Weathering and clay formation in semi-arid calcareous soils from Northeastern Brazil

Catena xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Catena journal homepage: www.elsevier.com/locate/catena

Weathering and clay formation in semi-arid calcareous soils from Northeastern Brazil D.P. Oliveiraa, L.R. Sartorb, V.S. Souza Júniorc, M.M. Corrêac,d, R.E. Romeroa, G.R.P. Andradee, ⁎ T.O. Ferreirab, a

Departamento de Ciências do Solo, Universidade Federal do Ceará, UFC, M.B.12168, Fortaleza, Ceará, Brazil Departamento de Ciência do Solo, Universidade de São Paulo, ESALQ/USP, Piracicaba, São Paulo, Brazil c Departamento de Agronomia, Universidade Federal Rural de Pernambuco, UFRPE, Recife, Pernambuco, Brazil d Departamento de Ciência do Solo, Universidade Federal Rural de Pernambuco (UFRPE) - Unidade Acadêmica de Garanhuns, Garanhuns, Pernambuco, Brazil e Laboratório de Solos/CCTA, Universidade Estadual do Norte Fluminense Darcy Ribeiro, Campos dos Goytacazes, RJ, Brazil b

A R T I C L E I N F O

A B S T R A C T

Keywords: Microrelief Newmod II TG-DTG Kaolinite Mixed-layered clay minerals

Despite the great number of studies exploring the inter-profile genetic relationships in a catena, the microrelief and microtopographic features have been overlooked. This study investigated the role of microrelief surfaces (convex, flat and concave) on the mineral formation and pedogenesis in soils derived from limestones in NE Brazil. Soil profiles were studied in different micro-landscape surfaces, and soil samples were taken by horizon for chemical (weathering indexes) and mineralogical analyses. Mineralogy was determined in samples from diagnostic subsurface horizons, which were subjected to the proper chemical treatment for identification (X-ray diffraction - XRD) and quantification (Newmod II and Thermogravimetry - TG) of the phyllosilicates. The XRD patterns showed similar composition among the parental material of the studied profiles. The chemical also showed similar dynamics along the surfaces; however, the concave surface had soils with a higher degree of weathering, when compared to others. The clay minerals identified on the convex and flat surfaces were vermiculite, illite, smectite and kaolinite; while on the concave surface, only illite and kaolinite were found. Based on the information generated by the Newmod II, there was an increase in the amount of kaolinite towards the concave surface, from 21%, on the convex surface, to > 50% of pure kaolinite crystals on the concave surface. These data were supported by the TG curves. Our results evidence that microrelief features, marked by low topographic amplitudes, are determinant for the formation and distribution of phyllosilicates in calcareous soils; and, thus, may be of key importance for predicting the occurrence of contrasting mineralogy in soils derived from calcareous rock.

1. Introduction The notion that the local relief and its forms (slopes and curvatures; both vertically and laterally) control water dynamics and the processes of material redistribution (i.e. colloidal material and solutes) on a landscape dates back to the catena concept coined by Milne (1935); which was complemented by others (Conacher and Dalrymple, 1977; Huggett, 1975; Ruhe, 1975). Several studies explored the inter-profile genetic relationships in a catena, and how they govern the distribution of different soils on landscapes. However, historically, the vast majority of these studies, including those on calcareous material (Atalay, 1997; Silva et al., 2017; Van de Wauw et al., 2008), have focused and high slope gradients, while microrelief and microtopographic features have been often overlooked.

⁎

The origins of these small-scale landforms may be diverse and complex, and several different processes have been proved to create them, such as treethrow disturbances (tree uprooting; EmbletonHamann, 2004; Schaetzl, 1990), faunalturbation (pedoturbation promoted by animals; Soyer, 1983; Phillips, 2007), argillipedoturbation (shrinking and swelling of clays; Knight, 1980; Schaetzl, 2008), and cryoturbation (freeze–thaw activity; Bockheim and Tarnocai, 1998). However, irrespective of its origins, microrelief features may trigger or hinder certain pedogenetic processes, depending on local conditions (i.e. climate, drainage, water table behavior) and govern the soil distribution patterns at local and regional scales. In northeastern Brazil, where the semi-arid region covers an area of about 750,000 km2 (Ab'Saber, 1977), pedogenesis is mostly governed by the combination of high temperatures and irregular rainfall, acting

Corresponding author. E-mail address: [email protected] (T.O. Ferreira).

https://doi.org/10.1016/j.catena.2017.10.030 Received 31 March 2017; Received in revised form 17 October 2017; Accepted 24 October 2017 0341-8162/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Oliveira, D.P., Catena (2017), http://dx.doi.org/10.1016/j.catena.2017.10.030

