The study was conducted from January 2015 to December 2016 in an experimental pequi orchard (Caryocar brasiliense C.) established in January 2009. Planting details can be obtained in Alves et al. (2015). 120 pequi seedlings were transplanted, in a 5 x 5 m spacing, into a 3000 m² experimental area at the School of Agriculture of the Federal University of Goiás (EA/UFG) (16°35’12"S, 49°21’14"W, altitude: 730 m), Goiânia, Goiás (GO) state. The soil of the area was classified as a dystrophic Red Latosol with a loam-clayey texture (Table 1). The climate of the region, according to the Köppen-Geiger classification, is Aw: hot and semi-humid with a well-defined dry season. The average annual air temperature is 22.9°C. The dry season is concentrated from May to September, and the rainy season lasts from October to April, with an annual average of 1520 mm of rainfalls.

Table 1

The experimental design was randomized blocks with 6 plots, analyzed by subdivided plots. The plots consisted of two irrigation levels (with and without irrigation) and the subplots consisted of two fertilization levels (with and without fertilization). Each block contained sixteen plants, eight of which were irrigated and eight were not irrigated. Each subplot was composed of eight plants: four plants grown with and four plants grown without fertilization. The treatments were defined as:

Irrigated and fertilized (IF): Application of irrigation using a water depth calculated to restore 100% of the evapotranspiration (ETc) during the whole dry season, and organic fertilization with a dose corresponding to the general recommendations made for other fruit plants cultivated in the region (Crisostomo and Naumov, 2009).

The irrigation treatments (IF and IWF) were irrigated by microsprinklers containing one microsprinkler per plant (working pressure (WP): 10 mca, flow rate (q) = 43 L h^-1, wet ray (r) = 2 m), weekly in 2015 and biweekly in 2016. Irrigation intervals were adopted seeking to deepen the volume of water in the soil applied by irrigation. Irrigation water management was carried out by atmospheric monitoring. The meteorological variables were obtained by the automatic agrometeorological station (HOBO) of the EA/UFG located 350 m from the experimental orchard. ET_o was estimated by the Penman-Monteith model (Allen et al., 1998), and the volume of water applied by irrigation (microsprinklers) was calculated by adopting a crop coefficient (K_c) of 0.9 because it is representative of aged fruit trees cultivated in a spacing similar to that of the pequi orchard (Doorembos and Kassam, 1977; Doorembos and Pruit, 1979). As only 50% of the useful area of each plant was wet, the ET_o was corrected for the calculation of water volume supplied per plant in each irrigation using the correction coefficient Kloc = 0.85 (Faci and Fereres, 1980). The irrigation criterion was no water restriction, replacing 100% of the maximum evapotranspiration estimated for the orchard.

To monitor the soil water content, two ECH₂O sensor batteries (model: EC-5, brand: Decagon) were installed in the canopy projection (0.50 m from the stem) of two pequi plants, one representative of the irrigation treatment and the other of the dry land. The sensors were calibrated according to the manufacturer’s recommendation (Decagon) and installed at the depths 0.50, 1.00, 1.50, 2.00 and 2.50 m connected to an Em50 Datalogger. The recordings of the readings were performed every one hour from Sep/2015 to Dec/2016.

In order to adjust the water retention curve in the soil, 15 undisturbed soil samples were collected at the same installation depths of the EC-5 sensor (0.50, 1.00, 1.50, 2.00 and 2.50 m). The samples were saturated, taken to a centrifuge and subjected to rotations corresponding to the tensions -6, -8, -10, -33, -60, -100 and -1500 kPa. Data were adjusted to the Van Genuchten (1980) model using the SWRC software (Dourado et al., 1990). The tensions of -10 and -1500 kPa were considered for water content at field capacity and permanent wilting point, respectively.

Fertilizations were carried out in November. In 2014, the fertilization consisted of the application of 2.5 kg of compost material and 1.0 kg of Yoorin; in 2015, 5.0 kg of cattle manure and 2.5 kg of chicken litter were applied (Tables 2 and 3). The applications were performed within a 0.50 m radius of the plant stem.

Table 2

Table 3

After 30 days of application of the organic fertilizer, in December 2015, an analysis of macro and micronutrients was conducted (Malavolta et al., 1997; Embrapa, 1997) to perform a nutritional evaluation of pequi trees.

Variations in stem perimeter were monitored monthly between Jan/2015 and Dec/2016. A tape measure was used to measure the stem perimeter at 0.10 m from the soil surface. Monthly, plant height was also measured using an electronic forest clinometer from the soil surface to the highest branch of the plant.

The canopy area was measured in Aug/2015. For this, all the trees of the orchard were individually photographed, placing some object with a known height next to the tree for comparison purposes. The photographs were processed in the software AutoCad 2015 for scale correction using the photographed object as a reference and canopy delimitation using the outline of the tree in the photograph. Then, the area of the object formed by the delimited canopy was obtained.

