Introduction

Climate change has led to increase spring and summer temperatures, significantly altering the biological cycle of grapevines, particularly in Mediterranean regions (Jones et al. 2005). The period from budburst to leaf fall has lengthened, veraison is occurring earlier, and the berry ripening phase has shortened (Petrie and Sadras 2007; Prats-Llinàs et al. 2020). High temperatures during ripening decouple the accumulation of sugars and anthocyanins (Sadras and Moran 2012) and accelerate malic acid degradation. This results in a decline in grape and wine quality (Mira de Orduña 2010).

In response to the decline in grape quality caused by high temperatures, recent research has focused on management techniques to delay veraison and harvest, allowing these stages to occur under cooler conditions (Palliotti et al. 2014; Gutiérrez-Gamboa et al. 2021). Techniques explored include leaf removal (Buesa et al. 2019), late pruning (Gatti et al. 2016), crop forcing (also known as forcing regrowth) (Gu et al. 2012), double cropping (De Toda 2021; Poni et al. 2021) and even the use of chemical treatments (Palliotti et al. 2013). Among these techniques, the crop forcing technique—summer pruning which unlocks dormant buds and initiates a new biological cycle—has recently attracted much research interest as it can greatly delay the ripening period (Gutiérrez-Gamboa et al. 2021). Studies have demonstrated various delay outcomes: late pruning can postpone the harvest by 6 days (Gatti et al. 2016), leaf removal at veraison by 5–10 days (Buesa et al. 2019), and trimming after fruit set by one month (Zheng et al. 2017). However, the crop forcing technique stands out by potentially delaying the harvest by up to two months (Gu et al. 2012).

The main challenge with the crop forcing technique is the frequently observed reduction in yield (Martinez De Toda et al. 2019; Martínez-Moreno et al. 2019). This reduction could be partly attributed to the availability of carbohydrates in the permanent structure of the vine (roots and trunk) at the time of forcing pruning. Additionally, there is a strong competition for the vine’s resources from fast-growing organs, which is stimulated by high temperatures occurring after the forced pruning (Martínez-Moreno et al. 2019; Oliver-Manera et al. 2022).

The biomass partitioning to aboveground organs and carbohydrate reserves dynamics in forced vines have already been studied (Oliver-Manera et al. 2022; Lavado et al. 2023b). However, little is known about the dry mass allocated to the roots, which could introduce bias into estimations of total carbohydrate reserves (Holzapfel et al. 2010). Factors that may have an important influence on, and/or alter, seasonal root growth and root growth patterns include environmental conditions, canopy management, carbohydrate availability, vine water availability (Keller 2010), and competition with sink organs, and particularly bunches (Petrie et al. 2000). The crop forcing technique shifts the entire biological cycle of the vine to completely different environmental conditions from those associated with “traditional” management (Gu et al. 2012). In addition, seasonal net photosynthesis is lower in forced vines than in unforced vines (Oliver-Manera et al. 2022). However, a leaf removal experiment observed that although the reduction in total leaf area led to a reduction in root dry mass, root biomass increased when the crop was removed at veraison (Petrie et al. 2000b). The yield reduction in forced vines may therefore compensate for the seasonal decrease in photosynthesis and maintain root growth at a similar rate to that in unforced vines.

On the other hand, although root growth appears to be less sensitive to water stress than aboveground organs growth (Boyer and Silk 2004), water stress before forcing can affect vegetative growth and even reduce the number of forced bunches per vine (Oliver-Manera et al. 2023). In addition, water stress applied soon after forcing may reduce vegetative growth and yield, thus altering biomass partitioning (Lavado et al. 2023b). Comas et al. (2005) reported that heavily pruned vines tended to produce fewer roots than those which were minimally pruned, and that the interaction of exogenous factors, such as environmental conditions and soil moisture, affected root growth.

This study is based on two main hypotheses: (i) forced vines reduce yield, total biomass and new root biomass and alter biomass partitioning; and (ii) yield, total biomass, root production and biomass partitioning are sensitive to the timing of the application of water deficit to forced vines. The aim of this study was therefore to evaluate the combined effects of applying the crop forcing technique and water stress and to assess the seasonal sensitivity of forced vines to water stress, as measured in terms of yield, biomass and biomass partitioning, and root production.

