Soil Formation on the Namaqualand Coastal Plain by Michele Louise Francis Submitted to the Department of Soil Science in part fulfilment of the requirements for the degree of Doctor of Philosophy in Agriculture UNIVERSITY OF STELLENBOSCH Supervisors: M. V. Fey and F. Ellis March 2008 c? University of Stellenbosch 2008 i Declaration I, Michele Louise Francis, declare that this dissertation is my own work, except where otherwise stated. It is being submitted for the degree Doctor of Philosophy in Agriculture at the University of Stellenbosch and it has not been submitted before for any degree or examination, at any other university. Signature of Author . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stellenbosch March 2008 ii Abstract The (semi-)arid Namaqualand region on the west coast of South Africa is well- known 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 con- tinues 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, iii 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 remain- ing 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 associ- ated 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 ce- mented, 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 al- kali, 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 dor- bank (petroduric)) to these ?sepiocrete? horizons. Both silica and sepiolite are iv present in the matrix, although the degree to which silica and sepiolite domi- nate 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 dis- tinguish 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 ?m 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 v 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 ?m 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 de- creasing 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 re- mains 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 satura- tion index. The sepiolite saturation index is independent of Mg2+ and H4SiO4 and only increases with increasing pH. Evidence of the pH control on sepio- lite 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 geochem- ical 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. vi The hardpan horizons in heuweltjies commonly grade from a ?sepiolitic? petro- calcic 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 sub- strate) 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 sil- ica 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 precipita- tion, 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. vii Uittreksel Die (semi) ariede Namakwaland streek teen die Weskus van Suid-Afrika is bek- end vir sy pragtige lenteblomme. Vanwee? die dorre geaardheid van die streek het grondnavorsing, wat veral op landboukundig-produktiewe dele van Suid-Afrika gefokus het, hier agterwee gebly. Nietemin, rehabilitasie pogings na die sowat honderd jaar se myn van diamante, gekoppel aan die toenemende belangstelling in ekologiese- en biodiversiteitsnavorsing, het die huidige belangstelling in Na- makwalandse grond as ?n substraat vir plantegroei, aangewakker. Die streek is ook bekend vir sy oorvloed van ?heuweltjies? (lae hope of knoppe in die landskap), wat algemeen aanvaar word dat dit deur die grasdraertermiet Mi- crohodotermes viator, met debatteerbare bydraes deur ander diere, veroorsaak word. Meeste van die vorige werk op heuweltjies het op biogenetiese aspekte soos spasie?ring, oorsprong en ouderdom gefokus. Alhoewel heuweltjies grootliks ?n grondgeaardheid het, is min gepubliseerde werk aangaande die grondvorm- ingsprosesse hiervan gedoen. Die geskiedenis van die sediment afsettings van die Namakwalandse kusgebied is redelik bekend, in teenstelling met die skaarste aan data oor hulle daaropvolgende pedogenese. Omdat die regoliet gedurende som- mige tye van die Neogeen aan die atmosfeer blootgestel was, maak grondvorming ?n belangrike deel uit van die geskiedenis van die sedimente. Die hoofdoel van hierdie proefskrif was daarom om die grondkenmerke van die sedimente op die Namakwalandse kusgebied te ondersoek sodat hulle pedogenese beter verstaan kan word. Die regoliet van die noordelike Namakwalandse kusgebied, dikwels tien of meer meters dik, bestaan uit opeenvolgende lae van mariene oorsprong wat afgesit is gedurende seevlak terugtrekkings in die Laat Tersie?re tydperk. Die opper- vlak grondhorisonte het vanaf eoliese moedermateriaal gevorm. Die relatiewe lae CaCO3 inhoud in die eoliese sande het die pedogenetiese rigting in die af- viii settings bepaal wat gelei het tot ?n meer kleiryke, apedale horison wat sterker struktuurontwikkeling met diepte toon. Soortgelyke horisonte is teenwoordig in paleosols onder ?n duinveld wat tot die Laaste Glasiale Maksimum dateer is. Die apedale horisonte rus f direk op mariene sande via ?n subtiele diskontinu??teit wat beteken dat pedogenese ook in die onderliggende mariene materiale plaasgevind het of op kalkreet, dikwels via ?n kliplyn of geronde rolstene. Dieper in die profiel is daar ?n algemene dog gereelde verandering van die verskillende sedimente?re lae, met die growwer lae wat meer rooi en die fyner sande wat bleek kleure vertoon. In sommige van die lae het die oksidasie van gloukoniet-ryke sedimente ?n oranje kleur veroorsaak. Die Namakwalandse kusvlakte, met ekstensiewe areas van kalkreet ontwikkeling, is moontlik ?n teksboek voorbeeld vir kalkreet ontwikkeling deur anorganiese prosesse. Nietemin toon hierdie kalkrete ook mikroskaal biogenetiese kenmerke. Dit sluit in M stawe, MA stawe en swamfilamente. Dit wil voorkom asof abiotiese alfa-raamwerk oorheers in die ouer kalkreet horisonte terwyl beta-raamwerk ge- assosieer word met die kalkhoudende en kalsiese B horisonte wat bokant kalkreet voorkom. Die skynbare afwesigheid van Mg-ryke kalkreet en dolomiet en die volopheid van sepioliet in die kalkrete van die Namakwalandse kusgebied, dui aan dat hierdie Mg-ryke kleimineraal die hoof Mg-draende fase verteenwoordig. Deformasie (vals antikliene) in die kalkreet blyk die resultaat te wees van die verplasende effek van kalsiet kristallisering. Alhoewel bewyse van swel/krimp gedrag voorkom in die vorm van akkommoderende vlakke, wil dit voorkom dat dit nie so volumetries belangrik is as die van verplasende kalkreet. Verharde liggekleurde horisonte wat lyk soos kalkreet maar wat nie- tot swak kalkhoudend is, wat in konko??dale patrone verbrokkel, wat nie verslemp in suur of alkali, wat pers verkleur in metiel-oranje en in ?n bulk grondmonster sepioliet XSD pieke vertoon, skyn vergelykbaar te wees met die paligorskiet-gesementeerde materiaal (?palikreet?) wat in Spanje en Portugal beskryf is. Dit word hier voor- lopig ?sepiokreet? genoem. Beide silika en sepioliet is teenwoordig in die matriks, alhoewel die mate wat silika en sepioliet domineer selfs binne die dieselfde ho- rison skyn te verskil. Dit is baie moontlik dat beide van hulle bydra tot die struktuureienskappe van die horison. Sepiolitiese horisonte vorm nie ?n diagnos- tiese horison in die World Reference Base, Soil Taxonomy of die Suid-Afrikaanse Taksonomiese Grondklassifikasiesisteem nie. Om in die bestaande grondklassi- fikasiesisteme geakommodeer te word, word voorgestel dat die terme ?sepiolities? ix en hardebank sepioliet (op dieselfde wyse as ?kalsies? en ?petrokalsies? of harde- bank karbonaathorison tans), gebruik word. Sepioliet pieke is meer prominent as die van paligorskiet op die XSD aanwysers van die kleiner as 2 ?m fraksie. Die <0.08 ?m fraksie was die enigste groot- tefraksie waarin paligorskiet bespeur kon word voor asetaat behandeling. Dit is onwaarskynklik dat hierdie veselkleiminerale geerf kon gewees het van of die mariene of die eoliese moedermateriale. Hulle blyk van pedologiese oorsprong te wees. Tendense in die MgO, Al2O3 en SiO2 inhoude toon dat die grond se kleifraksie le? op ?n gemengde lyn tussen sepioliet en mika endlede, met ?n bydrae van smektiet, wat in ooreenstemming is met XSD en TEM resultate. Daar is ?n sterk korrelasie tussen Fe2O3 en TiO2 inhoude, wat toegeskryf kan word aan die alomteenwoordige mika. Daar was geen TEM bewys van veselmineraal afbraak na plaatsilikate; ook nie die evolusie van mika lateraal na ?n veselmineraal toe nie. Daar was geen bevestigende bewyse vir of teen die teenwoordigheid van keroliet in die kleifraksie nie, alhoewel dit wil voorkom dat dit nie ?n dominante fase van die kleifraksie uitmaak nie. Die hipotese word voorgestel dat die deurlatende boonste horisonte in gronde van Namakwaland eintlik ?n vlak kortstondige waterdraer (akwifer) kan vorm, wat die pedogenetiese analoog is met ?n sout-meer omgewing waarin sepioliet tipies sal vorm. Die chemiese evolusie van die grondoplossing kan daarom in dieselfde terme as die geochemiese evolusie van geslote kom pekel(sout)water beskou word. Haliet het onderversadig gebly by alle konsentrasies in die versadigde pasta ek- straksies. By hoe?r konsentrasie het gips versadiging bereik en sulfaat word dan uit die oplossing verwyder. Die aktiwiteit van H4SiO4 het onveranderd gebly by alle vlakke van verdamping. Kalsiet bly naby aan versadiging en is slegs afhank- lik van die HCO?3 aktiwiteit en pH vir die reeks van Cl ? aktiwiteite wat aangetref is. Meeste van die gronde waarin sepioliet aangetref is, het ?n positiewe sepioliet versadigingsindeks getoon. Die sepioliet versadigingsindeks is onafhanklik van Mg2+ en H4SiO4 en neem slegs toe met toename in pH. Bewyse van pH beheer by sepioliet versadiging is dat sepioliet slegs met kalkhoudende horisonte geas- sosieer word. Sepioliet presipitasie sal daarom meer geredelik plaasvind wanneer ?n oplossing ?n pH verandering ondervind as by die konsentrasie van ione deur verdamping. Die mariene-geaffekteerde hoe? Mg gekoppel met die Hardie-Eugster model van pekelwater gedrag, bied ?n verklaring vir die oorheersing van sepioliet teen die kus en paligorskiet verder binneland toe. Kusgebiede, in teenstelling x met kontinentale gebiede, het Mg>HCO?3 aanvanklik, Mg vlakke sal styg wat die (Mg+Si)/Al verhouding sal bevoordeel vir die presipitasie van sepioliet. Dit beteken dat as sepioliet eers begin presipiteer, die gepaardgaande saliniteit, met verhoging in Mg, betekenisvolle paligorskiet vorming onwaarskynlik sal maak. Die hardebank horisonte in heuweltjies algemeen wissel van ?n ?sepiolitiese? petrokalsies naby die middel, deur ?sepiolities?