Can Sustained Deﬁcit Irrigation Save Water and Meet the Quality Characteristics of Mango?

: Mango is one of the most cultivated tropical fruits worldwide and one of few drought-tolerant plants. Thus, in this study the effect of a sustained deﬁcit irrigation (SDI) strategy on mango yield and quality was assessed with the aim of reducing irrigation water in mango crop. A randomized block design with four treatments was developed: (i) full irrigation (FI), assuring the crop’s water needs, and three levels of SDI receiving 75%, 50%, and 33% of irrigation water (SDI 75 , SDI 50 , and SDI 33 ). Yield, morphology, color, titratable acidity (TA), total soluble solids (TSS), organic acids (OA), sugars, minerals, ﬁber, antioxidant activity (AA), and total phenolic content (TPC) were analyzed. The yield was reduced in SDI conditions (8%, 11%, and 20% for SDI 75 , SDI 50 , and SDI 33 , respectively), but the irrigation water productivity was higher in all SDI regimes. SDI signiﬁcantly reduced the mango size, with SDI 33 generating the smallest mangoes. Peel color signiﬁcantly changed after 13 days of ripening, with SDI 75 being the least ripe. The TA, AA, and citric acid were higher in SDI 75 , while the TPC and ﬁber increased in all SDI levels. Consequently, SDI reduced the mango size but increased the functionality of samples, without a severe detrimental effect on the yield.


Introduction
Food production demand has been continuously on the rise and is projected to increase by up to 50% by 2050 [1]. This has mainly happened due to global population growth, which despite slowing down in some regions continues to expand in others, with global population predictions of almost 10 billion by 2050 and 11.2 billion by 2100 [1]. This means that the natural resource base upon which agriculture depends will be increasingly stressed.

Plant Material and Experimental Design
The experiment was carried out in 2020, in Almuñécar (Granada, SE Spain, 36 • 48 00 N, 3 • 38 0 W). The average annual rainfall is 449 mm, and the climate can be classified as Mediterranean subtropical [12]. During the experimental irrigation period, climatic conditions were characteristic of the study area, with air temperature values that ranged between 10 and 31 ºC, and an average maximum relative humidity of 94.2%.
The soil can be classified as typical xerorthent, which is shallow, varied in depth, and formed from weathered slates, with 684, 235, and 81 g kg −1 of sand, silt, and clay, respectively, leading to good drainage.
The experimental mango plantation is located in typical terraces of the area, with a single row of 16-year-old mango trees (Mangifera indica L. cv. "Osteen" grafted onto "Gomera-1") in each bench, with trees spaced 3 m apart, and with a distance between terraces of 5 m (around 600 trees per ha). Taking into account the lack of literature about this strategy and the ease of implementation for farmers, three sustained deficit irrigation (SDI) regimes were applied, which corresponded to 33%, 50%, and 75% of crop evapotranspiration (ET C ). This strategy allows the crop to adapt to the stressful situation gradually as water becomes scarce and adapt the vegetative development of trees to water deficits. Additionally, a control treatment was established, in which trees were fully irrigated at 100% ET C . The irrigation requirements were estimated using the reference evapotranspiration (ET 0 ) calculated according to the Penman-Monteith methodology, and local crop coefficients, K C , previously defined in the studied area [13]. For this, a weather station located at the same experimental farm was used, obtaining the ET 0 and rainfall. Taking into consideration the crop's water requirements, total amounts of irrigation water of 562.1, 411.1, 290.8, and 247.5 mm were applied in the control, SDI 75 , SDI 50 , and SDI 33 treatments, respectively.
The fruits were harvested manually at the pre-climacteric hard-green stage and the yield per tree was estimated. After harvest, 4 batches of 4 randomly selected fruits per treatment were stored for 13 days at 22 ± 1.9 • C and 47 ± 4% relative humidity for further analysis.
Morphological characterization included the size and weight of 16 fruits per treatment (4 whole fruits × 4 repetitions × treatment), which was done at the ripening stage (day 13). Moreover, the specific weight of the pulp, peel, pit (endocarp), and seed was also recorded. The size included the fruit length, width, and thickness measured with a digital caliper (Mitutoyo 500-197-20, Kawasaki, Japan), while a digital Gibertini scale (EU-C LCD series, Milan, Italy) was used for the weight.
