Effects of Preharvest Deficit Irrigation and Postharvest Rewatering on Leaf Physiology, Carbon Partitioning, and Fruit Productivity of Sweet Cherry cv. ‘Regina’ under Field Conditions

1. Introduction

Cherries (Prunus avium L.) are highly demanded by consumers for their organoleptic and nutritional characteristics [1,2]. Although world production of fresh cherries has increased 42% in the last decade [3], many cultivated areas are facing increasing water scarcity for irrigation associated with global climate change [4,5]. To mitigate the impact of limited water availability on fruit production, several strategies have been proposed, encompassing from crop breeding to agronomical practices such as regulated deficit irrigation and the use of biostimulants and growth regulators [6,7,8,9].

From a physiological point of view, plants have evolved several mechanisms that allow them to adjust growth and leaf functioning under water-limited conditions. Leaf water potential (Ψ_L) is one of the most widely used indicators of plant water status since it integrates soil water availability, atmospheric demand and plant hydraulic functioning (e.g., [10]). As water deficit develops, reductions in Ψ_L has been largely related to stomatal closure and lower photosynthesis rates due to less CO₂ diffusion into carboxylation sites [11,12]. However, reductions in Ψ_L do not always reflect in a proportional decrease in gas exchange, suggesting that factors other than plant water status may contribute to regulation of leaf physiological activity under water deficit conditions.

In this regard, considerable evidence shows that carbon source-sink relationships may also modulate photosynthetic activity and stomata regulation through changes in assimilate export, carbohydrate partitioning and osmotic adjustment (e.g., [13]). Indeed, responses to water deficit may vary according to sink demand as previously reported fruit-bearing sweet cherry branches by Jorquera-Fontena et al. [14]. In parallel, prolonged water deficit may induce changes in leaf structural traits, including changes in CO₂ diffusion and photosynthetic performance [15,16,17].

Sweet cherry has been reported to be sensitive to preharvest water deficit [18]; however, its effect on productive and vegetative performance is primarily determined by the severity of water limitation and scion/rootstock combination’s ability to tolerate the lack of water. In this line, Blanco et al. [19] found that sustained deficit irrigation during preharvest (at 85% of crop evapotranspiration) had no effect on yield and some physiological traits of ‘Prime Giant’ sweet cherry but stem water potential and shoot growth significantly fell. In similar way, Küçükyumuk [20] reported that a fruit yield in ‘0900 Ziraat’ sweet cherry was unaffected by sustained deficit irrigation at 50% of field capacity applied from 30 days after bloom to leaf fall in the first year of study, but reduced trunk cross-sectional area and stomatal conductance. Furthermore, Toro et al. [21] showed contrasting transpiration responses among different scion/rootstock combinations and highlighted the significance of sucrose and sorbitol mediating the drought responses in young cherry trees. Collectively, these studies indicate that sweet cherry responses to water deficit are dependent on both irrigation intensity and genotype and suggest that physiological acclimation seems to involve both water-status and carbon-related processes.

While existing studies have documented the effect of deficit irrigation on sweet cherry water relations, gas exchange and productivity [19,22,23], considerably less attention has been given to the relationships among tree water status, carbon partitioning and leaf structural adjustments during preharvest period. Moreover, little information is available on how these responses evolve following fruit harvest and irrigation recovery, despite major changes in sink demand that occur during this stage.

The aim of this study was to assess how preharvest irrigation deficit affects the coordination between tree water status, gas exchange, carbon partitioning, leaf structural traits and productivity in field-grown sweet cherry cv. ‘Regina’ grafted onto ‘Gisela^® 6’. In addition, postharvest physiological responses following irrigation recovery were assessed to gain insight into the role of changing fruit sink demand on leaf physiology.

2. Materials and Methods

2.1. Experimental Site and Plant Material

The experiment was carried out during the 2023/24 season in a commercial orchard located in Lumaco, La Araucanía Region, Chile (38°06’28.5” S, 72°51’45.0” W). The site is characterized by warm and dry climatic conditions from November to March. The soil is classified as an Inceptisol (Lumaco series, Fluventic Dystrudepts) [24]. It is a deep, gently sloping, moderately slow permeability, and well drained with a sandy clay loam texture (46% sand, 26% silt, and 28% clay) [25]. The Lumaco soil (0-20 cm depth) contains 4% organic matter, a slightly acidic pH (pH_W 6.29), and a field capacity and wilting point of 0.25 m³ m⁻³ and 0.15 m³ m⁻³, respectively.

The plant material used consisted of 7-year-old cherry trees cv. ‘Regina’ grafted onto ‘Gisela^® 6’, established at a spacing of 4 × 1 m in two north–south oriented rows and using a tall spindle axis training system. Pest control was carried out following the technical recommendations of the export fruit industry. The preharvest fertilization requirements were supplied via fertigation before applying the deficit irrigation treatment (see below).

2.2. Treatments

A total of 36 trees distributed among two orchard rows (18 trees per row) were selected and assigned to two irrigation treatments. Trees were chosen based on similar trunk cross-sectional area to ensure homogeneous vigor and minimize within-orchard variability. The irrigation treatments were a full irrigated control (FI), receiving 100% of crop evapotranspiration (ETc), and a deficit irrigated treatment (DI), receiving 55% of ETc, as described by Blanco et al. [19] and Jorquera-Fontena et al. [14]. To reach these conditions, trees were irrigated at two-days interval using two irrigation lines per row with drippers spaced at 70 cm, with a flow rate of 4.7 L h^-1 for FI and 2.5 L h^-1 for DI. The DI was applied from the second rapid fruit growth stage (BBCH 76; November 20, 2023) until four days after harvest (January 8, 2024), from when trees were water-recovered at 100% ETc. Irrigation treatments were separated by two buffer trees.

The ETc was determined according to FAO 56 [26], and the reference evapotranspiration obtained from the nearest agroclimatic station (Gaby Ranquilco) located 4 km from the experimental site (https://agrometeorologia.cl). Recorded mean temperatures, reference evapotranspiration (ETo) and rainfall are shown in Figure 1. As observed, four precipitation events were recorded accumulating 17.7 mm, during the study (Figure 1).

2.3. Measurements

2.3.1. Soil Water Content

Volumetric water content was recorded throughout the season using soil moisture/temperature sensors (Hobo model RXW-T11-922, Massachusetts, USA) installed at 30 cm depth and spaced at 50 cm from four representative trees of each treatment [27].

2.3.2. Leaf Sampling

Full expanded, healthy and sun-exposed leaves were randomly selected on 2-year-old productive branches located in the mid-portion of the trees. Sampling followed a fixed sequence to ensure consistency among measured variables. The exception was stem water potential (Ψ_S), which was assessed on separate leaves collected from the same canopy position and measured at seven-day intervals throughout the experimental period according to an independent sampling schedule. Measurements of the remaining variables were conducted before harvest at 1, 9, 24, and 37 days after the beginning of treatments (DABT), and again seven days after harvest (51 DABT). Prior to the final assessment, trees were rewatered at 100% of the irrigation treatment for three days.

For all measurements excepting Ψ_S, one leaf was sampled from 18 randomly selected trees within irrigation treatments on every sampling date. The Ψ_S was evaluated on all experimental trees along the study. Gas exchange was measured first, followed immediately by leaf water potential (Ψ_L) determination on the same leaf. This sequence was consistently applied across all sampling dates excepting at 37 and 51 DABT, where chlorophyll index was additionally measured between gas exchange and Ψ_L determinations. On these dates, after Ψ_L measurement, two 2 cm² sections were collected from each measured leaf: one for leaf mass per unit area determination and the other for simultaneous assessment of lignin fluorescence and stomatal density. In addition, at 37 DABT, four extra 2 cm² leaf sections were collected from the same leaf for soluble sugar analysis.

2.3.3. Gas Exchange Variables

Gas exchange variables were measured with a Portable Photosynthesis System LI-6800 equipped with a fluorometer (LI-COR, Inc., NE, USA). The leaf chamber conditions were established at 1500 µmol m^-2 s^-1 PAR saturation light, at 400 µmol mol^-1 of CO₂ concentration, at 21 °C of air temperature with 60% relative air humidity, and with an air flow of 700 µmol s^-1. Measurements were carried out after the leaves reached the steady state, between 10:00 h and 16:00 h of the day. The mesophyll conductance (gm) was estimated according to Harley et al. [28] as:

$$\mathrm{gm} ={\displaystyle \frac{\mathrm{A}}{\mathrm{Ci} - \frac{\mathsf{\Gamma }* \text{×}( \mathrm{J}+8 \text{×} \left(\mathrm{A} + {\mathrm{R}}_{\mathrm{L}}\right))}{\mathrm{J} - 4 \text{×} (\mathrm{A} + {\mathrm{R}}_{\mathrm{L}})}}}(\mathrm{Equation} 1).$$

where J is the electron transport rate, Г* is the non-photorespiratory CO₂ compensation point, and R_L is the day respiration. J was derived from the electron transport rate (ETR) calculated from chlorophyll fluorescence measurements using the instrument’s default equations, the Г* was taken from Bernacchi et al. [29] and R_L was derived from measurements of dark respiration (R_d) such as ${\mathrm{R}}_{\mathrm{L}}={\mathrm{R}}_{\mathrm{d}}/2$ [30]. Briefly, R_d was determined by additional gas exchange measurement where six leaves from each treatment were used. For determinations, conditions in the leaf chamber were similar to those mentioned above but using a PPFD set at 0 μmol m⁻² s⁻¹ PAR. Leaves were exposed to dark in the chamber conditions for 5 min before measuring.

2.3.4. Ψ_L and Ψ_S Determination

The Ψ_L and Ψ_S were determined using a Scholander pressure chamber PMS 600 (PMS Instruments, Corvallis, OR). For Ψ_L, the same leaves previously used for gas exchange measurements were detached, wrapped in a wet paper towel to minimize dehydration, and immediately measured in the pressure chamber [31,32]. For Ψ_S leaves were enclosed with an opaque plastic bag for 1 h before measurement, as suggested by Choné et al. [33].

2.3.5. Leaf Mass per Unit Area (LMA)

The LMA was determined in a 2 cm² piece of leaf, avoiding the central vein. Subsamples were then stored in a cooler to be transported to the laboratory to obtain its dry mass by drying for 48 h at 60 °C in an oven. The LMA was calculated as the ratio of leaf dry mass (g) to leaf area (m²).

2.3.6. Lignin Fluorescence and Stomatal Density

Subsamples were stored in the FAA fixative solution (formaldehyde, alcohol, acetic acid) in the field. In the laboratory, leaf sections were stained with safranin O and analyzed by confocal laser scanning microscopy (CLSM; Olympus FV1000, Arquimed, Japan) at an emission/excitation λ of 543/590 nm according to Sant’Anna et al. [34] to visualize the stomata and lignin distribution. To quantify lignin fluorescence and stomata density the obtained images were processed by using image processing software (FV10-ASW software v0.200c; Arquimed).

2.3.7. Soluble Sugars

Leaf subsamples were immediately frozen in liquid N₂ and stored at −20 °C in the laboratory. Frozen leaves were ground to a fine powder under liquid N₂. Soluble sugars were extracted from 0.1 g of fresh leaf tissue with 1.5 mL of chromatographic-grade water. Samples were vortexed for 30 s, shaken for 30 min at room temperature, centrifuged at 10,000 rpm for 10 min, and filtered through a 0.22 µm membrane [35]. Supernatant was subsequently analyzed using an HPLC system (MODEL LC 2050C, JAPAN) with a RID-20A refractive index detector, following Usenik et al. [35] protocol. Briefly, a Hi-Plex Pb 300 x 7.7 mm column was used for separation, operated at 65 °C, with HPLC grade water. The injection volume was 20 μL, the flow rate was 0.6 mL/min, and the run time was 30 minutes. Quantification of the identified compounds, including sucrose, fructose, glucose, and sorbitol, was based on peak area expressed in mg g^-1 on fresh weight basis (FW).

2.3.8. Yield, Fruit Quality and Water Productivity

Six representative trees per treatment were harvested at 44 DABT and yield was expressed as kg tree^-1. Subsequently, a 1 kg sample was randomly selected from each replicate and stored in a cooler for further analysis. In the laboratory, samples were stored for 12 h at 4 °C and then fruit diameter was determined using a digital caliper (±0.01 mm). A total of 54 fruits per treatment were analyzed. Water productivity was calculated as the ratio between fruit yield (kg ha^-1) and total water supply (Irrigated plus precipitation; m³ ha^-1) accumulated from the beginning of the irrigation season (November 1, 2023) to the harvest.

2.4. Experimental Design and Data Analysis

Irrigation treatments were randomly assigned to 36 trees distributed across two parallel orchard rows, with nine trees per irrigation treatment in each row. Trees subjected to contrasting irrigation regimes were separated by two buffer trees. For statistical analyses, each tree was treated as an experimental unit following a completely randomized design. The data analysis was done by using the JAMOVI 2.3.16 statistical program. First, data were subjected to the analysis of normality and equality of variances using the Shapiro-Wilk and Levene tests, respectively. To compare treatments in each measurement date, a Welch’s t-test was performed when the variable had normality, but unequal variances. If the opposite occurred, a Mann-Whitney U test was performed. For variables having normality and homogeneity of variances, Student’s t-test was performed (p ≤ 0.05). Subsequently, the variables were analyzed in the time course using an analysis of variance (ANOVA) and Tukey’s multiple comparison test (p ≤ 0.05). For non-parametric variables, a Kruskal-Wallis ANOVA was used with a Dwass-Steel-Critchlow-Fligner two-by-two comparison test.

3. Results

3.1. Soil Water Content and Stem Water Potential

Throughout the experiment, the irrigation supply corresponded to 1326 and 707 m³ ha⁻¹ for FI and DI, respectively. The sensors successfully recorded temporal variation in soil water content between treatments (Figure 2).

Soil water content under DI remained lower than under FI throughout the experiment, while remaining above the permanent wilting point. The largest differences occurred between 35 and 50 DABT, when DI and FI averaged approximately 0.17-0.18 and 0.21-0.23 m³ m-³, respectively.

The Ψ_S progressively decreased in both treatments during the experiment, although values under DI were consistently lower than under FI (Figure 3). Significant differences were detected from 8 DABT onwards. The decrease in Ψ_S was more pronounced under DI reaching a minimum about of -1.7 MPa at 38-44 DABT, whereas FI trees stabilized near to -1.3 MPa during the same period. Following rewatering (51 DABT), Ψ_S was recovered with respect to immediately previous measurement, reaching similar values between treatments. Despite irrigation in FI treatment was supplied as demanded along the experiment, Ψ_S also increased at 51 DABT (Figure 3).

3.2. Leaf Water Potential and Gas Exchange Parameters

Deficit irrigation treatment induced a marked reduction in Ψ_L, averaging 32% lower values than FI throughout the experiment (-1.24 and -0.94 MPa for DI and FI, respectively) before rewatering. The Ψ_L significantly differed between treatments at each sampling date and exhibited a significant temporal response within treatment (Figure 4A).

The largest treatments differences occurred between 24 and 37 DABT, when Ψ_L in DI declined to values below -1.5 MPa, reaching a minimum close to -1.8 MPa at 38 DABT. Following rewatering, Ψ_L increased by about 31% in DI and 18% in FI relative to immediately previous sampling date. However, values remained below those measured at 1 DABT, being 34% and 14% lower in DI and FI (Figure 4A).

The DI consistently reduced A_SAT compared to FI, resulting in values that were 25% lower on average throughout deficit irrigated period (Figure 4B). Under DI, A_SAT decreased during the first 24 DABT, and then again from 37 DABT until 51 DABT. Conversely, FI trees exhibited a progressive increase in A_SAT, reaching values close to 50% higher at 37 DABT than at beginning of the experiment. Following postharvest rewatering, A_SAT in FI returned to rates comparable to those found at treatment imposition, while in DI it remained about 30% lower than the initial value (Figure 4B).

Changes in A_SAT were accompanied by coordinated responses of the other gas exchange variables (Figure 4). Prior to rewatering DI reduced gs, Ci, and gm by 37%, 12%, and 13%, respectively, compared to FI, while increasing WUEi by approximately 18%. After rewatering, gs fell by about 25% and 42% in DI and FI, respectively, relative to the previous sampling date (Figure 4C). Likewise, gm decreased by about 42% and 34% in DI and FI, respectively, whereas Ci remained relatively stable in DI but decreased by 10% in FI (Figure 4D and F). The WUEi increased by 20% in FI and remained unchanged in DI (Figure 4E).

The Ψ_L explained a small fraction of the A_SAT and gs variability (Figure 5A and B). In contrast, A_SAT was strongly correlated to gs (Figure 5C), while Ci was positively correlated to gs (Figure 5D). The gm also contributed to A_SAT variation, showing a moderate relationship between them, although gm was poorly related to Ci (Figure 5E and F).

