Introduction

The comprehensive utilization of resources is based on the principle that nature must be understood as a whole; that the water, earth, air, plants, animal and human components interact between one another and that modifications in any one of them have direct or indirect impacts on the rest. Another principle is that the diversity of components of the unit, of the conditions of production and of animals and plants must be increased, and according to the latter principle, the earth, water and air must be kept clean, and their enhancement or recycling through various practices must be emphasized (^{Nair, 1997}). Natural resources are lost as a consequence of the need to produce food in order to meet the demand of the population, which leads to the expansion of the agricultural frontier and the reduction of the fallow periods (^{Young, 1990}).

The soil is very important to human beings for purposes of agricultural production, as its state determines the activities that need to be carried out and the corrections required to attain the desired levels of output. The soils that develop beneath natural vegetation function as active, stable balanced systems and provide environmental services (^{Fisher and Binkley 2000}). On the other hand, agricultural soils usually show a degradation process and contain less organic matter, total nitrogen and exchangeable bases, as well as a lower cationic exchange capacity (^{Geissert et al., 2000}). Their structural stability is different, and the edaphic material is removed as a result of the erosion caused by tillage, surface runoff and wind deflation (^{Meza and Geissert, 2006}).

Various physical and chemical properties lend the soil the necessary qualities to hold life and maintain the productive capacity of agricultural land -which functions are negatively affected by degradation phenomena such as erosion and the decrease of basic components like fertility and biodiversity (^{UNCCD, 1996}).

Agricultural soils often provide low-quality environmental services, and technical and economic efforts are required to preserve or restore their ecological functions. Agroforestry systems, such as coffee plantations under shade, meet intermediate conditions compared to agricultural soils, being more disrupted than the forests, but less so than annual crops.

According to ^{Nair (1993}), the woodlands produce biomass, which maintains and enhances the levels of organic matter; certain species fix nitrogen, while others protect against water- and wind-related erosion, which reduces the loss of nutrients; at the same time, they favor an increase in soil fertility, apart from the fact that they can take nutrients from the deepest layers of the soil. Furthermore, they can enhance various physical properties, generate a favorable micro-climate beneath the tree canopy, and promote increased activity by the microorganisms in charge of breaking up organic matter.

However, adverse effects are also generated through poor agroforestry planning, such as competition for water and nutrients, inhibition of growth due to excess shade, and difficulty to practice mechanical harvesting (^{Mahecha, 2003}).

This study utilized four systems: a natural system, two agroforestry alternatives, and a monoculture, each of which has a different impact on the soil as a result of its particular vegetal cover. These effects are reflected in the physical and chemical properties of the substrate, so that the capacity of agroforestry to maintain or increase fertility and enhance efficiency in carbon sequestration and nitrogen fixation contrasts with that of the montane cloud forest. Therefore, this research sought to prove or disprove the claims in regard to agroforestry, as well as to gain knowledge of how the essential elements vary with increasing soil depths in these systems.

Based on all of the above, the objective of this study was to determine the state of the texture, apparent density (DAP), field capacity (FC), and permanent wilting point (PWP), organic matter (OM), inorganic N (ammonium, nitrates), total N, P, exchangeable bases, pH and Cationic Exchange Capacity (CEC) of the natural forest systems, two agroforestry systems with coffee trees, and a coffee monoculture, in order to establish the nutritional efficiency of their soils, in the municipality of Huatusco, Veracruz.

Materials and Methods

Soils in the traditional system (coffee and multiple-use trees), commercial system (coffee-macadamia-avocado polyculture), sun coffee plantations (monoculture), and in the montane cloud forest or mountain mesophile system, in the municipality of Huatusco, located at the center of the state of Veracruz, at the geographical coordinates 19°10’25” N and 96°57’43”W, and at an altitude of 1345 masl. The soil types are Cambisole and Luvisole. The former is characterized by a rocky subsoil and susceptibility to erosion; the latter has a clayey subsoil and is particularly prone to erosion. The climate is warm humid, with an average temperature of 19.1 °C; its mean annual precipitation is 1 825 mm (^{Cano et al., 1998}). Below is a description of each system.

Traditional coffee production system (TCPS). It consists of a traditional coffee plantation in the shade, which comprises a wide variety of native trees or natural vegetation, as well as Persea schiedeana Nees. (chinene), Inga spp. (inga) y Grevillea robusta A. Cunn. ex R. Br. (grevillea). The coffee trees are distributed in a 2 x 2 m topological arrangement and have a vegetal cover with a mean carbon sequestration of 53.56 Mg ha-¹ (^{Hernández, 2013}).