Material and methods

Experimental site, design, and phenology

The trial was conducted in Lleida (Catalonia, Spain) (41.65°N, 0.52°E; 320 m above sea level) in 2020 and 2021, with potted (80 L usable volume) Tempranillo vines grafted onto Richter 110 rootstock which had been planted in 2018. To minimize potential damage from the soil overheating, a double pot was used. The inner pot contained the soil and the plant, while the outer pot was coloured white. The combination of the white colour and a layer of air between the two pots helped to prevent heat damage to the roots. The vines were oriented in a north–south direction and spaced 0.8 m apart, with 3 m between rows. However, root growth was restricted to the volume of the pot. The vines were trained using a vertically positioned canopy. The criterion for winter pruning was to leave about 6 spurs on each vine and two buds on each spur. The soil was a mixture of peat (30%) and local soil (22.7% clay, 49.3% silt, 28% sand) with a layer of gravel at the bottom of the pot to ensure drainage. The total available water per pot was estimated at about 12 L pot−1. The vines were drip irrigated, using two 2.3 L h−1 pressure compensating emitters per vine. The vines were irrigated in the morning, except during periods when the daily irrigation requirement exceeded 4 L vine−1. In such cases, they were irrigated twice a day (morning and noon) to ensure soil water availability and minimize drainage loss. The local climate is characterized by cold winters, with fog, and hot summers, with the possibility of heat waves and temperatures above 40 °C. The average annual precipitation and reference evapotranspiration are 340 mm and 1064 mm, respectively (data from 2007 to 2017). Rainfall events mainly occur in spring and autumn: mostly outside the main growing season. Weather data were collected from a weather station located 6.8 km from the plot, which forms part of the meteorological network of the Government of Catalonia (Servei Meteorològic de Catalunya). In the trial, two different irrigation strategies were applied to unforced (UF) vines and six more were applied to forced (F) vines, resulting in a total of eight treatments. The trial was replicated four times. An elemental plot included six vines for each treatment in each replicate (6 vines × 4 replications = 24 vines per treatment). The treatments were randomized in each replicate (see Fig. 1 in S1). The irrigation strategies were applied using leaf water potential (ΨL) thresholds based on Girona et al., (2006). The ΨL thresholds used for each phenological stage are summarized in Table 1. The fully irrigated treatments (UF-FI and F-FI in Table 1) were irrigated at 100% of the estimated crop evapotranspiration (ETc) and served as the reference for the other treatments. For these other treatments, the irrigation dose was adjusted to align leaf water potential with the predetermined thresholds for each phenological stage. Forced pruning was performed manually, 18 days after full bloom in 2020 and 20 days after full bloom in 2021. The pruning criterion for forced pruning was that 4–6 buds were left on each shoot. All leaves, bunches and dormant lateral buds were removed manually.

Fig. 1
figure 1

Daily rainfall, reference evapotranspiration (ETo) and average temperature (T), for 2020 (a) and 2021 (b). Vertical arrows indicate the most relevant phenological stages for the unforced (UF) and forced (F) treatments BB  budburst, F  forcing pruning, V veraison, H   harvest

Full size image
Table 1 Treatments and leaf water potential thresholds at noon (ΨL). The thresholds values (MPa) were obtained from the literature for each period between two phenological stages
Full size table

The phenological stages were monitored on a weekly basis according to a modified version of the system developed by Eichhorn and Lorenz (E–L) (Coombe 1995): budburst (E-L 4), fruit set (E-L 27), veraison (E-L 35) and harvest (E-L 38). A certain phenological stage was considered to have been reached once 50% of the population had passed that stage.

Irrigation, water status and leaf net photosynthesis

Each treatment was equipped with a water meter (CZ 3000, Contazara S.A., Zaragoza, Spain) to monitor the amount of water applied. The irrigation dose for fully irrigated treatments (L day−1 vine−1) was calculated on a weekly basis, following Eq. (1):

$$\text{Irrigation Dose}=(\text{ETc}-{\text{P}}_{\text{eff}})/0.9$$
(1)

where ETc (L day−1 vine−1) is the crop evapotranspiration for fully irrigated vines, calculated according to the FAO-56 Penman–Monteith formula: ETc = ETo Kc (Allen et al. 1998), with maximum crop coefficients (Kc from mid-season to harvest) of 0.25 for 2020 and 0.45 for 2021, derived from previous trials conducted in the same orchard (Girona et al. 2006; Marsal et al. 2008). Peff (L day−1 vine−1): the effective rainfall, was estimated as 15% (corresponding to the ground area covered by one pot) of the rainfall for a day with more than 2 mm of precipitation. It was otherwise set to zero. A factor of 0.9 was assumed for the efficiency of the irrigation system. To prevent excessive soil dryness after forced pruning in the F treatments, small irrigation events were applied until the forced buds opened. Drainage (D) was also collected from two vines per treatment, using a handmade container placed underneath the pot. D (L vine−1) was recorded by weighing, twice per week. The amount of water used by the vine (WU) in L vine−1 week−1 was then calculated for each treatment using Eq. (2):

$${\text{WU}} = {\text{Irrigation}} + {\text{P}}_{{{\text{eff}}}} {-} {\text{D}}$$
(2)

To monitor the vine water status, ΨL was measured once a week, from May to October, using the method described by Girona et al. (2006). On one vine per replicate, a mature leaf that had been exposed to the sun was placed in a plastic bag just before measurement. Measurements were taken at between 11:30 and 12:30 GMT, using a pressure chamber (Model 3005, Soil Moisture Equipment Co. Sta. Barbara, CA, USA).

To account for both the intensity and duration of water stress, ΨL was integrated as a water stress integral (IntΨ). IntΨ (Mpa day) was calculated according to the equation defined by Myers (1988), Eq. (3):

$${\text{Int}}\Psi = \left| {\mathop \sum \limits_{{{\text{i}} = 0}}^{{{\text{i}} = {\text{t}}}} (\overline{\Psi }_{{{\text{i}},{\text{i}} + 1}} - {\text{c}}){\text{n}}} \right|$$
(3)

where Ψi,i+1 is the average ΨL between two measurements, n is the number of days between the two measurement times, and c is the maximum ΨL measured during the study period (c =  − 0.5 MPa at budburst and -0.7 MPa when the forced buds opened in the F treatments).