Instrumental color and photography. The peel color of mango samples was measured using a Minolta CR-300 colorimeter (Minolta, Osaka, Japan). Three evenly distributed places along the green side of the fruit were selected and a mean value was used; the same procedure was done for the red side of the fruit. Peel color was measured every day during 13 days of ripening and was expressed using the CIEL*a*b* system. These coordinates determine color in a tridimensional space. Numeric values include the luminosity (L* = 0 black: 100 white), red (a*)-green (−a*), and yellow (b*)-blue (−b*) coordinates. Additionally, photographs were also taken every day using a Nikon D3400 camera (Tokio, Japan) and a light box to uniform the lighting for all pictures.
Titratable acidity and total soluble solids. Both measurements were done in mango juice with an acid-base potentiometer (877 Titrino plus, Metrohm ion analyses CH9101; Herisau, Switzerland) for the titratable acidity (0.1 mol L −1 NaOH), and an Atago digital refractometer (model N-20; Atago, Bellevue, WA, USA) to determine the total soluble solids at 20 • C. The values were expressed as g citric acid L −1 and • Brix, respectively.
Organic acids and sugars were measured as previously described by Carbonell-Barrachina et al. [14] with some modifications. Briefly, 5 mL of mango juice was centrifuged at 10,000 rpm during 15 min. One milliliter aliquot of the supernatant was filtered through a 0.45 µm millipore membrane filter (Billerica, MA, USA) and injected into a high-performance liquid chromatograph (HPLC) Hewlett Packard HP 1100 (Wilmington, DE, USA) with a Supelcogel TM C-610H column (30 cm × 7.8 mm) with a precolumn (Supelguard 5 cm × 4.6 mm; 219 Supelco, Bellefonte, PA, USA). A refraction index detector (RID) and a diode-array detector (DAD) were used for sugar and organic acid measurements. Analyses were done in quadruplicate and the results were expressed as concentrations g L −1 of fresh weight (fw).
The mineral contents were determined as previously described by Cano-Lamadrid et al. [15] using a microwave digestion unit Ethos Easy, Milestone (Milestone, Sorisole, Italy) equipped with a rotor for ten TFM (chemically modified PTFE) vessels for sample mineralization and an inductively coupled plasma mass spectrometry (ICP-MS) instrument Agilent 7500× Octopole Reaction System (ORS) (Agilent Technologies, Tokyo, Japan) for mineral determination. The measurements were done in lyophilized samples and the results (mean of 4 replications) were expressed as mg kg −1 freeze dried mango.
Antioxidant activity and total phenolic content. To measure these parameters, the methodology previously reported in the literature [3,16] was used, with the slight modification that for the extraction step, 5 mL extractant (MeOH/H 2 O 2 (80:20, v/v) + 1% HCl at 20 • C) for 0.5 g of lyophilized mango was used. All measurements were performed in an ultraviolet-visible (UV-vis) spectrophotometer (Helios Gamma model, UVG 1002E; Helios, Cambridge, UK).
Finally, total dietary fiber was determined following the AOAC Official Method 985.29 [17] using 1 g of lyophilized mango and the TDF-100 kit (Sigma-Aldrich, St. Louis, MO, USA). Together with mango samples, blank and reference samples were also analyzed simultaneously in triplicate for comparison.
Statistical Analysis. A randomized block design with 4 replications per treatment was used, monitoring the 5 central trees per replication. One-way analysis of variance (ANOVA) followed by Tukey's multiple range test was used to check the statistical differences among samples, and only those parameters significantly affected by the treatments were further considered for Pearson's correlation analysis. Statistical differences were considered significant when p < 0.05. To run the analyses, the XLSTAT Premium 2016 software was used, while Sigma Plot 11 software was used for figure preparation.

