Structure of the Landscape and Quality of Dystrophic Red Oxisols

Modeling the soil system using geostatistical techniques, geographic information systems and remote sensing enables the identification, monitoring and the sustainable use of soil resources. In this way, the interactions between the landscape attributes (or independent variables) and the physical and chemical attributes (or dependent variables) of dystrophic red oxisols in southern Minas Gerais State were evaluated. Twenty-three soil samples were correlated using linear and nonlinear regression models by overlapping landscape variables, such as the composition, configuration and relief, with the physical and chemical attributes of the soils. The models were ranked according to the Akaike information criterion, which indicates direct relationships between the organic matter contents and average slope gradient and between the geometric mean diameter and the adopted management and slope gradient. The clay content, percentage of native forest, landscape average slope gradient, and sum of exchangeable bases were conditioned primarily by the slope of the sampled area and the shape of the forest fragments. Thus, the attributes are explained largely by the relief, which restricts the use of forests and enables their preservation, as well as the types of management adopted in the different uses and the relative abundance of native forests. Therefore, the use of conservationist practices and the improvement of management practices, as well as compliance with environmental law, tend over the long term to result in better-preserved and more sustainable soils for agricultural activities, thereby reducing morphogenesis processes in relation to those of pedogenesis.


Introduction
The environmental modeling of factors that condition the natural resource quality is essential for identifying, preserving and monitoring those resources.Therefore, ecosystem studies are essential for addressing the increasing degradation caused by unsustainable production techniques, in addition to the impact related to the projection of demographic growth and the consumption of goods, which implies more pressure on those resources (UNFPA, 2012).
Soils are highly affected by traditional management practices that increase surface runoff and erosion and increase the soil loss rate to values above the Soil Loss Tolerance (SLT) limit.This management approach reduces the productive capacity and downgrades soil attributes by breaking structures, changing the soil texture, reducing organic matter (OM) contents and causing compaction (Bertoni and Lombardi Neto, 2012).It also leads to the siltation and the contamination of water resources and the trophic chain by the surface runoff of agricultural inputs, the eutrophication of water bodies (Londres, 2011), the decrease in infiltration and percolation and the increase of carbon dioxide liberation by eroded soils (Pimentel et al., 1995;Morgan and Nearing, 2011).
In this scenario, the soil quality includes the ability to regulate and distribute runoff from water and chemical elements, to promote and supporting the growth of roots, to maintain an appropriate biological habitat and to respond to management by resisting degradation and keeping the soil at productive capacity (Vezzani and Mielniczuk, 2009).These attributes vary with differences in the configuration of landscapes and reflect directly on their use capacity.
From this viewpoint, a systemic approach is recommended because it considers the landscape diversity and interactions to explain and assess the quality of the different environments as well as the ecodynamic balance over the years.In addition, systemic conception using relationships among environmental factors such as the insertion (ranks), juxtaposition (proximity/continuity) and functionality (cause) allow us to work with data through more effective geoprocessing, in addition to contributing to a more diligent interpretation of the results, avoiding researcher bias (Chorley and Kennedy, 1971).These techniques allow for the outline of soil evolution with respect to changes in natural conditioning facts by management techniques such as plowing, harrowing, scarification, mechanical seeding, using agrochemicals, permanent grazing systems without fertilization, intensive irrigation, the absence of conservationist management practices such as planting in contour lines and/or directly, and the absence of environmental protection areas (EPA), permanent preservation areas (PPA) and legal reserves (LR).
Therefore, the biophysiographic features that control the morphogenesis of dystrophic red oxisols in an area with a homogeneous climate and geology in southern Minas Gerais State were assessed and modeled in this study.The aim was to acquire a systemic understanding of the way that natural conditioning factors affect the quality of soil and change the hydrosedimentological dynamics.

