weathering of granitic rocks

The weathering of granitic rocks in a hyper-arid and hypothermal environment: A case study from the Sør-Rondane Mountains, East Antarctica

Introduction

The weathering process in Antarctica is key to our understanding of the geomorphological processes that take place in hyper-arid and hypothermal environments on Earth. Although rock disintegration and soil formation are caused mainly by physical weathering processes, chemical weathering is of considerable importance to soil differentiation in Antarctica (Campbell and Claridge, 1987). Recently, having studied element composition changes in basaltic glass from Beacon Valley in the McMurdo Dry Valleys, Antarctica, Salvatore et al. (2013) suggested that the initial chemical weathering process is controlled by cation migration in response to the oxidizing environment. These studies have contributed significantly to our knowledge of weathering processes in Antarctica; however, they focused mainly on basaltic rocks in the McMurdo Dry Valleys. On the other hand, quartzofeldspathic rocks are exposed in many parts of Antarctica. As granitic rocks (a type of quartzofeldspathic rock) have a relatively homogeneous chemistry and isotropic structure, they are a suitable rock type with which to investigate the physical and/or chemical weathering processes of quartzofeldspathic rocks. Furthermore, quartzofeldspathic rocks are potentially common on parts of Mars (e.g., Wray et al., 2013); therefore, this knowledge can also be used as an analog that will help to improve our understanding of geomorphological processes and landscape evolution on Mars (e.g., Marchant and Head, 2007).

The degree of weathering of the rocks in Antarctica has been used to reconstruct the history of Antarctic ice retreat. For example, a significant decrease in the thickness of the East Antarctic Ice Sheet (EAIS) has been reconstructed using the Moriwaki index (e.g., Moriwaki et al., 1991, Moriwaki et al., 1994; White et al., 2009; Suganuma et al., 2014), which is based mainly on the color change of surface of rocks that is probably related to the degree of oxidation. However, the nature of the weathering processes that affect glacial till, including the physical and chemical processes that produce color changes, remain unclear. This uncertainty limits our ability to obtain quantitative data that can be used to identify the spatial and altitudinal distributions of the mode of weathering and thereby create geomorphological maps that can be used to reconstruct the glacial history of the area. Consequently, we must improve our understanding of the weathering process as well as the nature of weathering products.

In this study, we investigated the weathering processes associated with granitic rocks under the hyper-arid and hypothermal conditions that prevail in Dronning Maud Land, East Antarctica, using a number of approaches, including degree of weathering, color measurement, whole-rock chemistry, mineralogical experiments, rock hardness tests, and microscopic observations. We focused on the differences between the weathered and relatively fresh parts of the rock samples collected from the glacial till. Our findings will lead to a new quantitative method that can be used to evaluate the degree of weathering based on in situ and/or satellite measurements of exposed rocks in Antarctica.

Study area

The Sør-Rondane Mountains are exposed as glaciated landforms and cover an area of 2000 km2 within Dronning Maud Land, East Antarctica (Fig. 1). The elevation of the ice sheet surface at the northern side of the Sør-Rondane Mountains is about 1000 m, rising to 2500 m in the south (Fig. 2). The mountains are composed of low- to high-grade metamorphic rocks and various plutonic rocks that intruded the metamorphic bedrock (e.g., Shiraishi et al., 1997). The tectonothermal events that affected the mountains occurred at 550–1000 Ma, based on sensitive high-resolution ion microprobe U–Pb zircon ages (e.g., Shiraishi et al., 2008; Kamei et al., 2013; Owada et al., 2013).

Fig. 1. Location of the Sør-Rondane Mountains in Dronning Maud Land, East Antarctica. The box indicates the location shown in Fig. 2. Antarctic stations are shown by red circles.
Fig. 2. (a) Map of sample locations (red stars). Colored lines correspond to profiles of glacial elevation. (b) Glacial profiles of central and western parts of the Sør-Rondane Mountains. Stars indicate the latitude and elevation of sampling sites. Relative heights used in this study between each sampling site and the adjacent current ice sheet were calculated from these profiles. The satellite image used is ©JAXA.

