Abstract

Key Words

Terra Rossa, marble, pedogenesis, Mediterranean climate, Fleurieu peninsula, FESEM.

Introduction

Terra Rossa soils are common in areas with a Mediterranean climate (i.e. cool, wet winters alternating with warm, dry summers). They characteristically overlie hard calcareous bedrock such as crystalline limestone or marble or unconsolidated calcareous deposits of terrestrial or marine origin (Kubiena 1953). The genesis of these highly structured, uniform textured red soils has long been a matter of controversy. Two hypotheses have currency: (1) In the ‘residual theory’ Terra Rossa development is the result of carbonate dissolution and subsequent accumulation and transformation of limestone residue (Bronger and Bruhn-Lobin 1997; Gal 1967; Moresi and Mongelli 1988). (2) The other hypothesis is that the soil is unrelated to the underlying rock and is allochthonous in origin. In Many cases aeolian dust has been shown as the parent material of red soils in the Mediterranean region (MacLeod 1980; Yaalon 1997).

Preliminary fieldwork in the southern and central Mount Lofty Ranges revealed a morphologically similar group of Terra Rossa soils over thin, elongate marble lenses of the lower Cambrian Normanville Group. Although Terra Rossa soils of South Australia have been studied (McIntyre 1956; Norrish and Rogers 1956; Stace 1956), researchers have focused on regions other than the southern Fleurieu peninsula and have been mainly concerned with mineralogical development. This region, however, provided an ideal scenario in which to test the hypotheses stated above. First, preliminary examination suggested that the marbles of the Normanville Group are quite pure, i.e. they have little argillaceous material. Large thicknesses of parent material would need to be weathered to account for the thickness of the solum. In a study of the genesis of a Terra Rossa in Epirus, Greece, MacLeod (1980) calculated that for 40 cm of soil to develop, of the order of 130 m of limestone would need to be weathered. Given Yaalon and Ganor's (1975) calculation of 1-2 cm per 10³ years for limestone denudation in Judea and Galilee, MacLeod (1980) suggested that residue released from limestone during denudation must accumulate at a rate of 8×10^-6cm/a for 5×10⁶ years. Even the oldest surfaces of the Delamere region are much younger than this (Wellman and Greenhalgh 1988). Should such estimates apply here, given the tectonic instability of the Adelaide Geosyncline (Firman 1969; Webb 1957; Wellman and Greenhalgh 1988), it is unlikely that a soil formed during the Tertiary could survive erosion. This implies that the residual theory is an unlikely explanation.

In addition, the marble lenses are found side by side with shales of the Normanville Group on which brown podsolic soils with strong texture contrast have developed. The transition between soil types is quite abrupt and seems coincident with lithological change. This implies that lithology is the dominant influence on the soil type and that the accession of material from upslope and/or aeolian deposition also seems an unlikely explanation.

The purpose of the study reported here was to establish the relationship of the Terra Rossa soil to the underlying marble and the extent and origin of outside influences. Analytical tools include morphological study of the soils, X-ray diffraction analysis of soil fractions and field emission scanning electron microscopy.

Geology and landscape

The lithology of the Delamere region comprises sandstones, siltstones and carbonates of the Cambrian Kanmantoo and Normanville Groups that belong to the Adelaidean sequence (Preiss 1987), as well as the Permo-Carboniferous glacial deposits of the Cape Jervis Formation (Bourman and Alley 1988). Terra Rossa soils are found overlying the hard Cambrian marbles/dolomites known as the Angaston marble which is a member of the Normanville Group (Figure 1).

Figure 1. Sampling locality and regional distribution of the Terra Rossa and related geology.

Delamere is located on the Fleurieu peninsula approximately 100km south of Adelaide. The annual mean minimum temperatures are between 9-12°C and the corresponding maximum temperatures between 18-21°C. Annual rainfall is between 400 and 900 mm. The mean number of rainy days per year is 126.5 (Commonwealth Bureau of Meteorology, 2003). The climate – cool wet winters with excess water, alternating with warm dry summers – produces a xeric moisture regime in the soil profile. Undulating rises and low hills with slopes between 10 and 30% characterise the landform.

