Clay Fraction Mineralogy and Structural Soil Attributes of Two Soil Classes under the Semi-Arid Climate of Brazil

Abstract

Mineralogical studies are incipient and necessary in the Apodi Plateau, Brazil. This study aimed to evaluate the interrelationship between the mineralogy of the clay fraction and its structural and chemical attributes and to differentiate two important pedo-environments in the semi-arid region of northeastern Brazil (Ferralsol and Cambisols of the Apodi Plateau-RN) using the multivariate technique. We evaluated the interrelationships between mineralogy and the structural and chemical attributes of the soil and differentiated between agroenvironments. We collected soil samples from six profiles in diagnostic horizons of the Ferralsols and Cambisols. In the mineralogical analysis of the clay fraction, X-Ray Diffraction was used to identify mineral peaks of clay and iron oxides. The physical and chemical properties of the soils were determined. The multivariate statistical technique was applied to process the data. The clay minerals identified were hematite, goethite, kaolinite, and illite. The principal component analysis allowed for the separation of environments according to soil density, Fed and Mg²⁺ for developed soils, and potassium, weighted average diameter, microporosity, total organic carbon, sodium, the sum of bases, calcium, total porosity, aeration, potential acidity, and Feo discriminated developing soils. This study revealed that the clay fraction influenced the structural attributes of the soil according to the degree of soil development. Two profiles presented atypical situations: High contents of crystallized iron in Cambisols and illite peaks in Ferralsolos. These mineralogical results are not commonly found in the literature, highlighting the potential for further studies. The answers concerning the mineralogy of semiarid soils in Brazil and the world show similarity.

1. Introduction

The Apodi Plateau covers 2146 km² between the states of Ceará and Rio Grande do Norte. This vast geomorphological unit has a flat relief with dominant slopes of <2% [1]. Moreover, the region has karst activity with soils derived from the Jandaíra formation [2], determining morphogenetic processes that are decisive in the dynamics and evolution of the landscape [3]. The Jandaíra formation belongs to the post-rift sequence of the Potiguar Basin and is part of the Apodi group [4]. It rests on the sandstones of the Açu formation [5], and in some areas, the karst features of the Jandaíra formation are covered by neogeneous sediments from the Barreiras group [6].

This formation of the Cretaceous period (80–110 million years) consists of marine carbonate sediments, characteristic of shallow and agitated water environments, in addition to deposited bioclastic calcarenites and calcilutites, forming a carbonate platform that covered the entire emerged portion of the Potiguar Basin during the Turonian and the Mesocampanian ages [7,8]. The shallow marine carbonate rocks of the Jandaíra formation have been subject to significant variations in permeability over time due to various fracture events and calcite cementation [7].

Soils formed on carbonate rocks show high levels of calcium carbonate and predominant minerals such as calcite and dolomite, which are typical of arid and semi-arid regions [3,9]. Cambisols are the predominant soi class in the Brazilian semi-arid region (Apodi Plateau-RN), even though Ferralsols, Argisols (Acrisols), and Chernosols (Chernozems) are also identified [1,2,10]. Among soil attributes, the mineralogical composition is one of the most important features since it influences physical and chemical phenomena and is related to soil formation, offering indications about the degree of soil development and the advance of weathering [11].

The mineralogy of the soil clay fraction varies depending on the soil parent material and soil position in the landscape [12]. In highly weathered soils (Ferralsols), typical of tropical regions, kaolinite peaks and iron and aluminum oxides (hematite, goethite, and gibbsite) predominate. Iron and aluminum oxides have different degrees of crystallinity and are evaluated according to mineral groups based on the iron content (Feo/Fed ratio); they are extractable forms of iron, Feo quantification of Iron in amorphous forms extracted by oxalate and Fed- quantification of iron in crystalline forms, extracted by dithionite-citrate-sodium bicarbonate [12,13]. The Feo/Fed ratio is an important soil weathering indicator, with values below one indicating developed soils [14]. However, Cambisols have a predominant mineralogical composition of 2:1 clays (illite, vermiculite, smectites, and micas) formed by the process of bisialitization that is typical of low to intermediate soils (about pedogenetic evolution) in the semi-arid region of Brazil [1].

From this perspective, the mineralogy of the clay fraction is closely related to physical and chemical soil properties, influencing the soil structure due to the formation and maintenance of aggregates and soil density [15]. In addition, cohesion and adhesion forces influence mechanical processes that govern macroscopic soil properties and are determined by the processes of flocculation and dispersion, which depend on molecular interactions at the particle level [16,17]. Moreover, soil mineralogy influences phosphate sorption phenomena and geochemical processes, e.g., nutrient cycling and the soil availability of nutrients and heavy metals [18,19,20].

The literature contains several studies on the clay fraction conducted with various types of soils [1,18,21,22,23]. However, there are still gaps to be filled to better understand clay mineralogy, especially in arid and semi-arid regions. Studies of this nature can improve planning activities for low-carbon agriculture, predicting the effects of land use associated with weathering and pedogenesis. These studies are essential for soil surveys, ecological/biogeochemical modeling, and decision-making for sustainable actions. Therefore, this study aimed to evaluate the interrelationship between the mineralogy of the clay fraction and its structural and chemical attributes and to differentiate two important pedo-environments in the semi-arid region of northeastern Brazil (Ferralsol and Cambisols of the Apodi Plateau-RN) using the multivariate technique. The hypothesis stated that the mineralogy of the clay fraction would influence the structural attributes according to the degree of development of the studied soil classes.

2. Materials and Methods

2.1. Study Area and Sample Collection

This study was carried out in two Settlement Projects (SP) in the Apodi Plateau micro-region of the Potiguar West mesoregion, State of Rio Grande do Norte, Brazil (Figure 1). The studied settlements were called “Terra da Esperança” (Governor Dix-Sept Rosado) and “Moacir Lucena” (Apodi). The climate of the region is classified as BSw’h (hot semi-arid) according to Köppen [24], and the natural vegetation is the Caatinga dry forest biome.

The soils of the studied areas are classified as Cambisols and Ferralsols according to [25] and to the World Reference Base—WRB [26] published by the Food and Agriculture Organization of the United Nations—FAO. The research was carried out in six Agroecosystems (

Table 1. Uses and classification and location of the environments.

Profile	Horizont (m)	Soil Class/WRB	Land Use	Settlement
P1	Bw (0.48–1.20)	Ferralsol	Recovering—RC	Moacir Lucena
P2	Bi (0.15–0.54)	Haplic Cambisol	Lake—L
P3	Bw (0.25–0.97)	Ferralsol	Reserve—R
P4	Bi (0.08–0.37)	Haplic Cambisol	Native Forest—NF	Terra da Esperança
P5	Bi (0.07–0.22)	Haplic Cambisol	Colective—CC
P6	Bi (0.10–0.42)	Haplic Cambisol	Agroecological—A

), two Ferralsols, and four Cambisols by investigating different diagnostic horizons (Bi/Cambisols and Bw/Ferralsol) of the soil class profiles in both SP.

Soil samples were collected for laboratory analysis in the diagnostic horizons of the Cambisol (Bi) and Ferralsol (Bw) classes, with deformed and undeformed structures [27]. The incipient B horizon (Bi) is a sub-surface diagnostic horizon underlying the A, Ap, or AB horizons and shows physical and chemical modifications in a not very advanced degree, although sufficient for developing color and structural units. In contrast, the Ferralsols B horizon (Bw) is a sub-surface mineral horizon that exhibits advanced weathering with the transformation of easily changeable minerals, followed by intense desilication, base leaching, and residual concentrations of 1:1 sesquioxide and clay minerals and weathering-resistant minerals [25].

A deformed sample was collected from each diagnostic horizon, with the aid of a shovel and tray, placed in plastic bags, identified, and sent to the laboratory. The soil was air-dried, ground, and passed through 2 mm sieves to obtain fine air-dried soil for measuring its texture and chemical and mineralogical attributes. The laboratory procedure was performed in four replicates.

The undeformed samples, totaling ten samples for each horizon of the respective soil profiles, were collected to determine the Ma (soil macroporosity), Mi (soil microporosity), TP (total porosity and aeration), and BD (bulk density). For the extraction of the samples, the horizons prior to the diagnosis were removed with the aid of a straight shovel. The Uhland extractor and rings measuring 0.05 m in both height and diameter were used for that purpose, resulting in impacts sufficient to promote penetration of the ring into the ground. After collection, the rings were coated with aluminum foil and taken to the laboratory to preserve the soil structure and moisture.

2.2. Soil Analyses

2.2.1. Physical Analyses and Aggregate Stability

The physical parameters presented in

Table 2. Physical characteristics of soils with their respective methods and standards.

Physical Attributes	Unity	Method
Sand Total (2–0.05 mm)	g.kg−1	Screening
Clay (<0.002 mm)	g.kg−1	Sedimentation
Silt (0.05–0.002 mm)	g.kg−1	Sedimentation
BD	kg.dm−3	Volumetric ring
Mi 1	cm3.cm−3	Tension table
TP 2	cm3.cm−3	Tension table
Paeration	cm3.cm−3	Tension table
Ma	cm3.cm−3	Macroporosity = (TP − Microporosity)
WMD	mm	[29]

¹ Microporosity soil was determined by the water content of the soil at a height value equal to 60 cm of water (approximately a pore radius of 25 μm). ² The volumetric rings were saturated for 48 h and weighed to determine the total porosity. BD: bulk density; Mi: microporosity soil; TP: total porosity; P_aeration: aeration porosity; Ma: macroporosity soil; WMD: Weighted mean diameter.

were determined and subjected to multivariate statistical analysis to assess their influence on the studied soil characteristics. Furthermore, as a physical procedure, the clay fraction of the samples collected from the diagnostic horizons was prepared for further mineralogical characterization. The separation of the clay fraction from the other granulometric fractions was performed using chemical dispersion (0.025 mol.L⁻¹ sodium hexametaphosphate) and mechanical stirring (with a “Wagner” agitator for 16 h). After each collection, the volume of each beaker was completed with sodium carbonate solution at pH 10, thus maintaining the pH between 8 and 8.5. This procedure was repeated until the clay fraction was totally removed [28].

The blocks extracted in the soil profiles were passed through 4 mm and −2 mm sieves, thus preserving the structure of the aggregates. The aggregates retained in the 2 mm sieve were placed in cans, after which the wet sieving method was employed [29]. Four 25 g replicates were used for each profile. After pre-wetting, the samples were transferred to a set of 4.76, 2, 1, 0.5, and 0.25 sieves in the vertical oscillator working at 42 oscillations per minute [30] and then taken to the oven at 105 °C. Subsequently, the samples were placed in an oven (105 °C), and the aggregates were quantified. Based on the dry mass, the sand fraction was discounted using a dispersing solution (NaOH mol.L⁻¹ solution), and the aggregate size distribution and geometric and weighted average diameters were obtained.

2.2.2. Chemical and Mineralogical Analyses

The chemical attributes determined are shown in

Table 3. Chemical characteristics of soils and their respective methods.

