Inheritance and formation of smectite in a texture contrast soil in the Pilliga State Forests, New South Wales.
In a study examining the relationship between the orientation and spacing of coarse, columnar peds in an acid, texture-contrast soil and jointing in the underlying bedrock, Walsh and Humphreys (2008) reported the presence of large amounts of smectite in the saprolite of an in situ weathered soil in the Pilliga State Forests in north-western New South Wales, Australia. Smectite genesis is generally considered to require an alkaline environment, for in acid environments it is reported to be unstable and will tend to alter to chlorite or kaolinite (e.g. Williams et al. 1953; Grim 1968; Millot 1970). However, Douglas (1981) argued that this viewpoint, held by many clay mineralogists, is only partially supported by the literature. He reported smectite in 3 highly acidic soils and sediments, and suggested that its formation is dependant on both the type of parent material and a soil environment that inhibits the formation of aluminium polymers, conditions which would be met in a soil either with a low pH, or where organic Al-complexing agents are present. Karathanasis and Hajek (1984) reported that smectite can be formed and preserved in soil with pH as low as 3.6, but only where the concentrations of Si and Mg in the soil solution remain high and leaching is restricted. Righi et al. (1999) found that smectite was the end product of mica alteration in strongly leached and acidified E horizons of Podzols in moraines in Switzerland, while Egli et al. (2004, 2006) reported the development of smectite in acid Alpine soils in northern Italy.
The elucidation of the smectite's origin in the saprolite of the study soil was seen therefore as an opportunity to provide further evidence for its development in acid environments outside of those regions where is has been previously reported, and to demonstrate that the study soil has developed in situ from the weathering of the underlying sandstone. Accordingly, 3 modes of formation of smectite in the saprolite will be presented here. The first mode of formation involves direct inheritance from the underlying sandstone, whereby primary minerals and volcanic glass fragments have been pseudomorphed (i.e. replaced) by smectite during diagenesis and weathering of the sandstone. The second mode of formation involves precipitation from solution of the products of hydrolysis; thus, the smectite formed in this manner constitutes neogenetic (cf. Borchardt 1989), orthochemical void-filling cement. The third mode also occurs as a void-filling coating and involves deposition of smectite from suspension. The clay coatings formed in this manner are similar to the mechanically infiltrated (MI) clay coatings described in soils (e.g. Brewer 1964; Sullivan 1994) and rock (e.g. Wilson and Pittman 1977; Matlack et al. 1989; Moraes and De Ros 1990, 1992).
Study area and geology
The location of the study area, its geomorphic setting, and a full description of the soil is provided elsewhere (Walsh and Humphreys 2008). The soil is acid throughout (pH 4.8-5.8) but has a sodic subsoil (ESP increases to 18-23% with depth) with magnesium the dominant exchangeable cation (Walsh and Humphreys 2008). The soil is classified as a 'grey, magnesicnatric, bleached Kurosol' (Isbell 1996), and a 'fine-loamy over clayey, mixed semi-active, acid, thermic, torretric Natrustalf' (Soil Survey Staff 1998).
The main surface geology of the Pilliga Forests comprises the Upper Jurassic Pilliga Sandstone. Arditto (1982) described this unit in outcrop east of the study area. He found that lenticular, cross-stratified beds of coarse sandstone dominated the sequence, with siltstone and shale either very subordinate or absent from most sections. Where present, they occurred both as individual layers and finely interbedded with sandstone. Throughout the sequence, lenticular conglomerate horizons occurred, ranging in thickness from 5 to 60cm with pebble sizes ranging from 1 to 12 cm. The rock fragments comprising this unit included chert, jasper, quartzite, phyllite, vein quartz, and acid volcanics. In a study of the clay mineralogy of the Pilliga Sandstone, Arditto (1983) demonstrated an abundance of low birefringent clay material filling detrital grain fractures and pore spaces. X-ray diffraction (XRD) showed this material to be predominantly well-ordered kaolinite. He attributed the origin of the kaolinite matrix to the in situ alteration of detrital potassium feldspar and mica fragments in contact with percolating groundwater, i.e. the kaolinite was authigenic. Using mineral phase equilibria diagrams, he showed that the majority of bore water analyses from known Pilliga Sandstone aquifers fell within the kaolinite field, and concluded that minerals such as microcline or biotite mica would be chemically unstable when in contact with such water and tend to alter to kaolinite. However, his diagrams also showed that some data points fell within the montmorillonite field, a mineral not detected in the matrix of his samples.
Subsequently, Slansky (1984) described the clay mineralogy of the Jurassic strata in the Surat Basin using samples collected from diamond drill holes and outcrop to the north of our study area. He noted that in the eastern portion of the unit, there was considerable lateral variation in the clay minerals associated with kaolinite in the Pilliga Sandstone. From south to north he observed the following clay mineral assemblages: kaolinite-illite, kaolinite-illite-mixed layer mica-smectite, kaolinite mixed layer mica-smectite, smectite-kaolinite-illite, and kaolinite-smectite-illite. Bourke (1974), as reported by Slansky (1984), divided a 100-m section of Pilliga Sandstone into 4 subunits from a core also collected to the north of our study site. His first subunit was predominantly kaolinitic, which averaged 85% of the clay fraction. The second subunit contained lesser kaolinite and more smectite with a kaolinite/smectite ratio around 1.7. In the third subunit he distinguished between 2 lithologies, a coarse-grained member and a fine-grained member in which kaolinite was dominant in the former and smectite in the latter. The fourth subunit at the base of the formation was predominantly kaolinitic. Slansky (1984) noted that in the Jurassic sequence, lithology exerted a significant control on the relative occurrence of kaolinite and smectite, with a tendency for kaolinite to predominate in coarser sediments and smectite in finer sediments. In the Pilliga Sandstone, 2 smectite peaks in an XRD analysis corresponded to the occurrence of lithic sandstone and mudstone, respectively. He also found that the kaolinite/smectite ratio increased when the lithic component of the rock decreased.
