1 Introduction

The avocado tree (Persea americana Mill.) is cultivated throughout tropical and subtropical regions worldwide. Avocado belongs to the Lauraceae family, which is composed of 50 genera and roughly 2500–3000 species with three main horticultural races: Mexican race (P. americana var. drymifolia) cultivated in the tropical highlands, Guatemalan race (P. americana var. guatemalensis, L.O. Williams) adapted to medium elevations in the tropics and West-Indian race (P. americana var. americana) cultivated in the lowland humid tropics (Chanderbali et al. 2008; Ibarra et al. 2015). The subtropical origin is responsible for its high sensitivity to heat radiation and water stress, although the general success of commercial avocado plantations over a wide range of environmental conditions suggests some degree of morphological, anatomical and physiological adaptation.

The “Hass” is the prevailing avocado cultivar grown throughout the world and is associated with buttery and nutty flavours and a spherical shape that turns from bold green to dark purplish black when ripening. Economically important cultivars, such as Hass and Fuerte, are often grown under Mediterranean climate conditions, but the main commercial avocado plantations are located in Mexico, Dominican Republic, Peru, Indonesia, Colombia, USA, south-western Australia and parts of the European Mediterranean basin (FAOSTAT 2020; Fresch Plaza 2020). Recently, the demand for avocado has increased by more than 8% yearly due to an increased interest in healthy foods. Avocados are often perceived as a healthy food choice due to their nutritional content. The most important nutritional content is the avocados monounsaturated fatty acids (MUFA), which can potentially decrease the incidence of many diseases (Kawagishi et al. 2001; Ding et al. 2007; Woolf et al. 2009). The high lipid content of avocado fruit is due to the high number of fatty acids (oleic, palmitic, palmitoleic, linoleic, linolenic, etc.), particularly of the unsaturated type: monounsaturated and polyunsaturated (PUFA) fatty acids (Villa et al. 2011; Dreher and Davenport 2013; Flores 2019) and a potential source of bioactive compounds (Tremocoldi et al. 2018; Salazar et al. 2020; Cervantes et al. 2021).

The avocado tree is unique from a compositional perspective and displays particular and complex physiology in relation to other woody fruit trees. The flowering stage can last up to 3 months; therefore, fruit from the same tree can vary significantly in physiological age. The fruit set is low, and the fruits can stay in the tree for several months, far beyond the time needed to achieve physiological maturity to ripen when they are detached.

Avocado productivity in many producing regions is consistently low in comparison to the potential yield of 30.0 t ha−1 claimed by Wolstenholme (1986) or Silber et al. (2019) of 25–31 t ha−1, which is attributed to high flower and fruit abscission rates and typical alternate bearing of this crop (Garner and Lovatt 2016). Under a Mediterranean climate with hot and dry summers and prevailing humid winters, irrigation is vital to restore water deficits arising from high crop evapotranspiration. In the Mediterranean climate, fruit set occurs in spring at the end of the rainfall period; however, during the early summer, the fruitlet abscission limits productivity; therefore, efficient irrigation management is essential in inducing a higher yield. On the other hand, the alternate and irregular bearing in avocado trees is significant because they generate annual production fluctuations, with negative impacts on the avocado plantations (Mickelbart et al. 2012). Also, the alternate bearing in avocado can mask the impact of water stress on fruit yield (Carr 2013).

In Spain, the area devoted to subtropical crops is approximately 4.3% of the total area of fruit trees and 2.2% of the total fruit production (ESYRCE 2019). In 2019, there were 17,158 ha of irrigated avocado, of which 14,427 ha (84%) were in Andalusia (SE Spain), mainly on the coast of Malaga and Granada (Fig. 1). The main destination exports (95% of total production) of Spanish avocados are the European Union, with France as the main importer (39,000 t) with 39%, followed by Germany, Netherlands and the UK with 13, 12 and 8%, respectively (OPM 2020). The main commercially appreciated cultivar in this area and the predominant avocado is Hass, which represents 85% of the devoted area, followed by Bacon, Fuerte, Pinkerton, Reed, Gween, etc.

Fig. 1
figure 1

(A) Irrigated avocado plantation in terraces growing in Almuñécar (Granada, SE Spain) and (B) avocado fruit cv. Hass (photographs by the authors)

Full size image

The impact of climate change will be most apparent primarily through water availability (Rocha et al. 2020), with consequences that could be large and uneven, significantly impacting agricultural activities. For the Mediterranean basin, the reduction in water availability is predicted to considerably increase at the end of this century (Gorguner and Kavvas 2020). This trend, together with the expansion and intensification of land use for agriculture has promoted an increase in water demand (De Fraiture and Wichelns 2010). Today more than ever, arid and semiarid regions suffer from water shortages for agricultural use, which is the most important environmental impact on woody fruit crops, limiting productivity and sustainable development (García-Tejero and Durán 2018). That is, the water availability and its scarcity leads definitely to admit that irrigated agriculture in Mediterranean regions needs to appraise the implementation of alternative strategies to conventional practices in order to attain a higher resilience in terms of water shortage. Thus, the current circumstances denote a period of appreciable changes and water resource uncertainty for agricultural activity, making it necessary to move away from high-water-allocation cropping towards irrigation systems that maximize water savings. Efforts must be made to develop efficient water management systems in avocado plantations and adapted irrigation scheduling under a challenging environment (García-Tejero et al. 2011; Rodríguez et al. 2018).

In this study, we examined the impact of sustained-deficit irrigation on avocado cv. Hass productivity during three monitored seasons at the highest atmospheric water demand period in a commercial orchard located in the semiarid Mediterranean environment in south-eastern Spain. The main objective was to evaluate the response in terms of plant–water relations (leaf water potential and stomatal conductance) during specific developmental stages (flowering, fruit set and fruit growth) and the effects on yield and fruit quality (lipid profile) under water stress conditions.