3.3. Structural and Anatomical Parameters

Complementary measurements of lignin, stomatal density, LMA and SPAD were conducted at 37 DABT and after postharvest rewatering (51 DABT) to provide additional insight into the responses associated with changes in leaf water status and gas exchange (Table 1).

Lignin as expressed in relative fluorescence units was 8% lower in DI than in FI at 37 DABT, although this difference disappeared after rewatering due to a 19% increase under DI. The LMA increased from 37 DABT to 51 DABT in both treatments and was 8% higher in DI than FI after rewatering. In contrast, stomatal density and chlorophyll index were unaffected by irrigation treatment at either sampling date.

3.4. Leaf Soluble Sugars

Leaf soluble sugars were determined at 37 DABT, coinciding with the period of maximum treatment differences in Ψ_L and gas exchange variables (Table 2).

Sorbitol and fructose were the predominant soluble sugars, accounting for approximately 41% and 35% of the total pool, respectively. The DI increased sorbitol content by 27% and reduced sucrose content by 57%. In contrast, glucose, fructose and total soluble sugar did not differ between irrigation treatments. Consequently, despite significant changes in individual sugar fractions, total soluble sugar concentration remained unchanged under DI.

3.5. Tree Yield, Fruit Size and Water Productivity

The DI treatment did not significantly affect fruit yield but reduced fruit diameter by about 8% compared with FI trees (Table 3).

Despite reductions in fruit size, DI received 619 m³ ha^-1 less irrigation than FI throughout the experiment. Consequently, water productivity increased substantially under DI, reaching values 67% higher than those observed in FI trees (Table 3).

4. Discussion

Irrigation treatments were effective in producing significant differences in tree water status as reflected in Ψ_S (Figure 3). The Ψ_S has been useful for discriminating different irrigation doses in fruit crops [36,37,38,39] and is well correlated to soil water content (e.g., [40]). Although FI trees received 100% ETc, Ψ_S progressively declined in preharvest from approximately -0.6 MPa at the beginning of the experiment to values close to -1.3 MPa, concomitantly with soil water content remaining mostly below field capacity (Figure 2). The Ψ_S values were lower than those reported for well-irrigated ‘Prime Giant’ cherry trees by Blanco et al. [19] whose control trees, irrigated at 110% ETc to safeguard non-limiting soil water conditions, maintained Ψ_S ranging between -0.5 to -0.75 MPa approximately in preharvest. Thus, our findings suggest that replenishment of 100% ETc may not have fully compensated for tree water demand under environmental and soil conditions of the present study, possibly reflecting uncertainties in Kc under local orchard conditions.

The Ψ_L varied in a similar way than Ψ_S, (Figure 3 and 4A). Temporal changes in Ψ_L explained a small proportion of the variability in A_SAT and gs (R²= 0.08 and 0.05, respectively), whereas A_SAT was strongly correlated to gs and gm (R²= 0.87 and 0.55, respectively), suggesting that reductions in carbon assimilation under deficit irrigation were primarily led by diffusional limitations than by effect of leaf water status. Nevertheless, the transient increase in Ci observed at 24 DABT in DI treatment suggests that non-diffusional limitations may also have contributed during specific period of fruit growth phase.

Following harvest and rewatering, Ψ_L increased in both treatments, although values remained lower in DI than FI trees (Figure 4A). The increases in Ψ_L in FI relative to the previous sampling date may be attributed to the cessation of fruit water consumption after harvest.

Despite the increase in leaf water status, A_SAT, gs and gm declined in both treatments, contrasting with preharvest period, during which these variables fluctuated within a relatively narrow range in DI and even increased in FI despite progressive decrease in Ψ_L. Taken together, these results suggest that sink demand seemed to play an important role in regulating cherry water status and leaf gas exchange as observed by Plavcová et al. [41] in apple trees. During fruit growth strong fruit sink activity may have stimulated carbon assimilation through sustained export of assimilates from source leaves, thereby promoting stomatal opening and photosynthetic activity [13,42,43]. In contrast, the abrupt reduction in sink demand following harvest coincided with decreases in A_SAT, gs and gm despite improved Ψ_L. Similar postharvest reductions in gas exchange activity have previously been reported in ‘Sylvia’ and ‘Kordia’ sweet cherry by Quentin et al. [44], further supporting a contribution of source-sink relationships to the physiological regulation.