Material and methods
The mineralogy of the saprolite and underlying sandstone were studied by integration of petrographic and XRD data. Two sandstone samples were taken at a depth of 1.3m in the trench excavation described in Walsh and Humphreys (2008), while 2 saprolite samples (dome and basal) were taken at depths of 0.25 and 0.40 m, respectively, from a pit 5 m from the trench excavation.
The XRD analysis was performed by the CSIRO Land and Water Laboratory in Adelaide as follows. A small subsample of each of the samples was ground in an agate mortar and pestle before being lightly back-pressed to reduce orientation effects into steel samples holders for XRD analysis. A 10-g subsample of each soil was shaken with 10 mL of a solution of sodium hexametaphosphate and sodium carbonate for 10 min, repeatedly dispersed, and centrifuged to recover all of the <2-[micro]m fraction. The <2-[micro]m suspensions were treated with acetic acid to remove any carbonate minerals, calcium-saturated using 1 M Ca[Cl.sub.2], washed with alcohol, and dried at 105[degrees]C. The <2-[micro]m powders were then lightly pressed into aluminium sample holders to achieve random orientation of the mineral particles for XRD analysis. A 120-mg subsample of the <2-[micro]m powders was re-dispersed in deionised water and the suspensions were sucked under vacuum onto each of 2 cellulose nitrate filter discs to produce maximum orientation of the platy clay mineral particles. One of these orientated samples was saturated with magnesium using 1 M Mg[Cl.sub.2] and glycerol to aid identification of the clay mineral components using XRD. XRD patterns were recorded with a Philips PW1800 microprocessor controlled diffractometer using Co K[alpha] radiation, variable divergence slit, and graphite monochromator. The diffraction patterns were recorded in steps of 0.05[degrees] 2[theta] with a counting time of 1.0s per step, and logged to data files on an IBM-compatible PC for analysis. Quantitative XRD analysis was performed on the samples using the commercial package SIROQUANT from Sietronics Pty Ltd. The data were first background-subtracted and calibrated for the use of the variable divergence slit.
Petrographic thin-sections were prepared from a block of sandstone and a portion of a columnar ped was extracted from the pit where the saprolite samples described above were taken. A BX50 Olympus petrological microscope and camera system was used to describe and identify minerals and photograph the thin-sections.
Microscopic attributes of clay minerals
The recognition of clay minerals by optical techniques is difficult due to their fine-grained nature. Because they are all optically negative, and the slow vibration direction invariably lies in the plane of cleavage, many of the microscopic properties of the various clay minerals are similar (Williams et al. 1953). The diagnostic optical properties that can be used to identify the different clay minerals are refractive indices and birefringence. Kaolinite has low birefringence and commonly shows only first-order grey colours. In thin-section, kaolinite commonly shows a characteristic high relief and small black and white crystallites. In contrast, smectite has a relatively high birefringence, and tends to show at least second-order colours. Smectite is the only common clay mineral with a refractive index below balsam, and because of this it appears spurious brown in thin-section (Folk 1974). In thin-sections of sandstones, Carrigy and Mellon (1964) found smectite to have a brownish cast in plane polarised light and first-order white to yellow interference colour under crossed nicols. Folk (1974) suggested that while clay minerals are difficult to identify with the petrological microscope, an educated guess is possible with careful work and the assistance of other techniques such as XRD. As the XRD analysis only identified kaolinite and smectite in the sandstone and saprolite samples, the optical properties that can be used to identify other common clay minerals are not presented here.
Distinguishing between mechanically infiltrated and authigenic clays
To assist in the interpretation of the clay mineral phases in the saprolite and sandstone at the study site, the following 5 textural criterion developed by Moraes and De Ros (1990) to distinguish detrital mechanically infiltrated (MI) clay (i.e. decanted from water) from authigenic neoformed clay (i.e. precipitated from solution) in a petrographic study of the reservoir quality of Jurassic sandstones in north-eastern Brazil will be used here:
(i) Ridges and bridges. These are aggregates of clay platelets oriented roughly normal to grain surfaces, and consist of small ridges projecting from the grain surface or elongated aggregates connecting 2 adjacent grains.
(ii) Geopetal fabrics. These consist of clay accumulations at the bottom of large pores and/or attached to the lower surface of grains.
(iii) Loose aggregates. These are chaotically flocculated aggregate in which the clay platelets compose an open flamework without any recognisable orientation.
(iv) Cutans. These are clay coatings along pore walls and grain surfaces. They have an anisopachous character (i.e. their thickness varies along the grain surface) and are oriented parallel to the grain surface. In contrast, neogenetic clays are usually oriented normal to grain surfaces and form isopachous coatings.
(v) Massive aggregates. These are aggregates that completely fill intergranular pores, with an internal structure that is characterised by thick coatings near the grain surface, and a dense and chaotic mass of clay platelets in the centre of pores. As with the loose aggregates, they lack internal organisation.
They also describe the following 3 characteristics of MI clays that were unrelated to the clay arrangement itself but could also be used to distinguish them from neogenetic clays:
(a) Impurities. As MI clays are detrital in origin, they commonly contain impurities such as oxides and fine organic debris mixed with the clay platelets.
(b) Dehydration and shrinkage detachment of the cutans from the grain surfaces and fragmentation of the massive aggregates. Both of these features appear as contracted replicas of the original pore fill or as curled sheets.
(c) Appearance under scanning electron microscopy (SEM). MI clay cutans typically show a smooth appearance in SEM due to the small crystal size of the dental smectites and their tangential organisation. They also have a pasty appearance in contrast to the well-defined crystalline aggregates of plates of neoformed clays.
The first of these 3 characteristics of MI clays will be also adopted here. The dehydration and shrinkage criterion will not be adopted, as shrinkage features can be produced artificially near the surface by alternate wetting and drying, conditions which are met at the study site. Appearance under SEM was not used.
Chemical analyses of the dome and basal saprolite samples were performed by Hill laboratories in New Zealand and the methods used for the various analyses are provided in Table 1.
Results and discussion
Table 1 shows the results of the chemical analysis for the 2 saprolite samples. Both of the samples are strongly acid (terminology of Isbell 1996, where strongly acid is defined as pH <5.5) and magnesium is clearly the dominant exchangeable cation.
Table 2 shows the results of quantitative XRD analyses of the saprolite and sandstone samples. Both materials show appreciable quantities of smectite in both the fine earth fraction (<2 mm) and the <2 [micro]m fraction (Table 2). When the standard errors are taken into account, there is little or no difference in the smectite content between the saprolite and sandstone samples for the fine earth fraction, which implies that the smectite in the saprolite is inherited from the sandstone. For the <2 [micro]m fraction, there is an increase of 16-18% in the amount of smectite from the saprolite into the sandstone (Table 2). A reverse trend is apparent in the kaolinite. In both fractions there is a 15% decrease with depth in the amount of kaolinite from the saprolite into the sandstone. The kaolinite/smectite ratio for the respective fractions shows that there is a strong increase in smectite content from the saprolite into the sandstone, particularly for the <2-[micro]m fraction (Table 3). Rice et al. (1985), in a study of soil saprolite profiles derived from mafic rocks, found that smectite decreased upward in the profile, which they attributed to destruction of the smectite or its alteration to hydroxl-interlayered vermiculite. Smectite is commonly altered to kaolinite where weathering, particularly leaching, is intense due to the removal of bases and/or silica which is usually accompanied by a drop in pH (Karathanasis and Hajek 1983). While the upward decrease in smecite content observed in the study soil suggests that kaolinite may be forming at the expense of smectite or, alternatively, that smectite is altering to kaolinite towards the surface of the profile, there was no evidence for this in the petrological interpretation of thin-sections from the saprolite. Further work incorporating SEM would assist in resolving this issue.
The presence and amount of smectite in the sandstone concurs with the observations of Slansky (1984) on the clay mineralogy of the Pilliga Sandstone, from which he reported a value of 50% as being typical of the amount of smectite present in the clay fraction of the finer grained units (Table 2). That Arditto (1983) did not detect smectite in the Pilliga Sandstone was evidently due to the sampling strategy; the study was confined to outcrop along the north-eastern edge of the Coonamble Lobe of the Great Australian Basin, whereas Slansky (1984) analysed subsurface samples from boreholes away from the margins of the basin. The finer grained, smectite-rich units of the Pilliga Sandstone are recessive, and therefore are not exposed as outcrop in the region.
The amount of clay in the sandstone reached a maximum of 36% (Table 3), an amount similar to that found in the saprolite at a depth of 25 cm. From a grain-size analysis of the Pilliga Sandstone, Arditto (1982) reported the average sediment to be a muddy sand (after Folk 1974). However, the majority of those samples contained >80% sand (>63 [micro]m), an amount far greater than the sandstone unit at our study site, which, based on the analysis of the above, gives an average clay content of 32% in the 2 sandstone samples (Table 3). This suggests that the sandstone unit at our study site is much finer than the outcrop sampled by Arditto (1982).
Organisation of the petrological interpretation
The petrological analysis of the saprolite and sandstone is organised in the following way. The first description and interpretation is focused on the nature of the relatively undisturbed/intact sandstone. The second and third are focused on the changes in the domains of a subhorizontal and vertical crack system in the sandstone, respectively, while the fourth focuses on the nature of a clay-lense in the sandstone. The fifth and sixth descriptions and interpretations are focused on 2 areas of the saprolite with distinct differences in their texture and mineralogy. The descriptions of the thin-sections follows the classification systems of Brewer (1964) and Bullock et al. (1985). The following abbreviations have been used in the text: PPL, plane polarised light; XPL, crossed polarised light; PXPL, partially crossed polarised light; MI, mechanically infiltrated. All of the photomicrographs of the thin-sections presented here are aligned so that stratigraphic up is towards the top of the page.
