Study area

The study area is located within the Costa de Jalisco Experimental Site, belonging to INIFAP, located between the municipalities of La Huerta and Casimiro Castillo in the State of Jalisco, between coordinates 19° 31’ 15” north latitude and 104° 32’ 00” west longitude, at an average altitude of 298 masl. The climate is AW2 (w) warm sub-humid type with rains in summer and an average annual precipitation of 1 100 mm., according to the Köppen classification modified by ^{García (2004}).

Data collection

The field information was collected in 30 plots from a plantation with four forest species: Tabebuia rosea (Bertol.) DC. (purple rose), Tectona grandis L.f. (teak), Gmelina arborea Roxb. ex Sm. (melina) and Enterolobium cyclocarpum (Jacq.) Griseb. (parota), made in real frame with a distance of 4 x 4 m of separation between trees, which was established in 1994.

The sampling was carried out completely at random in two management scenarios. On the one hand, the plots with management (removal of herbaceous plants and shrubs), where three conditions were evaluated with the presence of the following species: purple rose, teak, melina, establishing six sampling sites for each plot. On the other hand, the plots without management where there is the presence of herbaceous plants and shrubs, from which biomass data were obtained from both components (herbaceous and shrubs) as one, in this plot 6 sampling sites were also established for each plot with presence of purple rose and parota.

For the evaluation of the carbon content, a modification of the methodology proposed by ^{Marquez (1997}) was used to estimate carbon and evaluate the variables of main interest, in this case two components: aerial (bush, herbaceous, leaf litter) and underground (soil at two depths).

Carbon quantification in shrubs and herbaceous plants

To determine the carbon content, six random sampling points were established for each plot, using 1 x 1 m (1 m²) square sites as the sampling unit. All the plant material belonging to the bushes and herbs within the quadrant was obtained by cutting all the existing biomass at ground level. Of the plants that were partially outside the sampling site, only the branches or stems that entered the quadrant were considered (^{Acosta et al., 2009}).

Leaf litter carbon quantification

Sampling sites established for grasses and shrubs were also used to measure litter amount; however, in this case the quadrants were 0.25 m² (50 x 50 cm squares). Biomass belonging only to leaf litter or mulch was collected in that area. Once all the grass and shrub material had been collected, and litter was placed in a separate blanket bag to be sent to the INIFAP soil laboratory at the Tecoman Experimental Field, Colima. Where they were dried in an oven at 75 ºC for 48 h to obtain the dry weight. To convert the biomass data to carbon, the total biomass per hectare obtained by the constant of 0.5 was multiplied, since, in most species, half of the biomass is carbon in any of its parts (^{Masera et al., 2000}).

Soil carbon quantification

The soil samples were obtained at the same sampling point where the previous components were sampled. To do this, it was performed with an 8 cm diameter auger (AMS - Soil Sampling Equipment) and two soil samples were obtained for each depth level (0-15 and 15-30 cm). In this way, twelve soil samples were obtained for each plot at the previously established sampling sites. Knowing the diameter of the cylinder used and its length, the volume of soil sampled was determined, then when analyzing the soil and determining the percentage of carbon, inferences were made of the amount of carbon per unit area and according to the density of soil samples.

All soil samples were placed individually in plastic bags with their respective label and transferred to the laboratory, where they were dried in a drying oven at a temperature of 60 ºC for 48 h, the roots, stones and other materials such as plant waste and wildlife waste.

Subsequently, the soil samples, already free of roots and other materials, were sieved with a wooden hammer until they passed through a 2 mm mesh, after completely homogenizing them, a subsample of approximately 5 g was extracted, dried at 105 ºC and it was ground for three minutes in a Speck 8000 Mixer/Mill^® type mill. Carbon analysis was determined using the direct combustion method at 1 000 ºC, with an automatic total carbon analyzer, equipped with a unit to process solid samples.

The analysis to determine the carbon in the soil samples was performed in the INIFAP soil laboratory in the Tecoman Experimental Field, Colima. To determine the carbon content in quantity, it was necessary to know the apparent density of the soil, for this, its calculation was made, by obtaining the total volume of the soil and then the adjustment, subtracting the volume occupied by roots, stones or any other residue.

Analysis of the information

Tests were carried out to verify that the residual assumptions were fulfilled, normality with the Shapiro-Wilk test and homogeneity of variances using the Levene test, both with a degree of significance (p< 0.05), with the help of the software R^© ver 3.6. Since the assumptions were fulfilled, an analysis of variance (Anova) of a factor (plots) was carried out for the carbon values in bushes, leaf litter, soil from 0 to 15 and 15 to 30 cm depth, the Duncan test was used to determine if there were significant differences (p< 0.05) of the variables analyzed by plot (^{Zar, 2010}).

