Clay mineralogy and genesis of fragipan in soils from Southeast Brazil

Catena 135 (2015) 22–28

Contents lists available at ScienceDirect

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

Clay mineralogy and genesis of fragipan in soils from Southeast Brazil Michele Ribeiro Ramos a, Vander Freitas Melo b,⁎, Alexandre Uhlmann a, Renato Antonio Dedecek a, Gustavo Ribas Curcio a a b

Brazilian Agricultural Research Corporation - Embrapa Florestas, Estrada da Ribeira km. 111, CP 319, Colombo, PR, 83411–000, Brazil Department of Soil Science, Federal University of Paraná - UFPR, Rua dos Funcionários, 1540, Curitiba, PR, 80035–050, Brazil

a r t i c l e

i n f o

Article history: Received 21 October 2014 Received in revised form 20 June 2015 Accepted 26 June 2015 Available online xxxx Keywords: Fragipan horizons Kaolinite Amorphous materials Goethite Soil bulk density Soil porosity

a b s t r a c t Fragipan horizon has a hard consistency when dry but is brittle when moist. Such a horizon restricts root growth and water infiltration due to the low volume of macropores and discontinuous voids. In Rio de Janeiro state in Brazil, neighboring soils were developed from different materials (sediments and granite/gneiss) and were subject to the same environmental conditions; one manifested fragipan characteristics in the subsurface and the other did not. The main objective of this study was to characterize and quantify clay minerals and relate their properties to the genesis of fragipan horizons. The fragipan in the studied soils show high bulk density (1.67 g cm−3) and low average macroporosity (0.03 cm cm−3), total porosity (0.42 cm cm−3) and hydraulic conductivity (1.43 cm h−1). Sequential and interrelated causes favored the formation of fragipan horizons: 1) face to face adjustment of kaolinite (Ka) filling of larger spaces occurring between sand grains, favored by low goethite contents and the absence of hematite and gibbsite; and 2) mineral binding and cementation of sand, silt and clay fractions by amorphous materials. Higher amounts of goethite and lower amounts of amorphous materials in the clay fraction were associated with horizons with higher total porosity. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Cohesive behavior is a characteristic found in soils around the world (Chartres et al., 1990; Dexter, 2004; Ley et al., 1989; McKyes et al., 1994; Mullins et al., 1990; Prandel et al., 2014; Schjønning and Thomsen, 2013), but fragipan presents different cohesive features, as cementation is minimized by the presence of water (Embrapa, 2013; FAO, 2006; Soil Taxonomy, 1999). However, some cohesive horizons can be hard even under wet conditions as a result of iron oxide (laterite), calcium carbonate (petrocalcic horizons) and Si (duripan) accumulation (Franzmeier et al., 1996). Young (1992) identified cohesive soils in the United Kingdom based on the following features: 1) a texture that is typically within the sandy clay to loamy sand range; 2) low organic matter content (typically below 20 g kg− 1); 3) weakly structured and prone to slumping when wetted; 4) hard setting upon drying with an apedal structure, consequently exhibiting high tensile and shear strength values; 5) low content of shrink-swell clays; and 6) friability only over a limited moisture range. The notion of apedality was restored by including the term ‘structureless’ into the definition: ‘cohesive soils are soils that set to a hard, structureless mass during drying and are thereafter difficult or impossible to cultivate until the profile is rewetted’ (Mullins et al., 1990). ⁎ Corresponding author. E-mail address: [email protected] (V.F. Melo).

http://dx.doi.org/10.1016/j.catena.2015.06.016 0341-8162/© 2015 Elsevier B.V. All rights reserved.

The occurrence of cohesive horizons creates problems from a practical point of view, such as seedling emergence, blocking and resistance to root penetration, decreased productivity and greater effort for soil management (Daniells, 2012; Dexter, 2004; Greene et al., 2002; Mullins et al., 1987, 1990; Schjønning and Thomsen, 2013; Young, 1992). Bulk density of fragipan horizons can reach values as high as 1.7 Mg m− 3 (Young, 1992), and root growth and distribution are restricted to the plow layer, owing to the high penetration resistance exhibited in this soil (greater than 6 MPa) (Young et al., 1991). Under field conditions fragipan soils can be too hard to cultivate, producing a cloddy structure in a slightly moist state when mechanically plowed (McKyes et al., 1994). Some management techniques have been used to minimize the expression of the cohesive character in these soils: incorporation of maize residues (Mullins et al., 1987), deep tillage (Mead and Chan, 1992), an increased root system (permanent-pasture) (Chan, 1989), combined deep moldboard plowing and gypsum application (Hall et al., 1994), addition of ferrihydrite and Al oxides (Breuer and Schwertmann, 1999), treatment with anionic polymer (Chan and Sivapragasam, 1996) and no-till cropping (Ley et al., 1989). For a complete understanding of fragipan soils, studies of effects on physical characteristics, crop management, and plant growth should be accompanied by the genetic causes of cohesion. In general, the causes of cohesion of fragipan horizons vary with local soil and climate conditions and are associated with both physical and chemical processes. In Australia, it was associated with cycles of wetting and drying