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

hyperxerophilic Caatinga; however, the studied area is currently under banana cultivation. The studied site covers an area of 102 ha, located the municipality of Limoeiro do Norte (state of Ceará, NE Brazil). The terrain contains slope gradients predominantly < 3%, which covers 97% of the studied area, with a slight inclination in a northwest-southeast direction (Fig. 1). At the northwest end of the study site the higher elevations were recorded (~ 103 m) while minimum elevations (~ 101 m) were obtained at the opposite end (southeast). In previous studies by Oliveira (2009) and Oliveira et al. (2013), both a detailed soil survey and a (micro)topographic survey were conducted at the study site. A digital elevation model was produced (DEM) and based on the shape of the contour lines (microslope features) the studied area was divided into three different surfaces (Fig. 1): convex (42 ha), flat (22 ha) and concave (38 ha). Based on the chemical and physical data and morphological descriptions (according to Schoeneberger et al., 2012). The particle-size analysis was determined by the pipette method (Gee and Bauder, 1986). The total organic carbon was determined by wet oxidation (Walkley and Black, 1934). The soil pH was measured in water (1:1.25) using a glass electrode. The calcium carbonate equivalent (CCE) was determined according to Allison and Moodie (1965). The soils were classified according to WRB (IUSS Working Group WRB, 2015). Although the profiles were classified as Calcaric Cambisol (loamic) (convex and flat surfaces) and Calcaric Cambisols (clayic) (concave surface), soils showed a wide variation with respect to their morphological characteristics (i.e. soil depth) and mineral composition of clay fraction; with shallower soils and high-activity clays occurring in the convex surface. According with criteria established by WRB (IUSS Working Group WRB, 2015), the soils profiles do not show diagnostic properties that characterize a lithological discontinuities. Al2O3 and Fe2O3 were determined in the air-dried fine earth (< 2 mm; ADFE), using an extraction solution of sulfuric acid (1:1). SiO2 was extracted with 30%NaOH (EMBRAPA, 2011). Aluminum and iron contents were determined in the extracts by atomic absorption spectrophotometry, and the results were represented as oxides (Al2O3 and Fe2O3). Silica content was determined by colorimetry (Kilmer, 1965). The results were used to calculate a weathering index (the Ki index) based on the contents of SiO2 and Al2O3 (Ki = [SiO2 / Al2O3] x1.7; see Demattê et al., 2004). The mineralogical analysis was carried out in the diagnostic subsurface horizons (X-ray diffraction modeling and thermal analysis). The bulk mineralogy of the parent materials was analyzed by XRD as

predominantly on rocks from the crystalline basement (Precambrian Era).These conditions usually lead to the formation of slightly weathered soils, with high-activity clays and high base status. Despite the prevalence of low weathering rates in the Brazilian semiarid region, there are some exceptions where the typical processes from tropical humid environment take place (i.e. altimontane soils; see Barbosa et al., 2015). Among these unique environments is the Apodi Plateau, a vast geomorphological unit in NE Brazil, where the flat relief with dominant slopes of < 2% shows no dissection. Although the soil survey of the Apodi Plateau (Brasil, 1973) indicates a predominance of high-activity clay mineral (e.g. smectite and vermiculite; Ernesto Sobrinho, 1980; Lemos et al., 1997) and high base status (e.g. Cambisols), recent studies have reported the occurrence of low-activity clay minerals in soils mostly dominated by kaolinite; i.e. Lixisols and Ferralsols (Alencar, 2002; Ferreira et al., 2016; Mota et al., 2007). These contrasting soils form a continuum on the regional landscape with unclear boundaries, mostly due to the absence of clear landform distinction. The determination of relative abundance of each mineral as a function of the different (micro)relief features can serve as an important tool for predicting the occurrence of different soil types on these fairly flat, monotonous, karst landscapes. Additionally, it may assist with decision making for the management systems to be adopted, and can generate valuable information for the evaluation of pedogenesis. Thus, the aim of the present study is to assess how weathering rates, clay formation and hydrological conditions are ruled by rather slight variations in (micro)relief in these environments.

2. Material and methods The Apodi Plateau covers 2146 km2 between the states of Ceará and Rio Grande do Norte (NE Brazil; Fig. 1). This geomorphological unit belongs to the Potiguar Basin that is composed stratigraphcally, at emerged portion by the Jandaíra Formation (250–300 m thick) above the Açu formation. The first is composed by calcitic and dolomitic limestone rocks, whereas the. Açu Formation is composed of finegrained sandstone, with micaceous and kaolinitic clay mineralogy (Girão et al., 2014; Brasil, 1973; Sampaio and Schaller, 1968). The climate in the region is BSw'h, hot and semi-arid, with a mean annual temperature of approximately 28 °C and relative air humidity varying from 50 to 84%, with a mean rainfall of 750 mm (Fig. 1; Ceará, 1980). The native vegetation in the studied area was composed of

Fig. 1. (A) Brazil, Ceará State (NE), and climate conditions (mean temperature and rainfall) at the study site; (B) The digital elevation model and the three soil profiles representative of each surface (convex, flat and concave); arrows indicate the direction of preferential water fluxes at the land surface.

2

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

(28.2% to 44.9%) positions. On the other hand, clay contents increase in soils at Convex to Concave positions. The pH values vary from 7.2 to 9.2 in all horizons, and the high values are clearly associated with carbonate content (CCE in Table 1), which are higher in C horizons of soil at the convex position and in Bwck and BCkc horizons in the soil at Flat position. The observed variations are interpreted as an effect of terrain morphology in surface water dynamics. At the concave position, the accumulation of water flux during wet seasons leads to an increase in weathering rates, promoting the transformation of sand into clay-sized particles (by physical and chemical alteration), the dissolution of carbonate inherited from parent material (CCE) (Da Costa et al., 2014) and a subsequent decrease in pH values. This effect of the landscape in water circulation and its impact on soil attributes is in agreement with geochemical and mineralogical results, discussed below.

randomly oriented powder mount. In addition, the parent materials of each soil profile were subjected to carbonate removal using HCl 1 M, and the residues were analyzed by XRD. The samples were scanned from 3 to 50 °2θ, with 0.02 °2θ step size and count time of 1.0 s/step. Phyllosilicates were studied by the X-ray diffraction (XRD). The clay fraction (< 2 μm size) from the diagnostic subsurface horizons was obtained through particle sedimentation. Carbonates were removed using 1 mol L− 1 sodium acetate solution, buffered at pH 5. Organic matter was removed with 30% (v/v) H2O2. Pedogenic Fe oxyhydroxides were removed by the dithionite-citrate-bicarbonate (DCB) method (Mehra and Jackson, 1960). To identify different clay minerals, the samples were saturated with KCl and MgCl2. The K-saturated samples analyzed as air-dried samples (25 °C; K-25 °C) and after heating them at 550 °C (K-550 °C). Mg-saturated samples were analyzed as air-dried (Mg-25 °C) and after proceeding the solvated with ethylene glycol (Mggl). The saturated samples were analyzed as oriented aggregate specimens using a Shimadzu XRD 6000, with CuKα radiation (operated at 40 kV and 20 mA). The samples were scanned from 3 to 35 °2θ interval, with step size of a 0.02 °2θ step size and a count time of 1.0 s/step. The XRD patterns of the glycolated samples were modeled using the software Newmod II (Reynolds and Reynolds, 1996) in order to characterize and quantify the clay assemblage of the < 2 μm fraction. The main parameters used were: % of the different phases, % of the different layers in the mixed-layered minerals, stacking layer-ordering (R parameter), octahedral Fe content, maximum (Nmax) and average (Nave) number of layers in the coherent scattering domain, and the orientation factor (σ*, set as 12 for all samples, which means best particle orientation). To achieve a better fit of the diffraction peaks intensities for kaolinite, it was necessary to introduce layers of serpentinite to allow the manipulation of the Fe contents (Dudek et al., 2006, 2007). The XRD patterns were calculated without a °2θ slit compensator and assuming Mg2 + as the interlayer cation. For thermo gravimetric (TG) analysis, the samples were subjected to same pre-treatments to described for mineralogical analysis (organic matter and pedogenic Fe oxyhdroxides removal) Then, the samples were saturated with Mg2 + (1 mol L− 1 MgCl2 solution) and left for one night in a desiccator with relative air humidity of approximately 52% in a solution saturated by Mg(NO3)2 (Soukup et al., 2008). TG analysis was performed using a thermal analyzer (Netzsch - STA 449 F3) in the temperature range from 35 and 1100 °C and a heating rate of 10 °C min−1 in a N2 atmosphere. The obtained data were processed using the program Proteus® Version 5.1, Netzsch. The amount of kaolinite was determined based on the dihydroxylation event between 450 and 550 °C, where the % of kaolinite = 100 × (MLs / MLkt), in which MLs is the mass loss of the sample and MLkt is the mass loss of pure kaolinite. For the calculation, the water loss in the thermal events occurring prior to 450 °C was subtracted (Karathanasis, 2008).

3.2. Geochemical characterization and Ki index The semi-total chemical composition of soil samples is displayed in Table 1. The highest Al2O3 contents are superior in soil at the Concave position and are significantly higher in A and B horizons of all soils when compared to C horizons. A similar trend is observed for Fe2O3: highest values concentrated in A and B horizons of soil at the Concave position, reaching 19.6% in Apkc horizon. In general, the values range ~11–12% in B and A horizons, decreasing to very low contents in C horizons. The method used to characterize the samples dissolves only the fine fractions of soils and, thus, these results are important to describe how weathering and pedogenesis affect clay mineralogy. This increase in Al2O3 and Fe2O3 contents, therefore, the consequence of high weathering rates in topsoils. The Ki values, represented by the SiO2/Al2O3 molecular ratio, point out to different weathering pathways in the three soils. At Convex and Flat positions, Ki values are higher than in soil at the Concave position, reaching values close to 3.0. These high values denote low weathering rates and higher contents of 2:1 clays minerals, at the expenses of kaolinite and Fe/Al oxides. The contrasting values found in concave position represent the opposite situation: the influence of higher weathering rates and prevalence of kaolinite in clay fraction. This geochemical behavior is in agreement with the variation of general soil attributes described in the previous section and with the investigation of clay mineralogy. The Ki index is used in the Brazilian Soil Classification System to characterize and identify B Ferralic horizons, which must show values lower than 2.2, as an indication of high weathering rate (EMBRAPA, 2013). According to Mota et al. (2007), the high Ki values in these calcareous soils may be related to the presence of illite in the clay fraction, that probably was originated by mica weathering present in parental material (Fig. 2B). The lower values observed in A and B horizons at the concave surface indicate that, in this part of the terrain, there is an environment favorable to monosialitization, and the formation of 1:1 clay minerals. In this case, the concave form of the surface would favor convergent water fluxes, which drives a higher weathering rate. Conversely, at the convex part of the surface, the loss of water by local runoff (Fig. 1) would favor the occurrence of less weathered soils. In fact, studies have shown that smooth relief forms are able to influence water dynamics (King et al., 1983, 1999) imposing different processes reflected in soil geochemical attributes.

3. Results and discussion 3.1. Soil general characteristics The three studied soils have some common morphological features, attributed to the influence of calcareous parent material, although important differences can be noticed and related to other chemical and physical characteristics (Table 1).The sequence of A and Bwkc horizons (above Ck or CRk horizons) is ubiquitous in all soils, but do not reach 50 cm in depth in the soil at Convex positions. In soils at Flat and Concave positions, the solum thickness reaches up to 149 cm. Soil colors are predominantly yellow (hues varying from 7.5 YR to 2.5 Y), and slight variations in value and chroma are observed in surface horizons, due to the effect of organic matter. The soil particle distribution seems to be sensitive to the landscape position, but does not vary significantly within soil profiles. Higher sand contents prevail in the A and B horizons of the soil in the Convex position (from 65.3% to 67.8%), progressively decreasing in the soil of Flat (33.9% to 44.6%) and Concave

3.3. Clay fraction mineralogy The XRD analysis of calcareous parent material studied as randomly mounts, reveal the prevalence of calcite, recognized by d-space values at 3.86, 3.03, 2.49, 2.28 and 2.09 Å (Fig. 2A). After carbonate removal in the same samples, it was possible to visualize residual minerals not affected by the acidic treatment (Fig. 2B). Mica-group minerals, quartz and some minor amounts of an undistinguishable expansive 2:1 clay 3