Methods

Preliminary mapping and sampling led to the identification of several possible sites for the study. These were selected on the ease of access and the extent and uniformity of the Terra Rossa. The main site was selected on the basis of recent soil-landscape mapping by the Department of Primary Industries and Resources S.A. (Figure 1). Using a backhoe, two trenches were dug, one in the Terra Rossa and the other in a Brown Podsolic soil of strong texture contrast from further upslope. Hereafter this soil profile will be referred to as the Upslope Soil. Following the photography of the profiles, 1-2kg bulk samples of each horizon to bedrock were collected from the face of the trench. Sampling depths were based on changes in colour and texture down the profile. The morphology of the profiles were described using the Munsell colour system, along with methods outlined in McDonald et al. (1984) for texture and structure.

Large undisturbed clods of soil were collected for micromorphological analysis. Sections of soil were carefully excised and packed for transport to the Canberra laboratories of C.S.I.R.O. Land and Water where they were impregnated under vacuum with a polyester resin by the standard method of Brewer (1976). The impregnated blocks of soil were then sectioned to produce 7.5×5cm slides for micromorphological analysis. Fabric and mineralogy were observed and photographed using a Leitz microscope under polarized light according to the method of Brewer and Sleeman (1988).

Separation of the fine earth (<2000μm) into clay (<2μm), silt (2-20μm), very fine sand (20-53μm), fine sand (53- 125μm) and coarse sand (125-2000μm) fractions was carried out by sedimentation and sieving. The insoluble residue of the marble was obtained through dissolution via 0.5м HCl solution. The same process allowed the calculation of the percentage of insoluble residue by comparing the initial and final weights of the material. The residue was separated following the same procedure as that described above.

Cation exchange capacity was measured using the alcoholic ammonium chloride procedure based on the method described by Rayment and Higgson (1992). Exchangeable bases were determined using atomic absorption for calcium and magnesium and flame photometry for potassium and sodium. Electrical conductivity and pH were measured in a solution of 8g of sample and 40mL of water.

Oriented disks were prepared for x-ray diffraction by pipetting a suspension of clays, under suction, onto silver filters of pore size 0.45μm. Clays were saturated with 1м MgCl₂, washed free of excess electrolyte and saturated with 10% glycerol. X-ray diffractograms were generated using a Philips PW 1800 microprocessor-controlled diffractometer with CoKα radiation, variable divergence slit and a graphite monochromator. Patterns were acquired from 3 to 33° 2θ in 0.02° steps and analysed using the software program X-Plot. A sample from the silt fraction of the Terra Rossa from depth 40-47cm received treatment with 1м KCl and 5% glycerol to determine the presence of vermiculite.

Mineral types and morphology of grains in the coarse sand fractions were observed using a Philips XL30 Field Emission Scanning Electron Microscope (FESEM) with the acceleration voltage set to 20kV. The samples were fixed to the sample holder with double-sided tape and a carbon coating was applied prior to analysis. Mineralogical analysis conducted with the FESEM required use of the EDAX program iDXI(maps). The concentrations of elements in each grain were measured by EDAX. The mineralogy of each grain was assigned manually by scanning its elemental distribution. Each sample was analysed in 56 fields. Within each field iDXI(maps) distinguished individual grains and, using x-rays, determine the elements present. Due to the volume of data, the mineralogy of each sample was determined from only 5 randomly selected fields out a possible 56.

Results and Discussion

Morphological analysis

The Terra Rossa profile was dark reddish brown with hues ranging from 2.5YR to 5YR. The C horizon (weathered bedrock) was strong brown with a hue of 7.5YR and is distinctly calcareous in contrast to the A and B horizons that contained no noticeable carbonate. The profile had a strong subangular to angular blocky structure and medium texture (clay loam) throughout. The soil was a gradational clay loam, Petrocalcic Red Dermosol (Isbell 1996).

The upslope profile was dark reddish brown to very dark brown in the A and upper B horizons and strong to light brown in the lower B and C horizons. The hues ranged from 5YR in the A to 10YR in the C horizon. There was a strong texture contrast between A and B horizons unlike the texture of the Terra Rossa profile, which was more uniform throughout. The soil was a loam over brown clay and was classified as a Eutrophic Brown Chromosol (Isbell 1996).