M.R. Ramos et al. / Catena 135 (2015) 22–28

(Chartres et al., 1990); in soils from Denmark and Switzerland, by packing of sand grains in sandy soils with low organic carbon (Schjønning and Thomsen, 2013); in Nigeria (Ley et al., 1989) and Western Australia (Harper and Gilkes, 1994), to the increase of clay due to illuviation joining coarse soil particles. Some authors have concluded that the type of mineral in the clay fraction is more important than the amount of clay in promoting cohesion. The predominance of kaolinite in relation to the 2:1 expandable mineral hampers the development of structural cracks on drying and favors the formation of fragipan horizons (Mullins et al., 1987; Prandel et al., 2014). The juxtaposition of high crystallinity kaolinite was the cause of cohesion in kaolinitic soils formed from sediments in Brazil (Giarola et al., 2009). Mullins and Panayiotopoulos (1984) demonstrated that artificial mixtures of sand and as little as 2% kaolinite could exhibit cohesive behavior. Other studies reported the importance of amorphous materials and fresh gels of elements, especially Si and Fe in the cohesion of subsurface layers (Brown and Mahler, 1988). Chartres and Norton (1994) pointed to silica and less crystalline Fe oxide (ferrihydrite) bonds on the clay surfaces as a determination factor of cohesion. This chemical effect was increased by the presence of very fine sand and silt particles and low soil biological activity. The crust formed by the precipitation of amorphous silica gels on mineral surfaces under reduced humidity promotes the reduction of porosity and hydraulic conductivity of soils (Daniells, 2012; Mckyes et al., 1994; Mullins et al., 1990). In fragipan soils of Australia, the strength was correlated positively with extractable Si and negatively with extractable Al by ammonium oxalate (Franzmeier et al., 1996). This apparent divergence of information about crystalline and amorphous materials justifies the need for further investigation of the genesis of cohesive soils, especially under varying geological and soil conditions. The main objective of this work was to study the clay fraction and its properties (quantity, chemical composition and

23

crystallography) related to the genesis of fragipan horizons in two landscape positions. 2. Material and methods 2.1. Study area and soil sampling The study area is located in Itaboraí county, state of Rio de Janeiro, Brazil (Fig. 1). Regional climate is tropical, with a dry season in winter (one to three dry months) and a rainy season in summer from November to April. Average air temperature in the coldest month is above 18 °C and annual rainfall ranges from 1000 to 1500 mm (Ramos et al., 2013). Regional basin stratigraphy is formed by Macacu Formation sediments from Eocene/Oligocene period overlapping granite/gneiss layers of St. Fidelis Formation of the Proterozoic age (Ramos et al., 2013). The thickness of the sediments package is approximately 200 m. Two toposequences were studied (Table 1): toposequence 1 (fragipan soils) — Macacu Formation sediments; toposequence 2 (non-fragipan soils) — St Fidelis Formation (area without overlapping Macacu sediments). Non-fragipan soils were used as reference for discriminating soil attributes in the expression of cohesive character. Due to the small distance between toposequences (1 km), it can be assumed that the studied soils were subject to similar environmental conditions. The morphology of soil profiles was described according to FAO (2006). Disturbed and undisturbed samples were collected from all horizons of soil profiles. Disturbed samples were air dried and sieved through a 2 mm mesh for soil physical and chemical characterization (Embrapa, 1997): pH in 0.01 mol L−1 CaCl2 (soil/solution ratio 1: 2.5), nonexchangeable potential acidity (H) extracted with 0.5 mol L−1 pH 7 Ca acetate, exchangeable Ca2+, Mg2+ and Al3+ extracted with 1 mol L−1 KCl, exchangeable K+ in 0.05 mol L−1 H2SO4 and 0.025 mol L−1 HCl, organic carbon (OC) extraction with H2SO4 and potassium dichromate. Soil texture was determined using the pipette method.

Fig. 1. Location of the study area.

24

M.R. Ramos et al. / Catena 135 (2015) 22–28

Table 1 General characteristics and morphology of soil profiles. Parent material

Altitude above sea level (m)

Landscape position/ slope

Soil classesa

Horiz.

29

Shoulder 12%

Typic Haplustult

Ap BAx Btx1 Btx2

19

Foot slope 28%

Typic Haplustox

Ap BAx Bwx1 Bwx2

42

Shoulder 20%

Typic Haplustult

Ap BA Bt1 Bt2

20

Foot slope 29%

Typic Haplustult

Ap BA Bt1 Bt2

Depth (cm)

Moist color

Field textureb

Structurec

Consistencyd

10YR 3/2 10YR 4/5 10YR 5/7 9YR 5/8

SL SCL SCL SC

PT3; PG1/PG2; PZ2 PT6; PG1/PG2; PZ3/PZ4 PT6; PG1/PG2; PZ4 PT6; PG1; PZ4

CM2; CD1; CP1; CS1 CM2/CM3; CD5; CP2; CS2 CM2/CM3; CD4; CP2; CS2 CM2/CM3; CD4; CP2; CS2

10YR 3/2 10YR 4/5 10YR 5/7 10YR 5/8

SCL SCL SCL SCL

PT6; PG2; PZ2/PZ3 PT6; PG1/PG2; PZ4 PT6; PG1; PZ4 PT6; PG1; PZ4

CM2; CD1; CP1; CS1 CM2/CM3; CD5; CP2; CS2 CM2/CM3; CD4; CP2; CS2 CM2/CM3;CD4; CP2; CS2

10YR 3/3 7.5YR 4/6 7.5YR 5/7 7.5YR 5/7

SCL SC SC C

PT6; PG2; PZ2/PZ3 PT4; PG2; PZ2/PZ3 PT4; PG2; PZ1/PZ2 PT6; PG2; PZ1/PZ2

CM2; CD1; CP1; CS1 CM2/CM3; CD2; CP2; CS2 CM1/CM2; CD2; CP2; CS2 CM2; CD2; CP2; CS2

10YR 3/3 6YR 4/6 7.5YR 4.5/6 7.5YR 5/6

SC C C C

PT6; PG2; PZ1/PZ2 PT6; PG2; PZ1/PZ2 PT6; PG2; PZ2 PT6; PG2; PZ3/ PZ4

CM2; CD1; CP1; CS1 CM2; CD2; CP2; CS2 CM2; CD2; CP2; CS2 CM1/CM2; CD2; CP2; CS2

Toposequence 1 Macacu Formation (sediments)

Profile 1 0–18 18–32 32–61 61–75+ Profile 2 0–22 22–36 36–74 74–102+

Toposequence 2 St. Fidelis Formation (granite/gneiss)