Chemical Attributes	Unity	Method
TOC	g.kg−1	Digestion of organic matter 1.
pH Calcium (Ca2+)	mg.dm−3	Agitation of a soil-water suspension (1:5) for 2 h Titrometry 2.
Sodium (Na+) Magnesium (Mg2+) Potassium (K+) Potential acidity (H + Al)	mg.dm−3 mg.dm−3 mg.dm−3	Titrometry 2. Titrometry 2. Titrometry 2. calcium acetate 0.5 mol L−1
Fes	g.kg−1	Atomic Absorption Spectrophotometry, with a sulfuric attack.
Fed	g.kg−1	Atomic Absorption Spectrophotometry, using Dithionite-Citrate-Bicarbonate.
Feo	g.kg−1	Atomic Absorption Spectrophotometry, using ammonium oxalate.

¹ The samples were crushed, which were passed through an 80-mesh sieve, then added potassium dichromate and carried on a hot plate. Afterwards, orthophosphoric acid and distilled water were added, being titrated with ferrous ammonium sulfate (0.05 M). ² Determined using potassium chloride extractor, using ascorbic acid and titrated with ethylenediamine tetra acetic acid (0.0125 M).

. The variables of TOC, exchangeable Na⁺, Ca²⁺ Mg²⁺ and K⁺ (H+Al), pH, and SB (sum of exchangeable basic cations (K⁺, Ca²⁺, Mg²⁺ and Na⁺). Furthermore, the iron oxides were quantified to support the mineralogical analysis of the collected soil samples. For this, three different analyses were carried out in the diagnostic horizons of the studied soils.

Sulfuric attack was initially carried out [31] to quantify the iron content present in the secondary minerals (Fes). For this, about 0.5 g of air-dried fine earth was used in 75 mL digestion tubes along with 20 mL of H₂SO₄ (at a 1:1 ratio with deionized H₂O) by taking the digester block to the temperature of approximately 180 °C for one hour. For cooling, 50 mL of deionized water was added, and the material was filtered through blue strip filter paper into 250 mL volumetric flasks [28].

Subsequently, the quantification of Fe in crystalline forms (Fed) was performed. For this, 0.2 g of clay was weighed and placed in 50 mL centrifuge tubes. Then, 10 mL of the 0.2 mol. L⁻¹ citrate solution and 0.5 g of powdered sodium dithionite were added to the tubes. Next, the tubes were taken to a water bath (at 50 °C) for 30 min, after which the tubes were removed, cooled at room temperature, and taken to the centrifuge at 2000 rpm. Finally, the extract was placed in a 50 mL volumetric flask. All steps were repeated three times. Then, the volumetric flask was filled with deionized water. A blank test containing the reagents and the final extract was carried out in parallel to the analysis, and the iron content was read with an atomic absorption spectrophotometer.

The iron content was determined by the oxalate method to quantify amorphous iron forms (Feo). For this, 0.2 g of clay was weighed and placed in centrifuge tubes covered with aluminum foil. Then, 10 mL of the “Tamm reagent” solution (0.2 mol.L⁻¹ ammonium oxalate + 0.2 mol.L⁻¹ oxalic acid at pH 3) was added, after which the tubes were sealed and taken to a horizontal shaker, where they were fixed parallel to the direction of movement and agitated for two hours. Then, the samples were centrifuged for 10 min at 2000 rpm. The final extract was transferred to 50-mL volumetric flasks and supplemented with deionized water. A blank test containing the reagents was carried out in parallel to the analysis. The extracts were also read with an atomic absorption spectrophotometer.

2.2.3. Mineralogical Analyses

The mineral peaks present in the clay fraction were identified by X-Ray Diffraction (XRD) with a SHIMADZU diffractometer, model XRD—6000, using kα1 emission from copper (Cu). The source potential was 40 kV, and the current was 30 mA. A scanning speed with a step of 0.02° was applied every second. The scanning range (2θ) was established from 5 to 70° using the program X-Ray v. 1.0.0.37, and the minerals were identified according to [32].

2.3. Statistical Analyses

After obtaining the data on the attributes with a deformed structure, expressed by the mean of four repetitions in the laboratory, the data were subjected to statistical analysis through multivariate analysis to detect the most sensitive attributes using the software Statistica 7.0 [33]. Pearson’s Correlation Matrix (P ≤ 0.05), Factor Analysis, Principal Component Analysis, and Cluster Analysis were used during this step. Pearson’s correlation analysis was adopted for 19 variables to ensure that they showed sufficient minimal correlations to justify their use in the data matrix.

3. Results

3.1. Physical and Chemical Attributes

Table 4. Mean values of physical, chemical, structural and mineralogical attributes in soil classes, Apodi Plateau, Rio Grande do Norte, Brazil.

