Advertisement
Research Report

Annual Nutrient Removal from Irrigated Vineyards in Dry Climates

Nataliya Shcherbatyuk, Pierre Davadant, Markus Keller
Am J Enol Vitic.  2026  77: 0770009  ; DOI: 10.5344/ajev.2026.25041

Abstract

Background and goals Vineyard nutrient budgeting as a basis for efficient nutrient management is essential to maintain grapevine productivity and fruit quality. This study quantified nutrient removal through harvested fruit, senescent leaves, and pruned canes across four wine and juice grape cultivars in arid eastern Washington over 3 yr.

Methods and key findings Vine tissue samples were collected from Chardonnay, Sauvignon blanc, Syrah, and Concord vineyards from 2021 through 2023. Biomass and concentrations of macro- and micronutrients were measured in harvested fruit, abscised leaves and, in winegrapes, pruned canes. Overall, potassium and nitrogen accounted for the largest macronutrient losses. Fruit harvest was responsible for most of the nutrient removal, which varied with crop yield. Abscised leaves and pruned canes contributed additional potential nutrient losses, with higher biomass production amplifying the removal of macronutrients such as nitrogen and calcium. Early fall frost events disrupted the remobilization of foliar nutrients, especially nitrogen and phosphorus, leading to increased nutrient retention in senescent leaves.

Conclusions and significance This study highlights the importance of properly accounting for annual nutrient losses in vineyards. The findings underscore the need for holistic nutrient budgets that incorporate all loss pathways to support long-term vineyard sustainability, providing actionable insights for optimizing fertilization strategies and improving nutrient use efficiency.

Introduction

Sufficient mineral nutrition characterized by a well-balanced supply of macro- and micronutrients is important for the optimal growth, development, productivity, and fruit quality of grapevines. Each mineral nutrient plays distinct roles in various physiological processes and inadequate replenishment can lead to imbalances and deficiencies, jeopardizing vine health, yield, and grape quality (Keller 2025). Nitrogen (N) is the primary nutrient influencing grapevine vigor and is correlated with yield (Spayd and Morris 1978, Amiri and Fallahi 2007, Vrignon-Brenas et al. 2022) and aromatic compounds in wine (Rapp and Versini 1996, Lacroux et al. 2008, Yuan et al. 2018). Both N deficiency and excess can disrupt vine growth and productivity, with deficiency limiting vegetative development and canopy function, and oversupply promoting excessive vigor, delayed fruit ripening, and reduced fruit quality (Keller 2025). Potassium (K) is one of the most abundant cations in plants, and K deficiency can impair photosynthesis, reduce berry size, limit sugar accumulation, and increase susceptibility to drought stress, whereas elevated amounts of K in grape berries are associated with high juice pH (Mpelasoka et al. 2003, Assimakopoulou and Tsougrianis 2012, Keller 2025). Phosphorus (P) deficiency may not only obstruct cell division, leading to restricted leaf and root growth, but it can also impair cluster initiation or differentiation (Tulloch and Harris 1970, Skinner et al. 1988, Skinner and Matthews 1989, Balzergue et al. 2017). Calcium (Ca) stands out among macronutrients because a significant portion of its ions (Ca2+) are concentrated in cell walls. Ca deficiency can reduce shoot and root growth, while excessive Ca availability may disrupt nutrient balance (e.g., by reducing K and magnesium [Mg] uptake), potentially affecting vine vigor and fruit composition (Keller 2025). Mg deficiency often leads to interveinal leaf chlorosis and decreased photosynthetic capacity, ultimately limiting carbohydrate production and berry sugar accumulation, whereas excessive Mg can disrupt cation balance (e.g., by reducing K and Ca uptake), potentially influencing vine vigor and fruit composition (Keller 2025). Sulfur (S) deficiency can increase a plant’s vulnerability to pathogen attack and reduce vine vigor, while excessive S may acidify the soil and interfere with nutrient availability, potentially affecting root growth and nutrient uptake (Cooper and Williams 2004, Keller 2025).

Micronutrients, though needed in much smaller quantities than macronutrients, are essential for grapevine growth, development, and fruit quality. Micronutrients such as iron (Fe), zinc (Zn), manganese (Mn), boron (B), and copper (Cu) support vital physiological functions (Marschner 2012, Keller 2025). Micronutrient deficiencies or imbalances can negatively affect vine health, yield, and berry composition. However, fulfilling plant requirements for micronutrients can be challenging, as they are generally present in low concentrations within the soil solution (Cass 2005).

Vineyard nutrient management involves considering both input and output of nutrients in a vineyard. Annual nutrient inputs predominantly derive from fertilizer application (both organic and inorganic), fixation by leguminous cover crops, weathering of rocks, and atmospheric deposition through rainfall. In contrast, nutrient output or loss includes leaching, denitrification, and runoff, as well as nutrient removal during harvest and leaf fall (Williams 1991, Whitehead 2000, Niinemets and Tamm 2005, Ludemann et al. 2024). By estimating inputs and key nutrient losses, vineyard managers can assess the balance between nutrient supply and demand and make informed decisions about potential adjustments to nutrient management practices.

Arguably the largest nutrient export from a vineyard—and the one most commonly considered in nutrient budgets and fertilization guidelines—is the removal of nutrients with the harvested fruit each year (Christensen and Peacock 2000, Schreiner et al. 2006, Arrobas et al. 2014). Current estimates for key macronutrients vary considerably and are on the order of 1 to 3 kg N, 0.2 to 0.4 kg P, 1.5 to 4 kg K, 0.2 to 1 kg Ca, and 0.05 to 0.2 kg Mg per ton of fruit (Keller 2025). However, leaf fall, the annual process of leaf senescence and abscission in plants, also plays a crucial role in the life cycle of grapevines. Leaf senescence is a regulated biological process corresponding to the final stage of leaf development (Guo and Gan 2005, Lim et al. 2007). During this phase, nutrients accumulated in leaves throughout the growing season are retrieved and transferred to the perennial storage organs. Before leaf abscission, leaves redistribute ~50 to 80% of their N, P, and K to other plant parts (Keskitalo et al. 2005, Niklas 2006). This redistribution supports the renewed vegetative and reproductive growth in spring. However, if senescence is interrupted by a frost event (temperature < −2°C; Fennell 2004), the nutrients might be trapped in detached leaves.

Therefore, in addition to fruit harvest, leaf fall contributes to nutrient removal from the vines. When senescent leaves stay in the vineyard, they are gradually incorporated into soil organic matter by soil microorganisms and nutrients will be slowly mineralized and released for roots to take up. For example, N that is present in fallen leaves may gradually become accessible to roots as nitrate (NO3−) or ammonium (NH4+), or susceptible to loss through ammonia (NH3) volatilization, leaching, or denitrification (Wetselaar and Farquhar 1980, Arrobas et al. 2014, Eichert and Fernández 2023). Some N (and perhaps other nutrients as well) may also be transferred more directly from leaf litter to vine roots by arbuscular mycorrhizal fungi (Xiao et al. 2025). However, in vineyards located in windy areas, leaves are prone to being displaced, which may result in nutrient loss from a vineyard (Figure 1A). Even if leaves remain in the vineyard, they can sometimes be trapped by midrow vegetation, so that their nutrients may become available to the cover crop rather than the vines, especially where vine roots are concentrated in the rows (Figure 1B; see also Schreiner 2005). Of course, it is also possible for leaves to be blown into a vineyard from neighboring vineyards or other crops or natural vegetation. Unlike the nutrient loss in harvested fruit that can be estimated from the crop yield, leaf biomass at the end of the growing season is not routinely measured and therefore this metric is generally not considered in fertilizer programs.

Two photographs show fallen grapevine leaves on pasture vegetation and along a vineyard midrow. The two photographs are arranged side by side and labeled A and B. Photograph A shows a ground surface covered with many fallen grapevine leaves scattered across grass and low pasture vegetation. The leaves form dense patches across the center and lower areas, with green vegetation visible between and around the leaves. Photograph B shows a vineyard row with grapevines on both sides and a narrow midrow path through the center. Many fallen grapevine leaves are trapped along the middle of the path, with exposed soil, dry ground cover, and irrigation lines visible beside the vine rows.
Figure 1

Wind may displace abscised leaves from a vineyard, making their nutrients unavailable to grapevines. Grapevine leaves in a pasture next to a vineyard (A), and grapevine leaves trapped by resident vegetation in the midrows of a vineyard (B).

Finally, during winter, grapegrowers generally prune the previous season’s fruiting shoots, called dormant canes after leaf abscission. When left in the vineyard, these canes experience the same fate as the fallen leaves, albeit much more slowly, and over time their nutrients may also become available to vine roots for uptake. Nevertheless, because pruned canes are often moved to the midrow, the same caveat applies to them as that for leaves trapped by midrow vegetation, especially in drip-irrigated vineyards in dry regions, where grapevine roots are typically concentrated in the vine row. Moreover, in some instances the canes are removed from the vineyard to minimize the presence of pathogen inoculum and reduce disease pressure (Wise et al. 2007). This practice also accounts for nutrient removal that should be included in fertilization programs.