Micromorphological analysis of the soils under thin section showed the Terra Rossa to have an iron-rich, asepic plasmic fabric i.e. the clay particles were highly flocculated, demonstrating no visible clay separations. This strong structure, characteristic of Terra Rossas, was attributed to the high levels of exchangeable Ca²⁺ released from the carbonates during dissolution (Table 1). High flocculation of particles due the to Ca²⁺, as well as iron oxide and organic matter, would inhibit the movement of clay through the profile, suggesting clay illuviation should not be evident in the Terra Rossa profile. Nevertheless, particle size data showed a relative increase in clay-sized particles from the A to the B horizon (Figure 2a). The Terra Rossa also demonstrated a high degree of micro aggregation of particles. Analysis of the marble underlying the Terra Rossa revealed highly variable particle sizes that occurred as bands of like-sized grains through the rock and contained argillaceous material in stylolites. The upslope profile had a high degree of clay mobility, indicated by the formation of argillans in cracks and voids of the soil. The lower Ca²⁺ levels in the Upslope profile suggest clay particles could move more freely through the profile allowing clay illuviation to take place. The greater proportion of clay sized particles in the B horizon than the A is consistent with this (Figure 2b).

Table 1. Chemistry of the Terra Rossa and Upslope Profiles.

Profile	Depth (cm)	Horizon	CEC (meq/100g)	pH	EC (mS)	Exchangeable Bases (meq/100g)
						K	Na	Ca	Mg
Terra Rossa	0-10	A	12.1	6.2	0.45	0.8	1.4	22.6	3.8
	20-28	B	12.1	5.83	0.25	0.6	1.2	20	2.9
	40-47	B	15	5.88	0.23	0.8	1.4	20.4	3.5
	47-55	C	5.0	8.38	0.15	0.5	1.4	25.4	1.4
Upslope	0-10	A	8.5	5.57	0.28	0.7	1.0	14.8	3.0
	23-42	B	9.2	5.51	0.18	1.2	2.8	10	5.4
	42-63	B	11.1	5.83	0.14	1.2	1.4	11.4	7.3
	83-100	C	17.3	6.94	0.09	1.0	1.6	9.6	7.3

Further, the Ca²⁺ input from the marble provided an explanation as to why the Terra Rossa was observed as coincident with lithology, despite contributions from upslope. It is possible that Ca²⁺ gave structure throughout the profile, promoting air flow and thus ferric iron oxide development (particularly haematite), giving the soil its characteristic red colour, regardless of the contributions from upslope (Boero and Schwertmann 1989; Bronger 1983; Singer et al. 1998; Torrent et al. 1983; Torrent and Schwertmann 1987). The upslope profile could also be been high in iron, however, due to the lack of Ca and subsequent structure, goethite may have formed preferentially to haematite and so the soil remained brown (Boero and Schwertmann 1989).

Particle size analysis and chemistry

Both the Terra Rossa and Upslope profiles had increasing levels of clay with depth (Figure 2). Despite this, given the apparent immobility of the clay one would expect there to be a relatively homogenous amount of clay throughout the Terra Rossa profile (Figure 2a). This implies that the Terra Rossa is receiving contributions of coarse material from some outside source (i.e. allochthony). The micromorphology of the Upslope profile suggests that perhaps clay is mobile here and that the process of illuviation could account for greater clay in the B horizon than the A horizon (Figure 2b).

After dissolution of carbonate, the marble consisted of approximately 6.5% insoluble residue that is predominantly of sand size (Figure 2a). The depth of marble required to produce 40cm of soil was calculated as 3-4m with 6.5% insoluble residue. This calculation assumed the bulk density of the soil and the marble were 1.3g/cm and 2.65g/cm respectively (MacLeod 1980), as these data were not available. This figure is much less than MacLeod’s (1980) Mediterranean example. Assuming a weathering rate of approximately 0.3cm/ka and therefore a formation time of 1.5×10⁵ years an autochthonous explanation for the origin of the Terra Rossa is acceptable.

Figure 2. Depth plots of the (a) Terra Rossa and (b) Upslope profiles showing relative proportions of grain size in the fine earth. Depths greater than 55cm in the Terra Rossa profile represent the particle size analysis of the marble insoluble residue.

Mineralogical analysis

The XRD analyses shown in figure 3a revealed that the larger amounts of quartz and the comparatively smaller amounts of primary mica in the upper horizons of the Terra Rossa silt (2-20μm) fraction were distinct from the subsoil. There were greater levels of illite compared to quartz in the marble and an obvious relationship with the lower B horizon (40-47cm) by way of the mica-vermiculite association, and the similarly small amounts of quartz. The presence of well ordered vermiculite in the lower B horizon was indicative of the weathering of primary mica from the marble. It is possible that weathering of mica resulted in the release of K⁺ into the soil solution and the dissolution of the marble resulted in the release of Ca²⁺. Thus, the exchange of Ca²⁺ for K⁺ in the interlayer spaces of the clay converted the mica into vermiculite (Bassett 1959). In the A (0-10cm) and upper B horizons (20-28cm) no vermiculite was present and considerably larger amounts of quartz were found. This suggested perhaps an allochthonous origin for the upper horizons of the profile whereas closer to the marble the soil was largely autochthonous.

