Description of the study area

The study was conducted on 40-year-old pecan nut trees, Western variety, from July to October 2016 and 2017. The trees are located in the municipality of Viesca, Coahuila, Mexico, with geographical coordinates of 25° 20’ 28’’ N and 102° 10’ and 48º 16’ 16' W, and altitude of 1 100 m. The climate is BWhw (e) type, which is interpreted as very arid, semi-warm with rainfall in summer and extreme thermal amplitude with an average annual temperature between 18 and 22 °C. Average annual precipitation ranging between 200 and 300 mm, with rainfall from April to November and scarce rainfall during the rest of the year (^{Servicio Meteorológico Nacional [SMN], 2010}). Soil has a clay loam texture, with a field capacity of 0.35 m³∙m^-3 (water content weight values, calculated at water/soil volume) and permanent wilting point of 0.19 m³∙m^-3. Spacing between trees is 12 m x 12 m and trees are planted under a true frame design. The irrigation system is drip irrigation with two lines on each side of the tree placed 2 and 3 m from the trunk at a 40 cm depth.

Experimental units and treatments

Representative trees were selected from the orchard with an average trunk circumference of 90 cm, measured at 60 cm from the ground, with an approximate height of 15 m and a leaf density of 3.5 m²∙m^-3. Each tree was irrigated with 40 emitters with an expense of 3.2 L∙h^-1. Irrigations of 3.5 h per day started in March, during the bud break phenological stage. Irrigation time increased as the season progressed, up to 6 h per day. In the phenological stage of kernel filling, irrigation time was 4, 6 and 8 h in three irrigation sections, respectively to establish moisture levels (treatments) of 0.257, 0.327 and 0.380 m³ of water per m³ of soil. The sampling units (three trees) were selected in the middle part of each irrigation section, which included 50 trees respectively, regarding the criteria suggested by ^{Prodan (1968}) for sampling tree species. The soil moisture level was recorded based on volumetric water content, expressed in cubic meters of water per cubic meter of soil, at a depth of 80 cm, where each moisture sensor was placed. Sensors were type capacitive (EC-5 DECAGON DEVICES Inc., Pullman WA, USA) duly calibrated, which measure the dielectric constant of soil, attached to a datalogger (Em50 DECAGON DEVICES Inc., Pullman WA, USA) to record the readings, which were subsequently correlated with moisture content. The recording was done daily, and data were stored in spreadsheet format file, from July to October 2016 and 2017.

Response variables and statistical analysis

Variables evaluated were nut production (kg) per tree at the time of harvest, which was carried out with mechanical vibrator on September 11 and 16, 2016 and 2017, respectively; fruit size, expressed in length and diameter, measured with vernier in a sample of 50 nuts per tree; percentage of germinated nut and commercial nut derived from the quotient of the number of nuts for each condition (n) and the total of collected nuts (N), multiplied by 100; kernel percentage from a sample of 20 nuts per tree, which were broken with a nutcracker and the kernel was separated from the shell. The weight (g) of the kernel was divided by the total weight of the sample and multiplied by 100. The experimental design was completely randomized in repeated measures over time; the study factor corresponded to the available soil moisture content (0.380, 0.327 and 0.257 m³∙m^-3). The experimental unit was an individual tree. The normal distribution of the data was analyzed using the Shapiro-Wilk test; then, an ANOVA test was carried out to determine the significance of treatments and Tukey's test (P ≤ 0.05) for mean difference. For this purpose, SAS version 9.0 (^{SAS Institute, 2002}) and Microsoft Excel version 2013 were used.

Soil moisture dynamics

Figure 1 shows the behavior of soil moisture under three volumetric water contents, at 80 cm soil depth from July to October 2016. The soil irrigated with 0.380 m³∙m^-3 had higher water moisture than at field capacity (FC), starting from the phenological stage of husk hardening occurring in the last week of July (Figure 1). Moisture content in this treatment showed no saturation values during the season; however, at the end of September higher water content was observed. The area irrigated with 0.327 m³∙m^-3, which corresponded to the irrigation applied by the producer, kept available moisture close to field capacity during July and August; starting in September, it decreased as the (0.257 m³∙m^-3) had the lowest values during the season and decreased from September onwards, as in the treatment of 0.327 m³∙m^-3.

Figure 1 Soil moisture dynamics under three volumetric water contents at a depth of 80 cm in pecan nut (Carya illinoinensis) during the production cycle of 2016. FC: field capacity, PWP: permanent wilting point. 

In 2017, moisture curves similar to those of 2016 were observed. Accumulated precipitation in 2016 was 352 mm; the amount of water recorded was 114 mm from January to June and 238 mm from July to October. In 2017, the amount of accumulated water was 200 mm, much lower than in 2016, with a record of 26 mm from January to June and 131 mm from July to September, while the rest, equivalent to 43 mm, was recorded in December.