The current mean annual air temperature at Asuka Station (970 m above sea level), located at the northern margin of the Sør-Rondane Mountains, is −18.4 °C, and in summer the air temperature rarely exceeds freezing point (Matsuoka et al., 2006). As the elevation of the ice sheet increases towards the south, most of the mountains are assumed to remain below freezing point permanently. Snow falls throughout the year, but the mountains are essentially snow-free in summer, probably because of sublimation. This view is supported by Takahashi et al. (1992), who reported a high sublimation rate (200–280 mm y−1) on a bare ice field near Asuka Station. These very low atmospheric temperatures and the limited availability of liquid water are similar to the McMurdo Dry Valleys, which have long been considered the most Mars-like environment on Earth (e.g., Anderson et al., 1972; Gibson et al., 1983; Mahaney et al., 2001; Wentworth et al., 2005; Marchant and Head, 2007; Levy et al., 2009; Levy, 2015).

This study targets the Austkampane, Lunckeryggen, Mefjell, Dufek, and Nils Larsen ranges in the central and western Sør-Rondane Mountains (Fig. 2). There is geomorphological evidence that a large part of these mountains was covered by an expanded ice sheet (e.g., Van Autenboer, 1964; Iwata, 1987; Hirakawa et al., 1988; Hayashi and Miura, 1989; Aniya, 1989; Hirakawa and Moriwaki, 1990; Moriwaki et al., 1991; Suganuma et al., 2014). Therefore, the Sør-Rondane Mountains provide a valuable opportunity to investigate elevation changes of the EAIS and the progress of physical and chemical weathering of rocks that were exposed after lowering of the ice sheet. Many parts of the mountain range are flat-topped ridges, with flanks that are covered mostly by glacial tills (Supplementary Fig. 1). There are also several moraine fields fringing the mountain blocks at almost the same level as the present ice surface (Supplementary Fig. 1). Lichen developments, possibly having influence for weathering of rocks, were not noticeable in the study area.

Methods

The weathering of glacial tills in the study area is generally controlled by elevation (Suganuma et al., 2014), and the glacial till deposits in the Sør-Rondane Mountains have been divided into four stages based on the degree of weathering (the Moriwaki index; Moriwaki et al., 1991, Moriwaki et al., 1994). The 10Be surface exposure ages from this area reveal that the degree of weathering generally increases with relative height to the present ice sheet as the period of exposure also increases (Suganuma et al., 2014). Thus, we selected nine sites from various elevations throughout the study area, and took 1–2 granitic rock samples (total 14) from each location as representative of the weathering at each site (Table 1, Table 2; Supplementary Fig.1; Fig. 3a, b). The height differences between these rock samples and the ice sheet nearby (relative height) were determined via handheld GPS and corrected using elevation data from Technical Report of the Geospatial Information Authority of Japan (GSI), C1-No.419.

We selected a medium-grained granite for this study (Fig. 3a–h). The rock types of the samples studied are listed in Table 2. The mineral assemblages of all samples were mainly quartz, plagioclase, alkali-feldspar, biotite, and opaque minerals. Several rock samples also contained muscovite and tourmaline. Under the microscope, the difference in the degree of alteration between the core and crust was indistinguishable in all samples, except for fracture development, oxidization features, and vein-like stains along grain boundaries and fractures (Fig. 3e–h). Chlorite, dusty feldspar, and fine sericite within the plagioclase were seen in both the core and crust of several samples. No secondary calcite was found in the samples (Fig. 3e–h). The material used for whole-rock chemical analysis and color measurements was removed from the crust and core of the halved rock samples (Fig. 3a, b). The crust and core were defined as the outer layer, which extended from the rock surface to a depth of 1 cm, and the relatively fresh material in the rock interior, respectively. These subsamples were manually crushed into pebble- to cobble-sized fragments. The fragments were washed in an ultrasonic bath with filtered hot water (70 °C) for 30 min to remove any dust and salt, and then with pure water for 20 min. They were then dried in an incubator at 105 °C for at least 15 h, and a smaller representative subsample of about 500 g was then ground into a powder, using an agate ball mill for 30 min, prior to conducting the following analyses.

Table 2. Weathering characteristics of glacial tills at each site. F and W at the end of each specimen number indicate core and crust of the rock sample, respectively.

