The experimental trial was carried out during the 2022 growing season over a 4-week period from 23 August to 19 September (DOE=days of the experiment) in a 9-year-old mature commercial kiwifruit (Actinidia chinensis var. chinensis, cv. Zesy002) orchard in Metaponto, southern Italy (latitude 40°24'N, longitude 16°46'E). The yellow-fleshed cultivar was grafted onto D1 rootstock (A. chinensis var. deliciosa), cultivated at 5 m x 2 m spacing and trained as a pergola system. The soil profile consisted of a uniform sandy clay loam texture (58% sand, 14% silt and 28% clay).

Irrigation was provided by micro-sprinklers, using a single emitter per vine spaced 1 m from the trunk, at a flow rate of 40 L h^-1 and with a wetted radius of 0.9 m. Soil volume wetted by irrigation, expressed as total volume per hectare (m³ ha^-1), was assumed as a cylinder and calculated considering the sprinkler wetting radius, the depth of the wetting pattern (0.6 m) and the number of emitters per hectare. All vines were grown under optimal water availability conditions during the season until the beginning of the experimental trial. Daily vine water requirement was estimated through the compilation of the water balance, taking into account environmental data and crop coefficient (K_c) reported in the FAO Irrigation and Drainage paper Nr. 66 (Steduto et al., 2012). The scheduled irrigation volume was adjusted throughout the season through the soil water balance, operating feedback adjustment mechanisms based on the monitoring of the water content in the soil volume wetted by irrigation by soil moisture probes, in order to maintain the soil moisture level close to or slightly below the field capacity (FC) (Vera et al., 2019; Mininni et al., 2022). Thus, irrigation drainage in this study was considered negligible. The experimental irrigation-managed plot, consisting of 63 vines, was equipped with digital flow meters and an automatic irrigation controller for remote control of water supply ("Irrifarm" system, Pan. Agri s.r.l., Basilicata, South Italy). Daily irrigation volumes ranged from 30 to 60 m³ ha^-1 during the season.

From 23 August on, 12 vines within the experimental plot and distributed along a row, underwent restricted irrigation (WS) with irrigation volumes reduced by 30% and ceased, adjusted taking into account rainfall contribution, to allow the soil to achieve a progressive available water content (AW) depletion. The soil about the control vines (C) received optimal irrigation, ensuring a soil water content level close to or slightly below the FC, so that it remained well-watered throughout the experiment. Half of the water-stressed vines were re-irrigated after 15 days (R_I) following the irrigation strategy kept for the control vines, restoring the available water content in the soil profile, while the remaining half of the vines continued to be subjected to restricted irrigation for other 5 days (WS), subsequently entering the recovery phase (R_II).

A meteorological station was installed nearby the experimental orchard block and mounted above the kiwifruit canopy, outside the hail net protection. The sensors recorded hourly averages of wind speed, air temperature, relative humidity, evapotranspiration and daily rainfall. Hourly air vapor pressure deficit (VPD) was computed using air temperature and relative humidity.

Frequency Domain Reflectometry (FDR) 90 cm multi-profile soil moisture probes (Drill & Drop, Sentek Sensor Technologies Stepney, Australia) were used to continuously monitor the volumetric soil water content in 10 cm increments throughout the soil vertical profile, allowing the remote detection of instantaneous oscillations and trends in each soil layer. A soil layer of 0-60 cm was considered useful for irrigation scheduling in kiwifruit (Xiloyannis et al., 2023). Sensors were installed on the row at a distance of 0.5 m from the vine stem and approximately in the center of the volume of wet soil. Measured values of soil moisture were stored in 15-min-intervals using a data logger (Model CR1000, Campbell Scientific, USA), hourly means were considered for analyses.

The soil physico-hydrological properties were evaluated through the soil-water retention curve obtained with the Richards' pressure plates apparatus (Richards, 1948) to define the soil-water relation and, in particular, the quantity of water held at different values ​​of soil matric potential, and therefore the FC, the wilting point (WP) and the AW in the whole profile (0-60 cm) ( Table 1 ).