Microscopic investigation of the sandstone in the upper left-side comer of the thin-section illustrated in Fig. 1 shows that the coarse fraction comprises mainly subangular to subrounded grains of quartz and feldspar with subordinate volcanic-rock fragments (Fig. la). The quartz and feldspar grains display a similarity in grain-size and in their degree of roundness. Arditto (1982) reported a similar grain-size relationship between quartz and feldspar in the Pilliga Sandstone, and suggested that this grain-size similarity was indicative of the feldspar not having been transported great distances from the source area. In XPL, some quartz grains display undulose and composite extinction patterns (Fig. lb), which is indicative of strain at one or more stages in their history. From an analysis of the distribution of the types of detrital quartz grains in the Pilliga Sandstone, Arditto (1980) concluded that most of the quartz grains were derived from granitic rocks, and that quartz grains sourced from low- to medium-grade regional metamorphic rocks are of secondary importance.
[FIGURE 1 OMITTED]
As can be seen in thin-section (Fig. 1a), grains of feldspar, quartz, and rock-fragments are sporadically in contact, but are generally separated by clayey intergranular bridges and pseudomorphed grains. The related distribution pattern of the coarse and fine constituents tends to be chitonic (i.e. coarser units are surrounded by a cover of smaller units) with a local tendency towards gefuric (coarser grains are linked by braces of finer material; cf. Bullock et al. 1985). The degree of sorting is generally moderate to poor. Bright yellow, highly birefringent smectite forms argillan rims around the mineralogically stable epiclastic grains (mainly quartz) and pseudomorphs them. Kaolinite is also present and is recognisable by its characteristic high-relief and small black and white crystallites (Fig. lc).
Under higher magnification, some quartz grains show embayed margins occupied by kaolinite (Fig. 1c). The boundaries of some quartz grains also appear to display reaction with the clay matrix. Arditto (1983) observed embayed quartz grains being invaded by kaolinite in thin-sections of the Pilliga Sandstone. He suggested that the embayed grains-segments corresponded to former potassium-feldspar subgrains, the original polymineralogical grain having constituted a granite fragment. The initial breakdown of a feldspar grain (probably microcline) along the cleavage planes and partial replacement by smectite along its bottom edge is evident under XPL towards the top-right of the field-of-view in Fig. 1b. Mica (possibly biotite) and volcanic glass fragments are also present in the intact sandstone, but to a much lesser degree than quartz and feldspar grains, presumably because of their greater susceptibility to chemical alteration.
Evidence for the complete and partial pseudomorphing of the more labile grains (mica and volcanic glass fragments) by smectite can be seen in the mid-left and mid-upper fight regions of Fig. la. Under XPL these replaced grains impart a bright yellow colour (Fig. 1b). Under PPL, the shattering of a mica grain and partial replacement by smectite along its edges is evident slightly to the left-of-centre of the field-of-view in Fig. la. This shattering of the mica fragment is more clearly shown under XPL (Fig. 1b). Another feature of the thin-section depicted in Fig. lb is the numerous isopachous smectite rims growing in the fabric discontinuity between the grains. These isopachous smectite rims are associated with smectite-occluded microfractures. Away from these smectite-occluded microfractures, the grain coating argillan phase is absent (Fig. lb). The isopachous nature of the smectite that is growing in the fabric discontinuity between grains and along walls in the sandstone shown in Fig. lb suggests that is has precipitated from solution.
Sub-horizontal and vertical crack system in the sandstone
A photomicrograph of a sub-horizontal crack system towards the top of the thin-section in the sandstone (Fig. 2) shows smectite growing along the subhorizontal and horizontal crack network (Fig. 2a). The top-right of the slide is predominantly unreplaced sandstone, whereas the bottom-left of the slide is mainly solid smectite. Under PPL, the outlines of former grains that have been pseudomorphed by smectite are evident in the centre of the field-of-view in Fig. 2a. Under XPL, layering of the smectite along the dominant subhorizontal crack traversing the thin-section is evident (Fig. 2b). The layering of smectite along the network of first-order cracks and its penetration into the sandstone along second-order cracks that mainly follow grain boundaries to form a channel argillan (cf. Brewer 1964) is further illustrated at higher magnification under PPL and XPL in Fig. 3a and b.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
The smectite adjacent the wall rock on either side of the subvertical fracture shown in Fig. 3a and b is optically coherent, whereas the smectite towards the centre of the crack is incoherent and its murky appearance suggests some degree of contamination. These contaminants may interfere with the ability of the system to lay down the more coherent material typical of the smectite along the walls of the crack. Along the centre of the crack in Fig. 3b are several individual pellets of clay which have probably been produced through shrinkage of the less coherent smectite in the crack, for which further evidence can be seen in the centre of the crack in Fig. 3a. These different textures of the smectite may be associated with crack widening. During the initial stages of the crack development the fabric is so tight that MI clays cannot be deposited, and the pore water precipitates smectite. This is supported by the isopachous nature of the smectite growing along the second-order cracks and grain boundaries in the rock adjacent to the crack wall. As the crack widens through expansion and contraction of the rock, the incoherent material is able to penetrate the crack and is deposited from suspension. The layering of the smectite, particularly along the fight-hand margin of the crack, suggests that the system is mechanically dynamic and is leading to textural changes in the smectite lining the walls of the crack. This layering is also evidence of episodic accretion of smectite along the crack walls.
A large area of kaolinite is also present adjacent to the smectite-lined crack in Fig. 3a. Along the edge of the kaolinite adjacent to a quartz grain towards the lower left-side of the field-of-view, smectite is pseudomorphing a fragment of volcanic glass. Thus, in the space of a few microns, smectite is forming by transformation, and quite possibly neogenesis and mechanical infiltration, alongside apparently stable kaolinite. This implies that the kaolinite predates the smectite.