Leaf net photosynthesis (An) was determined using a portable infrared gas analyser system (Li-6400, Li-Cor, Inc., Lincoln, NE, USA) equipped with a 6 cm2 chamber. Measurements were performed on one fully developed, sunlit leaf (PAR > 1200 µmol s−1 m−2) per vine and replicate (four measurements per treatment), with this leaf being at mid-canopy height. Measurements were taken on a weekly or bi-weekly basis, at between 11:00 and 12:30 GMT. Due to technical issues, An measurements were not conducted between day-of-year (DOY) 219 and DOY 240 in 2020 or during the post-harvest period in 2021. To account for the interaction between An, leaf age and canopy size, the parameter IntCanopyAn (Ψmol day s−1 vine−1) was calculated using Eq. (4):

$$\text{IntCanopyAn}=\sum_{\text{i}=0}^{\text{i}=\text{t}}{\text{An}}_{\text{i},\text{ t}+1} {\text{LA}}_{\text{i},\text{ t}+1}\text{ n}$$
(4)

where LAi,i+1 is the average leaf area per vine (m2 vine−1) taken from two consecutive measurements, Ani,i+1 is the average An for the same period, and n is the number of days during this period. LA and An at the time of budburst were considered to be 0.

Leaf area, yield, biomass and seasonal increase in root biomass

Two vines per replicate were selected to measure LA (eight vines per treatment). LA was determined once every three weeks, on two representative fruiting shoots per vine, according to the methodology established by Lopes and Pinto (Lopes and Pinto 2005). Following this methodology, total LA is estimated by multiplying the total leaf area of individual shoots by the number of shoots per vine. The leaf area of each individual shoot was calculated as the mean value of the leaves with the largest and smallest areas on the shoot, which was multiplied by the number of leaves per shoot. The individual leaf area was then estimated from central vein (CV) measurements, performing a linear regression between the two parameters (\(LA=17.22CV-55.48; {R}^{2}=0.89\)). To determine the regression, 150 leaves (75 from UF and 75 from F vines) were sampled in July 2020. For each sampled leaf, the CV was measured, using a tape measure, and the individual leaf area was measured with the Li3000 (Li-3000, Li-Cor, Inc., Lincoln, NE, USA). Only leaves whose CV was longer than 4.5 cm were considered.

The yield was weighed and the number of bunches per vine was counted in the field on the day of harvest. To determine fruit dry weight, a fruit sample was taken from each treatment and replicate. This was then weighed and dried in an oven at 67 °C until its weight remained constant.

On the forced pruning dates, all the biomass which had been removed (leaves, shoots and bunches) was dried in an oven at 67 °C until its weight remained constant. On DOY 274, one vine per treatment and replicate (4 vines per treatment) was bagged, using a net, in order to sample falling leaves. On DOY 305 for the UF treatments and on DOY 311 for the F treatments, all the biomass (including leaves, shoots, roots and trunk) was removed from the previously bagged vines. The roots were then carefully cleaned and classified as thick (root diameter > 5 mm), medium (root diameter < 5 mm and > 2 mm) or thin (root diameter < 2 mm). The leaf area was measured based on a sample of 25 leaves per vine, using the Li3000 meter (Li-3000, Li-Cor, Inc., Lincoln, NE, USA). The leaves were then oven-dried at 67 °C until their weight remained constant in order to determine the leaf specific dry weight (gDW m−2). The total biomass was also oven-dried at 67 °C until its weight remained constant, and total vine LA was then estimated by dividing the total leaf dry weight by the specific leaf dry weight of each treatment. If any treatment lost a significant amount of LA prior to wrapping, the dry weight of the lost leaves was estimated as the product of the lost LA (which was estimated as previously described) and the specific leaf dry weight.

In January 2020, when the vines were still dormant, two vines per replicate were selected (eight vines in total) to determine the initial root weight (Rt0). The roots of these eight vines were subsequently processed as previously described. The seasonal increase in root biomass (∆R), expressed in dry weight (gDW), was then determined using Eq. 5:

$$\Delta {R}_{t}= {R}_{t}-{R}_{t-1}$$
(5)

For the ∆R analysis, only roots with a diameter of < 5 mm was considered. The initial values for Rt0 are shown in Table 1 in S1.

Statistical analysis

To evaluate the effect of the combination of canopy management (CM) and irrigation strategy (I), a two-way ANOVA (P < 0.05) was conducted with CM as factor A and I as factor B. We only analysed treatments UF-FI, UF-DI, F-FI and F-DI. A univariate analysis of variance (ANOVA) was performed to show differences between the F treatments (P < 0.05). The only exception was the ∆R analysis when only F treatments were considered, in which a two-factorial analysis was performed that included the year (Y) as factor B. The normal distribution of experimental errors was assessed using the Shapiro–Wilk test and the homogeneity of the error variances was tested with Levene’s test (P < 0.05). The differences between the mean values were determined using Tukey’s test. All the statistical analyses were performed using JMP14 software (SAS Institute Inc., Cary, NC, 1989–2021).