2 Material and methods

2.1 Experimental site

The study was undertaken over three growing seasons (2017–2019) in Almuñécar, Granada coast (SE Spain, 36° 48′00″N, 3° 38′0″W). The local temperatures are subtropical to semi-hot within the Mediterranean subtropical climatic category, with an average annual rainfall of 449.0 mm (Elias and Ruiz 1977). The rainfall shows a complex spatial and seasonal variability pattern, which is intensified by its unpredictability from year to year, within the year, and spatially during unique rainfall events.

At the study site, there are shallow soils from weathered slates, and some soils are rocky, providing good drainage, and are classified as Typical Xerorthent (Soil Survey Staff 2014), with 605, 243 and 152 g kg−1 of sand, silt and clay, soil pH 7.25 (1:2.5) and bulk density of 1.04 g cm−3. The content of soil–organic carbon and total N was 8.7 and 0.68 g kg−1, respectively, and extractable P (Olsen) and available K were 7.6 and 89.7 mg kg−1, respectively.

The experimental avocado plantation is located in terraces commonly found in the study area at an altitude of 170 m.a.s.l. Each platform had a single row of 33-year-old avocado trees (Persea americana Mill. cv. Hass grafted onto Lula), which were healthy and uniform in size, and spaced 8 m apart (~ 250 trees per hectare) (Fig. 1). The entire experimental avocado orchard was managed with the same fertilization (992 g N, 520 g P2O5 and 460 g K2O). Common local cultivation techniques (pruning, weeds and pest control) were applied uniformly across all irrigation treatments.

2.2 Irrigation treatments and experimental design

Three sustained-deficit irrigation (SDI) treatments were implemented, receiving 33% (SDI33), 50% (SDI50) and 75% (SDI75) of ETC (crop evapotranspiration), and a control was established with a fully irrigated plot at 100% ETC (C100), which is the amount required to meet the evapotranspiration demand for the irrigation period. To estimate the irrigation requirements, the reference evapotranspiration (ET0) was calculated according to the Penman–Monteith methodology, using a weather station installed in the same experimental orchard; crop coefficients (KC) values were based on guidelines provided by Allen et al. (1998). The irrigation system for each treatment consisted of two lateral lines per row located 30 cm on each side of the trunk with a combination of self-regulating drip emitters per tree, providing 4 and 8 L h−1, and separated by 50 cm. In general, the experimental plots were irrigated on alternate days (3–4 h) during flowering, fruit set and early fruit development (April–June), daily during the maximum evaporative demand (4–5 h) at fruit growth (July to September), and after this period, the frequency (2–3 h) was reduced to one or two times per week (October–March). That is, the water stress was induced by decreasing water irrigation doses, being the average irrigation per season during the three monitoring period for control trees C100, SDI75, SDI50 and SDI33 amounted to 7868 ± 286, 5837 ± 420, 4057 ± 320 and 2714 ± 150 m3 ha−1, respectively.

The experimental design was a completely randomized block design with three replications per irrigation treatment; the five central trees from each row were included for growth, physiological and fruit yield measurements; the others served as border trees.

2.3 Field measurements

During the irrigation period (from April to October), the crop water status was monitored by measurements of the midday leaf water potential (Ψleaf) and stomatal conductance to water vapor (gS) between 12:00 and 13:30 GTM, and with a periodicity of 7–10 days. The gS was measured with a porometer SC-1 (Decagon Devices, INC, WA, USA) in three selected leaves from the medium portion of fruitless shoots at 1.5 m of height and with south-eastern facing that was completely exposed to the sun (monitoring 5 trees per irrigation treatment). Similarly, the Ψleaf was measured using a pressure chamber (Soil Moisture Equipment Corp., Sta. Barbara, CA, USA), monitoring three leaves per tree. Measurements for both physiological parameters were made throughout the flowering, fruit set and fruit growth phenological periods during the growing season.

After harvesting the five monitored central trees for each treatment, the height, canopy diameter (at 1.75 m from the soil surface) and trunk circumference were measured. Canopy volume was calculated by assuming the tree approximated to the shape of one-half of a prolate spheroid by using the following equation (Castel and Phillips 1980; Avilán et al. 1997):

$$CV=\frac{4}{3}\pi {r}^{2 }\frac{1}{2} H$$

where CV (m3) is canopy volume; r is canopy radius; H is tree height.

The trunk circumference (cm2) was converted into trunk cross-sectional area (TCSA):

$$TCSA=\frac{{C}^{2 }}{4\pi }$$

where C is the trunk circumference.

The canopy efficiency (kg m−3) was calculated by dividing the yield of the tree by its canopy volume and water use efficiency (WUE) as fresh avocado yield divided by total seasonal irrigation water applied (kg m−3). At harvest throughout the three-year monitoring period (February–March), the total fruit yield per tree was registered for each treatment. Twenty-five fruits per tree were randomly collected to measure both their individual weight and their length and width.

2.4 Quality parameters

2.4.1 Morphological analysis and lyophilization process

The harvested fruits were kept in room temperature conditions (20 ± 1 °C) and analyzed after 13 days of ripening. Once reaching the maturity stage, 5 fruits were randomly selected from each repetition of treatments, and their size (length and width) and weight (whole fruit, seed and flesh) were measured using a digital calliper (Mitutoyo 500–197-20, Kawasaki, Japan) and a balance (Mettler Toledo model AG204, Barcelona, Spain), respectively.

The mature avocado fruits were cut into thin slices (seeds and peel were removed) and frozen immediately at − 80 °C for 24 h, followed by the freeze-dried process using a lyophilizer (Christ Alpha 2–4; Braum Biotech Int., Melsungen, Germany) for 24 h at a pressure of 0.220 mbar. The samples were weighed before and after lyophilization to calculate the dry weight of the sample. The dried avocado samples were then ground (Moulinex grinder AR110830) to a fine powder, vacuum packed in plastic bags and stored at − 20 °C prior to extraction and analysis.

2.4.2 Fat content and fatty acid profile

The Soxhlet technique was used for fat content determinations according to AOAC 963.15 (AOAC 1973). The thimble containing the lyophilized mesocarp tissue (2 g) was placed in the Soxhlet equipment and was extracted with diethyl ether for 1.5 h. Results were calculated and expressed in percentage of dry weight.