Yield and Morphology
The average yield for trees under SDI 33 , SDI 50 and SDI 75 was 26.2, 28.8, and 32.3 kg per tree, respectively, compared to 35.8 kg per tree for the control treatment. As was expected, the effect of water stress induced by deficit irrigation treatments was evident by lowering the productivity with respect to non-stressed control trees (100% ET C ). Taking into account the irrigation water applied per tree in each treatment, an irrigation water productivity of 3.8, 4.7, 5.9, and 6.4 kg m 3 was reached for control, SDI 75 , SDI 50 and SDI 33 , respectively. Table 1 shows the morphological characteristics of mango cv. Osteen for each irrigation treatment. As observed, the control mangoes (FI) presented the highest weight and size compared to those fruits grown under DI conditions. For instance, FI mangoes reached a longitudinal diameter of 155 mm with a width and thickness of 98 and 82 mm, followed by SDI 75 (146, 95, an1 mm, respectively) and SDI 50 (143, 91 and 77 mm) with similar length between them, even though the SDI 75 was more similar to the control in terms of width and thickness. The most affected treatment in terms of size was SDI 33, the one which received the lowest amount of irrigation water. Similar results were also observed for the weight, where FI is the treatment with the highest weight (681 g), followed by SDI 75 (586 g) and SDI 50 (540 g), and the lowest value being that of SDI 33 (455 g). However, in terms of relative values of each part of the fruit (Figure 1), it can be observed that an increase in the intensity of the DI treatment led to a greater pit fruit, with no differences being found for the weights of the peel or flesh. The main mango producing countries, such as Malaysia with 80,841 t year −1 and more precisely the Perlis Department of Agriculture, have established grading standards which have been developed according to qualitative and quantitative criteria for mango cv. Harumanis. They established three grades: (i) A (weight >400 g), (ii) B (weight 351-399 g), and (iii) C (weight <350 g), requiring a standardize shape and size for all categories and no peel defects, except for the grade C. As observed, all the studied irrigation treatments (cv. Osteen) reached the top grade for mangoes (A) in terms of weight [18]. Additionally, according to international marketing standards (CODEX STAN 184-1993) for mango of Codex Alimentarus, the mangoes cultivated under FI and SDI 75 fall under the C-size group (representing the highest weight ranging 551-800 g), while SDI 50 and SDI 33 belonged to the B-size group (351-550 g) [19]. It is important to highlight that this standard tolerates a minimum weight of mangoes of 200 g, which is the lowest group (A) with values between 200 and 350 g. Other authors working with cv. Osteen of 11-year old trees cultivated in Sicily (Italy) reported lower fruit weights (462 g) than those of FI, SDI 75 and SDI 50 , but similar to the most stressed treatment SDI 33 [20]. Lower weight and size values were also reported by other authors for the same cultivar (400 g), but also for 27 other cultivars under study [21]. Among the scarce literature regarding the influence of DI on mango quality parameters, the morphological parameters of mango in RDI, PRD and non-irrigated conditions over 3 years have been reported by Spreer et al. [5]. They concluded that the weight was mainly affected in the second year of treatment and by RDI and non-irrigated strategies, while PRD samples remained similar to the control. These findings were reported on the Chok Anan cultivar and the mean values ranged from 223 g in the 1st season to 319 and 313 g in the 2nd and 3rd season, respectively. It has been clearly demonstrated that the effect of the irrigation strategies are cultivar-and season-dependent. However, not only drought conditions can affect the final production and quality; other climate variables such as air temperature or solar radiation are important factors that ultimately will determine the plant growth and development through the seasons. However, in our case, the climatic conditions registered during the monitoring period were very similar to those traditionally registered in the grown area of SE for this crop. This fact reinforces the obtained results in this work, evidencing how important is the irrigation water on the fruit development, and how severe water stress generated through the SDI strategy can significantly decrease both the size and weight, as well as increase the by-products of mangoes such as the pit.

