Soil Formation on the Namaqualand Coastal Plain

Francis, Michele Louise (2008-03)

Thesis (PhD (Soil Science))--Univ ersity of Stellenbosch, 2008.

The (semi-)arid Namaqualand region on the west coast of South Africa is wellknown for its spring flower displays. Due to the aridity of the region, soils research has lagged behind that of the more agriculturally productive parts of South Africa. However, rehabilitation efforts after the hundred or so years of mining, coupled with the increasing ecology and biodiversity research, have prompted a recent interest in Namaqualand soils as a substrate for plant growth. The area is also notable for the abundance of heuweltjies. Much of the previous heuweltjie-work focussed on biogenic aspects such as their spacing, origin and age, but although heuweltjies are in fact a soil feature, there have been few published studies on the soil forming processes within heuweltjies. However, the depositional history of the sediments on the Namaqualand coastal plain is well constrained, which is in stark contrast to the paucity of data on their subsequent pedogenesis. Given that the regolith has been subaerially exposed in some parts for much of the Neogene, the soil formation forms an important part of the sediments’ history. The primary aim of this thesis, therefore, was to examine the soil features of the Namaqualand coastal plain to further the understanding of pedogenesis in the region. The regolith of the northern Namaqualand coastal plain, often ten or more metres deep, comprises successive late Tertiary marine packages, each deposited during sea-level regression. The surface soil horizons formed from an aeolian parent material. The relatively low CaCO3 in the aeolian sands dictated the pedogenic pathway in these deposits. The non-calcareous pathway lead to clay-rich, redder apedal horizons that show a stronger structure with depth, and generally rest directly on marine sands via a subtle discontinuity that suggests pedogenesis continues into the underlying marine facies. The calcareous pathway lead to similar clay-rich, redder apedal B horizons, but which differ in that they are calcareous, and rest on a calcrete horizon often via a stoneline of rounded pebbles. Deeper in the profile, there is generally a regular alteration of sedimentary units, with the upper shoreface facies showing reddening, and the lower shoreface sands remaining pale. This seems to be a function of the grain size, since the upper shoreface materials are coarser, and the redder parts of the lower shoreface are also associated with slightly coarser sands. In some strata the oxidation of glauconite-rich sediments resulted in an orange colour. In an area with abundant heuweltjies, a strongly-cemented calcretized nest was present about 2 m deep within a silica cemented, locally calcareous dorbank profile. Vertical termite burrows are present up to 12 m deep, and appear to have been conduits for preferential vertical flow. Soil formation and termite activity is at least as old as the Last Interglacial. E horizons may have formed in a wetter Last Interglacial paleoclimate, but they are still active in the present day. The Namaqualand coastal plain, with its extensive areas of calcrete development, is almost a textbook setting for calcrete development by inorganic processes. However, these calcretes also show microscale biogenic features. These include M rods, MA rods, and fungal filaments. Abiotic alpha-fabric seems dominant in mature calcrete horizons, and beta-fabric in calcareous nodules in a calcic B horizon above calcrete. The apparent absence of Mg-calcite and dolomite, and abundance of sepiolite in the calcretes of coastal Namaqualand suggests that these Mg-rich clay minerals are the main Mg-bearing phase. Deformation (pseudo-anticlines) in the calcrete appear to result primarily from the displacive effect of calcite crystallization. Although evidence of shrink/swell behaviour is present in the form of accommodating planes, it does not appear to be as volumetrically significant as displacive calcite. Indurated light-coloured horizons that resembled calcrete but are non- to mildly calcareous, break with a conchoidal fracture, resist slaking in both acid and alkali, turn methyl-orange purple, and show a bulk-soil sepiolite XRD peak are similar to palygorskite-cemented material (‘palycrete’) from Spain and Portugal, and so were tentatively named ‘sepiocrete’. Sepiolite and palygorskite are often reported from arid region soils but there has been no recorded cementation of soils by sepiolite. The degree of induration in some of these horizons suggest that amorphous silica could play a role in cementation, and so this thesis compares the two silica-cemented horizons encountered in Namaqualand (silcrete and dorbank (petroduric)) to these ‘sepiocrete’ horizons. Both silica and sepiolite are present in the matrix, although the degree to which silica and sepiolite dominate seems to vary even within same horizon. It seems most probable that both contribute to the structural properties of the horizon. Sepiolitic horizons do not form a diagnostic horizon in the World Reference Base, Soil Taxonomy, or the South African system. To fit the existing soil classification schemes, the terms ‘sepiolitic’ and ‘petrosepiolitic’ (in the same sense as ‘calcic’ and ‘petrocalcic’) would be appropriate. The term ‘sepiolitic’ should be used for horizons which: contain sepiolite in amounts great enough for it to be detected by XRD in the bulk soil, peds (a fractured surface and not just the cutan) cling strongly to the wetted tongue, and methyl orange turns from orange to purple-pink over most of a fragmented surface. The term can be easily be applied as a adjective to other hardpans where sepiolite is significant but not necessarily cementing, such as ‘sepiolitic’ petrocalcic/petroduric. If the horizon is in addition to the above criteria cemented to such a degree that it will slake neither in acid (so cannot be classified as petrocalcic) nor in alkali (and so cannot be classified as petroduric) then the term ‘petrosepiolitic’ would be appropriate. The ‘sepiolitic’ criteria distinguish the ‘petrosepiolitic’ horizon from a ‘silcrete’, a silica-cemented horizon which does not