomposition after leaffall and pruning. However, the findings indicate that some of the N in the leaves has already been moved into the shoots and the perennial structure, as reported in earlier studies (Conradie, 1981). The N in the crop is usually removed from the vineyard system and needs to be replaced by fertilizer or appropriate cover crops.
All nutrients showed a pattern of decline and replenishment in the roots and wood tissues and most nutrients were taken up in the four weeks prior to flowering and bloom and then again during the four weeks prior to veraison. Mg was also strongly accumulated four weeks after veraison, while P was predominantly taken up and accumulated four weeks before and after flowering. These nutrients were both further accumulated in the post-harvest period-Mg in the first weeks after harvest and P in the later part-following a decline prior to this accumulation. The amounts of nutrients allocated to the perennial structure and annual parts varied between the nutrients, but most important is that the removal by the crop is highest for K at 16 g vine, followed by Ca at 4 g, P at 2 g, and S and Mg with about 1 g each. Remobilization from the leaves to the perennial structure and shoots did not appear to take place for Ca and Mg: both have been described as taken up after harvest, but the accumulated nutrients are lost to the vine by leaf-fall (Conradie, 1981). As for N, the nutrients remaining in the leaves and shoots will most likely be recycled in the vineyard system over time after leaf-fall and pruning.
Overall there was a decline in all nutrients in the perennial structure from bud-burst to flowering, when an increase in the annual organs takes place (Table 2) ; from flowering to veraison there were only small changes in the perennial sections. A considerable accumulation in annual organs was present in this period, indicating that a considerable uptake of macro nutrients occurred. During the rest of the season more macronutrients were accumulated in the roots and wood, and N was particularly elevated in the post-harvest period, which is seen as an important time to store N for the following season (Conradie, 1992). There is a decline of nutrients during grape maturation and from harvest to leaf-fall in the annual organs, even without including the removal of nutrients due to grape harvest. Leaf loss was observed during maturation, likely due to the high temperatures frequently present, and after harvest and close to leaf-fall most nutrients are commonly remobilized from the leaves to the perennial structure. However, any Ca and Mg accumulated in the leaves is most likely being lost due to the poor mobility of these nutrients (Conradie, 1981). At the end of the season, the grapevine perennial structure nutrient reserves had increased from the previous dormancy, due to the growth of this structure. The nutrient requirements for the development of the wood and the roots must be considered for the perennial structure growth in every season. In addition, the removal of nutrients by fruit and the  varying seasonal demands should be included in requirements to ensure a vineyard's productivity level.

CONCLUSION
This research, undertaken on 10-year-old ownrooted Shiraz grapevines excavated from an irrigated hot climate vineyard, shows the nutrient concentrations and content over a growing season for all macronutrients. The study demonstrates the decline of nutrients in the root and wood tissue, the utilization of these mobilized nutrients for the annual growth and development, and their replenishment toward the end of the season. The observed decline of vine nutrients one month after bud-break indicates significant nutrient reserves are being used to support a root growth flush for nutrient uptake prior to flowering. The amount accumulated during the growing season in the annual parts at harvest is most pronounced for K, Ca and N with a range 26-30 g per vine, while for the nutrients P, Mg and S the range is 2-5 g. However, further accumulation occurred after harvest in the perennial structure for essentially all macronutrients. This is a period to replenish the reserve tissue to the level of the start of the season, indicating that the total uptake is higher than just the amount accumulated at harvest. The main periods of nutrient uptake are four weeks before flowering and before veraison, with uptake peaks for P four weeks before and after bloom. Mg is also taken up significantly four weeks after veraison. The reason behind this difference in uptake is not clear at this point, but the increase in P accumulation after bloom is in the developing bunch while for Mg this accumulation is predominantly in the roots. However, the accumulation pattern indicates that nutrient supply is important: generally prior to bloom and veraison for the annual parts and after harvest to replenish the reserves if required. Ca (g) Mg (g) N (g) K (g) P+(g) S (g)