Impact of Storage Time and Deficit Irrigation on the Appearance (Color) of Mango Fruits
Mango peel color represents one of the most important quality attributes with a key role in consumer acceptance, because it is the first trait perceived by consumers [20]. Figure 2 presents the mangoes from day 1 after harvest until day 13 of ripening storage. Mangoes are a climacteric tropical fruit in which the ripening process is associated with an increase in cellular respiration and ethylene production [22]. This means that the fruits were harvested when mature, but before ripening has started. In this figure, the changes in color with time can be easily observed for all treatments over 13 days of ripening, in both sides of the fruit. A color change from green to yellow was observed on the shaded side of the fruit (the green one) and from a red goji berry to a coral pantone on the sun-exposed side (the red one). The visual differences that could be observed among the treatments led to more yellow-coral fruits at the end of the ripening for FI and SDI 33 samples, while more green-orange color was found for the SDI 75 and SDI 50 fruits. However, these differences can be attributed to the mango conditions at the beginning of ripening, because as seen, those from the FI and SDI 33 already contained more reddish notes, while those from SDI 75 and SDI 50 contained more green notes. It is important to highlight that only one fruit per treatment was chosen to prepare the visual color changes represented in Figure 2; however, these fruits were selected to represent instrumental color coordinates presented in Figure 3, where values represented the mean of eight mangoes per treatment.   Percentage of parts of mango fruit as affected by deficit irrigation. Different letters mean significant differences (p < 0.05) among treatments, according to Tukey's least significant difference test.

Impact of Storage Time and Deficit Irrigation on the Appearance (Color) of Mango Fruits
Mango peel color represents one of the most important quality attributes with a key role in consumer acceptance, because it is the first trait perceived by consumers [20]. Figure 2 presents the mangoes from day 1 after harvest until day 13 of ripening storage. Mangoes are a climacteric tropical fruit in which the ripening process is associated with an increase in cellular respiration and ethylene production [22]. This means that the fruits were harvested when mature, but before ripening has started. In this figure, the changes in color with time can be easily observed for all treatments over 13 days of ripening, in both sides of the fruit. A color change from green to yellow was observed on the shaded side of the fruit (the green one) and from a red goji berry to a coral pantone on the sunexposed side (the red one). The visual differences that could be observed among the treatments led to more yellow-coral fruits at the end of the ripening for FI and SDI33 samples, while more green-orange color was found for the SDI75 and SDI50 fruits. However, Figure 1. Percentage of parts of mango fruit as affected by deficit irrigation. Different letters mean significant differences (p < 0.05) among treatments, according to Tukey's least significant difference test.
The mean values of color change for each treatment during 13 days of storage are represented in Figure 3, using the CIEL*a*b* color coordinates. As observed in both figures, the examined mangoes presented a wide heterogeneity for the peel color, represented as high error bars in Figure 3. Regarding the shade side of the fruit (green side), it was observed a slight increase in L* values from day 1 to 13 for all the treatments (FI: 53-57; SDI 75 : 46-47; SDI 50 = 48-53; and SDI 33 = 52-54). This means that all treatments become lighter with storage time. Significant differences were also observed among the irrigation treatments: SDI 75 and SDI 50 were both darker at the beginning and at the end of storage, as can be seen in Figure 3. The a* color coordinate significantly changed with the storage time, from green values at the beginning of storage to red notes at the end. No significant differences among irrigation treatments at the beginning of storage were observed, but at the end, SDI 75 recorded the highest values of green notes, followed by SDI 50 . Finally, the b* color coordinate, which represents blue colors when the values are negative and yellow colors when the values are positive, also changed completely within the storage time registering mean values of all treatments between 24 (1st day) and 37 (13th day). The differences were also significant among treatments in SDI 75 , where mangoes had the lowest values of b* coordinates both at the beginning and end of storage, which means that this treatment has lower yellowish notes. These values also showed that each treatment ripens differently; for instance, SDI 50 was the treatment with low values of b* color coordinate at the beginning of storage and raised to values similar to SDI 33 and FI at the end of storage. The same trend was observed for the a* color coordinate, which at the beginning of storage was similar for all treatments, but at the end showed significant differences with SDI 75 fruits being the less reddish ones. However, these differences can be attributed to the mango conditions at the beginning of ripening, because as seen, those from the FI and SDI33 already contained more reddish notes, while those from SDI75 and SDI50 contained more green notes. It is important to highlight that only one fruit per treatment was chosen to prepare the visual color changes represented in Figure 2; however, these fruits were selected to represent instrumental color coordinates presented in Figure 3, where values represented the mean of eight mangoes per treatment.  Evaluating the sun side of the fruit (red side), an increase in all color coordinates was observed during storage from day 1 to 13 (mean values for all treatments L* = 41-44; a* = 13-22; and b* = 14-26). As seen, the mangoes become sightly lighter with storage time, and with higher yellow-red notes. Regarding the irrigation treatments, significant differences were observed at the beginning of the ripening process, being SDI 75 treatment that with the darkest fruits, while the control and SDI 33 treatments led to the most yellowish fruits. However, at the end of ripening, the differences were not significant for any of the color coordinates.
