Study site

The study was carried out at the Experimental Campus of the Forest School of the Universidad Autónoma de Nuevo León (Autonomous University of Nuevo León), located 8 km south of Linares municipality, at 24°47' N and 99°32' W, and 350 masl.

The climate of the region is subtropical and semiarid, with a warm summer, rains from April to November, and an intraestival drought period. The mean monthly temperature ranges between 14.7 °C and 22.3 °C, with maximum temperatures of 45 °C during the summer. The mean annual precipitation is 805 mm, with a bimodal distribution (^{González et al., 2004}). The dominant soils are Vertisols with montmorillonite, with a depth of 0.80 to 1.50 m, a dark gray color, and a silty clayey texture; they contract and expand visibly in response to the changes in soil moisture content (^{López, 2014}).

Vegetation

The main vegetation type of the studied ecosystem corresponds to the Tamaulipan thorn scrub, a shrub/subarboreal formation with dominant floristic elements measuring 4 to 6 m in height, which are perennial, mostly thorny, with small, deciduous leaves. The most representative species are: Acacia amentacea DC., Acacia farnesiana (L.) Willd., Acacia schaffneri (S. Watson) F. J. Herm., Castela erecta Turpin, Celtis pallida Torr., Condalia hookeri M. C. Johnst., Cordia boissieri A. D C., Diospyros texana Scheele, Eysenhardtia texana Scheele, Forestiera angustifolia Torr., Havardia pallens (Benth.) Britton & Rose, Lantana macropoda Torr., Leucophyllum frutescens (Berland.) I. M. Johnst. and Zanthoxylum fagara (L.) Sarg. The floristic characterization of this type of community has been documented before, by ^{Domínguez et al. (2013)}.

Litterfall decomposition and collection frequency

The so-called litter bag method (^{Bocock and Gilbert, 1957}) was used to collect newly fallen dead leaves, branches and inflorescences in bags. Each 10 g bag reflected a relative proportion of the structural components of the original material: leaves, branches, reproductive structures and other components of litterfall. A total of 80 bags were placed at random, in winter and spring, in an undisturbed plot in the Tamaulipan thorn scrub. 10 samples were retained in the laboratory as controls. Subsequently, 10 bags were collected at random at 5, 10, 30, 45, 60, 75 and 90 days, after having been placed on the soil in each season. The samples were dried in a 292AD Felisa^TM convection oven at 70 °C during 72 hours. This value was used to calculate the loss of the baseline biomass equivalent to decay. The remaining mass values were adjusted to the simple exponential regression model (^{Olson, 1963}) as follows:

Where:

X _t = Litterfall mass at a given time (years)

X ₀ = Baseline litterfall mass

β ₀ = Value at time zero

e = Napier’s constant

k = Decomposition constant

t = Time expressed as days

Also, the theoretical time (years) required for 50 %, 95 % and 99 % of the litterfall to decay was determined according to the following formulas:

Once the residual dry weight of each sample was estimated, the litterfall was ground in a 3383 Thomas Willey mini mill with a No. 20 (0.85 mm²) mesh. The ground litterfall was subsequently stored in Ziploc^TM type bags labeled by date, repetition and sampling season for purposes of chemical analysis.

Nutrients content

By way of a sample, 1.0 g were used to determine the concentration of minerals (Ca, K, P, Mg, Cu, Fe, Mn and Zn); they were incinerated in a F-48010 Thermo Scientific muffle furnace at 550 °C, during 4 h; the ashes were digested in a hydrochloric acid solution (HCl) and nitric acid (HNO₃), by humid digestion (^{Cherney, 2000}). The Ca, K, Mg, Cu, Fe, Mn and Zn contents were analyzed using atomic absorption spectrophotometry, with a PinAAcle 900F Perkin-Elmer ^TM spectrophotometer. The calcium concentrations were quantified with nitrous oxide-acetylene flame, and potassium, magnesium, copper, manganese, iron and zinc, with air-acetylene flame. Phosphorus (P) was quantified by colorimetry, using a Lambda 25 Perkin-Elmer ^TM spectrometer.