Coffee-Macadamia-Avocado production system (CMAPS). This polyculture commercial system is characterized by the production of coffee with market-oriented associated species, as a productive diversification strategy. In this case, it includes coffee (Coffea spp.), macadamia (Macadamia tenuifolia F. Muell.) and avocado (Persea americana Mill.). Its vegetal cover provides a mean carbon sequestration of 17.10 Mg ha-¹ (^{Hernández, 2013}).

Sun coffee production system (SCPS). As described by its name, this is a coffee culture without shading. It seeks to maximize the production of coffee as a monoculture, as it is rather uncommon, amounting to little less than 10 %. Its vegetal cover has a mean carbon sequestration of 9.18 Mg ha-1 (^{Hernández, 2013}).

Natural montane cloud forest (Forest). This association corresponds to the cloud forest, which in the study area has been extremely disturbed by the introduction of extensive livestock and coffee plantations. This vegetal community has three strata: the higher stratum, with a height of 28 to 30 m; the intermediate stratum, with 18 to 20 m, and the lower stratum, consisting of a single species, with 4 to 6 m. The prevalent genera are Liquidambar, Quercus, Meliosoma, Cornus, Ilex and Clethra. The area includes three microrelief conditions: river lowland, mountainside and plateau (^{Pérez, 2004}). It has a vegetal cover with a mean carbon sequestration of 421.01 Mg ha-1 (^{Hernández, 2013}).

Four systems, regarded as levels of a first factor of treatments, were selected. Three sites were chosen at random within each selected system, and samplings were carried out at three different depths -0-10 cm, 10-20 cm and 20-30 cm- in order to evaluate the properties of the soil. The systems were: montane cloud forest (Forest), coffee-macadamia-avocado production system (CMAPS), traditional coffee production system (TCPS), and sun coffee system (SCPS).

In order to determine the nutrient efficiency of the soils of the four systems, the following dependent variables were evaluated: Texture, Apparent Density (DAP), Field Capacity (FC), Permanent Wilting Point (PWP), Organic Matter (OM), inorganic Nitrogen (ammonium, nitrates), total Nitrogen, Phosphorus, Exchangeable Bases, Microelements, Hydrogen Potential and Cationic Exchange Capacity.

The physical parameters of the soil were determined in order to characterize each system; the assessments were carried out from October to December 2014, at the Laboratorio de Física de Suelos of the Universidad Autónoma Chapingo, under the standards and methods of the Mexican official norm NOM- 021-RECNAT-2000 (^{Semarnat, 2002}). Thus, the apparent density DAP was estimated by using the cylinder; the field capacity (FC), through water retention at 33 kPa; the permanent wilting point (PWP), at 1 500 kPa, with a pressure cooker and membranes, and the texture, using a pipette, with organic matter destruction and dispersion with sodium hexametaphosphate.

The purpose of the assessment of the chemical parameters of the soil was to describe each system and establish the nutritional conditions available to the plants; these assessments were carried out in November, 2014, at the Laboratorio Central of the Universidad Autónoma Chapingo, also according to the criteria established by the Mexican Official Norm NOM-021-RECNAT-2000 (^{Semarnat, 2002}).The analyzed samples were air-dried and sifted through a 2 mm sieve in order to determine the values of: exchangeable potassium (K), extracted with ammonium acetate (1.0N pH 7.0), using flame emission spectrophotometry; exchangeable calcium (Ca) and magnesium (Mg) obtained with ammonium acetate (1.0N pH 7.0), using atomic absorption spectrophotometry; organic matter (OM), using the ^{Walkley and Black method (1934}); total nitrogen (N tot), estimated through steam distillation: Kjeldahl; inorganic nitrogen (N min), with potassium chloride (2N), using steam distillation; available phosphorus (P), with Bray p-1; cationic exchange capacity (CEC), with ammonium acetate (1.0 n pH 7.0), using steam distillation; pH, through potentiometric titration in a 1:2 sample/water ratio, and available water (FC - PWP).

The data were analyzed according to a 4 x 3 factorial experiment with two factors, corresponding to four systems and three soil depths. Three repetitions for each of the 12 combinations of treatments were considered. The corresponding statistical model is the following:

Where:

εijt´s ~ N (0, δ2),

εijt´s = Mutually independent

t = 1, …

rij; i = 1, …

a; j = 1, …, b

Yijt = Response value in the t^th replication of the treatment with the ith production system and at the j^th soil depth

μ = Constant (overall average)

αi = Effect of the i^th system

βj = Effect of the j^th depth

( αβ)ij = Effect of the interaction for the i^th system and

the j^th depth

εijt = Random errors

The collected data were averaged by site and by soil depth; all the assessed characteristics were analyzed using the ^{SAS 9.1 statistical software (2002)}. Variance analyses and multiple mean comparison tests were carried out using Tukey’s method (^{Steel and Torrie, 1998}) in order to determine significant differences between the four production systems and between the three different soil depths, as well as the potential interactions between these two factors.

Results and Discussion

Physical properties

Concentrations of the various physical parameters were significantly different in the studied systems. The mean comparison tests for the physical properties of the soils are shown in Table 1.

Table 1 Mean comparison test of soil physical parameters in the studied systems in Huatusco, Veracruz, Mexico. 

Forest = Montane cloud forest; TCPS = Traditional coffee production system; CMAPS = Coffee-Macadamia-Avocado production system; SCPS = Sun coffee system; SD = Soil depth; N = Number of samples; DAP = Apparent density; FC (%) = Field capacity; PWP (%) = Permanent wilting point; means with the same letter are not significantly different (Tukey, α ≤ 0.05)

The soil texture classes were loamy-sandy in the forest, and loamy to loamy-clayey in the coffee plantations; at all depths they had a loamy texture with variations. Furthermore, they are acidic and their concentrations of exchangeable bases are low, despite having a surface layer rich in organic matter.

The variance analyses showed the presence of statistical differences (α < 0.05) for apparent density (DAP), with P <0.0001 in the four evaluated systems. The system with the lowest apparent density was the Forest, followed by the TCPS and the SCPS. CMAPS was the system with the highest apparent density. This accounts for the fact that the Forest and the TCPS had the highest concentrations of organic matter. As is well known, systems with the highest content of carbon and organic matter have lower values of apparent density (^{Monroy, 2009}). The vegetal cover is scarcer in the SCPS than in the CMAPS, the TCPS and the Forest, which favored a lower contribution of organic matter, and a higher oxidation thereof, since this component is closely related to apparent density and porous space. In the presence of higher contents of organic matter, the soils have lower densities and, consequently, favorable physical conditions, compared to higher densities (^{Cavazos and Rodríguez, 1992}; ^{Carter, 2002}).

The apparent density increased in direct proportion to soil depth in the Forest, CMAPS and TCPS, due to the low content of organic matter at greater depths, where there is less aggregation and more compacting. However, the SCPS showed the opposite behavior.

According to the ^{NOM-021-RECNAT-2000} norm, which establishes the specifications for fertility, salinity and soil classification, as well as the apparent density values, the Forest and the TCPS were classified among the organic and volcanic soils, while the CMAPS and the SCPS were classified among the mineral soils. The studied systems did not reflect changes in the apparent density, and the small differences cannot be ascribed to these; they may depend on the texture, the organic matter contents and the state of compacting of the soil (^{Alvarado, 2007}). The structural conditions are better in the Forest and the TCPS, thanks to the relationship between organic matter and apparent density.

The variance analyses showed significant statistical differences (α < 0.05) for the FC, or field capacity, with P = 0.0003, and for the PWP, the permanent wilting point, PWP, with P < 0.0001; the highest values were found in the Forest, followed by the TCPS and CMAPS; instead, the SCPS had the lowest values.

As depth increased, these variables diminished in the Forest, CMAPS and TCPS. Their behavior varied in the SCPS.

As for the usable humidity (UH) or actually available water, the highest values were registered in the SCPS, while the Forest had lower values, of 1.01 to 13.22 % in the agroforestry soils, 1.37 to 12.98 % in the forest, and 13.42 to 14.17 % in sun coffee plantations. The significantly high values for the permanent wilting point indicate that a considerable amount of water is retained under strong tension in the soil and cannot be absorbed by the plants.

Chemical properties

The concentrations of the various chemical elements differed significantly among the studied systems. Table 2 shows the mean comparison tests for the chemical properties of the soils.

Table 2 Mean comparison test for chemical soil parameters in the systems studied in Huatusco, Veracruz, Mexico. 

Forest = Montane cloud forest; TCPS = Traditional coffee production system; CMAPS = Coffee-Macadamia-Avocado production system; SCPS = Sun coffee system; SD = Soil depth; N = Number of samples; pH = Hydrogen potential; CEC (Cmol (+( kg-1) = Cationic exchange capacity; OM (%) = Organic matter; tot N (%) = Total nitrogen; N min (mg kg-1) = Mineral nitrogen; P (mg kg-1) = Phosphorus; K (mg kg-1) = Potassium; Ca (mg kg-1) = Calcium. Measures with the same letter are not significantly different (Tukey, α ≤ 0.05).

The chemical properties of the soil were analyzed in order to characterize the nutritional condition at various depths in each of the systems.