/?petrosepiolities? na ?n petro- duriese (dorbank) horison op die rand. Opmerksaam is dat sepioliet-verwante pe- domaaksels in kalkreet soos olietkorrels insluit. Hulle vorming kan mees waarskyn- lik verklaar word deur die hidrofiliese/hidrofobiese membraanteorie vir stilstaande water te gebruik, gegewe die opeenvolgende sepioliet (hidofilies en daarom ?n pre- sipiterende substraat) en mikriet/naaldvormige kalsiet lae in die bedekkings, en sterk hidrofobiese organiese materiaal in die heuweltjie. Ietwat van ?n raaisel is die deurskynende geel nodules met vals-negatiewe eenassige inteferensiefigure. Laasgenoemde lyk soos sferoliete in vars termiet uitwerpsels. Hulle veselagtige geaardheid en lae dubbelbreking, tesame met die lae Ca en hoe? Mg en Si inhoude en hoe? molare Mg/Si verhouding, is ooreenstemmend met di van sepioliet. Dit word voorgestel dat die vorming van die hardebanke in die heuweltjie as volg plaasgevind het: Verryking met katione soos Ca en Mg in die middel van die heuweltjie deur termiete wat plantmateriaal ingebring het en wat kalsiet en klei deur outigenese in die middel van die heuweltjie veroorsaak het. Die afbrekende organiese materiaal wat aangesamel het in die middel van die heuweltjie was genoeg om die komponente vir kalsiet presipitasie te kon verskaf. Na hierdie or- ganiese materiaal ge??nisieerde kalsiet presipitasie, was die pH geskik vir sepioliet presipetasie. Die beweging van die Mg-Si-verrykte water verder heuwel afwaarts, gekoppel aan die verlaging van die HCO?3 en toename in Mg 2+ a.g.v. sepioliet presipitasie, veroorsaak die presipetasie van die sepiolitiese sone nader aand die buitenste sone van die heuweltjie met die presipitasie van silika op die rand van die heuweltjie. Die Namakwalandse kusgebied is goed geposisioneer vir verdere werk betref- fende die regoliet, veral a.g.v. die mynaktiwiteite waar diep uitgrawings goeie blootstelling van goed-gedefinieerde lae tot op vaste rots blootle?. Grondvorming en termiet-aktiwiteite is tenminste so oud as die Laaste Interglasiale Tydperk. Verdere werk hierop sal die geskiedenis van die sub-blootgestelde grondopper- vlak se veranderings in Suidelike Afrika help verklaar. Dit kan ook help met ?n beter begrip van die dikwels verstrengelde informasie wat aangaande Na- xi makwaland bestaan, asook help met landgebruiksbestuur en biosfeer studies. Bykomend sal mikromorfologiese studies wat die verwantskap tussen kalsiet, se- pioliet/paligorskiet en silika fases in meer heuweltjies bestudeer, die algemene model wat in hierdie proefskrif gestel word, verder verfyn. xii Acknowledgements This thesis spanned two continents and six years from its beginnings as a Master?s degree, and so the list of people who have provided support directly and indirectly is long. I particularly need to thank my family, two members of which have known life no other way, for their patience with it. Thanks are due to De Beers Namaqualand Mines, in particular the Environmen- tal Department for facilitating this thesis project. People who went out of their way to help me during the sampling stages, and are owed a special thanks, are: Paul Kruger for his enthusiastic help in the field; Charlaine Kruger for making this project possible; Jenny van der Westhuizen for co-ordinating all aspects of the Kleinzee visits; Russel Nicol for his assistance with my visits to the Ko- ingnaas Complex; John Fulton for producing the map of sample locations; the Geology Department for permission to publish the results and their help with the stratigraphy, in particular Andy Grills, Neale Baartjies and Marthinus Prinsloo for taking the time to assist me; John Pether for agreeing to spend a day in the field with me at very short notice; and the late Hendrik Prinsloo for the many thought-provoking discussions and his assistance in the field. Many people from outside the Soil Science department were involved in the data collection. Electron Microscope images were taken by Mr N. Steenkamp and Esme? Spicer and myself at the University of Stellenbosch Geology Depart- ment, and Belinda White and Vijay Bandu at the University of KwaZulu-Natal (Pietermaritzburg campus). XRF analyses were done by Esme? Spicer at the Uni- versity of Stellenbosch Geology Department. The thin sections were prepared by Montserrat Antu?nez Pujol from the University of Lleida, Spain, and David Wil- son from the University of Cape Town. Thanks are also due to Remy Bucher at the iThemba Labs Materials Research Group, Somerset West, for running many, many XRD samples. xiii Thanks are also due to the staff in the Soil Science Department at the University of Stellenboch: Annatjie French for the Admin, the Laboratory Staff Judy Smith, Kamiela Crombie and Herschel Achilles for their help with some of the analyses; Tanya Medinski for the carbon and nitrogen analyses; Matt Gordon for the cation and anion analyses, and his patience with the samples and my repeated requests for sometimes large amounts of chemicals; and Herschel Achilles and Kenneth Davidse Jr. for moving the sample crates around, particularly given my refusal to down-size the sample volumes. I would like to particularly thank Richard O?Brien for his generous assistance with the XRD, his advice on all aspects of the clay extraction and identification, and for kindly running some of my samples when I could not. He was the first one to identify almost mono-minerallic sepiolite in the clay fraction. Friends who helped freely with some of the more labor-intensive chemical anal- yses are Cathy Dowding, Ailsa Hardie, Arne Henrichsen, Tanya Medinski, and Jason Salzwedel. Students who helped during their vacation work in the Soil Sci- ence department are Nicolette van der Merwe, who did acetate pre-treatments and clay separations, and Thami Vilakhazi, who did particle size analysis, clay separations and ran some of the samples on the XRD at the University of Stel- lenbosch Geology Department. Thanks are also due to Willem de Clercq for constructing the resin impregnator in the University of Stellenbosch Soil Science department; Jeffrey Hughes for taking the time to give a lot of advice about clay minerals; John Compton for advice on pCO2 conversions; David L. Parkhurst for answering queries about setting up the PHREEQC runs; Georges Stoops for remotely diagnosing a problem with the microscope set-up; Andrei Rozanov for helping to identify phytoliths; Anthony Mills and Antoni Milewski for their assistance with the question of oxalate in the vegetation, and Tim Francis for his help with soil forming processes in sediments on the East Coast of South Africa. Much credit is due to Rosa Poch for her generous assistance with the sampling and thin section descriptions; to Jan Lambrechts and Freddie Ellis for locating the ?sepiocretes? and heuweltjies, the sampling trips, and many discussions; to Martin Fey for his seemingly never-ending patience with questions of an editorial nature. I am indebted to Freddie Ellis for translating the Abstract. xiv I thank my supervisors, Professor Martin Fey and Dr Freddie Ellis, for the op- portunity they gave me to work in the Soil Science Department and for their continuing enthusiasm, a lot of advice and guidance, and most of all their faith in my ability to finish. Financial support for this project was received from De Beers Namaqualand Mines Limited, the National Research Foundation (Grant number 2047381), and Eskom?s Tertiary Support Program (TESP). xv I dedicate this thesis to Arne Henrichsen. Without his help with everything else, this thesis would not have been written. xvi Contents Introduction 1 1 Silica- and sepiolite- cemented duricrusts on the west coast of South Africa 4 1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.1 Silica-cemented horizons . . . . . . . . . . . . . . . . . . . 5 1.1.2 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . 8 1.2.1 Site description . . . . . . . . . . . . . . . . . . . . . . . . 8 Coastal plain; BMC: profiles SK11 and KV196T . . . . . . 8 Knersvlakte: profile KNERS . . . . . . . . . . . . . . . . . 13 1.2.2 Laboratory methods . . . . . . . . . . . . . . . . . . . . . 14 1.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 15 1.3.1 Slaking behaviour . . . . . . . . . . . . . . . . . . . . . . . 15 Dorbank (petroduric) . . . . . . . . . . . . . . . . . . . . . 15 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 18 1.3.2 Mineralogy . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Dorbank (petroduric) . . . . . . . . . . . . . . . . . . . . . 19 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 22 xvii 1.3.3 Micromorphology . . . . . . . . . . . . . . . . . . . . . . . 24 Dorbank (petroduric) . . . . . . . . . . . . . . . . . . . . . 24 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 27 1.3.4 Genesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Dorbank (petroduric) . . . . . . . . . . . . . . . . . . . . . 32 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 35 1.3.5 Classification . . . . . . . . . . . . . . . . . . . . . . . . . 37 Dorbank (petroduric) . . . . . . . . . . . . . . . . . . . . . 37 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 38 1.4 Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . 40 1.4.1 Dorbank (petroduric) . . . . . . . . . . . . . . . . . . . . . 40 1.4.2 ?Sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . . . . 41 2 Soil formation in sediments of the Namaqualand coastal plain 43 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 2.2 Site description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 2.2.1 Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Moisture . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Rainfall . . . . . . . . . . . . . . . . . . . . . . . . 46 Evaporation . . . . . . . . . . . . . . . . . . . . . . 47 Topography . . . . . . . . . . . . . . . . . . . . . . 47 Microrelief . . . . . . . . . . . . . . . . . . . . . . . 47 Soil Permeability . . . . . . . . . . . . . . . . . . . 49 2.2.2 Regolith . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 xviii 2.2.3 Bedrock geology and morphology . . . . . . . . . . . . . . 50 2.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 2.4 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 52 2.4.1 A horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 2.4.2 B horizons . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Age . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.4.3 Eluvial horizons . . . . . . . . . . . . . . . . . . . . . . . . 63 2.4.4 Cemented horizons . . . . . . . . . . . . . . . . . . . . . . 65 Calcite-cemented horizons . . . . . . . . . . . . . . . . . . 65 Silica-cemented horizons . . . . . . . . . . . . . . . . . . . 67 Sepiocrete and sil-sepiocrete intergrades . . . . . . . . . . 67 2.4.5 Marine sediments . . . . . . . . . . . . . . . . . . . . . . . 68 2.4.6 Biological activity . . . . . . . . . . . . . . . . . . . . . . . 70 Tunnels . . . . . . . . . . . . . . . . . . . . . . . . 71 Roots . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Termites . . . . . . . . . . . . . . . . . . . . . . . . 71 2.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 3 ?Pseudo-anticlines? and biogenic features in sepiolite-containing calcrete in Namaqualand 76 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 3.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . 78 3.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 79 3.3.1 Composition . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Carbonates . . . . . . . . . . . . . . . . . . . . . . . . . . 79 xix Clays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 3.3.2 Morphology and micromorphology . . . . . . . . . . . . . 82 ?Pseudo-anticlines? . . . . . . . . . . . . . . . . . . . . . . 82 Biogenic fabric . . . . . . . . . . . . . . . . . . . . . . . . 85 3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 4 Clay mineral occurrences in Aridisols on the west coast of South Africa (Namaqualand) 90 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . 92 4.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 94 4.3.1 Effect of milling and acetate pretreatments on clay fraction 94 4.3.2 Clay mineral variations in soil profiles . . . . . . . . . . . . 95 4.3.3 Kerolite . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4.3.4 Chemical composition of clay fraction . . . . . . . . . . . . 106 4.3.5 Sepiolite/palygorskite modes of occurrence . . . . . . . . . 109 4.3.6 Palygorskite/sepiolite transformation . . . . . . . . . . . . 109 4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 5 Brine evolution and geochemical barriers: mineral formation in Namaqualand soils 115 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 5.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . 119 5.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 122 5.3.1 Chloride as a proxy for evaporative concentration . . . . . 125 5.3.2 Marine-influenced initial ratios and subsequent evapora- tion trends . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 xx 5.3.3 pH control of sepiolite equilibrium . . . . . . . . . . . . . . 134 5.3.4 Coastal sepiolite and inland palygorskite in Namaqualand 137 5.3.5 Summary of mineral genesis on the Namaqualand coastal plain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Halite and gypsum . . . . . . . . . . . . . . . . . . . . . . 140 Calcite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Sepiolite-palygorskite . . . . . . . . . . . . . . . . . . . . . 