Study area
Twenty-three (23) samples were collected from landscapes in Alfenas, in southern Minas Gerais State (Figure 1).According to Köppen, the climate in this region is CwA (Tropical mesothermic or Tropical high-altitude) (Sparovek et al., 2007).This area belongs to the phytogeographic domain transition from Mata Atlântica to Cerrado, and it has only 8% native vegetal cover (Fundação SOS Mata Atlântica and INPE, 2013).
This region has an old geological surface with elevations displaying rounded contours, long slopes, open valleys and local discontinuities.Proterozoic charnockitic and granulitic polymetamorphic gneisses form the southern Minas Plateau (Hasui, 2010).There is a prevalence of tropical oxisols resulting from weathering that are associated with an udic hydric regimen with a high mean precipitation, which, over time, has led to the hydrolyzation and lixiviation of the bases and part of the silicon in the soils, in turn leading to the relative enrichment of iron and aluminum from the parental rocks.This action lowers the pH and raises the clay flocculation, which results in more developed structures and surpasses the effect of the impermeable clayed texture of oxisols, increasing infiltration to the detriment of surface runoff (Vitorino et al., 2003).However, traditional management practices have degraded these structures, increasing the surface runoff and reducing water infiltration (Ayer et al., 2015).For geostatistical evaluation, more recurrent biophysiographical variables were obtained from pedogenetic studies (Grunwald, 2006).From those results, linear and nonlinear regression models were generated using landscape biophysiographical features as independent variables and soil physical and chemical attributes as dependent variables.These models were ranked and selected using the Akaike Information Criterion (AIC), which compares the models among themselves and defines which of them have more consistent and explicative variables (Burnham and Anderson, 2002).This approach allowed for the characterization of the independent variables, that is, the biophysiographic factors that most affect local soil attributes.
For the experimental design, landscapes were considered as areas that after evaluation in loco, presented criteria including i) homogeneous areas of dystrophic red oxisols under anthropic use and native forest; ii) homogeneous climate and geological areas, in accordance with Vitorino et al. (2003) and Brasil (2010); iii) the primary land uses and management in the region; and iv) hydrosedimentologic flows converging to the sampling areas.From these criteria, a circular sampling area with a 250 m radius was adopted for analysis.
In each landscape, the metrics and landscape variables were estimated and used to build the regression models.The explicative variables are composed of the following two groups: elements of the landscape and elements of the relief.The response variables are the physical and chemical attributes of dystrophic red oxisols.The aim of these criteria was to define a homogeneous area in which the variation in soil attributes could be explained by the use and management variation and by their effect on the different relief configurations.These attributes were evaluated for the landscape and for the sampled area to explain the variable by using the synergistic effects of the interaction with landscape elements and/or with local features only.
To evaluate the management practices in question, we created the Management Quality Index (MQI), which was derived from the soil use and management factor "C" (Table 1) of the Universal Soil Loss Equation (USLE).According to Wischmeier and Smith (1978), factor "C" incorporates the soil coverage effect by using the canopy, the vegetation density and the management system for reducing eroding rain forces and increasing infiltration.Factor "C" is nondimensional and varies from 0 to 1, with 1 being used for bare soils in which surface runoff predominates, and 0 being used when infiltration predominates and erosion reduction occurs (Bertoni and Lombardi, 2012).The MQI was created from equation 1, which uses factor "C" values for the most important uses and management types employed in the region.Therefore, the higher the index, the lower the capacity of the landscape at retaining the soil and environment quality.MQI = ∑ (C) Ac / At (Equation 1) where MQI: Management Quality Index; C: soil use and management factor with values from 0 to 1 (Table 1); Ac: total area of use and management (ha); and At: total area of landscape (ha).0.00 *The coverage factor for beans was not found in the literature.The soy factor value was adopted by Bertol et al. (2001), whose natural characteristics are similar to those of beans (Roloff and Bertol, 1998).
Relief variables were obtained from the Topodata Digital Elevation Model (DEM), at 30 m of resolution.The DEM was filtered with the ArcGIS 10.5 Sink tool to remove the pixels with anomalous altitudes.The slopes (Figure 2A) for the landscapes around the sampled areas were then generated.The values of variables from the landscape composition and relief are presented in Table 2.The RGB composition in band 543 and the map for soil use and management for each landscape (buffer) using visual classification (Figure 2B) were also produced from Rapid Eye Satellite imagery.For the landscape composition, the ArcGIS 10.5 Patch Analyst Tool was used to calculate the use percentages for coffee, sugarcane, pasture, and native forest over a radius of 250 m.
In addition to these variables, the Shannon Diversity Index of the landscape, which defines the variety of landscape uses, was also considered.This index is calculated by using the quantity of soil use classes and the proportion of each one in the landscape (Table 2).The Shannon Index is zero when there is only one type of class in a landscape; for example, when a landscape is composed of 90% or 100% pasture.This index increases in proportion to the number of classes and their percentage in the landscape.For example, with 81% pasture, 9% native forest and 8% floodplain areas, the index is 0.6.However, with 40% pasture, 35% native forest and 15% floodplains, the index value is 0.9 (McGarigal and Marks, 1995;Fahrig et al., 2011).The shapes of the types of use were calculated by using AWMPFD metrics (Area-Weighted Mean Patch Fractal Dimension) from the ArcGIS 10.5 Patch Analyst Tool.The fragment mean fractal dimension is calculated by using the perimeter and native forest area to describe the shape of the forest patch.Because larger fragments generally present higher values for the shape complexity, the Area-Weighted Mean was used.This parameter varies from 1 to 2, and it has values close to 1 when they present simpler shapes such as circles and squares, and close to 2 when they present more complex and irregular shapes (McGarigal and Marks, 1995;Fahrig et al., 2011).
An evaluation of soil attributes was performed in 23 soil samples that were collected from 0 to 30 cm deep; 6 were from native forest (NF) and 17 were from spots under anthropic use (AU).They were used for the determination of the physical, chemical and morphological properties.These dependent variables were ranked and overlapped with the independent composition and relief variables.The soil attributes were described in loco, in accordance with Lemos et al. (2005).The soil analyses were as follows: texture, by the densimeter method (Black, 1986), with and without NaOH solution; flocculation index (Embrapa, 1998), besides sand, silt and clay classes, content of G > 0.25 = grains sum > 0.25 mm (g kg -1 ); G < 0.25 mm = grains sum < 0.25 mm (g kg -1 ); and aggregate stability, by the sieving method in water (Kemper and Rosenau, 1986).From these values, the Geometric Mean Diameters were calculated (GMD) (Kemper and Chepil, 1965).From the sorptive complex, the pH in water with KCl and CaCl2, relation 1:2.5;Ca-Mg-Al with extractor KC1 1 mol L -1 ; OM by oxidizing with Na2Cr2O7 2 mol L -1 + H2SO4 5 mol L -1 ; sum of exchangeable bases (BS) and bases saturation index (V) (Embrapa, 1998) were determined.The handbook of soil analysis was used and the resulting values were classified in accordance with Epamig (2012) (Table 3).