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

Table 1 Morphological, physical and chemical characteristics of soils. Fe2O3, Al2O3and SiO2 contents determined with the sulfuric acid attack (in the air-dried fine earth) and the weathering index (Ki = 1.7xSiO2 / Al2O3) results for the soil profiles representative of each surface (convex, flat and concave). Horiz

Depth

Color

cm

Munsell

Sand

Clay

OC

%

pH

CCE

Water

g kg− 1

Fe2O3

Al2O3

SiO2

Ki

%

Convex – Calcaric Cambisol (loamic) Apkc 0–12 10 YR 3/4 Bwkc1 12–24 10 YR 4/4 Bwkc2 24–47 10 YR 4/4 CBk 47–70 10 YR 5/6 Ck 70–95 2,5 Y 8/1 CRk 95–119 2.5 Y 8/1 CRk1 119–154 10 Y 7/1 CRk2 154–173 + 10 Y 6/1

65.3 63.8 67.8 – – – – –

19.4 22.3 21.0 – – – – –

3.4 1.8 1.2 0.3 0.0 0.3 0.9 0.3

8.0 7.5 7.7 8.4 8.7 8.7 8.9 9.2

25 25 33 249.5 249.8 249.6 249.8 249.3

19.6 12.4 11.7 4.5 1.2 1.2 1.2 1.2

24.4 21.8 22.2 9.3 4.8 3.7 3.4 3.5

36.2 36.4 31.4 11.3 3.3 3.0 6.6 10.9

Flat – Calcaric Cambisol (loamic) Apkc 0–13 Bwkc1 13–52 Bwkc2 52–68 Bwkc3 68–83 BCkc 83–105 +

10 YR 3/4 10 YR 4/6 10 YR 5/8 10 YR 5/8 2.5 Y 5/6

44.6 42.3 33.9 43.5 –

29.2 28.4 29.6 29.9 –

4.3 1.2 1.0 1.4 1.2

8.2 8.4 8.4 8.3 8.3

75 126.5 206.0 174.5 248.5

11.8 9.7 9.4 8.4 2.8

20.6 20.0 20.1 20.5 7.8

29.5 29.6 39.9 29.3 20.9

Concave – Calcaric Cambisol (clayic) Apkc 0–6 7.5 YR 4/4 ABkc 6–31 7.5 YR 4/4 Bwkc1 31–56 10 YR 5/6 Bwkc2 56–92 10 YR 5/6 Bwkc3 92–119 10 YR 5/6 Bwkc4 119–149 + 10 YR 5/6

44.9 37.6 29.8 28.2 30.5 40.5

33.1 40.5 54.4 55.2 50.1 39.9

3.5 2.5 1.4 0.5 1.8 0.0

8.3 8.0 7.3 7.3 7.2 7.3

45.0 43.5 26.0 37.0 42.5 32.5

11.3 11.3 11.0 10.5 10.6 10.9

26.0 24.2 30.2 30.6 30.8 33.7

28.6 42.5 46.6 31.5 34.8 31.9

2.8 2.4

2.5 3.4 2.4

2.6 1.8 1.9 1.6

Legend: OC - organic carbon; CCE – calcium carbonate equivalent.

and Mg-samples, not affected by thermal treatment or glycol solvation. The most of the peaks related to 2:1 clays are unresolved and asymmetric. The intensity of main three phyllosilicate groups varies among soil samples. Kaolinite peaks increase their intensity from Convex to Concave surfaces, denoting higher proportions in the more weathered soils at lower positions of the terrain. Conversely, 2:1 clay peaks decrease from Convex to Concave positions, as expected because of the noticeable progression of weathering rates.

mineral (see the low 2θ angles in Fig. 2b) comprise the mineral assemblage of rock bulk samples. Considering the detailed investigation of clay fraction after XRD traces modeling (discussed in XRD modeling results section), the mica-group mineral is probably illite, derived from muscovite weathering or directly inherited from the calcareous parent material. This mineral assemblage in bedrock is similar in samples from the three studied soils. The mineral suite of the clay fraction in representative B horizons of the three soils indicates the presence of kaolinitic, smectitic and illitic minerals. The experimental XRD patterns of K/Mg-saturated and glycol solvated samples are displayed in Fig. 3, organized below the digital model elevation of the terrain, which helps to visualize differences in terrain morphology. Kaolinite was recognized in all treatments by ~ 7.15 and 3.60 Å peaks, which collapsed after K-saturated 500 °C treatment. Smectitic minerals were identified by combining the analysis of K/Mg-saturated samples: in K-samples, by peaks at 12 Å, collapsed to 10 Å after thermal treatment; in Mg-saturated samples, by unresolved peaks at 14 Å, which shifted towards 17 Å after glycol solvation. Illitic minerals were confirmed by the sequence of 10, 5 and 3.33 Å peaks in K

3.4. XRD modeling results The XRD modeling results confirm the general weathering pathways observed in the three different terrain surfaces and suggest a complex sequence of mineral transformations involving mixed-layered minerals as weathering rates become more intense, at the concave position. The parameters used for each mineral phase during XRD modeling procedure are shown in Table 2 and the calculated lines (fitted into experimental patterns) in Fig. 4. In all samples, a mix of discrete kaolinite and illite minerals, and kaolinite-smectite (K-S) and I-S mixed-layered Fig. 2. XRD patterns from the parent material natural (A) and residue (B) of each soil profile (Ca = calcite; Qz = quartz; Mi = Mica). The values displayed are dspacing in angstrom (Å).