140 Silica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 pH variation and its potential to affect the genesis of calcic, sepiolitic and duric soils . . . . . . . . . . . . . . 142 5.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 6 Micromorphology, mineralogy and genesis of soils associated with a Namaqualand ?heuweltjie? 145 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 6.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . 148 6.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 151 6.3.1 Profile descriptions . . . . . . . . . . . . . . . . . . . . . . 151 6.3.2 Clay mineralogy . . . . . . . . . . . . . . . . . . . . . . . . 153 Calcrete in centre of heuweltjie: samples 2A, 2B, 2C . . . . 153 ?Sepiolitic? pedocutanic (argic) 7.5 m from centre of heuweltjie: sample 2D . . . . . . . . . . . . . . . . . . . . . . 154 Dorbank (petroduric) at edge of heuweltjie . . . . . . . . . 154 6.3.3 Micromorphology of excrement and exit towers: Stellen- bosch heuweltjie . . . . . . . . . . . . . . . . . . . . . . . . 156 6.3.4 Micromorphology of the heuweltjie at Papendorp . . . . . 159 Biological activity . . . . . . . . . . . . . . . . . . . . . . . 182 xxi Horizon over calcrete in the heuweltjie centre . . . . . . . . 182 Calcrete in the heuweltjie centre . . . . . . . . . . . . . . . 183 Calcareous ooids and peloids . . . . . . . . . . . . . . . . . 183 Limpid nodules . . . . . . . . . . . . . . . . . . . . . . . . 186 Pseudo-negative uniaxial ?spherulites? . . . . . . . . 186 Possible Other Morphological Forms . . . . . . . . . 187 Composition . . . . . . . . . . . . . . . . . . . . . . 188 ?Sepiolitic? pedocutanic (argic) midway to edge of heuweltjie (sample 2D) . . . . . . . . . . . . . . . . . . . . . 190 Dorbank (petroduric) at heuweltjie edge; sample 1.3 . . . . 191 6.3.5 Heuweltjie formation . . . . . . . . . . . . . . . . . . . . . 192 Organic matter as a source of ions for calcrete formation . 192 Source of ions for sepiolite/palygorskite formation . . . . . 194 Leaching and redistribution of ions within the heuweltjie: zoning of sepiolite and dorbank (petroduric) . . . 194 6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 Summary of findings and recommendations for further work 198 References 203 A Photographs of soil profiles: BMC and KNC 231 xxii List of Figures 1.1 Location of sample areas . . . . . . . . . . . . . . . . . . . . . . . 9 1.2 Location of profiles KV196T, SK11 and their farm names within the Buffels Marine Complex (BMC) mining area . . . . . . . . . 12 1.3 Photograph: dorbank (petroduric) cutan on ?sepiocrete? . . . . . . 17 1.4 Photograph: ?sepiocrete? . . . . . . . . . . . . . . . . . . . . . . . 18 1.5 XRD traces of clay fractions from profile KV196T . . . . . . . . 20 1.6 XRD traces of clay fractions from the Knersvlakte profile Kners 21 1.7 TEM image: amorphous material coating clay grain . . . . . . . . 22 1.8 XRD traces of clay fractions from profile SK11 . . . . . . . . . . 23 1.9 Photomicrographs: typical views of silica in dorbank (petroduric) horizons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1.10 Optical and SEM images of neoformed and illuviated smectite and silica in dorbank (petroduric) fragments . . . . . . . . . . . . . . 28 1.11 Optical and SEM images of granular microstructure in ?sepiocrete? 29 1.12 SEM images: HCl etching of ?sepiocrete? . . . . . . . . . . . . . . 31 1.13 SEM images: HCl etching of ?sepiocrete? continued . . . . . . . . 32 1.14 Optical and SEM images of silica in ?sepiocrete? . . . . . . . . . . 33 1.15 SEM: broken sponge spicule in dorbank (petroduric) fragments. . 35 2.1 Location of field area. . . . . . . . . . . . . . . . . . . . . . . . . . 45 xxiii 2.2 Location of sample areas . . . . . . . . . . . . . . . . . . . . . . . 48 2.3 Generalized regolith profile . . . . . . . . . . . . . . . . . . . . . . 53 2.4 Archaeological site buried by regic sands . . . . . . . . . . . . . . 60 2.5 Dune plume north of Swartlintjies River: interpretation. . . . . . 62 2.6 Photograph: Possible vertical termite burrows that have become conduits for preferential flow. . . . . . . . . . . . . . . . . . . . . 66 2.7 Photograph: probable hematite ?halos? in lower shoreface sands . . 69 2.8 Photograph: termite nest, 1.7 m deep below calcrete. . . . . . . . 72 2.9 Photograph: strongly-cemented, calcareous hive 2 m deep within dorbank. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 2.10 Photograph: Possible vertical termite burrows now conduits for preferential flow in lower shoreface sands. . . . . . . . . . . . . . . 73 3.1 XRD traces: Nature of the carbonates in calcrete . . . . . . . . . 80 3.2 XRD traces: clay fraction mineralogy . . . . . . . . . . . . . . . . 81 3.3 XRD trace: sepiolite cutan on calcrete . . . . . . . . . . . . . . . 83 3.4 Photomicrographs: displacive, replacive and needle fibre calcite; clay textures in calcrete . . . . . . . . . . . . . . . . . . . . . . . 84 3.5 SEM images: Biogenic calcite in calcite nodules from a neocar- bonate B horizon overlying calcrete . . . . . . . . . . . . . . . . . 87 3.6 SEM images: HCl-etched filaments and fibrous clay minerals . . . 88 4.1 XRD traces of clay fraction from profile 6868 . . . . . . . . . . . 97 4.2 XRD traces of clay fraction from profile AK61H . . . . . . . . . 100 4.3 XRD traces of clay fraction from profile AK1/TRIG . . . . . . . 101 4.4 XRD traces of clay fraction from profile LKC1-5 . . . . . . . . . 102 4.5 XRD traces of clay fraction from profile OBT . . . . . . . . . . . 103 xxiv 4.6 XRD traces of clay fraction from profile SL4-1 . . . . . . . . . . . 104 4.7 TEM images of clay fractions . . . . . . . . . . . . . . . . . . . . 105 4.8 Trends in total element composition of clay fractions: MgO+Al2O3 vs SiO2; Al2O3 vs K2O; Fe2O3 vs TiO2 . . . . . . . . . . . . . . . 108 4.9 SEM images of sepiolite/palygorskite occurrences in calcareous nodules from 6869/iv . . . . . . . . . . . . . . . . . . . . . . . . . 110 4.10 TEM images of sepiolite/palygorskite fibres and fibre bundles in 6869/iv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 5.1 Variation of ion activities with Cl? as a proxy for evaporative concentration in saturated paste extracts . . . . . . . . . . . . . . 123 5.2 Variation of mineral saturation indices with Cl? as a proxy for evaporative concentration in saturated paste extracts . . . . . . . 126 5.3 Variation of mineral saturation indices (SI) with pH in saturated paste extracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 5.4 Variation of mineral saturation indices (SI) with activity of com- ponent ion in saturated paste extracts . . . . . . . . . . . . . . . 129 5.5 Variation of mineral saturation indices (SI) with activity of com- ponent ion in saturated paste extracts (continued) . . . . . . . . . 136 5.6 Variations in saturation indices with pCO2 . . . . . . . . . . . . . 138 6.1 Location of Papendorp heuweltjie . . . . . . . . . . . . . . . . . . 148 6.2 Photographs of sample locations within heuweltjie . . . . . . . . . 150 6.3 XRD traces of white palygorskite/sepiolite cutans and associated calcrete from the centre of heuweltjies from Worcester and Oudt- shoorn. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 6.4 XRD traces of clay fraction from Papendorp heuweltjie . . . . . . 155 6.5 TEM image of ?sepiolitic? pedocutanic (argic) sample 2D and dor- bank (petroduric) sample 1.3. . . . . . . . . . . . . . . . . . . . . 156 xxv 6.6 Photomicrographs of termite excrement from an active heuweltjie on the south side of Papegaaiberg (Stellenbosch) . . . . . . . . . . 157 6.7 Photomicrographs of termite exit tunnels from active heuweltjie on the south side of Papegaaiberg (Stellenbosch) . . . . . . . . . . 158 6.8 Photomicrographs from the centre of the heuweltjie: neocutanic B (sample 2.2) and calcrete (2A). . . . . . . . . . . . . . . . . . . 166 6.9 Continued - photomicrographs from the centre of the heuweltjie: calcrete (2A). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 6.10 Continued - photomicrographs from the centre of the heuweltjie: calcrete (2A). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 6.11 HCl etching of peloids and limpid nodule with pseudo-negative uniaxial interference figures in sample 2A, calcrete in centre of heuweltjie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 6.12 Photomicrographs and corresponding SEM (backscatter electron) images from calcrete (2A) in centre of heuweltjie . . . . . . . . . . 170 6.13 SEM (backscatter electron) images of fragments and thin sections from calcrete (2A) in centre of heuweltjie . . . . . . . . . . . . . . 171 6.14 Photomicrographs from 2 m away from the centre of the heuweltjie: calcrete (2B). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 6.15 Continued - photomicrographs from 2 m away from the centre of the heuweltjie: calcrete (2B). . . . . . . . . . . . . . . . . . . . . . 173 6.16 Continued - photomicrographs from 2 m away from the centre of the heuweltjie: calcrete (2B). . . . . . . . . . . . . . . . . . . . . . 174 6.17 Continued - photomicrographs from 2 m away from the centre of the heuweltjie: calcrete (2B). . . . . . . . . . . . . . . . . . . . . . 175 6.18 Photomicrographs from 3.2 m away from the centre of the heuweltjie: calcrete (2C). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 6.19 Continued - photomicrographs from 3.2 m away from the centre of the heuweltjie: calcrete (2C). . . . . . . . . . . . . . . . . . . . 177 xxvi 6.20 Photomicrographs from 7.5 m away from the centre of the heuweltjie: non-calcareous ?sepiolitic? pedocutanic (argic) (2D). . . . . . . . . 178 6.21 Continued - photomicrographs from 7.5 m away from the centre of the heuweltjie: non-calcareous ?sepiolitic? pedocutanic (argic) (2D). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 6.22 Continued - photomicrographs from 7.5 m away from the centre of the heuweltjie: non-calcareous ?sepiolitic? pedocutanic (argic) (2D). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 6.23 Photomicrographs of dorbank (petroduric, sample 1.3) at the edge of the heuweltjie (8 m from 2D) . . . . . . . . . . . . . . . . . . . 181 6.24 Inferred water movement through the heuweltjie. . . . . . . . . . 195 Plate: Profile 6869 . . . . . . . . . . . . . . . . . . . . . . . . . 232 Plate: Profile AK1 . . . . . . . . . . . . . . . . . . . . . . . . . 233 Plate: Profile AK61H . . . . . . . . . . . . . . . . . . . . . . . 234 Plate: Profile DL88 . . . . . . . . . . . . . . . . . . . . . . . . 235 Plate: Profile KV196T . . . . . . . . . . . . . . . . . . . . . . 236 Plate: Profile KV220PRN . . . . . . . . . . . . . . . . . . . . 237 Plate: Profile LKC1-5 . . . . . . . . . . . . . . . . . . . . . . . 238 Plate: Profile OBT . . . . . . . . . . . . . . . . . . . . . . . . 239 Plate: Profile SK11 . . . . . . . . . . . . . . . . . . . . . . . . 240 Plate: Profile SL4-1 . . . . . . . . . . . . . . . . . . . . . . . . 241 Plate: Profile SN30 . . . . . . . . . . . . . . . . . . . . . . . . 242 Plate: Profile SNT60 . . . . . . . . . . . . . . . . . . . . . . . 243 Plate: Profile SPNT . . . . . . . . . . . . . . . . . . . . . . . . 244 Plate: Profile TP231L . . . . . . . . . . . . . . . . . . . . . . . 245 Plate: Profile TP266Q . . . . . . . . . . . . . . . . . . . . . . 246 xxvii List of Tables 1.1 Profile descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.2 Thin section descriptions . . . . . . . . . . . . . . . . . . . . . . . 25 1.3 Profile chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 1.4 XRF analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 2.1 Profile descriptions. . . . . . . . . . . . . . . . . . . . . . . . . . . 54 2.2 Redness rating: soil profile comparison . . . . . . . . . . . . . . . 64 4.1 Clay mineral variation down profiles. . . . . . . . . . . . . . . . . 96 5.1 Composition of saturated paste extracts . . . . . . . . . . . . . . 