Statistical Analyses
Linear and non-linear regression models were generated.They were used to associate the independent landscape variables, composition and relief with the dependent ones; that is, the soil physical and chemical attributes.These models were ranked by the Akaike Information Criterion (AIC) and the derivative parameters AICc, ΔAIC and WAICc.The AIC is calculated on the basis of the variance in the residues of each regressive model, number of variables and intercept of the modeling.This criterion generates a large variation in the values.Each hypothetical model generates an AIC value, and the models with lower values are considered more appropriate for explaining the relationship between the landscape and the soil quality (Burham and Anderson, 2004).The AICc is the correction factor of the Akaike Information Criterion.It sets the bias of this criterion.This parameter also sets the bias for a number of variables in the model and for small sample models (n < 40).In addition, with large variations in values, it follows AIC logic in which lower values point to the more appropriate models.The individual values of the AIC parameters have little significance and should be interpreted by comparing the models.
For the evaluation of the models used here, the AIC-derivative parameters were generated; ΔAICc or delta AICc is the variation in the lowest AICc value generated here, that is, the value from the best model, minus the AICc value of the model being analyzed.Therefore, the best model will be set to zero and the other models will have positive values.This analysis allows for a comparison in relation to the best model.Models with values equal to or less than 2 will be considered consistent (Burham and Anderson, 2004).Finally, the WAICc derivative is the weight of the model generated in comparison to the others.This parameter is calculated from the Δ AICc, and it indicates that the probability of the model being analyzed should be better than the concurrent probability.This measure varies from 0 to 1 or from 0 to 100%.If a model presents a WAICc = 0.5, then it has a 50% chance of being the best model among the evaluated models or if the comparison is repeated several times (Burnham and Anderson, 2002).These analyses were run on R software using the tools 'bbmle' and "gam" (Bolker, 2008).From the generated models, the use and management factors were evaluated to find how to strengthen the effects of the relief and reflect the soil attributes by addressing the change in hydrosedimentologic dynamics.