4

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

Fig. 3. (A) XRD patterns (Vr = vermiculite; Il = ilite; ka = kaolinite; Sm = smectite; the values displayed are d-spacing in angstrom); (B) XRD modeling results for the representative soil profiles from convex, flat and concave surfaces (samples from the diagnostics subsurface horizons; convex: Bwkc2 and concave: Bwkc2), (C) and the TG/DTG curves.

Table 2 Parameters used in NEWMOD II calculations for Bwck2 horizons. Phase

Layer proportion (%)

Fe kaolinite

Fe smectite

Fe illite

Nave

N

% wt

Convex microrelief Ka 100 Il 100 Ka-Sm 80–20 Ka-Sm 65–35 Il-Sm 84–16 Il-Sm 71–29

< 0.01 – < 0.01 < 0.01 – –

– – 0.5 0.5 0.5 0.5

– 0.15 – – 0.15 0.15

5.0 4.8 1.7 1.0 2.5 2.5

17 17 12 10 14 12

21.2 19.6 33.0 6.7 10.7 8.8

Flat microrelief Ka 100 Il 100 Ka-Sm 90–10 Ka-Sm 65–35 Il-Sm 85–15

< 0.01 – < 0.01 < 0.01 –

– – 0.5 0.5 0.5

– 0.15 – – 0.15

5.0 4.8 1.8 1.0 1.0

17 17 17 10 10

31.5 17.2 38.6 5.8 6.9

Concave microrelief Ka 100 Il 100 Ka-Sm 95–5 Il-Sm 85–15

0.63 – 0.2 –

– – 0.5 0.5

– 0.15 – 0.15

4.8 4.0 1.5 1.0

17 12 10 10

50.5 11.0 31.1 7.4 Fig. 4. Experimental and calculated XRD patterns of the oriented glycolated specimens.

Legend: Fe = in the corresponding phase per O10(OH)2. N, Nave = range of number of layers and average number of layers in the coherent diffraction domain, respectively. Other parameters: R0, K+ in illite = 0.7 per O10(OH)2. Ka = Kaolinite; Il = Illite; KaSm = Kaolinite-smectite; Il-Sm = Illite-smectite.

5

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

minerals were detected. In the soil at the convex surface, six mineral phases are present, including the mixed-layered minerals K-S and I-S with different layer proportions (K-S: 80–20% and 65–35%; I-S: 84–16% and 71–29%). In the soil at Flat Surface, five minerals were recognized: two K-S mixed-layered minerals (90% and 65% of kaolinite layers, respectively), an I-S mixed-layered (85% of illite layers), and discrete kaolinite and illite. The soil sample located at the concave surface has four phases, including K-S (95–5%) and I-S (85–15%), discrete illite and the prevalence of discrete kaolinite. The modeling confirms that there is an increase of discrete kaolinite contents from soil at Convex surface towards the concave surface, from 21.2% to over 50%. For discrete illite, an opposite behavior is observed: the contents decrease from 19.6% on the convex surface to 11.0% on the concave surface. The types and quantities of mixed-layered minerals also vary in accordance to the position in the landscape. The soil sample of the convex surface has two different K-S minerals, which represent together 39.7% of all clay minerals (with layer proportions of 80–20% and 65–35%), and other two I-S minerals (with layer proportions 84–16% and 71–29%), representing 19.5% of all clay minerals. In flat surface, the K-S with intermediary layer proportions (65–35%) remains, but I-S minerals are represented by only one phase (layer proportions of 85–15%). At the concave position, there is only one K-S mineral, with a prevalence of kaolinite layers (95–5%), and the same I-S mineral observed at the Flat surface, rich in illite layers (85–15%). The normalized amounts of kaolinite, smectite and illite layers (in wt%), calculated from their proportions within each mixed-layered or discrete mineral, and its respective quantities in mineral assemblages of the clay fraction (Table 2), show the progressive enrichment of kaolinite layers in samples from Convex to Concave positions (Fig. 3B). At the same time, smectite and illite layers become less common, suggesting a gradual substitution as the weathering progress at the Concave position. The existence of several mixed-layered minerals with intermediary compositions and the progressive kaolinization observed towards Concave position are clearly associated with the weathering pathways linked to microrelief variations. The local changes in water circulation control the formation of pedogenetic minerals (Righi et al., 1999; Vingiani et al., 2004), and mixed-layered minerals are considered transitory minerals during sequential weathering mineral reactions in this context (Srivastava and Parkash, 1998; Fisher and Ryan, 2006; Dudek et al., 2006; Ryan and Huertas, 2009, 2013; Pincus et al., 2017). At Apodi Plateau, the semiarid climate may favor the formation of transitory mixed-layered minerals if water circulation is constrained. It takes place in the soil at the convex surface, where water availability is more limited (Fig. 1), resulting in higher amounts of intermediate species (K-S and I-S). As the microrelief enables a more efficient water infiltration (at the Flat and Concave positions), even in semiarid conditions, mixed-layered minerals with transitory compositions disappear from mineral assemblages, and compositions close to end-members prevail. The chemical mechanisms during mixed-layering reactions have been described in many works and involve local modifications at the phyllosilicate layer level. Illite or other mica-group minerals are usually transformed into vermiculite via this process (Banfield and Eggleton, 1988), but the direct transformation into smectite group minerals was also observed (Santiago Buey et al., 1998; Aldega et al., 2009). This is the pathway detected in the three soils, and explains the formation of detected I-S phases during transitions between the two end members. In the case of smectite-kaolinization, the mechanism involves the gradual tetrahedra stripping and inversion, followed by substitution of Al by Mg or Fe, originally present in smectite layers. It creates kaolinite patches in the original 2:1 layers, leading to the formation of K-S structure with increasing number of kaolinite layers as the chemical modifications proceed (Dudek et al., 2006). This model can explain the compositional variation observed for K-S minerals as the weathering rates progress

Table 3 Contents of hydration and structural water (DTG/TG curves) of clay samples from the representative soil profiles (convex, flat and concave surfaces). Microrelief

Hydration water

Structural water %

Convex surface Flat surface Concave surface

5.12 3.64 1.57

6.34 6.75 8.37

towards concave surfaces: at Convex and Flat positions, there is a K-S mineral with an intermediary composition between kaolinite and smectite, which disappears where the weathering rates are supposed to be higher (at concave). Part of smectite layers still remain in such conditions and are not completely transformed into kaolinite. At the concave position, where weathering rates are more pronounced, the transition between smectite and kaolinite seems to be almost complete, as intermediate compositions were not observed.