Horizontes (cm)	Sand	Silt	Clay	BD	Mi	Ma	TP	PA	WMD	pH	TOC	P	K	Na	Ca	Mg	SB	V%	H+Al
	g kg−1			g cm3	cm3 cm3			(%)	mm		g kg−1	mg kg−1	cmolc kg
Profile 1—Ferralsol
A (0–4)	660	87	253	1.60	0.27	0.14	0.41	9.40	0.94	7.24	4.80	7.19	0.47	0.03	4.30	3.80	8.60	81.47	1.96
AB (4–17)	492	92	416	1.61	0.39	0.04	0.43	15.30	0.96	6.68	3.73	2.56	0.31	0.03	3.10	3.10	6.54	71.12	2.66
BA (17–48)	371	81	548	1.85	0.20	0.13	0.33	10.50	0.98	6.28	7.84	2.67	0.10	0.04	3.30	2.60	6.04	71.29	2.43
Bw (48–120)	397	97	506	1.51	0.38	0.03	0.41	12.50	0.94	6.80	4.93	2.88	0.05	0.03	3.40	3.00	6.48	82.26	1.40
BC (120–155+)	382	148	470	1.60	0.39	0.05	0.44	14.10	1.95	6.58	1.77	2.25	0.05	0.02	2.50	3.80	6.37	82.90	1.31
Profile 2—Cambissol
A (0–3)	653	96	251	0.72	0.30	0.07	0.37	15.20	1.85	7.64	6.23	4.56	0.42	0.08	7.10	3.30	10.90	89.65	1.26
BA (3–15)	415	135	450	1.16	0.20	0.21	0.41	21.10	1.89	6.75	3.20	2.14	0.32	0.04	4.90	3.30	8.56	79.74	2.17
Bi (15–54)	311	153	536	1.37	0.41	0.03	0.44	27.40	1.96	6.51	3.54	2.14	0.05	0.03	6.40	3.40	9.88	81.93	2.18
BC (54–75)	179	230	591	1.29	0.47	0.02	0.49	23.70	1.99	7.10	2.95	3.09	0.04	0.06	13.40	6.20	19.70	91.78	1.77
C (75–85+)	322	137	541	1.39	0.45	0.03	0.48	19.50		7.21	1.58	7.82	0.04	0.16	20.10	8.50	28.80	99.54	0.13
Profile 3—Ferralsol
A (0–7)	656	57	287	1.67	0.28	0.08	0.36	22.21	0.81	7.30	7.83	2.88	0.37	0.0	3.30	3.00	6.69	85.05	1.18
BA (7–25)	556	37	407	1.72	0.29	0.08	0.37	21.10	0.99	6.51	4.38	1.09	0.21	0.0	4.10	0.70	5.04	75.96	1.6
Bw (25–97)	444	44	512	1.62	0.35	0.09	0.44	20.00	0.99	6.33	4.38	2.56	0.13	0.0	4.60	0.30	5.06	77.92	1.44
BC (97–140+)	479	49	472	1.58	0.37	0.05	0.42	19.85	1.02	6.93	2.19	1.40	0.10	0.0	4.30	1.00	5.43	83.81	1.05
Profile 4—Cambissol
A (0–8)	489	174	337	0.96	0.53	0.08	0.61	27.32	1.76	7.06	32.10	1.50	1.15	0.28	24.30	0.15	25.88	100	0.00
Bi (8–37)	537	97	366	1.23	0.47	0.08	0.55	23.32	1.19	7.64	9.19	1.50	0.23	0.41	32.30	0.70	33.64	100	0.00
C (37–52)	642	106	252	1.31	0.38	0.08	0.46	18.89	1.36	8.58	5.81	2.50	0.21	0.32	24.80	3.70	29.03	100	0.00
R (52+)	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
Profile 5—Cambissol
A (0–7)	673	94	233	1.32	0.43	0.02	0.45	22.84	1.56	8.15	48.35	3.70	2.30	0.16	17.80	0.60	20.86	100	0.00
Bi (7–25)	569	151	281	1.35	0.42	0.06	0.48	21.16	1.18	8.48	9.18	0.90	0.41	0.21	15.80	1.30	17.72	100	0.00
Bi/C (25–50)	560	170	270	1.39	0.41	0.06	0.47	20.39	1.21	8.52	8.52	1.90	0.25	0.20	17.30	1.40	19.15	100	0.00
CB (50–200+)	536	182	282	1.42	0.43	0.03	0.46	21.15	1.49	8.48	5.00	2.50	0.18	0.21	15.80	1.20	17.39	100	0.00
Profile 6—Cambissol
A (0–4)	691	90	219	1.53	0.37	0.02	0.39	18.55	1.84	7.72	20.34	2.20	1.92	0.16	15.80	5.70	23.58	100	0.00
BA (4–10)	679	77	244	1.59	0.39	0.03	0.42	19.21	1.95	7.60	15.86	5.08	0.07	0.05	9.43	2.46	12.01	100	0.00
Bi (10–42)	587	115	298	1.48	0.40	0.03	0.43	19.69	1.44	7.29	5.27	3.40	0.84	0.15	12.30	1.60	14.89	100	0.00
BiC (42–70)	539	206	255	1.49	0.41	0.05	0.46	19.68	1.44	7.50	3.91	2.59	0.27	0.05	12.13	2.02	14.47	100	0.00
CB (70–76)	519	207	274	-	0.41	0.06	0.47	19.47	1.28	7.40	3.59	0.90	0.15	0.06	14.47	1.91	16.59	100	0.00
C (76–91+)	674	63	264	-	0.41	0.06	0.47	19.49	0.84	7.84	0.95	4.70	0.53	0.16	13.30	5.00	18.99	100	0.00

PA—Aeration porosity.

presents the average physical, chemical, structural, and mineralogical values in soil classes for each diagnostic horizon. The sand fraction predominates in the surface soil classes, except for profile 4. The explicit silt contents in the Cambissols (P4, P5, and P6) indicate less weathered soils. The clay fraction increased in depth in all soil classes, except profile 4 (Bi 0.08–0.37 m). The highest levels of iron (Fes) occurred in profiles 1 (Ferralsol) and 6 (Cambisol). Soil density was higher in Ferralsols (P1 and P3). Profiles 2, 4, 5, and 6 (Cambisols) showed more excellent resistance to the action of the active agent’s water and wind due to the higher values of WMD, as well as microporosity and total porosity of the soil. To pH, all classes had an alkaline character, including the weathered ones (Ferralsol). Total Organic Carbon (TOC) decreased in depth in all profiles. However, considering low contents, higher in Cambisols. In general, occurred low levels of phosphorus (P), low levels of potassium (K⁺), and no restrictions on sodium (Na⁺). The bases Ca²⁺ and Mg²⁺ and the sum of bases were higher in Cambisols. The classes showed a eutrophic character (V ≥ 50%). The highest values of potential acidity (H+Al) occurred in the Ferralsol without restrictions.