The available water (AW, m³ ha^-1) that can be contained in each layer, and thus in the whole 0-60 cm soil profile considered in the present study, was calculated taking into account soil water retention characteristics derived from the soil-water retention curve (FC and WP values, Table 1 ) and soil volume of each 10 cm layer wetted by irrigation calculated on a per hectare basis, which is influenced by the irrigation system and the wetted radius. The volumetric soil water content values recorded by the soil moisture probes in 10 cm increments were used to estimate the actual available water (AWact, m³ ha^-1) contained in each soil layer and in the whole 0-60 cm profile on a per hectare basis. The depletion of AW (AWdepletion, %) was calculated using the following equation:

where AW is the potential available water contained in each soil layer (10 cm increments) or in the whole profile (0-60 cm) and AWact is the actual value of the available water calculated for each day of the experiment.

FDR technologies, which allowed for routine measurements and detection of temporal changes in the soil water content over time, were used to deduce the spatial pattern of water uptake by the roots in the soil profile. In Mediterranean environments, the volume of soil not wetted by irrigation tends to dry out during the warmer months, so it is not affected by root water uptake (Green and Clothier, 1995). Drainage of water in the deeper layer was nullified by the precise irrigation management, which was based on the continuous monitoring of soil water content and the distribution of daily irrigation volumes in two/three events per day, thus it was considered negligible in the present study. Soil evaporation plus understorey transpiration were minimized due to localized micro-irrigation, high vine cover and absence of cover crops, making irrigation requirements almost dominated by vine transpiration (Goodwin et al., 2004; Goodwin et al., 2015). Therefore, they were assumed to be not relevant for irrigation management purposes in the present study.

Diurnal courses of physiological stress indicators were measured to investigate plant-water relations by analyzing their trends influenced by changing soil water status conditions, both for stressed and well-watered vines, and considering the peculiar physiological and anatomical traits of kiwifruit vines. Physiological measurements were performed during the increasing soil water deficit conditions (WS vines) and the subsequent recovery phases (R_I and R_II vines) compared to C vines.

Plant water status and stress level were monitored through the stem water potential (ψ_stem, MPa) on six fully developed, expanded and not exposed to sunlight leaves chosen from the middle part of the canopy of three plants per treatment. The selected leaves were wrapped in a tin foil for about half an hour and then excised for immediate measurement with a Scholander pressure chamber (Model 1000, PMS Instrument Company, Corvallis, OR, USA) pressurized with nitrogen according to the procedure recommended by Begg and Turner (1970). The ψ_stem daily trends were monitored on seven days of the experiment at two-hourly intervals throughout the day, from 5:30 (pre-dawn values) to 16:30 (afternoon values) hours (European daylight savings time, DST), in order to understand the synchronization of ψ_stem diurnal pattern with soil volumetric water content and VPD environmental conditions. Stomatal conductance (gsw, mol m^-2 s^-1), transpiration (E, mmol m^-2 s^-1) and leaf temperatures (T_leaf, °C) were similarly monitored during the same days with a portable handheld LI-600 porometer system integrated with a fluorometer (LI-COR Biosciences, Lincoln, USA). Values obtained were averages of nine measurements taken on mature and sun-exposed leaves, near the top of the canopy, with no visual symptoms of biotic or abiotic stress, from three vines per treatment. During the other days (five) of trial observation, measurements of ψ_stem, gsw, E and T_leaf were conducted at midday (between 12:30 and 14:30 hours) on sunny days. Leaf to air temperature differences (ΔT (°C) = T_leaf - T_air) were calculated. Measurements of ψ_stem, gsw and E were carried out at 1, 2, 5, 7, 12 and 2, 7 days after re-watering for the first and second recovery phases (R_I and R_II) compared to C. Measurements of photosynthesis (A, µmol m^-2 s^-1) were carried out on six fully expanded, developed and sun-exposed leaves from three vines during the drought (diurnal trends for four days) and re-watering (midday values for two days) cycles with an open gas exchange system unit (ADC LCA-4, Analytical Development Co., Hoddesdon, UK).