Results and Discussion
From the regression models and the AIC, the landscape variables were identified; these variables explain and interfere with soil attributes.Most of the explanatory models presented nonlinear relationships, and only three presented linear relationships (Table 4).The relief metrics were more relevant to the analyses, once they presented interference in every soil attribute.The landscape composition metrics were also important for explaining the variable responses in the generated models.s =Variable referring to the use or slope gradient at the sampled spot, regardless of the 250 m-radius landscape; DMG = Geometric Mean Diameter; OM = Organic Matter; BS = Exchangeable Base Sum; V = Base saturation index; AWM = area-weighted mean; Clay = clay (g kg -1 ); G > 0.25 mm = sum of grains > 0.25 mm (g kg -1 ); and G < 0.25 = sum of grains < 0.25 mm (g kg -1 ).
The GMD variable was explained by the MQI (Figure 3A) that integrated the primary nonlinear model that explained this attribute (ΔAIC = 0; WAICc = 0.45) (Figure 3A), followed by a model integrated by the same variable in synergy with the average landscape slope (ΔAIC = 1.3;Ayer, J.E.B., Raniero, M., Servidoni, L.E., Olivetti, D., Silva, A.L.N., Mincato, R.L. WAICc = 0.24).In accordance with the quadratic relationship, the GMD is higher in areas with intermediate MQI values and in the strongest slopes (Figure 3B).In general, areas with strong slopes had cultures with smaller C factors, such as native forest, pastures and coffee predominating, due to the natural difficulties of mechanization.
Therefore, the higher the MQI, the higher the surface runoff, that is, the lower the capacity of the soil use to stabilize soil aggregates, which may explain the lower GMD values (Conceição et al., 2005;Blanco and Lal, 2008;Vezzani and Mielniczuk, 2009).Figure 3. A) Relationship between the Geometric Mean Diameter (GMD) and the Management Quality Index (MQI); and B) the relationship between the GMD in relation to the synergy between the slope gradient and MQI.
The organic matter contents were considered low under both native forest and agropastoral use, although a higher content was found under native forest.This variable presented a nonlinear regression model, which was integrated only by the slope of the sampled area (ΔAICc = 0; WAICc = 0.82).Using the polynomial relationship, it is possible to observe that the smoother the slope is, the larger the organic matter content in the soil (Figure 4A).In general, areas with smooth slopes have less turbulent and slower flows, in addition to probably taking material that is carried from the neighboring area by surface runoff during the rains, resulting in the accumulation of organic matter in these areas, usually floodplains.Vezzani and Mielniczuk (2009) and Mohammad and Adam (2010) showed the variation in OM content in situations similar to the findings observed in this area, where there are lower concentrations of OM under agricultural uses as a result of the scarce vegetation cover.This trend increases movement by surface runoff to lower areas with smooth slopes.However, the low OM content under native forest is the result of fragments that were historically used as pastures, and only recently, there was a legal need for fencing the legal reserve (LR) and the PPA.The previous uses of native forests ended with the impaired capacity to regulate and distribute the flow of chemical elements, endangering their role in preserving the ecosystem.
Soils in the area vary from very acidic to acidic, with pH values between 4.2 and 6.5.The explanatory pH model (Figure 4B) presented a nonlinear relationship due to the synergistic slope effect and to the type of use in the sampled area (ΔAICc = 0; WAICc = 0.99).In applying the quadratic relationship, the pH was observed to be lower under native forest in areas with smooth slopes (Figure 4B).These soils result from weathering under udic hydric regimens with high average rainfall, which, over time, led to the hydrolyzation and lixiviation of the bases and silicon in the soil, causing relative enrichment with iron and aluminum from the parental rocks.This feature is the principal pH conditioning.In addition, when we compare the pH averages in the areas under native forest and under anthropic use of 4.77 and 5.24, respectively, it is obvious that the lower pH under native forest results from both the weak organic acids from the OM decomposition and from water erosion in the area that causes H + and Al 3+ concentration increases.However, the higher pH under agropastoral uses primarily reflects the soil correction for planting.Within the metrics of the exchangeable base sum (BS), the slope of the sampled area is the only explanatory model (ΔAICc = 0; WAICc = 1).It was checked by using the polynomial relationship, in which the smoother the slope is, the higher the base sum (Figure 5A).It is worth noting that the area soils are, on average, naturally poor in Ca, Mg and K, which are part of the BS calculation.Therefore, these findings result primarily from the bases added for fertility correction, and only in small part from the OM decomposition.In areas with smooth slopes, the potential gravitational energy is lower and results in less turbulent and slower flows, reducing erosion and increasing infiltration during rains.This reduced flow results in the accumulation of fertilizers and OM from higher altitudes and contributes to the increased BS in the soil (Blanco and Lal, 2008).
The base saturation index values (V) presented large variations from very low values, such as 6.56%, up to very high values, such as 90.78%.This variable was explained by the synergy between the slope and the use in the sampled areas (ΔAICc = 0; WAICc = 0.7).The result was checked by using the quadratic relationship for the smooth slopes, in which pastures and sugarcane predominate, and higher values for V occur (Figure 5B).Thus, in general, areas with smooth slopes tend to accumulate sediments and nutrients that flow from upstream areas (Blanco and Lal, 2008).
Figure 5. A) Relationship between the sum of exchangeable bases and the slope gradient of the sampled spot; and B) the relationship between the saturation by bases (V) and the synergistic effect of the use plus slope gradient at the sampled spot.