The objective of this study was to determine the annual nutrient removal (i.e., output) from irrigated, commercial wine and juice grape vineyards in arid eastern Washington to support nutrient budgeting for replacement fertilizer application programs. Though the growth of grapevine roots and aboveground perennial organs also permanently removes nutrients from the soil over the lifespan of a vineyard, those outputs are typically minor in mature vines (e.g., Schreiner et al. 2006, Pradubsuk and Davenport 2010, 2011, Verdenal et al. 2021) and were not considered in our study. We first hypothesized that the total amount of nutrient removal from a vineyard is dominated by the harvested fruit, with a non-negligible amount of nutrients lost in fallen leaves and pruned canes that are removed from the vineyard. Then, we hypothesized that early frost damage to leaves in fall may disrupt the normal nutrient remobilization pathway, leading to a more substantial loss of nutrients with the abscised leaves.

Materials and Methods

Vineyard sites and management

Samples were collected in southeastern Washington from own-rooted winegrapes, Vitis vinifera L., cvs. Chardonnay, Sauvignon blanc, and Syrah, and own-rooted juice grape cv. Concord (an interspecific hybrid cultivar with Vitis labrusca L. and V. vinifera ancestry). The climate of the region is arid (~200 mm annual precipitation) with warm summers and moderately cold winters. Winegrape blocks were located in the Columbia Crest vineyard of Ste. Michelle Wine Estates (45°93′N; 119°63′W; 160 m asl). Chardonnay, Sauvignon blanc, and Syrah were planted in 2010, 2015, and 1998, respectively. The soil type across all three vineyard blocks is a Quincy loamy sand with pH 7.9 and <1% organic matter, volumetric water content (θv) at field capacity (FC) of 11.4% (v/v), and θv at permanent wilting point (PWP) of 3% (https://websoil-survey.sc.egov.usda.gov). A caliche layer above basalt occurs at depths between 90 and 120 cm.

One of the two Concord vineyards used was planted in 2003 at Washington State University’s (WSU) Irrigated Agriculture Research and Extension Center (46°29′N; 119°74′W; 360 m asl). The soil is a Warden silt loam with pH 7.2, <1% organic matter, θv at FC of 22.7%, and θv at PWP of 7.1% (https://websoilsurvey.sc.egov.usda.gov). A caliche layer over basalt at depths varying from 50 to 120 cm limits rooting depth. The other Concord vineyard was planted in 2018 at Schilperoort Farms near Sunnyside, WA (46°32′N; 120°01′W; 229 m asl). The soil is a Warden silt loam with pH 7.5, <1% organic matter, θv at FC of 22.4%, and θv at PWP of 6.8% (https://websoilsurvey.sc.egov.usda.gov). A caliche layer over basalt occurs below 90 cm depth.

Chardonnay and Sauvignon blanc vines were trained to spur-pruned bilateral cordons positioned ~1 m above the vineyard floor, utilizing a vertical shoot positioning (VSP) system supported by catch wires. Vine spacing within rows was 1.83 m, with rows spaced 3.05 m apart. Syrah vines were similarly trained at a height of 1 m but utilized a spur-pruned unilateral cordon VSP system. The spacing for Syrah vines was 1.83 m within rows and 2.74 m between rows. Concord vines in both vineyards were trained to a single cordon wire at 1.83 m and planted at 2.74 m between rows and 1.83 m within rows. The WSU vineyard was machine-pruned and the Schilperoort vineyard was prepruned by machine, with a hand follow-up targeting 120 buds per vine. Winter ryegrass was used between vineyard rows at Schilperoort Farms and frequently mowed, while permanent but summer-dormant resident vegetation was maintained in all other blocks. Undervine vegetation was controlled using 1.2-m herbicide strips.

In accordance with standard regional practice, all vineyards were drip-irrigated, with water applied to replenish the soil profile to near FC around budbreak. During the growing season, irrigation scheduling was determined by each grower and water was applied weekly based on canopy observation, soil moisture measurements, weather conditions, evapotranspiration rates (ET), and phenological stage. No water stress was applied prior to fruit set. Between fruit set and harvest, mild water stress was applied to Chardonnay and Sauvignon blanc (irrigated at 60% to 80% of crop ET), and mild-to-moderate water stress was applied to Syrah (irrigated at 30% to 50% of crop ET to stop shoot growth). In the juice grape blocks, irrigation water was applied once or twice a week to keep θv above 15% (WSU vineyard) or 18% (Schilperoort Farms) and prevent vine water stress (Keller et al. 2023). After harvest, the soil profile was refilled to near FC in all blocks.

Because sample collection occurred within existing fertilizer trials, N (Sauvignon blanc, Syrah, Concord) or K (Chardonnay) fertilizer rates varied across each block depending on the cultivar, but no P was applied during the study years. Though these trials are not the focus of this paper, they are briefly described below under “Experimental design.” All other vineyard management practices, including pest and disease control, were applied by the collaborating growers according to regional industry standards. No S-based fungicides were applied in the juice grape blocks, and the last S sprays in the winegrape blocks were applied well before veraison (typically before July; see Supplemental Table 1). No Mn- or Zn-based fungicides were applied during the study years.

Experimental design

Each fertilizer trial was arranged in a randomized block design, with four replicate blocks comprising 45 vines each (three rows wide, 15 vines long) per treatment. Fertilizer was applied by fertigation from 2021 through 2023 as follows: three rates of N (22, 45, 90 kg N/ha, as urea ammonium nitrate [UAN32]) were split-applied at the six-leaf stage, full bloom, and fruit set in Sauvignon blanc and Syrah. Two rates of N (45, 90 kg N/ha, as UAN32) were applied at the same stages in the Schilperoort Farms Concord block. Three rates of K (0, 45, 90 kg K/ha, as liquid potash) were applied at the six-leaf stage in Chardonnay. Tissue samples were collected at harvest (fruit), after the first killing frost (< −2°C) at the end of the growing season (leaves), and during winter pruning (canes) from 2021 through 2023 (sampling dates listed in Supplemental Table 1). Fruit and cane samples were collected from three (2021) or five (2022 and 2023) vines located at the center of each treatment replicate. Canes were not collected in the machine-pruned Concord vineyards. Leaf samples were collected as described below under “Leaf collection.”

Weather data collection

Daily weather data were obtained from the AgWeatherNet system (https://weather.wsu.edu), using data from weather stations located in Paterson (station Paterson.W) for the winegrape vineyards, and from Prosser (station Roza.2) and Sunnyside (station Sunnyside.N) for the juice grape vineyards at WSU and Schilperoort Farms, respectively.

Yield and pruning weight measurements

Data were collected from the same vines that were used for fruit and cane sampling. Harvest time was determined by each grower using fruit target total soluble solids (TSS) of 22 Brix for Chardonnay and Sauvignon blanc, 24 Brix for Syrah, and 16 Brix for Concord. All vines within a block were harvested on the same day. Yield was determined by weighing all clusters and used to calculate the amount of nutrients in the fruit. Pruning weight was determined during winter pruning of winegrapes and used to calculate the amount of nutrients in dormant canes. Pruning weight was not measured in the machine-pruned Concord vineyards.

Fruit and juice collection

Four clusters were collected from each data vine a day before harvest and brought to the laboratory. A total of 50 berries were collected from the four clusters and processed as outlined below in the “Nutrient analysis” section. A subset of samples was weighed to determine fresh weight (FW), then oven-dried at 60°C to obtain dry weight (DW): the percentage DW was 19.6 ± 0.3% for Concord (n = 80), 29.1 ± 0.2% for Syrah (n = 300), 25.9 ± 0.2% for Chardonnay (n = 180), and 24.9 ± 0.4% for Sauvignon blanc (n = 120). Given the relatively high variability, a rounded average DW percentage of 20% for Concord, 30% for Syrah, and 25% for both Chardonnay and Sauvignon blanc was used to estimate nutrient removal per ton of fruit (i.e., FW basis). Additionally, in 2021 and 2022, nine Chardonnay and 15 Syrah grape juice samples (1 L each) were collected in glass bottles after pressing in the research winery at the WSU Wine Science Center and processed as described below (under “Nutrient analysis”) to measure K, the predominant cation in grape juice, with a substantial proportion localized in the skins and seeds (Mpelasoka et al. 2003, Rogiers et al. 2017). Fruit and juice K concentration values were used to estimate the potential quantity of K that could be recycled back into the vineyard system through the reintegration of pomace. To estimate the amount of K retained in pomace, the difference between the nutrient content in whole fruit and the corresponding content in juice was determined. For this calculation, we assumed 1 t of fresh grapes to yield ~750 L of juice, based on standard commercial extraction rates.

Leaf collection

Vine canopies were enclosed (netted) after harvest with 2-cm mesh polyethylene bird netting (Orchard Valley Supply). In all winegrape blocks, 12 random vines were netted in 2021 and one vine located at the center of each treatment replicate (i.e., n = 12) was netted in 2022 and 2023. For juice grapes, four random groups of four consecutive vines were netted in the WSU vineyard in 2021 and at Schilperoort Farms in 2022 and 2023. All leaves captured in the nets and any remaining leaves still on the canopy were collected after the first killing frost event (< −2°C; see Supplemental Tables 1 and 2). A subset of samples was weighed to determine FW, then oven-dried to obtain DW, and processed as described below under “Nutrient analysis.” The percentage DW was 89.9 ± 1.1% for Syrah (n = 12), 97.1 ± 2.5% for Chardonnay (n = 12), 101.2 ± 1.3% for Sauvignon blanc (n = 12), and 90.1 ± 1.1% for Concord (n = 4), indicating that Chardonnay and Sauvignon blanc leaves had already completely dried out in the vineyard, while Syrah and Concord leaves retained some residual moisture. In addition, an experiment was conducted in the Chardonnay block in 2022 to estimate the amount of nutrients remaining in leaves after a simulated early fall frost event. All leaves were collected separately from nine random vines on 21 Sept, another nine vines on 21 Oct, and a final set of nine vines on 22 Nov.