All statistical analyses were performed using RStudio statistical software (4.2.2 version, Posit Software, PBC, Boston, MA) and results were plotted with SigmaPlot 15.0 (Systat Software Inc., San Jose, CA, USA). Data were reported as mean value and standard error of the mean (± SE). The data were subjected to an analysis of variance (one-way ANOVA) to examine the differences between irrigation treatments at each sampling time (p value<0.05 was considered significant) after testing for normality distribution of residuals (Shapiro-Wilk test) and homogeneity of variances (Levene's test). Dunnett's test was used for mean comparison of WS and R_I against the control. To investigate the relationships between physiological indicators (ψ_stem, gsw, A) and soil water status (SWC), polynomial regression models were fitted to averaged data from each irrigation treatment.

The experimental trial coincided with a period characterized by a quite high evaporative demand. The weather was generally sunny and dry with the exception of a few days when rainfall occurred. 19 days out of the 28 days had maximum air temperatures that exceeded 30°C ( Figure 1 ), while the minimum air temperatures varied between 9 and 24°C during the experimental trial. VPD was quite variable throughout the experimental days, reaching maximum hourly values equal to 3.65 kPa at DOE 3. Daily ET₀ ranged between 3.5 and 6.0 mm. Daily net solar radiation reached a maximum value of 24.2 MJ m^-2 (DOE 3) and ranged between 16.2 and 24.2 MJ m^-2 during the experimental trial. Irradiance, air temperature and humidity were typical of Mediterranean late summer conditions.

Soil water content gradually decreased in the field area subjected to water restriction and trends of AW depletion in different soil layers and in the whole soil profile are shown in Figure 2A . In particular, AW was reduced by approx. 61%, 19% and 14% in the 0-20, 20-40 and 40-60cm soil layers, respectively, at the end of the first phase of restricted irrigation (DOE 15) before the stress was relieved by re-watering vines entering the R_I recovery phase. An AW depletion of approx. 69%, 21% and 19% computed in the 0-20, 20-40 and 40-60 cm soil layers, respectively, was achieved at the maximum level of stress (DOE 20) before the stress was relieved by re-watering vines entering the R_II recovery phase ( Figure 2A ). In C vines soil water content was on average maintained close to or slightly below the FC. As water stress increased and in response to progressive soil water depletion, ψ_stem of WS vines became more negative, reaching pre-dawn values (ψ_predawn) of -0.5 and -1.0 MPa, and midday values (ψ_midday) of -1.4 and -1.8 MPa, respectively at the end of the first and the second phase of restricted irrigation, when about 31% and 36.5% of AW in the whole soil profile (0-60 cm) was depleted ( Figures 2A, B ). C vines maintained ψ_stem values indicative of an optimal plant water status, ranging from -0.1/-0.2 MPa at pre-dawn to -0.4/-0.7 MPa at midday throughout the days of the experiment of restricted irrigation ( Figure 2B ).

Under non-limiting soil water conditions, ψ_stem and ΔT observations correlated with midday VPD more tightly than E, whereas the relation between gsw and VPD was not significant ( Figure 3 ). Relationships between physiological indicators and air VPD indicated that, under non-limiting soil water conditions, vines did not implement any stomatal regulation with the daily increase in environmental evaporative demand (i.e. VPD) preferring to transpire, losing water derived from root uptake, and slightly lowered the plant water potentials.