124 5.2 Ion-concentration trends in maritime and continental water systems132 6.1 Profile descriptions from Papendorp heuweltjie . . . . . . . . . . . 152 6.2 Thin section descriptions of samples from Table 6.1. Abbrevia- tions: c/f: coarse/fine; diam.: diameter; incl.: including; max.: maximum; occ.: occasional; OIL: oblique incident light; PPL: plane polarised light; XPL: cross polarised light. . . . . . . . . . . 160 1 Introduction This study was conceived in order to examine in greater detail the 100 km stretch of coast that was mapped by F. Ellis and J. J. N. Lambrechts for the De Beers Namaqualand Mines rehabilitation program, as characterisation and classifica- tion of all horizons within the regolith before disturbance assists in understand- ing the new soil forming processes that have begun on the surfaces of the mine dumps. The study area was subsequently expanded to encompass additional heuweltjies and ?sepiocretes?. The Namaqualand region has been receiving an increasing amount of attention since the commencement of this thesis. This has culminated in a Special Issue of the Journal of Arid Environments on Sustainable Land Use in Namaqualand (September 2007, Volume 70, Issue 4, Pages 561-846), and although the body of literature on Namaqualand is by no means limited to this volume, the reader unfamiliar with Namaqualand is urged to consult it as an introduction to the present state of knowledge. Many of the premises (par- ticularly in terms of the influence of climate on floral kingdoms and ecosystem functioning) outlined in the Special Issue were presupposed in this thesis. The thesis is presented as a collection of papers that are intended to be submitted for publication. Each Chapter has therefore been formulated that it can be read as a ?stand-alone? paper, with reference to the work in previous Chapters. This means, however, that there is some overlap in the ?Materials and Methods? section of each chapter. The aims of each Chapter is outlined below. During the initial field work light-coloured horizons were encountered that re- sembled calcrete but were non-calcareous, broke with a conchoidal fracture and were resistant to slaking in both acid and alkali. Further examination showed them to contain a large amount of sepiolite, and they were tentatively named ?sepiocretes?. The aim of Chapter 1 (Silica- and sepiolite-cemented duricrusts on the west coast of South Africa), was to derive a name for these horizons based 2 on their cementing agent. Since amorphous silica could also have played a role in cementing the ?sepiocrete? horizons, Chapter 1 compares the two known types of silica-cemented horizons encountered in Namaqualand to these ?sepiocrete? horizons. The aim of Chapter 2 (Soil formation in sediments of the Namaqualand coastal plain), was to describe and characterise the general soil forming features, both current soils and paleosols, in the regolith down to bedrock. The mining exca- vations tens of metres deep down to bedrock have provided an opportunity to examine soil formation in a regolith that has been subaerially exposed in some parts for much of the Neogene. The aims of Chapter 3 (?Pseudo-anticlines? and biogenic features in sepiolite- containing calcrete in Namaqualand) were to describe the micromorphology, mineralogy and genesis of calcretes on the Namaqualand coastal plain, including the development of pseudo-anticline structures. Biogenic processes, however, are increasingly being seen as important to calcrete genesis, and so a particular aim was to establish whether there was a biogenic contribution to calcrete formation, and to distinguish the fibrous clay minerals sepiolite and/or palygorskite from similar biogenic needle fibre calcite. Chapter 4 (Clay mineral occurrences in Aridisols on the west coast of South Africa (Namaqualand)) focuses on the clay mineralogy of the soils. The aims of this Chapter were to evaluate the effect of pretreatments on the sepiolite and palygorskite peaks, and particularly to concentrate kerolite if present; ex- amine the clay mineral associations throughout the profile including the marine parent materials to establish which, if any, were inherited phases; and to ex- amine the morphology and modes of occurrence of the fibrous clay minerals in Namaqualand soils, especially for evidence of disintegration or transformation into sheet silicates. Chapter 5 (Brine evolution and geochemical barriers: mineral formation in Namaqualand soils) is based on the hypothesis that the permeable upper horizons in Namaqualand soils constitute a shallow ephemeral aquifer, which can be considered as a pe- dogenic 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 aim was to apply the principles of brine evolution, particularly the Hardie-Eugster model, to sepiolite genesis in arid soils of the Namaqualand 3 coastal plain. The Hardie-Eugster model is based on the effect mineral precip- itation has on brine evolution, which is determined by the initial ratio of the component ions in the solution. It was conceived as a means of understanding the evolution of closed-basin brines, and has since been applied to a diverse range of environments, from saline soil seeps to coastal pans and calcrete genesis. Chapter 6 (Micromorphology, mineralogy and genesis of soils associated with a Namaqualand Heuweltjie) is the first detailed micromorphological study through one of the heuweltjie (Mima-like) mounds that are ubiquitous in the region. The aim was to use soil micromorphology and the principles of brine evolution from Chapter 5 to help explain soil formation within the heuweltjie, particulary the re- lationship between the (petro)calcic, ?(petro)sepiolitic? and petroduric horizons. 4 Chapter 1 Silica- and sepiolite- cemented duricrusts on the west coast of South Africa 1.1 Introduction The arid Namaqualand region on the west coast of South Africa has been recog- nised for its exceptionally high species diversity and endemism (Cowling and Hilton-Taylor, 1999; Cowling et al., 1999), and the region has been receiving increasing attention as a result. It is the focus of a Special Issue of the Journal of Arid Environments (2007). A summary of the landscape and climate and their effect on vegetation diversity is given by Desmet (2007). Soils knowledge, however, has lagged behind the biodiversity interest. With the exception of a study of the clay mineralogy of Namaqualand soils (Singer et al., 1995), soils information was generally limited to Land Type Surveys (Land Type Survey Staff, 1987), unpublished irrigation reports, work by Ellis (1988), and an outline by Watkeys (1999). Francis et al. (2007) summarised the key characteristics of Namaqualand soils. They noted the presence of three types of hardpans: reddish brown, silica- cemented horizons which slake in alkali; calcium carbonate-cemented horizons which slake in acid; and a cemented horizon they referred to as ?sepiocrete?, an indurated light-coloured horizon that resembles calcrete but is non- to mildly 5 calcareous, breaks with a conchoidal fracture, is resistant to slaking in both acid and alkali and reacts positively to methyl orange in the field (Mifsud et al., 1979). XRD analysis of the bulk soil shows a sepiolite peak. They are commonly asso- ciated with calcrete in ?heuweltjies? (hillocks; circular features having a different vegetation pattern and raised (1 to 2.5 m) surface, 10 to 20 m diameter, usually attributed to termites). The material is similar to the palygorskite-cemented material (?palycrete?) described by Rodas et al. (1994) and Stahr et al. (2000), and Sauer et al. (submitted). Sepiolite and palygorskite are often reported from arid region soils and is commonly found in calcretes (Vanden Heuvel, 1964; Singer and Norrish, 1974; Yaalon and Wieder, 1976; Elprince et al., 1979; Hay and Wig- gins, 1980; Watts, 1980; Singer and Galan, 1984; Singer, 1989; Blank and Fos- berg, 1991; Monger and Daugherty, 1991; Verrecchia and Le Coustumer, 1996; Singer, 2002; Neaman and Singer, 2004; Owliaie et al., 2006), but there has been no recorded cementation of soils by sepiolite. This Chapter describes these ?sepiocretes? and their genesis, and attempts to derive a name for these horizons based on their cementing agent. Since Francis et al. (2007) and F. Netterberg (pers. comm., 2003) suggested that amorphous silica could also play a role in cementing the ?sepiocrete? horizons, this Chapter describes and compares the two types of silica-cemented horizons encountered in Namaqualand to these ?sepiocrete? horizons. 1.1.1 Silica-cemented horizons Two morphologically distinct silica-cemented horizons occur in southern Africa: ?dorbank? in the South African Classification System (Soil Classification Work- ing Group, 1991), equivalent to the ?duripan? of Soil Taxonomy (Soil Survey Staff, 1999) and ?petroduric? of the World Reference Base (1998), and ?silcretes?. Although both are the result of silica cementation, their general morphology and genesis in southern Africa differ (Ellis and Lambrechts, 1994). There is much less information available on the genesis and morphology of ?dorbank? (Ellis and Schloms, 1982; Ellis, 1988; Ellis and Lambrechts, 1994) than ?silcrete? (such as Summerfield, 1983a; Partridge and Maud, 1987). Two morphological types of dorbank (petroduric horizons) occur: a massive type, and one with a platy or laminated structure (Ellis and Lambrechts, 1994). The layers in the platy type are from a few millimetres to 0.3 m thick, and may show vertical cracks. Of- 6 ten coatings or thin layers of calcium carbonate occur between the plates. The massive dorbank varies in thickness from 1 to 5 m or more. In regions where both silcrete and dorbank are found, silcrete occurs as cappings on the oldest landscape positions, with dorbank on lower lying erosion surfaces (Ellis and Lambrechts, 1994). Ellis and Lambrechts (1994) found that the high correlation (r2 = 63) between dorbank distribution and the aridity index may indicate that dorbank developed under present arid climates, or similar paleo- climates, in contrast to silcrete (r2 = 19). Dorbank is strikingly similar to the ?red-brown hardpans? of Australia described by both Litchfield and Mabbutt (1962) and Chartres (1985), both of whom also draw a distinction between these and silcrete. Unlike the dorbanks, South African silcretes are characteristically pale in colour, which suggests silica accumulation and cementation under hydromorphic condi- tions in sandy materials (Ellis and Lambrechts, 1994). They are usually 1 to 2 m thick, massive, and very hard. They are commonly associated with Ter- tiary erosion surfaces and occur as remnant cappings (flat topped mesas) in the landscape (Partridge and Maud, 1987), where they are referred to as weathering profile silcretes by Summerfield (1983a). These are usually associated with pale coloured deeply weathered saprolite (?pallid zone?) up to 20 m thick, composed primarily of kaolinite and illite with variable amounts of detrital quartz. Silcrete also occurs as non-weathering profile silcretes (Summerfield, 1983a), such as the Cenozoic Kalahari Beds. These have formed through silicification of various host sediments, especially calcretes and playa sediments (Summerfield, 1983a). The author observed similar silicification of lower shoreface facies marine sands (Pether et al., 2000) near the base of deep regolith in mining excavations (De Beers Namaqualand Mines) on the Namaqualand coastal plain. 