Clay lense in the sandstone
A photomicrograph of a clay lense in the sandstone under XPL shows that these lenses are predominantly smectitic, with kaolinite occurring as remnant grains within the matrix (Fig. 4). The large grain of kaolinite towards the fight-of-centre of the field-of-view is enclosed by optically coherent smectite, which is precipitating as argillans in cavities around the edges of grains as well as through the replacement of primary minerals and/or labile rock fragments. While it has been demonstrated petrographically that transformation of some of the primary labile minerals and rock-fragments to smectite is contemporary (i.e. is occurring through weathering of the sandstone under near surface conditions), it is possible, nevertheless, that the bulk of the smectite in the clay matrix of the sandstone was formed during diagenesis. Slansky (1984) noted that on account of their shallow burial, the diagenetic changes in the sediments in the NSW portion of the Great Australian Basin was minimal. According to Folk (1974), these conditions are conducive to the precipitation from solution of kaolinite, chlorite, smectite, or large flaky sericite crystals as authigenic pore-fillings in sandstones.
Arditto (1983) suggested that much of the kaolinite matrix in the Pilliga Sandstone resulted from the post-depositional alteration of unstable detrital minerals by reaction with mobile groundwater. While he did not detect smectite in his sandstone samples, the chemical composition of the groundwaters sourced from Andrews (1975) did not preclude its formation as a stable matrix-forming mineral, a circumstance that is supported by the petrographic evidence presented herein. In fact Slansky (1984), in a study of the clay mineralogy of the Surat Basin, suggested that the relative stabilities of kaolinite and smectite in rivers and sea waters were not unambiguous. His isoplethic sections showed that while smectite has a slightly higher stability in sea water, in fiver water there is an equal chance for both minerals to be stable with respect to each other. Thus, the odds of one mineral forming over the other in the different geological units in the basin is not determined by pH, but by the chemical composition of the individual river waters.
[FIGURE 4 OMITTED]
Microscopic investigation of the saprolite shows that both quartz and smectite were abundant, with the latter occurring as both argillans around grains and as matrix-like material between the grains (Fig. 5). Kaolinite is also present predominantly as replacement grains and appears to be subordinate to the smectite. The saprolite is poorly sorted with abundant silt-sized quartz grains interposed between the larger quartz grains. The shape of the quartz grains is predominantly semi-equant, and their degree of roundness is variable from angular/subangular to subrounded. The presence of silt-sized grains in the interstitial spaces is typical of an enaulic-related distribution (cf. Bullock et al. 1985); however, as with the sandstone, there was a tendency towards a chitonic-related distribution where the coarser fragments are coated by clay argillans or cutans. At higher magnification, considerable interparticle porosity is evident between the smectite-coated quartz grains (Fig. 5b). The smectite argillans coating the grains are anisopachous and have strong continuous parallel orientation, suggesting that they are MI clay platelets that have progressively accreted the epiclast-grain surfaces in a geometrically ordered arrangement as described by numerous authors; for example, see Matlack et al. 1989 and Moraes and De Ros 1990.
[FIGURE 5 OMITTED]
The absence of a visible silt-sized quartz component in the sandstone suggests that this fraction in the saprolite is allochthonous with respect to the study site. A photomicrograph of the saprolite 8 mm below the photomicrographs in Fig. 5 shows abundant silt-sized grains occurring along crack margins and in the smectite matrix (Fig. 6). Hesse et al. (1998) reported a significant aeolian input to some of the Pilliga soils. They identified a silt-sized quartz fraction with a distinct mode of 30-40 [micro]m in a soil formed in a sand dune near (<3 km) the study site. This fraction formed ~20% by weight of the soil imparting to it an earthy, coherent fabric. For the less permeable soils formed on hillslopes (the study soil included), Hesse et al. (1998) suggested that the majority of the dust had been washed into drainage lines and creeks and is stored in the colluvial and alluvial deposits.
There are 2 possible mechanisms by which the allochthonous quartz has been incorporated into the clay matrix of the saprolite shown in Figs 5 and 6. The first mechanism involves mixing or ingestion of the soil by ants, termites, or earthworms. Hart (1995) reported significant ant and termite activity in the study soil, with evidence of tunnelling extending from the topsoil into the subsoil (saprolite), particularly in the domed tops of the columnar peds. Evidence for extensive bioturbation within a columnar ped was reported by Hart (1995) and Walsh and Humphreys (2008). This subsoil mixing by soil fauna can also explain the presence of smectite in the topsoil, whereby aggregates of clay are transported to the surface and incorporated into the topsoil.
The second mechanism involves the mechanical infiltration of the silt-sized quartz from the surface into the saprolite along microscopic cracks (Fig. 6). During differential wetting and drying of the saprolite this material in the cracks could be incorporated into the clay matrix through heaving and contraction of the soil in a manner similar to the mechanism described in Vertisols by Nettleton et al. (1981). They attributed the presence of silt-sized aggregates of organic material in the plasma of upper horizons in several Vertisols to mixing processes (pedoturbation) caused by changes in the moisture content of the soil. The presence of a silt-sized quartz component in the saprolite of the study soil suggests that at least some of this wind-blown material is retained by the hillslope soils in the Pilliga Forests.
