Materials and methods

Study area

The work was carried out in the municipality of Matamoros (area of 80 799.21 ha), Coahuila; between the parallels 25° 23’ and 25° 48’ north and the meridians 103° 03’ and 103° 23’ west (Figure 1), at an altitude of 1 120 m. The zone is characterized by a very dry or desert climate, semi warm with summer rains and cool winter with extreme oscillation, whose annual total precipitation is 250 mm and average annual total evaporations that reach 2 500 mm (^{García, 2004}). The geological elements belong to the Quaternary and Cretaceous periods, composed of alluvial and aeolian sediments, gypsum limestone and conglomerates. On the other hand, according to ^{INEGI (2016)}, six groups of soils (Regosols, Cambisols, Calcisols, Leptosols, Fluvisols and Vertisols) have been identified, being their use mainly agricultural, with shrub vegetation, pastures and urban areas (^{INEGI, 2010}).

Figure 1 Location of the study area and distribution of sampling sites. 

The work was divided into four stages as described below:

Site selection and soil sampling. Samples of the surface layer (0-30 cm) of different soils (two kilograms approximately) were collected in the municipality of Matamoros, Coahuila. The soils corresponded to sites that were selected by superimposing a 2.5 x 2.5 km grid in the polygon of the study area (Figure 1), where the vertices of the grid corresponded to potential sampling sites, whose geographic coordinates were recorded to later go to the field, as indicated by ^{Segura-Castruita et al. (2005)}. However, in the final choice of the sites, the accessibility of the places (paved roads, dirt roads and gaps) was taken as criteria, using the topographic charts of the ^{INEGI (2016)}. Each site was georeferenced with a Garmin Etrex GPS. Subsequently, with the location data and different characteristics of the soils, a database was generated. Likewise, an interview was conducted with the producers who owned the plots or places where the sampling sites were located.

Obtaining organic carbon and apparent density. The bulk density (Da) was obtained by the paraffin method; while the percentage of organic carbon (CO) in the soil was determined by the wet combustion method of ^{Walkley and Black (1934)}, as established in Mexican Standard NOM-021-RECNAT-2000 (^{SEMARNAT, 2002}).

The COS calculation and its variation in the region was obtained with the equivalent mass method (COS calculation stored in a reference soil mass), taking into account the effect of the change in Da on the carbon of the stored soil, with equation (1) proposed by (^{Eller et al., 2006}).

Where: SCDi represents the COS store in layer i (COSDi in Mg ha^-1 of C); SCi represents the COSi content (g kg^-1) in layer i; BDi is the apparent density (g cm^-3) in layer i; and Hi is the thickness (cm) of layer i. The term η_i represents the percentage of particles with a size >2 mm in layer i. It will be reported for the surface layer (0-30 cm).

Spatial distribution of COS and comparison. The spatial distribution of the COS concentrations was obtained by performing an interpolation with the Ordinary Kriging geostatistical method (^{Cruz-Cardenas et al., 2010}) in the ArcGIS 10.3 © software, where equation (2) was applied.

Where: Z_(s) is the variable of interest (COS); μ_(s) is the determinant and random trend in which the variable of interest (COS) is found; ε_(s) is the autocorrelated error; and s indicates the location. This interpolation was useful for obtaining the thematic map of organic carbon from the surface layer of Matamoros soils, for the year 2015.

Besides, to analyze the COS contents of the map obtained in this work and its behavior over time, the data of percentage of CO and Da of the points obtained by ^{Segura-Castruita et al. (2005)} and the COS was calculated again with equation 1, it is important to clarify that the map values that these authors reported were not used, since the interpolation was done with the inverse distance interpolator (IDW), reason why which, to bring about a level playing field, the interpolation of the COS content of 2002 with the Ordinary Kriging method was carried out again as was done for the 2015 data.

The comparison of the two thematic maps was carried out with a cross-validation table, for which the CROSSTAB module of the Idrisi Selva software was used; also, the Kappa concordance index (ICK) was obtained between the two maps (^{Landis and Koch, 1977}).

Characteristics of the floors

The soil samples of the superficial layer that were collected in the municipality of Matamoros were 92, since some places were not allowed access or were paved. The Da of these soils varied from 1.03 to 1.57 g cm^-3 (Table 1). The soils are used for agricultural purposes, urban and without apparent use. The agricultural soils have crops of alfalfa, cotton, beans, fodder maize, walnut and vegetables, some of intensive use with two annual crops (monocultures or intercrops), others with an annual crop (spring-summer cycle) and semi-perennial or perennial crops. These areas are irrigated, which can be irrigated with a gravity system or with deep well water. Also, some soils receive organic manure amendments and others do not. In such a way, that the percentages of organic carbon of the soils varied from 0.15 to 2.16% (Table 1).

The Da that was found in the study area, corresponded to soils with clay-to-sand textures, not compacted (^{Porta et al., 2010}) that according to ^{Heuscher et al. (2005)}, generally occur in soils with low percentages of CO. Situation that could be related to the type of vegetation that supports the soils, which is composed of xerophilous and rosethophilous scrub (^{INEGI, 2010}). ^{Montaño et al. (2016)}, indicated that this type of vegetation is common in the climates that prevail in the arid and semi-arid zones, and influence the contribution of carbon to the soil, as well as the intensive agricultural activities that influence the CO content (^{Becerra et al., 2014}).