Profile 3 0–11 11–34 34–55 55–98+ Profile 4 0–16 16–40 40–65 65–113

a

Soil Taxonomy (1999). Field Texture: SCL — sandy clay loam; SL — sandy loam; SC — sandy clay; C — clay. c Structure: type (PT3 — granular, PT6 — subangular blocky, PT4 — blocky); development degree (PG1 — Weak, PG2 — moderate); size (PZ1 — very fine, PZ2 — fine, PZ3 — medium, PZ4 — coarse). d Consistency: When moist (CM1 — very friable, CM2 — friable, CM3 — firm, CM4 — very firm, CM5 — extremely firm); When dry (CD1 — soft, CD2 — slightly hard, CD3 — hard, CD4 — very hard, CD5 — extremely hard); When wet: Plasticity (CP1 — slightly plastic, CP2 — plastic); Stickiness (CS1 — slightly sticky, CS2 — sticky). b

For the hydro-physical analysis, three undisturbed samples (replicates) were collected from each horizon using rings with 65 cm3 of volume (4 soils × 4 horizons × 3 rings = 48 undisturbed samples). These samples were used for the determination of bulk density, total porosity, macro- and microporosity, aeration porosity, available water, moisture at field capacity and saturated hydraulic conductivity. Except for saturated hydraulic conductivity, bulk density and porosity, the other parameters were determined in Richards Chamber after full water saturation of rings for 48 h (Embrapa, 1997). Soil field capacity was determined at 10 kPa tension and permanent wilting point at 1500 kPa. Soil macroporosity was determined at a tension table by using a water height column of 0.60 m. Total porosity was estimated by the difference between the weight of saturated soils and soil samples dried at 105 °C. The available water was defined as the difference in moisture volume between field capacity (−10 kPa) and the permanent wilting point (− 1500 kPa). Aeration porosity was determined by the difference between total porosity and moisture content at field capacity. To determine saturated hydraulic conductivity (Ks), undisturbed soil samples were placed in a constant-head permeameter and Ks was estimated from the Darcy equation (Embrapa, 1997).

2.2. Mineralogical analysis of the clay fraction Soil samples were treated with 30% (v/v) H2O2 to remove organic matter, and 0.2 mol L− 1 NaOH was added to promote dispersion (Jackson, 1979). Sand fraction was retained on a 0.05 mm sieve. Clay and silt fractions were separated by sedimentation based on Stokes law (Gee and Bauder, 1986). Identification of clay fraction minerals was conducted by X-ray diffraction (XRD) (non-oriented samples). The XRD patterns were obtained on a Philips vertical goniometer with a speed of 1 °2θ min−1 and a range of 2–65 °2θ. The diffractometer, equipped with a copper lamp and nickel filter, was operated at 20 kV and 40 mA using CuKα radiation.

Content and chemical composition of iron oxides of low and high crystallinity were respectively determined by the 0.2 mol L−1 pH 3.0 ammonium oxalate (AO) (McKeague, 1978) and citrate–bicarbonate– dithionite (CBD) (Mehra and Jackson, 1960) methods. Fe was determined by atomic absorption spectrophotometry. After washing residues of AO and CBD with a solution of 1 mol L−1 (NH4)2CO3 and deionized water to remove excess salt, the amount of material extracted was determined by the difference between dry weight (24 h at 60 °C) before and after treatments. The content of amorphous Fe oxides was estimated by the reduction of sample weight by treatment with AO (Melo et al., 2001a). Iron oxides in the clay fraction (hematite — Hm, and goethite — Gt) were concentrated with hot 5 mol L−1 NaOH (Norrish and Taylor, 1961) and 0.5 mol L− 1 HCl for the removal of residual sodalite (Singh and Gilkes, 1991). The residues were analyzed by XRD and NaCl was used as an internal standard for correction of instrumental distortions. The mixture (approximately 6%) was made by grinding the sample in a mortar in the presence of NaCl. In all samples only peaks of Gt were observed (ratio Gt/(Gt + Hm) = 1.0) (Torrent and Cabedo, 1986). The isomorphic substitution of Fe3+ by Al3+ (IS) in Gt was determined according to Schulze (1984). Mean crystal diameter (MCD) of Gt was calculated from the width at half height (WHH) of the reflections (110) and (111) using Scherrer equation (Klug and Alexander, 1954). The IS and the MCD of Hm were not determined because in all samples the Gt/(Gt + Hm) ratio was equal to 1.0. Gt contents in the clay fraction were estimated by allocating the amounts of crystalline Fe2O3 (Fe2O3CBD–Fe2O3AO) and considering the empirical formula of the mineral and the IS values (Melo et al., 2001a). CBD residue (kaolinite (Ka) concentrated) was analyzed by thermal analysis, using atmospheric N2 and a heating rate of 10 °C min−1. Quantification of Ka was performed according to the mass reduction of samples as a result of dehydroxylation of the mineral (Jackson, 1979). For transformation of Ka content of iron-free sample to natural clay fraction, we took the mass loss of the sample from CBD treatment into account.

M.R. Ramos et al. / Catena 135 (2015) 22–28

Ka crystallinity index (HBCI) was calculated according to recommendations of Hughes and Brown (1979). The MCD of Ka (001) was obtained using the Scherrer equation (Klug and Alexander, 1954). 2.3. Statistical analysis Mean values for chemical, physical and mineralogical parameters of subsurface BA and B horizons of both profiles along toposequences were compared by t test (5%), taking fragipan and non-fragipan soils from each area as a source of variation. Parameters showing significant difference were selected to form the original data matrix A (n × p); n being samples equivalent to observations and p being selected parameters significant to the variables. The matrix of correlations ρ (p × p) for these parameters was the basis for the application of principal component analysis (PCA), generating the associated eigenvalues and eigenvectors. Statistical analyses were performed in Statistica 6.0 software (Statsoft, 2007). 3. Results and discussion 3.1. Morphology and physical properties The morphological field data from subsurface BA and B soil horizons on toposequence 1 (Table 1) fell into the classical definition of fragipan (Daniells, 2012; Embrapa, 2013; Mullins et al., 1990; Soil Taxonomy, 1999; Young, 1992): low and weak degrees of structuring and development, hard to extremely hard consistency when dry and friable or firm when moist. The subsurface horizons of soils on toposequence 2 had higher total porosity and lower soil bulk density than the subsurface horizons of soils from toposequence 1 (significant at 5%) (Table 2) and were not classified as fragipan (Table 1). The physical parameters also showed the cohesive behavior of subsurface horizons on toposequence 1 (profiles 1 and 2, respectively) (Table 2): average soil bulk density (1.78 Mg m−3 and 1.56 Mg m−3), macroporosity (0.05 cm3 cm− 3 and 0.02 cm3 cm−3), total porosity (0.41 cm3 cm− 3 and 0.44 cm3 cm− 3) and hydraulic conductivity (0.63 cm h− 1 and 2.21 cm h− 1). These values were similar to those observed in other studies with fragipan horizons (Giarola et al., 2003; Young, 1992). The hydraulic conductivity can decrease from 37.7 cm h−1 in horizon A (non-cohesive) to 0.8 cm h−1 in the cohesive BAx horizon (Ramos et al., 2013). Ramos et al. (2013) used the same soil samples of the current study and attributed the greatest manifestation of fragipan character on profile1 (shoulder) compared to profile 2