In contrast to the saprolite shown in Fig. 5a and b, the saprolite shown in Fig. 6 is dominated by replacement smectite forming a matrix within which are embedded grains of quartz and feldspar. As with the sandstone, the smectite lining the crack walls and coating grain boundaries is isopachous, suggesting that it is neogenetic. Several grains that have been pseudomorphed by smectite to form the clay matrix are indicated by the white arrows in Fig. 6. There was some evidence of alternating zones of clay-rich and clay-poor seams in the saprolite similar to that found in the sandstone, suggesting that along with mineralogy, the fabric of the sandstone is also inherited by the saprolite. Stace et al. (1968) reported a similar fabric in a soloth formed in Quaternary calcareous sands, and suggested that it was evidence for the sedimentary origin of the parent material. Further petrological work over a larger area of a ped is, however, required to substantiate this suggestion of a fabric inheritance by the saprolite.
The petrological analysis of the sandstone and overlying saprolite at the study site demonstrates that smectite is forming in both of these materials through replacement of labile primary minerals and rock-fragments, and through precipitation from solution and/or mechanical infiltration of clay platelets in suspension. In the case of the former process, the evidence for partial pseudomorphing of grains in both the sandstone and saprolite suggests that this process is contemporary, and is occurring through weathering of the sandstone under nearsurface conditions. However, it is not possible to determine the proportion of completely pseudomorphed grains that are due to near-surface weathering or diagenesis. Irrespective of this, the petrological evidence suggests that the smectite that is formed through replacement of labile primary minerals and rock fragments in the sandstone is inherited by the saprolite.
[FIGURE 6 OMITTED]
The presence of neogenetic and mechanically infiltrated smectite in both the sandstone and saprolite requires some discussion. Millot (1970) suggested that smectite neogenesis can only occur in soils where drainage is moderate enough to maintain an alkaline environment. However, this study has demonstrated that apparent neogenetic and mechanically infiltrated smectite is forming and being maintained in an acid environment. Borchardt (1989) suggested that, under restricted drainage, smectite may precipitate from soil solution in almost any parent material capable of supplying its elemental constituents regardless of pH, providing there is an adequate supply of Mg. He argued that as the amount and rate of water percolation increases, only the micro-environments surrounding individual grains of Mg-bearing minerals maintain high enough solution concentrations to form smectite. He found that well-drained soils formed on granitic bedrock seldom contained neogenetic smectite, but the joints and fractures beneath them do, and that smectite neogenesis is common within fault fissures and along geologic contacts beneath otherwise well-drained-soils.
Walsh (2003) showed that the matrix of a columnar ped in the study soil is dense, with few planar voids penetrating into the centre of the ped. These characteristics are conducive to poor internal drainage, which favours the retention in solution of the products of hydrolysis. When evaporated the solution precipitates smectite. The chemical data presented in Table 1 show that Mg is the dominant cation in the saprolite of the study soil, which further supports this model of smectite genesis. Thus, it would appear that the critical factors for the formation of the neogenetic smectite in the study soil are restricted drainage and a parent rock capable of supplying sufficient amounts of Mg.
The absence of MI clays in the thin-sections of Arditto's (1980) Pilliga sandstone samples suggests that their presence in the sandstone at the study site may be due to near-surface weathering, where material within the clay lenses and cracks is being re-worked and deposited against apparent neogenetic smectite. This process may also be occurring in the saprolite. Fitzpatrick (1980) suggested that local reorganisation of clay material is a very active mechanism in soils, and described a complete cycle whereby a given clay particle forming part of the matrix can be translocated to form part of a coating which later disintegrates to again form part of the matrix.
This study has shown that the bulk of the smectite in the saprolite of the study soil is inherited from the underlying sandstone through the transformation of labile primary minerals and rock fragments. The greater part of this transformation has probably occurred during diagenesis of the sandstone; however, there is ample evidence that primary minerals are still being altered to smectite today under near-surface weathering conditions. These near-surface weathering conditions are also facilitating the apparent precipitation of smectite and/or its mechanical infiltration in suspension along grain boundaries and cracks in both the sandstone and saprolite.
Evaporation of water saturated with the products of hydrolysis is producing neogenetic smectite in an acid environment. Restricted drainage coupled with a parent material capable of supplying the elemental constituents of smectite are the critical factors in this process. Further, more detailed work using SEM is required to establish beyond doubt whether the smectite-forming argillan along crack walls and grain boundaries in the sandstone and saprolite is precipitating from solution, or being mechanically infiltrated in suspension, or both.
While there is a small allochthonous component in the saprolite in the form of silt-sized quartz grains, it has been demonstrated qualitatively that the bulk of the saprolite of the study soil is derived from the weathering of the underlying sandstone. This qualitative assessment needs to be substantiated by a quantitative measure such as point counting to establish the proportions of the various phases.
The study was partially funded by a postgraduate research grant from Macquarie University, NSW. Thanks to Dr Pat Conaghan for his invaluable assistance with the petrological interpretation and for taking photomicrographs, Forests NSW for permission to work in the Pilliga East State Forest, Mr Tom Bradley for the manufacture of the thin-sections, and the 3 anonymous reviewers for their helpful comments on an earlier version of the manuscript.
Manuscript received 7 April 2009, accepted 21 September 2009
Andrews HE (1975) Artesian water quality in the New South Wales portion of the Great Australian Basin. Report Geological Survey NSW South Wales, GS 1975/327.