The view of an allochthonous origin for the upper portion of the profile was further strengthened by the XRD traces from the texture contrast profile upslope (Figure 3b). In the silt fraction it was clear that quartz and primary mica were the dominant minerals, with quartz the more abundant. This was identical to the mineralogy of the A and upper B horizons of the Terra Rossa. Furthermore, the considerable levels of quartz in the clay fraction throughout the profile were similar to the upper horizons of the Terra Rossa profile.

Figure 3. Mineralogy of the silt fraction (2-20µm) of selected horizons of the (a) Terra Rossa and (b) Upslope soils.

Coarse particle morphology and elemental chemistry

The FESEM counts of quartz, mica and K feldspar from the coarse sand fraction deep in the Upslope profile showed roughly equal proportions. Proportions changed gradually such that quartz became dominant in the upper horizons (Table 2). This suggested that as the soil developed, the less resilient minerals (i.e. mica and K feldspar) weathered leaving relatively more quartz.

The insoluble residue from the underlying marble of the Terra Rossa had considerably more mica than the upslope profile. Similar results were observed in the B horizon, whereas the A horizon was clearly dominated by quartz (Table 2). One explanation was that perhaps the mica and feldspar have weathered out of the profile, leaving proportionately more quartz. There was, however, a lack of evidence for a much more chemically aggressive environment in the A horizon, relative to the B horizon, that would predispose mica and feldspar of silt and sand size to break down to such a degree and yet leave the quartz grains with similar morphology in the A horizon as the grains in lower horizons. Delgado et al. (2003) came to a similar conclusion: to quote these authors, Terra Rossas have a “relatively calm pedoenvironment”. A more likely explanation is that the upper horizon had received contributions from upslope, i.e. at least the A horizon of the soil was allochthonous in origin.

Table 2. FESEM mineralogy counts of 125-2000 μm fraction.

Profile	Depth (cm)	Mineral Type (%)a
		Quartz	Mica	K Feldspar
Terra Rossa	0-10	65	32	3
	20-28	30	50	20
	Dolomite (IR)	38	61	1
Upslope	0-10	82	8	10
	42-63	70	18	12
	85-100	28	33	39

^aPercent calculated relative to each other.

This was supported by the morphological analysis of the grains under the FESEM. The angular morphology of quartz grains (Figure 4b) from the insoluble residue was indicative of dissolution of silica and was consistent with the alkaline conditions associated with chemical weathering of carbonate (Dove & Nix, 1997). This morphology was not a feature that was retained through the Terra Rossa profile. Whereas very few grains show this morphology in the B horizon, the A horizon was comprised of solely rounded grains. Should the grains from the A and B horizon have been derived from the marble, an angular, dissolution-scarred morphology should predominate (Figure 4b). Instead, subrounded to well-rounded quartz grains (Figure 4a), that showed a similar morphology to the grains from the upslope profile (Figure 4c), were dominant. The severe weathering required to round the grains to this degree is achieved by sedimentological processes, thus grains must have been rounded prior to deposition. It is highly improbable that the dissolution scarring of the quartz in the marble would disappear in the Terra Rossa soil environment. The most likely explanation was that, while the marble residue contributed a minor amount of sand to the upper horizons, the major contribution of sand was from upslope.

Figure 4. (a) Angular quartz grain showing dissolution scars from the dolomite. (b) Subrounded quartz grain from the A horizon of the Terra Rossa. The surface is relatively smooth with some pits and grooves. (c) Well rounded quartz grain from the A horizon of the Uplslope profile.

Conclusion

The results of the physico-chemical, mineralogical and morphological studies leads us to conclude that the Terra Rossa soil was partly autochthonous in origin, having inherited material from the underlying marble, and partly allochthonous. The external source was most likely silt and fine sand eroded from the A horizon of a Brown Chromosol soil upslope. The soil can, however, be regarded as a lithomorphic soil as it depends on the Ca from the marble to give it the features unique to the Terra Rossa, i.e. porous, highly structured, red, medium-textured, shallow soils (Bronger 1983).

References

Bassett WA (1959) Origin of the vermiculite deposit at Libby, Montana. American Mineralogist 44, 282-300.

Boero V, Schwertmann U (1989) Iron oxide mineralogy of terra rossa and its genetic implications. Geoderma 44, 319-327.