25

(foot slope) on toposequence 1 to higher contents of available water throughout the year at the foot slope position. Sand content in fragipan soils was significantly higher compared to non-fragipan soils (Table 2). Most soils described in the literature as fragipan soils were presented as sandy loam, loamy and clayey-sandy texture (Daniells, 2012; Schjønning and Thomsen, 2013; Young, 1992). 3.2. Mineralogical properties Mineralogy of the clay fraction of subsurface horizons on both toposequences is kaolinitic. However, there were no significant differences in average contents of kaolinite (Ka) between non-fragipan soils of St. Fidelis Formation (722 g kg− 1) and fragipan soils of Macacu Formation (756 g kg−1) (Table 3). The low content of ferromagnesian minerals in the parent material of these soils, the abundance of feldspars in the granite/gneiss and the possibility of direct transformation of such minerals into Ka under tropical conditions (Melo et al., 2001b; Nwadialo and Lietzke, 1989; Rebertus et al., 1986) favored the concentration of 1:1 mineral in the clay fraction of all profiles (Fig. 2). The average content of goethite (Gt) was higher in the clay fraction of non-fragipan soils (Table 3). Gt peaks in (110) (0.417 nm) and (130) (0.266 nm) directions (21.3 and 33.6 o2θ – CuKα radiation – respectively) were more intense in the two non-cohesive soils (Fig. 2). The soils in the higher landscape positions (edges of the basin) originated from granite/gneiss (St. Fidelis Formation) erosion, and the sediments were deposited in humid conditions in the lower parts of the basin, resulting in the Macacu Formation (Ramos et al., 2013). The same removal behavior of pedogenic Fe oxides during a wet sedimentation process was observed by Melo et al. (2001a) in fragipan soils formed from Tertiary sediments of Barreiras Formation on the coast of Brazil. The clay fraction did not show characteristic reflections of hematite (Hm) and gibbsite (Gb) by X-ray diffraction (Fig. 2), indicating that the environment preferentially synthesizes Gt and Ka. The main factors that favor Gt formation in soil are lower levels of iron in the parent material, low temperature, high moisture and organic matter content and low pH values (Schwertmann and Kämpf, 1985). Average amounts of amorphous materials were significantly higher in the subsurface horizons of fragipan soils compared with non-fragipan soils (Table 3). The lowest macroporosity and hydraulic conductivity values in fragipan horizons (Table 2) favor the accumulation of water, and the formation and retention of amorphous materials in the soil. Hydromorphic conditions hinder crystallization of Fe and Al oxides (Duarte et al., 2012; Melo et al., 2001a). In fragipan soils, which tend to have impeded lower drainage, Fe oxide minerals periodically

Table 2 Physical properties of fragipan (1 and 2) and non-fragipan (3 and 4) soils. Profile

Horizon

Porosity Total 3

(cm cm P1

P2

P3

P4

BAx Btx1 Btx2 BAx Bwx1 Bwx2 Mean BA Bt1 Bt2 BA Bt1 Bt2 Mean p

0.41 0.40 0.41 0.42 0.42 0.47 0.42 0.48 0.50 0.52 0.49 0.52 0.52 0.50 0.00⁎

Macro

Aeration

FCa

AWa

−3

)

SBDa (g cm

0.06 0.05 0.04 0.02 0.01 0.02 0.03 0.09 0.94 0.12 0.07 0.12 0.09 0.24 0.18

0.10 0.07 0.06 0.11 0.11 0.15 0.10 0.10 0.13 0.14 0.08 0.12 0.09 0.11 0.59

0.31 0.33 0.34 0.31 0.31 0.32 0.32 0.38 0.37 0.38 0.41 0.40 0.43 0.39 0.00⁎

0.08 0.07 0.06 0.05 0.06 0.08 0.07 0.04 0.04 0.05 0.05 0.06 0.05 0.05 0.01⁎

a FC — Field capacity; AW — Available water; SBD — Soil bulk density; HC — Hydraulic conductivity. ⁎ Significant difference between means (p b 0.05).

1.78 1.78 1.77 1.58 1.57 1.52 1.67 1.42 1.29 1.31 1.46 1.22 1.34 1.34 0.00⁎

HCa −3

)

(cm h 0.83 0.48 0.65 2.35 2.62 1.66 1.43 0.71 3.75 4.93 1.63 4.65 5.42 3.51 0.04⁎

Clay −1

)

(g kg

Silt

Sand

174 116 31 121 90 95 105 137 121 139 146 118 56 120 0.53

600 559 519 529 585 555 557 338 338 311 304 207 269 294 0.00⁎

−1

226 325 450 350 325 350 338 525 541 550 550 675 675 586 0.00⁎

)

26

M.R. Ramos et al. / Catena 135 (2015) 22–28

Table 3 Mineralogical properties of fragipan (1 and 2) and non-fragipan (3 and 4) soils. Profile

