Characterization of Physical-Chemical and Structural Soil Attributes in the Semiarid Region of the Rio Grande do Norte State , Brazil

The study of the soil characterization and the relation of its attributes allows a systematic proposal of the local particularities, leading to adequate practices for maintenance and/or preservation of its productive capacity. In this sense, the aim of this study was to evaluate the influence of structural attributes in association with physical and chemical soil classes, using the multivariate statistical techniques to differentiate environments. The research was carried out in the Moacir Lucena Project, located in the municipality of Apodi, RN, Brazil. Three representative environments were chosen as follows: Profile 1 (P1)-Red-yellow Latosol-Area in recovery (1AR), P2-Haplic Cambisol-Lake Area, (2AL) and P3-Eutrophic Yellow Latosol-Cashew Tree Area (3AC). The soil samples were collected in the horizons of the studied areas. Ten (10) samples were collected per horizons in volumetric rings and in soil blocks (aggregate analysis), which resulted in triplicates in the laboratory. Structural, physical and chemical attributes were evaluated. The data were analyzed using multivariate statistical techniques, with correlation matrix, clustering analysis and factorial analysis performed by the extraction of the factors into principal components. The use of clustering analysis allowed the formation of four groups for soil classes and attributes; the inorganic fractions were determinant for environmental differentiation, where the sand was discriminant for the Red-yellow Latosol and the Eutrophic Yellow Latosol, and the clay and silt for the Haplic Cambisol. Higher similarity was observed in the transition horizons of the Latosols Class. The physical and structural attributes were determinant in the dissimilarity for the Haplic Cambisol, reflecting in physical restrictions to the plant growth. The factor analysis revealed that the variables particle density (Dp), Ca, Mg, sum of bases (SB) and cation exchange capacity (CEC) for factor 1, followed by pH, P, K, total Sand, Clay and soil density (Ds) for factor 2 are important soil attributes to distinguish the studied environments.


Introduction
The Brazilian semiarid region presents variability in terms of lithology, climatic pattern, landscape and vegetation, resulting in peculiar characteristics of soil attributes (morphological, physical, chemical and mineralogical), and consequently its structural properties, being considered as one of the most important characteristic from the agricultural point of view, by indirectly influencing plant growth factors, since it controls air changes, water and heat, nutrient availability, and mechanical resistance to root development (Carvalho et al., 2017).
Since the soil structure is a dynamic and functional property, easily modified by anthropic action, its study is necessary as an aid to decision making, regarding the soil management practices and the choice of the agricultural crops, taking into consideration their potentialities and/or restrictions.
The activity of irrigated fruit growing became one of the prominent segments in the semiarid areas of the West of the Rio Grande Norte State, Brazil, in order to meet the demand for fruit exports to international markets and the increasing incorporation of new areas.It is important to take into consideration that the anthropic action in these environments by irrigated agriculture has led to soil degradation, reflecting in the reduction of productive capacity, since it is being used improperly and without observing the basic principles of soil conservation, not evaluating the local particularities, breaking the soil dynamics, avoiding the soil properly use, unbalancing the biodiversity and socially impacting the local farmers (Conceição et al., 2014).
It is important to notice that the studied areas are part of the Moacir Lucena Settlement Project and that the locals adopt natural resource conservation systems and are being invaded by agribusiness activities, which reflects negatively on the preservation of biodiversity and of every form of life.Thus, studies of the characterization of the structural attributes associated with physical and chemical resources are scarce in the region.In this perspective, it is necessary to study the environment characteristics, the use of soil attributes and their relations, with the aid of multivariate techniques as an efficient tool in distinguishing the most sensitive attributes.
Thus, the aim of the present study was to characterize the environments, through the study of the structural and chemical attributes in association with the inorganic fractions, aiming the separation of the environments through the most sensitive attributes, using the technique of multivariate analysis in soil classes in the Moacir Lucena settlement, at the Rio Grande do Norte State mesoregion, Brazil.
The 1AR refers to an area without planting, within the limits of the Permanent Preservation Area (APP) with approximately 2.5 hectares, where the settlers planted cotton, but has been in rest for 16 years to recover the native forest and soil.The 2AL is located within the boundaries of the APP area with an area around 4 to 5 hectares where, due to a difference of slope, present flooding in rainy periods.The 3CAJU area was previously used to plant cotton, beans and other crops and was inserted in the collective area, in 2007, with 20 hectares.Cashew trees were planted and at the beginning of winter beans are planted as an associate crop, with pruning and other traits on every year.
In each of the three studied environments, trench openings were carried out aiming at the description of the soil classes, collecting samples with deformed and not deformed structure in the soil horizons.The samples were placed in plastic bags and analyzed in the Laboratory of Soil, Water and Plant analysis of the Center of Agricultural Sciences of the Federal Rural Semiarid University (LASAP-CCA-UFERSA), following the methodology proposed by Teixeira et al. (2017).Inorganic fractions were determined: texture (sand, silt and clay), as well as particle density (dp) and the consistency limits: Liquidity Limit (LL); Plasticity Limit (LP); Plasticity index (IP) and the soil structural atributes: density, macroporosity, microporosity, total porosity, volumetric humidity and aggregate stability (Average Weighted Diameter-DMP, Geometric Mean Diameter-DMG).The evaluated chemical attributes were: electrical conductivity (EC), pH in water, electrical conductivity of soil saturation extract (EC), exchangeable sodium percentage (PST), total organic carbon (TOC), phosphorus (P), potassium (K + ), calcium (Ca 2+ ), magnesium (Mg 2+ ), cation exchange capacity (CEC) and base saturation (V).
The granulometry was obtained using chemical dispersant (sodium hexametaphosphate) and distilled water in 20 g of the air dried fine soil (TFSA), with slow mechanical agitation (Wagner 50 rpm) for 16 hours, the sand fraction (2 to 0.05 mm) was quantified by sieving, the clay fraction (< 0.002 mm) by sedimentation and the silt (0.05 to 0.002 mm) by the difference between the total sand and clay fractions.The particle density analysis was carried out using the volumetric flask method, using the oven-dried fine soil (TFSE) at 105 °C and ethyl alcohol.
Soil consistency tests were determined based on liquidity limits (LL) using the Casagrande apparatus.The plasticity limit (LP) was determined with three replicates and the representative sample was obtained from the central part of the soil shear in the metallic sphere of the equipment, being derived from the determination of the liquidity limit and forming a sphere, which was compressed on a plate to form a cylindrical rod with 3.0 to 4.0 mm in diameter, without breaking or flowing.The plasticity index (IP) was obtained by the difference between LL and LP.Gravimetric moisture was determined in the plasticity condition for the soil rods.
The intact samples were collected using the Uhland apparatus and rings with dimensions of 0.05 m in height and 0.05 m in diameter, with 10 (ten) samples per horizon in the respective profiles, for evaluation of the soil density attributes, macro and microporosity, total and determined porosity.Soil density was determined by the volumetric ring method, described by Forsythe (1975), with stablished volume and with the mean of the obtained values, being represented by the mass quotient of the soil solid particles by the total soil volume, and expressed in kg dm -3 .
For the analysis of macroporosity, microporosity and total porosity, the undeformed samples in the volumetric rings were saturated for 48 hours and weighed (to determine the total porosity).The method used to determine these properties was the "tension table" (Kiehl, 1979), at the 6 kPa (microporosity) stress, with the total soil porosity (Pts) determined by measuring the samples saturation humidity, according to the equation: Pt = [(Msat -Ms)/VT] × 100, where Msat is the mass of the soil in the saturation condition; Ms is the mass of the dry soil and VT is the volume of the sample.The microporosity was determined by measuring the retained water content in the soil, for 'h' of 60 cm of water (approximately a pore radius of 25 μm).Soil macroporosity was determined by the difference between total porosity and soil microporosity.
For the aggregates study, block of the soil profiles were extracted in the respective horizons and in sieves with mesh opening of 4.00 and 2.00 mm (together).The wet sieving method was used with mesh sieves of 4.76; 2.00; 1.00; 0.50 and 0.25 in the vertical oscillation apparatus (Kemper & Rosenau, 1986).After separation of the aggregates obtained by shaking in water using the vertical oscillation apparatus (42 oscillations/minute), the samples were dried at 105 °C.After obtaining the dry mass, the sand content was discarded and the distribution of the aggregate size, percentage of aggregation and water-stable aggregates and DMP and DMG were obtained.
In order to perform the chemical analyzes, the following parameters were used: pH in water, EC in water, PST, TOC by the digestion of organic matter and macronutrients: exchangeable calcium (Ca 2+ ) and exchangeable magnesium (Mg 2+ ) with potassium chloride extractor, phosphorus (P), sodium (Na + ) and potassium (K + ), with the Mehlich extractor 1, being calculated the CEC, V and PST, evaluated according to the Manual of Recommendations for the use of correctives and fertilizers from Minas Gerais State, Brazil (Ribeiro & Guimarães, 1999).
The data of the deformed structure attributes were expressed using four replications and submitted to statistical analysis using the multivariate analysis technique, as the main tool used to detect the most sensitive attributes in the distinction of soil environments under different uses, with the aid of the Statistica 7.0 Software (StatiSoft, 2004).
The analytical results were standardized by the Pearson's correlation matrix and submitted to multivariate techniques such as cluster analysis (AA), factorial analysis (FA) and principal component analysis (PCA).
Pearson's correlation analysis (p ≤ 0.05) was used for the 28 variables in order to guarantee the sufficient minimum correlations to justify their use in FA data matrix.For FA, factors with eigenvalues above 1 were extracted by principal components and the factorial axes were rotated by the Varimax method.For this study the value of 0.65 was established for significant factorial weights (Hair Jr. et al., 2009).
The cluster analysis (AA) was represented by the vertical dendogram of the distance matrix.The Euclidean distance was adopted as a measure of similarity and Ward's method for linking cases.In AA, the importance of each variable in the distinction of the environments is measured as a function of its smaller distance from the reference axis, the x or y axis, and the axis that contains the largest value of accumulated variance explains part of the causes of variation (Sá Paye et al., 2012).
Assigning the value 0.1 for the binding distance in the dendrogram for the environments and 2 for the attributes obtained from the cluster analysis, it was possible to identify four groups for both (Figure 1A and B).
In FA, factors with eigenvalues above 1.0 were extracted by the principal components, and the factorial axes were rotated by the Varimax method.The value of 0.65 for significant factor loads was established for this study (Hair Jr. et al., 2009).
In FA the contribution of each variable and to each factor was observed, being the factors defined by the differentiating attributes of the environments.As a tool for distinguishing environments and their different uses, six diagrams of the principal components (Factor 1 and 2, Factor 3 and 4, Factor 5 and 6) were performed for the physical-chemical attributes.From these data, two-dimensional diagrams were created to distinguish the areas and diagrams of vector projection to verify the soil attributes that distinguished most in the evaluated areas.
As described by Souza (2018), the association of soil physical and chemical attributes in different uses in agroecosystems evidences the correlation between the variables using the joint analysis with numerous factors and characteristics, being possible to explain the variability of the original set of variables, distinguishing environments through the most sensitive attributes.
All statistical analysis were performed using the STATISTICA software version 7.0 (StatSoft, 2004).
After the standardization of the data by the correlation matrix, the cluster analysis was performed, represented by the dendrogram (Figure 1), with significant variation in the Euclidean distance values between the areas for considered variables.With the analysis of grouping of the analytical results, four groups were formed for the soil classes and for the attributes of the studied soils, showing greater similarity between them.
From the selected attributes, a first classification of the areas was performed according to the similarity between the studied environments.The first group (G1) was defined with reference to the surface horizon of each profile (1AR.A, 2AL.A, and 3CAJU.Ap), which indicates that the soil intrinsic characteristics was a determinant factor.This can be justified as a function of soil texture, where the sand variable presented higher concentration in G1, presented in considerable amounts on the surface of the three evaluated environments (Figure 1A).The 3CAJU (Latosol) presented greater dissimilarity in G1 when compared to 1AR.A and 2AL.A, with a greater percentage of the sand fraction (813 g cm -3 ) when compared to other areas.The physical attributes were: AGRI (%) > 2 mm, DMP, DMG, PST, Macrop and LP, being in agreement with the correlation values between the variables presented in the matrix (Table 1) and in the dendrogram (Figure 2A).
Source: The authors.
The factors F1 and F2 explained 65.89% of the total variation of the attributes (physical, chemical and structural), being the variables that most characterize the studied soil classes (Table 2).F1 explained 38.04% of the total variation, and the highest correlation coefficients (≥ |0.65|) were identified for the variables Dp, Ca 2+ , Mg 2+ , SB and CEC, according to Figure 2. The variable Dp, which refers to solids (soil matrix), was highlighted by the fact that the soils under study are considered as minerals, ranging from 2.51 to 3.04 g.cm-3, associated with chemical and structural characteristics (F1 to F6).For the chemical attributes Ca 2+ , Mg 2+ , SB and CEC, the representativeness in F1 was for the Cambisol (P2) as a function of the mineralogical characteristics of the clays from the source material of the Jandaíra limestone (Dnocs, 1978)

Table 1 .
Correlation of soil atributes variables of the evaluated areas at the Moacir Lucena Settlement, Rio Grande do Norte, Brazil

Table 2 .
Factorial axes for the soil atributes and their respective factorial weights, eigenvalues, total and accumulated variance . According toRonquim et al., (2010), the CEC of these soils is occupied by essential cations such as Ca 2+ , Mg 2+ with predominantly high activity of clay, ilite