The fatty acid methyl esters were determined using methylation according to Lipan et al. (2019) followed by the procedure of Tuberoso et al. (2007). After extraction, FAMEs were separated in a Shimadzu GC17 gas chromatography coupled with a flame ionization detector and CP Sill 88 capillary column. The carrier gas was He (at a constant pressure o 316 kPa, the initial flow rate was 1.2 mL min−1), and the detector gasses were H2 (30 mL min−1), air (350 mL min−1) and He (35 mL−1) as a make-up gas. The injector temperature was 250 °C and the detector had 260 °C. The preliminary temperature was 175 °C for 10 min, with a temperature gradient of 3 °C min−1 until 215 °C, and then was kept at 215 °C for 15 min. The injection volume was 0.6 uL and the split ratio was 1:10. The identification of methylated fatty acid (FAME) peaks was conducted by comparing the retention times of the standards (FAME Supelco MIX-37). Results were expressed quantitatively as gram per kilogram, using a methyl tridecanoate as an internal standard.

2.5 Statistical analysis

Analysis of variance (ANOVA) was used to characterize the data followed by Tukey’s multiple range tests in order to ascertain differences in physiological indicators, growth rates, fruit yield and quality parameters among the different irrigation strategies. XLSTAT Premium 2016 (Addinsoft, New York, USA) was used to perform statistically significant differences, with significant levels p < 0.05, 0.01 and 0.001. Finally, a correlation analysis among physiological parameters, fruit yield and irrigation water was calculated.

3 Results and discussion

3.1 Physiological response at the highest atmosphere water demand period

3.1.1 Midday leaf water potential

Imposing water deficit and measuring the level of water stress is a common technique to evaluate the water–plant status during the growing cycle, particularly throughout the period of maximum evaporative demand. The earliest exhibition of water stress signal on deficit irrigation treatments arrived from Ψleaf measurements, and no symptoms of leaf chlorosis were found in any of the SDI strategies. Table 1 shows the response of Ψleaf and gS throughout the growing cycle of avocado cv. Hass subjected to different SDI strategies in a Mediterranean subtropical environment. The significant variations in the dynamics of both physiological parameters were in accordance with the amount of water irrigation applied, especially to non-stressed (100% ETC) and severely water-stressed (33% ETC) trees. The water requirements were augmented considerably during flowering due to the formation of new structures, which leads to an increase in the area susceptible to water loss via transpiration. This pattern helps explain the physiological reaction to water deficit. Firstly, there is a clear indication, particularly for the period of the increasing trend of Ψleaf that occurred during the early fruit set development phase in spring, with considerably higher stress levels in the following fruit growth development period. Secondly, throughout the fruit growth stage, the data can be divided into two groups, from day of the year (DOY) 153 to 223 with higher readings of Ψleaf compared to the previous period in one, and from DOY 245 to 289 with a decreasing trend in the other. This pattern can be seen as an indication of a reduction in net water loss from the avocado leaves during the early period of fruit growth (Table 1).

Table 1 Average dynamics of the midday leaf water potential and stomatal conductance during the growing period. DOY, day of year; Ψleaf, leaf water potential; gs, stomatal conductance; SDI, sustained-deficit irrigation; ETC: crop evapotranspiration; SDI33, SDI at 33% ETC; SDI50, SDI at 50% ETC; SDI75, SDI at 75% ETC; C100, control at 100% ETC; †NS, not significant; **, significant at p < 0.05. Values followed by the same letter within the same column are not significantly different by Tukey’s least significant difference test
Full size table

During the growing season, the different evaporative demand reached values ranging from − 0.60 to − 1.06 MPa in control trees (100% ETC), which differed from the severely water-stressed trees that reached values ranging from − 0.98 to − 1.78 MPa (33% ETC). Despite the great variability of Ψleaf, the behaviour of the most severe SDI treatment contrasted markedly to that of fully irrigated trees. During the fruit growth stage, higher average daily values for Ψleaf were found at 16:00 h (between − 1.29 and − 1.86 MPa), when the evaporative demand reached its daily peak. According to Moreno et al. (2019) in the subtropical coast of Spain for fully irrigated (non-water stressed) avocado trees, the Ψleaf values were mostly over − 0.30 MPa except in July, in which values decreased significantly until − 0.74 MPa. By contrast, the Ψleaf values for water-stressed trees ranged between − 0.24 and − 1.48 MPa. It is well known that the water stress expands in the tree canopy at this period because of an excess of transpiration over uptake and conduction of water from the soil. Under the daily stress, the water needed by the tree can be partially supplied by the water stored in its tissues; therefore, severe water stress at an early stage of fruit set and fruit growth periods can result in strong competition for water between fruits and leaves. As stated by Borys (1986), under such circumstances, the leaves extract water from the fruits, which shrink severely, and if they do not regain their turgidity, abortion eventually takes place; this is especially true for the most demanding stages: flowering and fruit set, where a large amount of non-structural carbohydrates are mobilized from the leaves to the fruits, and the fruit growth period where a rapid use of carbohydrates for fatty acids formation occurs (Grant et al. 2011; Burdon et al. 2011).

The Ψleaf values in water-stressed trees with respect to control during flowering were significantly different from − 1.05 to − 0.75 MPa; this response was more marked in the following stages of fruit set (− 1.35 and − 0.91 MPa) and especially during fruit growth (− 1.78 and − 0.57 MPa) due to the evaporative demand; therefore, this increasing trend of water stress can trigger alterations throughout the growing season and ultimately affect the avocado productivity. In this sense, according to Whiley et al. (1988), during flowering, the evaporative surface increases by up to 90% due to the abundance of small flowers with a high evaporation rate, leading to an increase in the total tree transpiration rate (from 13 to 15%). Consequently, the water stress at this phenological stage may cause flower abortion, fruitlet abscission and early leaf drop, as stated by Whiley and Schaffer (1994). That is, under experimental conditions for deficit irrigation treatments, the water stress restricts the photoassimilate availability to support fruit set and disrupts the productive potential, which is consistent with Lahav et al. (2013), who reported that a limited water supply at flowering and fruit set might also lead to smaller fruit size and deteriorated fruit quality.