No direct relationship was verified between the hydrogen potential (pH) and the soil depth. The variance analysis showed significant statistical differences between the four evaluated systems (α < 0.05), with P < 0.0001). The system with the lowest pH was the Forest, followed by the TCPS, and the CMAPS had the highest values; this parameter behaved similarly to the apparent density of the soil.

According to ^{Brady and Weil (1999}), the capacity of the vegetation to absorb mineral elements exerts an important influence on the characteristics of the soils where it develops, and the pH affects the assimilation of nutrients directly.

Acidity was associated with a low saturation of the adsorption complex in exchangeable bases, mainly in calcium, but also in potassium and magnesium. The lowest values obtained in the present study are accounted for by the differences in the contents of organic matter, since this acidifies the soil (^{Ortiz and Ortiz, 1990}), as shown in the Forest and the TCPS. Based on the norm ^{NOM-021-RECNAT-2000} and on the pH values, all the soils of the systems are classified as moderately acidic.

For the cationic exchange capacity (CEC), the variance analysis assumes significant statistical differences (α < 0.05), with P = 0.0303 in the four systems. The interval in the present study was 41.70 to 46.68 Cmol (+) kg-1, both of which are extremely high values, but diminished as the soil depth increased. The system with the lowest capacity was SCPS, followed by the CMAPS and the TCPS; the Forest had the highest values. Therefore, it may be concluded that the natural chemical environment of the studied soils has been proven to favor a good availability of various nutrients for the plants. The CEC depends on the texture of the soil and on its content of organic matter (^{Osman, 2013}). In general, the more clay and organic matter there is in the soil, the higher the exchange capacity (^{Essington, 2004}; ^{Roy et al., 2006}). This variable was prominent in soils with a major organic matter concentration, namely those of the Forest and the TCPS.

The same is true of the soil depth, and therefore of the cationic exchange capacity, since lower organic matter concentrations were found as the soil depth increased. In general, the CEC increases in almost every soil type when the pH becomes alkaline; however, this was not the case in the study involved. Based on the norm NOM-021-RECNAT-2000 and on the CEC values, all the systems were classified as “extremely high”.

As for the organic matter (OM), the variance analysis revealed significant statistical differences (α < 0.05), with P < 0.0001, in the four assessed systems. The one with the lowest percentage was the SCPS, followed by the CMAPS and the TCPS. In contrast, the Forest registered the highest percentages for this variable, which is associated with the vegetal cover. As the soil depth increased, the organic matter diminished in all systems, since the largest input for this variable occurs at surface level (^{Soto et al., 2007}). In this case, the presence of dead leaves fallen from the canopy is confirmed, reflecting the management practices used in all the systems, the SCPS included, which have not been intense enough to affect the organic matter contents. When the ^{NOM-021- RECNAT-2000} norm was applied to the results obtained for this variable, the SCPS and the CMAPS had a low percentage of OM, unlike the Forest and the TCSP. Organic matter diminishes with a more intense use of the soil. Forest soil disturbance reduces the organic matter contents, since the vegetal cover tends to be removed, which favors an increase in temperature and, consequently, in microbial activity, as well as a rapid mineralization of the substrate, that eventually cause its loss.

Significant statistical differences (α < 0.05), with P < 0.001, were found for total nitrogen in the four evaluated systems. The system with the lowest percentage of total nitrogen was SCPS, followed by CMAPS and TCPS; the opposite is true for the Forest, due to the behavior of organic matter. As the soil depth increases, the total nitrogen diminishes considerably. The mineral nitrogen has a greater dynamic range than the total nitrogen, and therefore it manifestation varies sometimes. However, mineral nitrogen -like total nitrogen- was higher in the Forest soil than in the other systems. This is because the concentrations of mineral and total nitrogen in less disturbed systems are determined by the organic matter. According to the ^{NOM-021-RECNAT-2000} and to the total nitrogen values, the SCPS was shown to contain a high percentage of total nitrogen, while the percentages contained in the Forest, CMAPS and TCPS were rated as extremely high.

There were statistically significant differences (α < 0.05), with P = 0.001, in the mineral nitrogen concentrations of the four assessed systems, according to the variance analysis. The system with the lowest percentage of mineral nitrogen was CMAPS, followed by the SCPS and TCPS. The system with the highest values for mineral nitrogen was the Forest. The concentrations of mineral nitrogen were observed to decrease as the soil depth increased in the Forest and the TCPS. In the SCPS and the CMAPS, the behavior of this component did not follow a clear pattern. At the depth of 0-10 cm, the results suggest the existence of an association with the organic matter and total nitrogen contents. Based on the ^{NOM-021-RECNAT-2000}, the SCPS had a low content of mineral nitrogen, while the Forest, the CMAPS and the TCPS had a medium content.