1.1.2 ?Sepiocrete? Similar materials have been encountered in southern Africa before, but they have not been classified and no systematic study has been undertaken. Palygorskite- rich hardpans that macroscopically resembled calcretes, but were generally non- calcareous, and did not slake in 30% NaOH, were documented from Botswana (Runtu) by Netterberg (1969), who noted that the cementing medium was prob- 7 ably chalcedonic and opaline silica. There are also materials which look like calcretes but contain no carbonate at the Elandsfontein (near Hopefield) fossil site (Netterbeg, pers. comm., 2003), a white disintegrating material associated with dorbank in a heuweltjie near Ebenhaeser (31o35?50??S 18o16?45??E, road to Papendorp), with minor calcite and very abundant sepiolite (1.25 nm peak on a bulk sample XRD trace) (Ellis, unpublished data), and is present in some heuweltjies around the Worcester area of the Western Cape and Oudtshoorn in the Eastern Cape. Francis et al. (2007) suggested interlocking fibrous sepiolite crystals could pro- vide an explanation for the cementation of the ?sepiocrete horizons. Brecciated sepiolite lutites in the Esquivias deposit (Madrid Basin) are ?cemented with sepiolite? (Pozo and Casas, 1999, p. 400). The addition of 10% sepiolite to ce- ment increased both the compressive and bending strengths of the mortar (Kavas et al., 2004). Bain (1971) noted that massive authigenic sepiolite formed in playa lakes dries to an extremely tough and coherent material. Mosaddeghi et al. (2006) evaluated the tensile strength of artificial mixtures of sand?palygorskite?calcite, and found that the effect of 5% palygorskite was approximately similar to that of 30% CaCO3. They suggested that this might be due to the fibrous structure and the higher surface area of palygorskite as compared to CaCO3, and that the fi- brous units could act as binding bridges around and between the sand particles. This strong and non-friable nature is not restricted to sepiolite/palygorskite: non-slaking kaolin (flint clay) that breaks with a conchoidal fracture was de- scribed by Keller (1982, p. 30) which ?when immersed in water remains intact, firm ?rock-like? indefinitely . . . it withstands weathering for years on the outcrop or in a stockpile.? Keller (1982) attributed its resistance to slaking to tightly interlocking, randomly oriented, compact packages and sheaves of kaolin crystals exhibiting a mutual-boundary texture. Francis et al. (2007) and Netterberg (pers. comm., 2003) suggested that amor- phous silica could also play a role in cementing the ?sepiocrete? horizons. Inter- estingly, Keller (1982) noted that although siliceous cement may have increased certain hydrothermal kaolins? resistance to slaking, no independent field or lab- oratory evidence had demonstrated that free silica served as a cement in typical flint clay of sedimentary origin. 8 Massive sepiolite is sometimes called ?meerschaum? (from German meaning ?sea- froth? (Hay et al., 1995)), especially in the gem trade when it occurs in compact masses or nodules (Sariiz and Isik, 1995). In Eskis?ehir, Turkey, high-quality sepiolite nodules were formed by alteration of magnesite nodules at shallow burial under alkaline conditions (Ece and C?oban, 1994). Other major sepiolite deposits include the Hekimhan region of Turkey (Yalc?in and Bozkaya, 1995), Kenya (Hay et al., 1995), Somalia (Singer et al., 1998) and Spain, all of which formed in a paleolacustrine environment. There is abundant published information on these and other sepiolite-palygorskite occurrences (Weaver and Beck, 1977; Singer and Galan, 1984; Jones and Galan, 1988; Singer et al., 1992; Torres-Ru??z et al., 1994; Akbulut and Kadir, 2003; Jamoussi et al., 2003; Zaaboub et al., 2005, among others). 1.2 Materials and methods One typical dorbank (petroduric horizon) and two ?sepiocrete? profiles were used to examine the genesis and cementing agents of these duricrust samples. Sam- pling was done in the Knersvlakte (Kners, Figure 1.1), and in the Buffels Marine Complex (BMC) (De Beers Consolidated Mines Ltd - Namaqualand Mines) in the Kleinzee-Port Nolloth area in the northern part of the (semi-)arid Namaqualand coastal plain (KV196T and SK11). 1.2.1 Site description Coastal plain; BMC: profiles SK11 and KV196T On the Namaqualand coast, the prevailing winds are very strong southerly (oc- casionally gale force), with occasional berg (east) wind conditions, especially during winter. Temperatures increase markedly during berg wind conditions, which may persist for longer than a week. As a result, the average maximum temperature from February to October (late summer to spring) is over 30o C, with April and October experiencing temperatures of 38o C. In contrast, the average maximum temperature from November to January is 25o C (late spring to summer). Minimum temperatures from March to October are around 6o C, 9 Figure 1.1: Location of Buffels Marine Complex (BMC) mining area, KV196T, SK11, and the Knersvlakte sampling site (Kners). The escarpment (arrowed, dark lines) forms the eastern edge of the coastal plain. Modified from Ellis (1988). with July dropping to 3o C. From November to April the minimum is around 10o C (D.B.C.M., 2000, data collected at Kleinzee, July 1995 to March 1997). The Namaqualand coastal plain receives less than 150 mm winter rainfall an- nually, with the figure increasing southwards. Crucially for both plant growth and water movement through the soil, the climate is characterised by highly reliable rainfall when compared to other arid regions with similar mean annual precipitation (Desmet, 2007). It typically arrives as widespread, gentle show- ers (average rainfall event 6 mm, P. Carrick unpublished data, cited by Desmet (2007)). Drought conditions are rare, and rainfall is higher than average about once every 10 years, causing ephemeral rivers to flow. Flood events occurred in the Buffels River, for example, in 1945, 1961, 1962, 1963, 1976, 1980, 1986, 1996 and 1997 (D.B.C.M., 2000). Paleoclimates of the northern Namaqualand region are not that well constrained, but the consensus for the greater area seems to be that is was wetter during the Last Glacial (Van Zinderen Bakker, 1976; Tankard and Rogers, 1978; Parkington et al., 2000). 10 Coastal fog adds significantly to the total precipitation. Data from the Pilot Fog Study at Kleinzee for the period July to October 1995 showed that the highest precipitation due to fog occurred during August. The total fog recorded during this period was 1001 mm (D.B.C.M., 2000). Tinely (1985; in D.B.C.M., 2000) estimated that the plants may trap up to 300 mm of fog per annum. That evapotranspiration greatly exceeds precipitation in Namaqualand is re- flected by a vegetation cover (Sandveld bioregion (Desmet, 2007)) of approxi- mately 60%, as well as the presence of saline pans and ephemeral rivers, and areas with salt efflorescence and crusts at the soil surface. The evaporation rate is of the order of 2 m per year. The evaporation rate is higher than expected for a coastal area due to the wind regime, but is reduced by the regular occurrence of coastal fog (D.B.C.M., 2000). The regolith of the Buffels Marine Complex comprises successive late Tertiary marine packages, each deposited during sea-level regression (Pether, 1994). Some rest on kaolinized paleochannel sediments, described in detail by Pether (1994), who concluded that they were laid down as a quartzo-feldspathic sediment in a fluvial environment, and subsequently deeply weathered, with extensive alter- ation of feldspar to kaolinite. These sediments are described in greater detail by Tankard (1966), Tankard (1975), Pether (1994), and Pether et al. (2000). De Vil- liers and Cadman (2001) favoured an early Tertiary age for this channel north of the Swartlintjies River, although it may also contain reworked Cretaceous material (Pether et al., 2000; De Villiers and Cadman, 2001). The marine packages overlying the kaolinised paleochannel sediments are ar- ranged en echelon down the bedrock gradient, from oldest and highest inland to youngest and lowest at the coast. Each package is named after the eleva- tion of its transgressive maximum as represented in the Hondeklip Bay area (Pether, 1994; Pether et al., 2000). The 90 m package is ca. 18 - 16 Ma, the 50 m package early Pliocene, and the 30 m package not well constrained, but ca. 3.3 Ma or younger (Pether et al., 2000). The 30 m package is transgressed by younger littoral deposits up to about 10 m a.m.s.l. (Pether et al., 2000). The sand fraction is dominated by quartz and feldspar, lesser glauconite and phosphatic shell fragments, and variable amounts of heavy minerals (garnet, magnetite, il- menite, biotite, sphene, amphibole, epidote, kyanite (rare) and zircon (rare)) (Pether, 1994). The marine packages are capped by recent aeolian deposits. 11 Coastal dunefields originate at the mouth of the rivers and extend inland in a northerly direction. In places, older soil profiles are overlain by aeolian sands. Soil formation has taken place in aeolian, ?sheetwash?, and marine-deposited sed- iments. Soils within the broader context of Namaqualand are summarized by Francis et al. (2007) and Desmet (2007). Soils on the Namaqualand coastal plain in the vicinity of the Buffels Marine Complex sampling sites can be classified into the Aridisol and Entisol orders of Soil Taxonomy (Soil Survey Staff, 1999). Greater detail is presented in Chapter 2. In summary, the Entisol order, in which the only expression of soil formation is the presence of vegetation and a darkening of the A horizon, is represented by Quaternary aeolian sands. These regic sands are classified as the Namib Form in the South African classification system (Soil Classification Working Group, 1991). The grey sands are of most recent aeo- lian origin and closest to the coast, the red sands are the oldest and furthest inland, and the yellow sands usually occupy an intermediate zone between red and grey. Although described as deep sands there is abundant evidence (even though mostly incipient) of differentiation into horizons (for example bleaching, clay illuviation and secondary cementing by silica and carbonate), such that the more stable landscapes display many typical Aridisol features noted by Buol et al. (1997): surface soil crusts a few millimetres thick are common; somewhat altered subsurface horizons where the original stratification has been obliterated by mixing; argillic subsurface horizons with a higher clay content than the over- lying horizon; eluvial horizons; horizons of carbonate accumulation that vary from calcic to petrocalcic (which are generally on the order of 1 m thickness); red-brown silica-cemented horizons (petroduric in Soil Taxonomy or dorbank in the South African classification system), which may contain a significant amount of calcium carbonate and gypsum and are commonly 1 to 2 m thick. Gypsic and petrogypsic soils occur in the northern part. Profile KV196T is classified as the Garies Form (Orthic A - Red apedal B - Dorbank) in the South African classification system (Soil Classification Working Group, 1991). It is located adjacent to and in a mine excavation in the Buffels Marine Complex on the farm Kareedoornvlei 177 (see Figure 1.2), on the 95 m terrace cut into silcrete, 118 m a.m.s.l. (GPS), Terrain unit: crest. It is further inland than SK11 (Figure 1.1). The profile is pictured in Appendix A. 12 Figure 1.2: Location of profiles KV196T, SK11 and their farm names within the Buffels Marine Complex (BMC) mining area Profile SK11 is classified as the Pinedene Form (Orthic A - Yellow-brown apedal B - Signs of Wetness) in the South African classification system. It is located adjacent to and in a mine excavation in the Buffels Marine Complex on the farm Sand Kop 322 (see Figure 1.2), on the 65 m terrace, 60 m a.m.s.l. (GPS). Terrain unit: footslope. It is pictured in Appendix A. Fieldwork took place in February 2002. Profiles to bedrock were examined and sampled. Samples were named according to the mining excavation to which they were closest, the prefix being the farm name. Sample sites were selected to be representative of each terrace and marine package. Samples (approximately 2 kg) of each horizon were taken from the sidewalls of the mining excavations and