3.1.2 Leaf stomatal conductance

In relation to gS, this physiological indicator reacted dynamically to environmental conditions for the Ψleaf pattern. Similarly, the gS rate registered a decreasing trend in all water-stressed treatments during different days of the year. The gS showed a marked and significant decrease in water-stressed trees compared with the control trees (Table 1). Comparable significant differences for gS between non-stressed and water-stressed trees were first detected at flowering after the fruit set reached maximum rates at the fruit growth stage. The maximum gS rate at midday was fixed in control trees ranging from 129.4 to 155.2 mmol m−2 s−1, whereas the severe SDI treatment ranged from 115.9 to 135.0 mmol m−2 s−1, as water stress caused stomatal closure, this could be provoked by an increase of leaf temperature. In this line, during the fruit growth stage (DOY 260), the most severe water-stressed treatment (SDI33) displayed Ψleaf and gS values that dropped to − 1.78 MPa and 116.5 mmol m−2 s−1, respectively. In addition, at the flowering stage, the reduction in the gS rate (around 33%) in the severe water-stressed treatment with respect to the control trees would lead to a significant decline in net photosynthetic rates.

In general, the fully irrigated trees showed the least stomatal closure, whereas the deficit irrigated trees registered lower rates, although with different levels that were dependent on the water stress applied in each treatment. There was some evidence of a shift in this early closure tendency during flowering in the case of water-stressed trees; however, a peak in stomatal resistance occurred during the fruit growth stage at DOY from 216 to 275, which was prominent for the most severe water-stressed treatment (SDI33). During the latter period, there was a better acclimation to summer conditions manifested in the fully irrigated trees but not in the water-stressed treatments, resulting in an altered reaction of gS rates to environmental conditions.

An adequate water supply is crucial at the flowering stage because the transpiration rate of flowers is generally higher than that of leaves, and fruit growth is largely associated with water accumulation (Blanke and Lovatt 1993). In addition, flowering and fruit growth are key phenological stages during plant ontogeny, and both are expensive in energy terms. In this line, the present study corroborated that the avocado tree was sensitive to water deficit for the flowering and early fruiting period. Water demand increased with a rise in canopy surface area caused by flowering and when fruit grew swiftly during early fruit development. Consequently, our findings are somewhat expected and show that avocado can tolerate water stress to some degree despite their predominantly mesic adaptation, as was stated by Wolstenholme and Whiley (1999). In the present study, the water stress in the avocado tree induced stomatal closure with a presumably subsequent downregulation of the photosynthetic activity. Closing the stomata lessens the plant’s ability to interact with the surrounding environment, thus reducing photosynthetic enzyme substrate and lowering the net CO2 assimilation rate. It is well known that the stomatal aperture is governed by complex biochemical processes, including enzymes, leaf-carbohydrate concentrations, water and nutrient availability and climatic conditions (Paul and Pellny 2003; Hetherington and Woodward 2003). The findings of the present study showed higher gS rates in non-stressed control trees (fruited) than those under water-stressed trees (defruited), which agreed with previous studies regarding water stress and reduction in avocado leaf gS during the reproductive periods (Schaffer and Whiley 2002; Chartzoulakis et al. 2002).

Figure 3A presents the relationship between physiological indicators during the study period; being under experimental conditions was found to have an acceptable correlation (r = 0.624, p < 0.05) between Ψleaf and gS. In this experiment, the Ψleaf and gS values changed consistently with the continuous decrease in the amount of water irrigation for trees indicating that water was taken up by the roots and released through the stomata at different concordance rates with water stress level imposed. Under experimental conditions, the effect of continual water stress throughout the growing cycle in deficit treatments in physiological terms was less severe than during water stress periods imposed in summer; therefore, the avocado tree triggered a mechanism for adaptation to constant water deficiency conditions by lessening plant evapotranspiration.

In this context, a close relationship between a water stress indicator value and gS is a precondition for a plant-based water stress indicator. According to Jones (2004), gS was independent of Ψleaf in a few cases and, compared with Ψstem, was less correlated with gS in many cases, as was revealed by Naor (2006).

3.2 Water stress impact on productivity and tree development

3.2.1 Fruit yield

Over the three monitoring seasons, the average fruit yield from the water-stressed treatments of SDI33, SDI50 and SDI75 reached 31.6, 38.3 and 43.8 kg tree−1, respectively, which significantly decreased with respect to non-stressed control trees C100 with 47.5 kg tree−1 (Table 2). In terms of productivity per area in such terraced avocado plantations, the SDI33, SDI50, SDI75 and C100 yielded 7.9, 9.6, 11.0 and 11.9 t ha−1, respectively. Because of the alternate bearing habit of the avocado tree, the impact of water stress induced by deficit irrigation on fruit yield seems to be better determining with averaged data throughout the studied seasons; therefore, the reduction provoked by SDI33, SDI50 and SDI75 in relation to C100 was of 33, 19 and 8%, respectively. The decrease in yield only is surprising in light of the well-known sensitivity of the avocado to water supply. Similarly, Gil et al. (2011) reported the fruit yield for avocado cv. Hass irrigated with 0.65 and 0.77 ET0 of 13.4 and 15.4 t ha−1, respectively. Holzapfel et al. (2017) revealed the effect of water deficit using four applied water treatments based on 25, 50, 75 and 100% of ET0, yielding 14.5, 17.8, 22.1 and 23.0 t ha−1, respectively. In a study by Kiggundu et al. (2012), the reduction in the amount of water irrigation relative to ET0 over 4 years exhibited an improvement of water use efficiency by 93–87% without any adverse effects on fruit yield. Bayram et al. (2012), without irrigation restrictions, reported a mean average fruit yield of 76.9 kg tree−1 (15.4 t ha−1) in Turkey’s Mediterranean coastal region. Comparable, Moreno et al. (2019) claimed an average avocado fruit yield throughout a 6-year monitoring period under namely conventional irrigation (about 6503 m3 ha−1) of 10.34 ± 1.98 t ha−1, producing for on-crop and off-crop seasons of 16.5 and 8.7 t ha−1, respectively. Recently, Silber et al. (2019) reported high yields for no water-stressed trees from 25 to 31 t ha−1, contrasting with significantly lower yields for water-stressed trees between 16 and 21 t ha−1. Under the experimental conditions of the present study, subjecting the mild water stress based on 75% of the ETC can be considered a suitable alternative for water irrigation savings without considerably affecting the productivity of avocado fruit cv. Hass.

