Water-Yield Relationship and Vegetative Growth of Wonderful Young Pomegranate Trees under Deficit Irrigation Conditions in Southeastern Italy

This investigation was carried out through three successive seasons (2017, 2018, and 2019) on young pomegranate trees of the Wonderful cultivar to study the effect of four different irrigation treatments (100%, 75%, 50%, and 25% of crop evapotranspiration—ETc) on vegetative growth and the water–yield relationship. The study was conducted in Foggia (Apulia region, Southern Italy), an agricultural area characterized by strong wind speeds and scarce water resources. The results showed the effects of the different irrigation levels and seasons on the vegetative growth and fruit yield characteristics. The cumulative trunk diameter, the annual shoot growth, the number of fruits per tree, and the yield decreased from the full water restitution (100% ETc) to the severe water restriction (25% ETc). The weight and the size of fruits decreased significantly with the restriction of water volumes applied to the crop. A linear relationship between water consumption and yield (R = 1.00 in 2018 and 1.21 in 2019; n = 12) was found. The water use efficiency (WUE) gave no statistical differences among irrigation treatments. The yield response factors (Ky: 1.06 in 2018 and 0.99 in 2019) showed the sensitivity of pomegranates to water deficits.


Introduction
Modern water-saving agriculture aims to develop sustainable agriculture by focusing on improving water use efficiency and on the use of appropriate irrigation variables. Different plant species have different drought tolerances, the identification and evaluation of which is subject to environmental conditions. Pomegranate (Punica granatum L., Punicaceae) is fairly drought resistant, and cultivation is mainly confined to semi-arid, mild-temperate areas including most of the Mediterranean countries, which are characterized by water scarcity [1,2]. It is a domesticated tree with relatively short juvenility [3] whose fruits are of high nutritional and economic value [4]. In Italy, the pomegranate is cultivated in such southern semiarid regions as Sicily, Apulia, Basilicata, and Calabria, where water is a scarce resource. Particularly in the Apulia region, over the last decade, the area of pomegranate cultivation has increased significantly, passing from 8 ha in 2009 [5] to about 1000 ha (60% of the total national cultivation area) in cultivation area today [6], in which Wonderful is the cultivar most planted. To cope with water scarcity for this crop, there is an increasing move to more efficient technological innovation and irrigation management approaches such as the drip irrigation method and soil plastic mulching applied around pomegranate plants, which produces a positive effect on fruit yield and quality [7][8][9]. Moreover, many authors have already studied the pomegranate response under deficit irrigation (DI) strategies, with the aim of improving its development under water stress conditions, minimizing production losses, and keeping or even improving fruit quality. Some studies have demonstrated that restricting irrigation produced a decrease in photosynthesis and

Experimental and Climatic Conditions
The field trial was carried out at Foggia (41 • 27 08 N, 15 • 31 56 E and an altitude of 54 m above see level. during three consecutive seasons (2017, 2018, and 2019) with the Wonderful pomegranate cultivar, which corresponded to plants 2, 3, and 4 years old. The commercial orchard was grown from cutting and trained to a vase shape and was drip-fertirrigated and mulched with a plastic sheet along the tree rows. The drip irrigation system comprised a single pipe with drippers at with the 21 h −1 flow rate spaced every 40 cm. During the growing seasons, standard agronomic practices were applied. The soil texture was clay-loam [24] with an effective depth over 120 cm. The site of the research was in a typical semi-arid zone, characterized by a Mediterranean climate, which is classified as an accentuated thermomediterranean climate [25], with temperatures that may fall below 0 • C in the winter and exceed 40 • C in the summer. Rainfall is unevenly distributed throughout the year and is mostly concentrated in the winter months with a long-term annual average of 559 mm [26].
During each experimental season, from April to October, daily climate data were recorded at the nearest meteorological station, which was a few kilometers from the experimental area. The data were used to calculate the reference evapotranspiration (ET O ) by the Penman-Monteith equation [27]. The shaded area (SA) of the trees and the crop coefficient (Kc) were also estimated and used to calculate the maximum crop evapotranspiration (ET C ), the actual crop evapotranspiration (ETa), and other irrigation parameters.