from soil pits dug in the undisturbed areas nearby. The soil morphology was described in the field. A 10% solution of HCl was used to test for carbonate, and 30% H2O2 for manganese. Colour was described using a Munsell colour chart (Munsell Color Company, 1975). Horizons falling within 1.5 m of the surface were treated as diagnostic horizons for classification purposes. Those occurring deeper 13 in the profile were described, but treated as non-diagnostic and no consideration given to their position in the profile. Knersvlakte: profile KNERS Like the Buffels Marine Complex to the north, the Knersvlakte receives a reliable winter rainfall (MayAugust) of between 100 and 175 mm (Schmiedel and Ju?rgens (1999), using data from Weather Bureau (1988)). However, unlike the Buffels Marine Complex, models show the climate of the Knersvlakte remained as dry as present back to the Last Glacial Maximum (Desmet (2007), citing data from Midgley et al. (2001)). The Knersvlakte forms the southern part of Namaqualand, just north of the Olifant?s River (Figure 1.1). It is a part of the the Succulent Karoo bioregion (Desmet, 2007), in contrast to the Sandveld bioregion of the area around the Buffels Marine Complex. This area occupies a broad plain created by the proto- Orange River some 20 million years ago and almost completely cuts the continuity of the African escarpment, creating a very distinct physical boundary between the sandstone and shale sedimentary rocks of the Cape Fold Mountains to the south, and the predominately igneous landscape of Namaqualand to the north (Desmet, 2007). The Knersvlakte is consists of fluviatile sediments (Pickford and Senut, 1997), and gypsum at depth that is mined in some places. The area is underlain by shales, phyllites and limestones of the Nama Group and is streaked by numerous quartz veins (Schmiedel and Ju?rgens, 1999), that have weathered to give rise to the distinctive, white, quartz gravel surface (pictured in (Schmiedel and Ju?rgens, 1999)). It is characterised by shallow soils overlying dorbank horizons (Francis et al., 2007). The Knersvlakte profile (Kners) was located 53 km east of Kliprand turnoff on N7, at 30? 48? 47.3?? S, 18? 43? 25.3?? E, 554 m a.m.s.l. (GPS). Terrain unit: ex- tensive plain. Borrow pit, upper horizons removed (disturbed). It was associated with an old (inactive) heuweltjie, and in the vicinity of an alkali intrusive. Field- work took place in June 2004. Samples (approximately 2 kg) were taken of each hardpan horizon. The soil morphology was described in the field. A 10% solution of HCl was used to test for carbonate, and 30% H2O2 for manganese. Colour was described using a Munsell colour chart (Munsell Color Company, 1975). 14 1.2.2 Laboratory methods Slaking tests were devised to distinguish silica- from carbonate-cemented hard- pan horizons. These tests were based on the WRB (1998) definition of petroduric, petrocalcic and fragic horizons. A fragment a few centimetres in diameter was submerged in water, 5M HCl or 6M NaOH and gently heated on a waterbath for four days. More solution was added when necessary. Samples which did not slake were then subjected to alternating 5M HCl and 6M NaOH treatments. Thin sections for optical microscopy from the Knersvlakte profile were impreg- nated with a polyester resin and ground without water, and the samples from the KV196T profile impregnated with an epoxy resin. Fine grained calcite was identified by effervescence in 1M HCl under the optical microscope. Uncoated fragments and thin sections were observed before and after etching in 1M HCl using a Philips Xl30 environmental scanning electron microscope (ESEM) in low vacuum and with an energy dispersive X-ray analysis system (EDX). High vac- uum SEM was done on Au-coated fragments using a Leo 1430VP SEM-EDX system. Major element composition of bulk samples was determined by x-ray fluorescence (XRF). For chemical and mineralogical (x-ray diffraction, XRD) analysis, samples were air-dried, crushed in a pestle and mortar (unless specified as milled, where samples were milled for 5 minutes using a stainless steel ball mill) and passed through a 2 mm sieve. The >2 mm cement was sorted from the >2 mm pebbles and recrushed, to minimise shattering of detrital pebbles/gravel. For XRD analysis, the <2 ?m fractions of the bulk samples were separated by dispersion (shaking briefly by hand, raising the pH to approximately 10 with Na2CO3) and settling. The 1 to 0.08 ?m and <0.08 ?m fractions were sepa- rated by centrifugation. The clay suspension was flocculated by the addition of MgCl2, after lowering the pH to 5 to 7 with HCl to prevent the precipitation of brucite and/or clay destruction. The clay suspension was then Mg-, Ca- or K-saturated, concentrated by centrifugation, and sedimented (or smeared, many of the seopiolite-rich samples developed ?mudcracks?) onto a glass slide. XRD analyses were done with a stepsize of 0.05 degrees and steptime of 40 seconds, using a Bruker D8 Advance Powder Diffractometer with a graphite monochro- mator, 40 kV and 40 mA. Ethylene glycol was sprayed lightly onto the surface of the Mg-saturated sample slides. The <2 mm fraction was used for chemical analyses. The pH and electrical con- 15 ductivity (EC) were measured in a 20 g to 50 mL soil-water mixture. Saturated paste extracts were prepared for all samples except the Knersvlakte profile, since it was truncated so only the duricrust was exposed at the surface. They were prepared with a 300 g sample, to which approximately 100 g (depending on the clay content) of water was added to form a saturated paste, and equilibrated for 24 hours. The pH of the saturated paste was measured directly. The solution from the saturated paste was extracted under vacuum. Electrical conductivity (EC) was measured with a calibrated conductivity meter; alkalinity by potentio- metric titration to pH 4.5 with 0.01M HCl; major cations by atomic absorption spectroscopy; major anions by ion chromatography; Si using the blue silicimolyb- dous acid procedure of Weaver et al. (1968), and dithionite-citrate-bicarbonate extractable Fe using the method of Mehra and Jackson (1960). 1.3 Results and discussion Descriptions of a typical dorbank (petroduric) profile (KV196T), and two ?se- piocrete? profiles (SK11; Kners) are given in Table 1.1. Typical examples of dorbank and ?sepiocrete? (from the Knersvlakte profile) are shown in Figures 1.3 and 1.4. Dorbank (petroduric) occurring in the KV196T profile is the massive type, and the Knersvlakte profile the platy type, and forms prominent cutans on the ?sepiocrete?. In the ?sepiocrete? samples, water applied from a dropper spreads nearly instantly away along very fine cracks, usually in a dendritic pat- tern. This tendency for water to flow along preferential flow paths may explain the development of the sepiolite veins (arrowed in Figure 1.4). 1.3.1 Slaking behaviour Dorbank (petroduric) In this study, even if they effervesced in HCl, the dorbank horizons consis- tently slaked only in concentrated NaOH (Table 1.1). In comparison, a typi- cal weathering-profile silcrete (non-calcareous) capping a hill near the town of Kleinzee did not slake in prolonged contact with heated concentrated NaOH. This suggests the cement in the dorbank is amorphous silica, compared to a more crystalline form in the silcrete. It also reveals that iron oxides are not the 16 T abl e 1. 1: Profil e descripti on s. Roma n nu m er al s in di ca te diff er en t faci es . Bla nk fiel ds : no t det ec te d, ex cep t sla kin g te st s: no t tes te d, text ur e an d consi st en ce : no t app licable . A bb re vi atio ns : ab un.-ab un da nt ; alt .- al te rnati ng ; ang.-a ngula r; bio.-b iological ; c. -coars e; co nt.-co nti nu ous ; diam.- dia m eter ; f. -fi ne ; fr ag.-fragme nt ; m ed .- m edi um ; m od .-m oderate ; oc c.- oc casio nal ; O. M .-organ ic matter ; qt z -q uartz ; se c. -s eco nds ; sl .- slig ht ; v.- ve ry ; w. -wi th . Water Dry Moist Dry Moist Wet HCl NaOH H2O Silicic Duricrust KV196T: Garies Form (Orthic A - Red apedal B - Dorbank). BMC (farm Kareedoornvlei 177), 95m terrace cut into silcrete, 118mamsl (GPS). Terrain unit: crest. Adjacent to mine excavation. /A 0- A Orthic A 5YR 4/6 5YR 3/4 F. med. sand Weak, med. ang. blocky Loose Loose Non- plastic, non- sticky None >20 Clear, smooth Abun. O.M.; many roots /B 0.2->0.45 B1 Red apedal B 7.5YR 5/6 5YR 3/4 F. med. sand Weak, med. ang. blocky Loose Loose Non- plastic, non- sticky None 1 Not reached Few roots KV196T: New profile, upper material removed by mining. May be continuous with sample B. Mining excavation /1 ?1- B2 Red apedal B 5YR 5/6 5YR 4/6 F. med. sand Weak to mod., c., ang. blocky Loose Loose Non- plastic, non- sticky None 1 Diffuse, smooth Bedding or wind erosion /2 1.3- C1 Dorbank 5YR 5/6 5YR 4/6 Med. sand Nodular Hard Firm None 3 Abrupt, smooth /3 1.45- C2 Dorbank 5YR 6/4 5YR 4/6 Med. sand Nodular Hard Firm None 0 2 8 Gradual, smooth wavy Mn on, in peds; abun. white dis- persed non-calcareous material /4 1.6- C3 Calcareous dorbank 5YR 5/6 5YR 4/6 Nodular Sl. hard Sl. firm Strong 0 3-4 3 Abrupt, wavy /5 1.8- C4 Hardpan + Soft carbonate 5YR 8/1 5YR 6/4 Laminar to massive Loose Loose Sl. plastic, non-/sl. sticky Strong 2 Abrupt, tonguing /6 2 - ? C5 Soft dorbank 5YR 5/6 5YR 4/6 Nodular to massive Sl. hard- hard Friable-v. sl. firm Sl. plastic, sticky None 3-4 4 5 2 Obscured Gypsum spots /7 ? - 3.5 C6 Soft dorbank 5YR 5/8 5YR 5/6 Massive Soft-hard Friable Non- plastic, non- sticky None 5 2 Obscured /8 ? 3.5- C7 Dorbank 5YR 5/6 5YR 5/6 Massive V. hard V. firm None 0 2-4 15 Clear, tonguing Gypsum along fractures /9 ?4.5- C8 Dorbank 7.5YR 6/4 7.5YR 5/6 Massive to vesicular Hard Firm-v. firm None 0 3 12 Gradual, unclear /10 ? 6- ? 10 II C9 Marine deposit 5YR 5/6 5YR 5/6 Med. sand Mod., med.-c.ang. blocky Loose Loose Non- plastic, sl. sticky None 3 Abrupt to bedrock Salty taste Sepiolitic Duricrust a) SK11: Pinedene Form (Orthic A - Yellow-brown apedal B - Signs of Wetness). BMC (farm Sand Kop 322, 65m terrace, 60mamsl (GPS). Terrain unit: footslope. Adjacent to mine. /A 0- A Orthic A 7.5YR 5/5 7.5YR 3/4 Med.-f. sand Single grain apedal Loose Loose Non- plastic, non- sticky None 4 Abrupt, smooth Bio. crust, O.M., occ. droppings; many roots /B 0.2- B Yellow-brown apedal B 7.5YR 6/6 7.5YR 4/6 Med.-f. sand Few Weak, f. blocky Loose Loose Non- plastic, non- sticky None 2 Clear, smooth Few roots /C 0.6- E Gleyed sand 7.5YR 6/4 7.5YR 5/4 Med. sand Apedal single grain Loose Loose Non- plastic, non- sticky None 1 Abrupt, smooth /D 0.65- B2 Neocutanic 7.5YR 5/6 7.5YR 4/6 Loamy sand Weak, c., ang. blocky Sl. hard Loose Sl. plastic, sticky None 3 Abrupt, smooth Few roots; few clay cutans /E 0.7- II C1 'Sepiocrete' None 2-3 2-3 2-3 SK11: New profile 30 m away, upper material removed by mining. Mining excavation. /1 0.01- II C1 'Sepiocrete' 10YR 7/3 10YR 6/4 Laminar calcrete V. hard V. firm Strong 0 0-1 2-3 >20 Gradual, wavy Diffuse Mn coating some peds /6 0.45- II C2 Blocky calcretized sediment 10YR 7/4 7.5YR 5/6 Loamy sand Strong, c., blocky to prismatic Hard-v. hard Firm-v. firm V. strong 12 Gradual, wavy Mn spots on peds; calcite cutans /3 0.9-1.1 II C3 Hardpan carbonate 10YR 8/3 7.5YR 7/6 Laminar V. hard Firm-v. firm Strong 2-3 3 4 12 Clear, wavy Mn spots on some peds /9 0.9-1.1 III C3 Hardpan carbonate 10YR 8/3 7.5YR 7/6 Laminar V. hard Firm-v. firm Strong 3 2-3 2 12 Clear, wavy Mn spots on some peds /4 1.1-1.15 III C4 Neocarbonate 7.5YR 6/4 7.5YR 5/6 Med. sand Weak, ang. blocky Loose Loose Non- plastic, non- sticky Strong 1 /4b 1.15-1.5 III C5 Neocarbonate 10YR 7/4 10YR 6/4 Loamy sand Weak., c.,blocky Loose Loose Sl. plastic, sl. sticky Mod. 1 Clear, wavy to sample 7 /5 0.8-1.5 Termite nest 10YR 6/4 10YR 5/6 Loamy sand Mod., c., blocky Firm Loose Sl. plastic, sl. sticky Local 3 Diffuse Occ.Mn;chert frag.;calcite cutans /7 1.5->2.2 IV C6 Sedimentary 5Y 7/4 10YR 6/3 Sandy loam Local Mod., c., blocky Firm Loose Sl. plastic, sl. sticky Local 3 Not reached b) KNERS: Knersvlakte, 53 km E of Kliprand turnoff on N7. 