There are significant statistical differences (α < 0.05), with P = 0.0054, for phosphorus (P) in the four evaluated systems. The system with the lowest concentration of this element was the Forest, followed by the CMAPS and TCPS, in contrast with the SCPS. As in other cases, there is an opposite reaction of this system in regard to soil depth to that occurring in the CMAPS and SCPS; however, no regular pattern was determined for the Forest of the TCPS. The phosphorus contents in all systems were classified as low. ^{Alvarado (2007}) detected considerable variation in the phosphorus contents according to the climatic conditions in which the samples are collected in field. The SCPS had a medium content of phosphorus, while the Forest, the CMAPS and the TCPS had a low content. No relationship was established between P and the organic matter variables, apparent density, pH and nitrogen. In the sun coffee system, the values for phosphorus were high because coffee plantations in Veracruz receive extensive management, including regular application of fertilizers, which results in important changes in the content of nutrients (^{Geissert and Ibáñez, 2008}).

Significant statistical differences (α < 0.05) were also confirmed for exchangeable bases (K, Ca, Mg) -with P = 0.0029 for potassium and P < 0.0001 for calcium and magnesium- in the four assessed systems. The Forest registered the lowest concentration of all three. The system with the highest values for potassium and calcium was the SCPS, while the CMAPS had the highest values for magnesium. The effect of the soil depth on both elements was shown to be the reverse in all systems without a definite pattern, and magnesium concentrations particularly were verified in the Forest and the TCPS.

According to ^{Vergara (2003}), there is a close relationship between the exchangeable potassium, calcium and magnesium ions and the pH. In this study, this association is confirmed based on the data for the Forest and the TCPs with the lowest pH, but also with the lowest concentrations of exchangeable bases. These are indicative of a loss of nutrients, which are swept along by huge amounts of rain water that cause lixiviation (^{Geissert and Ibáñez, 2008}). According to the NOM-021- RECNAT-2000 and to the values of potassium, a mean content of potassium was identified in the Forest (0.41 Cmol(+)kg¹), while the CMAPS (0.67 Cmol(+) kg¹), the TCPS (0.76 Cmol(+) kg¹) and the SCPS (0.93 Cmol(+) kg¹) had high concentrations. The calcium content was extremely low in the Forest (0.74 Cmol(+) kg¹), low in the TCPS (1.86 Cmol(+) kg¹) and the CMAPS (4.94 Cmol(+) kg¹), and medium in the SCPS (5.01 Cmol(+) kg¹). The content of magnesium in the Forest (0.24 Cmol(+) kg¹) was also extremely low; it was low in the TCPS (0.52 Cmol(+) kg¹), medium in the CMAPS (2.22 Cmol(+) kg¹), and high in the SCPS (3.25 Cmol(+) kg¹).

Conclusions

The soil of the sun coffee production system had similar apparent densities and textures to those of the natural and agroforestry systems with coffee trees, as well as a lower water retention capacity expressed in low field capacity and permanent wilting point; in contrast, the Forest had the highest values for these parameters; intermediate values correspond to the soils of the agroforestry and the traditional coffee and Coffee- Macadamia-Avocado production systems.

The pH of the soil under the Forest, the agroforestry systems and the coffee monocultures was found to be acidic, a condition that increased with higher contents of organic matter. In those soils where this was abundant on the surface layers as a result of a broad vegetal cover, the cationic exchange capacity was found to be greater. However, organic matter becomes less abundant at greater soil depths.

Total nitrogen in the studied systems was found to range between high and extremely high, in direct proportion to the presence of organic matter; the behavior of mineral nitrogen varied.

In general, we conclude that the Forest was the best of the four studied systems, followed by the traditional coffee and Coffee-Macadamia-Avocado production systems; conversely, the sun coffee system proved to be the most deficient. The TCPS and CMAPS enhanced the physical and chemical properties of the soil, compared to the sun coffee system. The establishment of the coffee monoculture favored a reduction of organic matter and nitrogen concentrations, as well as a lower water retention capacity and an increase in the content of exchangeable bases.

Conflict of interests

The authors declare that they have no conflict of interests.

Contribution by author

Paul René Fernández Ojeda: sampling, sample preparation, data systematization, and discussion of the results; David Cristóbal Acevedo: assistance in sample preparation and review of the results of the laboratory tests; Antonio Villanueva Morales: statistical analysis of the data; Miguel Uribe Gómez; drafting of the paper and discussion of the results.