Gr: granite, QSy: quartz syenite, Gd: granodiorite, Ad: adamerite, Qd: quartz diorite, Mz: monzonite, AFG: alkali feldspar granite, CIA: chemical index of alteration (Nesbitt and Young, 1982, Nesbitt and Young, 1984), DW: degree of weathering. A: rock freshness (0: fresh to 3: not fresh). B: staining (0: not stained to 3: strongly stained). C: presence of cavernous weathering and/or ventifacts. D: presence of crumbling features.
Fig. 3. Photographs of representative rock samples. (a) Photograph of a relatively fresh rock sample (in situ) without a fragile crust from Lunckeryggen (Lu–02a). (b) Photograph of a rock sample (in situ) with a stained and fragile crust from Dufek (Du–01b). (c) Halved section of sample (a). The 15 tick marks on the halved rock samples indicate measurement points for rock hardness tests. (d) Halved section of sample (b). (e) Photomicrograph of the crust of sample (a). (f) Photomicrograph of the crust of sample (b). (g) Photomicrograph of the core of sample (a). (h) Photomicrograph of the core of sample (b). The white scale bar in (a) and (b) is 20 cm long. Photomicrographs were taken under cross-polarized light. Qz: quartz, Pl: plagioclase, Afd: alkali feldspar, Bt: biotite, Ms.: muscovite. Thin dashed lines in (e) and (f) indicate the rock surface.

Classification of geomorphic surfaces (Moriwaki index)

The weathering characteristics of glacial deposits were classified in the field using the Moriwaki index (Moriwaki et al., 1991, Moriwaki et al., 1994), which assigns the degree of weathering (DW) on the basis of: (1) freshness and staining, (2) cavernous weathering, and (3) ventifact development of the largest 100 gravel clasts from a 10 × 10 m2 area of the surface of each site. The method distinguishes five degrees of weathering:

DW 0 = a fresh clast (i.e., showing an original rock color without any weathered features); DW 1 = a stained clast without cavernous weathering, ventifacts and crumbling; DW 2 = a stained, cavernously weathered and/or wind-faceted, not crumbled clast; DW 3 = a distinctly stained and somewhat crumbled clast; and DW 4 = a strongly stained and crumbled clast.

The DW classification of selected samples was re-checked in the laboratory based on an inter-comparison among the sites, and we changed the DW value of one sample (from Lunckeryggen, sample Me–01a) from 4 to 3, accordingly. The degree of weathering for the whole site, the weathering index (WI), was calculated by summing the number of individual gravel clasts in each DW category weighted by their respective DW values, as follows:

where N0, N1, N2, N3, and N4 denote the number of gravel crusts of DW 0–4, respectively.

The weathering characteristics of the glacial tills at each site are summarized in Table 2.

Color strength index

An objective estimation of rock staining is difficult when using only visual means; therefore, to estimate the staining of our samples the color of powdered specimens was measured five times for each specimen under laboratory conditions using a handheld color reader (Konica Minolta SPAD-503) that employs the tristimulus method. We used the average values here, which are displayed in the L⁎a⁎b⁎ color space defined by the International Commission on Illumination (CIE 1976). In this space, the L⁎, a⁎, and b⁎ color values indicate lightness, green (−) to red (+), and blue (−) to yellow (+), respectively. The L⁎, a⁎, and b⁎ values were calculated using the tristimulus values measured by the color reader. We defined the color strength index (CSI) to quantify the degree of staining of the powdered specimens. The CSI shows the chroma of the color and simply indicates the distance from the origin to the a⁎ and b⁎ value on the coordinate system, and was calculated as follows:

Use of the color reader and CSI evaluation allows an objective estimation of stain, whereas the observer’s subjective estimation of stain is required when using the Moriwaki index.

Loss on ignition

In this study, loss on ignition (LOI) measurements were used to estimate the differences in the degree of weathering between the crust and core of the samples. LOI is generally used as a proxy for the weathering intensity of rocks and is based on the change in mass of the sample before and after ignition in a furnace (e.g. Kamei et al., 2012; Bazilevskaya et al., 2015). To measure the LOI, we placed 2 g of each of the powdered specimens from the crust and core of halved rock samples into a furnace at 1000 °C for 2 h and recorded the change in mass. The chemical alteration of rock-forming minerals ultimately forms hydrous clay minerals. Moreover, nano- to micro-scale cracks in rocks probably contain absorbed water. In addition, secondary calcite can precipitate from released calcium through the alteration of rocks (Bazilevskaya et al., 2015). The LOI procedure releases volatiles (e.g., water in hydrous clay minerals and nano- to micro-scale cracks), as well as CO2 derived from secondary calcite, from the weathered rocks.