Growth data were submitted to analysis of variance. The treatments were compared by Snedecor F test and, when necessary, by Tukey test at a 5% level of significance.

As a sign of water stress, leaf temperature and sap flow were evaluated in pequi trees. Leaf temperature was monitored biweekly from Nov/2015 to Aug/2016, when the orchard aged 6.9-7.7 years, using a digital infrared laser thermometer (Fluke, model: 59 max+). The readings were always performed in the morning, between 08:00 and 10:00, by sampling three irrigated plants and three other plants in an unirrigated area. Leaf temperature values indicated by the thermometer were compared to the mean air temperature provided by the EA/UFG agrometeorological station at the time of the evaluation (Trentin et al., 2011).

Sap flow was monitored between Oct/2015 and May/2016 using Granier heat dissipation probes (HDPs) (1987). In Oct/2015, seven pairs of HDPs were installed: three in pequi plants in the irrigation treatment, three in unirrigated areas, and one unheated HDP to record the natural thermal difference in unirrigated plants. The measurements of thermal differences were performed every 30 seconds and the averages of the readings were stored every 15 minutes in a data acquisition storage system ("datalogger" CR 21X Campbell SCi.). The estimation of sap flow by HDP readings was performed using the equation proposed by Granier (1987).

For the dry period of 2015, the volumes of water replenished by irrigation were 84.50, 105.63, 126.75, 126.75 and 137.31 L d^-1 per plant from June to October, respectively. In 2016, 88.73, 105.63, 73.94, 95.06 and 95.06 L d^-1 of water per plant were replenished from May to October, respectively (Figure 1).

Figure 1

On average, the volume of water the orchard received by irrigation was 116.95 L d^-1 per plant in 2015 and 92.24 L d^-1 per plant in 2016, corresponding to a water depth of 9.36 and 7.38 mm, respectively. During this period, the contribution of rainfalls to the water volume received by the orchard was 0.55 mm (2015) and 1.48 mm (2016), while evapotranspiration (ETo) presented values of 5.67 and 5.22 mm d^-1, respectively. Thus, the balance between the water volume received by the orchard and evapotranspiration during the same period presented a negative balance for non-irrigated plants.

On the other hand, irrigations performed between 7 and 15 days were sufficient to keep only the 0.5 m surface layer wet, while deeper layers received only rain water (Figure 6).

At the depth of 0.50 m (in the canopy projection of the irrigated plant), the soil water content was replaced by irrigation, and the contribution of rainfalls ranged between θ_cc and the water content considered critical for the crop (θ_critical). Between April and May 2016, the water content was close to the θ_critical (f = 0.7). Between Aug/16 and Sep/16, a biweekly irrigation period, the same behavior was observed. However, the contribution of rainfalls in Oct/16 was sufficient to raise the soil water content above θ_critical. At the depth 1.00 m, irrespective of irrigation, the water volume remained below the θ_pmp. At 2.00 m, such volume ranged above and below the θ_pmp, while at 2.50 m, although not at θ_pmp, soil water content remained below θ_critical during the evaluation period.

In the non-irrigated area, at a 0.50 m depth, soil water content remained below the θ_pmp until Oct/15. Thereafter, it presented increasing values, remaining above the θ_critical, then decreasing again until Apr/16, when it remained below the θ_critical. This was due to rainfalls. With the beginning of rains in Oct/16, the soil water volume at the 0.50 m depth presented the same increasing tendency. At the 1.00 m depth, there was an analogous behavior. However, the contribution of rainfalls was sufficient to raise the soil water content little above θ_pmp between February and April and November and December 2016. At 1.50 and 2.50 m, the soil water content, although below θ_pmp in Sep/2015 with the onset of rains in October of that year, remained above the θ_critical between December 2015 and April 2016 at the depth 1.50 m and between February and April 2016 at the depth 2.50 m.

The organic fertilization provided a high K content in pequi leaves. Regarding other nutrients, no significant differences were observed (Figure 2). In fertilized plants, the K content was 78%. In plants without fertilization, it was 60%.

Figure 2

The results obtained showed that there was no significant interaction (P>≥0.5) between irrigation and fertilization management for the variables plant height and stem perimeter, but that there was a significant effect of irrigation on canopy area (P<0.5). At 24 months of evaluation, plant height ranged from 5.26 to 6.41 m (IF), 5.16 to 6.13 m (IWF), 5.16 to 6.04 m (WIF) and 5.25 to 5.98 m (WIWF). The stem perimeter varied from 52.88 to 58.74 cm, 47.17 to 56.82 cm, 48.06 to 55.47 cm and 47.78 to 56.11 cm in the treatments IF, IWF, WIF and WIWF, respectively (Figure 3).