To simplify the ΨL and An analysis, some forced treatments were regrouped and analysed as a single treatment:

Before forcing:

$$\left[ {{\text{F}} - {\text{PreF}};{\text{ F}} - {\text{PreF}} - {\text{V}}} \right] \, = {\text{ F}} - {\text{PreF}}$$

From forced crop veraison to harvest:

$$\left[ {{\text{F}} - {\text{V}};{\text{ F}} - {\text{PreF}} - {\text{V}}} \right] \, = {\text{ F}} - {\text{V}}$$

Before regrouping, an ANOVA test (P < 0.01) was performed between treatments in each group to ensure that there were no statistical differences in ΨL or An between members of the same group.

Results

Phenology, weather data and water used by vines

The F treatments delayed veraison by 48 days in 2020 and 44 days in 2021 (Table 2), shifting the grape ripening phase to a period with cooler conditions (Fig. 1a, b). The average temperature recorded during the grape ripening phase decreased by 7.5 °C (2020) and 4.3 °C (2021) for the F treatments compared to the UF treatments. However, during the period from budburst to fruit set mean temperature increased by 7.5 °C in 2020 and 8.3 °C in 2021 in F treatments compared to the same phenological period for UF treatments. Similarly and during the same phenological period, ETo also increased by 2 mm in 2020 and 1.5 mm in 2021 in F treatments compared to UF treatments. Before forced pruning, the year 2020 (Fig. 1a) was significantly wetter (324 mm) than during the same period in 2021 (151 mm) (Fig. 1b). In contrast, 2020 (Fig. 1a) was drier than 2021 (Fig. 1b) from the time of forced pruning to the F treatments harvest (63.4 mm vs. 128.7 mm).

Table 2 Forced pruning date (day of year) and date of the phenological stages for unforced (UF) and forced (F) treatments in 2020 and 2021
Full size table

Throughout the season, the amount of water used by the vines was 37% (2020) and 27% (2021) higher in the UF-FI than in the F-FI treatment. However, when we compared the UF-DI and F-DI treatments, we found that the differences were negligible in 2020, but that the amount of water used was 26% higher in the UF-DI than in the F-DI treatment in 2021 (Table 3). The amount of water used decreased by 52% for F-DI, 28% for F-PreF-V, 25% for F-V-PH and F-V and 18% for F-PreF in comparison with F-FI.

Table 3 Total water used (L vine−1) for each treatment in the 2020 and 2021 seasons and the sum for both (Total)
Full size table

Plant water status and leaf net photosynthesis

Analysing ΨL from budburst (DOY 93) to forced pruning (DOY 160) in 2020, we noticed that there were only differences with respect to the FI treatments (F-FI and UF-FI) on DOY 147 (UF-DI, F-DI and F-PreF) and DOY 154 (F-PreF) (Fig. 2a). While ΨL ranged from −0.7 MPa to −0.9 MPa in UF-FI throughout the season, ΨL ranged from −0.9 MPa to −1.5 MPa (from fruit set to harvest), with a minimum value of −2.0 MPa after harvest in UF-DI. From budburst (DOY 171) to veraison (DOY 247) in the F treatments, the average ΨL for F-FI (−0.85 MPa) was 19% higher than the average ΨL for F-DI (−1.01 MPa). From veraison to harvest (DOY 283), ΨL averaged −0.82 MPa for F-FI, −0.97 MPa for F-DI, and −1.00 MPa for F-V. Significant differences among the three strategies were observed only at DOY 254 and DOY 266, when ΨL was significantly lower in F-V compared to both F-FI and F-DI. After harvest, ΨL did not differ between F-V (−1.36 MPa) and F-V-PH (−1.25 MPa), but there were significant differences compared to F-FI (−0.76 MPa) and F-DI (−0.78 MPa).

Fig. 2
figure 2

Measured ΨL in 2020 (a) and 2021 (b). Statistical analysis was performed using ANOVA at P < 0.05 (ns = not significant; * = significant). The vertical bars for each symbol show the standard deviation. The vertical bars at the bottom of each plot relate to the Tukey’s test (P < 0.05). Note that some forced treatments are grouped and they are only shown when the irrigation treatment was different from the F-FI treatment. DOY = Day of the year according to the Julian calendar