Pecan nut yield

According to Table 1, treatments with higher moisture increased nut yield per tree (P ≤ 0.05). This result was mainly due to the increase in nut size and higher kernel percentage, but not due to the number of nuts. In 2016, considered a productive year, the increase was 12.7 % when soil moisture level increased from 0.257 to 0.327 m³∙m^-3, and 24.1 % when it increased to 0.380 m³∙m^-3. A similar response was reported in 2017, in which lower production was observed due to alternation that is common in this species. Production per tree increased 12.7 % when available soil moisture increased from 0.257 to 0.327 m³∙m^-3, and 23.6 % in trees with the highest available moisture.

Premature germination of pecan nut

In the two production cycles (2016 and 2017), prematurely germinated nut production in trees (Figure 2) increased significantly (P < 0.05) under moisture contents of 0.327 m³∙m^-3 and 0.380 m³∙m^-3. According to Table 2, in 2016, trees with higher soil moisture had 12 % more germinated nut production than trees with the lowest level. In 2017, from the shell hardening stage (late July) to the end of the production cycle, the highest soil moisture level caused significant increase (P < 0.05) of 10.5 % of germinated nut. The increase in vivipary negatively affected the percentage of commercial nut (Table 2); therefore, the soil moisture level of 0.257 m³∙m^-3 proved to be the most effective in decreasing vivipary without affecting the percentage of commercial nut.

Figure 2 Germinated pecan nut inside Carya illinoinensis husk, showing radicle during ripening period (A) and harvested nut (B). 

Fruit size

In the production cycle of 2016, fruit length and diameter increased significantly (P < 0.05) in trees with the two higher moisture levels (Table 3). In this cycle, the year of highest production, nut length and diameter in trees with higher moisture content (0.38 m³∙m^-3) were 30 % and 12 % greater than in trees subjected to lower moisture levels. A similar effect was observed in 2017; the treatment with the lowest soil moisture content (0.257 m³∙m^-3) caused a 1.0 cm (20.6 %) and 0.23 cm (8 %) reduction in fruit length and diameter, respectively, in relation to the high moisture level. From fruit hardening, final length and diameter at harvest were reduced in trees with lower water supply.

Kernel percentage

Table 4 indicates that, in 2016, kernel percentage increased significantly (P < 0.05) in treatments with higher soil moisture; kernel percentages were 1.95 and 0.62 % higher than in pecan nut0 trees subjected to lower moisture levels. In 2017, moisture treatments had a similar effect on this variable; soil moisture level of 0.257 m³∙m^-3 had the lowest kernel percentage, which was exceeded by 0.58 and 1.82 percentage points because of trees with moisture levels of 0.327 and 0.380 m³∙m^-3, respectively.

Pecan nut yield

Adequate environmental conditions and proper management during pecan nut production cycle, particularly in each phenological phase, lead to good production (^{Reyes Vázquez & Morales Landa, 2019}). Furthermore, available soil moisture (^{Grauke, Wood, & Harris, 2016}; ^{Thompson, 2005}), related to the supply of fertilizations, is important for the tree to manifest its productive potential (^{Walworth, White, & Comeau, 2017}). Therefore, it is essential to avoid water stress conditions, during the period of nut growth and kernel filling stage (^{Godoy & López, 2000}). The results of the two years of evaluation indicate that the higher soil moisture level, from fruit shell hardening occurred at the end of July, had a positive effect on nut production per tree. This exceeded the average nut yield (1.1 t∙ha^-1) in the Comarca Lagunera, reported by ^{Santamaría, Medina Morales, Rivera-González, and Faz Contreras (2002}), especially in the two treatments with higher moisture content. The increase in nut yield with increasing soil moisture has also been observed by ^{Marco et al. (2021}) and supports that established by ^{Sparks (2005}) regarding the high-water requirement of the species (Walls, 2017). It is concluded that high water requirement was inherited by its ancestors, which developed in river margins and deep and fertile soils (^{Babuin, Echeverría, Menedez, & Maiale 2016}; ^{Poletto, Poletto, Morales, Briao, Silveira, & Richards 2020}). In this regard, ^{Orona Castillo, Sangerman-Jarquín, Cervantes Vázquez, Espinoza Arellano, and Núñez Moreno (2019}) determined that pecan nut yields per unit area, in pecan nut orchards in northern Mexico, are directly related to irrigation availability and technical attention. Regarding the above, it is possible to explain the low production in trees that had less water volume in the soil, reflected in the lower number of smaller fruits produced, as observed in other species (^{Wiegand & Swanson, 1982}).