Table 2 Impact of deficit irrigation on avocado yield and tree growth during the monitoring period. SDI, sustained-deficit irrigation; ETC: crop evapotranspiration; SDI33, SDI at 33% ETC; SDI50, SDI at 50% ETC; SDI75, SDI at 75% ETC; C100, control at 100% ETC; †NS, not significant; **, significant at p < 0.05; TCSA, trunk cross-sectional area. Values followed by the same letter within the same column are not significantly different by Tukey’s least significant difference test (p < 0.05), ± standard deviation.
Full size table

The average avocado production per tree for each studied season is shown in Fig. 2A, denoting an alternate bearing pattern with off-season (2017/2018) followed by on-season (2018/2019) with yields that averaged 49% more than the previous season, and finally by off-season (2019/2020), which had lowered yields of 27% with respect to the former season (2018/2019).

Fig. 2
figure 2

Average yield of avocado cv. Hass (A) and cumulative yield throughout the 3 years monitored (B). SDI, sustained-deficit irrigation; ETC, crop evapotranspiration; SDI33, SDI at 33% ETC; SDI50, SDI at 50% ETC; SDI75, SDI at 75% ETC; C100, control at 100% ETC. Bars are standard deviations.

Full size image

It is well known that increased avocado flower and fruit abscission can happen due to several factors, including extreme temperature, nutritional deficiencies and congenital factors, because even under optimal conditions, the flower and fruit abscission is still excessive. In this study, the most fruit abscission occurred within a month after flowering, with high fruit abscission rates during spring and summer followed by gradually declining rates and negligible abscission through autumn as developing fruit approached maturity (data not shown). The highest demand for carbohydrates during rapid fruit growth and oil accumulation coincided with peaks of high temperatures and high evaporative demand under water deficit circumstances; therefore, this fact presumably provoked the high rates of abscised fruits. In this line, Silber et al. (2013) pointed out that leaf carbohydrates play an important role in the complex framework of gS in avocado trees, stating the significant role of fruit load on gas-exchange properties and water management, which is consistent with gS behaviour found in this experiment.

The yield was highly variable in control trees during 2018/2019, varying from 25.3 to 106.0 kg tree−1 and for SDI33 during 2019/2020 between 10.7 and 54.8 kg tree−1. The avocado production differences among water deficit treatments for each season were found, showing that water stress equally affected production, in the same way, each season. Thus, according to the findings, the yield from non-stressed trees was slightly less susceptible to alternate bearing in contrast with the yield of water-stressed trees that was considerably lowered during off-seasons.

Figure 2B displays a complete tally of yearly cumulative avocado yield during the 3-year monitoring period, amounting to 23.7, 28.7, 32.8 and 35.7 t ha−1 for SDI33, SDI50, SDI75 and C100, respectively, at the end of this study. These values can demonstrate the effects of the irrigation water deficit programme on avocado productivity by duration and reiteration of seasons under such conditions. The most promising strategy seems to be the SDI75, which had no statistically significant productivity losses (p < 0.05) respect to the control trees. The average WUE for SDI33, SDI50, SDI75 and C100 during the study period amounted to 2.90, 2.34, 1.87 and 1.51 kg m−3, with minimum values during the off-season (2017/2018); however, this increased trend for WUE is negatively related to fruit yield.

The relationship between fruit yield of the treatments and water consumption from irrigation gave a good fit to a binomial yield function for each of the three study seasons (Fig. 3B). When the three study years were taken into account, the fit to a binomial function was also good (\(y=- 0.0024{x}^{2}+0.5578x+16.05; {R}^{2}=0.997)\). These types of binomial functions that relate yield and irrigation water have also been established for avocado by Holzapfel et al. (2017) as well as for other fruit crops (Gattan et al. 2006; Durán et al. 2011).

Fig. 3
figure 3

Relationship between leaf water potential (Ψleaf) and stomatal conductance (gS) during the water stress cycle for the study period (y = 24.306 x + 158.130, R2 = 0.624) (A) and production functions for the monitoring period relating fruit yield and applied irrigation water (2017/2018, R2 = 0.989; 2018/2019, R2 = 0.994; 2019/2020, R2 = 0.997) (B). Each point represents the average of 16 studied trees at different levels of sustained-deficit irrigations and control trees.

Full size image

3.2.2 Tree growth

The tree height, canopy diameter and canopy volume were significantly lessened (p < 0.05) in trees subjected to water stress in relation to control (Table 2). The tree growth as a result of water stress imposed was affected particularly in the most severe water-stressed trees; this effect is in line with the amount of irrigation water applied in each treatment. It should also be considered that avocado trees cultivated on terraces are usually smaller than those on flat areas, mainly due to the available soil, the latter being grown in a much larger active soil layer than those on terraces.

The tree canopy volume was lowered by SDI33, SDI50 and SDI75 with respect to the control trees C100 of 47, 26 and 18%, respectively. The average canopy efficiency for SDI33, SDI50, SDI75 and C100 during the study period amounted to 0.24, 0.21, 0.22 and 0.19 kg m−3, respectively. The canopy efficiency followed an alternate pattern that reflected total yield, which was higher overall in the on-season than in the off-season. The well-irrigated control trees had low canopy efficiency relative to the water-stressed trees examined. This fact suggests that water stress as a rootstock effect has been shown to affect canopy efficiency in avocado (Kremer and Köhne 1995). By contrast, a higher value than those found in this study for control trees was determined by Bayram et al. (2012) of 0.98 kg m−3 in a Mediterranean environment. Similarly, the trunk cross-sectional area was altered by water stress, showing a maximum reduction of 166.8 cm2 for the most severe treatment concerning non-stressed trees. These findings confirm the evidence of the high influence of the amount of irrigation water on avocado tree development.