3.2. Mineralogical Attributes

The identification of the mineral peaks of the clay fraction allowed us to separate the soil classes into two groups. One of the groups included developed soils belonging to the Red-Yellow Ferralsol (P1), Haplic Cambisol (P2), and Yellow Ferralsol (P3) classes. The other comprised developing soils (P4, P5, and P6), both classified as Haplic Cambisols (

Table 5. Mean values Fe₂O₃ contents extracted by sulfuric attack, dithionite-citrate-bicarbonate and ammonium oxalate, and the Feo/Fed ratio.

Profile/Horizont	Fes	Fed	Feo	Feo/Fed	Clay
	g/kg
P1—Bw (Ferralsol)	13.31	2.89	0.54	0.19	506
P2—Bi (Cambisol)	7.81	2.08	1.65	0.79	536
P3—Bw (Ferralsol)	8.82	3.25	0.90	0.28	512
P4—Bi (Cambisol)	8.96	0.62	0.36	0.57	366
P5—Bi (Cambisol)	8.64	0.75	0.43	0.58	281
P6—Bi (Cambisol)	10.36	0.75	0.21	0.28	298

Fes—Iron by sulfuric attack; Fed—Iron by dithionite; Feo—Iron by oxalate; Feo/Fed: Feo/Fed ratio.

).

The degree of crystallinity of iron oxides (Feo/Fed) ranged from 0.19 (Red-Yellow Ferralsol) to 0.79 (Haplic Cambisol). P3 (Yellow Ferralsol) showed the highest Fed content (3.25 g.kg⁻¹) among the studied soils. For the Feo content (amorphous iron), which ranged from 0.21 to 1.65 g.kg⁻¹, the mean value was 0.90 g.kg⁻¹. P4 (Haplic Cambisol) had the lowest Fed content (0.62 g.kg⁻¹) among the studied soils. With regard to the Feo/Fed ratio, most soil classes showed values below 1, corroborating their weathered characteristics.

The XRD of the natural clay from the diagnostic horizons is shown in Figure 2 and Figure 3. Among the clay mineral groups, Fe oxides (goethite and hematite), type 1:1 clay minerals (kaolinite), and illite (2:1 non-expansive clay mineral derived from mica) were observed in all studied soils.

The mineralogy of developed soils (Ferralsols) (Figure 2) is consistent with their evolutionary stage. Accordingly, all profiles showed well-defined kaolinite (Ct) peaks, implying an advanced crystallization degree. In developing soils (Cambisols) (Figure 3), most profiles had two expressive illite peaks (Il), implying fewer weathered minerals.

3.3. Multivariate Statistical Analysis

The correlation matrix revealed significant correlations between the clay fraction and the Fed and Feo contents and the properties of WMD, Mi, Ma, Mg²⁺ (H+Al), pH, and TOC. The same was observed for Mi, Ma, Na⁺, Ca²⁺, Mg²⁺, SB, pH, and TOC (

Table 6. Correlation matrix between the variables of physical, chemical and mineralogical attributes of the studied soils.

Atributes	Clay	Fes	Fed	Feo	Feo/Fed	TP	P.A	BD	WMD	Mi	Ma	Na+	Ca2+	Mg2+	K+	SB	(H+Al)	pH
Fes	0.10
Fed	0.88	0.29
Feo	0.75	−0.45	0.51
Feo/Fed	−0.02	−0.77	−0.39	0.55
TP	0.06	−0.02	−0.33	0.32	0.64
P.A.	0.47	0.15	0.10	0.48	0.46	0.88
BD	0.53	0.36	0.84	0.17	−0.69	−0.67	−0.38
WMD	−0.87	0.17	−0.65	−0.91	−0.41	−0.37	−0.65	−0.21
Mi	−0.93	−0.20	−0.94	−0.71	0.17	0.06	−0.31	−0.73	0.80
Ma	−0.98	−0.16	−0.84	−0.65	0.07	−0.04	−0.47	−0.48	0.79	0.86
Na+	−0.66	−0.24	−0.79	−0.57	0.27	0.14	−0.07	−0.80	0.57	0.88	0.54
Ca2+	−0.60	−0.31	−0.78	−0.47	0.36	0.20	0.00	−0.83	0.49	0.84	0.47	0.99
Mg2+	0.86	0.05	0.94	0.69	−0.21	−0.28	0.05	0.82	−0.70	−0.96	−0.78	−0.89	−0.86
K+	−0.13	−0.19	0.18	−0.20	−0.47	−0.93	−0.93	0.58	0.40	0.01	0.14	−0.15	−0.18	0.26
SB	−0.56	−0.35	−0.74	−0.45	0.35	0.14	−0.04	−0.79	0.48	0.81	0.43	0.98	1.00	−0.81	−0.12
(H+Al)	0.70	0.12	0.43	0.76	0.34	0.68	0.84	0.06	−0.83	−0.68	−0.63	−0.58	−0.50	0.50	−0.66	−0.53	1.00
pH	−0.84	−0.23	−0.67	−0.62	0.12	−0.16	−0.45	−0.47	0.62	0.79	0.83	0.66	0.58	−0.72	0.11	0.56	−0.72	1.00
TOC	−0.74	−0.16	−0.72	−0.62	0.22	0.07	−0.14	−0.70	0.54	0.84	0.66	0.88	0.83	−0.85	−0.17	0.81	−0.60	0.89