Leaf soluble sugars were determined at 37 DABT, matching with the period of maximum treatment differences in Ψ_L and gas exchange variables. The DI treatment did not affect total soluble sugars but significantly altered carbon partitioning. Thus, sorbitol content increased by 27%, while sucrose decreased by 57%. In Prunus species, sorbitol is one of the main translocated carbohydrates and plays a substantial role in leaf osmotic adjustment under water deficit [18,44,45], improving osmoprotection and enhancing expression of drought-responsive genes [46,47]. Therefore, the observed shift in carbon partitioning may have contributed to maintaining stomata open and thus, leaf physiological activity relatively high under water-limited conditions. The absence of treatment effect on total soluble sugars in correspondence with significant differences in A_SAT indicates that water deficit did not promote downregulation of carbon assimilation by sugar feedback at 37 DABT.

Lignin fluorescence was significantly reduced in DI treatment at 37 DABT, which was in line with reductions in both Ψ_L and gm. This suggests that water-induced modifications in leaf cell wall traits potentially contributed to internal CO₂ diffusion limitations. Notably, LMA remained unaffected at the same time, indicating that the response involved specific cell wall components rather than major changes in overall leaf structure. Additional analysis must be conducted to verify this assumption. Although lignin content has been commonly reported to increase under water deficit, decreases have also been observed in leaves of water-stressed maize [48]. Our results also showed that lignin fluorescence increased from 37 to 51 DABT in both irrigation conditions, although the increase was greater in DI trees (+19%) than FI trees (+8%), resulting in similar postharvest values between treatments. This response occurred in parallel to changes in LMA (16% and 12%, in DI and FI, respectively), suggesting continued leaf structural development (e.g., [49]), regardless of water regime. Notably, increases in lignin fluorescence and LMA agreed with reductions in A_SAT, gs, and gm despite recovery of Ψ_L, indicating a shift from sustaining physiological activity during fruit growth towards greater investment in structural leaf components after harvest. Therefore, postharvest responses likely reflected not only recovery from water deficit but also leaf maturation and a change in source-sink balance following fruit removal.

Neither stomatal density nor chlorophyll index were affected by irrigation treatment, indicating that the observed changes in gas exchange variables were not associated with changes in stomatal abundance, chlorophyll degradation or accelerated leaf senescence. Stomatal density has been reported as less responsive to short-time changes in soil water availability and set mainly during leaf expansion [50,51].

Despite that physiological activity was reduced in DI leaves, fruit yield was not significantly affected, suggesting that overall carbon supply by photosynthesis was enough to sustain fruit production under deficit irrigation conditions. Similar responses have been reported in sweet cherry and other fruit crops exposed to mild to moderate water deficit, where reproductive growth is maintained at the expense of vegetative growth [19,52,53,54]. Under these conditions, fruits act as the dominant carbon sink when assimilate availability becomes limiting (e.g., [55]). In contrast, DI produced smaller cherries, although the reduction in fruit size was not large enough to affect yield, as indicated. Because deficit irrigation was imposed when fruit growth is mainly driven by cell expansion (BBCH 76), the reduction in fruit diameter likely reflected limitations in fruit expansion rather than carbon limitation, as reported for tomato fruit by Medyouni et al. [56]. Fruit size is a key quality trait determining crop value in sweet cherry, thus, reduction in fruit diameter under DI may have economic implications for growers despite the absence of significant effect on yield.

Maintaining fruit yield while reducing irrigation increased water productivity in DI trees by 67%. This indicates that the decrease in water supply was proportionally greater than the reduction in physiological activity, allowing fruit production to be sustained with considerably less water. However, the potential gain in water productivity should be weighed against possible reductions in fruit size, given its importance in crop value.

5. Conclusions

The physiological response of ‘Regina’ trees to deficit irrigation followed by postharvest rewatering was studied. Deficit irrigation reduced tree water status, gas exchange and fruit size, but did not affect fruit yield. Carbon assimilation responses were more closely associated with stomatal and mesophyll conductance than with leaf water potential, indicating that physiological regulation under water deficit was not driven exclusively by tree water status. Changes in leaf carbon partitioning and structural traits suggest that source-sink relationships contribute to this response. Although deficit irrigation substantially increased water productivity, the reduction in cherry size highlights a potential trade-off between water savings and crop value in sweet cherry orchards.