Irrigation Treatments
Four different irrigation treatments (full water restitution of 100% of the ETc and deficit irrigation of 75%, 50%, and 25% of the ETc), starting from the bud break stage to fruit maturity, were performed. Each watering was applied by keeping a constant interval (7 days) and by changing the water volume from 100% to 75%, 50%, and 25% of the ETc net of rains [28,29] The irrigation volumes applied were measured using in-line water meters installed on each experimental plot. The experimental design was a randomized complete block design with three replicates per treatment. Each plot had three rows with 8 trees in each. The three central trees in the middle row were used for data collection. All experimental trees were surrounded by guard trees in order to minimize the effect of irrigation.

Determination of the ETc
The ETc was calculated weekly by a climatological approach, multiplying ET O by Kc [27]. Considering that the orchard had not reached the final size, the Kc values were corrected by the shaded area (SA) fraction of the soil at solar noon as the actual surface subjected to irrigation, and they were calculated by Equation (1): where, SA a = the actual shaded area; R = the average radius of the canopy's circumference, as measured on selected plants; N = the number of trees per hectare. The corrected Kc a was calculated by Equation (2) [30,31]: where, Kc a = the actual crop coefficient; Kc = the FAO crop coefficient; SA a = the actual percentage of the shaded area; SA max = the potential percentage of the soil shaded area (80%) of the mature pomegranate orchard (N = the planning density equal to 606 trees per ha). Ultimately, the actual evapotranspiration (ETa) was calculated by Equation (3): Weekly values of ET O , Kc a , and SA a were used to obtain the weekly values of ETc, starting from the bud break stage (April) to the fruit maturity stage (September). A water balance was carried out to determine the irrigation required.

Determination of the Yield Water Use Efficiency (YWUE) and the Yield Response Factor (Ky)
The yield water use efficiency (YWUE) was calculated in each treatment from the ratio of the total yield (in kilograms) to the amount of irrigation water applied (including rainfall) to the crop (Kg·m −3 water). The yield response factor (Ky), which links the relative yield decrease to the relative evapotranspiration deficit, is expressed by the following Formulation (4) [22]: where, Ya (Kg ha −1 ) and Ym (Kg ha −1 ) are the actual and maximum yields (obtained from the fully irrigated treatment), respectively; ETa (m 3 ·ha −1 ) and ETm (m 3 ·ha −1 ) are the actual and maximum water consumptions (of the fully irrigated treatment), respectively.

Vegetative Growth Parameters
The trunk diameter (TD) was determined in three seasons, whereas the shoot length (SL) was only determined in the second and third seasons. Both were measured on three plants in each irrigation plot. The TD was measured two times in each year, at bud break (in April) and at the end of growth season (in October). It was measured at a marked point, 50 cm above the ground level, using the Vernier digital caliper, and expressed in millimeters. The annual trunk growth (ATG) and the trunk cumulative growth (TCG) were also calculated. The SL was determined at the end of each growing season on four tagged one-year-old shoots in the four direction points (East, West, North, and South) using a tape and expressed in centimeters.

Yield, Morpho-Pomological Characteristics of the Fruits
In the first experimental season (2017), all the flowers were thinned by hand, to guarantee the adequate development of trees, whereas, in the second and third years, fruiting was allowed, and the harvests were carried out on 18 October, 2018 and on 29 October, 2019, respectively. At harvest time, the fruits per tree for each plot were counted and weighted. The average number of fruits per tree and the yield (ton·ha −1 ) were calculated. Samples of 12 fruits from each irrigation treatment of each replicate were randomly collected. Each fruit was weighed (g) and measured in length (mm) and diameter (mm).

Statistical Analysis
The results were evaluated with one-way ANOVA using JMP ® software version 8 (SAS Institute Inc., Cary, NC, USA) and average values were compared with the Tukey test. Standard deviations (SD) were calculated using Excel software of the Office 2007 ® suite (Microsoft Corporation, Redmond, WA, USA). Percentage values were transformed to arcsine prior to analysis of variance.