ψ_stem of C vines showed daily trends reflecting daily VPD conditions, reaching minimum values during the hottest hours (e.g., around midday), but remaining in an optimal range when soil water availability was not limiting ( Figure 4 ), allowing the plant's physiological activities not to be compromised. Vines under limited soil water availability conditions showed a fast drop in ψ_stem at midday and a gradual incapacity to recover plant water status in the afternoon hours, as highlighted on DOE 15 ( Figure 4A ), affecting the ability to recover during the nighttime hours, as shown by the progressive lowering of the ψ_predawn values ( Figures 2B , 4 ). Stomatal conductance remained relatively constant between 0.16 and 0.20 mol m^-2 s^-1, throughout the day and between different days in C vines ( Figure 4 ), supporting the results from Figure 3 . Stomatal conductance of WS vines was higher early in the morning, especially during the first phase of restricted irrigation ( Figures 4A, B ) when ψ_stem was not yet the seriously limiting factor, then gradually decreased and remained low throughout the day, reduced by approx. 55-75% compared with C vines during the days of experiment of significant stress ( Figures 4C, D ), reaching values relatively constant between 0.05 and 0.08 mol m^-2 s^-1, probably due to the implementation of an adaptive strategy involving stomatal control as water is depleted in the root zone. However, stomatal closure, occurring in the WS vines after the early morning hours, did not prevent ψ_stem from a further drop in the afternoon hours ( Figures 4A, B ). Conversely to gsw, E measured in kiwifruit vines under non-limiting soil water conditions was affected by VPD, increasing when environmental conditions became more water-demanding ( Figure 3B ). Transpiration rate in C vines was low in the early morning due to lower light intensity and air VPD. With increasing light intensity and temperature, E rose reaching higher values at midday and then gradually decreased ( Figure 4 ). In WS vines, E generally remained at very low values throughout the day, reflecting the same daily trend as in C vines. In particular, E of WS vines measured at early morning were similar to those of C vines in the first phase of restricted irrigation, also aided by low environmental demand, while they significantly differed during hours of higher evaporative demand ( Figures 4A, B ). Instead, E of WS vines differed since early morning from that of C vines in the second phase of restricted irrigation ( Figures 4C, D ), achieving a reduction of approx. 50-65% compared with control, almost constant over the days.

Leaf temperature of sun exposed leaves of kiwifruit vines under non-limiting soil water conditions, monitored during the midday hours (12:30-14:30 hours), was higher than air temperature in the range of approx. 2-4.5°C, depending on daily VPD conditions. In particular, leaf to air temperature differences (ΔT) decreased with increasing VPD and E, due to the potential cooling effect of transpiration ( Figure 3D ). Otherwise, in WS vines T_leaf were nearly 4°C higher than in C vines on some days of experiment (data not shown) and ΔT, on average equal to approx. 6°C, increased due to the reduced transpiration rate and consequent loss of thermoregulatory capacity under water stress conditions ( Figure 5B ). Reference values of ψ_midday, allowing maximum stomatal activity and opening, ranged from -0.4 down to -0.7 MPa, under non-limiting soil water conditions ( Figure 5A ). Results showed that under a threshold value of ψ_midday, generally corresponding to the minimum reached during the day, of approx. -0.8 MPa, gsw began to decrease significantly while ΔT increased ( Figure 5 ), indicating a change in plant behavior when subjected to water stress conditions.

Vines were then subjected to a recovery phase (R_I and R_II) from a different level of water stress, indicated by ψ_midday and ψ_pre-dawn values of approx. -1.4 and -0.5 MPa reached at the end of the first phase of restricted irrigation, and -1.8 and -1.0 MPa reached at the end of the second phase of restricted irrigation, respectively. Recovery dynamics of physiological indicators are shown in Table 2 . Upon re-watering, recovery of ψ_stem was rapid both in the first and in the second recovery (R_I and R_II), restoring values equal to those of C vines in the immediate 24 hours, regardless of the water stress level achieved. However, gsw of WS vines showed a delay in recovery, reaching 86% and 51% of the value of C vines one week after re-watering from the first and the second phase of restricted irrigation, respectively, showing a different recovery trend depending on the stress level achieved. Similarly, E of WS vines recovered slowly, but not as slowly as gsw, showing no differences with C vines and reaching 76% of the value of C vines 5 days and one week after re-watering from the first and the second phase of restricted irrigation, respectively ( Table 2 ).