Clay; Fes—Iron by sulfuric attack; Fed—Iron by dithionite; Feo—Feo/Fed: Feo/Fed ratio; Iron by oxalate; TP—total porosity determined; PA—Aeration porosity; BD—Bulk Density; WMD—weighted mean diameter; Mi—Microporosity; Ma—Macroporosity; Na⁺—sodium; Ca²⁺—calcium; Mg²⁺—Magnesium; K⁺—Potassium; SB: Sum of base; (H+Al): Potential acidez; pH: Hydrogen potential; TOC—Total Organic Carbon.

).

The variables of Feo/Fed, P_aeration, and K⁺ did not correlate with the other variables and were considered independent (Table 6). The high correlation observed corroborates the adequacy of the data for using the cluster, principal component, and factorial analyses.

Two groups were formed when the reading was performed from right to left in the vertical dendrogram obtained by cluster analysis (Figure 4), one of which was composed of soil classes with developed pedogenesis (P1 and P3), both Ferralsols. These profiles showed high dissimilarity and shorter Euclidean distances and were inserted into the first group, indicating a distinction of the variables in relation to the data set.

The cumulative proportion of factors extracted from the 19 variables studied (factors 1 to 3) explained 91.14% of the total variability of the results (

Table 7. Factorial axes extracted for soil attributes and their factor loadings, eigenvalues, total and cumulative variance.

Atributes	Factor 1	Factor 2	Factor 3
Clay	−0.90	0.23	−0.07
Fes	−0.22	−0.26	0.94
Fed	−0.92	−0.21	0.01
Feo	−0.70	0.51	−0.49
Feo/Fed	0.20	0.80	−0.54
TP	0.06	0.93	0.28
P. aeration	−0.28	0.89	0.34
BD	−0.75	−0.66	0.02
WMD	0.76	−0.56	0.22
Mi	0.99	−0.03	0.02
Ma	0.83	−0.22	0.01
Na+	0.91	0.18	−0.02
Ca2+	0.87	0.26	−0.08
Mg2+	−0.96	−0.16	−0.22
K+	−0.03	−0.87	−0.45
SB	0.85	0.23	−0.14
(H+Al)	−0.68	0.68	0.18
pH	0.84	−0.19	−0.12
TOC	0.90	0.08	0.02
Eigenvalues	10.37	4.97	1.98
Total Variance (%)	54.59	26.14	10.41
Accumulated Total Variance (%)	54.59	80.73	91.14

Clay; Fes—Iron by sulfuric attack; Fed—Iron by dithionite; Feo—Iron by oxalate; Feo/Fed: Feo/Fed ratio; TP—total porosity determined; P_aeration—Aeration porosity; BD—Bulk Density; WMD—weighted mean diameter; Mi—Microporosity; Ma—Macroporosity; Na⁺—sodium; Ca²⁺—calcium; Mg²⁺—Magnesium; K⁺—Potassium; SB: Sum of base; (H+Al): Potential acidez; pH: Hydrogen potential; TOC—Total Organic Carbon.

). F1 allowed for estimating the influence of expressive variables with significant factor loads in the differentiation of the environments, especially clay, Fed, Feo, BD, WMD, Mi, Ma, Na⁺, Ca²⁺, Mg²⁺, SB, pH, and TOC. Factor analysis revealed that most causes of variation (F1 = 54.59%) are due to the relationship between clay, Fe oxides, and structural attributes (Ma, Mi, WMD) and soil chemistry (Na⁺, Ca²⁺, Mg²⁺, SB, pH, and TOC). About F2, the variables that stood out were Feo/Fed, TP, P_aeration, and K⁺ (26.14%) (Table 7). F3 was the least expressive parameter, with the lowest value of the selected attributes highlighting the Fes variable, which describes the data related to iron oxides (10.41%). In addition, the sum of the cumulative variances (F1 and F2) explained 80.73% of the variation and contributed the most to distinguishing the studied environments (Table 7).

The WMD, K⁺, BD, Fed, Mg²⁺ (H+Al), TP, and P_aeration variables are close to the unit circle (Figure 5A), meaning greater sensitivity in differentiating between environments in relation to the other variables. According to the point cloud distribution (Figure 5B), the most prominent variables in factor 1 are WMD, Mi, TOC, TP, Ca²⁺, and Na⁺. Factor 2 included BD, Fed, Mg²⁺, clay, (H+Al), and P_aeration. Therefore, the significant variables for components 1 and 2 (F1 and F2) are those that most represent the studied soils due to their total accumulated variance in relation to the total of the three factors (91.14%).

The most important variables for F3 were Fes, Fed, Mg²⁺, and Na⁺, highlighting the interrelationship between iron oxides and structural attributes since they are part of soils with different levels and types of clay, acting differently as aggregating agents (Figure 6A). In addition, P2 (Haplic Cambisol) and P3 (Yellow Ferralsol) were sensitive to Feo, whereas P1 (Red-Yellow Ferralsol) was sensitive to Fes (Figure 6B). However, for soils with a less accentuated degree, e.g., Cambisols (P4, P5, and P6) (Figure 6A), the discriminating variables were TOC, WMD, Mi, Na⁺, and Ma (Figure 6B).