Dormant cane collection

During winter pruning, 10 to 15 dormant canes were collected from each data vine in the winegrape blocks. Samples were processed as described below under “Nutrient analysis.” Following Schreiner (2021), all nutrient data for the canes were derived from node tissue samples, each consisting of the dormant bud and ~1 cm of internode on either side, even though this approach would have slightly overestimated whole-cane nutrient concentrations for some nutrients (especially Mg). A subset of samples was weighed to determine FW, then oven-dried to obtain DW: the percentage DW was 57.0 ± 0.2% for Syrah (n = 200), 60.5 ± 0.2% for Chardonnay (n = 120), and 57.8 ± 0.3% for Sauvignon blanc (n = 120). Considering this variability and the variation in cane water content throughout the winter (Keller 2025), a rounded average DW percentage of 60% was used for all cultivars to estimate nutrient removal per unit vineyard area, using pruning weights (i.e., FW basis).

Nutrient analysis

Tissue samples were not washed for this study. We considered S, Mn, and Zn contamination from fungicide applications to be low because the last S spray was applied between 78 and 126 days before harvest across the winegrape blocks (Supplemental Table 1; cf. Kwasniewski et al. 2014); no S was applied to juice grapes, and no Mn- and Zn-based fungicides were applied in any blocks. All samples were ovendried at 60°C to constant weight, ground to a fine powder (0.25 mm) in a ZM 200 ultracentrifugal mill (Retsch), and sent to a commercial laboratory (Soiltest Farm Consultants) for macro- and micronutrient analysis. Dried fruit samples were sent without grinding and processed directly by the commercial lab. Juice samples were also delivered directly to the laboratory for K analysis. Total N was analyzed by dry combustion (Dumas method, LECO CN628); NO3− was analyzed following KCl extraction (FIA Lab, SM4500NH3, colorimetric); and P, K, S, Ca, Mg, B, Zn, Mn, Cu, and Fe were analyzed by inductively coupled plasma optical emission spectroscopy (ICP-OES; Perkin Elmer Optima DV7300) following nitric/perchloric acid digestion (Gavlak et al. 2003). The Fe data were excluded from further analysis because the analytical data were highly variable, likely due to dust contamination.

Data analysis

Data were analyzed using Statgraphics Centurion 18 (Statgraphics Technologies) and Excel ver. 2206 (Microsoft). As cultivars could not be statistically compared in our study (each cultivar was at a different site), one-way analysis of variance (ANOVA) was applied to evaluate the effect of year within each cultivar (α = 0.05). The effect of fertilizer rate was tested on pooled samples from the 3 yr using one-way ANOVA. The effects of defoliation timing and pomace recycling potential were also analyzed using one-way ANOVA. Post-hoc means separations were carried out using Tukey’s honest significant difference test. Mean values are reported with their standard error (SE). Pearson correlation analysis and linear regression analysis was conducted for select pairs of response variables, such as crop yield and nutrient concentrations.

Results

Weather

The three locations differ in seasonal heat accumulation (Supplemental Table 2). Prosser is the coolest site, with 1471 growing degree days (GDD > 10°C), followed by Sunnyside (+3.4%) and Paterson (+16.8%). Paterson also records twice as many days above 35°C compared to the other locations. Precipitation is very low across all sites (<200 mm annually). On average, 45% of annual precipitation occurs between April and October at all sites, and the first killing frost typically occurs in early November. All three study years in all locations were warmer than the long-term average (Supplemental Table 2). The 2021 and 2023 growing seasons were particularly warm across all locations. In contrast, 2022 experienced a cooler-than-average start to the season, followed by a sharp temperature rise in mid-August, allowing seasonal GDD to surpass the long-term average.

Nutrient removal with harvested fruit

The heaviest crop was harvested in 2022 in all winegrapes and in 2023 in Concord (Table 1). The fruit dominated the amount of biomass (i.e., DW) that was potentially removed from each vineyard on an annual basis (Table 1), on average accounting for 76, 83, and 69% of the total DW in Chardonnay, Sauvignon blanc, and Syrah, respectively (Concord is excluded here because pruning weight was not measured). In the fruit (which included skin, pulp, and seeds) at harvest, K was by far the most abundant macronutrient, followed by N, then Ca and P, and finally, Mg and S (Figures 2A and 3A, Supplemental Table 3). Nitrate-N consistently contributed <1% to the total fruit N. Although we were unable to statistically compare cultivars because they were grown at different sites, there was a trend of Concord berries, despite being substantially bigger (3.3 ± 0.04 g) than winegrape berries (Chardonnay and Sauvignon blanc average 1.6 ± 0.03 g; Syrah 1.4 ± 0.01 g), having higher nutrient concentrations in their dry matter compared with the winegrapes (Supplemental Table 3).

View this table:
Table 1

Crop yield, end-of-season leaf weight, and pruning weight, and their respective biomass, per unit vineyard area of Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr. Fresh weight (FW) and dry weight (DW) were determined from single-vine measurements and plant spacing. Yield was recorded at harvest, leaves were collected after the first killing frost after each growing season, and pruning weight was collected in late winter (i.e., next year). Total carbon can be estimated according to Song et al. (2023) for fruit (41% of DW), leaves (43%), and canes (45%).

Three pie graphs labeled A, B, and C compare nitrogen, phosphorus, potassium, calcium, magnesium, and sulfur percentages. The three circular pie graphs are arranged vertically and labeled A at the top, B in the middle, and C at the bottom. In panel A, the segments are labeled nitrogen 25.4 percent, phosphorus 5.8 percent, potassium 54.4 percent, magnesium 2.5 percent, calcium 9.6 percent, and sulfur 2.4 percent. In panel B, the segments are labeled nitrogen 19.7 percent, phosphorus 1.4 percent, potassium 8.9 percent, magnesium 10.1 percent, calcium 57.8 percent, and sulfur 2.0 percent. In panel C, the segments are labeled nitrogen 24.0 percent, phosphorus 3.8 percent, potassium 27.0 percent, magnesium 8.5 percent, calcium 34.7 percent, and sulfur 2.0 percent.
Figure 2

Proportions of different macronutrients in harvested fruit (A), senescent leaves (B), and dormant canes (C) averaged for Chardonnay, Sauvignon blanc, and Syrah grapevines in vineyards in southeastern Washington over 3 yr. Berries were collected at harvest, leaves were collected after the first killing frost, and canes were collected at pruning in late winter (i.e., next year).

Two pie graphs labeled A and B compare nitrogen, phosphorus, potassium, calcium, magnesium, and sulfur percentages. The two circular pie graphs are arranged vertically and labeled A at the top and B at the bottom. In panel A, the segments are labeled nitrogen 27.0 percent, phosphorus 5.6 percent, potassium 57.1 percent, magnesium 1.6 percent, calcium 6.3 percent, and sulfur 2.4 percent. In panel B, the segments are labeled nitrogen 32.6 percent, phosphorus 2.9 percent, potassium 10.7 percent, magnesium 5.6 percent, calcium 46.3 percent, and sulfur 2.0 percent.
Figure 3

Proportions of different macronutrients in harvested fruit (A) and senescent leaves (B) of Concord grapevines in vineyards in southeastern Washington over 3 yr. Berries were collected at harvest and leaves were collected after the first killing frost.

Across all cultivars, the fruit N concentration was highest in 2021. Higher K concentrations were generally found in 2023 in the winegrapes, but Concord fruit had lower K that year. The high K corresponds to a low cropping year in the winegrapes, and the low K corresponds to a high cropping year in Concord (Table 1). Correlation analysis confirmed a negative trend for K with increasing yield that was not consistently observed for other macronutrients (Supplemental Figure 1). Yet even the trend for K was weak (r2 ≤ 22%), and linear regression analysis showed that the fruit K concentration decreased by only 0.03% DW (Concord) to 0.2% DW (Syrah) for each 10 t/ha increase in yield (Supplemental Table 4). Comparing Supplemental Tables 3 and 5 shows that the annual variation in fruit N and K was similar to that introduced by varying fertilizer rates. With very few minor exceptions, however, only the mineral nutrient whose supply was directly manipulated by fertigation in each trial (K in Chardonnay, N in all other cultivars) was significantly altered in the fruit. Exceptions included S (0.06 ± 0.002 and 0.05 ± 0.001% at 90 and 22 kg N/ha, respectively) and B (26.6 ± 1.2 and 24.3 ± 0.9 mg/kg at 90 and 22 kg N/ha, respectively) in Sauvignon blanc, and Ca (0.19 ± 0.01 and 0.18 ± 0.01% at 90 and 45 kg N/ha, respectively) in Concord (p < 0.05). All other macro- and micronutrients remained unaffected by the fertilizer treatments (p > 0.05). By comparison with N, the annual variation in P concentration, though statistically significant in Chardonnay and Concord, was rather small. Similar year-to-year variation also occurred for Ca, Mg, and S, though the direction and magnitude of the variation varied among the different macronutrients (Supplemental Table 3). Among micronutrients in the harvested fruit, B was present at much higher concentrations than Cu and, finally, Mn and Zn (Supplemental Table 3). The micronutrient concentrations varied across cultivars and years, with shared trends and notable differences. Zn and Cu concentrations generally declined over time in all cultivars, while Mn and B displayed more variable patterns.

Macronutrient removal per ton of harvested fruit varied across years (Table 2). More N (1.7 ± 0.02 kg/t across 3 yr) and K (3.6 ± 0.02 kg/t) were removed with Concord fruit compared to the winegrapes (1.4 ± 0.07 kg N/t, 3.0 kg ± 0.02 K/t), but other macronutrients were similar across cultivars or sites. On average over the 3 yr and four cultivars, fruit harvest removed 0.4 ± 0.03 kg Ca, 0.3 ± 0.02 kg P, 0.1 ± 0.01 kg Mg, and 0.1 ± 0.02 kg S per ton of fruit annually. The amounts of K and N consistently accounted for the largest proportions of nutrient removal across all cultivars, with Ca, P, Mg, and S contributing smaller amounts. While N and K removal, and to a lesser degree Ca removal, varied considerably from year to year, the other macronutrients were rather stable among years (Table 2). Micronutrient removal with harvested fruit also showed significant variation, following patterns similar to macronutrients (Table 2). Except for B, micronutrient removal tended to be lower in Concord than in the winegrape cultivars. On average across the 3 yr, micronutrient losses were 6.2 ± 0.7 g B, 1.2 ± 0.07 g Mn, 1.4 ± 0.08 g Cu, and 1.0 ± 0.07 g Zn per ton of winegrapes. Juice grape harvest, on average, removed 5.3 ± 0.2 g B, 0.7 ± 0.08 g Mn, 1.0 ± 0.05 g Cu, and 0.4 ± 0.06 g Zn per ton of fruit.

View this table:
Table 2

Mineral nutrient removal per ton of harvested fruit in Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr. Removal was determined from single-vine measurements.

Pomace K return potential

The K concentration in Chardonnay pomace was 3.13 ± 0.09 and 2.31 ± 0.04 kg/t in 2021 and 2022 (p = 0.004), respectively, whereas the K concentration in Syrah pomace was 1.23 ± 0.04 and 1.99 ± 0.06 kg/t in 2021 and 2022, respectively (p < 0.001). Thus, the amount of K estimated to be potentially returned to the vineyard with pomace after juice pressing varied between cultivars and years. Despite similar K concentrations in the harvested fruit (Supplemental Table 3), Chardonnay pomace consistently had higher K content than Syrah pomace. Year-to-year variation in potential K return was significant for both cultivars.

Nutrient removal with end-of-season leaf fall

The total weight of abscised leaves varied across years (Table 1). Concord had much higher leaf DW (average 643 ± 43 g/vine) compared to the winegrape cultivars (Chardonnay 321 ± 17 g/vine, Sauvignon blanc 317 ± 30 g/vine, Syrah 423 ± 20 g/vine), with the highest values recorded in 2023. Macronutrient concentrations in senescent leaves also varied among years, with Ca being the dominant macronutrient, followed by N, then K and Mg, while P and S contributed smaller but important amounts (Figures 2B and 3B, Supplemental Table 6). The senescent leaves retained some NO3− but with the exception of Concord in 2022 and 2023, the contribution of NO3−-N to total N was usually less than 0.1%. In winegrape leaves, all macronutrients except Mg varied from year to year, though S concentrations also remained relatively stable across years. Concord leaves separated into two groups (2021 versus 2022 to 2023), clearly showing a vineyard effect; leaves were collected in different vineyards in 2021 than in the last 2 yr. Leaf N, NO3−, and K were markedly lower, and Mg was much higher, when leaves were collected in the WSU vineyard, while other macronutrients were comparatively less affected by vineyard location. Concord also stood out as having consistently higher leaf P concentrations than the winegrape cultivars. Contrasting with the fruit, the micronutrients in senescent leaves were dominated by Mn. Micronutrients in abscised leaves also varied among years (Supplemental Table 6). For instance, similar to what was observed for Ca, 2022 was an unusually low Mn year across all cultivars; Zn was also low in 2022, but only in winegrape leaves. The N and K fertilizer treatments generally did not alter the concentration of mineral nutrients in abscised leaves (p > 0.05), with few exceptions (mean ± SE, in % DW, for lowest versus highest fertilizer rate): K (0.87 ± 0.07 versus 1.12 ± 0.09, p = 0.04) and Mg (0.65 ± 0.02 versus 0.57 ± 0.02, p = 0.02) in Chardonnay; and Mg (0.4 ± 0.02 versus 0.5 ± 0.02, p = 0.02) in Syrah.

When calculated on the basis of vineyard surface area, macronutrient losses through fallen leaves after the first killing frost represented a meaningful fraction of total nutrient loss; annual Ca losses were especially high (≤58 kg/ha), followed by N (≤30 kg/ha), K (≤14 kg/ha), and Mg (≤7.5 kg/ha), while P and S losses were generally smaller (≤3 kg/ha). Losses varied significantly among years (Table 3). This variation in foliar nutrient loss per hectare reflected both the variation in leaf DW (Table 1) and that in leaf nutrient concentration (Supplemental Table 6), which was partly a result of the 25-day variation in the first frost date (Supplemental Table 2). Owing to its bigger canopy, nutrient losses with senescent leaves were generally higher in Concord compared to the winegrape cultivars. Apparent nutrient losses in Sauvignon blanc were unusually low in 2022. Though the cause of this anomaly remains unclear, visual inspection suggested that some leaves fragmented and passed through the mesh openings, leading to an underestimation of nutrient loss (see also Table 1). While the vineyard location dominated the annual variation in Concord, nutrient losses via leaves in winegrapes tended to be higher when the first frost occurred in October (2023), though there were some exceptions to this trend. Among micronutrients, Mn incurred the greatest annual losses (≤310 g/ha) due to leaf fall, followed by Zn, B, and Cu (Table 3). Additionally, Mn and B losses were especially variable from year to year and among cultivars or vineyard sites.

View this table:
Table 3

Mineral nutrient removal with abscised leaves per unit vineyard area in Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr. Leaves were collected after the first killing frost after each growing season. Removal was determined from single-vine measurements and plant spacing.

During the late 2021 growing season, different sets of Chardonnay vines were completely defoliated at monthly intervals to simulate different dates of first killing frost. The leaf nutrient concentrations showed notable changes: N and P decreased ~50% from September to November, S and Cu decreased only ~25%, while K increased and then decreased again (Table 4). However, NO3−, Ca, Mg, Zn, and B did not change significantly over the same period.

View this table:
Table 4

Mineral nutrient concentrations per unit dry weight in leaves sampled monthly for 3 mo in 2021 from different sets of fully defoliated Chardonnay grapevines in a vineyard in southeastern Washington.

Nutrient removal with pruned canes

There was marked year-to-year variation in pruning weight for Chardonnay and Syrah, while Sauvignon blanc showed no significant variation (Table 1); no pruning weight data were collected in the mechanically (pre-)pruned Concord vineyards. Macronutrient concentrations in dormant canes also varied somewhat from year to year (Supplemental Table 7). Compared with other macronutrients, P and S concentrations remained more stable across years. Though the canes contained some NO3−-N, it generally comprised <0.2% of the total N. Overall, Ca was the major nutrient in dormant canes, followed by N and K (Figure 2C). Among the micronutrients analyzed, Zn occurred at the highest concentrations in dormant canes. Micronutrients also varied across seasons (Supplemental Table 7). The N and K fertilizer treatments generally did not influence the cane nutrient concentrations (p > 0.05), although there were a few exceptions of minor magnitude (mean ± SE, in % DW except Mn in mg/kg, for lowest versus highest fertilizer rate): S (0.046 ± 0.001 versus 0.050 ± 0.001, p = 0.02) in Chardonnay; K (0.51 ± 0.01 versus 0.54 ± 0.01, p = 0.02), Ca (0.73 ± 0.02 versus 0.79 ± 0.02, p = 0.02), and Mn (18.8 ± 1.1 versus 24.2 ± 1.3, p = 0.002) in Sauvignon blanc; and K (0.64 ± 0.02 versus 0.74 ± 0.01, p < 0.001) in Syrah.

The potential nutrient removal with pruned canes varied from year to year (Table 5), mostly as a consequence of variation in pruning weight. Temporal variation was especially evident for Ca, N, and K removal, which increased in both Chardonnay and Syrah from 2021 to 2023. Ca removal was high in 2023 as a result of both higher pruning weights and higher Ca concentrations that year. Micronutrient losses followed similar trends, with Syrah and Chardonnay showing greater losses in 2023 compared to previous years. For most nutrients (Ca, Mn, and Zn were exceptions) however, the potential removal with canes was small compared with the losses due to fruit harvest, and even compared with the potential losses in abscised leaves (Table 6).

View this table:
Table 5

Mineral nutrient removal with pruned canes per unit vineyard area in Chardonnay, Sauvignon blanc, and Syrah grapevines in vineyards in southeastern Washington over 3 yr. Cane samples were collected in late winter (i.e., next year). Removal was determined from single-vine measurements and plant spacing.

View this table:
Table 6

Summary of potential annual mineral nutrient removal with fruit harvest, leaf fall, and winter pruning per unit vineyard area in Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr. Removal was determined from single-vine measurements of biomass per organ type, nutrient concentration, and plant spacing; Concord canes were not collected. Overall mean (± standard error) yields were 18.8 ± 0.8, 21.5 ± 0.7, 12.8 ± 0.5, and 29.5 ± 1.5 t/ha for Chardonnay, Sauvignon blanc, Syrah, and Concord, respectively.

Discussion

This study quantified the potential removal and loss of nutrients caused by fruit harvest, leaf abscission, and winter pruning from irrigated vineyards in arid eastern Washington over 3 yr. Our findings highlight considerable variation in nutrient export associated with growing season and the combined effects of cultivar, site, and management style. While fruit harvest was the dominant export pathway for biomass and most mineral nutrients, substantial potential losses also occurred through end-of-season leaf fall and pruned canes. Among macronutrients and when expressed per unit vineyard area, Ca loss was dominated by abscised leaves and Mg loss in leaves was similar to that with harvested fruit; among micronutrients, Mn loss was dominated by abscised leaves. Both senescent leaves and dormant canes also contributed substantially to Zn losses. Actual and simulated early fall frosts particularly hindered N and P retrieval from senescent leaves. These data validate both of our initial hypotheses: nutrient losses from vineyards extend beyond harvested fruit, and early fall frost events interrupt nutrient remobilization from leaves, increasing potential losses due to leaf abscission. When leaves and pruned canes remain in the vineyard as litter for decomposition and mineralization, or for mining by mycorrhizal fungi, their nutrients will eventually become available for uptake by vine roots again. But if they are removed from the vineyard, they must be accounted for in replacement fertilizer budgets to avoid the gradual buildup of nutrient deficiency.

Using data provided by Song et al. (2023), we estimated that the average annual removal of carbon (C) from the vineyard with the harvested fruit in our study amounted to 1.9, 2.2, 1.6, and 2.4 t/ha in Chardonnay, Sauvignon blanc, Syrah, and Concord, respectively. The C in abscised leaves accounted for only 0.2 to 0.3 t/ha in winegrapes and 0.6 t/ha in juice grapes, while the C in pruned canes accounted for 0.2 to 0.4 t/ha in winegrapes (we did not collect canes in juice grapes). For the most part, the mineral nutrient composition of the harvested fruit was quite similar among our tested cultivars, even though they were grown at different sites and managed differently. As in other studies (Rogiers et al. 2006, Schreiner et al. 2006), the nutrient composition was dominated by K (>50%) followed by N (~25%), with B being the most abundant micronutrient in all cultivars. Though Concord berries had somewhat higher K and N concentrations in their DW than winegrape berries, these differences were mostly balanced by the lower DW of Concord berries, which are typically harvested at considerably lower TSS (target 16 Brix in this study) than winegrapes (targets 22 or 24 Brix here). On average over the 3 yr, winegrape harvest removed 3 kg K and 1.4 kg N per ton of fruit annually, and juice grape harvest removed 3.6 kg K and 1.7 kg N per ton of fruit. The other macronutrients were similar across cultivars, and harvest removed 0.3 kg P, 0.4 kg Ca, 0.1 kg Mg, and 0.1 kg S per ton of fruit. These numbers are within the range of previous estimates for winegrapes; i.e., 1.5 to 4 kg K, 1 to 3 kg N, 0.2 to 0.4 kg P, 0.2 to 1 kg Ca, and 0.05 to 0.2 kg Mg per ton of fruit (Keller 2025).

In the fruit, differences in fertilizer rates (0 to 90 kg K/ha in Chardonnay, 22 to 90 kg N/ha in Sauvignon blanc and Syrah, 45 or 90 kg N/ha in Concord) applied over 3 yr slightly changed the concentration of the macronutrient that was altered by fertigation in each trial (i.e., K in Chardonnay, N in all other cultivars). Yet the variability in fruit N or K concentration due to different fertilizer rates was similar to the annual variability (which was also marked for Ca and some micronutrients). Some of the annual variation, especially for K, was related to yield variation, though fruit K tended to decrease by only 0.03% DW (Concord) to 0.2% DW (Syrah) for each 10 t/ha increase in yield. Consequently, nutrient export from vineyards at harvest appears to be mostly a function of crop yield and thus can be reasonably estimated using annual yield per block multiplied by average nutrient amounts per ton of fruit. Other sources of the annual variation in fruit nutrient concentrations, in both dry and humid climates, remain a subject for future research but may be related to differences in temperature, as well as soil water and nutrient availability, which alter vine growth, leaf and berry transpiration, and nutrient uptake by the roots. For instance, rapid shoot growth and high leaf transpiration rates favor root nutrient uptake (Keller 2025), and berry transpiration is highly dependent on vapor pressure deficit (Zhang and Keller 2015), which in turn may affect the rate of vascular flow (and hence nutrient transport) to the berries.

Our findings indicate that grape pomace has the potential to serve as a source for K recycling back to the vineyard, with the recoverable amount influenced by grape processing methods, as Syrah pomace contained less K than Chardonnay pomace. Given that the two cultivars had similar K concentrations in the harvested fruit, K was likely extracted due to skin contact during Syrah fermentation, while Chardonnay grapes were pressed before fermentation. Earlier studies found that short skin contact times (12 to 72 hr) can lead to increased K concentrations in wine (Test et al. 1986, Walker et al. 1998). In contrast, Harbertson and Harwood (2009) reported that extended skin contact durations (168 to 480 hr) resulted in a decline in wine K levels over time, with K instead accumulating in the pomace. Considering that ~25% of harvested fruit mass remains as pomace after processing (based on typical commercial extraction rates), we estimated that 10 to 20% of the K removed with harvested fruit could be returned to the vineyard through pomace recycling. Pomace residues contribute not only to meeting the nutrient demands of grapevines but also enhance soil health by supplying organic matter (Nistor et al. 2014, Doni et al. 2024, Kokkonen et al. 2025). Such additions may be especially important in arid regions, where soil organic matter is naturally very low.

In addition to fruit harvest, leaf fall after the first killing frost contributed to nutrient losses in our study. In Concord, which typically has a much larger canopy than winegrapes (e.g., Keller et al. 2023), N loss with senescent leaves reached nearly 30 kg/ha in some years, and K loss was ≤14 kg/ha. The highest yearly losses of N and K in winegrapes were only half of these amounts. Ca dominated the annual nutrient loss in abscised leaves, reaching ≤60 kg/ha, but even Mg losses could exceed 7 kg/ha and Mn losses, though highly variable, amounted to ≤0.3 kg/ha, by far the highest of any micronutrient. While high K fertilizer supply in Chardonnay slightly increased K and decreased Mg in senescent leaves, and high N supply in Syrah slightly increased leaf Mg, those effects were small compared with the yearly variation in foliar nutrient concentrations. Across all cultivars, the concentration of Ca in the leaves was low and that of K was high in 2022, which was the year with the lowest seasonal GDD accumulation (but at the winegrape site, also the highest number of hot and very hot days) and the highest seasonal rainfall. These factors may have combined to limit canopy transpiration via reduced stomatal conductance, which would have led to lower Ca uptake by the roots (Keller 2025). As Ca and K often compete for root uptake, this would also explain the high leaf K that year.

However, the timing of the first killing frost was important in determining the nutrient concentrations of abscised leaves. First, the date of the first frost varied by 25 days in our study and explained some of the annual variation in foliar nutrient concentrations. Second, simulating freeze events by manual defoliation of Chardonnay vines confirmed that leaf P concentrations declined three-fold and N concentrations by nearly 50% from September to November, while the effects on other nutrients were less pronounced or variable. The rise in leaf K after harvest is in agreement with (statistically untested) data presented by Conradie (1981), but even K declined by >30% after the postharvest spike in our study. Therefore, early frosts may curtail remobilization of nutrients from leaves and their retrieval into the vine’s perennial structures for storage, resulting in elevated nutrient concentrations in abscised leaves and, presumably, lower vine nutrient reserves for the following year. These observations are consistent with prior studies reporting that 50 to 80% of leaf N, P, and K can be remobilized under optimal senescence conditions (Keskitalo et al. 2005, Pradubsuk and Davenport 2010, Hendgen et al. 2021). In windy areas, physical displacement of fallen leaves from the vineyard floor may contribute to nutrient loss, especially following an early fall frost, and should be considered when estimating nutrient budgets for fertilization programs.

Pruned canes accounted for a third, albeit generally smaller, component of potential nutrient export. With the exception of Ca, cane macronutrient concentrations were relatively stable from year to year, while micronutrient concentrations were somewhat more variable. Across cultivars, our cane nutrient concentrations were similar to those reported for Pinot noir in Oregon, except for the lower Ca and Mg, and higher Mn, levels in Pinot noir, which was grafted on rootstock and grown in a sand-dominated medium (pH 6.2) supplied with nutrient solution (Schreiner 2021). Unlike that study, we found few effects of the N and K fertilizer treatments on cane nutrient concentrations; while they slightly altered different nutrients in different cultivars (S in Chardonnay; K, Ca, Mn in Sauvignon blanc; K in Syrah), all of these changes were minor. Because the pruning weight varied substantially among years, the variation in potential nutrient removal with pruned canes was mostly a function of the variation in pruning weight. These results emphasize the importance of considering aboveground woody biomass in nutrient budgets, particularly in years or locations of high vine vigor, and for vineyards where pruned canes are removed—unless the goal is to reduce vigor by depleting soil nutrients where vigor cannot be controlled through water deficit (c.f. Vinci et al. 2026).