Horizon

Kaa

P1

BAx Btx1 Btx2 BAx Bwx1 Bwx2 Mean BA Bt1 Bt2 BA Bt1 Bt2 Mean p

761 760 762 776 695 783 756 679 661 811 785 778 619 722 0.35

Gta

AMa

KaCIa

54 54 59 51 49 48 53 109 128 97 98 102 97 105 0.00⁎

70 66 59 55 51 65 61 21 14 19 35 34 24 25 0.00⁎

13.2 12.2 14.8 13.9 13.2 13.5 13.5 15.8 13.4 12.4 14.6 12.6 13.1 13.7 0.78

(g kg−1)

P2

P3

P4

DTKaa

MCDKaa(001)

(°C)

(nm)

496 487 487 413 491 491 477 497 491 492 493 495 492 493 0.25

15.2 20.5 14.8 14.4 14.4 15.2 15.7 14.8 15.2 15.2 14.8 15.2 20.5 15.9 0.88

MCD Gta (110)

MCD Gta (111)

13.0 19.1 19.1 28.3 28.3 19.1 21.1 19.7 19.1 19.1 19.1 19.1 20.4 19.4 0.50

19.8 15.0 17.1 23.6 29.3 19.8 20.7 19.8 17.1 19.8 21.2 21.2 20.5 19.9 0.71

WHHKaa(001)

WHH Gta (111)

(mol mol−1)

(o2θ) 0.65 0.53 0.66 0.67 0.67 0.65 0.64 0.66 0.65 0.65 0.66 0.65 0.53 0.63 0.84

Gt ISa

0.59 0.70 0.65 0.53 0.47 0.59 0.59 0.59 0.65 0.59 0.56 0.56 0.57 0.59 0.96

0.13 0.12 0.21 0.15 0.03 0.26 0.15 0.21 0.29 0.24 0.21 0.20 0.29 0.24 0.04⁎

a Ka — kaolinite; Gt — goethite; AM — amorphous materials extracted with ammonium oxalate; KaCI — kaolinite crystalline index (values without units); DTKa — dehydroxylation temperature of kaolinite; MCDKa — mean crystal diameter of kaolinite; MCDGt — mean crystal diameter of goethite; WHHKa — width at half high of kaolinite peak; WHHGt — width at half high of goethite peak; GtIS — isomorphous substitution of Fe by Al in goethite structure. ⁎ Significant difference between means (p b 0.05).

precipitate and dissolve (Franzmeier et al., 1996). In our fragipan soil profiles, these conditions were not evident enough to form redoximorphic pedofeatures (Table 1). In relation to crystallographic parameters of clay fraction minerals, only the isomorphic substitution of Fe3+ by Al3+ (IS) in the Gt structure was significantly lower in fragipan soils compared to non-fragipan soils (Table 3). Iron oxides with higher IS have a lower degree of crystallinity (Fontes and Weed, 1991; Melo et al., 2001a), which should have favored the preferential removal of this group of Gt during wet sedimentation processes (Macacu Formation) and the concentration of Gt with lower IS values in the area of fragipan soils. Water excess during sediment deposition favors removal of Fe oxides and the dissolution process is

facilitated by the presence of lower crystallinity minerals of higher surface area (higher IS). 3.3. Chemical properties All subsurface horizons showed organic carbon concentrations below 20 g kg−1 (Table 4), considered the upper limit usually found in fragipan horizons (Young, 1992). There was only a tendency of non-fragipan soils towards having higher amounts of organic carbon (no significant difference). On the other hand, fragipan soils showed higher CEC at pH 7.0 (Table 4). The higher Ka (although not significant) contents favor the formation of negative charges in fragipan soils, as this mineral has a pH point of zero charges (pHPCZ) value as low as 3.0 (Brian and Sposito, 1997). Soil pH in CaCl2 for the two toposequences ranged from 3.6 to 3.9. Further, beyond higher CEC, the fragipan soils store more water for longer time due to their higher amount of micropores and lower hydraulic conductivity, making nutrient leaching slower, and the fragipan system more conservative than the neighboring areas of non-fragipan soils. The higher Fe2O3–CBD for non-fragipan soils (Table 4) is coherent with their higher Gt contents (Table 3). CBD extraction dissolves all pedogenic iron oxides from the clay fraction (Ghidin et al., 2006; Mehra and Jackson, 1960). As the iron from Gt dissolution prevails in relation to amorphous Fe oxides (AO extraction), Fe2O3-crystalline was also higher in non-fragipan soils (p b 0.05). 3.4. Fragipan genesis

Fig. 2. X-ray diffraction (CuKα radiation) patterns of clay fractions of BAx horizon of fragipan (P1 and P2) and BA horizon of non-fragipan (P3 and P4) soils. Ka — kaolinite, Gt — goethite, Mi — mica.

Principal component analysis (PCA) of the hydro-physical, chemical and mineralogical variables resulted in the extraction of 11 major axes, or components, associated with the correlation matrix. The first two components explained 83.8% of the total variance (eigenvalue 1 = 75.1%, eigenvalue 2 = 8.7%) (Fig. 3). The first axis of PCA showed a strong division into two soil groups. The non-fragipan profiles were set in the negative quadrants of axis 1, and the positive quadrant of the same axis contained the fragipan profiles. The vectors may be interpreted according to their length and parallelism in relation to the vertical axis. The longer vectors roughly parallel to the axis have higher correlations with the first principal component, which was responsible for the segregation of soil into two groups. Thus, it can be interpreted that the variables Gt contents, total porosity, field capacity, clay contents, IS in Gt and saturated hydraulic conductivity are higher for non-fragipan soils. On the other hand, variables that were higher for samples of fragipan horizons were sand content,

M.R. Ramos et al. / Catena 135 (2015) 22–28 Table 4 Chemical properties of fragipan (1 and 2) and non-fragipan (3 and 4) soils. Profile

P1

P2

P3

P4

Horizon

BAx Btx1 Btx2 BAx Bwx1 Bwx2 Mean BA Bt1 Bt2 BA Bt1 Bt2 Mean p

CECa

SBa

OCa

Fe2O3 CBDa

(cmolc kg−3)

(g kg−3)

(g kg−1)

7.7 7.5 6.0 7.5 6.5 6.5 7.0 6.5 6.5 6.0 6.1 6.0 5.2 6.1 0.03⁎

11.5 5.1 2.4 12.4 2.4 3.3 6.2 12.4 9.6 6.0 6.0 7.8 9.6 8.6 0.29

47.2 50.7 54.2 46.6 45.4 43.4 47.9 98.4 113.9 93.8 88.8 91.8 86.8 95.6 0.00⁎

0.45 0.33 0.22 0.31 0.31 0.31 0.32 0.33 0.31 0.21 0.32 0.21 0.21 0.27 0.18

F2O3 AOa

Fe2O3 crystallinea

4.55 2.90 2.00 1.51 1.44 1.43 2.31 1.78 1.47 1.39 1.65 1.35 0.99 1.44 0.12

42.7 47.8 52.1 45.0 43.9 42.0 45.6 96.6 112.4 92.4 87.2 90.4 85.8 94.1 0.00⁎

a CEC — cation exchange capacity; SB — Sum of bases; OC — organic carbon; CBD — citrate–bicarbonate–dithionite; AO — ammonium oxalate; Fe2O3 crystalline = Fe 2O3 CBD–Fe2O3 AO. ⁎ Significant difference between means (p b 0.05).

amorphous material contents, bulk density, available water and cation exchange capacity. The two main mineralogical variables responsible for the separation of soil samples into two groups were the higher Gt contents in non-fragipan soils and the higher amorphous mineral contents (AO extraction) in fragipan soils (Fig. 3). High contents of Ka in soils (Table 3 and Fig. 2) favor the face to face adjustment of this phyllosilicate mineral, the packing of micromass and the expression of the fragipan character (Giarola et al., 2003, 2009; Mullins and Panayiotopoulos, 1984). The higher levels of Gt in non-fragipan soils (Table 3) hinder this juxtaposition of crystals in these clay fraction minerals. Thus, the highest Gt content corresponds to the highest degree of microscopic disorganization and, consequently, to better structuring and lower soil cohesion in the dry state (Ghidin et al., 2006). The correlation coefficient between Gt contents and total porosity in subsurface soil horizons was positive and significant (r = 0.80,

Fig. 3. Ordination diagram and correlations of the clay mineralogical and physical and chemical parameters of subsurface BA and B horizons of fragipan (Toposequence 1 — profiles P1 and P2) (●) and non-fragipan (Toposequence 2 — P3 and P4 profiles) soils (○) obtained by principal component analysis. Axis 1: 76% of the total variance; Axis 2: 9% of the total variance. CEC — cation exchange capacity at pH 7.0; Gt — goethite.

27

p b 0.05). This disorganizing effect (Ferreira et al., 1999) is facilitated by the isodimensional shape of Fe oxides (Fontes and Weed, 1991; Schwertmann and Kämpf, 1985). In fragipan soils with lower Gt contents, the higher amorphous mineral contents promote bonds between mineral particles (Brown and Mahler, 1988; Chartres and Norton, 1994; Daniells, 2012). The correlation between contents of amorphous minerals and total porosity was negative (r = −0.82, p b 0.05). The significant positive correlation between sand contents and fragipan soil density values (vectors almost coincident and in the same direction in Fig. 3) was expected because sand grains reduce microporosity and total porosity of soils (Chaudhari et al., 2013). Aside from the absence of micropores, sand grains in fragipan soils usually have very heterogeneous sizes, which facilitate the filling of larger pores and the reduction of total porosity (Lima et al., 2004). Consequently, sand content was also important to discriminate between fragipan and non-fragipan soils. As a result, sequential and interrelated causes favored the formation of fragipan horizons on toposequence 1: 1) face to face adjustment of Ka filling larger spaces between sand grains, favored by low Gt contents and absence of Hm and Gb; and 2) mineral binding and cementation of sand, silt and clay fraction by amorphous materials. Other researchers identified such mechanisms in the genesis of fragipan soils in a more isolated form: more sandy soils and sand grain size dispersion (Lima et al., 2004), ratio between phyllosilicate minerals and Fe oxides (Giarola et al., 2003, 2009; Mullins and Panayiotopoulos, 1984), fresh gels of Si and Fe and amorphous materials (Brown and Mahler, 1988; Chartres and Norton, 1994; McKyes et al., 1994). Based on results from the higher closure matrix, the van der Waals attractive forces between adjusted Ka layers become more intense and cohesion forces more expressive, making the consistency of fragipan horizons very hard or extremely hard in the dry state (Zhang and Low, 1989). With increasing soil moisture, the amorphous materials bonds are weakened and the entry of water among individual juxtaposed Ka particles reduces the cohesion forces and makes fragipan horizons friable/firm. These processes of increasing and decreasing cohesion are reversible depending on soil moisture (Daniells, 2012; Embrapa, 2013; FAO, 2006; Soil Taxonomy, 1999). The only crystallographic parameter that was important for the separation of non-fragipan and fragipan soils was the IS in Gt (Fig. 3). From the position and direction of vectors it can be concluded that Gt with greater presence of Al3+ in the structure contributed to an increase in total soil porosity. Due to the smaller ionic radius of Al3+ compared to Fe3+ there is a reduction in the size of the unit cell with the IS, especially in the direction of the Z axis (Fitzpatrick and Schwertmann, 1982). Smaller Gt particles (higher specific surface area) should be more efficient in hindering the juxtaposition of Ka particles in non-fragipan soils. There was no effect of Ka crystal size and crystallinity on the separation of fragipan and non-fragipan soils into groups (Fig. 3). The values for Hughes and Brown crystallinity index (HBCI) were similar between both groups of soils (Table 3). Giarola et al. (2009) studied various soils on the Brazilian coast by using the Rietveld method and also found no differences in crystallinity of kaolinites of fragipan and non-fragipan soils. The median value for the HBCI (13.6) is significantly higher than the values reported for kaolinite from highly weathered soils. For example, median values of 5.4 and 5.8 were reported for kaolinites from West Australian soils (Koppi and Skjemstad, 1981). According to Melo et al. (2001b), in a clay fraction with absolute predominance of Ka, independent of the parent material, the interference from other elements and minerals is lower and the crystallinity of this mineral becomes larger. This behavior is easily observed in kaolin mines, where the crystals are larger, have high crystallinity and present a hexagonal morphology (6 euhedral faces) (Melo et al., 2001b). HBCI values (Fig. 3), however, were lower than for standard kaolinites. Values between 38 and 83 were found for standard materials, including Georgia kaolinite (Hughes and Brown, 1979).