3.3 Avocado fruit quality response to water stress

3.3.1 Fruit size and weight

Table 3 shows the average avocado fruit size due to the implementation of SDI strategies, corroborating the overall impact of water stress on avocado trees as a whole. Concretely, the length and width were explicitly reduced under SDI33 with respect to control trees (79.4 vs. 97.3 and 54.9 vs. 66.2 mm, respectively) (p < 0.001), reaching intermediate values under SDI50 and SDI75 strategies. The fruit components pulp, peel and seed were affected by irrigation, showing a similar trend that was found for weight. The highest pulp:seed ratio was for SDI33 (4.9) and the lowest for C100 fruits (3.0); SDI50 was close to the latter ratio, with 3.7. The pulp loss for SDI75, SDI50 and SDI33 for non-stressed fruits amounted to 20, 13 and 32%, respectively; however, the proportion of pulp (68%) and peel (18%) in the severely stressed fruits (SDI33) increased compared to those found in control (C100) fruits (63 and 16%, respectively). Recently, Melgar et al. (2018) and Salazar et al. (2020) reported the importance of avocado fruit and its by-products, such as seeds and peels, as rich sources of nutrients and phytochemicals, although the by-products are rarely used and commonly discarded. Under such stress circumstances, the entire canopy growth was slowed down due to the water stress applied, provoking a reduced fruit carrying capacity, in which the water stress indicators denoted this impact throughout previous phenological stages. Particularly, the most severe deficit irrigation (SDI33) triggered the abscission of fruits during the early fruit set period (data not presented) that was not sufficient in ameliorating the growth rate in the remaining fruits.

Table 3 Average morphological parameters and fat content of avocado fruits. SDI, sustained-deficit irrigation; ETC: crop evapotranspiration; SDI33, SDI at 33% ETC; SDI50, SDI at 50% ETC; SDI75, SDI at 75% ETC; C100, control at 100% ETC; †NS, not significant at p < 0.05; ** and *** significant at p < 0.05 and 0.001, respectively. Values followed by the same letter within the same column are not significantly different by Tukey’s least significant difference test.
Full size table

The process of cell enlargement and division requires a water supply during the fruit growth period, and the lack of water inhibited fruit growth and development in most of the severe deficit treatments. That is, the fruit weight and its components were lowered in all water-stressed treatments for non-stressed control fruits; its impact was according to the water stress level applied. The fruit weight loss due to water stress under SDI33, SDI50 and SDI75 regarding fruits from control trees C100 was 43, 22 and 23%, respectively. This decrease is accounted for by the influence of the irrigation regime on fruit size; the fruits at SDI75 treatment were reasonably less affected by irrigation and achieved important water savings.

On the other hand, an increasing trend for dry matter and fat content of the most severe water-stressed fruits with respect to control fruits was found (Table 3). Concretely, the dry matter content for all treatments averaged about 36%, which is higher than those reported for avocado cv. Hass by Ferreyra et al. (2016) and Carvalho et al. (2015) of 26 and 27%, respectively. Similarly, Bayram et al. (2012), for non-water-stressed avocado fruits, determined the value of 28% for dry matter. The fat content obtained by extraction with organic solvent (diethyl ether) through the Soxhlet method ranged between 71 and 74% fat content with no significant differences among treatments.

In short, the avocado fruit enlargement requires the accumulation of water and hydrocarbons, and this rate depends on the balance of incoming and out-going fluxes, particularly the environmental conditions and agronomical practices crucial in this solute movement and accumulation in the fruit. The water-stressed trees presumably activated an adaptation mechanism in response to the persistent water stress conditions that involved the growth adjustment of the fruits in order to face the low water availability. This fact was corroborated with the growth rate of avocado fruit in the present study, which was according to water stress applied. This is in accordance with Lawlor and Cornic (2002), who stated that the water stress reduced fruit growth by slowing down the rate of cell division and expansion due to loss of turgor and increased synthesis of abscisic acid.

3.3.2 Fatty acid composition of fruit and health indices

Table 4 shows the fatty acid profile formed from C10 to C22 carbon chains in relation to the different deficit irrigation treatments. The oleic acid (18:1) was the predominant fatty acid, followed by palmitic (16:0), linoleic (18:2), palmitoleic (16:1) and cis-vaccenic (C18:1n7) averaging for all treatments with 107, 51, 37, 32 and 21 g kg−1, respectively. All these major compounds were significantly (p < 0.01) increased by deficit irrigation treatments except for the SDI50 strategy. The fruits from SDI75 and SDI33 improved their oleic acid content (118 and 131 g kg−1, respectively) compared to the control fruits (95.5 g kg−1). Similarly, palmitic acid content amounted to 55 and 59 g kg−1 for SDI75 and SDI33, respectively, whereas SDI50 fruits reached the lowest content (42 g kg−1). Water stress also significantly (p < 0.001) increased the third major linoleic acid (C18:2), which was higher in avocado fruits from SDI75 and SDI33 (44 and 43 g kg−1, respectively), and lower but almost similar values between SDI50 and C100 (30 and 33 g kg−1, respectively). On the other hand, the significant reduction (p < 0.01) of the cis-vaccenic acid (C18:1n7) by water stress was noteworthy, particularly under SDI50 with 19 g kg−1 compared to the other treatments (average of 22 g kg−1).