Area soils have a medium clayed texture.The variable clay in the soil is explained by the slope linear model (ΔAICc = 0; WAICc = 0.27) (Figure 6A), with a percentage of pasture (ΔAICc = 0.6; WAICc = 0.20) (Figure 6B), followed by the shape of the landscape index AWMPFD (ΔAICc = 1.3;WAICc = 0.14) (Figure 6C) and, at last, the synergistic effect of the average slope gradient with the percentage of pasture in the landscape (ΔAICc = 1.8;WAICc = 0.11) (Figure 6D).In the linear relationship, it is very clear that the smoother the slope is, the higher the quantity of clay in the soil.According to the quadratic relation, the higher the percentage of pasture is, the higher the quantity of clay. Figure 6.A) Relationship between the clay content in the soil and the landscape average slope gradient; B) the relationship between the clay content in the soil and the average slope and the percentage of native forest areas in the landscape; C) the relationship between the clay content in the soil and the shape of landscape index AWMPFD; and D) the relationship between the clay content of the soil and the synergistic effect of the average slope and the percentage of pasture in the landscape.
When compared to other agricultural uses, the average use of pasture leads to less soil structural degradation, probably because plowing and harrowing are not used.In addition, gramineous plants act as fixing structures through the action of their roots, and they obstruct the direct impact of eroding rains, with tufts of grass acting as barriers to runoff and fixing small particles in the soil.
In addition, the model indicated that larger AWMPFD values were associated with a higher quantity of clay in the soil.This parameter indicates the complexity of native forest shapes and may be associated with the borders of forest fragments with irregular perimeters, which act as barriers, retaining particles carried from other areas.Finally, stronger slopes with a high percentage of native forest, or smooth slopes with smaller areas of native forest, presented larger quantities of clay in the soil according to the quadratic relationship.From this relationship, native forest was observed to compensate for the high slope, and the low slope compensates for the small quantity of native forest in clay retention.Therefore, in the native forest areas, OM works as a cement in the aggregation of soil particles in addition to reducing the splash effect of the raindrops and the surface runoff.In addition, the areas with stronger slopes are natural impediments to agricultural use due to the difficulties of mechanization, in addition to being PPA in many cases (Brasil, 2012).These scenarios explain the larger quantity of native forests along the strongest slopes and the quadratic relationship.
The contents of larger grains in the soil (G > 0.25 mm) were expressed by integrating the explanatory model, in the first place, by using the shape of the native areas, the AWMPFD index (ΔAICc = 0; WAICc = 0.26), followed by the landscape average slope gradient (ΔAICc = 0.7; WAICc = 0.17) and the synergistic effect of the average slope gradient, which were added to the pasture percentage (ΔAICc = 1.7;WAICc = 0.11).From this linear relationship, the simpler and the more circular the shape of the native forest fragments are, the larger the soil granulometry content > 0.25 mm (Figure 7A).In the soils under native forest fragments, it is important to note that they were cleared and downgraded by deforestation and cattle pastures for a long time, which negatively influenced their structural and granulometric attributes.However, most of the fragments are small, with those with irregular perimeters tending to increase the retention of the finest particles in their borders in relation to the more regular borders.In general, these borders are more compact, which reduces the interception of the flow and consequently leads to a lower accumulation of carried sediments with larger quantities of particles < 0.25 mm (Blanco and Lal, 2008).Thus, according to the linear relationship, the stronger the slope is, the higher the larger grain content in the soil (Figure 7B). Figure 7. A) Relationship between the larger grain contents in the soil > 0.25 mm and the index for the shape of the native forest areas (AWMPFD) in the landscape; and B) the relationship between the finer grain contents in the soil < 0.25 mm and the average slope in the landscape.
The increase in the relative concentration of larger particles is due to the larger particles being heavier than the smaller ones of the same composition, and, in general, they were formed by minerals that were more resistant to weathering, such as quartz, which is more stable and barely solubilized and removed from the profile.
The explanatory models of the small grain (< 0.25 mm) contents were also linear.The model that better explains this attribute is integrated by the synergistic effect between the landscape average slope, the percentage of native forest and the MQI (ΔAICc = 0; WAICc = 0.28) (Figure 8A).In addition, the second and third-ranked models (Figure 8B) were integrated by using the synergistic effect between the landscape average slope gradient and the percent of pasture (ΔAICc = 0.5; WAICc = 0.22) and by using the landscape average slope gradient (ΔAICc = 0.7; WAICc = 0.20).Thus, according to the quadratic relationship, the stronger the slope and the percent of native forest and the smaller the MQI, the higher the content of grains smaller than 0.25 mm will be in the soil.As observed in the field, stronger slopes are less frequently used for temporary crops (corn, beans, sugarcane, etc.) because they present the worst MQI and the most intense surface flow.Thus, in the areas with stronger slopes, the native forest and permanent crops predominate (coffee and pastures) because they are less susceptible to erosion, resulting in a larger retention of smaller grains due to the chelating and adsorbent function of the OM and the greater binding power between particles resulting from the ions present in the clays.