30?48'47.3'' S, 18?43'25.3'' E, 554 mamsl (GPS). Terrain unit: extensive plain. Borrow pit, upper horizons removed . Profile 1 /A ~0.1 'Sepiocrete' 5YR 8/2 7.5YR 6.5/6 Abun. Massive V. hard V. firm Mod.-sl.; locally none 1 0 5 Sepiolite cutan /1 ~0.15 5YR 7/6 5YR 5/6 Laminar coating V. hard Firm None 0 5 >20 Cutan to Kners 2 /2 ~0.15 'Sepiocrete' 5YR 8/3 7.5YR 6/6 Mod. Massive V. hard Firm- v. firm Slight 2-3 0 5 Many cracks Profile 2: From opposite side of borrow pit. /5 ~0.15 Dorbank 5YR 6/6 5YR 5/6 Abun. qtz Platy to blocky Hard Firm None 0 5 15 Abun. cracks; occ. Mn; few roots; sepiolite cutan /4 ~1 'Sepiocrete' 5YR 8/2 7.5YR6.5/6 Abun. qtz Massive V. hard V. firm Mod.-sl.; locally none 0 0-1 5 Sepiolite cutan Transition Other Features Dorbank cutan few cm thick on Kners2 Alt. HCl/ NaOH absorp- tion (sec.) Consistence Slake (0=no, 5=full)Structure Efferves- cence HCl Depth (m) Horizon (Soil Classification Working Group, 1991) Field Texture >2mm Frag- ments Munsell colour 17 (a) Profile showing area magnified in (b) (b) Scale bar 1 cm Figure 1.3: Reddish-brown dorbank (petroduric) forming cutan (Kners/1) on white ?sepiocrete? (Kners/2) from the Knersvlakte (Profile 1, described in Table 1.1). Hammer is 30 cm long. 18 Figure 1.4: White ?sepiocrete? from the Knersvlakte (Profile 2), Kners/4 de- scribed in Table 1.1. Sepiolite veins arrowed. Lower fragment partially coated by white, non-calcareous sepiolite cutan (C). Hammer handle 15 cm long. cementing agent in dorbank. Despite their red-brown colour (Table 1.1), the dithionite-citrate-bicarbonate extractable Fe (Mehra and Jackson, 1960) con- tent of KV196T/8 was only 0.4%. This is consistent with the range found by Chartres (1985, p. 329) for red-brown hardpans in Australia. ?Sepiocrete? SK11/E was the only example where partial slaking occurring in prolonged water treatments. The remaining ?sepiocrete? horizons were generally non- to locally calcareous (Table 1.1), suggesting calcite is not the primary cement. HCl treatments did not result in slaking, nor did NaOH treatments. Partial slaking was observed when alkali/acid treatments were alternated. This is in contrast to the dorbank samples, which slaked in NaOH even if calcareous. This may 19 suggest that the cement in ?sepiocrete? is not solely amorphous silica. However, Mart??nez-Ram??rez et al. (1996) noted that the addition of sepiolite to a pH 12.4 NaOH solution caused the pH to be reduced to 8.6. Although the experimental methodologies were different in that the NaOH solution was refreshed on drying, and was more concentrated compared to Mart??nez-Ram??rez et al. (1996), the pH was not measured and the possibility exists that the presence of sepiolite caused a reduction in pH sufficient to reduce the solubility of silica. However, a weathering profile silcrete (similar to that described by Summerfield (1983a)) also did not slake in concentrated NaOH under the same experimental conditions. This suggests that in addition to sepiolite content, the degree of silica crystallinity and/or impregnation and/or total amount of secondary silica also plays a role in dictating whether slaking of a particular sample will occur. The slaking tests established that calcite was not the primary cement in ?sepi- ocrete?, but were inconclusive in determining whether silica or sepiolite formed the cementing agent. 1.3.2 Mineralogy Dorbank (petroduric) The clay fraction of dorbank profile KV196T contains smectite, and also mica/illite (Figure 1.5). The dorbank samples in the Knersvlakte profile (Kners/1 and Kners/5 in Figure 1.6) contain sepiolite, and possibly palygorskite. With the exception of the non-calcareous KV196T/4, the dorbank horizons show broad, low peaks of the clay minerals, a characteristic noted by both Flach et al. (1969) and Blank and Fosberg (1991), who attributed it to disordered siliceous coatings masking the clay minerals. Figure 1.7 confirms that very fine material coats some clay grains. Amorphous silica (broad opal-A peak at 0.41-0.403 nm (Drees et al., 1989)) does not appear in any quantity on the diffraction patterns of profile KV196T, despite its presence as cementing agent being suggested by slaking only in strong alkali. KV196T/2 is the only horizon that shows a broad rise of the background in the 0.4 nm vicinity. The amorphous silica peak is more apparent in the Knersvlakte 20 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a 1 .0 n m Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Op al C T/A 0.4 1 n m Kao linit e 0 .35 7nm Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; p aly. 0.4 49 nm Q u artz ; se pio lite ; m ica 0.3 34 nm Ca lcite 0.3 04 nm Milled; 1-0.08 ?m; Ca Milled; <0.08 ?m; Ca Milled; <0.08 ?m; Ca; EG Milled; 1-0.08 ?m; Ca; EG KV196T/3 Dorbank 1.45-1.6 m <2 ?m; Mg <2 ?m; Mg; EG KV196T/5 Hardpan carbonate 1.8-2 m <2 ?m; Mg <2 um; Mg; EG KV196T/2 Dorbank 1.3-1.45 m <2 ?m; Mg KV196T/4 Dorbank (calcareous) 1.6-1.8 m <2 ?m; Mg; EG Op al A 0.4 03 nm Mic a; s epi olite ( di ff.) 0.5 04 nm Milled; <2 ?m; Mg KV196T/8 Dorbank 3.5-4.5 m Milled; <2 ?m; Mg; EG Milled; <2 ?m; Mg Milled; <2 ?m; Mg; EG -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 1.5: XRD traces of Mg- or Ca- saturated clay fractions from selected horizons in profile KV196T. Diff: diffuse; EG: ethlyene glycol. 21 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a 1 .0 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Op al C T/A 0.4 1 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; p aly. 0.4 49 nm Q u artz ; se pio lite 0.3 34 nm Ca lcite 0.3 04 nm Milled; <2 ?m; Mg KNERS/4 'Sepiocrete' Milled, <2 ?m; Mg; EG Sep iolit e ( d iff.) 0.5 04 nm Sep iolit e 0 .37 4 n m HCl-slaked; milled; <2 ?m; Mg HCl-slaked; milled <2 ?m; Mg; EG NaOH-slaked; milled; < 2 ?m; Mg NaOH-slaked; milled; <2 ?m; Mg; EG Milled; 1-0.08 ?m; Ca Milled; 1-0.08 ?m; Ca; EG Milled; <2 ?m; Mg Milled; <2 ?m; EG; Mg KNERS/5 Dorbank Milled; <2 ?m; Mg Milled; <2 ?m; EG; Mg KNERS/1 Dorbank cutan on KNERS/2 KNERS/2 'Sepiocrete' Milled; <2 ?m; Mg Milled; <2 ?m; Mg; EG Op al A 0.4 03 nm -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 1.6: XRD traces of Mg- or Ca- saturated clay fractions from selected horizons in the Knersvlakte profile Kners. Diff: diffuse; EG: ethlyene glycol. 22 Figure 1.7: TEM image showing very fine material (such as on the left) coating clay grains (top right) in dorbank (petroduric) Kners/5 (milled). profile than KV196T. There seems to be a small opal-C/T peak at 0.41 nm in KV196T/8 and /3, and Kners/1 and possibly /5. Quartz is present in the clay fraction. Since quartz is ubiquitous in the marine sands that form the parent material, and the clay fraction was liberated by milling, there is a large detrital component which cannot be distinguished from authigenic silica. ?Sepiocrete? SK11/E, the non-calcareous ?sepiocrete? that slaked partially after prolonged soaking in water (Table 1.1), is the only horizon within the SK11 profile that contains monominerallic sepiolite in the clay fraction (Figure 1.8). Amorphous silica does not appear in the clay fraction. The other horizons in the profile contain smectite, kaolinite and mica/illite in addition to sepiolite. Palygorskite appears to be a minor constituent, with the 0.664 nm peak best defined in the <1 ?m fractions (Figure 1.8). 23 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a 1 .0 n m Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Op al C T/A 0.4 1 n m Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; p aly. 0.4 49 nm Q u artz ; se pio lite; mi ca 0.3 34 nm Ca lcite 0.3 04 nm Mic a; s epi olit e ( d iff.) 0. 504 nm Sep iolit e 0 .37 4 n m <2 ?m; Mg <2 ?m; Mg; EG SK11/5 Termite nest 0.8-1.5 m <2 ?m; Mg SK11/E 'Sepiocrete' 0.7 m - <2 ?m; Mg; EG <2 ?m; Mg <2 ?m; Mg; EG 1-0.08 ?m; Ca SK11/D Neocutanic 0.65-0.7 m 1-0.08 ?m; Ca; EG <0.08 ?m; Ca <0.08 ?m; Ca; EG Op al A 0.4 03 nm <2 ?m; Mg SK11/9 Petrocalcic 0.9-1.1 m <2 ?m; Mg; EG <2 ?m; K <2 ?m; K; 300?C -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 1.8: XRD traces of Mg-, K- or Ca- saturated clay fractions from selected horizons in profile SK11. EG: ethlyene glycol. 24 XRD traces of the clay fraction of the ?sepiocrete? samples in the Knersvlakte profile (Figure 1.6) show strong sepiolite peaks in all size fractions. These are not as clear on the dorbank horizons Kners/1 and Kners/5, but this may be a func- tion of disordered siliceous coatings masking the clay minerals in dorbank (Flach et al., 1969; Blank and Fosberg, 1991). Palygorskite is much less prominant that sepiolite in the <2 ?m fraction. Both the dorbank and ?sepiocrete? horizons in the Knersvlakte profile show a very broad amorphous silica peak emerging on the background at around 0.4 nm, although it is slightly less expressed in Kners/1 and Kners/2. This amorphous silica peak is more apparent in this profile than either KV196T (Figure 1.5) or SK11 (Figure 1.8). As with the dorbank, the quartz peaks are likely to have a largely detrital component which cannot be distinguished from authigenic silica. The <2 ?m fraction of partially slaked ?sepiocrete? from Kners/4 was milled and analysed. The HCl treatment caused a significant broadening of the sepiolite peak (Figure 1.6), suggesting the sepiolite was partially destroyed by the acid attack, and consistent with the findings of Goza?lez et al. (1984) and De?ka?ny et al. (1999). The 0.4 nm amorphous silica peak was not affected by acid treatment. The NaOH treatment also caused a slight broadening of the sepiolite peak, but not to the same degree as the acid treatment. The alkali treatment did not seem to have affected the 0.4 nm amorphous silica peak either, suggesting the alkali treatment was insufficient to remove the silica from these samples. 1.3.3 Micromorphology Dorbank (petroduric) The dorbank micromass is generally dominated by the reddish orange colour imparted by Fe (hydr)oxides, with strongly oriented clay more prominent than amorphous silica in many places (Table 1.2). The amorphous silica is generally clear and yellowish, and is generally isotropic. It may be than it has been masked by the colour of the Fe. Spherulites (crystal aggregations with an approximately circular outline and a permanent cross of extinction in crossed polarized light (Canti, 1997)) are present, some have an isotropic core and appear to be silica (Figures 1.9(a)-(b)), others show pseudo-negative uniaxial interference colours and concentric banding under crossed-polarisers (Figures 1.9(c)-(e)). Summer- 25 field (1983a) observed spherulites with pseudo-uniaxial extinction crosses com- prised of length-fast chalcedony in a vugh-fill in South African silcrete. The orientation of the crystals making up the spherulites in the Namaqualand sam- ples, however, could not be resolved with certainty. Laminated and layer-coated nodules and aggregates are present in Kners/5 (Figure 1.9(f)-(g)), which are remarkably similar to the compound ?glaebular aggregates? from silcretes in the Riversdale-Albertina area shown by Summerfield (1983a, p. 903). Table 1.2: Thin section descriptions. Abbreviations: diam.: diameter; incl.: including; max.: maximum; occ.: occasional; OIL: oblique incident light; ppl: plane polarised light; xpl: cross polarised light. Kleinzee, dorbank, sample KV196T/4. 1 vertical section 20 x 40 mm; thickness ~ 30-35 ?m. Knersvlakte, dorbank cutan, sample KNERS/1. 3 vertical sections 50 x 40 mm; thickness ~ 30-35 ?m. Knersvlakte, 'sepiocrete' sample KNERS/2. 3 vertical sections 50 x 40 mm; thickness ~ 30-35 ?m. Knersvlakte, dorbank, sample KNERS/5. 