Full size image

In 2021, no differences in ΨL were observed between UF-FI (mean ΨL of -0.85 MPa) and UF-DI (mean ΨL of −0.90 MPa) from budburst (DOY 84) to fruit set (DOY 147) (Fig. 2b). In the F treatments, from unforced budburst (DOY 84) to forced pruning (DOY 161), there were only significant differences between F-FI (mean ΨL of −0.81 MPa) and F-PreF (mean ΨL of −0.98 MPa and minimum ΨL of −1.23 MPa) on DOYs 140 and 147. When we compared the UF treatments from fruit set (DOY 147) to veraison (DOY 201), we observed that the ΨL was between −0.7 and (exceptionally) −1.0 MPa for UF-FI and between −0.8 and −1.15 MPa for UF-DI. From veraison to harvest (DOY 252), the ΨL for UF-FI ranged from −0.8 to −1.2 MPa and only on DOY 219 and DOY 246 were there any differences with respect to UF-DI, whose ΨL ranged from −1.15 to −1.44 MPa. However, in the post-harvest period, the UF-FI treatment recovered (mean ΨL of −0.71 MPa), while this was not the case for UF-DI (mean ΨL of −1.6 MPa). When comparing the F treatments, the mean ΨL for F-FI (−0.90 MPa) from forced budburst (DOY 168) to forced crop veraison (DOY 244) was 11% higher than the mean for F-DI (−1.01 MPa), with significant differences only being observed on DOY 223. From forced veraison to harvest (DOY 281), the mean value of ΨL was −0.73 MPa for F-FI, −1.06 MPa for F-DI and −0.9 MPa for F-V, with differences between F-DI and F-V only being observed on DOY 258 and those between F-FI and F-DI on DOY 270 and DOY 280. After harvest, only F-V-PH (mean ΨL of −1.65 MPa) differed from the other F treatments.

In general, higher IntΨL was observed for UF-DI than for UF-FI at all phenological stages in both experimental years (Table 4). Only the F-Pre and F-PreF-V treatments showed higher IntΨL than the other F treatments before the forced pruning. Water stress was evident in the F-DI treatment shortly after budburst in 2020, whereas in 2021 a higher IntΨL was only observed after fruit set. From veraison to harvest, the F-V, F-Pre-V, F-DI and F-V-PH treatments showed higher IntΨL than the other F treatments. After harvest, only the 2021 F-V-PH treatment showed higher IntΨL than the other F treatments. Throughout the season, UF-DI was the treatment with the highest IntΨL. In contrast to 2020, IntΨL was lower in F-FI than in UF-FI in 2021. Amongst the F treatments, only F-Pre-F (2020 and 2021), F-DI (2020 and 2021), F-V (2020), and F-V-PH (2021) showed higher IntΨL values than the F-FI treatment.

Table 4 Integral of water stress (IntΨL) in MPa day for each phenological stage (BB = budburst; FS = fruit set; V = Veraison; H = harvest; PH = post-harvest)
Full size table

In 2020, and before forced pruning, only the F-PreF treatments exhibited reduced An compared to the FI treatments on DOY 149 and DOY 157 (25% less, on average) (Fig. 3a). From DOY 177 to DOY 272, water stress applied to UF-DI greatly reduced An, with observed values of below 1 µmol m−2 s−1. Although the differences were only significant on DOY 219, a trend toward higher An in F-FI than in UF-FI was observed from DOY 212 to the end of the season. From DOY 199 to DOY 219, F-DI reduced An by 32%, compared to F-FI, and after DOY 219, An tended to be lower in F-DI, although not significantly so, except on DOY 279. After F treatments veraison, in the F-V treatments, An was reduced to an average of 62% of the An observed in F-FI. However, no differences were observed between the F-V-PH and F-V treatments after harvest.

Fig. 3
figure 3

Measured An in 2020 (a) and 2021 (b). Statistical analysis was performed using ANOVA at P < 0.05 (ns = not significant; * = significant). The vertical bars for each symbol show the standard deviation. The vertical bars at the bottom of each plot relate to the Tukey’s test (P < 0.05). Note that some forced treatments are grouped and they are only shown when the irrigation treatment was different from the F-FI treatment. DOY = Day of the year according to the Julain calendar

Full size image

In 2021, from UF fruit set (DOY 147) to UF harvest (DOY 252), An was reduced by 29% in UF-DI compared to UF-FI (Fig. 3b). After harvest, An in UF-DI decreased to a minimum value of 0.6 µmol m−2 s−1 on DOY 280. In the F treatments, An was reduced in both F-DI and F-PreF compared to F-FI before forced pruning. After forced pruning, and until F treatments veraison (DOY 244), An in F-DI was, on average, 35% lower than in the F-FI treatment. After F treatments veraison, An reduction was 50% for F-DI and 57% for F-V treatments, in comparison with F-FI. During grape ripening (from DOY 201 to DOY 252), the average An in UF-FI was 13.6 µmol m−2 s−1, which was 33% higher than the average An value registered during the same phenological period (from DOY 244 to DOY 281) for F-FI.

Leaf area, yield, biomass, and seasonal increase in root biomass

In 2020, no differences in LA were observed before DOY 160 (Fig. 4a). On DOY 199, the only significant differences observed were between the UF and F treatments. However, on DOY 219, LA values recovered in all the F treatments except for F-DI compared with UF-FI. In F-DI, LA was, on average, reduced to 50% of the F-FI LA. Notably, the UF-DI treatment reduced LA by 40% from DOY 240 to DOY 272; this was attributable to defoliation. In 2021, no significant differences were observed before forcing (DOY 161) (Fig. 4b). When comparing UF treatments, LA was reduced in UF-DI from DOY 176 to DOY 274 compared to UF-FI. All the F treatments, except F-DI, recovered their LA values to levels comparable to UF-FI from DOY 223 until the final measurement (DOY 301).