3.5. Thermal analysis of the clay minerals TG curves, interpolated with the DTG curves (Fig. 3C), reinforce the XRD modeling results. The sample from the convex surface showed the highest contents of hydration water (Table 3), which can be observed through the first pronounced peak in the DTG curve, until approximately 110 °C, related to the first inflection of the TG curve. In addition, this sample shows low amounts of structural water. Both characteristics can be associated with a large number of 2:1 layers existing in the material (structural water contents in 2:1 clay minerals are approximately 5%; see Mackenzie, 1957), evidenced in the modeled XRD patterns. In the sample from the flat surface relief, there is a decrease in the contents of water of hydration and a gradual increase in the value of structural water (Table 3), which is consistent with the reduction of expansive layers distributed in the various minerals present in the soil (see XRD modeling results). Differently from samples of convex and flat surface, samples of concave surface shows a considerable reduction in the contents of hydration water and an increase in structural water (in between discrete kaolinite and 2:1 clay minerals; Table 3), confirming the high contents of kaolinite layers in agreement with the XRD modeling results. The consistence between the data for the loss of hydration water in the clay minerals and the contents of expansive layers also reinforces the reliability of the modeling data of XRD patterns (Cuadros et al., 2013). The regions of the loss of structural water in the samples (250 °C and 750 °C) provide evidence for the differences between the soil formation processes with respect to microrelief, showing a decrease in the content of expansive layers from the convex to the concave surface. The sample of the convex surface shows a more asymmetric DTG curve, especially towards high temperatures. Although the peak of the endothermic reaction at ~480 °C is attributed to the loss of OH in the octahedral positions of the kaolinite layers, the previously highlighted asymmetry demonstrates the influence of smectite and illite layers on the samples (water loss events between 550 °C and 700 °C, see Drits et al., 1995, 2012), as well as the K-S phases, which are present in large amounts in this sample and can have a loss of structural water at approximately 550 °C (Dudek et al., 2006). The DTG curve of the samples at flat surface sample is similar to those of the convex surface relief sample, but with smaller asymmetries in the region of temperatures higher than 550 °C. This denotes the smaller influence of the smectite and illite layers, which are found in smaller amounts. However, the DTG curve from the concave surface soil is more symmetric, with a maximum at 550 °C, denoting less control of 2:1 layers, but also influenced by K-S phases. 6