fit the definition of petroduric. Sepiolite is more prominent than palygorskite in the XRD traces. The <0.08 μm fraction is the only size fraction where palygorskite could be detected before acetate treatment. It is unlikely that these fibrous clay minerals are inherited from either the marine or aeolian parent materials, they appear to be pedogenic in origin. Sepiolite and palygorskite are associated with the presence of calcite in the soil profile. Trends in MgO, Al2O3 and SiO2 show that the soil clay fractions lie on a mixing line between sepiolite and mica end-members, with a contribution from smectite, and is consistent with the XRD and TEM results. There is a good correlation between Fe2O3 and TiO2, which can be attributed to the ubiquitously presence of mica. There was no TEM evidence of fibrous mineral degradation to sheet silicates, nor for the evolution of mica laterally to a fibrous mineral. SEM analyses show that much of the sepiolite/palygorskite occurs as fringed sheets, but higher magnification often revealed these sheets to be composed of fibres. These are found coating (rather than evolving from) mica/illite particles, as free-standing mats, and are common on the grain-side of cutans. Some of these textures suggest illuviation of the fibrous clay minerals, but another explanation may be that sites such as that immediately adjacent to silicate grains have the highest concentration of silica for their formation. There was no conclusive evidence for or against the presence of kerolite in the clay fraction, although it does not appear to be a dominant phase in the <2 μm fraction. The hypothesis was that the permeable upper horizons in Namaqualand soils constitute a shallow ephemeral aquifer, which can be considered the pedogenic analogue of the saline lake environments in which sepiolite typically forms. The chemical evolution of the soil solution and clay mineral genesis could therefore be considered in the same terms as the geochemical evolution of closed-basin brines. The Namaqualand coastal plain, like other maritime areas, shows a trend of decreasing pH, increasing Ca and increasing Mg with increasing evaporation. This can be explained by their seawater-influenced initial ratios, and is consistent with the ‘chemical divides’ of the Hardie-Eugster model of brine evolution. Halite remains undersaturated at all concentrations in the saturated paste extracts. At higher concentrations, gypsum reaches saturation, and sulfate is removed from solution. H4SiO4 activity remains unchanged for all levels of evaporation and pH. Calcite remains close to saturation, and is only dependent on the HCO−3 activity and pH for the range of Cl− activity encountered. Most of the soils for which there is a positive sepiolite identification show a positive sepiolite saturation index. The sepiolite saturation index is independent of Mg2+ and H4SiO4 and only increases with increasing pH. Evidence of the pH control on sepiolite saturation is that sepiolite is commonly associated with calcareous horizons. Sepiolite precipitation is therefore more likely to be triggered when a solution encounters a pH barrier than by the concentration of ions by evaporation. The effect of a pH change on the sepiolite saturation index is much greater than that of the effect on calcite. The marine-influenced high Mg coupled with the Hardie- Eugster model of brine evolution offers an explanation for sepiolite-dominance at the coast, and palygorskite-dominance inland. Coastal areas, unlike continental areas, have Mg>HCO−3 initially, which results in an increasing Mg trend with evaporation during the precipitation of sepiolite according to the Hardie-Eugster scheme. The result is that after sepiolite precipitation is initiated by a geochemical pH-barrier, Mg levels will rise causing the increasing (Mg+Si)/Al ratio to continue to favour sepiolite precipitation. This suggests that once sepiolite has begun to precipitate, the subsequent salinity with its accompanying Mg increase makes substantial palygorskite formation unlikely to follow. The hardpan horizons in heuweltjies commonly grade from a ‘sepiolitic’ petrocalcic in the centre through ‘sepiolitic’/‘petrosepiolitic’ to the petroduric horizon on the edges. Noteworthy sepiolite-related pedofeatures in the calcrete include ‘ooids’ with successive sepiolite (hydrophilic and therefore a precipitational substrate) and micrite/acicular calcite layers in the coatings; and limpid yellow nodules with pseudo-negative uniaxial interference figures. They superficially resemble the spherulites in the fresh termite frass. Their fibrous nature and low birefringence, together with the low Ca, high Mg, Si composition, and molar Mg/Si ratios consistent with sepiolite. The pedogenesis of the hardpans in the heuweltjie is proposed to be as follows: enrichment of cations such as Ca and Mg in the heuweltjie centre caused by termite foraging results in calcite and clay authigenesis in the centre of the heuweltjie, leaving the precipitation of pure silica to occur on the periphery. The decaying organic matter concentrated in the centre of the mound by the termites is sufficient to supply the components for calcite precipitation in the centre of the heuweltjie. Following calcite precipitation, the pH is suitable for sepiolite precipitation. The movement of the Mg-Si enriched water downslope, coupled with the decrease in HCO−3 and increase in Mg2+ due to sepiolite precipitation, allows for the precipitation of the ‘sepiolitic’ zone on the outer side of the calcrete, and extend beyond the calcrete in some heuweltjies. The Namaqualand coastal plain is well positioned for further work on its regolith, particularly because of the mining excavations which provide excellent exposures of well-defined layers of the regolith down to bedrock. Soil formation and termite activity is at least as old as the Last Interglacial, and so more detailed work would further the understanding of the subaerial alteration history in southern Africa, as well as providing better-constrained information on the Namaqualand soils that can be used by land-use management and biosphere studies.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/1182
This item appears in the following collections: