Physical Indicators of Cambisols Under Agricultural Uses in Chapada do Apodi, Semiarid Region of Brazil

Soil physical structure is related to porous space dynamics, which is affected by pedogenetic conditions, land uses, and agricultural practices. Thus, the objective of this work was to evaluate physical and structural attributes of Cambissolos under different uses in the Terra de Esperança Settlement Project, in Chapada do Apodi, Governor Dix-Sept Rosado, Rio Grande do Norte, Brazil, and detect the most sensitive attributes for the distinction of environments using multivariate analysis. The study areas with different land uses were: Native Forest Area 1 (NFA1), Native Forest Area 2 (NFA2), Native Forest Area 3 (NFA3) (reference areas), Conventional Management Area (CMA) Agroecological Area (AEA), and Cajaraneira (Spondia sp.) Orchard Area (COA). Areas with agricultural uses were characterized through physical and structural analyses, using disturbed and undisturbed soil samples collected from their 0.00-0.10, 0.10-0.20, and 0.20-0.30 m soil layers. The soil classes of the areas, according to the Brazilian Soil Classification System (SiBCS) were Cambissolo Haplico Carbonatico vertissolico (NFA1); Cambissolo Haplico Ta Eutrofico tipico (NFA2 and COA); Cambissolo Haplico Ta Eutrofico vertissolico (NFA3); and Cambissolo Haplico Carbonático tipico (CMA and AEA). The results of the attributes analyzed were expressed as mean of three replications per soil layer of each area, using multivariate analysis. Soil textures varied from sandy clay loam to clay. The total sand fraction presented negative correlations with clay dispersed in water, gravimetric moisture (GM), volumetric moisture (VM), total porosity determined (TPd) and microporosity (MiP); and positive correlations with soil density (SD), and basic infiltration rate (BIR), denoting pedogenetic influence. The SD presented significant correlation with GM, VM, TPd, MiP, macroporosity (MaP), aeration porosity (AP) and BIR, denoting its importance for the physical structure of the soil, and its dynamics. The most relevant attributes for the discrimination of the soil physical structure were the inorganic fractions clay and sand, porosity, degree of flocculation, aggregates, and soil mechanical resistance to penetration. The physical and structural attributes of the Cambissolos Haplicos were generally preserved, when compared to the current conditions of the NFAs, despite the different land use and managements. However, the soils of NFA1 (0.20-0.30 m layer), CMA, and AEA areas indicate fragility in aggregate stability and degree of flocculation due to the predominance of the sand fraction. The COA presented more favorable physical and structural conditions to the development of agricultural crops, especially on the surface layers, mostly influenced by their clay, TPd, AP, GM, VM, and soil mechanical resistance to penetration.


Introduction
Soil dynamics are affected by pedogenetic conditions, land uses, and managements due to the traffic of machines and implements used in conventional soil preparation for crops (Oliveira et al., 2013). Physical and structural attributes of adequate soils allow processes of water infiltration, retention, and availability to plants, and gas and heat exchanges with the atmosphere and roots of plants; these soils respond to management and resist to degradation, providing adequate conditions for the growth and development of plants (Reichert et al., 2003). The performance of these soils and the maintenance of their productive capacity cannot be directly measured, however, it can be estimated using some physical attributes, which are used as indicators when they are sensitive Six sites were chosen for the research (Figure 1):  Native Forest 1, 2, and 3 (NFA1, NFA2, and NFA3): environmental reserve areas with preserved hyper-xerophilous Caatinga vegetation, which includes deciduous species, favoring the input of organic matter; this area had some wood removed for fencing of other areas. In the dry periods, these areas were used for grazing animals (goats) raised by the settlers. These areas were used as reference areas because of the reduced anthropogenic influence, to compare changes in soil attributes, simulating natural environment conditions.
 Conventional Management Area (CMA): area with maize, string beans, sesame, pumpkin, and sorghum crops grown in the rainy season, with no irrigation. No burning practices and no chemical fertilizers had been applied in this area since 2005, however, soil management were conducted using plowing and harrowing. This area was fallow during the study period.
 Agroecological area (AEA): area established in 2005 to produce fruits and forage for the subsistence of the family farmers, and animals, and had production of honey by Africanized bees. Some agroecological soil conservation practices were used in this area, such as: absence of burning practices, preservation of natural succession of the plants (using fruit species that are adapted to the semiarid), and control of erosive processes by using ridges that soften the surface runoff.
 Cajaraneira (Spondia sp.) orchard area (COA): area with cajaraneiras, a fruit species of the genus Spondia, which were planted by the former owner in the 1970s. This species had great economic importance for the settlers; it produces about 2.8 to 3.2 Mg of fruits per week at the peak harvest. This area had great contribution of organic matter to the soil during the dry season due to the leaf loss the of the Spondia sp.

Soil Sampling
Disturbed and undisturbed soil samples from the 0.00-0.10, 0.10-0.20, and 0.20-0.30 m layers were collected in the areas considering 1 ha for each area, and taken to the Soil Physics for Analysis of Soil, Water, and Plant of the Agricultural Sciences Center of the Federal Rural University of the Semi-arid Region. Five composed samples of disturbed soil were formed from 15 subsamples that were collected using a Dutch auger and packed in identified plastic bags. These samples were air dried, disaggregated and passed through a 2 mm mesh sieve to obtain the air-dried fine earth (TFSA). Undisturbed samples were collected by opening four small random trenches in each area; two samples were collected in each of the three soil layers, using the volumetric ring method (rings of 5.0 cm in height and 5.0 cm in diameter), totaling 144 samples, and one undisturbed block by trench (04) per area (06) per layer (03), totaling 72 samples, which were packed in plastic bags.

Physical Indicators
The results of the physical and chemical analyzes were expressed as means of three replications. Disturbed samples were analyzed for particle size, particle density, dispersed clay in water, degree of flocculation, and silt to clay ratio; undisturbed samples were analyzed for soil density, macroporosity, microporosity (tension table at 6 kPa), total porosity determined, soil aggregate, and stability of the undisturbed soil block to determine its weighted mean diameter (WMD), and geometric mean diameter (GMD), and soil water content (gravimetric, and volumetric moistures), according to Teixeira et al. (2017).
The soil granulometry was evaluated by the pipette method using chemical dispersant (sodium hexametaphosphate) and distilled water in 20 g of the air-dried fine earth under slow mechanical agitation on a shaker (Wagner 50 rpm) for 16 hours. The sand fraction (2 to 0.05 mm) was determined by sieving, clay (< 0.002 mm) by sedimentation, and silt (0.05 to 0.002 mm) by the difference between the sand and clay fractions; the silt to clay ratio was also calculated. Particle density analysis was performed using the volumetric flask method using the air-dried fine earth in an oven at 105 °C, and ethyl alcohol.
Dispersed clay in water was determined by slow mechanical dispersion in water with shaking for 16 h at 50 rpm, followed by separation of the clay fraction by sedimentation of the silt. The degree of flocculation was obtained by the naturally dispersed clay to total clay ratio.
Soil density was determined by the volumetric ring method, using a ring with known volume, with means represented by the quotient of the soil solid particle weight by the total soil volume.
Undisturbed samples were saturated for 48 hours and weighed to determine total porosity. Subsequently, they were placed in a tension table saturated with water and without air bubbles in the column at tension of 6 kPa to determine microporosity. The soil macroporosity was determined by the difference between total porosity and microporosity.
The soil aggregate distribution was determined by the wet sieving technique. The samples were sieved in 4.76, 2.00, 1.00, 0.50, 0.25, 0.105, and 0.053 mm mesh sieves, immersed in water and shaken in a mechanical oscillator for 15 min. Soil aggregates were separated into size classes of 4.76 to 2.0, 2.0 to 1.0, 1.0 to 0.50, 0.50 to 0.25 mm, and < 0.25 mm. The WMD, and GMD of soil aggregates were calculated based on these results.
Soil mechanical resistance to penetration (SRP) was determined using a penetrometer (SoloStar PLG 5500, Falker) in 15 points of each study area. This device has an automatic measuring system, cone diameter of 12.83 mm, resolution of 0.02 MPa; maximum of 90 kgf supported by the rod, and maximum depth of 40 cm, complying with the ASAE S.313.3 standards (ASAE, 2004), and simultaneous use of GPSMAP (Garmin 64s) for georeferencing, considering the mean SRP in the 0.0-0.10 m, 0.10-0.20, and 0.20-0.30 m soil layers.
Soil water infiltration was evaluated by the ring infiltrator method with three replications per study site, using two concentric cylinders-height of 40 cm and diameters of 30 cm, and 50 cm for the internal and external rings, respectively-with manual water supply for the cylinders, and readings of the water height (cm) at 0, 1, 2, 3, 4, 5, 10, 20, 30, 45, 60, 90, and 120 minutes, or up to the equilibrium point to determine the basic infiltration rate (BIR) of water into the soil, according to Bernardo et al. (2008).
Deformed samples were collected from the 0.00-0.10, 0.10-0.20, and 0.20-0.30 m soil layers to perform gravimetric moisture (GM) analysis, since SRP and BIR are dependent on cohesion and adhesion forces, according Teixeira et al. (2017).

Statistical Analysis
Data of physical attributes were expressed by averages of three replications per layer and subjected to multivariate statistical analysis techniques to detect the most sensitive attributes for the distinction of soil environments under different uses, in the Statistica 7.0 program ( Reference ranges of the most important variables of factorial analysis were used to interpret the results, according to Kiehl (1979), Arshad et al. (1996), Prevedello (1996), Reynolds et al. (2002), Reichert et al. (2003), Pereira et al. (2010), Ferreira (2010), and Prado (2013). The soils were classified according to their physical performance degree as good, regular, or poor.

Soil Physical Indicators
The textural classification of Cambissolos Haplicos (Table 1) showed sandy clay loam to clay soils, with granulometric composition with predominance of the sand fraction for the Native Forest Area 1 in the third soil layer (0.20-0.30 m) (NFA1), Conventional Management (CMA), and Agroecological (AEA) areas, presenting sand contents of 518 to 685 g kg -1 , characterizing them as sandy clay loam texture, denoting primary minerals that are more resistant to the weathering process. Marinho et al. (2016) evaluated organic matter, and physicochemical attributes of a Cambissolo under different agricultural uses in the Chapada do Apodi microregion and found predominance of the sand fraction in areas with native forest, collective intercrops under conventional soil management, and in an agroecological area, with similar results, ranging from 426.1 to 660.5 g kg -1 ( Table 1).
The soil of the NFA1-3, CMA, and AEA presented the highest particle densities due to the predominance of the quartz mineral in the solid fraction, with a density of 2.65 g cm -3 ; and the lowest total porosity, microporosity, and aggregate stability because solid particles are less susceptible to aggregation in coarser soils. Soil density is usually higher in sandy than in clayey soils; sandy soils have fewer micropores (internal pores of the aggregates) and thus, lower total porosity (Brady & Weil, 2013), mechanical resistance to penetration, and gravimetric, and jas.ccsenet.org Journal of Agricultural Science Vol. 12, No. 6; 2020 volumetric moistures. However, the soils presented higher rates of basic infiltration, especially the CMA and AEA, due to their homogeneity and predominance of the sand fraction in surface and subsurface layers (Table 1).  Note. CDW = clay dispersed in water; DF = degree of flocculation; Dp = particle density; SD = soil density; GM = gravimetric moisture; VM = volumetric moisture; TPd = total porosity determined; MiP = microporosity; MaP = macroporosity; AP = aeration porosity; WMD = weighted mean diameter; GMD = geometric mean diameter; SRP = soil mechanical resistance to penetration; BIR = basic infiltration rate of water into the soil.

Statistical Analysis
The correlation matrix between the physical variables (Table 2) showed that the sand fraction had a negative correlation with clay dispersed in water (CDW), GM, VM, TPd, and MiP, and positive correlation with SD and BIR. SD presented a high correlation with GM, VM, TPd, MiP, MaP, AP, and BIR, denoting the dynamics of the soil physical structure.  Note. CDW = clay dispersed in water; DF = degree of flocculation; Dp = particle density; SD = soil density; GM = gravimetric moisture; VM = volumetric moisture; TPd = total porosity determined; MiP = microporosity; MaP = macroporosity; AP = aeration porosity; WMD = weighted mean diameter; GMD = geometric mean diameter; SRP = soil mechanical resistance to penetration; BIR = basic infiltration speed of water into the soil.
The principal component, and factorial analyses were performed in the data matrix consisting of 16 variables, requiring the removal of the microporosity due to multicollinearity. Table 3 shows the factorial loads after rotation of the data of soil physical indicators. The eigenvalues indicate the relative importance of each factor in the explanation of the variance of the set of attributes analyzed, showing factors in order of significance-significant factor loads with opposite signs denote variation in opposite direction (Arcoverde et al., 2015).
Then, factors 1, 2, 3 and 4 were selected because they met the criterion of eigenvalues above 1. These factors accounted for 79.19 % of the variance (Table 3).
Factor 1 (soil porosity) explained the largest part of total variance of the data, i.e., had the greatest influence. It consisted of the clay content, CDW, TPd, AP, VM, GM, varying in opposite direction to the sand content and SD, explaining 45.42 % of the total data variance. Factor 2 (degree of flocculation; DF) consisted of DF and MaP, explaining 14.01 % of the total variance of the data. Factor 3 (Aggregate) consisted of the GMD, and Factor 4 consisted of the SRP, explaining 11.29 %, and 8.47 % of the total variance of the data, respectively (Table 3). Note. Factorial loads ≥ 0.65 were significant.
The principal component analysis (PCA) showed a graphic representation of the distribution of variables in the unitary correlation circle (Figure 2a), and the distribution of the cloud of points representing the relationship between Factors 1 and 2 of the study areas (Figure 2b).
Practically all variables were near the unit circle, indicating greater contribution of the principal components to the more distant variables (Figure 2a). The inorganic fractions (sand, silt, and clay) were not close to the circle of correlations because the studied soils presented different textures despite their same classification up to the second categorical level (SiBCS), and the graph shows the predominance of the variables that better discriminated the environments, such as sand and clay. The graphs denoted the interrelationship between the variables and the predominant characteristics of each study area (Figure 2). he description e on the PCA m) areas, differ nd content diff osed mostly by ement), and c e 1).
TPd, GM and yer, representin o (Figure 2b), by their redu hich contribute osited to the so f the soil mois ayer (Table 1)      According to the reference ranges of the ten physical indicators in the performance groups (Table 4), the three layers of all Cambissolos Haplicos under different land use and managements were classified as good, despite the soils of the NFA1 (0.20-0.30 m), CMA, and AEA presented the most altered attributes. This can be attributed to the pedogenetic conditions, which influenced their medium texture, with predominance of the sand fraction (606 to 691 g kg -1 > 600 to 800 g kg -1 , regular) and, consequently, affected the attributes Dp, SD and BIR (PCA), indicating low aggregate stability (GMD = 0.0 to 0.97 > 1 mm, poor), DF (2 to 30.95%, poor), and moisture (VM = 0.16 to 0.20 cm³ cm -³ < 0.22 cm³ cm -³, poor). However, they had good performance because the number of physical indicators classified as good was higher in Factor 1, which is the most important, representing 45.42% of the variance of the data, which includes SD, TPd, and AP.

Conclusions
The most relevant physical attributes to discriminate Cambissolos Haplicos are clay and sand fractions, and the structural attributes porosity, degree of flocculation, aggregates, and soil mechanical resistance to penetration.
The physical and structural attributes of the Cambissolos Haplicos were, in general, preserved when compared to the current condition of native forests, even under different land use and managements. However, the soils under the Native Forest Area 1 (0.20-0.30 m soil layer), Conventional Management Area, and Agroecological Area naturally denoted fragility in aggregate stability and degree of flocculation, mainly due to the predominance of the sand fraction.
The Cajaraneira (Spondia sp.) Orchard Area presented more favorable structural physical conditions to the development of agricultural crops, especially in the surface layer, and in relation to clay, total porosity, aeration porosity, gravimetric moisture, volumetric moisture, and soil mechanical resistance to penetration.
The sand fraction, and soil density discriminated the Agroecological, and Conventional Management Areas, representing the Cambissolo Haplico Carbonático tipico. The clay fraction, aeration porosity, total porosity, gravimetric moisture, and volumetric moisture discriminated the Cajaraneira (Spondia sp.) Orchard Area, Native Forest Area 2, and Native Forest Area 3, representing the Cambissolo Haplico Ta Eutrofico tipico, and the Cambissolo Háplico Ta Eutrófico vertissólico, the latter being also discriminated by the geometric mean diameter, and soil mechanical resistance to penetration. The degree of flocculation, and macropores discriminate the Native Forest Area 1, representing the Cambissolo Haplico Carbonatico vertissolico.