Figure 8. A) Relationship between the contents of finer grains in the soil < 0.25 mm and the synergistic effect of the moderate slope with the native forest in the landscape and the soil quality index; and B) the relationship between the contents of finer grains in the soil < 0.25 mm and the synergistic effect of the moderate slope and the pasture percentage in the landscape.
Regarding the clays, it is important to note that they directly influence the infiltration rate, the retention of humidity in the soil and the porosity, among other features.In this way, the increase in the quantity of grains larger than 0.25 mm may reduce the humidity retention of the soils, increasing the demand for irrigation, which intensifies the process of concentrating particles larger than 0.25 mm over the long term.This concentration increases the need to add bases to the plantation in a downgrading vicious cycle that might result in the loss of the productive capacity of the soil (Blanco and Lal, 2008).
In accordance with the models (Table 4), the influence of the relief group is remarkable when considering the gradient of the slope in the conditioning of the soil attributes.It is followed by the quantity and shape of native forest fragments and by the management type and quality (MQI).Thus, the variation of chemical attributes such as the BS, V and pH are explained primarily by the use and management characteristics of the sampled areas, and they largely result from the additions to and correction of the soil.With respect to the soil physical attributes, such as the size of the structures and the clay content, among other, it especially results from the interaction among landscape elements, which makes sustaining the quality of these attributes difficult.To maintain the ecosystem and productive services of these soils, more regional measures are necessary so that pedogenic and morphogenic cycles changes are limited, and the ecodynamic and geological landscape balance is preserved.
The hydrosedimentological dynamics of the area is associated with the history of the anthropic occupation of the region.This history was based on the extensive exploration of the area's natural resources, as illustrated by the municipality's deforestation, in which only 8% of the native forests remain (Ribeiro et al., 2011).This native forest is dispersed in fragments that have been badly downgraded and has experienced border effects resulting in the loss of ecosystem functions.These areas, which favored planting and mechanization, were gradually occupied, with native forests being restricted to areas considered less appropriate for agricultural activities.Under this scenario, the agricultural areas were subjected to a management system that employs plowing, harrowing, scarification, mechanized seeding, over-dosing with agrochemicals, continuous pasturing without fertilizing, intensive irrigation practices and the absence of conservationist management practices and PPA and LR (Ayer et al., 2015).These practices have changed the soil texture and caused an increase in the concentration of grains larger than 0.25 mm.They have reduced the OM content and the size of the soil structures, changing the geochemical dynamics of the landscape and impairing the hydro-and pedological resources.It is notable that the mechanical seeding techniques as well as the use of agrochemicals represent a gain when dealing with the soil and agro-silvi-pastoral production.
However, its inappropriate use could be harmful to the agropastoral productivity and to the environment.
Considering the above findings, it is necessary to adopt management practices that reduce the slope and soil use effects on erosion and the attributes of dystrophic red oxisols and also cause the hydrosedimentological imbalance of the landscape.The use of conservationist practices is recommended for this purpose (Ayer et al., 2015).For those authors, the adoption of conservationist management approaches as an alternative to traditional ones would greatly assist in the reduction of erosion rates.In calculating the erosion associated with traditional management and soil loss under conservationist management, it was concluded that the employment of conservationist practices would reduce the erosion rate from 8.40 to 2.84 Mg ha -1 year -1 and reduce the area experiencing erosion above the SLT from 34.8 to 4.0% of the total area.
In addition to the absence of PPA and LR, which have environmental roles in preserving water resources, the landscape, the geological stability, biodiversity and fauna and flora gene flow, they also contribute to the well-being of the population, which is remarkable (Brasil, 2012).This area is second to Blanco and Lal (2008), which are areas with larger quantities of native forests that undergo fewer attacks from insects and have higher average productivity.Bearing in mind that the increase in productivity is related to pollination by insects, the attenuation of the erosion effect was caused by the surface runoff, increased OM and humidity and improvements in soil structure, in addition to other factors.However, these ecosystem services that are provided by the fauna and flora in the region are being impaired by both deforestation and the traditional techniques of agropastoral activities, and by the use of pesticides without observing any control or regulation (Londres, 2011;Servidoni et al., 2016).
Thus, management improvements, PPA and LR preservation and reforestation, and the control of agrochemical use could help improve the ecodynamic balance of the area over the long term, with improvements in the average productivity per hectare and reductions in the costs of irrigation and agrochemicals.To disseminate these practices, credit lines should be created or promoted for properties that use conservationist managements and/or preserve PPA and LR areas, as shown by Zolin et al. (2011).Additionally, ecological tracking stamps could be created, which would add value to the products (Conceição and Barros, 2005).The manpower qualification and awareness and fiscalization programs for rural workers and for society should generally be more effective and occur more frequently.