Fig. 4
figure 4

Evolution of the whole vine leaf area (LA) for 2020 (a) and 2021 (b). Vertical bars show the standard error (n = 4). * = significant differences according to the Tukey test (P < 0.05). DOY = Day of the year according to the Julain calendar

Full size image

Applying the crop forcing technique and reducing irrigation had a negative impact on yield and bunch weight (Table 5). Furthermore, in 2021, the F treatments reduced the number of bunches per vine while the irrigation strategy had no apparent effect. No consistent results were found when we compared the different F treatments. The LA/yield ratios were higher in the F than the UF treatments, even in the case of F-DI.

Table 5 Effect of forcing and irrigation strategy on yield and yield components for 2020 and 2021
Full size table

The total amount of biomass produced was reduced in the F treatments in 2021 but not in 2020 (Table 6). In general, leaf biomass increased in the F treatments, while no consistent results were observed for shoot biomass. The F treatments did not reduce total root biomass, but the irrigation deficit treatments (UF-DI and F-DI) did. Whereas trunk biomass was insensitive to forced pruning and irrigation management in 2020, in 2021, the treatments involving reduced irrigation produced a reduction in trunk biomass. After two experimental years, a good, negative correlation was observed between total biomass and IntΨL when only the F treatments were analysed, but this was not the case when all the treatments were included in the analysis (Fig. 5a). On the other hand, good positive and significant correlations between total biomass and IntCanopyAn (Fig. 5b) and the total amount of water used (Fig. 5c) were obtained for 2020, for 2021, and for the sum of both experimental years, regardless of the canopy management strategy. The F treatments increased the relative amount of biomass partitioned to leaves and shoots and reduced the amount of relative biomass partitioned to fruit in both years of the experiment. On the other hand, the relative biomass partitioned to trunk and roots was insensitive to forced pruning in 2020, although they increased in F treatments with respect to UF treatments in 2021 (Table 7). Reducing irrigation only had a negative effect on the relative biomass partitioned to shoots in 2020 but had a positive effect on that partitioned to the trunk in both 2020 and 2021. Other parameters were insensitive to irrigation management.

Table 6 Biomass (g Dry Weight) for all the treatments for both the 2020 and 2021 seasons
Full size table
Fig. 5
figure 5

Linear regression between the sum of the total biomass and the IntΨL (a), IntCanopyAn (b) and Water Used (c) for the years 2020 (squares), 2021 (rhombuses) and the sum of both years (circles). The black trend lines and equations indicate that all the treatments were considered. The grey trend lines and equations in plot A indicate that only the forced treatments were analysed. * = significant (P < 0.05); ** = significant (P < 0.01); *** = significant (P < 0.001)

Full size image
Table 7 Biomass (pre-forcing + after forcing for F treatments) partition (%) for all the treatments for the 2020 and 2021 seasons
Full size table

No differences in canopy management (UF or F) were observed for ∆R in either of the experimental years or in total ∆R (the sum for both years) (Table 8). In contrast, the DI irrigation strategy reduced thin ∆R (< 2 mm) in 2020.

Table 8 Root seasonal increment (∆R) in dry weight (gDW) comparison, depending on the canopy management (CM) used (UF or F) and the irrigation (I) strategy (FI or DI) for 2020, 2021 and the sum of the two years of the experiment (Total)
Full size table

When comparing the F treatments, only F-DI reduced thin ∆R (< 2 mm) compared to F-FI, while medium ∆R was not significantly affected by the irrigation treatment (Table 9). However, a significant (P < 0.05) and negative correlation was found between IntΨL and ∆R for roots less than 5 mm in diameter (R2 = 0.96) (Fig. 6a), for medium diameter roots (R2 = 0.68) (Fig. 6b) and for thin diameter roots (R2 = 0.95) (Fig. 6c). In addition, highly significant (P < 0.01) and positive correlation was found between IntCanopyAn and ∆R for root diameters of less than 5 mm (R2 = 0.95) (Fig. 6d), medium diameter roots (R2 = 0.89) (Fig. 6e) and thin diameter roots (R2 = 0.88) (Fig. 6f). Finally, ∆R for roots with diameters of less than 5 mm (R2 = 0.93) (Fig. 6d), medium diameter roots (R2 = 0.75) (Fig. 6e) and thin diameter roots (R2 = 0.89) (Fig. 6f) were significantly correlated (P < 0.05) with the total amount of water used by the vines.

Table 9 Root seasonal increment (∆R) in dry weight (gDW) comparison between forced treatments
Full size table
Fig. 6
figure 6

Linear correlation between ∆R with diameters of less than 5 mm (a, d and g), with diameters of between 5 and 2 mm (b, e and h), and with diameters of less than 2 mm (c, f, and i) with IntΨL (a, b and c), IntCanopyAn (d, e, and f) and the total volume of water used (g, h and i). All the parameters consider the sum of the two years of the experiment, only considering F treatments. * = significant (P < 0.05); ** = significant (P < 0.01); *** = significant (P < 0.001)