Overall, the color change from green to yellow occurs during the ripening process when the transition of chlorophyll to carotenoids takes place, as well as other chemical reactions such as the biochemical conversion of starch into sugars, or the organic acids loss through oxidation [22]. The main phytochemicals responsible for pigmentation in mango peel are the carotenoids and anthocyanins, depending on the cultivar. The former imparts lighter and yellow to red colors, while the latter contributes to darker and pink-red to blue colors on the mango peel [23]. In this sense, the peel color can be used as an indicator of quality of the ripening stage, as the authors report a positive correlation between color and carotenoids or sugar content regardless storage temperature or harvest stage [22]. According to the shaded side of the fruit results, a soft deficit irrigation (SDI 75 ) led to more greenish fruits compared to the other treatments (yellow-orange fruits). This means that SDI 75 can contribute to a higher shelf-life of mangoes, helping to extend the marketing period of this cultivar.

Impact of Water Stress on Total Soluble Solids, Titrabale Acidity, Organic Acids and Sugars
Titratable acidity, organic acids, total soluble solids and sugars were recorded at the end of ripening and the results are presented in Table 2. These parameters are essential as mango flavor is primarily generated by acids, sugars, and volatile compounds providing the desirable gustatory quality of mangoes. The TA results showed that SDI 75 presented the most acid samples, while SDI 50 the lowest ones. Regarding the total content of organic acids, ranging from 4.70 to 7.20 g L −1 , an increase was observed for all DI treatments, with SDI 33 as the treatment with the highest content. On the other hand, regarding the TSS, only SDI 75 treatment reached similar values to those of the control, because in most stressed conditions, TSS values were slightly decreased. The total amount of sugars ranged from 149 to 168 g L −1 , with all treatments being statistically equivalent except for SDI 50 which registered the lowest values. Finally, Figure 4 represents the organic acid (a) and sugar (b) profiles for each irrigation strategy. As observed, the major organic acids were citric, representing 32% of total organic acids, followed by tartaric and quinic, with 22% each, and finally malic and shikimic acids representing 12% of the total acids. On the other hand, the main sugars found within this study were sucrose (66%), fructose (27%) and glucose (7%). These results agreed with those previously reported by other authors in mango fruits of different cultivars, such as Gleen, Mamme, Saigon, etc. [24,25]. It was observed that the SDI 75 strategy increased the citric acid content, which leads to a stronger sour taste in the fruit, while more severe DI treatments lead to more bitter and astringent fruits due to the increase in quinic and shikimic acids. On the other hand, sucrose was lower in SDI 75 , with respect to the other treatments, and a reduction of glucose and an increase of fructose was observed for SDI 50 treatment. During the ripening process, organic acid levels are often inversely related to sugar levels; this means that sugars are accumulated mainly due to the starch hydrolyzation from unripe mango to more simple sugars such as glucose [25,26]. On the other hand, organic acids that are usually accumulated in unripe fruits strongly decrease during fruit ripening. For instance, a reduction in mango acidity it reported after 15 days of storage, which was attributed to the enhanced activity of citric acid glyoxylase implicated in citric acid degradation, as well as to their conversion into sugars and additional consumption in the metabolic process of fruit [22,27]. Consequently, these might explain how the SDI 75 strategy led to mangoes less ripe after 13 days of storage, which might help to increase the mangoes shelf life on the fresh market. Other authors also reported similar values of TA and TSS, which were correlated with the ripening period obtaining the highest values of TSS (17 • Brix) and lowest acidity (0.2%) after 120 h of ripening [28]. The effect of DI on fruit quality was studied on different fruits such as almonds, olives, pistachios, tomatoes, grapes, etc., and authors reported a greater quality in hydro sustainable (hydroSOStainable) fruits [2,[29][30][31][32][33]. However, little information regarding the DI influence on mango quality exists, and almost none on cv. Osteen. Authors working with Chok Anan, studied the effect of different DI strategies (FI, RDI, PRD and non-irrigated) on the fruit quality both pre-and post-harvest and concluded that all DI produced fruits with desirable traits for the fresh market [34]. Additionally, they reported that those fruits cultivated under RDI conditions significantly increased the TSS sugars after 6 days of