Conclusions
Physical attributes exhibit a direct relation to the configuration and structure of landscapes, and chemical ones result primarily from soil use conditions.
Slopes have an essential role, either because of their direct effect on the infiltration rate and/or the surface runoff rate, affecting the removal, transportation and deposition of soil particles, or because their indirect affect impairs land use and mechanization in areas with irregular and steep reliefs.
The lack of native forest and permanent preservation areas results in changes to the hydrosedimentological cycle, in which smaller particles are removed from the soil and carried to bodies of water, increasing the concentration of the sand size grains on the Latosols.
The uses and management practices were considered factors that negatively change the landscape dynamics, organic matter contents, texture, and soil structure.

Figure 1 .
Figure 1.Map of the Alfenas municipality location and the 23 sampled soil spots.

Figure 2 .
Figure 2. A) Slope map and B) Soil use and management map in Alfenas -MG.
nond = nondimensional; Slope = Slope gradient; Slop s = Slope at the spot; Use s = Use and management at the spot; Shannon = Shannon Diversity Index; AWM = Shape Index; NV = Native forest; FP = Floodplain; Urb = Urban area; Temp = temporary crop area; Perm = permanent crop area; and MQI= Management quality index.

Figure 4 .
Figure 4. A) Relationship between organic matter (OM) and the slope gradient of the sampled area; and B) the relationship between the pH and use in the sampled area for the synergistic effect with the slope gradient of the sampled area.

Table 1 .
"C" factors for the uses and management types identified in the study area and employed for the Landscape Quality Index Calculation.

Table 2 .
Landscape independent variable composition and relief as calculated to build the regression models.

Table 3 .
Soil attributes (dependent variables) used to generate the regression.Models of the 23 sampled areas.

Table 4 .
Nonlinear and linear explanatory models of soil variables in relation to the landscape