4. Discussion

The hierarchical analysis allowed the grouping of the studied soils into advanced weathering (Ferralsols) and little-weathered categories (Cambisols), except for P2 (Cambisol), which was grouped in P1 and P3 (Ferralsols). This can be explained by the fact that P2 (Cambisol) had a high amount of pedogenic iron oxides (Fed), similar to Ferralsols. This finding is considered unusual in Cambisols of sedimentary origin and is little documented in the literature, especially in the studied region.

The high Fed content in Cambisols could indicate developed soils even in conditions of little pedogenesis, with the flat relief acting as an important factor in this process, similarly to the study conducted in [34]. In another study, Girão et al. [35] analyzed the Fed levels in Cambisol and Argisol profiles (Acrisols) in the Apodi Plateau and found that the Fed values increased in Cambisols, suggesting variation in the degree of crystallinity. However, greater homogeneity occurred in Argisols, which is characteristic of evolved soils.

Melfi et al. [36] point out that Cambisols have iron contents (Fe₂O₃) lower than 5% because they are poorly evolved. This low content is explained by the soil parent material and the degree of soil pedogenesis. However, in the context of the Brazilian semi-arid region, some soil classes have iron in the form of oxides. In addition, pedogeochemical processes, such as sialferritization, justify iron incorporation into 2:1 clay mineral.

Girão et al. [35] argues that the presence of iron oxides in limestone-derived soils usually occurs in reddish clayey soils (5YR) and with oxide (hematite and goethite), kaolinitic, and micaceous mineralogy. Anastacio et al. [37] pointed out that the mineralogy of the clay fraction of Cambisols in the region of Minas Gerais (Brazil) consists mainly of maghemite (Fe₂O₃) and superparamagnetic goethite (FeOOH), in addition to kaolinite, smectite, and smaller portions of anatase.

Ferralsols naturally show a greater presence of iron and aluminum oxides such as hematite, goethite, and gibbsite, depending on variations in the parent material, the intensity of weathering, and drainage conditions [38]. The oxidic constitution, virtually without phyllosilicates, gives these soils a developed granular structure, low cation exchange capacity, and excessive permeability, in addition to a relatively homogeneous color with reddish or yellowish hues, horizon uniformity, and greater depths [39].

Correlations occurred between Fed, clay, and other structural attributes related to soil aggregation (Mi, Ma, WMD, and Ca²⁺). Since soil texture contributes to stabilizing and forming aggregates, soils with high clay contents favor aggregation since the clay content favors particle adhesion, which, in turn, depends on complexing agents, silicate minerals, Fe and Al oxides, ionic strength, pH, exchangeable cations, and phosphate sorption [16,40]. Lu et al. [41] found strong and positive correlations between clay, iron oxides (Fed), and pore sizes, mentioning that iron oxides and clay minerals (main kaolinite) promote pedogenic cryptoporosity (<0.10 μm). In another study, Durn et al. [42] noted that iron oxides were important in the aggregation of Cambisols in Croatia.

Moreover, although obtaining a low correlation between iron oxides and aggregation, Yin et al. [43] pointed out that precipitated aluminum phosphates are important cementing agents that stabilize soil aggregates. Mota et al. [40] highlighted that this aggregation could be modified with soil preparation and biotic and abiotic factors. In addition to cementing agents (iron and aluminum oxides), significant contributions of the carbon and nitrogen cycles (aggregating agents) also promote the stability of soil aggregates (mainly macroaggregates) [44].

In Brazilian subtropical regions, higher contents of crystalline iron were detected in this soil class [35]. The levels of crystalline iron (Fed) vary according to the parent material. Therefore, compared to basic rocks, sedimentary rocks show lower values of this component [45,46]. Iron oxides (hematite and goethite) are correlated with soil structure and aggregation, influencing water permeability and resistance to erosion [47]. Furthermore, Fe₂O₃ was observed at higher values in subsurface horizons than at the surface due to the higher clay levels at those horizons, where the highest proportion of pedogenic Fe oxides are concentrated.

As for the structural attributes, soil texture was influenced by the more pronounced presence of the clay fraction since soils with higher clay levels resulted in better aggregation, increasing the contact between soil particles. The contrary was observed in sandy soils, which hinder the formation of primary aggregates [48], with organic matter playing an essential role in the formation of secondary aggregates [49].

The clay content supplies agglomerates of unitary particles resulting from the physical attraction between particles and the action of cementing agents, e.g., iron oxides and TOC [50,51]. Moreover, TOC is important for maintaining porosity and improving the apparent soil density [52] since soil porosity directly influences the dynamics of organic matter and the apparent density [53,54].

P3 (Ferralsol) had the highest soil density value, which could be due to the rearrangement of particles since it corresponds to subsurface horizons where mineral translocation occurs over short distances and the sand fraction predominates. Another factor that can increase soil density is the presence of iron oxides. However, studies claim that these factors have little influence but show effects when interacting with TOC and the soil clay fraction [55,56].

The discriminating attributes in P4 (Cambissol) were Ca²⁺ and Na⁺. This presence of sodium is characteristic of semi-arid regions due to factors such as high evapotranspiration [57], low rainfall [58], and anthropic interference [59]. Exchangeable calcium, on the other hand, can be attributed to the parent material and the formation mechanisms associated with pedogenetic processes, justifying the variability of soil classes in regions with a limestone predominance. This was also reported in [60], which stated that those soils show calcium accumulation and promote the formation of calcium or petrocalcic horizons. Similar results were found in the research carried out in [2], in which high values of exchangeable calcium were also detected.