In our study, K was the dominant nutrient in grape berries, accounting for >50% of total mineral content, followed by N at ~25%. This contrasts sharply with leaf tissue, where Ca was the predominant element (>50%) and N contributed ~20%. The nutrient profile of dormant canes showed a more balanced composition, with Ca, N, and K each comprising roughly 25 to 32% of total mineral content. These differences are consistent with the known mobility and transport mechanisms of the individual nutrients. For example, K is highly mobile in the phloem, allowing it to be continuously translocated into the fruit throughout berry development and ripening (Mpelasoka et al. 2003, Rogiers et al. 2017). In contrast, Ca transport occurs almost exclusively via the xylem and is driven by transpiration. Because berries have low transpiration rates, Ca is preferentially partitioned to the leaves, which have much higher transpiration demands. Additionally, grapevine roots have limited control over Ca uptake, unlike nutrients such as N or K, further contributing to its accumulation in transpiring tissues (Keller 2025).

Collectively, our findings confirm that nutrient export from vineyards is not always limited to the harvested crop, even though in most cases the fruit dominates the total amount of aboveground nutrient removal from vines. Leaf fall and winter pruning constituted additional potential nutrient losses and should be considered when estimating long-term nutrient budgets; for Ca, Mg, and Mn, leaves even dominate potential losses, as do canes for Zn. Across fruit, abscised leaves, and pruned canes of winegrapes, the average total annual N and K removal from vineyards was ~36 and 63 kg/ha, respectively. These values are in the same range as those reported by Marocke et al. (1976) for six winegrape cultivars in Alsace, France. Compared with our data, however, those authors found less export with fruit and more with vegetative tissues, likely reflecting lower yields and higher vigor in their study. In juice grapes, the total average N and K removal with fruit and leaves in our study was ~74 and 117 kg/ha, respectively. These losses are considerable and if unaccounted for could lead to nutrient mining and gradual declines in soil fertility. Sustained nutrient removal, especially from soils with low natural availability, can contribute to latent deficiencies if not monitored and corrected. This is particularly critical in dry regions where Mn, Zn, and B frequently limit perennial crop production (Marschner 2012, Kochian et al. 2015). Fertilization strategies that rely solely on fruit yield estimates may therefore be insufficient for sustaining long-term vine productivity, particularly in high-vigor systems or years of excessive vegetative growth, or in regions prone to early fall frosts combined with high winds. Nutrient management programs should integrate all loss pathways and be tailored to site-specific variables, including cultivar, annual crop yield and pruning biomass, tissue nutrient concentration, and seasonal frost timing. Where appropriate, recycling pruned canes and pomace into vineyard rows is encouraged as part of a circular nutrient economy. Though accounting for potential mineral nutrient removal from vineyards to balance inputs and outputs constitutes a valuable decision-aid tool for growers, other dynamic factors also affect the nutrient status of vineyard soils. Such factors include, but are not limited to, soil type (e.g., sandy soils being prone to leaching), soil organic matter, weather variability (e.g., rainfall leading to leaching, drought reducing nutrient availability, temperature altering mineralization rates), soil microbiome (e.g., N-fixing bacteria, arbuscular mycorrhizal fungi; see Schreiner et al. 2007), floor management and cover cropping (e.g., legumes favoring N fixation, competition reducing nutrient availability; see Tesic et al. 2007, Klodd et al. 2016), irrigation level (e.g., deficit irrigation reducing nutrient availability; see Keller 2005), and irrigation water source (e.g., minerality or salinity of ground- or surface water; see Downton 1985).

Conclusion

While harvested fruit represented the major source of nutrient removal from the wine and juice grape vineyards tested in this study, substantial quantities of nutrients were also lost annually from vines, and potentially from vineyards, through abscised leaves and pruned canes. These additional losses were particularly pronounced in years marked by early frost or increased vegetative growth. Early frost events, simulated by complete vine defoliation, disrupted foliar nutrient remobilization, leading to elevated N and P retention in fallen leaves. While the mineral nutrient composition of the fruit was dominated by K (>50%) followed by N (~25%), that of the leaves was dominated by Ca (>50%) followed by N (~20%), and that of the canes was more evenly distributed, with Ca, N, and K each making up 25 to 32%. Nutrient losses from vineyards are not limited to fruit harvest but are distributed across multiple pathways, each varying by season. Quantifying these pathways provides a more accurate foundation for developing effective, site-specific nutrient management strategies based on replacing the net export of macro- and micronutrients from a vineyard system.

CRediT Authorship Contributions

NS: Writing – Original Draft; NS and MK: Conceptualization, Methodology; NS and PD: Investigation; NS, PD, and MK: Formal Analysis; PD and MK: Writing – Review & Editing; MK: Project Administration, Resources, Supervision

Supplemental Data

The following supplemental materials are available for this article in the Supplemental tab above:

Supplemental Table 1 Dates of last sulfur (S) fungicide application and sampling for Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr. Dates for dormancy samples refer to the following year.

Supplemental Table 2 Weather conditions for vineyards at Columbia Crest (Paterson), Schilperoort Farms (Sunnyside), and Washington State University (WSU) Roza Farm over 3 yr. Weather data were obtained from the AgWeatherNet stations Paterson.W, Sunnyside.N, and Roza.2 (https://weather.wsu.edu). GDD, growing degree days.

Supplemental Table 3 Mineral nutrient concentration per unit dry weight in harvested fruit of Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr.

Supplemental Table 4 Parameters for linear regression between vineyard yield (Y) and concentrations of three macronutrients in harvested fruit of Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington. Data from 3 yr (2021 to 2023) were pooled.

Supplemental Table 5 Concentration of key macronutrients per unit dry weight (DW) in harvested fruit following addition of different fertilizer rates in Syrah, Sauvignon blanc, Concord (N), and Chardonnay (K) grapevines in vineyards in southeastern Washington. Values are means (± standard error) for the lowest and highest fertilizer rates averaged over 3 yr (2021 to 2023).

Supplemental Table 6 Mineral nutrient concentration per unit dry weight in senescent leaves of Chardonnay, Sauvignon blanc, Syrah, and Concord grapevines in vineyards in southeastern Washington over 3 yr. Leaves were collected after the first killing frost after each growing season.

Supplemental Table 7 Mineral nutrient concentration per unit dry weight in dormant canes (nodes and adjacent internode portions) of Chardonnay, Sauvignon blanc, and Syrah grapevines in vineyards in southeastern Washington over 3 yr. Cane samples were collected in late winter (i.e., next year).

Supplemental Figure 1 Relationship between yield and the concentration of three macronutrients (potassium, nitrogen, and phosphorus [K, N, P]) per unit dry weight (DW) in harvested fruit of Sauvignon blanc, Chardonnay, Syrah, and Concord grapevines in vineyards in southeastern Washington. Yield values were determined from single-vine measurements and plant spacing. Data from 3 yr (2021 to 2023) were pooled.

Data Availability

The data underlying this study are available on request from the corresponding author.

Footnotes

  • This study was funded by the USDA-NIFA Specialty Crop Research Initiative (award number 2020-51181-32159), USDA/WSDA Specialty Crop Block Grant Program (award number K3944), Washington State Grape and Wine Research Program, and Washington State Concord Grape Research Council. Ste. Michelle Wine Estates and Schilperoort Farms provided vineyards for field trials, Soiltest Farm Consultants provided services for tissue analysis, and Valley Wide Cooperative and Wilbur-Ellis donated fertilizer. We thank Lynn Mills, Alan Kawakami and Zilia Khaliullina for technical assistance and the Keller lab members for help with data collection. We also thank the Rippner lab members and Maria Mireles (Moyer lab) for field and lab work assistance, and Yun Zhang and Ryan McAdams for assistance in organizing field work at Ste. Michelle Wine Estates.

  • Shcherbatyuk N, Davadant P and Keller M. 2026. Annual nutrient removal from irrigated vineyards in dry climates. Am J Enol Vitic 77:0770009. DOI: 10.5344/ajev.2026.25041

  • By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. If you do not agree to the Disclaimers, do not download and/or accept this article.

  • Received August 2025.
  • Accepted February 2026.
  • Published online May 2026

This is an open access article distributed under the CC BY 4.0 license.