Figure 3

There was a greater stem perimeter in irrigated and fertilized plants (IF) compared to other treatments during the first half of 2015 (Figure 3B). However, during the second half of the year, plants submitted to the other treatments were similar regarding stem perimeter. This is due to a higher availability of energy (average air temperature and radiation) during this period (Figure 3C), in addition to a greater availability of soil water due to the rainfalls beginning on September and lasting up to May (Figure 3). Thus, the plant did not respond to irrigation and organic fertilization in relation to height and stem perimeter. Previous studies also found no significant effects of fertilization and irrigation on plant growth (Alves et al., 2013, 2015; Miranda et al., 2016). An explanation for such results is that the pequi tree, as a native species from the Cerrado, has drought tolerance mechanisms, such as osmotic adjustment and deep root systems, which maintain its slow growth even during dry season and dry conditions (Sternberg et al., 2004; Palhares et al., 2010).

In relation to canopy area, values of 13.28, 10.84, 8.34 and 10.17 m² were obtained for the treatments IF, IWF, WIF and WIWF, respectively (Figure 4). Ferreira et al. (2015) reports that there is a positive and significant correlation between the number of fruits and the canopy projection area of pequi trees.

Figure 4

The authors also evaluated other dendometric characteristics, such as height and diameter, but the canopy area explained the best plant productivity. In this sense, the effects of irrigation on the growth of the canopy area of pequi trees (Figure 4) deserves continued studies, given that it may make the implementation and the productivity of commercial pequi orchards feasible.

Leaf temperature evaluations indicated that treatment plants without irrigation presented water deficit (Figure 5). During the rainy season (Nov/2015-Mar/2016), there was no evidence of water stress. However, from April, a transition period between the dry and the rainy seasons in the Cerrado of Goiás, there is stress in plants without irrigation. In that month, the leaf temperature was up to 5.90 °C higher than the average air temperature. Even in the irrigated treatment, the interruption of irrigation during the rainy season made the plant show a tendency to water stress (leaf temperature equivalent to average air temperature) during that same period, which did not occur due to the restart of irrigation in May.

Figure 5

Even the supply of water from rainfalls (Figure 6) was not sufficient to avoid water stress of non-irrigated plants during the first dry months (May-June 2016) (Figure 6). In June and July, the thermal gradient between average leaf and air temperatures was 1.20 and 1.50 °C, respectively. Under irrigation, the mean thermal gradient for the evaluation period was -2.00 °C.

Figure 6

Although the evaluation of leaf temperature as an indicative for water stress (Khatum et al., 2015; Brunini and Turco, 2016) is rather widespread, little is discussed about the value of the gradient in which physiological changes occur, especially because this is an information that depends directly on plant sensitivity to water deficit (Taiz and Zeiger, 2004). For the conditions of this study, it was possible to observe water stress in pequi trees in gradients of temperature above 1.20 °C. However, the adaptation of the plant to the edaphoclimatic conditions of the Cerrado of Goiás allowed its growth even under water stress.

During the rainy season (calendar days 277 to 121), the mean sap flow was 39.30 and 23.29 L d^-1 per plant for irrigated and non-irrigated pequi trees, respectively. With the beginning of the dry season, the flow was 64.26 L d^-1 per plant for irrigated plants, whereas this value was 29.06 L d^-1 per plant for non-irrigated plants (Figure 7). The occurrence of a higher sap flow during drought is due to a shorter evaluation period under this condition (19 days) associated to the transition period between the end of rainfalls and the beginning of the drought period, ensuring the availability of soil water.

Figure 7

The raw sap flow is associated with plant transpiration. In order to carry sap, there is a need for water stress in the xylem (the main water transport cell in vascular plants) established by transpiration (Pimentel, 2004). Considering the whole evaluation period (calendar days 277 to 142), the mean orchard transpiration was 42.79 and 24.09 L d^-1 per plant for irrigated plants and plants cultivated without irrigation, respectively. Thus, if irrigated, pequi plants transpire approximately 56% more than under natural conditions in the Cerrado biome.

The effects of soil water content, although at the superficial layer and with a higher content of leaf K due to fertilization, on leaf temperature and sap flow (Figures 6 and 7) of pequi trees suggest that, under irrigation, the pequi tree presents more absorbent roots on the soil surface, which provides the plant with a greater absorption of water and nutrients. Well-supplied potassium plants have a high water use efficiency, while potassium-deficient plants have a low photosynthetic performance due to an irregular stomatal opening, reducing the intake of CO₂ (Silva et al., 2016). In guava trees, the appropriate level of K ranges from 1.7 to 2.0 dag kg^-1 (Natale et al., 1994). Although guava was initially considered as rustic as and morphologically similar to pequi trees, they are possibly physiologically different plants with different nutritional requirements, since the fertilized and irrigated pequi tree tolerated water stress.