Finally, P5 (Cambisol) showed low TOC and Mi values compared to the other classes. This can be attributed to local climatic conditions and rainfall irregularity in association with the type of vegetation of the semi-arid region [22]. In addition, the material of limestone origin in P6 contributes to discriminating the values of structural attributes, e.g., WMD and Mi, while providing Ca²⁺ and Na⁺.

The low levels of iron oxides in the soil classes can be explained by sedimentary rocks (Jandaíra Formation). This formation is characterized by sedimentary carbonate rocks with small-scale, sub-vertical fractures [7], forming the largest exposed carbonate platform of the Cretaceous on the Brazilian margin, where conspicuous outcrops form exposures up to 2 km wide [6]. Soils like these, derived from unconsolidated sediments, naturally have low Fe contents and do not provide conditions for forming ferrimagnetic minerals [61].

Fe oxides can usually be distinguished into two types: Amorphous (Feo) or crystalline (Fed) [62]. They have different degrees of crystallinity depending on the level of weathering, parent material, and the position in the landscape [12,13], considerably influencing the physical, chemical, and structural soil properties.

While Feo/Fed ratio values lower than one indicate more developed soils, values above one indicate younger soils [63]. Thus, Ferralsols (weathered soils) have a low Feo/Fed ratio, indicating higher iron contents in crystalline forms (Fed) due to greater weathering [12,14,64]. The Feo content (amorphous iron) in P3 indicates greater pedogenic development and fewer preserved characteristics of the parent material. In addition, the presence of iron was visualized in the field in the form of concretions with a mean diameter larger than 2 mm. These nodules are believed to be sources of information on pedogenetic processes [65].

P4 had the lowest Fed content among the studied soils. It is believed that its condition is characteristic of little-weathered soils and intrinsic to the parent material. Since Fe oxides are important indicators of pedogenetic development, much of the Fe detected by Fe oxides might be present as silicate clay minerals. In general, soils in semi-arid zones have low Fe₂O₃ levels, with few crystallized iron forms, mainly due to the climatic pattern and weathering processes [66]. In addition, those soils have poor drainage conditions, long water deficit periods, and inefficient leaching [67].

In addition, the diffractograms showed well-defined kaolinite (Ct) peaks, indicating that kaolinite is the dominant mineral in the studied soil classes, especially in Ferralsols. Ct is generally formed from primary minerals (micas and feldspars) [68] in hot and humid climates with intermediate mineral weathering and the partial removal of basic cations and silicon [69]. Oliveira et al. [2] and Ferreira et al. [1] also point out that kaolinite was the most common clay form found in the soils of the Apodi Plateau, whereas [70] stated that kaolinite was the main mineral found in the clay fraction of weathered soils in the Amazon basin. Similar results were described by [22] in Planosols.

Furthermore, the Cambisol class also showed kaolinite peaks (1:1) as well as iron and aluminum oxides (hematite and goethite). Concerning this, [71] points out as a probable hypothesis that muscovite will be present in the parent material (limestone rock), with silica solubilization occurring under high pH conditions. In association with the balance of the relationship between the activities of exchangeable bases (Ca²⁺, Mg²⁺, and K⁺), this condition favors the formation of kaolinite even in the absence of pronounced leaching. Furthermore, [72], in a study conducted in Cambisols of the Apodi Plateau, stated that higher pH values favor silica dissolution and higher kaolinite contents. In addition, the oxidation of 2:1 clay is not restricted to leached soils [21], as kaolinite formation can occur in Cambisols.

Diffractograms also showed illite peaks (2:1 clay). While less common, type 2:1 clay minerals such as illite can be found in lower proportions in Ferralsols compared to Cambisols [38]. The presence of type 2:1 clay minerals evidences the differentiation of the class of Cambisols derived from limestone rocks abundant in the Apodi Plateau region [2].

As a mineral of the mica group, illite derives from the weathering of muscovite or is inherited from calcareous parent materials [1]. Reference [2] described the presence of 2:1 illite and smectite clay minerals in Cambisols of the Apodi Plateau, corroborating the present study. In addition, [23], studying soils of calcareous origin in the semi-arid region of Morocco, pointed out that the mineralogy of the clay fraction was composed mainly of illite, smectite (2:1), and kaolinite (1:1), varying depending on the parent material, climate, and topography.

5. Conclusions

The present study investigated the mineralogical composition of the clay fraction of Ferralsols and Cambisols in the Apodi Plateau, Brazil. The multivariate analysis allowed differentiating agri-environments by indicating attributes with the most sensitive factor loads. The structural attributes were valuable for discriminating Cambisols and the apparent soil density and crystalline iron forms in Ferralsols. The clay fraction influenced the structural attributes and varied according to the degree of development of the studied classes, correlating with iron oxides. These iron oxides were positively correlated with soil aggregates, acting as cementing agents in the studied soil classes. Under the semi-arid climate, both soil classes showed clay fractions with the predominance of kaolinite (uncommon in Cambisols) and a lower proportion of illite (uncommon in Ferralsols). One of the Cambisol profiles (Profile 2) showed higher iron contents in the crystalline form, with kaolinite peaks and iron oxides, constituting an improvement in the degree of development considering the flat relief. The similarity of the formation factors help in the answers about the mineralogy of soils in the semiarid region in Brazil and in the world. The study will be able to subsidize the planning of programs and activities for development and carbon sequestration, effects of land use on soils and pedogenesis, soil surveys and ecological/biogeochemical modeling and decision-making in sustainable development actions.