Table 4 Fatty acids profile in avocado fruits affected by deficit irrigation during the study period. SDI, sustained-deficit irrigation; ETC: crop evapotranspiration; SDI33, SDI at 33% ETC; SDI50, SDI at 50% ETC; SDI75, SDI at 75% ETC; C100, control at 100% ETC; †NS, not significant at p < 0.05; *, ** and *** significant at p < 0.05, 0.01 and 0.001, respectively. Values followed by the same letter within the same column are not significantly different (p < 0.05) by Tukey’s least significant difference test.
Full size table

In terms of relative concentration (% of total lipid), the main oleic acid (C18:1) in fruits from SDI33 reached the highest value of 41.1% (at an altitude of 170 m.a.s.l.), which is similar to those reported by Carvalho et al. (2015) in Colombia with 42–46% in Hass avocados grown in plantations at 1100 m.a.s.l.; at a higher altitude, the values increased (59%). In a study by Mpai and Sivakumar (2020) in South Africa, the oleic acid content amounted to 45.7% in the middle of the season, although at the end of the season, it increased to a maximum of 59.4%. In this context, many studies highlighted the impact of the orchard’s location on the oil content and composition (Lu et al. 2009; Donetti and Terry 2014; Ferreyra et al. 2016). Important differences in fatty acid composition and amounts were reported in avocado fruits from different producing areas (Landahl et al. 2009; Donetti and Terry 2014), suggesting the use of oleic acid as a biomarker explicitly to determine the avocado-producing country.

Cervantes et al. (2021) reported similar values of fatty acids to those found in our study for avocado cv. Hass at different ripening stages; the oleic acid content was higher after 8 days (140 g kg−1), reduced after 12 days (78 g kg−1) and increased again after 16 days of ripening (113 g kg−1). The linoleic acid content was maintained after 0–4 days of ripening (18–20 g kg−1), increased after 8 days (29 g kg−1) and started to decrease between 12 and 16 days (25 and 20 g kg−1, respectively).

Among the minor fatty acids, the eicosenoic and alpha-linolenic acids showed significant differences among irrigation treatments that were higher in SDI75 and SDI33 following the same pattern as the main fatty acids, and the rest of minor fatty acids (capric, lauric, myristic, myristoleic, arachidic, stearic, pentadecanoic and pentadecanoic). In this context, Costa (2001) reported similar values for myristic acid but higher for stearic and arachidic acids (7.0, 0.30 and 0.60 g kg−1, respectively). Additionally, these authors found other fatty acids such as octadecanoic (5.6 g kg−1), eicosenoic (2.1 g kg−1), decaheonic (1.3 g kg−1) and octadecadienoic (0.4 g kg−1).

The impact of the deficit irrigation on the contents of saturated fatty acids (SFA), monounsaturated fatty acids (MUFA), polyunsaturated fatty acids (PUFA), ratios and health indices are shown in Table 4. The SFA concentration was increased in fruits under SDI75 and SDI33 (60 and 66 g kg−1, respectively) compared to SDI50 (46 g kg−1) and C100 (51 g kg−1). On the other hand, there were also significant differences (p < 0.001) for the MUFA and PUFA contents; fruits from trees subjected to SDI50 resulted in the lowest content of MUFA with 131 g kg−1 compared to the remaining treatments (C100, SDI75 and SDI33 of 149, 177 and 195 g kg−1, respectively). Similarly, regarding PUFA, the lowest content (39 g kg−1) was found with SDI50 fruits; in contrast, SDI75 and SDI33 augmented their contents (56 and 59 g kg−1, respectively) with respect to the control fruits (44 g kg−1). Regarding the different ratios and health indices, no significant differences among irrigation treatments were found; however, we highlight the averaged values for atherogenic (AI) and thrombogenic indices (TI) (0.25 and 0.48, respectively) found for avocado fruits in this study when compared to other products such as olive oil (AI = 0.31; TI = 0.52), fresh meat (AI = 0.68; TI = 0.69) and almond milk powder (AI = 0.09; TI = 0.25) (Attia et al. 2017; Laudadio et al. 2015; Lipan et al. 2021). Finally, it is important to point out that both SDI75 and SDI35 (293 and 319 g kg−1, respectively) raised the total fatty acids compared to the control avocado fruits. No other studies were found within the scientific literature that addressed the deficit irrigation effect on avocado fatty acid profile; however, studies on other fruit crops such as almonds showed an increase in these fatty acids under water stress conditions. In this line, Gutiérrez et al. (2020) stated that under SDI conditions, a rise in both MUFAs and PUFAs was found. Similarly, Lipan et al. (2020) claimed an increase in PUFAs but a decrease in MUFAs by regulated deficit irrigation. According to Hwang et al. (1997), a greater oleic/linoleic ratio in avocado oil is better for human health, particularly for cardiovascular problems; this ratio was not affected by water stress, although there was a trend towards increasing in SDI33 fruits (3.02) compared to remaining irrigation treatments.

Figure 4 shows the effect of deficit irrigation on the omega 3 (α-linolenic and eicosatrienoic) and omega 6 (linoleic, gamma-linolenic, eicosadienoic and docosadienoic) PUFAs and omega 9 (oleic and erucic) MUFAs. As observed, SDI75 and SDI35 strategies increased important fatty acids such as those from series omega 3, 6 and 9. These fatty acids are considered precursors of anti-inflammatory lipid mediators (Richard et al. 2008), associated with the risk reduction in cardiovascular diseases because they preserve high-density lipoprotein levels and antioxidant properties (Villa et al. 2011).

Fig. 4
figure 4

Average omega 3, 6 and 9 fatty acids content in avocado fruit affected by deficit irrigation (p < 0.01). SDI, sustained-deficit irrigation; ETC, crop evapotranspiration; SDI33, SDI at 33% ETC; SDI50, SDI at 50% ETC; SDI75, SDI at 75% ETC; C100, control at 100% ETC. Bars are standard deviations. Values followed by the same letter within the same column are not significantly different by Tukey’s least significant difference test.