ripening. The same happened with mangoes cv. Guifei grown under RDI at the maturity, which helped to raise the TSS and also the TA in mangoes in two consecutive seasons [4]. In this sense, the authors highly recommended the use of RDI during maturity stage because it was able to significantly increase the water use efficiency and quality parameters such as soluble solids, organic acids, and carotenoids in fruit. This phenomenon might occur because the maturity stage demands low water volume, and the RDI strategy might reduce the water potential of fruit cells, increasing the ability of cells to absorb nutrients enhancing in this way the soluble solids content, sugars, organic acids and other compounds [4]. As seen, there are pros and cons of the RDI strategy, because it can either negatively affect the weight or positively affect the fruit quality. In this case, the stress level of RDI must be optimized to assure both parameters as previously observed in other crops [3].  0.99ab 7.16a 15.3b 161a † * and *** significant at p < 0.05 and 0.001, respectively. ‡ Values (mean of 4 replication) followed by the same letter, within the same column and factor, were not significantly different (p < 0.05), according to Tukey's least significant difference test.   Table 3 shows the effect of SDI on both plant micro-(Fe, Cu, Mn and Zn) and macro-(K, P, Mg and Na) nutrients. Potassium represented almost 80% of the cations found in mangoes cv. Osteen, followed by P (8.1%), Mg (7.6%), Ca (5.8%), and Na (0.22%) within macroelements. On the other hand, the highest microelement was Fe (0.09%), followed by Zn (0.06%), Cu (0.05%), and Mn (0.04%). The present values agree with other authors and the United States Department of Agriculture (USDA, Washington, DC, USA) [25,35] database considering a dry weight of mango ~16% as calculated in the present study or  Table 3 shows the effect of SDI on both plant micro-(Fe, Cu, Mn and Zn) and macro-(K, P, Mg and Na) nutrients. Potassium represented almost 80% of the cations found in mangoes cv. Osteen, followed by P (8.1%), Mg (7.6%), Ca (5.8%), and Na (0.22%) within macroelements. On the other hand, the highest microelement was Fe (0.09%), followed by Zn (0.06%), Cu (0.05%), and Mn (0.04%). The present values agree with other authors and the United States Department of Agriculture (USDA, Washington, DC, USA) [25,35] database considering a dry weight of mango~16% as calculated in the present study or reported by other authors [36]. As observed, K (7188 mg kg −1 ) was the predominant element in these mangoes and together with P, Mg and Ca (750, 696, and 533 mg kg −1 ) make these fruit a good source of these macroelements [20]. Moreover, authors working with different mango cultivars stated that cv. Osteen was the one with the highest K content (2420 mg kg −1 fw) [20]. Deficits in irrigation affected each element in a different way depending on stress level; for instance, elements such as Na, Mg, K, Ca, Mn and Zn were increased with water stress, while P, Fe, and Cu were decreased in the most severe treatments. In general, SDI 50 was the most effective level for almost all elements except Mg, which only increased in the most severe conditions of SDI 33 . It is well known that water scarcity reduces the transports of mineral nutrients from root to shoot, due to the early closure of stomata and so a decrease in transpiration rate [37]. However, there are plants able to obtain and preserve more water, creating a greater resistance to water stress due to a better water use efficiency, and minerals plays a key role. In the present study, it might be said that the plant coped with the water stress because most of the elements were maintained or increased in those mangoes' growth with a lower amount of irrigation water. The effect of DI on the mineral composition of fruit and vegetables has been studied by several researchers, however it is contradictory and elusive. For instance, there are authors concluding that an increase in irrigation water helps to increase the mineral content in fruit due to the release of a greater amount of ions in the solution with the irrigation water, also increasing the rate of absorption by the plant roots [38]. On the other hand, other authors reported no differences between FI and DI fruit minerals or even higher content in those grown under DI strategies, on different crops such as olives, tomatoes, grapes, apples, almonds and pistachios [39][40][41][42]. Table 3. Effect of irrigation dose on mineral content (mg kg −1 ) on lyophilized mango pulp as affected by the irrigation treatment.  33 17.6c 792a 646b 7709a 655a 3.08bc 7.68b 2.76b 3.79b † ** and *** significant at p < 0.01 and 0.001, respectively. ‡ Values (mean of 4 replication) followed by the same letter, within the same column and factor, were not significantly different (p < 0.05), according to Tukey's least significant difference test.