Full size image

Discussion

The crop forcing technique successfully delayed veraison and harvest dates until cooler and less evaporatively demanding conditions prevailed. When comparing UF-FI and F-FI treatments, we noted that vine water requirements were reduced in the F treatments. We would attribute this reduction in water requirement to the forced pruning itself, which reduced the vine leaf area to zero in early summer when ETo was high. Additionally, the maximum vine leaf area in the forced treatments was reached when ETo was already decreasing. When comparing UF-DI and F-DI treatments, we found that the irrigation requirement for F-DI vines was reduced by 18%, though this reduction was only significant in 2021. This was because in 2020 the reduction in irrigation was too severe in UF-DI, with ΨL falling below −2.0 MPa, which even led to defoliation. Our results differed from those reported by Lavado et al. (2023a), where field-grown forced vines did not exhibit any water savings. This discrepancy could probably be attributed to Lavado et al. (2023a) used field-grown vines in a deep soil environment and with a high water storage capacity, which acted as buffer to maintain stem water potential at the specified threshold for well-hydrated vines over a longer period. In contrast, the potted vines used in our study had limited soil water capacity, necessitating more frequent irrigation to maintain good vine water status.

In this study, except for the F-PreF treatments, it was difficult to detect any differences in water stress before forced pruning was applied. At that time, the vapour pressure deficit, on which leaf water potential depends (Olivo et al. 2009), was still low. However, when we analysed IntΨL, we observed clearer signs of water stress. This suggests that the parameter IntΨL offers a useful tool for analysing water stress over long periods (Myers 1988). It has also been shown to be effective in correlating water stress and growth parameters (Pérez-Álvarez et la., 2021) or yield (Oliver-Manera et al. 2023). On the other hand, detecting post-veraison water stress in forced vines have often been a source of difficulty in experiments using field-grown vines, as the phenological shift caused by the forcing technique shifts the ripening phase to a cooler conditions and with a low atmospheric evaporative demand (Martínez-Moreno et al. 2019; Oliver-Manera et al. 2023). The use of potted vines overcomes this difficulty due to the extremely limited soil water capacity compared to the higher soil water capacity of field-grown vines. In fact, potted vines are a useful tool for water stress studies that require a rapid physiological response (Girona et al. 2009). As expected, leaf net photosynthesis was reduced in response to water stress (Chaves et al. 2009), constraining biomass production at the whole-vine level. Although mature but young leaves are considered more efficient in photosynthesis than older leaves (Poni et al. 1994), we did not observe any increase in An in the forced treatments. Even in the 2021 season, An decreased in forced vines as vine regrowth decreased. This observation aligns with Oliver-Manera et al. (2022), who linked the reduction of An in forced vines to an extremely high LA/fruit ratio. It has been reported that vines with low sink demand adjust their physiological activity (source) (Palliotti et al. 2012).

Our results were consistent with those from other studies showing that fully irrigated and forced vines could restore functional leaf area due to reduced competition between vegetative and fruit growth (Oliver-Manera et al. 2022). Only the UF-DI and F-DI treatments reduced functional leaf area because water stress was applied during active vegetative growth, a period when plants are highly sensitive to water stress (Hsiao 1973). In the F-DI treatment, applying water stress shortly after forced pruning, when atmospheric evaporative demand is high, likely triggered an irreversible water stress response by reducing vegetative growth (Hsiao 1973; Chaves et al. 2009). A significant yield reduction of between 20% (Lavado et al. 2023a) and 80% (Martinez De Toda et al. 2019; Martínez-Moreno et al. 2019; Oliver-Manera et al. 2022) was observed in field-grown mature Tempranillo vines forced on similar dates to those in our study. In our study, the yield reduction observed in forced, fully-irrigated vines compared to unforced, fully-irrigated vines was 90% in 2021. We attribute this exceptional yield reduction to the use of young vines, whose yield tends to be more sensitive to carbon restrictions, such as water stress, than older vines (Bou Nader et al. 2019). The yield reduction, combined with the ability to maintain high LA levels, resulted in forced vines having a higher LA/fruit ratio than unforced vines, well above the 0.8 m2 kg⁻1 value for balanced vines (Kliewer and Dokoozlian 2005). However, it should be noted that the LA/fruit ratio may lose reliability when vine water status significantly impacts source-sink relationships (Etchebarne et al. 2010).

The biomass partitioning of forced vines had been previously studied by Oliver-Manera et al. (2022), and Lavado et al. (2023b), who also considered the interaction between forcing and irrigation strategies. Similar to our findings, these studies observed that forced vines exhibit a reduction in total biomass (Table 6) and an increased biomass allocation to vegetative organs (shoots and leaves) at the expense of fruit production (Table 7). However, to our knowledge, this is the first study to examine biomass partitioning to the roots. Water stress can diminish biomass production by reducing individual leaf photosynthesis and vegetative growth (Hsiao 1973), thereby reducing the canopy’s ability to intercept light, which is closely related to the whole canopy photosynthetic capacity (Poni et al. 2003). Our observations indicate that total biomass is significantly and positively correlated with the whole vine's photosynthetic capacity and the amount of water used per vine, irrespective of the canopy management strategy employed (see Fig. 5b, c). The reduced seasonal photosynthetic capacity of the forced vines, attributable to complete leaf removal and the subsequent period required to recover a fully functional leaf area, reduced biomass in the forced treatments. In the UF-DI treatment, biomass was reduced due to the limited amount of water used. In the F-DI treatment, biomass reduction was caused by both factors.