Weather Conditions and Water Consumption
In Table 1, the climatic data and the ETc in the three experimental seasons (April-September) are reported. The weather conditions between seasons were quite similar, whose values of the total ETo (a representative climate evaporative parameter) were 1080.7, 1061.1, and 1044.7 mm in 2017, 2018, and 2019, respectively. During each growing season, the ETc increased with increasing canopy and evaporative air demand, reaching the peak in July. The total seasonal ETc also increased over the years, with increasing plant development from 310.4 to 375.4 to 518.9 mm in 2017, 2018, and 2019, respectively. In each season, the cumulated water requirement values (ETc-0.70 p) became positive in June. Therefore, the differentiation of irrigation treatments in each year began from the flower-petals-opening fall stage (fruit setting) to the fruit-ripening stage. Table 1. The monthly mean maximum and minimum temperatures (T max and T min ), the relative air humidity (RH max and RH min ), the wind speed (W s ), the radiation (Rad), the total precipitation (p), the reference evapotranspiration (ET O ), the percentage of the shaded area of the trees (SA a ), the crop coefficient (K C ), the crop evapotranspiration (ETc), and the cumulated water requirement (ETc-0.70 p) during the 2017, 2018, and 2019 seasons.   Table 2 shows the irrigation volumes, whose values obviously increased over the years as the canopy of the plants and the relative Kc and ETc increased. The seasonal irrigation volumes for the 100%, 75%, 50%, and 25% ETc treatments were: 101, 83, 55, and 27 mm, respectively, with 11 watering events in 2017; 140, 105, 70, and 35 mm, respectively, with 12 watering events in 2018; and 395, 296, 197 and 99 mm, respectively, with 14 watering events in 2019 (Table 2).

The Effects of Irrigation Treatments on Vegetative Growth
The annual trunk growth (Figure 1) was the lowest in the first year, whose diameter values were not statistically different among the irrigation treatments (average = 7.0 mm). In the second and third years, ETc 100 and ETc 75 treatments showed a significantly higher growth than in the first year. ETc 50 and ETc 25 , although showing some differences among years, were not statistically different.
Trunk cumulative growth (TCG) (Figure 2) gradually increased with some differences among the irrigation treatments and the two seasons. During the first season, the increase was almost uniform in all the treatments. It seems that all irrigation treatments supplied enough water for plants because of their small size and low water requirements. However, during the second season, and especially during the third season, the pomegranate trees had a different growth, showing a lower capacity for tolerating water restrictions [32,33]. After three consecutive seasons, the largest TCG (42.6 mm) was noticed in the ETc 100 treatment, although it was not significantly different from the ETc 75 (39.0 mm) and ETc 50 (38.8 mm) treatments but was statistically different from the ETc 25 (35.5 mm) treatment. These results, which show the sensitivity of stem growth of pomegranate trees to water restriction, were similar to those obtained by Bugueño et al. [34]. Additionally, previous studies [11,12,35] conducted on young almond, lemon, olive, and cherry trees showed that water restrictions reduced the cumulative trunk growth rate. As for the shoot development, monitored only in the second and third years, no significant differences were noted among the four spatial directions of the trees. A decrease in length as irrigation water application decreased was noted, although there were no significant differences between treatments because of the high variability of the data ( Table  As for the shoot development, monitored only in the second and third years, no significant differences were noted among the four spatial directions of the trees. A decrease in length as irrigation water application decreased was noted, although there were no significant differences between treatments because of the high variability of the data (Table 3). These results are in harmony with those obtained by Khattab et al. [13]. Significant higher values were obtained in 2019 (on average 42.1 cm) than in 2018 (on average 12.9 cm) ( Figure 3).

The Effect of Irrigation Treatments on Yield
The average number of fruits per tree (no) and the yields (t·ha −1 ) in all the irrigation treatments were significantly higher in 2019 (averages of 18.9 no and 3.6 t·ha −1 , respectively) than in 2018 (averages of 9.3 no and 2.5 t·ha −1 , respectively) ( Table 4). These differences, obviously, depended on the age of the plants. As regards the yield, it should be noted that its increase in 2019 compared to 2018 was less than proportional to the number of fruits per plant. This was due to the lower average weight of the fruit recorded in this last year, as reported in Table 5. As a general trend, the aforementioned parameters decreased with reducing irrigation levels, albeit with some differences between the years. In 2018, the ETc100 treatment gave the highest number of fruits per tree (11.6), which was not significantly different from the ETc75 (10.5) and the ETc50 (8.9) treatments, but which was statistically different from the ETc25 (6.5) treatment. Instead, in 2019, the ET100 treatment provided the highest number (26.3), which was significantly different from the ETc75 (19.7) treatment, which, in turn, was statistically different from the ETc50 (15.7) and the ETc25