The response of physiological indicators to SWC during the days of experiment is shown in Figure 6 . In particular, ψ_midday, gsw and photosynthesis (A) showed a positive correlation with SWC recorded by soil moisture probes in the 0-40 cm soil layer, indicating that a progressive decrease in SWC negatively affected plant behavior ( Figure 6 ). Relationships were carried out considering the 0-40 cm soil layer that was mainly affected by root water uptake dynamics as shown in Figures 2A , 8 . SWC was maintained approximately in the range between 40% and 50% for C vines in the 0-40 cm soil layer. Stomata were wide open, allowing maximum transpiration rates when SWC was maintained at optimal levels ( Figure 6B ). ψ_midday, gsw and A declined by 50% once SWC dropped to about 27%, 26% and 24%, respectively. At much lower SWC, the stomata were almost completely closed, which severely affected vine transpiration. Determination coefficients (R²) of the relationships between physiological indicators and SWC were higher for gsw, followed by ψ_midday and A, equal to 0.84, 0.73 and 0.64, respectively. The intersection of the mean value of SWC calculated for the control vines with the function trajectory resulted in a reference value of -0.67 MPa, 0.17 mol m^-2 s^-1 and 16.0 µmol m^-2 s^-1 for ψ_midday, gsw and A, respectively ( Figure 6 ).

The analysis of plant measurements, carried out in the WS vines during the days of experiment of restricted irrigation, in relation to the progressive AW depletion reached in different soil layers (0-20, 20-40 and 40-60 cm) is shown in Figure 7 . ψ_midday and gsw showed to be greatly affected by the dynamic of AW depletion according to soil depth, in particular, changes in plant water status and stomatal behavior are associated with higher reductions of the available water contained in the 0-20 cm soil layer, followed by the 20-40 cm and 40-60 cm soil layers ( Figures 7A, B ).

Daily SWC and AW depletion trends in C and WS vines, respectively, are shown in Figure 8 . The changing pattern of SWC around the C vines reflected the combined effects of multiple daily irrigation interventions and the root water uptake strategy used by the vine. Greater variations in SWC were observed at 15, 25 and 35 cm soil depths ( Figure 8A ), while at deeper depths an almost flat trend in SWC was observed (data not shown). Three daily irrigation interventions were carried out to allow the vine to meet more efficiently the transpiration demand, which was higher during the midday hours, as shown by the increased root water uptake in the first 15 cm of soil on both days, but particularly on DOE 13, probably due to the increased environmental evaporative demand ( Figures 8A , 1 ).

The progressive decrease in SWC around the WS vines reflected a differential reduction in AW according to different soil depths and time of the day. Water uptakes were mostly concentrated in the central hours of the day (i.e., from 6:00 to 18:00), as shown by the decrease in AW contents when the vine accessed water to meet transpiration demands ( Figure 8B ). Preferential root water uptakes occurred in the superficial soil layers, as significant reductions of approx. 3.6%, 2.4% and 1.5% were achieved at 15, 25 and 35 cm soil depths during DOE 13, indicating a decrease in AW depletion with increasing soil depth ( Figure 8B ). Water uptake in the deeper layers increased gradually and slowly as the AW in the superficial soil layers was depleted.

The spatio-temporal changing dynamics of AW and the contribution of each layer to the progressive lowering of AW in the whole soil profile during the days of the experiment is shown in Figure 9 . The AW was preferentially depleted in the first soil layers (up to 30 cm), which showed the largest reductions, contributing to a reduction in AW contained in the whole soil profile ranging from 10% to 27% during the days of the experiment. The progressive reduction of AW in the first soil layers (up to 30 cm) was slightly more limited from DOE 13 when the vines increased their water uptake in the deeper layers (from 30 to 60 cm), contributing to an AW reduction in the whole soil profile in the range of 0% to 9% during the days of the experiment ( Figure 9 ). In particular, the deeper layers showed no or small reductions in AW at DOE 2 and DOE 3, respectively. From DOE 13, soil water conditions also began to be limiting in the deeper layers (from 30 to 60 cm), as the vines gradually accessed AW after the superficial soil layers had exhausted the more freely available water.