Whole-rock major element composition and determination of Fe3+/Fe2+ values

To estimate the degree of chemical alteration caused by weathering, we measured the whole-rock major element composition and Fe3+/Fe2+ values. When rocks are altered, the whole-rock chemistry changes; e.g., alumina enrichment in the crust of the samples. Fe3+/Fe2+ values can then be used as a proxy for oxidation. To measure the major element composition of the rock samples, we carried out wavelength dispersive X-ray fluorescence (XRF) analysis on the powdered specimens using a Rigaku ZSX Primus II at the Earthquake Research Institute, University of Tokyo, Tokyo, Japan. Our analytical approach followed Hokanishi et al. (2015), which yields sufficient accuracy for evaluating the major element concentrations of silicate rocks, using a fused glass bead of silicate rock powder diluted with lithium-tetraborate at a ratio of 1:5. The minimum wt% of the calibration lines of the major elements used in this study were as follows: SiO2, 39.46; TiO2, 0.005; Al2O3, 0.68; Total FeO⁎, 0.05; MnO, 0.001; MgO, 0.006; CaO, 0.09; Na2O, 0.022; K2O, 0.003; and P2O5, 0.002. The correlation coefficients of the calibration lines used here for all major elements were >0.999.

The chemical index of alteration (CIA, Nesbitt and Young, 1982, Nesbitt and Young, 1984), which is based on the whole-rock chemical composition, can be used as an indicator of alumina enrichment caused by weathering. However, as Kamei et al. (2012) pointed out, the CIA includes chemical variations arising from granitic petrogenesis, and the value of the CIA itself for each rock has only limited meaning in terms of weathering. Therefore, we instead used the CIA ratio (CIAcrust/CIAcore) as an index of alumina enrichment and clay mineral formation.

The whole-rock concentrations of ferrous iron (Fe2+) were determined using the potassium permanganate titration method. Ferric iron (Fe3+) was calculated as the difference between the total iron content determined from the XRF analysis and the ferrous iron content determined by titration.

Rock hardness (Equotip rebound value)

An Equotip electronic rebound hardness tester was used to record rock hardness in relation to rock weathering (e.g., Viles et al., 2011). The Equotip is suitable for measuring the hardness of a wide range of stone and rock surfaces at different stages of weathering because of its low impact energy (Wilhelm et al., 2016). Various methods have been proposed to estimate the surface and subsurface conditions of samples using the Equotip, including the single impact method (SIM), the repeated impact method (RIM), and a combination of the two methods (Wilhelm et al., 2016). In this study, we used the Proceq Equotip 3 with a D-type impact device to measure rock hardness along three survey lines on the halved rock sample (Fig. 3c, d). Measurements were carried out on the core, crust, and at three equally spaced points in between, along the three survey lines (total = 15 points). Aoki and Matsukura (2007) reported that repeated rock hardness tests at the same point result in an artificial increase in rebound strength that is caused by compaction. Thus, we used the SIM. The tick marks on the halved rock sample in Fig. 3c and d illustrate our approach to these measurements. The average values from the equivalent points on the three survey lines were used as the values of rock hardness in this study. Thus, the average value of points 1 to 3 and 13 to 15 are the rock hardness values of the core and crust, respectively.

Laser Raman microspectrometry and electron probe analysis

The high resolution laser Raman microspectrometry is a powerful tool for identifying tiny materials. Consequently, we obtained the Raman spectra of thin and tiny minerals in a thin section from the crust of rock samples using laser Raman microspectrometry (HORIBA Jobin Yvon, LabRAM HR800) equipped with a 532 nm Nd:YAG laser (Showa Optronics, J100GS-16) with an irradiation power of 1.6 mW and a spectral resolution of about ±2.5 cm−1 at Kanazawa University, Japan. The Raman spectra were collected in the range between 100 and 1800 cm−1, and the analytical procedure followed Miura et al. (2011). Chemical analysis of opaque minerals was conducted using a JEOL-8800R electron probe microanalyzer (EPMA) at Nihon University, Japan, with an accelerating potential of 15 kV and a probe current of 12 nA. We used the ZAF matrix correction method.

Results

CSI values

A comparison of CSI values from the core to the crust of samples generally shows an increase with the DW value of the rock samples (Fig. 4a). This indicates that the CSI values of the crust of the rock samples with higher DW values tend to be greater than the CSI values of those samples with lower DW values. This phenomenon is clearly indicated by the CSI ratio (CSIcrust/CSIcore) of the samples (Fig. 4b), which generally increases with the DW value. However, there was one exception, from Dufek (Du–01a), that had a high CSI value even in the core, resulting in a lower CSI ratio.