References

  1. ↵
    1. Amiri ME and
    2. Fallahi E.
    2007. Inf luence of mineral nutrients on growth, yield, berry quality, and petiole mineral nutrient concentrations of table grape. J Plant Nutr 30:463-470. DOI: 10.1080/01904160601172031
    OpenUrlCrossRef
  2. ↵
    1. Arrobas M,
    2. Ferreira IQ,
    3. Freitas S,
    4. Verdial J and
    5. Rodrigues MÂ.
    2014. Guidelines for fertilizer use in vineyards based on nutrient content of grapevine parts. Sci Hortic 172:191-198. DOI: 10.1016/j.scienta.2014.04.016
    OpenUrlCrossRef
  3. ↵
    1. Assimakopoulou A and
    2. Tsougrianis C.
    2012. Correlation between yield, must attributes and nutritional status of the Greek red wine grape variety “Agiorgitiko.” J Plant Nutr 35:1022-1036. DOI: 10.1080/01904167.2012.671405
    OpenUrlCrossRef
  4. ↵
    1. Balzergue C,
    2. Dartevelle T,
    3. Godon C,
    4. Laugier E,
    5. Meisrimler C,
    6. Teulon J-M et al
    . 2017. Low phosphate activates STOP1-ALMT1 to rapidly inhibit root cell elongation. Nat Commun 8:15300. DOI: 10.1038/ncomms15300
    OpenUrlCrossRefPubMed
  5. ↵
    1. Cass A.
    2005. Effects of soil physical characteristics on mineral nutrient availability, movement, and uptake. In Proceedings for the Soil Environment and Vine Mineral Nutrition Symposium, American Society for Enology and Viticulture, pp. 3-11. San Diego, CA.
  6. ↵
    1. Christensen LP and
    2. Peacock WL.
    2000. Mineral nutrition and fertilization. In Raisin Production Manual. Christensen LP (ed.), pp. 102-114. University of California, Agriculture and Natural Resources, Oakland, CA.
  7. ↵
    1. Conradie WJ.
    1981. Nutrient consumption by Chenin blanc grown in sand culture and seasonal changes in the chemical composition of leaf blades and petioles. S Afr J Enol Vitic 2:15-18. DOI: 10.21548/2-1-2404
    OpenUrlCrossRef
  8. ↵
    1. Cooper RM and
    2. Williams JS.
    2004. Elemental sulphur as an induced antifungal substance in plant defense. J Exp Bot 55:1947-1953. DOI: 10.1093/jxb/erh179
    OpenUrlCrossRefPubMed
  9. ↵
    1. Doni S,
    2. Masciandaro G,
    3. Macci C,
    4. Manzi D,
    5. Mattii GB,
    6. Cataldo E et al
    . 2024. Zeolite and winery waste as innovative by-product for vineyard soil management. Environments 11:29. DOI: 10.3390/environments11020029
    OpenUrlCrossRef
  10. ↵
    1. Downton WJS.
    1985. Growth and mineral composition of the Sultana grapevine as influenced by salinity and rootstock. Aust J Agric Res 36:425-434. DOI: 10.1071/AR9850425
    OpenUrlCrossRef
  11. ↵
    1. Eichert T and
    2. Fernández V.
    2023. Uptake and release of elements by leaves and other aerial plant parts. In Marschner’s Mineral Nutrition of Plants. 4th ed. Rengel Z et al. (eds.), pp. 105-129. Elsevier Academic Press, London. DOI: 10.1016/C2019-0-00491-8
    OpenUrlCrossRef
  12. ↵
    1. Fennell A.
    2004. Freezing tolerance and injury in grapevines. J Crop Improv 10:201-235. DOI: 10.1300/J411v10n01_09
    OpenUrlCrossRef
  13. ↵
    1. Gavlak R,
    2. Horneck D,
    3. Miller RO and
    4. Kotuby-Amacher J.
    2003. Soil, Plant and Water Reference Methods for the Western Region. WCC-103 Publication, Fort Collins, CO.
  14. ↵
    1. Guo Y and
    2. Gan S.
    2005. Leaf senescence: Signals, execution, and regulation. Curr Top Dev Biol 71:83-112. DOI: 10.1016/S0070-2153(05)71003-6
    OpenUrlCrossRefPubMed
  15. ↵
    1. Harbertson JF and
    2. Harwood ED.
    2009. Partitioning of potassium during commercial-scale red wine fermentations and model wine extractions. Am J Enol Vitic 60:43-49. DOI: 10.5344/ajev.2009.60.1.43
    OpenUrlAbstract/FREE Full Text
  16. ↵
    1. Hendgen M,
    2. Schubert S and
    3. Löhnertz O.
    2021. Ethephon‐induced leaf senescence increases the concentration of yeast‐assimilable nitrogen in grape must (Vitis vinifera cv. Riesling). Aust J Grape Wine Res 27:472-482. DOI: 10.1111/ajgw.12500
    OpenUrlCrossRef
  17. ↵
    1. Keller M.
    2005. Deficit irrigation and vine mineral nutrition. Am J Enol Vitic 56:267-283. DOI: 10.5344/ajev.2005.56.3.267
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Keller M.
    2025. The Science of Grapevines. 4th ed. Elsevier Academic Press, London.
  19. ↵
    1. Keller M,
    2. Mills LJ and
    3. Kawakami AK.
    2023. Optimizing irrigation for mechanized Concord juice grape production. Am J Enol Vitic 74:0740008. DOI: 10.5344/ajev.2022.21050
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Keskitalo J,
    2. Bergquist G,
    3. Gardeström P and
    4. Jansson S.
    2005. A cellular timetable of autumn senescence. J Plant Physiol 139:16350-1648. DOI: 10.1104/pp.105.066845
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Klodd AE,
    2. Eissenstat DM,
    3. Wolf TK and
    4. Centinari M.
    2016. Coping with cover crop competition in mature grapevines. Plant Soil 400:391-402. DOI: 10.1007/s11104-015-2748-2
    OpenUrlCrossRef
  22. ↵
    1. Kochian LV,
    2. Piñeros MA,
    3. Liu J and
    4. Magalhaes JV.
    2015. Plant adaptation to acid soils: The molecular basis for crop aluminum resistance. Annu Rev Plant Biol 66:571-598. DOI: 10.1146/annurev-arplant-043014-114822
    OpenUrlCrossRefPubMed
  23. ↵
    1. Kokkonen AA,
    2. Tiecher TL,
    3. Loss A,
    4. Toselli M,
    5. Garlet LP,
    6. Marques ALL et al
    . 2025. The use of grape pomace residues as a nutrient source in subtropical viticulture. Agronomy 15:1010. DOI: 10.3390/agronomy15051010
    OpenUrlCrossRef
  24. ↵
    1. Kwasniewski MT,
    2. Sacks GL and
    3. Wilcox WF.
    2014. Persistence of elemental sulfur spray residue on grapes during ripening and vinification. Am J Enol Vitic 65:453-462. DOI: 10.5344/ajev.2014.14027
    OpenUrlAbstract/FREE Full Text
  25. ↵
    1. Lacroux F,
    2. Trégoat O,
    3. van Leeuwen C,
    4. Pons A,
    5. Tominaga T,
    6. Lavigne-Cruège V et al
    . 2008. Effect of foliar nitrogen and sulphur application on aromatic expression of Vitis vinifera L. cv. Sauvignon blanc. OENO One 42:125-132. DOI: 10.20870/oeno-one.2008.42.3.816
    OpenUrlCrossRef
  26. ↵
    1. Lim PO,
    2. Kim HJ and
    3. Nam HG.
    2007. Leaf senescence. Annu Rev Plant Biol 58:115-136. DOI: 10.1146/annurev.arplant.57.032905.105316
    OpenUrlCrossRefPubMed
  27. ↵
    1. Ludemann CI,
    2. Wanner N,
    3. Chivenge P,
    4. Dobermann A,
    5. Einarsson R,
    6. Grassini P et al
    . 2024. A global FAOSTAT reference database of cropland nutrient budgets and nutrient use efficiency (1961-2020): Nitrogen, phosphorus and potassium. Earth Syst Sci Data 16:525-541. DOI: 10.5194/essd-16-525-2024
    OpenUrlCrossRef
  28. ↵
    1. Marocke R,
    2. Balthazard J and
    3. Corrège G.
    1976. Exportations en éléments fertilisants des principaux cépages cultivés en Alsace. C R Séanc Acad Agric France 62:420-429. https://gallica.bnf.fr/ark:/12148/bpt6k64883190/f427.item
    OpenUrl
  29. ↵
    1. Marschner H.
    2012. Mineral Nutrition of Higher Plants. 3d ed. Elsevier Academic Press, London. DOI: 10.1016/C2009-0-63043-9
    OpenUrlCrossRef
  30. ↵
    1. Mpelasoka BS,
    2. Schachtman DP,
    3. Treeby MT and
    4. Thomas MR.
    2003. A review of potassium nutrition in grapevines with special emphasis on berry accumulation. Aust J Grape Wine Res 9:154-168. DOI: 10.1111/j.1755-0238.2003.tb00265.x
    OpenUrlCrossRef
  31. ↵
    1. Niinemets Ü and
    2. Tamm Ü.
    2005. Species differences in timing of leaf fall and foliage chemistry modify nutrient resorption efficiency in deciduous temperate forest stands. Tree Physiol 25:1001-1014. DOI: 10.1093/treephys/25.8.1001
    OpenUrlCrossRefPubMed