Arditto PA (1980) Stratigraphy and sedimentation of the Pilliga Sandstone in the southeastern portion of the Great Australian Basin, New South Wales. MSc Thesis, University New South Wales, Australia.
Arditto PA (1982) Deposition and diagenesis of the Jurassic Pilliga Sandstone in the southeastern Surat Basin, New South Wales. Journal of the Geological Society of Australia 29, 191-203.
Arditto PA (1983) Mineral-groundwater interactions and the formation of authigenic kaolinite within the southeastern intake beds of the Great Australian Basin, New South Wales, Australia. Sedimentary Geology 35, 249-261. doi:10.1016/0037-0738(83)90061-1
Borchardt G (1989) Smectites. In 'Minerals in soil environments'. 2nd edn (Eds JB Dixon, SB Weeds) pp. 675-718. (Soil Science Society of America: Madison, WI)
Bourke DJ (1974) Completion report of stratigraphic bore holes DM Rawdon Nos. 1 and 1A, Surat Basin. New South Wales Geological Survey-Report GS 1974/008.
Brewer R (1964) 'Fabric and mineral analysis of soils.' (John Wiley and Sons Inc.: Sydney)
Bullock P, Fedoroff N, Jongerius A, Stoops G, Tursina T, Babel G (1985) 'Handbook for soil thin-section description.' (International Society of Soil Science/Waine Research Publications: Wolverhampton, UK)
Carrigy MA, Mellon GB (1964) Authigenic clay mineral cements in Cretaceous and Tertiary sandstones of Alberta. Journal of Sedimentary Petrology 34(3), 461-472.
Douglas LA (1981) Smectites in acidic soils. In 'Developments in Sedimentology 35: International Clay Conference 1981. Proceedings of the International Clay Conference'. Bologna and Pavia, Italy, 6-12 September. (Eds H Van Olphen, F Veniale) pp. 635-640. (Elsevier: Amsterdam)
Egli M, Mirabella A, Mancabelli A, Sartori A (2004) Weathering of soils in alpine areas as influenced by climate and parent material. Clays and Clay Minerals 52(3), 287-303.
Egli M, Mirabella A, Sartori A, Zanelli R, Bischof S (2006) Effect of north and south exposure on weathering rates and clay mineral formation in Alpine soils. Catena 35(3), 155-174.
Fitzpatrick EA (1980) 'Soils: their formation, classification and distribution.' (Longman: New York)
Folk RL (1974) 'Petrology of sedimentary rocks.' (Hemphill Publishing Company: Austin, TX)
Grim RE (1968) 'Clay mineralogy.' 2nd edn (McGraw-Hill Book Company: New York)
Hart DM (1995) Litterfall and decomposition in the Pilliga State Forests, New South Wales, Australia. Australian Journal of Ecology 20, 266-272. doi:10.1111/j.1442-9993.1995.tb00538.x
Hazelton P, Murphy B (1992) 'Interpreting soil test results What do all the number mean?' (CSIRO Publishing: Collingwood, Vic.)
Hesse PP, Townsend FN, Humphreys GS, Kamper S (1998) Dust accession rates over the Late Pleistocene in northern NSW; the potential for the formation of aeolian mantles and reworking within the landscape. In 'Symposium on Aeolian Dust: Implications for Australian Mineral Exploration and Environmental Management'. (CRC LEME: Canberra)
Isbell RF (1996) 'The Australian Soil Classification.' (CSIRO Publishing: Collingwood, Vic.)
Karathauasis AD, Hajek BF (1983) Transformation of smectite to kaolinite in naturally acid soil systems: structural and thermodynamic considerations. Soil Science Society of America Journal 47, 158-163.
Karathanasis AD, Hajek BF (1984) Evaluation of aluminium-smectite stability equilibria in naturally acid soils. Soil Science Society of America Journal 48, 413-417.
Matlack KS, Houseknecht DW, Applin KR (1989) Emplacement of clay into sand by infiltration. Journal of Sedimentary Petrology 59, 77-87.
Millot G (1970) 'Geology of clays.' (Springer-Verlag: Berlin)
Moraes AS, De Ros LF (1990) Infiltrated clays in fluvial Jurassic sandstones of Reconcavo Basin, Northeastern Brazil. Journal of Sedimentary Petrology 60, 809-819.
Moraes AS, De Ros LF (1992) Depositional, infiltrated and authigenic clays in fluvial sandstones of the Jurassic formation, Reconcavo Basin, Northeastern Brazil. In 'Origin, diagenesis and petrophysics of clay minerals in sandstones'. (Eds DW Houseknecht, ED Pittman) SEPM Special Publication No. 47. pp. 197-208. (Society for Sedimentary Geology: Tulsa, OK)
Nettleton WD, Peterson FF, Borst G (1981) Micromorphological evidence of turbation in vertisols and soils in vertic subgroups. In 'Soil micromorphology. Vol. 2: Soil genesis'. (Eds P Bullock, CP Murphy) pp. 441-458. (AB Academic Publishers: Berkhamsted, UK)
Rice TJ, Buol SW, Weed SB (1985) Soil-saprolite profiles derived from mafic rocks in the North Carolina Peidmont: I. Chemical, morphological and mineralogical characteristics and transformation. Soil Science Society of America Journal 49, 171-178.
Righi D, Huber K, Keller C (1999) Clay formation and podzol development from postglacial moraines in Switzerland. Clay Minerals 34(2), 319-332.