The fraction of roots that are less than 5 mm in diameter is considered to be representative of young roots (Bennett 2002) and therefore that most sensitive to the treatments applied during the two-year trial. We had expected forced pruning to negatively affect the increase in young root biomass for two main reasons: i) after forced pruning, canopy photosynthesis is reduced to zero and only recovers after about one month (Oliver-Manera et al. 2022), and root growth is highly sensitive to the reduced availability of photoassimilates (Buttrose 1966); and ii) the removal of shoot tips can cause root auxin depletion, leading to a temporary inhibition of root growth (Ferguson and Beveridge 2009). However, in our experiment, we did not observe any negative effects of the forced pruning on the production of new root biomass (Table 8). Root growth can be reduced by competition with fruit for vine resources, as a reduction in the number of bunches per vine can lead to an increase in the amount of biomass partitioned to the roots, even if the crop load is reduced after veraison (Edson et al. 1995). Therefore, the reduced fruit sink activity in forced vines may lead to an increased allocation of carbon to the roots, potentially offsetting the hypothesised negative effect of forced pruning on root biomass production. In contrast, prolonged water stress, as in the case of the DI treatments, reduced new root biomass. The increase in thin root biomass was more sensitive to water stress than thicker roots, as fine roots (< 0.5 mm) are constantly renewed throughout the season (Comas et al. 2005). However, after two years of the experiment, a decrease in the growth of medium-sized root (2–5 mm) was also observed in the DI treatments. The treatments involving only water stress applied before forced pruning did not reduce root growth, although they did reduce the level of leaf photosynthesis. The absence of a negative response in leaf area development to pre-forcing water stress suggests that the water stress applied before forcing was neither intense enough (−1.0 MPa) nor long enough to affect root growth. This is because the growth of aboveground organs is generally more sensitive to water stress than root growth (Boyer and Silk 2004). Additionally, we did not observe any reduction in root growth in these forced treatments with either post-veraison or post-harvest water reduction (Table 9). Oliver-Manera et al. (2022) observed that at a high LA/fruit ratio, fully irrigated forced vines were able to replenish the carbohydrate reserves in their roots and trunks to a very high degree. We therefore hypothesise that, in F treatments, roots continued to grow despite water restrictions during the grape ripening because they compete less with fruit and because the reduced photosynthetic capacity is compensated by the supply of carbon from the permanent vine structure.

When we compared only the forced (F) treatments, we observed a reduction in the increase of root biomass exclusively in the F-DI treatment (Table 9). However, after the two-year experiment, we found a negative and significant correlation between root increment and IntΨL, as well as positive and significant correlations between root increment and both the IntCanopyAn parameter and the water used by vines (Fig. 6). This indicates that the increase in root biomass is sensitive to water stress over the long term, regardless of the timing of its application. The dynamics of vine carbohydrate reserves can be misinterpreted if total carbohydrate reserves, including total root biomass, are not considered (Holzapfel et al. 2010). Therefore, our findings present a novel approach and offer valuable insights for interpreting studies on carbohydrate reserves in forced vines or those subjected to similar canopy management techniques.

Conclusions

Under these experimental conditions, the forced vines reduced the total biomass and drastically reduced yield compared to unforced vines, while the leaf area and root biomass remained comparable and the leaf and shoot biomass even increased. It was thus confirmed that forced vines promote biomass partitioning to vegetative organs to the detriment of fruit production. This results in an extremely high source-to-sink ratio in the forced vines, which may even cause a reduction in leaf photosynthesis as a physiological response to the low sink activity. Reduced competition with fruit could explain the lack of a negative response to root growth because of forcing. Severe and continuous water stress negatively affected biomass and new root production but had minimal effect on biomass partitioning. Total biomass was found to depend on the whole canopy's photosynthetic capacity and the amount of water used by the vines, regardless of the canopy management strategy employed. Moreover, the crop forcing technique proved effective in saving water compared to unforced vines, contributing new insights to the field.

Regarding forced irrigation treatments, water stress applied shortly after forced pruning had an immediate, negative, and irreversible impact on vegetative growth, biomass, and new root production. Applying water stress shortly after forced pruning should therefore be avoided as an irrigation strategy. Other irrigation strategies had no negative effects on yield or biomass production and reduced the amount of water used compared to fully irrigated vines. Furthermore, post-harvest root growth was minimal, as reducing irrigation from veraison to the end of season (F-V-PH) did not reduce root biomass. However, in the long term, biomass and new root production were negatively correlated with water stress and positively correlated with both the photosynthetic capacity of vines and the amount of water they used, regardless of the irrigation strategy applied. Therefore, in the long term, even mild water restrictions can have a negative impact on whole vine productivity and root growth. However, further research using more intense seasonal water stress thresholds should be conducted to identify the short-term effects of water stress and to establish reliable water stress thresholds.

The use of potted vines and the associated limited soil water capacity make it necessary to conduct further research into the effects of applying the forcing technique and to investigate the impact of water stress on biomass partition and root growth using field-grown vines.