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

4. Conclusions

Embleton-Hamann, C., 2004. Processes responsible for the development of a pit and mound microrelief. Catena 57 (2), 175–188. EMBRAPA, 2011. Manual de métodos de análises de solo. Rio de Janeiro, 2nd ed. (230 pp.). EMBRAPA, 2013. Sistema Brasileiro de Classificação do Solo, 2013. Rio de Janeiro, 3rd ed. (353 pp.). Ernesto Sobrinho, F., 1980. Caracterização, gênese e interpretação para uso de solos derivados de calcário da região da Chapada do Apodi, Rio Grande do Norte. In: Mestrado em Fitotecnia. Universidade Federal de Viçosa, Viçosa (1980. M.Sc. Dissertation). Ferreira, E.P., Anjos, L.H.C.D., Pereira, M.G., Valladares, G.S., Cipriano-Silva, R., Azevedo, A.C.D., 2016. Genesis and classification of soils containing carbonate on the Apodi Plateau, Brazil. Rev. Bras. Ciênc. Solo 40. Fisher, G.B., Ryan, P., 2006. The smectite-to-disordered kaolinite transition in a tropical soil chronosequence, Pacific Coast, Costas Rica. Clay Clay Miner. 54 (5), 571–586. Gee, G.W., Bauder, J.W., 1986. Particle-size analysis. In: Klute, A. (Ed.), Methods of Soil Analysis: Part 1. Physical and Mineralogical Methods, 2nd ed. Soil Science Society of America, Madison, pp. 383–411. Girão, R.O., Moreira, L.J.S., Girão, A.L.A., Romero, R.E., Ferreira, T.O., 2014. Soil genesis and iron nodules in a karst environment of the Apodi Plateau. Rev. Ciênc. Agron. 45 (4), 683–695. Huggett, R.J., 1975. Soil landscape systems: a model of soil genesis. Geoderma 13 (1), 1–22. IUSS Working Group WRB, 2015. World Reference for Soil Resources 2014, Update 2015 International Soil Classification System for Naming Soils and Creating Legends for Soil Maps (World Soil Resources Reports No. 106). FAO, Rome. Karathanasis, A.D., 2008. Thermal analysis of soil minerals. In: Ulery, A.L., Dress, L.R. (Eds.), Methods of Soil Analysis, Part 5 – Mineralogical Methods. Soil Sci. Soc. Am. J., Madison, pp. 117–160. Kilmer, V.J., 1965. Silicon. In: Black, C.A. (Ed.), Methods of soil analysis: chemical and microbiological properties. Am. Soc. Agron., Madison, pp. 959–962 (Agronomy, 9). King, G.J., Acton, D.F., Arnaud, R.J.St., 1983. Soil-landscape analysis in relation to soil distribution and mapping at a site within the Weyburn association. Can. J. Soil Sci. 63, 657–670. King, D., Bourennane, H., Isambert, M., Macaire, J.J., 1999. Relationship of the presence of a non-calcareous clay-loam horizon to DEM attributes in a gently sloping area. Geoderma 89, 95–111. Knight, M.J., 1980. Structural analysis and mechanical origins of gilgai at Boorook, Victoria, Australia. Geoderma 23 (4), 245–283. Lemos, M. do S. da S., Curi, N., Marques, J.J.G. de S.M., Sobrinho, F.E., 1997. Evaluation of characteristics of cambisols derived from limestone in low tablelands in northeastern Brazil: implication for management. Pesq. Agrop. Brasileira 8 (32), 825–834. Mackenzie, R.C., 1957. The Differential Thermal Analysis of Clays. Oxford Press; Mineral. Soc., London (456 pp.). Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by dithionite citrate systems buffered with sodium bicarbonate. In: 7th Nat. Conf. Clay Miner, pp. 317–327. Milne, G., 1935. Some suggested units for classification and mapping, particularly for East African soils. Soil Res. 4, 183–198 (Berlin). Mota, J.C.A., Assis Júnior, R.N., Amaro Filho, J., Romero, R.E., Mota, F.O.B., Libardi, P.L., 2007. Mineralogical attributes of three soils under melon cultivation in the Apodi Tableland, Northeastern Brazil. Rev. Bras. Ciênc. Solo 31, 445–454. Oliveira, D.P., 2009. Caracterização, gênese e mineralogia de solos da Chapada do ApodiCE: o papel do microrrelevo. 2009.79 f. Monografia (Graduação em Agronomia). Centro de Ciências Agrárias, Universidade Federal do Ceará, Fortaleza. Oliveira, D.P., Ferreira, T.O., Romero, R.E., Farias, P.R.S., Costa, M.C.G., 2013. Microrelief and particle size fraction distribution in calcareous Inceptisols. Rev. Ciênc. Agron. 44 (4), 676–684. Phillips, J.D., 2007. Development of texture contrast soils by a combination of bioturbation and translocation. Catena 70 (1), 92–104. Pincus, L.N., Ryan, P.C., Huertas, F.J., Alvarado, G.E., 2017. The influence of soil age and regional climate on clay mineralogy and cation exchange capacity of moist tropical soils: a case study from Late Quarternary chronosequences in Costa Rica. Geoderma 308, 130–148. Reynolds, R.C., Reynolds, R.C., 1996. NEWMOD for Windows a Computer Program for the Calculation of One-dimensional Diffraction Patterns of Mixed Layered Clays. (Published by the author, 8 Brook Dr., Hanover, New Hampshire, USA). Righi, D., Terribile, F., Petit, S., 1999. Pedogenic formation of kaolinite-smectite mixed layers in a soil toposequence developed from basaltic parent material in Sardinia (Italy). Clay Clay Miner. 47 (4), 505–514. Ruhe, R.V., 1975. Geomorphology: Geomorphic Processes and Superficial Geology. Hought-Miffin, Boston. Ryan, P.C., Huertas, J., 2009. The temporal evolution of pedogenic Fe–smectite to Fe–kaolin via interstratified kaolin–smectite in a moist tropical soil chronosequence. Geoderma 151, 1–15. Ryan, P.C., Huertas, J., 2013. Reaction pathways of clay minerals in tropical soils: insights from kaolinite-smectite synthesis experiments. Clay Clay Miner. 61 (4), 303–318. Sampaio, A.V., Schaller, H., 1968. Introdução a Estratigrafia Cretácea da Bacia Potiguar. Boletim Interno da Petrobrás. 11. pp. 19–44. Santiago Buey, E., Suárez Barrios, M., García Romero, E., Doval Montoya, M., Domínguez Díaz, M.C., 1998. Electron microscopic study of the bentonites from “Cerro del Aguila” (Toledo, Spain). Clay Miner. 33, 501–510. Schaetzl, R.J., 1990. Effects of treethrow microtopography on the characteristics and genesis of Spodosols, Michigan, USA. Catena 17 (2), 111–126. Schaetzl, R.J., 2008. Pedoturbation. In: Encyclopedia of Soil Science. Springer, Netherlands, pp. 516–522.