Premature germination of pecan nut

Several studies indicate there are a number of plants that exhibit vivipary (^{Duermeyer et al., 2018}); however, there are few reports concerning this phenomenon in commercial fruiting species (^{Taylor, Kunene, & Pandor, 2020}). ^{Farnsworth (2000}) suggests that vivipary is more common in plants with a particular morphological profile, including large seeds observed only in one fruit; seeds with hard endocarp surrounded by fleshy tissue that keep internal water content; species with directed dispersal (a strategy that allows them to reach specific habitats favorable for survival); species with long adult life spans, as well as, those that are less specialized for microhabitats; and those adapted to moist habitats. According to ^{Wood (2015}), vivipary is also favored by high fruit moisture, together with the presence of temperatures of 30 to 35 °C during the final stage of development and low concentration of abscisic acid. Since pecan nut has most of these characteristics, the probability of premature germination, even remaining on the tree, is high, when temperature and moisture conditions are favorable (^{Gonçalves Bilharva et al., 2018}; Wood, 2015). The study region has an average maximum temperature of 33 °C and precipitation of 131 mm from July to September, conditions that can be conducive the occurrence of this phenomenon (^{Sifuentes-Ibarra et al., 2015}).

Vivipary is considered a plant adaptation to take advantage of certain conditions to increase survival in the wild; in cultivated pecan nut trees, it is a cause of economic losses (^{Wood, 2015}). The germination process uses stored carbohydrates for growth, and the process causes blackening of the region of the embryo that joins the two kernel sections; nuts with this characteristic are not marketable (Wood, 2015). ^{Taylor et al. (2020}) add that, in addition to the aforementioned conditions, tree vigor, light condition, and overfertilization in the late stages of nut development could be the factors leading to vivipary in C. illinoinensis.

In this study, the highest number of germinated nuts was observed in trees with higher soil moisture content. The trend toward greater vivipary as soil moisture levels increased confirms the results obtained by Godoy and López (2000), ^{León (2014}), and ^{Wood (2015}). Prior to ripening, nuts are not physiologically in dormancy, but are mechanically restricted to germinate (Wood, 2015). In this constraint, the nut shell, which appears to be a formidable barrier (^{Wells, 2017}), provides no impediment to the transfer of water or gases required for germination. As a result, the high amount of moisture between nut shell and husk, related to high temperatures, could consequently be a factor that neutralizes the mechanical effect of the shell, allowing the nut to germinate prematurely, even when it still on the tree, as happens in some susceptible varieties under suitable climatic conditions (^{Aguilar et al., 2015}). In 2016, the maximum, minimum, and mean temperatures in September were 32.7, 19.3, and 26 °C, respectively, and in October were 20.2, 16.2, and 23.3 °C. In 2017, the maximum, minimum and average temperatures in September were 32.4, 19.0 and 25.7 °C and in October were 30.2, 16.9 and 23.6 °C. The presence of temperatures above 30 °C during the nut ripening period, associated with high humidity, could have been an important factor on the induction of premature nut germination on the tree (^{García-Moreno et al., 2020}; ^{Nonogaki, Barrero, & Li, 2018}).

On the other hand, the number of germinated nuts increases in years of high production (^{Sparks, 2005}; ^{Stein, 1985}). In this study, the trend towards a higher presence of germinated nuts in 2016 (year of high production) was not evident in 2017, a year with lower production and vivipary.

Nut size and kernel percentage

Nut growth and kernel filling stage depend mainly on moisture content; therefore, its adequate availability in the soil is important, both in the first phase of development, which is characterized by nut growth, and during the second phase, which corresponds to the kernel filling stage (^{Ferreyra, Selles, & Lemus, 2002}; ^{Godoy & Huitrón, 1998}; ^{Godoy & López, 2000}). Moisture stress during fruit growth leads to the production of small nuts (^{Godoy & Huitrón, 1998}; ^{Herrera, 1990}). On the contrary, a water deficit during kernel filling stage induces a low kernel percentage (Herrera, 1990; ^{Godoy & Huitrón, 1998}). In the present study, nut size increased as moisture availability increased, without affecting kernel percentage, whose tendency to increase was greater in trees with better moisture conditions. Values for nut growth are similar to those reported by ^{Marco et al. (2021}), agreeing with the fact that the amount of water applied during the kernel filling stage influences fruit width and length, with significantly larger nuts as irrigation increases.

According to the USDA (^{United States Department of Agriculture, 2020}), the average nut length is 36.8 mm and 25.07 mm wide. In this study, nut length and width values were greater than those referred to and increased in relation to water availability. Thus, during the kernel filling stage, when embryo development occurs, the size of walnut response to increased irrigation water is evident; however, vivipary increases and the percentage of commercial pecan nut decreases. The positive response toward larger size and higher kernel percentage in higher moisture treatments is not consistent with the assumption that pecan nuts fill better or produce better developed kernels (^{Thompson, 2005}). In this study, kernel percentage ranged from 56 % to 58 % in the lower and higher moisture trees during the two production cycles. These records exceed the reference values for export (55 %) (^{Godoy & López, 2000}). The kernel percentage is used in the pecan nut industry to determine the price per unit weight of pecan nuts in shell (Thompson, 2005). Also, it has been noted that high yields affect kernel percentage (^{Smith, 2012}; ^{Smith, Reid, Carroll, & Cheary, 1993}). In this study, yield per tree, specifically in 2016, had no effect on kernel size and kernel percentage, which increased within the moisture levels evaluated.