28

M.R. Ramos et al. / Catena 135 (2015) 22–28

4. Conclusions 1. The clay mineralogy parameters that influenced positively the fragipan genesis in subsurface horizons were lower contents of goethite and goethite crystals with less isomorphic substitution of Fe3+ by Al3+, and higher contents of amorphous materials. 2. As a practical result of the packing and closure of the matrix, the fragipan horizons showed a consistency of very or extremely hard when dry, and lower macroporosity, total porosity and hydraulic conductivity in relation to non-fragipan subsurface horizons. Acknowledgments We thank Professor Valmiqui Costa Lima by supporting this work and Maria Aparecida Carvalho, soil lab technician, for helping with the preparation of soil samples and mineralogy analysis. References Breuer, J., Schwertmann, U., 1999. Changes to cohesive properties of soil by addition of metals hydroxides. Eur. J. Soil Sci. 50, 657–664. Brian, K.S., Sposito, G., 1997. Surface charge properties of kaolinite. Eur. J. Soil Sci. 45, 85–91. Brown, T.H., Mahler, R.L., 1988. Relationships between soluble silica and plow pans in Palouse silt loam soils. Soil Sci. 145, 359–364. Chan, K.Y., 1989. Friability of a cohesive soil under different tillage and land use practices. Soil Tillage Res. 13, 287–298. Chan, K.Y., Sivapragasam, S., 1996. Amelioration of a degraded cohesive soil using an anionic polymeric conditioner. Soil Technol. 9, 91–100. Chartres, C.J., Norton, L.D., 1994. Micromorphological and chemical properties of Australian soils with cohesive and duric horizons. Dev. Soil Sci. 22, 825–834. Chartres, C.J., Kirby, J.M., Raupach, M., 1990. Poorly ordered silica and aluminosilicates as temporary cementing agents in hard-setting soils. Soil Sci. Soc. Am. J. 54, 1060–1067. Chaudhari, P.R., Ahire, D.H., Ahire, V.D., Chkravarty, M., Maity, S., 2013. Soil bulk density as related to soil texture, organic matter content and available total nutrients of coimbatore soil. Int. J. Sci. Res. Publ. 3, 1–8. Daniells, I.G., 2012. Cohesive soils: a review. Soil Res. 50, 349–359. Dexter, A.R., 2004. Soil physical quality: Part II. Friability, tillage and hard-setting. Geoderma 120, 215–225. Duarte, A.P., Melo, V.F., Brown, G.G., Pauletti, V., 2012. Changes in the forms of lead and manganese in soils by passage through the gut of the tropical endogenic earthworm (Pontoscolex corethrurus). Eur. J. Soil Biol. 53, 32–39. Embrapa — Empresa Brasileira de Pesquisa Agropecuária, 2013. Sistema Brasileiro de Classificação de Solos. third ed. Embrapa, Rio de Janeiro, Brazil. Embrapa — Empresa Brasileira de Pesquisa Agropecuária, Centro Nacional de Pesquisa de Solos, 1997. Manual de métodos de análise de solos. second ed. Embrapa, Rio de Janeiro, Brazil. FAO, 2006. Guidelines for Soil Description. 4th edition. FAO, Rome. Ferreira, M.M., Fernandes, B., Curi, N., 1999. Mineralogia da fração argila e estrutura de latossolos da Região Sudeste do Brasil. Rev. Bras. Ciênc. Solo 23, 507–514. Fitzpatrick, R.W., Schwertmann, U., 1982. Al-Substituted goethite an indicator of pedogenic and other weathering environments in South Africa. Geoderma 27, 335–347. Fontes, M.P.F., Weed, S.B., 1991. Iron oxides in selected Brazilian Oxisols. I. Mineralogy. Soil Sci. Soc. Am. J. 55, 1143–1149. Franzmeier, D.P., Chartres, C.J., Wood, J.T., 1996. Cohesive soils in southeast Australia: landscape and profile processes. Soil Sci. Soc. Am. J. 60, 1178–1187. Gee, G.W.E., Bauder, J.W., 1986. Particle-size analysis. In: Klute, A. (Ed.), Methods of Soil Analysis. American Society of Agronomy, Madison WI, pp. 383–412. Ghidin, A.A., Melo, V.F., Lima, V.C., Lima, J.M.J.C., 2006. Toposseqüências de latossolos originados de rochas basálticas no Paraná: II Influência dos minerais da fração argila nas propriedades físicas dos solos. Rev. Bras. Ciênc. Solo 30, 307–319. Giarola, N.F.B., Silva, A.P., Imhoff, S., Dexter, A.R., 2003. Contribution of natural soil compaction on cohesive behavior. Geoderma 113, 95–108. Giarola, N.F.B., Lima, H.V., Romero, R.E., Brinatti, A.M., Silva, A.P., 2009. Mineralogia e cristalografia da fração argila de horizontes coesos de solos nos Tabuleiros Costeiros. Rev. Bras. Ciênc. Solo 33, 33–40. Greene, R.S.B., Eggleton, R.A., Rengasamy, P., 2002. Relationship between clay mineralogy and the cohesive properties of soils in the Carnarvon horticultural district of Western Australia. Appl. Clay Sci. 20, 211–223.