Fig. 4. Results of color measurement, whole-rock chemistry, and rock hardness tests with relation to degree of weathering (DW). (a) Color strength index (CSI). (b) CSI ratio. (c) Loss on ignition (LOI). (d) LOI ratio. (e) Chemical index of alteration (CIA). (f) CIA ratio. (g) Ferric/ferrous iron proportion (Fe3+/Fe2+). (h) Fe3+/Fe2+ ratio. (i) Result of hardness test for each rock sample. (j) Ratio of rock hardness. In (a), (c), (e), (g) and (i), values of the core and crust are shown by filled and open symbols, respectively. The core and crust of the same samples are connected by thin lines. The ratios are the quotients of the values from the crust divided by those from the core.

Chemical weathering: LOI and major element composition

LOI values were <2 wt% in most specimens for the crust and core of all samples (Fig. 4c). The generally lower values of LOI (<2%) found in this study indicate that the production of hydrous clay minerals within these rock samples was limited. The major element composition and CIA values of all samples from the crust and core shows no obvious enrichment of alumina content (Fig. 4e, 5). The CIA ratios of all samples were <2%, which indicates the limited production of hydrous clay minerals (Fig. 4f). Our LOI and major element composition data are consistent with the X-ray diffraction (XRD) data of Matsuoka (1995) that show a lack of clay minerals in the soil in this area.

Fig. 5. Al2O3–K2O–(Na2O + CaO) ternary diagram showing the chemical weathering trend inferred from the alteration of feldspar with hydrous clay mineral production, after Nesbitt and Young, 1984, Nesbitt and Young, 1989 and Rollinson (1993). Rectangle at right is a close-up view of the area indicated by the small rectangle in the ternary diagram. Numbers beside the diagrams indicate the concentration of Al2O3 (wt%).

However, the LOI values of the crust in the rock samples with a higher DW value tend to be slightly higher than those from the core (Fig. 4c). This phenomenon is also indicated by the LOI ratios (LOIcrust/LOIcore) of the samples (Fig. 4d). On the other hand, the CIA ratios show no obvious increase with increasing DW value (Fig. 4f). These data suggest that the crust of the rock samples with the higher DW values contains a small amount of water as liquid within the nano- or micro-scale cracks, or water in the crystals, such as H2O in biotite and goethite.

Oxidation: iron hydroxide formation and alteration of Fe–Ti oxides

Fig. 4g shows that the Fe3+/Fe2+ values increase from the core to the crust of the samples. As the original Fe3+/Fe2+ values of igneous rock are derived from its magmatic conditions (Gill, 2010), the contrasts in the Fe3+/Fe2+ ratios between the crust and the core in our samples simply reflect subsequent oxidation. The Fe3+/Fe2+ ratio between the crust and core of the samples probably increases with the DW value (Fig. 4h), indicating that oxidation is significant in those samples with higher DW values.

Microscopic observations of crust samples show that they contain brown reddish minerals in the veins and in the red opaque minerals found as inclusions in crystals or as independent crystals (Fig. 6a, b, d). Raman spectral analysis of the brown reddish minerals in the veins, which was only successful for relatively thick vein in the rock sample from Dufek (Du–01b), revealed that these minerals are goethite (e.g., Dünnwald and Otto, 1989; Faria et al., 1997; Fig. 6c). This indicates that the goethite grains, which are ferric iron (Fe3+) hydroxide, formed in the veins in the crust of the rock samples by oxidation.

Fig. 6. (a) Photomicrograph of stained crust of sample Au–03b with a micro-crack filled by a brown material (black arrow; plane polarized light). (b) Hematite (white arrow) in biotite cleavage in sample Au–01a (plane polarized light). (c) Raman spectra and photomicrographs of goethite (black arrows) in a micro-crack in stained crust of sample Du–01b. Photomicrographs of the veins were taken under plane polarized light (left) and reflected light (right). (d) Photomicrograph of hematite and ilmenite with a non-stoichiometric compound in sample Du–01a. White arrows indicate non-stoichiometric Fe–Ti compound. (e) TiO2–FeO–1/2Fe2O3 ternary diagram showing the chemical composition of the hematite and ilmenite shown in (d). Bt: biotite, Qz: quartz, Hem: hematite, Ilm: ilmenite.