Impact of Water Stress on the Antioxidant Activity (AA) and Total Phenolic Content (TPC)
The AA and TPC were also evaluated, and the results were embodied in Table 4. These parameters are important from a functional point of view because oxidative stress is reported to be the key factor for many diseases such as cardiovascular, hypertension, atherosclerosis, neurodegenerative or cancer, mainly caused by an imbalance between reactive oxygen species (ROS) and the antioxidative defense system [43,44]. The AA was measured using three spectrophotometric assays, ABTS •+ , DPPH • and FRAP, as each antioxidant compound has different mechanism of action, either by way of single-electron or hydrogen atom transfer [45]. Moreover, the main aspect that influenced these methods' potential are the concentration and structure of phenolic compounds in the specific plant material or extract under study. As seen, significant differences were observed among DI levels for both functional parameters. In the case of AA, this was reduced in DI mangoes by ABTS •+ and DPPH • methods, while an increase within the FRAP method was shown. The observed discrepancies occurred because the reagents used in these methods react in a different way with a wide range of phenolic compounds present in the samples due to differences in their structures and concentrations [45]. In this sense, a greater value of TPC was observed for those mangoes cultivated under DI conditions (2.8 g GAE kg −1 ) compared to the full irrigation ones (2.2 g GAE kg −1 ). These results agreed with those reported by other authors (2.6 g GAE kg −1 ) in their study about lipophilic and hydrophilic antioxidant capacity of common foods, including mangoes, in the United States [46]. As seen, SDI 75 and SDI 50 boosted the TPC, while more severe conditions started to decrease it. Here, it is important to investigate irrigation strategies with different stress levels to optimize the stress in plants and avoid loss of important bioactive compounds. It is known that drought is conducive an excessive production of very reactive ROS ( 1 O 2 , O 2 and H 2 O 2 ) which damage the plant cells, and the risk of irreparable cell damage is higher under water stress conditions [47]. Nevertheless, plants have the ability to readjust and grow under drought conditions due to their adaptation mechanisms such as the antioxidant defense system (phenolic compounds, enzymes, vitamins, etc.) and osmotic adjustments (sugars, sugar alcohols and amino acids, etc.). As observed in the present study, the plant defense system was still able to produce and accumulate phenolic compounds up to a 50% reduction of the irrigation water, but a 70% water reduction started the reduction of these compounds. Table 4. Effect of irrigation dose on the antioxidant activity (mmol Trolox kg −1 ) and total phenolic content (g GAE kg −1 ) on lyophilized mango pulp as affected by the irrigation treatment. 2.75ab † *, ** and *** significant at p < 0.05, 0.01and 0.001, respectively. ‡ Values (mean of 4 replication) followed by the same letter, within the same column and factor, were not significantly different (p < 0.05), according to Tukey's least significant difference test.