The Effect of Irrigation Treatments on Yield
The average number of fruits per tree (n o ) and the yields (t·ha −1 ) in all the irrigation treatments were significantly higher in 2019 (averages of 18.9 n o and 3.6 t·ha −1 , respectively) than in 2018 (averages of 9.3 n o and 2.5 t·ha −1 , respectively) ( Table 4). These differences, obviously, depended on the age of the plants. As regards the yield, it should be noted that its increase in 2019 compared to 2018 was less than proportional to the number of fruits per plant. This was due to the lower average weight of the fruit recorded in this last year, as reported in Table 5. As a general trend, the aforementioned parameters decreased with reducing irrigation levels, albeit with some differences between the years. In 2018, the ETc 100 treatment gave the highest number of fruits per tree (11.6), which was not significantly different from the ETc 75 (10.5) and the ETc 50 (8.9) treatments, but which was statistically different from the ETc 25 (6.5) treatment. Instead, in 2019, the ET 100 treatment provided the highest number (26.3), which was significantly different from the ETc 75 (19.7) treatment, which, in turn, was statistically different from the ETc 50 (15.7) and the ETc 25 (14.0) treatments. Similar to the number of fruit, in 2018, the highest yield of the ETc 100 treatment (3.67 t·ha −1 ) was not significantly different from that of the ETc 75 (2.79) treatment but was statistically different from the ETc 50 (2.24) and the ETc 25 (1.31) treatments. In 2019, the ETc 100 treatment's highest values (5.94 t·ha −1 ) were significantly different from the ETc 75 (4.10) treatment, which, in turn, was statistically different from the ETc 50 (2.97) and the ETc 25 (1.57) treatments. These results showed that, in the first productive year (2018), the pomegranate was not very sensitive to moderate (75% and 50%) water stress treatments but was very sensitive to severe stress (25%), which caused a significant yield reduction. In the successive seasons, any stress caused a significant yield reduction. These results are in harmony with those obtained in other studies [16,36], but are in contrast with those of Centofanti et al. [21], who reported that the DI strategies did not significantly affect the yield in the Wonderful cultivar.  Furthermore, as it is seen in Figure 4, a linear crop-water-consumption-yield relationship was found in both years (R = 0.85 in 2018 and 0.99 in 2019; n = 12), whose yield reduction was directly related to the reduction in water consumption. This results differ from those obtained from Aydin et al. [37] who found a second-degree polynomial relationship between the irrigation water and the yield.
Furthermore, as it is seen in Figure 4, a linear crop-water-consumption-yield relationship was found in both years (R = 0.85 in 2018 and 0.99 in 2019; n = 12), whose yield reduction was directly related to the reduction in water consumption. This results differ from those obtained from Aydin et al. [37] who found a second-degree polynomial relationship between the irrigation water and the yield.

The Effect of Irrigation Treatments on Morphological Characteristics of Fruit
The average weight, diameter, and length of fruits showed significantly lower values in 2019 compared to 2018. In both years, these parameters showed the highest values in 2019 compared to 2018. In both years these parameters showed the hhighest values in ETc 100 treatment, although these were not significantly different from the ETc 75 and the ETc 50 treatmetns and were not different from the ETc 25 ones (Table 5).These results are in accordance with a previous study [38] that reported a reduction in the fruit size in pronounced water deficit levels. These results obtained on the physical prosuctive response to irrigation can provide information relevant to production and industrial purposes. For the latter, in addition to the qualitative aspects of the juice (content of bioactive compounds, organic acids, sugars, etc.), reported in a prvious study [23], the knowledge of the morpholological properties of fruit is useful in designing optimal equipment for harvesting, trasporting, grading, cleaning, storing and packaging as well as processing the whole fruit into other derived products.