EPMA analysis revealed that the red opaque minerals are hematite (Fig. 6b, d, e) and that the non-stoichiometric Fe–Ti compound associated with ilmenite grains are only found in samples from the higher elevation sites at Nils Larsen (Ni–01a) and Dufek (Du–01a and Du–01b). The samples from these two higher-elevation sites yield a DW value of 4 (Table 2). Hematite and ilmenite cannot coexist in the same magma, because of the different oxygen fugacity conditions required for crystallization (Buddington and Lindsley, 1964). In addition, ilmenite cannot be crystalized by subsequent sub-aerial oxidation (i.e., high oxygen fugacity). Thus, the presence of ilmenite in the present rock samples indicates that hematite was not a primary product in the magma and most probably formed as a secondary product during subsequent weathering. This indicates that the original Fe–Ti oxide grains in the rock samples were affected by weathering and partly altered to hematite and to a non-stoichiometric Fe–Ti compound associated with ilmenite grains at the sites with increased DW values. These findings also explain why the contrast in CSI values from core to crust increases with increasing DW value; i.e., because of the intense formation of brown reddish and red opaque minerals in these rock samples.

Rock hardness

The hardness of the crust of our samples generally decreased with increasing DW value (Fig. 4i), although the ratio of the hardness between core and crust was not clearly related to the DW value (Fig. 4j). On the other hand, the hardness of the core of the samples tended to decrease slightly with the DW value (Fig. 4i). This suggests that the reduction in hardness occurred not only in the crust, but also in the core, in the case of the increased DW values. This is why the ratios of rock hardness between core and crust show a weak relationship with the DW value. Profiles of rock hardness from core to crust show that hardness generally decreases from core to crust (Fig. 7). These observations indicate that the crust of our rock samples was typically more fragile than the core.

Fig. 7. Results of rock hardness tests for each sample. Horizontal axis shows the relative distance from sample core to sample crust, increasing towards the crust.

Weathering in a hyper-arid and hypothermal environment

The type and rate of weathering, and the characteristics of weathered rock surfaces are controlled primary by local climatic conditions. Weathering in a hyper-arid and hypothermal environment is controlled mainly by physical (e.g., Campbell and Claridge, 1987; Hamblin and Christiansen, 2003) rather than chemical weathering. Generally, physical weathering breaks down rock and minerals by the thermal expansion and deflation of salt, ice, and quartz, which show high rates of thermal expansion. In addition, based on field observations of weathering style and XRD analyses of rock and soil samples, Matsuoka (1995) and Matsuoka et al. (2006) reported that salt weathering and oxidation are major weathering processes in the Sør-Rondane Mountains. Although they showed clear evidence of the weathering process in this area, the mechanism of oxidization and its products remained unclear.

The LOI value and major element composition of specimens from the crust and core of the rock samples obtained in this study indicate that the production of hydrous clay minerals was very limited (Fig. 4c, 5). In contrast, microscope observations and laser-Raman microspectroscopy showed that goethite grains had formed in the veins in the crust (Fig. 6). This is consistent with the widely differing Fe3+/Fe2+ ratios and CSI values from core to crust (Fig. 4b, g). A negative correlation between rock hardness and DW indicates that the crust of the weathered samples tends to be softer than the core (Fig. 4i). Matsuoka (1995) reported that the temperature of the rock surface in the study area sometimes exceeds 30 °C during the summer, despite the air temperature rarely rising above freezing point. This must cause efficient frost shattering and/or salt fretting. In addition, the thermal expansion and contraction of rock-forming minerals by daily or seasonal temperature changes may also be an effective physical weathering process (Hamblin and Christiansen, 2003). Differences in the rate of thermal expansion between rock-forming minerals may also contribute to the breakdown of rocks (Halsey et al., 1998; Weiss et al., 2004). The rate of thermal expansion of quartz, which is a common constituent of the rocks studied here, is two to three times higher than that the other rock-forming minerals (Skinner, 1966). Therefore, we conclude that cracking by the thermal expansion and contraction of rock-forming minerals, frost shattering and/or salt fretting, and subsequent vein formation causes a reduction in the hardness of the crust and its staining by the formation of goethite grains in the veins.

The increase in CSI values from core to crust also indicates the intense formation of reddish brown and red opaque minerals in these rock samples. EPMA analysis revealed that the red opaque minerals are hematite and that the non-stoichiometric Fe–Ti compounds are associated with ilmenite grains. Our microscope observations suggest that the goethite veins are generally associated with biotite (Fig. 6a). Similarly, the hematite grains and non-stoichiometric Fe–Ti compound associated with the ilmenite grains probably formed via alteration from magnetite and other opaque minerals, such as original ilmenite (Fig. 6e). Alteration of magnetite to hematite discharges excess iron from its structure stoichiometrically. Microscopic texture showed in Fig. 6d indicates that the breakdown of original ilmenite produces the non-stoichiometric Fe–Ti compound. These suggest that the iron sources of the iron hydroxide formation in the veins are biotite and/or original opaque minerals.

The oxidation of annite, which is the Fe end-member of the biotite group, induces dehydroxylation, the chemical reaction for which is as follows (White, 1990):

In general, granitic rocks contain abundant biotite, and a certain proportion of this biotite is the Fe end-member, annite (MacKenzie and Adams, 1994). This means that oxidation of granitic rocks, including annite, can supply H2O to form secondary hydrous minerals, such as goethite. This process can explain the formation of goethite within granitic rocks, even under hyper-arid and hypothermal conditions.

The slight increase in LOI from crust to core with the DW value also suggests that more water was captured in the nano- and micro-scale cracks and/or goethite in the vein in samples with a higher DW (Fig. 4b), because the production of hydrous clay minerals was not evident in the whole-rock chemistry (Fig. 5) and LOI values were generally low (<2 wt%; Fig. 4c). At this stage, the crust of the rock becomes fragile due to cracking and subsequent vein formation, allowing the Fe–Ti oxide grains to be strongly oxidized and converted to hematite, and the ilmenite grains to be converted to non-stoichiometric Fe–Ti compound. These features that appear when rock weathered in this area seems to have little relation with rock type. We propose that these are the major weathering processes that affect granitic rocks under the hyper-arid and hypothermal conditions of Antarctica.

Fig. 8 presents a schematic model of the weathering process in a hyper-arid and hypothermal environment. The mechanical disturbance caused by thermal expansion and contraction of rock-forming minerals, frost shattering and/or salt fretting forms micro cracks in the rock, which then develop into veins (Fig. 8a, b). Subsequently, goethite grains form in the veins. When the weathering process proceeds, the veins grow towards the core of the rock by further cracking, and the oxidation process also proceeds with the formation of hematite and the non-stoichiometric Fe–Ti compounds (Fig. 8c).

Fig. 8. Schematic illustration of the weathering process (oxidation) in the study area. (a) Fresh rock. (b) Goethite vein formation following nano- or micro-crack propagation by the mechanical disturbance caused by thermal expansion and contraction of rock-forming minerals, frost shattering and/or salt fretting. (c) Oxidation progress towards the sample core. Increased oxidation causes alteration of original Fe–Ti oxide to form hematite and non-stoichiometric Fe–Ti compounds associated with ilmenite grains. Gray and red diamonds indicate unaltered and altered Fe–Ti oxides, respectively.

Implications

Physical basis for the Moriwaki index and the weathering process

The negative correlation between rock hardness and CSI values was also confirmed (Fig. 9). We found that the CSI values and the Moriwaki index from the same site are positively correlated (Fig. 10a). These results indicate that the physical and chemical weathering processes described in this study provide a physical basis for the Moriwaki index.

Fig. 9. Plot of rock hardness versus CSI values.
Fig. 10. Plots of (a) crust CSI versus the Moriwaki index (WI) of the sampling site; (b) LOI ratio versus relative height; and (c) crust CSI versus relative height.

Suganuma et al. (2014) reported that the glacial till deposits in the Sør-Rondane Mountains can be divided into four stages using the Moriwaki index. The 10Be surface exposure ages from erratic boulders and bedrock vary from 3.39 to 2319 ka, and generally increase as a function of increasing relative height up to about 700 m. Thus, the increases in the LOI ratio and CSI values of the crust with relative height shown in Fig. 10b and c also correspond to a longer period of exposure. The samples showing the highest LOI ratio and CSI values have been exposed to the hyper-arid and hypothermal environment for millions of years.

These long exposure periods indicate that the weathering process in the Sør-Rondane Mountains is extremely slow compared with that of hot and humid locations (e.g., Hamblin and Christiansen, 2003). Matsuoka (1995) reported that salt weathering and oxidation is a major weathering process in this area. In this study, we concluded that the formation of iron hydroxide in the veins created by mechanical disruption is a major source of soil, and this is consistent with the geochemical analysis of Matsuoka et al. (2006). As the oxidation process is relatively quick in terms of the geological timescale, this indicates that the duration of the physical and chemical weathering is essentially dependent on the cracking caused by thermal expansion and contraction of rock-forming minerals, frost shattering and/or salt fretting. Therefore, we suggest that the restricted supply of liquid water and/or salt is the most likely cause of the extremely slow weathering rates in the Sør-Rondane Mountains. Consequently, as the rock color recorded in the field is directly related to the amount of iron hydroxide on the rock surface, in situ and satellite measurements of the color of exposed rocks in Antarctica can potentially provide quantitative information regarding the degree of weathering.