Impact of Water Stress on the Fiber Content of Mango
According to the Codex Alimentarus, dietary fiber is the indigestible cell wall component of the plant and plays a key role in the human diet and health because it contains mainly polysaccharides that cannot be completely broken down by the human gastrointestinal tract and are not absorbed in the human body. This means dietary fiber helps to maintain the gut balance by increasing the beneficial bacteria (Lactobacillus, Bifidobacterium) and reducing pathogen microorganisms (Clostridium, E. coli) preventing constipation and colon cancer [48]. In this sense, the dietary fiber of mango was analyzed within this study and the results are showed in Figure 5. The total dietary mango in cv. Osteen ranged between 12 and 21%, and an increase in mango fiber was observed within the water stress fruits; in this sense, FI registered the lowest amount 12% and SDI 75 and SDI 33 the highest one (19 and 21%, respectively). Similar values for dietary fiber were also shown by other authors in the pulp of unripe mango (cv. Ataulfo) [49]. Regarding the effect of water deficit, authors stated that water deficit can affect the fiber content in tomato fruits, for instance, a strong negative correlation (R 2 = 0.908) between the fiber content of the tomato fruits and the volume of water used was previously reported [38]. Additionally, drought has also been reported to affect the dietary fiber of wheat, because under natural conditions, drought usually occurs in combination with heat and these two stresses have been reported to act synergistically to increase the dietary fiber content of wheat [50]. Finally, Pearson's correlations were run among the stress integral and all the quality variables with significant differences among treatments. In this sense, a positive relationship among the stress integral with sucrose and total sugars (R = 0.98; p < 0.02; and R = 0.97; p < 0.02, respectively) was shown, implying that deficit irrigation can contribute to sweeter mangoes as previously reported in other crops such as tomatoes, thyme, nectarines or almonds [3,[51][52][53], due to an osmotic adjustment, activated by accumulation of solutes rich in hydroxyl groups (sugars, proline, etc.) in the cytoplasm [54]. Moreover, the irrigation water use productivity obtained using SDI strategy led to 4.7, 5.9, and 6.4 kg m 3 for SDI75, for SDI50, and SDI33, respectively compared to 3.8 kg m 3 of the control. Thus, different deficit irrigation regimes together with the fruit quality characteristics are important factors to be evaluated in order to reach an efficient use of water in agriculture, assuring at the same time optimal fruit traits. Because decision making in on-farm irrigation improves through the use of biophysical and economic water productivity indicators, thus helping to increase the fruit quality and save water as shown in the present study and as concluded by other authors working on the sustainability of water resources in cultivars such as peach and olives [55,56].

Conclusions
Finally, this study contributes to an efficient and sustainable management of water resources in mango farming, producing fruits with higher quality and functionality when a sustained deficit irrigation strategy is used. It was demonstrated that controlled deficit irrigation can increase the fruit quality and functionality leading to a higher content of sugars, minerals, fiber or total phenolics. Moreover, it was shown that SDI75 and SDI50 can lead to minimal losses on fruit yield and can increase the irrigation water productivity. Thus, these results might be used by the farmers and industry when the water availability is below the mango irrigation requirements, producing fruits with a greater quality and Finally, Pearson's correlations were run among the stress integral and all the quality variables with significant differences among treatments. In this sense, a positive relationship among the stress integral with sucrose and total sugars (R = 0.98; p < 0.02; and R = 0.97; p < 0.02, respectively) was shown, implying that deficit irrigation can contribute to sweeter mangoes as previously reported in other crops such as tomatoes, thyme, nectarines or almonds [3,[51][52][53], due to an osmotic adjustment, activated by accumulation of solutes rich in hydroxyl groups (sugars, proline, etc.) in the cytoplasm [54]. Moreover, the irrigation water use productivity obtained using SDI strategy led to 4.7, 5.9, and 6.4 kg m 3 for SDI 75 , for SDI 50 , and SDI 33 , respectively compared to 3.8 kg m 3 of the control. Thus, different deficit irrigation regimes together with the fruit quality characteristics are important factors to be evaluated in order to reach an efficient use of water in agriculture, assuring at the same time optimal fruit traits. Because decision making in on-farm irrigation improves through the use of biophysical and economic water productivity indicators, thus helping to increase the fruit quality and save water as shown in the present study and as concluded by other authors working on the sustainability of water resources in cultivars such as peach and olives [55,56].

Conclusions
Finally, this study contributes to an efficient and sustainable management of water resources in mango farming, producing fruits with higher quality and functionality when a sustained deficit irrigation strategy is used. It was demonstrated that controlled deficit irrigation can increase the fruit quality and functionality leading to a higher content of sugars, minerals, fiber or total phenolics. Moreover, it was shown that SDI 75 and SDI 50 can lead to minimal losses on fruit yield and can increase the irrigation water productivity.
Thus, these results might be used by the farmers and industry when the water availability is below the mango irrigation requirements, producing fruits with a greater quality and minimal yield loses. However, building on the present results which provide baseline information of SDI regimes on mango yield and quality, further research during several seasons of water stress is required to confirm the present findings, as well as transferring the most positive SDI strategy to other mango cultivars in order to establish the behavior of other cultivars.