Full size image

According to European Food Safety Authority (EFSA), α-linolenic can be referred to as a “source of omega-3 fatty acids” on food labelling, but only when the product contains more than 0.3 g of α-linolenic acid per 100 g of product. In this case, the product label can also receive the health claim, “α-Linolenic contributes to the maintenance of normal blood cholesterol levels” and information about the beneficial effect obtained with a daily intake of 2 g of α-linolenic acid (EFSA 2012). In our study, we determined that strategies such as SDI75 lead to levels higher than this requirement (0.45 g per 100 g), whereas the SDI50 is just below the threshold (0.29 g per 100 g). On the other hand, all the fruits (non-stressed and water-stressed) measured in this study can receive the health claim related to the omega 6 linoleic fatty acid “linoleic acid contributes to the maintenance of normal blood cholesterol levels” and the additional information about that the beneficial effect is obtained with a daily intake of 10 g of linoleic acid because these avocados (3.7 g per 100 g) are above the limit required by EFSA (1.5 g per 100 g) in order to use this health claim. Finally, none of the present samples could use the health claim related to omega 9 oleic fatty acid because EFSA requires at least 70% of the product fat content to be unsaturated; in the present study, the unsaturated fat was 61% of the total fat content; however, omega 9 attracts significant interest due to its health-promoting properties (Dreher and Davenport 2013). In addition, Salgado et al. (2008) highlighted that the fatty acid composition of avocado oil is comparable to olive oil in terms of the food industry due to its physicochemical similarity and taste.

On the other hand, the increase in unsaturated fatty acids in water-stressed fruits gives an added value to the product because, besides their essential and healthy effect on humans, they are also precursors of C6 and C9 aromatic compounds, which increase the aroma profile of the avocados, and play an important role in the plant response to biotic and abiotic stresses (Ju et al. 2018). That is, PUFAs such as linoleic (omega 6) and linolenic acid (omega 3) fatty acids are easily oxidized to the corresponding fatty acid hydroperoxides (Schaller and Stintzi 2009). These compounds can be oxidized, producing phytoprostanes and phytofurans (non-enzymatically), which were reported to be significantly higher in fruits cultivated under water stress conditions (Lipan et al. 2020).

3.4 Implications for irrigation management in avocado orchards

In general, according to the findings of the present study, the SDI75 or SDI33 could be feasible strategies from the perspective of the lipid profile since both improved their contents compared to the control treatment; however, the SDI33, as was discussed before, has a significant disadvantage in lowering the productivity and fruit size compared to the SDI75; based on this fact, we highly recommend this irrigation strategy because it achieves a balance among the yield, quality and water savings. In addition, the avocado fatty acid profile is considered a biomarker to determine the growing areas because of adaptation to the environment and to the pre-harvest practices applied during the production process of this crop (Donetti and Terry 2014; Ferreyra et al. 2016; Tan et al. 2017), which in our case was the semiarid Mediterranean subtropical environment and the amount of irrigation water. To our knowledge, some works investigated the pre-harvest factors such as climate (temperature), soil, altitude and plantation management on the fatty acid profile in avocado, but not the effect of irrigation. The increasing water shortage has caused us to consider proper strategies for the sustainable use of irrigation water. Physiological responses and adaptation to such factors are of relevance to irrigated avocado tree performance in semiarid environments. An efficient water-saving irrigation system without significantly detrimentally affecting productivity needs to be adopted in the context of sustainable intensification of agricultural development.

An admissible fruit yield loss for the sake of boosting water savings is critical when considering adaptation and mitigation tools to meet both current and future water shortage situations. Furthermore, as it was proved in the present study, the fruit quality in terms of bioactive compounds was improved under deficit irrigation, and this fact is crucial to develop environmentally friendly market diversification—that is, to produce avocado fruits with medium sizes with improved levels of health-enhancing substances (e.g. oleic, linoleic, palmitic, palmatoleic acids). Thus, in a drought context, to grow crops sufficiently, avocado farmers will need to adapt less available irrigation water and benefit from water shortage. In this sense, Ramakrishna and Ravishankar (2011) and Llanes et al. (2018) stated the positive effects linked with a drought that promote the called secondary metabolism, enhancing the concentrations of compounds involved in quality traits, mainly taste and health characteristics.

4 Conclusions

Given the potential impact of climate change on water resource availability in semiarid regions, adaptive and sustainable water-saving strategies are crucial. This means many threats and challenges for all crops; therefore, foreseeing a reduction in yield in which water shortage is one of the most significant and negative factors. In the present study, water stress was expected to negatively affect vegetative growth, flowering intensity, fruit setting and productivity of avocado trees. The low yield of the avocado trees resulted from the water stress imposed and the series of abscission events that occurred throughout reproductive development. The impact of water deficit on fruit size and weight was clearly evident with the severe irrigation treatment (SDI33), which was most pronounced in the last season. In general, there was a significant difference in fruit average weight and fruit size distribution based on the different SDI treatments.

In consonance with the water stress levels imposed, the Ψleaf and gS decreased significantly compared to the control trees. Both monitored physiological indicators were shown to be negatively affected by the SDI strategies; however, the mild water-stressed deficit irrigation (SDI75) in the lower level was affected in productive and physiological terms. By considering the saved water and the possible economic return of the fruit production (yield and fruit size), the most suitable SDI strategy was 75% ETC, resulting in acceptable margins of yield loss (8–10% regarding the control trees).

On the other hand, the SDI75 or SDI33 could be feasible strategies from the lipid profile perspective since both improved their contents compared to the control treatment. Both these strategies raised essential fatty acids such as omega 3 and 6 as well as unsaturated fatty acids, which are all important from a dietary point of view; however, SDI33 has a significant disadvantage in lowering the productivity compared to SDI75; we believe the latter to be most suitable, achieving a sustainable balance among fruit yield, quality and water savings. In this sense, the avocado fruits from water-stressed trees may be better, particularly regarding the content of health-promoting phytochemicals.

In consequence, this work highlights the environmental, nutritional and economic benefits for farmers and health benefits for consumers by implementing deficit irrigation as a part of sustainable intensification practices. That is, avocado cultivation in a Mediterranean environment offers promising possibilities for providing environmentally friendly, high-quality fresh fruits.