Analog for the weathering system on Mars

The present results indicate that during the weathering of granitic rocks in Antarctica, secondary hydrous minerals (e.g., goethite) can form following fracturing of the rock and subsequent oxidation, even under hyper-arid and hypothermal conditions. Our data suggest that physical and chemical weathering in a hyper-arid and hypothermal environment does not rely on water supplied from the atmosphere (snow or vapor) and that this process could be used as an analog for weathering on Mars. Previous studies of the weathering process on Mars have focused mainly on mafic rocks; however, the existence of felsic rocks on Mars has recently been reported (Wray et al., 2013). In addition, the Mars exploration rover detected goethite using Mössbauer spectrometry (Klingelhöfer et al., 2005), and minor amounts of biotite were found in a Martian meteorite (Bridges and Warren, 2006). These findings suggest that water-containing material, such as goethite and biotite, probably spread on the surface or subsurface of Mars. Moreover, it is possible that the oxidation of Fe/Mg phyllosilicates, which occur on the Martian surface (Mustard et al., 2008), supplies the water that enables the generation of goethite in Martian surface rocks. Therefore, our finding that the weathering of granitic rocks can occur without water from the atmosphere will potentially assist our understanding of landform evolution on Mars.

Conclusions

In this study, we have investigated the physical and chemical weathering of the granitic rocks of Dronning Maud Land in East Antarctica using color measurements, whole-rock chemistry, mineralogical experiments, rock hardness tests, and microscope observations. The contrast in CSI values between the core and crust of the samples generally increased with the DW value. The major element composition of all subsamples from crust and core showed no obvious enrichment in alumina content, which indicates little production of hydrous clay minerals in this area. Consequently, the contrast in LOI values between crust and core in the samples with a greater DW probably indicates that more water is captured in the nano- and micro-cracks or water in the crystals, such as goethite in these samples. The microscope observations and laser-Raman microspectroscopy of thin sections from the crust and core indicate that goethite grains form mainly in the veins in the crust, and this is consistent with the increased Fe3+/Fe2+ contrast between core and crust. Rock hardness tests show that the crust of the rock samples is generally softer than the core. A negative correlation between the rock hardness and CSI value also indicates that the crust of the rock samples tends to be softer than the core following the cracking of the rock and subsequent goethite formation. Furthermore, EPMA analysis indicated that the original Fe–Ti oxide grains in the core of the samples were affected by weathering and altered to hematite and non-stoichiometric Fe–Ti compounds associated with ilmenite grains, in the samples with a higher DW value.

Our findings reveal that the weathering of granitic rocks in a hyper-arid and hypothermal environment is controlled mainly by oxidation, including iron hydroxide formation in the veins created by mechanical disruption caused by thermal expansion and contraction of rock-forming minerals, frost shattering and/or salt fretting, and Fe–Ti oxide alteration in the rock interior. Importantly, these physical and chemical weathering processes probably do not need a liquid water supply from the environment (snow and vapor), and the restricted supply of liquid water and salt is the most probable reason for the extremely slow weathering rates observed in the Sør-Rondane Mountains. We therefore conclude that the physical and chemical weathering processes revealed in this study are a typical feature of the hyper-arid and hypothermal environment of Antarctica, and potentially also of Mars. The measurement of rock color essentially expresses the amount of iron hydroxide on the rock surface. This indicates that the physical and chemical weathering process provides a physical basis for the Moriwaki index so that this index can potentially be used to convert in situ and satellite measurements of the color of exposed rocks in Antarctica into a quantitative indicator of the degree of weathering.

The following are the supplementary data related to this article.

Supplementary Fig. 1. Photographs of sampling sites: (a) Austkampane: Au–01, (b) Austkampane: Au–02, (c) Austkampane: Au–03, (d) Dufek: Du–01, (e) Mefjell: Me–01, (f) Mefjell: Me–02, (g) Nils Larsen: Ni–01, (h) Lunckeryggen: Lu–01, and (i) Lunckeryggen: Lu–02.

Source: The weathering of granitic rocks in a hyper-arid and hypothermal environment: A case study from the Sør-Rondane Mountains, East Antarctica

Authors: Tatsuo Kanamaru, Yusuke Suganuma, Hisashi Oiwane, Hideki Miura, Makoto Miura, Jun’ichi Okuno, Hideaki Hayakawa

One thought on “The weathering of granitic rocks in a hyper-arid and hypothermal environment: A case study from the Sør-Rondane Mountains, East Antarctica

Leave a Reply

Your email address will not be published. Required fields are marked *