  32. ↵
    1. Niklas KJ.
    2006. Plant allometry, leaf nitrogen and phosphorus stoichiometry, and interspecific trends in annual growth rates. Ann Bot 97:155-163. DOI: 10.1093/aob/mcj021
    OpenUrlCrossRefPubMed
  33. ↵
    1. Nistor E,
    2. Dobrei A,
    3. Dobrei A,
    4. Kiss E and
    5. Ciolac V.
    2014. Grape pomace as fertilizer. J Hortic Forest Biotechnol 18:141-145.
    OpenUrl
  34. ↵
    1. Pradubsuk S and
    2. Davenport JR.
    2010. Seasonal uptake and partitioning of macronutrients in mature ‘Concord’ grape. J Am Soc Hortic Sci 135:474-483. DOI: 10.21273/JASHS.135.5.474
    OpenUrlCrossRef
  35. ↵
    1. Pradubsuk S and
    2. Davenport JR.
    2011. Seasonal distribution of micronutrients in mature ‘Concord’ grape: Boron, iron, manganese, copper, and zinc. J Am Soc Hortic Sci 136:69-77. DOI: 10.21273/JASHS.136.1.69
    OpenUrlCrossRef
  36. ↵
    1. Rapp A and
    2. Versini G.
    1996. Influence of nitrogen compounds in grapes on aroma compounds of wines. Dev Food Sci 51:193-203. DOI: 10.1016/S0167-4501(06)80257-8
    OpenUrlCrossRef
  37. ↵
    1. Rogiers SY,
    2. Greer DH,
    3. Hatfield JM,
    4. Orchard BA and
    5. Keller M.
    2006. Mineral sinks within ripening grape berries (Vitis vinifera L.). Vitis 45:115-123. DOI: 10.5073/vitis.2006.45.115-123
    OpenUrlCrossRef
  38. ↵
    1. Rogiers SY,
    2. Coetzee ZA,
    3. Walker RR,
    4. Deloire A and
    5. Tyerman SD.
    2017. Potassium in the grape (Vitis vinifera L.) berry: Transport and function. Front Plant Sci 8:1629. DOI: 10.3389/fpls.2017.01629
    OpenUrlCrossRef
  39. ↵
    1. Schreiner RP.
    2005. Spatial and temporal variation of roots, arbuscular mycorrhizal fungi, and plant and soil nutrients in a mature Pinot noir (Vitis vinifera L.) vineyard in Oregon, USA. Plant Soil 276:219-234. DOI: 10.1007/s11104-005-4895-0
    OpenUrlCrossRef
  40. ↵
    1. Schreiner RP.
    2021. Utility of dormant season pruning wood to predict nutrient status of grapevines. J Plant Nutr 44:238-251. DOI: 10.1080/01904167.2020.1806311
    OpenUrlCrossRef
  41. ↵
    1. Schreiner RP,
    2. Scagel CF and
    3. Baham J.
    2006. Nutrient uptake and distribution in a mature ‘Pinot noir’ Vineyard. HortScience 41:336-345. DOI: 10.21273/hortsci.41.2.336
    OpenUrlAbstract/FREE Full Text
  42. ↵
    1. Schreiner RP,
    2. Tarara JM and
    3. Smithyman RP.
    2007. Deficit irrigation promotes arbuscular colonization of fine roots by mycorrhizal fungi in grapevines (Vitis vinifera L.) in an arid climate. Mycorrhiza 17:551-562. DOI: 10.1007/s00572-007-0128-3
    OpenUrlCrossRefPubMed
  43. ↵
    1. Skinner PW and
    2. Matthews MA.
    1989. Reproductive development in grape (Vitis vinifera L.) under phosphorus-limited conditions. Sci Hortic 38:49-60. DOI: 10.1016/0304-4238(89)90019-8
    OpenUrlCrossRef
  44. ↵
    1. Skinner PW,
    2. Cook JA and
    3. Matthews MA.
    1988. Responses of grapevine cvs Chenin blanc and Chardonnay to phosphorus fertilizer applications under phosphorus-limited soil conditions. Vitis 27:95-109. DOI: 10.5073/vitis.1988.27.95-109
    OpenUrlCrossRef
  45. ↵
    1. Song R,
    2. Zhu Z,
    3. Zhang L,
    4. Li H and
    5. Wang H.
    2023. A simple method using an allometric model to quantify the carbon sequestration capacity in vineyards. Plants 12:997. DOI: 10.3390/plants12050997
    OpenUrlCrossRef
  46. ↵
    1. Spayd SE and
    2. Morris JR.
    1978. Influence of irrigation, pruning severity, and nitrogen on yield and quality of ‘Concord’ grapes in Arkansas. J Am Soc Hortic Sci 103:211-216. DOI: 10.21273/jashs.103.2.211
    OpenUrlCrossRef
  47. ↵
    1. Tesic D,
    2. Keller M and
    3. Hutton RJ.
    2007. Influence of vineyard floor management practices on grapevine vegetative growth, yield, and fruit composition. Am J Enol Vitic 58:1-11. DOI: 10.5344/ajev.2007.58.1.1
    OpenUrlAbstract/FREE Full Text
  48. ↵
    1. Test SL,
    2. Noble AC and
    3. Schmidt JO.
    1986. Effect of pomace contact on Chardonnay musts and wines. Am J Enol Vitic 37:133-136. DOI: 10.5344/ajev.1986.37.2.133
    OpenUrlAbstract/FREE Full Text
  49. ↵
    1. Tulloch HW and
    2. Harris WB.
    1970. Fertilizer responses with non-irrigated Shiraz grapevines, 1944-1966. Aust J Agric Res 21:243-252. DOI: 10.1071/AR9700243
    OpenUrlCrossRef
  50. ↵
    1. Verdenal T,
    2. Dienes-Nagy Á,
    3. Spangenberg JE,
    4. Zufferey V,
    5. Spring J-L,
    6. Viret O et al
    . 2021. Understanding and managing nitrogen nutrition in grapevines: A review. OENO One 55:1-43. DOI: 10.20870/oeno-one.2021.55.1.3866
    OpenUrlCrossRef
  51. ↵
    1. Vinci G,
    2. Calderan A,
    3. Lodovici A,
    4. Fasoli M,
    5. Sivilotti P and
    6. Zanin L.
    2026. Grapevine responses to water deficit and N x K fertilization: Seasonal variation of ‘Cabernet Sauvignon’ and ‘Grenache’ physiology. Environ Exp Bot 241:106294. DOI: 10.1016/j.envexp-bot.2025.106294
    OpenUrlCrossRef
  52. ↵
    1. Vrignon-Brenas S,
    2. Fontez B,
    3. Bisson A,
    4. Rolland G,
    5. Chopard J,
    6. Fumey D et al
    . 2022. Quantification of the pluriannual dynamics of grapevine growth responses to nitrogen supply using a Bayesian approach. J Exp Bot 73:1385-1401. DOI: 10.1093/jxb/erab469
    OpenUrlCrossRef
  53. ↵
    1. Walker RR,
    2. Clingeleffer PR,
    3. Kerridge GH,
    4. Rühl EH,
    5. Nicholas PR and
    6. Blackmore DH.
    1998. Effects of the rootstock Ramsey (Vitis champini) on ion and organic acid composition of grapes and wine, and on wine spectral characteristics. Aust J Grape Wine Res 4:100-110. DOI: 10.1111/j.1755-0238.1998.tb00138.x
    OpenUrlCrossRef
  54. ↵
    1. Wetselaar R and
    2. Farquhar GD.
    1980. Nitrogen losses from tops of plants. Adv Agron 33:263-302. DOI: 10.1016/S0065-2113(08)60169-8
    OpenUrlCrossRef
  55. ↵
    1. Whitehead DC.
    2000. Nutrient Elements in Grassland: Soil-Plant-Animal Relationships. CABI Publishing, Wallingford, UK.
  56. ↵
    1. Williams LE.
    1991. Vine nitrogen requirements – Utilization of N sources from soils, fertilizers, and reserves. In Proceedings of the International Symposium on Nitrogen in Grapes and Wine. Rantz JM (ed.), pp. 62-66. American Society for Enology and Viticulture, Davis, CA.
  57. ↵
    1. Wise A,
    2. Martinson TE,
    3. Hawk J,
    4. Weigle T and
    5. Tarleton L.
    2007. The New York Guide to Sustainable Viticulture Practices: Grower Self-Assessment Workbook. Cornell Cooperative Extension, Ithaca, NY.
  58. ↵
    1. Xiao J,
    2. Kao-Kniffin J and
    3. Zhu B.
    2025. Arbuscular mycorrhizal fungi enhance nitrogen acquisition from, but not carbon loss of, organic matter in soil. New Phytol 247:1415-1425. DOI: 10.1111/nph.70274
    OpenUrlCrossRef
  59. ↵
    1. Yuan F,
    2. Schreiner RP,
    3. Osborne J and
    4. Qian MC.
    2018. Effects of soil NPK supply on Pinot noir wine phenolics and aroma composition. Am J Enol Vitic 69:371-385. DOI: 10.5344/ajev.2018.17077
    OpenUrlAbstract/FREE Full Text
  60. ↵
    1. Zhang Y and
    2. Keller M.
    2015. Grape berry transpiration is determined by vapor pressure deficit, cuticular conductance, and berry size. Am J Enol Vitic 66:454-462. DOI: 10.5344/ajev.2015.15038
    OpenUrlAbstract/FREE Full Text
Back to top
Print
View full PDF
Email Article
Citation Tools

Open Access
Annual Nutrient Removal from Irrigated Vineyards in Dry Climates
Nataliya Shcherbatyuk, Pierre Davadant, Markus Keller
Am J Enol Vitic.  2026  77: 0770009  ; DOI: 10.5344/ajev.2026.25041
del.icio.us logo Twitter logo Facebook logo Mendeley logo
Save to my folders

Jump to section