Statistical analyses

When the data corresponding to each season and to each nutrient were tested, they were shown not to meet the normality assumptions (Kolmogorov-Smirnov and Shapiro Wilk tests) or the variance homogeneity assumptions (Levene’s test); therefore, they were subjected to the Kruskal-Wallis nonparametric test (^{Ott, 1993}) in order to detect differences between the seasons for each of the analyzed nutrients. The statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS), version 22 for Windows (^{SPSS, 2013}).

Environmental variables

The temperature (°C) and relative moisture (%) at the study site were registered (on an hourly basis) using automated HOBO sensors (H8 Family, Forestry Suppliers, Inc.) Precipitation (mm) was quantified at the site with a HOBO^TM automated pluviometer. Table 1 shows the mean monthly temperature and relative moisture and the monthly precipitation registered during the period at the study site.

Decomposition of the litterfall

The litterfall decomposition rate is shown in Table 2, which illustrates the value of the decomposition constant (k, gg^-1 year^-1) for winter (0.002) and spring (0.003); this is a low decomposition rate. A study carried out in the Tamaulipan thorn scrub estimated the k values at an interval of 0.42 and 0.50 (^{Marmolejo et al., 2013}). On the other hand, the k values for a different type of vegetation -Coffea arabica L., Pinus oocarpa Schiede ex Schltdl. and Eucalyptus grandis Hill ex Maiden plantations- were 0.87, 1.08 and 0.72, respectively (^{Farfán and Urrego, 2007}). ^{Rivera et al. (2013)} cite a value of k=0.00913 in fallowed areas aged over 20 years, and a time of 75.9 days for the decomposition of 50 % of the matter, and 504.3 days for 99 %. In the present study, the litterfall decomposition was shown to require 0.8 years in winter, and 0.6 years in spring. The time required for the decomposition of 99 % ranges between 5.2 years in winter and 3.7 years in spring, which means that the litterfall will take longer to decay in winter.

Table 3 shows the results of the Kruskal-Wallis test for detecting significant differences between the macro- and micro-nutrient contents of the litterfall in the two seasons of the year (winter and spring) during the decomposition process, at different incubation times. Macronutrients like Ca and Mg showed significant differences at seven of the nine considered incubation times; the element P exhibited significant differences only at one incubation time. As for micronutrients, Mn had significant differences at six times; micronutrient Zn exhibited differences only at two times.

The results indicate that a greater retention, and therefore a lesser release, of nutrients occurred in the spring, when the precipitation was more abundant. This agrees with the data cited by ^{Tripathi and Singh (1992)}, who suggested that most nutrients become immobilized during the rainy season.

Macronutrients release

In winter and spring, both macro- (Ca>P) and micronutrients (Fe>Mn>Zn>) tend to exhibit immobilization at the beginning, and are subsequently accumulated during the incubation time, as shown in figures 1a, 1c, 2a, 2c and 2d. The nutrients K>Mg were released in both seasons. The order of release registered in the present study differs from the data cited by ^{Kuruvilla et al. (2016)} in Munrochloa ritchei plants, for which the release pattern was Mg>K. Cu was the only element released in winter, while the release pattern of Ca was greater than the P in spring. For an incubation period ranging between 0 and 90 days, the accumulation of Ca amounted from 45.7 to 80.5 mg g^-1 in winter, and 76.9 to 87.8 mg g^-1 in the spring. There was no release in either season; however, an accumulation of 34.7 mg g^-1 occurred in the winter, and of 10.9 mg g^-1 in the spring (Figure 1a), and there was an increase of 76 % and 14 %, respectively, in relation to the baseline value. These results are contrary to those observed by ^{Farfán and Urrego (2007)} in Coffea arabica L. and Cordia alliodora (Ruiz & Pav.) Oken plantations, which exhibited a release of 34 % and 54 %, respectively, during an experimental period of 365 days. Thus, the incubation time, the environmental conditions and the quality of the litterfall were limiting factors for Ca release.

Figure 1 Litterfall content of Ca (a), K (b), P (c) and Mg (d) in the Tamaulipan thorn scrub during winter and spring at different incubation times during the decomposition process. 

Contenido de Cu, Fe, Mn, Zn = Content of Cu, Fe, Mn, Zn; Tiempo de incubación (días) = Incubation time (days); Invierno = Winter; Primavera = Spring.