Organic carbon from soils in 2015

The superficial layer (0-30 cm) of the soils of the municipality of Matamoros in 2015 presented COS with a variation of 2.48 to 5.63 Mg ha^-1, where the highest contents were found in the northwest and central part in a northwest-southeast direction of the municipality, which were grouped into classes >4.5 and 3.5-4.5 Mg ha^-1, respectively; while the lower contents are to the northeast between the hills that are in Matamoros and in the southern part of the municipality (Figure 2).

Figure 2 Spatial distribution of the organic carbon contents of the soils of the municipality of Matamoros, Coahuila. 

On the other hand, class 3.5-4.5 Mg ha^-1 occupied the largest area (46 266.66 ha) while the smallest area was occupied by soils with <2.5 Mg ha^-1 with 4 493. 45 ha (Table 2). The contents of COS in the municipality are low compared to the range of COS contents (2.1 to 72 Mg ha^-1) that have been reported for other arid regions of Mexico and the world (^{Díaz-Hernández et al., 2003}; ^{Elbasiouny et al., 2014}; ^{Montaño et al., 2016}). In this sense, the accumulation of COS is related to the type of vegetation and climatic characteristics such as the annual precipitation index (^{Bui et al., 2009}) and even soil management (^{Lents and Lersch, 2014}). However, up to this point it is not possible to establish the magnitude of change in relation to the past of these soils.

The thematic map that was obtained from the 2002 of COS data and the georeferences reported by ^{Segura-Castruita et al. (2005)}, it is different (Figure 3) that was obtained for 2015, since for 2002 only two classes of contents were established. The class 3.5 - 4.5 Mg ha^-1, had the largest area (51 639.49 ha) which represented around 154 919.54 Mg, while the class 2.5 - 3.5 Mg ha^-1 occupied an area of 29 159.22 ha, which had a content of 72 898.05 Mg, approximately; making a total of 227 817.54 Mg for the surface layer of Matamoros soils. The spatial distribution of the largest contents was found in the central part of the municipality. This content is lower than that obtained in our research (293 329.42 Mg), which means that in thirteen years the COS increased 65 511.88 Mg.

Figure 3 Distribution of the organic carbon contents of the soil of the superficial layer of Matamoros, Coahuila (adapted from ^{Segura et al., 2005}). 

When comparing the COS content maps of the year 2002 and 2015, graphically it can be seen that they are different, as was confirmed by the ICK (0.10) which indicates that the maps have a very low concordance. Thus, in the 2002 map, two kinds of content appear, while in 2015, four were detected. The differences reflected changes between the COS of the two maps, which could be established when the cross-tabulation was carried out, where some areas increased and others decreased (Figure 4).

Figure 4 Map of changes in the organic carbon contents of the soil between the years 2002 and 2015, in the municipality of Matamoros, Coahuila. 

The magnitude of the changes in surface of the COS classes was different in the whole study area in such a way that, the COS contents in the surface occupied by class 2.5 to 3.5 Mg ha^-1 in 2002 decreased by 2015, both negatively and positively; that is, in part of this surface the contents were reduced to <2.5 Mg ha^-1, appearing an area of 662.55 ha that in 2002 did not exist (Table 3).

In contrast, part of the surface occupied by class 2.5 to 3.5 Mg ha^-1, also contracted because the contents increased, both to the next higher class (3.5 to 4.5 Mg ha^-1) and to the >4.5 Mg ha^-1 (15 650.81 and 258.56 ha, respectively). Situation that also occurred in class 3.5 - 4.5 Mg ha^-1 of the map 2002, although the changes were larger. It gives such form that, in 2015, an area of 4 128.84 ha with COS contents <2.5 Mg ha^-1 was detected, which did not exist in 2002, in general the COS increased in the surface layer of the municipality’s soils.

These results are related to what was observed during the field trips for sampling. The agricultural areas were divided into two, areas that corresponded to the common and small property areas. Some small-property soils received manure amendments and others did not. They also had different management when supporting various crops, although the main one was forage maize. On the other hand, the vegetation that was observed in the non-cultivated areas, consisted of scrubland as governor, mesquite and huizache trees, grasses and organs. In this sense, ^{Fortis-Hernández et al. (2009)} indicated that the contribution of manure to agricultural land in the Lagunera Region is a common practice in some small properties, reaching up to 120 t ha^-1 of manure.

In this regard, the addition of manure increases short-term carbon storage (^{Lents and Lersch, 2014}). However, ^{Flores-Sánchez et al. (2015)} indicated that the traditional management of soils causes a minimum accumulation of CO in soils; since tillage can cause the loss of COS reserves (^{Olson, 2013}). Likewise, the arid climate and the vegetation of the region also influence the COS contents in those areas that are not cropland (^{Montaño et al., 2016}).

The previous thing, allows to deduce that the applications of organic amendments would be increasing the reserves of COS in the floors of Matamoros, but at the same time the conventional tillage systems would not allow the accumulation of a greater amount; also, in those places where these organic practices are not carried out, the carbon reserves could be lost quickly due to the effect of tillage and the opening of new areas to the crop.