Measurements of plant water status and stomatal conductance combined with soil observations and root water uptake dynamics during DOE 2 and DOE 3 are shown in Figure 10 . Based on the daily dynamics of ψ_stem and gsw, DOE 2 was identified as the day before the onset of the water stress condition, since these parameters were almost the same in both C and WS vines ( Figures 10A, B ). During DOE 2, AW started to decrease in the first soil layers (up to 30 cm), with approximately 31%, 22% and 10% of AW being depleted in the 0-10, 10-20 and 20-30 cm soil layers respectively, while deeper layers showed insignificant decreases in AW ( Figure 10C ). Similarly, DOE 3 was identified as the day of water stress onset, as both gsw and ψ_stem of WS vines were different from those of C vines ( Figures 10D, E ). In particular, ψ_stem of the WS vines was equal to that of the C vines during the early morning hours, while it decreased significantly at midday, reaching a value of approx. -1.0 MPa, due to the combination of a higher evaporative demand and a rapid decrease in SWC, suggesting that the reduced water availability exacerbated the midday depression phenomenon ( Figures 10D, E ). During DOE 3, AW was further reduced by approx. 16%, 11% and 5% compared to DOE 2 in the 0-10, 10-20 and 20-30 cm soil layers, respectively ( Figure 10F ).

The irrigation strategy adopted for the well-watered vines consisted in splitting the daily irrigation volume into several interventions, maintaining the SWC steadily close to the FC ( Figure 8A ). The timely synchronization of the vine's water needs during the day, as influenced by environmental demands, with the irrigation supply, made it possible to guarantee an optimal soil water status at any time of the day, ensuring maximum root water uptake, avoiding the restriction of leaf gas exchange and the establishment of stomatal control and promoting transpiration processes ( Figures 3B, C ). This result is in agreement with Torres-Ruiz et al. (2016), who found that providing water at times when environmental conditions are more stressful, i.e. during the hottest hours of the day, allows kiwifruit vines to benefit from an improved water status and higher leaf gas exchanges at this time of the day. Soil water status and plant transpiration rate, which is influenced by environmental evaporative demand, are two main factors influencing plant water status at a given time of the day (Tardieu and Simonneau, 1998) ( Figures 3 , 4 , 7 ). Stem water potential measured in the present study was used as the primary indicator of plant water status because it has been shown to be less variable than leaf water potential and more efficient in detecting small differences between irrigation treatments (McCutchan and Shackel, 1992). A gradient exists between leaves and stem that is proportional to the transpiration rate, resulting in lower leaf water potentials than stem water potentials, as previously reported in kiwifruit by Morandi et al. (2010). Therefore, when interpreting the data, it should be taken into account whether the results refer to leaf or stem water potential values. Maximum values of ψ_stem were observed at pre-dawn, when an equilibrium between plant and soil water potentials is achieved in the absence of water flux (Tardieu and Simonneau, 1998). ψ_stem reached lower values as the environment became more water-demanding ( Figure 4 ), but well-irrigated vines maintained their ψ_midday in the range between -0.4 and -0.7 MPa, below which vines began to experience a stress condition that significantly altered normal plant functioning. Previous studies reported daily trends in leaf water potential for well-irrigated 'Hayward' kiwifruit ranging from -0.1/-0.2 MPa at pre-dawn to -0.9/-1.2 MPa at midday (Gucci et al., 1997), -0.03/-0.08 MPa at pre-dawn to -0.4/-0.8 MPa at midday (Judd et al., 1989) and -0.6/-1.0 MPa at midday (Montanaro et al., 2009; Calderón-Orellana et al., 2021), with differences depending on environmental evaporative demand (i. e. VPD conditions), stomatal behavior and root system development.

The response of yellow-fleshed kiwifruit to progressive SWC lowering was investigated to evaluate the effects of SWC on plant water status and physiological parameters and identify thresholds below which plant performance began to be compromised, up to severe effects.