Hall, D.J.N., McKenzie, D.C., MacLeod, D.A., Barrett, A., 1994. Amelioration of a cohesive Alfisol throughdeep moldboard ploughing, gypsum application and double cropping. I. Soil physical and chemical properties. Soil Tillage Res. 28, 253–270. Harper, R.J., Gilkes, R.J., 1994. Cohesive in the surface horizons of sandy soils and its implications for soil classification and management. Aust. J. Soil Res. 32, 603–661. Hughes, J.C., Brown, G.A., 1979. Crystallinity index for soil kaolinite and its relation to parent rock, climate and soil maturity. J. Soil Sci. 30, 557–563. Jackson, M.L., 1979. Soil Chemical Analysis — Advanced Course. Prentice-Hall, Madison WI. Klug, H.P.E., Alexander, L.E., 1954. X-ray Diffraction Procedures for Polycrystalline and Amorphous Materials. John Wiley & Sons, New York. Koppi, A.J., Skjemstad, J.O., 1981. Soil kaolins and their genetic relationships in southeast Queensland. Aus. J. Soil Sci. 32, 661–672. Ley, G.J., Mullins, C.E., Lal, R., 1989. Cohesive behaviour of some structurally weak tropical soils. Soil Tillage Res. 13, 365–381. Lima, H.V., Silva, A.P., Jacomine, P.T.K., Romero, R.E., Libardi, P.L.L., 2004. Identificação e caracterização de solos coesos no Estado do Ceará. Rev. Bras. Ciênc. Solo 28, 467–476. McKeague, J.A., 1978. Manual on Soil Sampling and Methods of Analysis. Canadian Society Soil Science, Ottawa. McKyes, E., Nyamugafata, P., Nyamapfene, K.W., 1994. Characterization of cohesion, friction and sensitivity of two cohesive soil from Zimbabwe. Soil Tillage Res. 29, 357–366. Mead, J.A., Chan, K.H., 1992. Cultivation techniques and grazing affect surface structure of an Australian cohesive soil. Soil Tillage Res. 25, 217–230. Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clay by a dithionite– citrate system buffered with sodium bicarbonate. Clays Clay Miner. 7, 317–327. Melo, V.F., Singh, B., Schaefer, C.E.G.R., Novais, R.F., Fontes, M.P.F.F., 2001a. Chemical and mineralogical properties of kaolinite rich Brazilian soils. Soil Sci. Soc. Am. J. 65, 1324–1333. Melo, V.F., Fontes, M.P.F., Novais, R.F., Singh, B., Schaefer, C.E.G.R., 2001b. Características dos óxidos de ferro e de alumínio de diferentes classes de solos. Rev. Bras. Ciênc. Solo 25, 19–32. Mullins, C.E., Panayiotopoulos, K.L., 1984. The strength of unsaturated mixtures of sand and kaolin and the concept of effective stress. J. Soil Sci. 35, 459–468. Mullins, C.E., Young, I.M., Bengough, A.G., Ley, G.J., 1987. Cohesive soils. Soil Use Manag. 3, 79–83. Mullins, C.E., Macleod, D.A., Northcote, K.H., Tisdall, J.M., Young, I.M., 1990. Cohesive soils: behavior, occurrence and management. Adv. Soil Sci. 11, 37–108. Norrish, K., Taylor, M., 1961. The isomorphous replacement of iron by aluminium in soil goethites. J. Soil Sci. 12, 294–306. Nwadialo, B.E., Lietzke, D.A., 1989. Mineralogy and weathering of soils in the Tennessee Copper Basin. Soil Sci. 147, 162–173. Prandel, A.N., Saab, S.C., Brinatt, A.M., Giarola, N.F.B., Leite, W.C., Cassaro, F.A.M., 2014. Mineralogical analysis of clays in cohesive soil horizons by X-ray fluorescence and X-ray diffraction using Rietveld method. Radiat. Phys. Chem. 95, 65–68. Ramos, M.R., Curcio, R.G., Dedecek, R.A., Melo, V., de, F., Uhlmann, A., 2013. Influência da posição na encosta na manifestação do caráter coeso em solos da Formação Macacu no Estado do Rio de Janeiro. Rev. Bras. Ciênc. Solo 37, 837–845. Rebertus, R.A., Weed, S.B., Buol, S.W., 1986. Transformations of biotite to kaolinite during saprolite-soil weathering. Soil Sci. Soc. Am. J. 50, 810–819. Schjønning, P., Thomsen, I.K., 2013. Shallow tillage effects on soil properties for temperate-region hard-setting soils. Soil Tillage Res. 132, 12–20. Schulze, D.G., 1984. The influence of aluminium on iron oxides. VIII-unit-cell dimensions of Al-substituted goethites and estimation of Al from them. Clays Clay Miner. 32, 36–44. Schwertmann, U., Kämpf, N., 1985. Properties of goethite and hematite in kaolinitic soils of Southern and Central Brazil. Soil Sci. 139, 344–350. Singh, B., Gilkes, R.J., 1991. Concentration of iron oxides from soil clays by 5 M NaOH treatment: the complete removal of sodalite and kaolin. Clays Clay Miner. 26, 463–472. Soil Taxonomy, 1999. A basic system of soil classification for making and interpreting soil surveys. Agriculture Handbook No 436. United States Department of Agriculture, Washington DC. Statsoft, I., 2007. Statistica (Data Analysis Software System). Torrent, J., Cabedo, A., 1986. Sources of iron oxides in reddish brown soil profiles from calcarenites in Southern Spain. Geoderma 37, 5766. Young, L.M., 1992. Cohesive soils in the UK. Soil Tillage Res. 25, 187–193. Young, I.M., Mullins, C.E., Costigan, P.A., Bengough, A.G., 1991. Cohesive and structural regeneration in two unstable Bristish sand loams and their influence on crop growth. Soil Tillage Res. 19, 383–394. Zhang, Z.Z., Low, P.F., 1989. Low Relation between the heat of immersion and the initial water content of Li-, Na-, and K-montmorillonite. J. Colloid Interface Sci. 133, 461–472.