Bourman RP, Alley NF (1988) Permian glacial features, north western Fleurieu Peninsula, South Australia. Quarterly Geological Notes of the Geological Survey of South Australia 108, 2-6.

Brewer R (1976) Fabric and Mineral Analysis of Soils, pp. 482. Krieger, New York.

Brewer R, Sleeman JR (1988) Soil Structure and Fabric, pp. 173. CSIRO, Melbourne.

Bronger A (1983) Rubification of Terrae Rossae in Slovakia: A Mossbauer effect study. Clays and Clay Minerals 31, 269-276.

Bronger A, Bruhn-Lobin N (1997) Palaeopedology of Terrae rossae-Rhodoxeralfs from Quaternary calcarenites in NW Morocco. Catena 28, 279-295.

Commonwealth Bureau of Meteorology, 7 October 2003. Climate averages by number. <www.bom.gov.au/climate/averages/tables/cw_023751.shtml>.

Delgado R, Martin-Garcia JM, Oyonarte C, Delgado G (2003) Genesis of the terrae rossae of the Sierra Gador (Andalusia, Spain). European Journal of Soil Science 54, 1-16.

Dove PM, Nix CJ (1997) Influence of the alkaline earth cations magnesium, calcium and barium on the dissolution kinetics of quartz. Geochimica et Cosmochimica Acta, 61: 3329-3340.

Firman JB (1969) Quaternary period. In 'Handbook of South Australian Geology'. (Ed. LW Parkin) pp. 204-233. (Geological Survey of South Australia: Adelaide)

Gal M (1967) Clay mineralogy in the study of the genesis of terra rossa and rendzina soil originating from calcareous rocks. Proceedings of the International Clay Conference, Jerusalem, Israel, 1966 1, 199-207.

Isbell RF (1996) The Australian Soil Classification, CSIRO, Melbourne.

Kubiena WL (1953) 'The Soil of Europe.' (Thomas Murby and Company: London)

MacLeod DA (1980) The origin of the Red Mediterranean soils in Epirus, Greece. Journal of Soil Science 31, 125-136.

McDonald RC, Isbell RF, Speight JG, Walker J, Hopkins MS (1984) 'Australian Soil and Land Survey Field Handbook.' (Inkata Press: Melbourne)

McIntyre DS (1956) The effect of free ferric oxide on the structure of some terra rossa and rendzina soils. Journal of Soil Science 7, 302-306.

Moresi M, Mongelli G (1988) The relation between the terra rossa and the carbonate free residue of the underlying limestones and dolostones in Apulia, Italy. Clay Minerals 23, 439-446.

Norrish K, Rogers LER (1956) The mineralogy of some terra rossas and rendzinas of South Australia. Journal of Soil Science 7, 294-301.

Preiss WV (1987) 'The Adelaide Geosyncline - late proterozoic stratigraphy, sedimentology, palaeontology and tectonics.' (Bulletin of the Geological Society of South Australia: Adelaide)

Rayment GE, Higgson FR (1992) 'Australian Laboratory Handbook of Soil and Water Chemical Methods.' (Inkata Press: Melbourne)

Singer A, Schwertmann U, Friedl J (1998) Iron oxide mineralogy of terrae rossae and rendzinas in relation to their moisture and temperature regimes. European Journal of Soil Science 49, 385-395.

Stace HCT (1956) Chemical characteristics of terra rossas and rendzinas of South Australia. Journal of Soil Science 7, 280-293.

Torrent J, Schwertmann U, Fechter H, Alferez F (1983) Quantitative relationships between soil color and hematite content. Soil Science 136, 354-358.

Torrent J, Schwertmann U (1987) Influence of hematite on the color of red beds. Journal of Sedimentary Petrology 57, 682-686.

Webb BP (1957) Summary of tectonics and sedimentation. In 'The Geology of South Australia'. (Ed. LW Parkin) pp. 136-148. (Journal of the Geological Society of Australia)

Wellman P, Greenhalgh SJ (1988) Flinders/Mount Lofty Ranges, South Australia; their uplift, erosion and relationship to crustal structure. Transactions of the Royal Society of South Australia 112, 11-19.

Yaalon DH (1997) Soils in the Mediterranean region: what makes them different? Catena 28, 157-169.

Yaalon DH, Ganor E (1975) Rates of aeolian dust accretion in the Mediterranean and desert fringe environments of Israel. International Congress of Sedimentology, Nice 2, 169-174.