Monitoring soil water status, which is related to plant water status, is increasingly becoming a common approach to schedule and control irrigation in order to ensure optimal plant performances (Nolz et al., 2016). Fluctuations in soil AW vary throughout the soil profile, depending on the structure and development of the root system, the irrigation system wetting a specific soil volume, and the environmental conditions (such as reference evapotranspiration and VPD). Therefore, it is important to define in-situ soil moisture thresholds based on multi-layer sensor data collected in the field to manage irrigation from a practical point of view. The method proposed in this paper defines thresholds for AW by measuring plant responses to progressive lowering of SWC and analyzing their relationships, in order to identify the 'stress point' that serves to early detect a water stress condition and implement more sustainable and precise irrigation strategies. The present study found that plant water status and other physiological indicators of kiwifruit vines were sensitive to a progressive decrease in SWC ( Figure 6 ). This result is consistent with the findings of previous research carried out on other species (Garnier and Berger, 1987; Naor and Cohen, 2003; Al-Yahyai, 2012; Jamshidi et al., 2020), focusing on the response of water stress indicators to soil drying and whether these parameters are correlated with each other. The resulting correlation confirmed that, in addition to environmental variables, other factors such as soil moisture could have a great influence on plant water status. In the lower range of SWC, plant water status and physiological parameters were controlled by the root water uptake mechanism and water availability. The current work aimed to identify soil water content thresholds and define the amount of AW that can be used by the plant without compromising the physiological aspects, delimiting the range where no stress should occur. Results provide new information to define the depletion fraction of AW from the root zone before moisture stress occurs in yellow-fleshed kiwifruit, which was previously considered to be 0.35 for kiwifruit (cv 'Hayward') with a crop evapotranspiration (ET_c) of approx. 5 mm day^-1 (Allen et al., 1998). Vines, which absorbed water differentially in each soil layer, experienced a decrease in the AW amount of approx. 35 m³ ha^-1 from the soil profile achieved at DOE 2, identified as the day until which the plant's water status and physiology were maintained at optimal levels, which corresponds to a depletion of about 11% of the AW contained in the whole soil profile (0-60 cm) ( Figure 9 ). Instead, AW was overall reduced by approx. 17.8% at DOE 3, identified as the 'stress onset day' ( Figures 10D, E ), which implied a decrease of about 57 m³ ha^-1 of AW from the whole soil profile ( Figure 9 ). The depletion factor changes with the environmental evaporative demand, in particular at high ET_c it is 10-25% lower than the ordinary value, indicating that a lower AW fraction can be depleted before the onset of the stress (Allen et al., 1998). Values found in the present study highlighted that under the same ET_c conditions, the depletion factor of AW for yellow-fleshed kiwifruit was approx. 50% lower than that reported for 'Hayward' in the FAO-56 paper. A previous study has focused on defining the allowable soil water depletion for precise irrigation scheduling in other tree crops, and obtained a depletion factor lower than those tabulated in the FAO-56 paper (Puig-Sirera et al., 2021). The assessment of root water uptake through changes in soil moisture is promising (Jackisch et al., 2020) and provides further information on the amount of water depleted, which can be considered complementary in estimating transpiration and water needs of plants. The greater activity of surface roots is shown by the rapidly changing pattern of both SWC around the C vines and AW depletion around the WS vines, which followed a decreasing trend with increasing soil depth ( Figure 8 ). This is consistent with previous research investigating the pattern of SWC at different depths in the root zone of kiwifruit vines, which showed that vines preferentially use near-surface water when available (Clothier and Green, 1994; Green and Clothier, 1995; Green et al., 2006). Preferential root water uptake occurred initially from the superficial, wetter parts of the soil, and then increased in the deeper layers. Vines began to extract water from the depths where water was more freely available ( Figure 9 ), in accordance with Green and Clothier (1995) who confirmed the capacity of kiwifruit vines to shift their water uptake pattern in response to changes in soil water availability. In the present study, soil water depletion was used to understand the intensity and dynamics of root water uptake throughout the root zone. Considering the dynamics of root water uptake, soil moisture thresholds should be identified for each layer, which contribute differently to triggering water stress. Therefore, in Mediterranean environments where daily vine water requirement can reach about 60-70 m³ ha^-1 during the hottest days, it is advisable to split the daily irrigation volume in two or more interventions, according to the DOE 2 depletion indication. Ensuring RAW throughout the day in the first 0-30 cm soil layer, mainly interested by root water uptake, is essential to avoid the establishment of water stress and in particular of midday depression phenomena.