Carbon in shrubs and grasses

The concentration of carbon by the components of leaf litter, herbs and shrubs was higher in the plots without management because this component was not removed due to management effects, there being an approximate difference of 1.5 Mg ha^-1 in the carbon content of herbs and shrubs, while the litter concentration was approximately 0.95 Mg ha^-1. With respect to the plots without management, where the plot with parota and purple rose II were analyzed, the carbon content provided by the herbaceous and shrub vegetation was higher in the parota, by 0.25 Mg ha^-1; however, no significant differences were found (Figure 1a).

Figure 1 Means and standard error of carbon quantity (Mg ha^-1) a) shrubs and grasses; and b) leaf litter in the plantation of different species with and without management. Average values ± standard error followed by letter differences (a, b) indicate significant difference for p< 0.05. 

Leaf litter carbon

In the case of litter, the plots with unmanaged parota were those with the highest carbon concentration (6.36 Mg ha^-1), while the species with the lowest carbon concentration were the plots with purple rose I with management (2.9 Mg ha^-1), without there being a significant difference in this component, mainly due to its high variability in carbon content in the different plots sampled (Figure 1b). In terms of biomass litter estimation, the reported values are in the range that other authors report for tropical areas, which vary from 3.2 to 5.2 Mg ha^-1 of mulch biomass reported by (^{Klinge, 1968}; ^{Aguilar, 2004}; ^{De Miguel, 2004}; ^{Bergamini et al., 2009}).

Carbon in soil

The highest carbon content in soil was found at a depth of 0 to 15 cm, in the parota plot with a 37.68 Mg ha^-1 and the lowest concentration was registered in the plots of purple rose I with 26.94 Mg ha^-1, finding significant differences between the plot with parota and purple rose I, while the remaining plot s, no significant differences were found (Figure 2a).

At the depth of 15 to 30 cm, the carbon content with the highest concentration was the plots of purple rose II with a 28.75 Mg ha^-1, very similar to the plots with purple rose II; while the lowest were teak plots with a 14.54 Mg ha^-1. In this component, significant differences are found between managed and unmanaged plots, while there are no differences between them (Figure 2b).

The difference between these two depths reflects that in plots with management the carbon content decreases by approximately 50%, while in plots without management it only decreases by approximately 30%, as can be seen in Figure 2. This may be due to because the movement of the humus in the soil profile is downward and its concentration decreases as the depth of the soil increases (^{Braakhekke et al., 2013}; ^{Paz and Etchevers, 2016}), but it can also be attributed to the fact that in the soil layers superficial, the greater number of roots develops and the microbial activity is greater.

Figure 2 Means and standard error of carbon quantity in soil (Mg ha^-1) of two depths a) 0 to 15 cm; and b) 15 to 30 cm in plantation of different species with and without management. Average values ± standard error followed by letter differences (a, b) indicate significant difference for p< 0.05. 

After the carbon concentration analyzes in the soils of the different plots have been carried out, it can be affirmed that at a greater depth the carbon content decreases approximately 35%. This was determined after performing Duncan’s multiple comparisons test, which showed that there are significant differences with a p< 0.05 between the carbon content at the depth of 0 to 15 cm, compared to the amount of carbon measured at the depth. 15 to 30 cm (Table 2).

In addition, when performing an analysis of variance to determine if there are significant differences between the plots with management and without management, it was found that the plot with the highest carbon content was at depth from 0 to 15 without management with a 36.53 ±5.15 Mg ha^-1, while the lowest carbon concentration was in the plots with management and a depth of 15 to 30 cm (17.12 ±5.79 Mg ha^-1) (Table 2). Within the plots without management at different depths it was found that there are significant differences, as well as between the plots with management as can be seen in Figure 3.

Figure 3 Amount of carbon in soil (Mg ha^-1) at two depths a) 0 to 15 cm and b); 15 to 30 cm in plantation of different species with and without management. Average values ± standard error followed by letter differences (a, b) indicate significant difference for p< 0.05. 

When the sum of the components was carried out (shrubs and herbaceous, litter, and soil), it was found that the sites with the highest carbon content were the plots without management as can be seen in Figure 4, mainly due to the null herbaceous vegetation and bushy that is eliminated in the cultural works and maintenance of the place. Furthermore, the study reflected that the total carbon concentration, the plots without management had 30% more carbon than the plots with management, with a significant difference.

Figure 4 Means and standard error of total carbon quantity (Mg ha^-1) in the plantation of different species with and without management. Average values ± standard error followed by letter differences (a, b) indicate significant difference for p< 0.05.