Each charted value represents the mean (n = 10) ± standard error.

Figure 2 Litterfall Cu (a), Fe (b), Mn (c) and Zn (d) content in the Tamaulipan thorn scrub at different incubation times during the decomposition process. 

As for K, this nutrient exhibited a fluctuation of 11.4 to 4.2 mg g^-1 in winter, and 10.6 to 4.3 mg g^-1 in the spring. Therefore, according to Figure 1b, 63 % (7.2 mg g^-1) of this element was released in the winter, and 59 % (6.3 mg g^-1), in the spring. These observations suggest that K is a rapidly released element, as previously documented by ^{Arce (2007)} and by ^{Castellanos and León (2011)} for Ormosia coccinea (Aubl.) Jacks. and Vochysia lomatophylla Standl. plantations. This mobility of K has been cited; thus, we infer that it is not linked to the organic structure of leaf tissues but is freely located and therefore is easily washed or removed (^{Attiwill, 1968}; ^{Tukey, 1970}; ^{Parker, 1983}).

The concentration of P ranged between 0.3 and 0.6 mg g^-1 in winter and between 0.6 and 0.7mg g^-1 in spring. In either season, little mobility of this element was observed during the litterfall decomposition process, as the values remained relatively stable from day 5 to day 60 of incubation, when it began to accumulate, up to day 90. At the end of the study, 91 % P was accumulated in winter, and 15 % in spring, with respect to the baseline amount (Figure 1c). These results agree with those cited for Eucalyptus grandis litterfall, which exhibited P immobilization (^{Farfán and Urrego, 2007}). A similar tendency was observed by ^{Castellanos and León (2011)} in Acacia mangium Willd. litterfall; these authors point out P as the most restrictive element for the decomposition and mineralization of organic matter, with a tendency to immobilization, due to its scarce availability in the soil. According to ^{Will (1967)}, during the litterfall decomposition process, P is retained by microorganisms or by organic compounds.

Mg exhibited a fluctuation of 3.9 to 3.6 mg g^-1 in winter and of 3.3 to 2.9 mg g^-1 in the spring. The release of this element was slow in both seasons, amounting to 0.34 mg g^-1, equivalent to 9 % of the baseline quantity, in winter, and to 0.39 mg g^-1, or 12 % of the baseline quantity, in the spring (Figure 1d). These results indicate a low release of Mg compared to that of K. ^{Moro and Domingo (1996)} documented a similar pattern for Adenocarpus decorticans Boiss. litterfall in the Filabres mountain range in Spain, which exhibited a release of 0.96 % of Mg and 1.69 % of K. However, ^{Kuruvilla et al. (2016)} observed a greater release of Mg than of K in Munrochloa ritchei.

Micronutrient release

The Cu content ranged between 5.4 to 4.3 μg g^-1 in winter and 3.5 to 6.3 μg g^-1 in the spring. In winter, Cu increased to 0.8 μg g^-1 at 30 days; subsequently, at the end of the season, 1.0 μg g^-1 was released. Conversely, in the spring the accumulation was 2.7 μg g^-1 at 90 days of incubation. Therefore, release was only around 7 %, only in the winter (Figure 2a).

The Fe element ranged between 278.4 and 681.3 µg g^-1 in winter, and between 299.5 and 745.9 µg g^-1 in the spring. In either season, there was no release of Fe during the experimental period. Figure 2b illustrates how, at the end of the incubation, Fe had a similar average increase (424.55 µg g^-1) in winter and in the spring.

The content of Mn was 26.6 to 44.5 µg g^-1 in winter, and 33.0 to 39.3 µg g^-1 in the spring. Figure 2c shows a similar average increase in Mn, of 12.1 µg g^-1, at the end of the incubation period.

No significant variation was registered in winter in the Zn content, which remained virtually constant (25.2±5.22 μg g^-1) at 0, 30, 45 and 90 days, while it varied between 14.2 and 30.3 µg g^-1 in the spring. There was no release of this element during the incubation period. Figure 2d shows that Zn remained relatively stable in winter; however, in spring, at the end of the incubation period, it increased by 13 % (6.1 µg g^-1) with respect to the initial time.