The physico-chemical and geochemical differences observed in the three soils developed in a calcareous tropical semiarid environment are closely related to microrelief variations. Solum thickness, clay and sand contents and CCE values vary according to soil position, denoting higher weathering rates where surface water concentrates, at the Concave position. Semi-total chemical compositions and Ki index also follow the same trend, evidencing the effect of microrelief on hydrological conditions and weathering rates at a local scale. The clay mineralogy of representative B horizons of each soil is consistent with variations on the mentioned attributes. There is a progressive increase in kaolinite contents towards the concave position, where the weathering rates are higher. The sequential transformation involving clays takes place via mixed-layering, evidenced by the presence of I-S and K-S, interpreted as transitory minerals during illite to smectite, and smectite to kaolinite transitions. These results emphasize the role of local relief conditions on the dynamic of mineral transformation in soils developed in the studied conditions. Acknowledgments The authors thank the financial support offered by the National Council of Technological and Scientific Development (CNPq, grant number 308288/2014-9), Coordenação de Aperfeiçoamento de Pessoal de Nivel Superior (CAPES), São Paulo Research Foundation (FAPESP, grant number 2014/117762). The present study was partially funded by Banco do Nordeste do Brasil S.A. (BNB). We also thank Gabriel Nuto Nóbrega for the assistance with the elaboration of Figures. References Ab'Saber, N.A., 1977. Os domínios morfoclimáticos na América do Sul: primeira aproximação. USP, São Paulo. Aldega, L., Cuadros, J., Laurora, A., Rossi, A., 2009. Weathering of phlogopite to beidellite in a karstic environment. Am. J. Sci. 309, 689–710. Alencar, E.L.L., 2002. Química e mineralogia de três pedons originários de calcário da Chapada do Apodi–CE. In: Soils and Plant Nutrition. Centro de Ciências Agrárias, Universidade Federal do Ceará, Fortaleza (83 p. M.Sc Dissertation). Allison, L.E., Moodie, C.D., 1965. Carbonate. In: Black, C.A. (Ed.), Methods of Soil Analysis. Part 2. Agron. Monogr 9. ASA, Madison, W.I., pp. 1379–1396. Atalay, I., 1997. Red Mediterranean soils in some karstic regions of Taurus Mountains, Turkey. Catena 28 (3–4), 247–260. Banfield, J.F., Eggleton, R.A., 1988. Transmission electron microscope study of biotite weathering. Clay Clay Miner. 36, 47–60. Barbosa, W.R., Romero, R.E., de Souza Júnior, V.S., Cooper, M., Sartor, L.R., de Moya Partiti, C.S., Jorge, F.O., Cohen, R., Jesus, S.L., Ferreira, T.O., 2015. Effects of slope orientation on pedogenesis of altimontane soils from the Brazilian semi-arid region (Baturité massif, Ceará). Environ. Earth Sci. 73 (7), 3731–3743. Bockheim, J.G., Tarnocai, C., 1998. Recognition of cryoturbation for classifying permafrost-affected soils. Geoderma 81, 281–293. Brasil. Ministério da Agricultura, 1973. Levantamento Exploratório Reconhecimento de Solos do Estado do Ceará. Theatr. Rec. I (301 pp.). Ceará, 1980. Comissão de Planejamento Agrícola. In: Projeto de desenvolvimento rural integrado do Ceará -Situação Geográfica, Recursos Humanos e Recursos Naturais, 1. 32 CEPA, Fortaleza 272 p. Conacher, A.J., Dalrymple, J.B., 1977. The nine unit land surface model: an approach to pedogeomorphic research. Geoderma 18, 1–154. Cuadros, J., Michalski, J.R., Dekov, V., Bishop, J., Fiore, S., Dyar, M.D., 2013. Crystalchemistry of interstratified Mg/Fe-clay minerals from seafloor hydrothermal sites. Chem. Geol. 360, 142–158. Da Costa, P.A., Mota, J.C.A., Romero, R.E., Freire, A.G., Ferreira, T.O., 2014. Changes in soil pore network in response to twenty-three years of irrigation in a tropical semiarid pasture from northeast Brazil. Soil Tillage Res. 137, 23–32. Demattê, J.A., Campos, R.C., Alves, M.C., Fiorio, P.R., Nanni, M.R., 2004. Visible–NIR reflectance: a new approach on soil evaluation. Geoderma 121 (1), 95–112. Drits, V.A., Besson, G., Muller, F., 1995. An improved model for structural transformation of heat-treated aluminous dioctahedral 2:1 layer silicates. Clay Clay Miner. 43 (6), 718–731. Drits, V.A., Derkowski, A., McCarty, D.K., 2012. Kinetics of partial dehydroxylation in dioctahedral 2:1 layer clay minerals. Am. Mineral. 97 (5–6), 930–950. Dudek, T., Cuadros, J., Fiore, S., 2006. Interstratified kaolinite-smectite: nature of the layers and mechanism of smectite kaolinization. Am. Mineral. 91, 159–170. Dudek, T., Cuadros, J., Huertas, J., 2007. Structure of mixed-layer kaolinite-smectite and smectite-to-kaolinite transformation mechanism from synthesis experiments. Am. Mineral. 92, 179–192.

7

Catena xxx (xxxx) xxx–xxx

D.P. Oliveira et al.

Srivastava, P., Parkash, B., 1998. Clay minerals in soils as evidence of Holocene climatic change, central Indo-Gangetic plains, North-Central India. Quat. Res. 50, 230–239. Van de Wauw, J., Baert, G., Moeyersons, J., Nyssen, J., De Geyndt, K., Taha, N., Zenebe, A., Poesen, J., Deckers, J., 2008. Soil–landscape relationships in the basalt-dominated highlands of Tigray, Ethiopia. Catena 75 (1), 117–127. Vingiani, S., Righi, D., Petit, S., Terribile, F., 2004. Mixed-layer kaolinite-smectite minerals in a red-black soil sequence from basalt in Sardinia (Italy). Clay Clay Miner. 52 (4), 473–483. Walkley, A., Black, I.A., 1934. An examination of Degtjareff method for determining soil organic matter, and proposed modification of the chromic acid tritation method. Soil Sci. 37, 29–38.

Schoeneberger, P.J., Wysocki, D.A., Benham, E.C., 2012. Soil survey staff. In: Field Book for Describing and Sampling Soils, Version 3.0. 36 Natural Resources Conservation Service, National Soil Survey Center, Lincoln, NE. Silva, M.B., dos Anjos, L.H.C., Pereira, M.G., Schiavo, J.A., Cooper, M., de Souza Cavassani, R., 2017. Soils in the karst landscape of Bodoquena plateau in cerrado region of Brazil. Catena 154, 107–117. Soukup, D.A., Buck, B.J., Harris, W., 2008. Preparing soils for mineralogical analyses. In: Ulery, A.L., Drees, R. (Eds.), Methods of Soil Analysis. Part 5 – Mineralogical Methods. SSSA Book Seriespp. 13–32 (Madison). Soyer, J., 1983. Microrelief de buttes basses sur sols inondés saisonnièrement au SudShaba (Zaïre). Catena 10 (3), 253–265.

8