The Water use Efficiency (WUE) and the Yield Response Factor (Ky)
As a general trend, Table 4 indicates that the WUE values were slightly higher in 2018 (they varied between 1.85 and 2.08 kg/m 3 ) than those in 2019 (they varied between 1.31 and 1.41 kg/m 3 ). However, in both years, no statistical differences in the WUE among irrigation treatments were noted, although, in both years, slightly higher values were obtained in the fully irrigated treatment. These results differ from other previously obtained results [13,39], which indicated significantly higher WUE values in trees under highly irrigated levels.
As for the yield response factor, the results showed a slightly higher value in 2019 (1.21) than in 2018 (1.00) ( Figure 5). In any case, according to the Ky obtained in this study, it can be said that young Wonderful pomegranate plants are sensitive to drought stress during the growth period. Our Ky values were higher than that (0.81) reported by Taha [40], which expressed a moderate tolerance to water stress, but were much lower than that (1.57) reported by Aydin et al. [37] for the Zivzik variety, which expressed a high sensitivity to water stress. These differences may be due to factors other than water, such as nutrients, different environments, different cultivars, and the age of the trees, which also affects the yield response to water [22].

Conclusions
This study investigated the adoption of a DI strategy (restitution of 100%, 75%, 50%, and 25% of the ETC) during three successive growing seasons on pomegranates to evaluate the effects on vegetative growth, yield, the water use efficiency, and the yield response factor. The results showed that the plants at the end of the third experimental trial reached the significantly highest cumulative trunk diameter by the restitution of 100% of the ETc, while it gradually decreased with the increase in water stress treatments. Additionally, the shoot length, on average, decreased as irrigation water applications decreased, although insignificant differences were noted among treatments because of the high variability of the data. As a general trend, the average number of fruits per tree, the fruit size, and the yield decreased with reducing irrigation levels, among which the severe water restriction (25% ETc) always provided the statistically lowest values. A linear crop-waterconsumption-yield relationship was found, where yield reduction was directly related to reduced water use. Moreover, although slightly higher WUE values were obtained in the fully irrigated plants, no statistical differences among irrigation treatments were noted. The results of the yield response factor to the reduction in water use showed values of 1.00 and 1.21 in the 2018 and 2019 seasons, respectively. This means that, under our experimental conditions, the young Wonderful pomegranate plants were sensitive to drought stress during the growth period.
In conclusion, based on the above-mentioned results, young Wonderful pomegranate trees in semiarid areas, also characterized by strong speed winds, require full water restitution throughout the dry season to reach optimal growth and yield. However, considering that the effect of DI on pomegranate productive responses has been only recently studied and that results from this and several other previously cited studies are contradictory, longer-term studies are needed in semiarid areas to better predict physiological responses to DI strategies relative to the productivity of pomegranate.

Conclusions
This study investigated the adoption of a DI strategy (restitution of 100%, 75%, 50%, and 25% of the ET C ) during three successive growing seasons on pomegranates to evaluate the effects on vegetative growth, yield, the water use efficiency, and the yield response factor. The results showed that the plants at the end of the third experimental trial reached the significantly highest cumulative trunk diameter by the restitution of 100% of the ETc, while it gradually decreased with the increase in water stress treatments. Additionally, the shoot length, on average, decreased as irrigation water applications decreased, although insignificant differences were noted among treatments because of the high variability of the data. As a general trend, the average number of fruits per tree, the fruit size, and the yield decreased with reducing irrigation levels, among which the severe water restriction (25% ETc) always provided the statistically lowest values. A linear crop-water-consumptionyield relationship was found, where yield reduction was directly related to reduced water use. Moreover, although slightly higher WUE values were obtained in the fully irrigated plants, no statistical differences among irrigation treatments were noted. The results of the yield response factor to the reduction in water use showed values of 1.00 and 1.21 in the 2018 and 2019 seasons, respectively. This means that, under our experimental conditions, the young Wonderful pomegranate plants were sensitive to drought stress during the growth period.
In conclusion, based on the above-mentioned results, young Wonderful pomegranate trees in semiarid areas, also characterized by strong speed winds, require full water restitution throughout the dry season to reach optimal growth and yield. However, considering that the effect of DI on pomegranate productive responses has been only recently studied and that results from this and several other previously cited studies are contradictory, longer-term studies are needed in semiarid areas to better predict physiological responses to DI strategies relative to the productivity of pomegranate. Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.