Slansky E (1984) Clay mineralogy. In 'Contributions to the Geology of the Great Australian Basin in New South Wales: Geological Survey of N.S.W, Bulletin 31'. (Eds JM Hawke, JN Cramsie) pp. 179-203. (Department of Mineral Resources: Sydney)
Soil Survey Staff (1998) 'Keys to Soil Taxonomy.' (United States Department of Agriculture Natural Resources Conservation Service: Washington, DC)
Stace HCT, Hubble GD, Brewer R, Northcote KH, Sleeman JR, Mulcahy M J, Hallsworth EG (1968) 'A handbook of Australian soils.' (Rellim Technical Publications: Glenside, S. Aust.)
Sullivan LA (1994) Clay coating formation on impermeable materials: deposition by suspension retention. In 'Soil micromorphology: Studies in management and genesis. Proceedings IX International Working Meeting on Soil Micromorphology'. Townsville, Qld, July 1992; Developments in Soil Science 22. (Eds AJ Ringrose-Voase, GS Humphreys) pp. 373-380. Elsevier: Amsterdam)
Walsh P (2003) The development of columnar peds in a texture contrast soil in the Pilliga State Forests northwestern NSW. MSc Thesis, Department of Physical Geography, Macquarie University, Sydney, NSW, Australia.
Walsh P, Humphreys G (2008) Orientation and spacing of columnar peds in a sodic, texture contrast soil in Australia. In 'New trends in soil micromorphology'. (Eds S Kapur, A Mermut, G Stoops) pp. 163-198. (Springer-Verlag: Berlin, Heidelberg)
Williams H, Turner FJ, Gilbert CM (1953) 'Petrography: an introduction to the study of rocks in thin-sections.' (W. H. Freeman and Company: San Francisco, CA)
Wilson MD, Pittman ED (1977) Authigenic clays in sandstones: recognition and influence on reservoir properties and paleoenvironmental analysis. Journal of Sedimentary Petrology 47, 3-31.
Peter G. Walsh (A,B,E) and Geoff S. Humphreys (C,D)
(A) Forests NSW, PO Box 273, Eden, NSW 2551, Australia.
(B) School of Resources, Environment and Society, Australian National University, Canberra, ACT 0200, Australia.
(C) Department of Physical Geography, School of Environmental and Life Sciences, Macquarie University, NSW 2109, Australia.
(E) Corresponding author. Email: email@example.com
Table 1. Selected soil chemistry of the dome and basal saprolite samples Exch. cations EC Organic CEC (cmol/kg) (A) (dS/m) PH matter (cmol/ Na Mg Ca K (B) (C) (%) (D) kg) (E) Dome-saprolite at 0.25 m depth 1.39 8.41 <0.5 0.4 0.08 5.2 1.1 14.3 Basal-saprolite at 0.40m depth 4.17 14.08 <0.5 0.5 0.56 4.9 1.0 23.1 (A) [1.sub.M] Neutral ammonium acetate extraction. Atomic absorption (Mg, Ca) and atomic emission (K, Na). (B) Value obtained from Hazelton and Murphy (1992), where EC 1 :5 (dS/m)=TSS (g/100g)/0.32. (C) 1 :2 (v/v) soil water slurry. Potentiometrically using a pH electrode. (D) Walkley-Black oxidation, determined colourimetrically, converted from readily oxidisable C using a factor of 1.72. No correlation for incomplete oxidation (approx. 90%) has been applied. (E) SUM of extractable cations (K, Ca, Mg, Na) and the acidity determined from the change in the pH of the cation extraction solution. Table 2. Percent quantitative mineralogy of the saprolite and sandstone samples Shaded and unshaded values are results for the fine earth fraction (<2 mm) and the <2 gm sample respectively. Values in parentheses are the standard errors in the last significant figures derived from the Rietveld analysis Quartz Kaolinite Smectite Orthoclase Albite Anatase Dome saprolite at 0.25m 35 (1) 30 (2) 34 (2) <1 <1 <1 3 (1) 57 (2) 39 (2) -- -- 1 (1) Basal saprolite at 0.40m 39 (1) 25 (1) 33 (2) <1 <1 <1 4 (1) 49 (I) 47 (2) 1 (1) Sandstone 0.20m from clay root at 1.30 m 41 (2) 22 (2) 31 (2) 4 (l) 2 (1) <1 2 (1) 45 (3) 53 (3) -- -- l (1) Sandstone 0.30m from clay root at 1.30m 49 (2) 15 (2) 29 (2) 5 (1) 2 (l) <1 2 (1) 42 (3) 56 (3) -- -- l (1) Table 3. Percent clay content and the kaolinite/ smectite ratio in the fine earth fraction (<2 mm) and the <2 [micro] m fraction of the saprolite and sandstone samples Fine <2 %Clay earth [micro]m fraction fraction Dome saprolite at 37 0.88 1.46 0.25 m Basal saprolite at 40 0.76 1.04 0.40m Sandstone sample 1 36 0.71 0.84 at 1.30m Sandstone sample 2 28 0.52 0.75 at 1.30m
|Printer friendly Cite/link Email Feedback|
|Author:||Walsh, Peter G.; Humphreys, Geoff S.|
|Publication:||Australian Journal of Soil Research|
|Date:||Feb 1, 2010|
|Previous Article:||Surface charge characteristics and sorption properties of bauxite-processing residue sand.|
|Next Article:||Comment on: 'burning management and carbon sequestration of upland heather moorland in the UK'.|