3 vertical sections 50 x 40 mm; thickness ~ 30-35 ?m. Micro-structure Micro-structure Micro-structure Micro-structure Vughy microstructure. Channel and vughy microstructure. Weakly separated granular to massive microstructure. Granules 0.6 - 0.05 mm diam. Vughy microstructure. Groundmass Groundmass Groundmass Groundmass c/f limit: 5 ?m c/f limit: 5 ?m c/f limit: 5 ?m c/f limit: 5 ?m c/f ratio: 8:1 c/f ratio: 9:1 c/f ratio: 1:8 c/f ratio: 8:1 c/f-related distribution: chito- gefuric. c/f-related distribution: close porphyric. c/f related-distribution: open porphyric. c/f related-distribution: chito-gefuric to close porphyric. Coarse Material Coarse Material Coarse Material Coarse Material Subrounded. quartz, less feldspar, minor opaques, mica. Grains ~0.4 - 0.03mm. Predominantly subangular quartz, also opaques. 5mm diam. common, to 0.2 mm. Predominantly subangular quartz, also opaques, feldspar, garnet. 1mm diam. common, to 0.5 mm. Subrounded quartz, less feldspar, minor opaques, mica. Grains ~1cm - 0.02mm, orientated ~ with long axis horizontal. Micromass Micromass Micromass Micromass PPL: Limpid, orange-brown. Comprises Fe-stained silica, and/or clay, some pure silica. Although sample slakes only in NaOH, pure silica in the matrix is difficult to discern, it may be very fine-grained. XPL: Granostriated b - fabric. PPL: Banded dark orange predomin- antly speckled; some limpid layers. Generally orientated parallel to grain edges or to channels. Lamin-ated (with very fine dark bands) and entire lamination "crenulated". Dispersed fine calcite, grey. Mn, mammilate, dispersed. XPL: Granostriated b-fabric. PPL: Speckled grey-yellow. Seems to be predominantly sepiolite, locally covered by finely dispersed grey secondary calcite which projects into tiny voids giving a crystallitic b-fabric, and locally some silica. XPL: Speckled b-fabric dominant, locally cross-hatched and crystallitic. PPL: Limpid to speckled, reddish- orange-brown. Comprises Fe- stained silica, and/or clay, some pure silica interspersed. XPL: Granostriated b-fabric. Pedofeatures Pedofeatures Pedofeatures Pedofeatures Grains coated by amorphous Fe- oxides incl. hematite (red in OIL). Also over this are clay coatings, and amorphous (isotropic) silica, which tends to be concentrated on the outer edges and intergrain areas. Coatings are length-slow. Dense infillings of clays, forming microlaminated crescent coatings. In some cases they appear to seal the void and no further vertical illuviation occurs below. Silica is over the clay. Spherulites (with pseudo- negative uniaxial interference colours). Coatings: Grains coated by amorphous Fe-oxides incl. hematite (red in OIL). Clay, silica may be masked by the colour of the Fe oxides. The coatings are length- slow. Crescent coatings, not always associated with voids, the voids may have been completely filled. Spherical, rounded peds 0.5 - 1cm, also elongated and large irregular ( ~ 5cm) peds with laminated ( limpid to speckled with silt-sized quartz) reddish-orange-brown coatings. Some are aggregate features, laminated and layer-coated nodules enclosed with other matrix material within a larger laminated and layer- coated nodule. Spherulites: some show pseudo- negative uniaxial interference figures, some have an isotropic core, others concentric banding. Coatings: Some grains have an edge that seems to be coated with orientated silica. "Crenulated" clay layers very common, layers orientated parallel to grains and filling interstices, in some cases the inter-coating areas are filled with amorphous silica. Channels: generally branching downwards, may be root channels, <0.4 mm diam, accommodating faces, rough textured sides, sometimes side-branching chambers. Channels and chambers appear to be coated and/or filled with decomposed organic material, generally 0.02mm diam., often occurring as aggregates, and usually associated with fine grey calcite (often only distinguishable by effervescence with HCl), in some places resolved as needle fibre calcite. Channels lined and filled with sepiolite, arranged in fan-like bundles. Speckled, grey, laminated calcic pendants and cappings. Discreet subangular limpid yellow clay (silica?) nodules (300 - 70 um) which have a cross-striated or isotropic b-fabric and very low (first order) interference colours, occasionally will show pseudo- negative uniaxial interference colours, may be coated with micrite, orientated clay or silica (isotropic). Coatings: Quartz and clay particles coated with micrite (mixed with clay) hypocoatings, some are coated with isotropic silica and Fe. Illuviation of clay has been a dominant soil forming process, with strong ori- 26 (a) Spherulites (left), limpid crescent coating, amorphous silica (isotropic). Kners/3, ppl. Scale bar 0.1 mm. (b) Same view as (a), xpl. (c) Spherulites. Kners/5, ppl. Scale bar 0.1 mm. (d) Same as (c), xpl. (e) Same as (c), xpl, ? plate, fast direction indicated. (f) Portion of complex aggregate (border with matrix arrowed). Kners/5, ppl. Scale bar 0.2 mm. (g) Same view as (f), xpl. Figure 1.9: Typical views of silica in dorbank (petroduric) horizons. Way-up is top of page. entation of clay parallel to grains and crescent coatings on the lower part of voids common (Figures 1.9(a)-(b)). Figure 1.10(a)-(b) shows a void blocked by clay-coated grains, resulting in a crescent coating forming on top and no further illuviation into the lower part of the void . Sullivan (1994) obtained optically ori- 27 ented clay coatings on coarse quartz sand by passing smectitic clay suspensions through sand columns. These coatings were not removed by subsequent passage of water, suggesting that the dominant clay deposition mechanism involves the retention of suspensions on sand surfaces by capillary and adsorption forces. It appears than clay illuviation occurred before silica precipitation in the dorbank horizons, since the silica is generally (although not exclusively) concentrated on the void side of the clay coatings (Figures 1.9(a)-(b), 1.10(a)-(b)), similar to that described by Litchfield and Mabbutt (1962) for red-brown hardpans in Western Australia. Layering suggestive of illuviation is evident in Figure 1.10(c), the composition of which (Figure 1.10(d)) is consistent with montmorillonite clay, silica and the reddish orange colour imparted by Fe (hydr)oxides. Smectite neoformation also appears to be present. SEM images of a fragment of KV196T/3 (Figure 1.10(f)) are similar to images of neoformed smectite in Borchardt (1989, p. 680). The composition (Figure 1.10(e)) is very Mg-rich for a smectite: it is closer to that of saponite than montmorillonite. The XRD trace shows the broad peak of an expandable phase (Figure 1.5), as well as mica and kaolinite peaks. The mica and kaolinite alone could not account for the high Mg content. The horizon KV196T/3 occurs fairly high up in the profile (Table 1.1), and it may be that the clay is neoformed in the upper part, and illuviated downwards together with silica and more Al-rich smectite to form harder, more laminated horizons like KV196T/8 (Figure 1.10(c)). ?Sepiocrete? The micromorphology of typical ?sepiocrete is given in Table 1.2. They are distinctly different from dorbank horizons, the most obvious being the granular microstructure, the micrite-rich matrix, and the lack of crescent clay coatings (Figure 1.11). Prolate calcite crystals (length-slow) project into the inter-granule areas (Figures 1.11(a)-(b)). Although the colour and localised effervescence in HCl resemble calcrete in the field, calcite is not the primary cement: Figures 1.12 and 1.13 shows a fragment from Kners/4 before and after etching in 1M HCl for 30 seconds, until efferves- cence ceased. Effervescence was only observed in the lower part of the fragment where pitting developed (area C in Figure 1.12(b)). Before etching, the calcite was present in the form of finely dispersed, less than 2 ?m particles, consistent with the low Ca content (Figure 1.12(d)). After etching, most of the calcite was 28 (a) Crescent coating developed above blocked void (v). Scale bar 0.1 mm. Way-up is top of page. KV196T/4, ppl. (b) Same view as (a),xpl. (c) Layering in dorbank, KV196T/8. Scale bar 20 ?m. (d) Spectrum of square in (c). (e) Layering and neoformed clay, KV196T/3. Scale bar 20 ?m. C: charging. (f) Magnification of square indicated in (e). Scale bar 1 ?m. Figure 1.10: Neoformed and illuviated smectite and silica in dorbank (petroduric) fragments. removed. The sample remained very firm (Figure 1.12(b)), which shows calcite is not the primary cement in these samples, and the micrite causing the speckled b-fabric observed with the optical microscope is an accessory feature. 29 (a) Granular microstructure. Scale bar 0.1 mm, Kners/2, ppl. Way-up is top of page. (b) Same view as (a), xpl. (c) Thin section, SEM image of microstructure. Scale bar 100 ?m, Kners/2. (d) Magnification of (c). Scale bar 10 ?m. Figure 1.11: Optical and SEM images of granular microstructure in ?sepiocrete?. 30 The high Mg/Si ratio of the fibres both pre- and post etching (Figure 1.12(d), (f)) is consistent with molar Mg/Si of 0.67 for sepiolite (Stoessell, 1988). An interlocking, subparallel morphology is present over much of the sample (Figures 1.12(c), (e), 1.13(b), (c)), perhaps contributing to the induration as suggested by Francis et al. (2007). It is sometimes difficult to distinguish sepiolite from silica with certainty under the optical microscope, given the low birefringence of sepiolite (as in Figures 1.14(a)-(b). Sepiolite appears to form the matrix areas (Figures 1.11, 1.12, 1.13 1.14(d)), whereas the silica is localised: Figures 1.13(c)-(d) show morphology suggestive of localised amorphous silica. Elemental mapping in Figure 1.14(c) shows the laminar area to be Si-dominant, and poorer in Mg than comparative mapping of sepiolite-rich areas like Figures 1.11(d) and 1.14(d). Pedofeatures similar to the sepiolite/silica indicated in Figure 1.14(a) have been recorded in Sardinia, Italy in >2.7 Ma buried paleosols (Usai and Dalrymple, 2003). They suggested that the ?milky pedofeatures? as they termed them, were made of different phases of silica such as inorganic opal-A, opal-C and opal-CT, but mainly composed of opal-C. They noted that slight and local birefringence in colourless white-milky pedofeatures has been previously described for opal-CT by Drees et al. (1989). Although silica was the main constituent of the ?milky pedofeatures? (reaching 67% of the total weight), the remaining constituents were ? 25% Al2O3, 5% MgO, 2 to 4.6% Fe2O3 and 2% CaO, which they attributed to chemisorbed impurities to silica (Drees et al., 1989) or analytical imprecision. The tapering texture of some sepiolite fibres (seen in Figures 1.12(c), 1.13(a), (c) and (d)) is similar to that observed by Khoury et al. (1982), Blank and Fosberg (1991) and Akbulut and Kadir (2003). The wider ?root? of the tapering structure suggests that that crystal growth is from the substrate into the void, with a greater number of shorter fibres growing at the base and fewer fibres reaching greater lengths growing into the void. The micromorphology of the ?sepiocrete? shows that calcite is not the cementing agent. Whether sepiolite or silica forms the primary cementing agent is not as clear: while most of the intergrain area seems to be composed of pure sepiolite, there are areas that seem to be composed of dense laminar silica. It remains speculative which has had the greater influence: the pervasive sepiolite matrix with its interlocking fibres, or the more localised silica with its ability to indurate a sample to a very firm consistency. 31 (a) Before etching. Scale bar 1 mm. (b) After etching (c) Before etching area B. Scale bar 2 ?m. Weight % C 32 O 39 Na 1.3 Mg 8.4 Al 0.7 Si 13 Cl 1.6 K 0.3 Ca 3.7 Fe 0.4 Total 100 m Mg/Si 0.74 (d) EDX analysis of (c) (e) After etching area B. Scale bar 5 ?m. Weight % C 34 O 37 Na 1.1 Mg 9.2 Al 0.9 Si 15 Cl 1.7 K 0.2 Ca 0.7 Fe 0.5 Total 100 m Mg/Si 0.72 (f) EDX analysis of (e) Figure 1.12: SEM images: 1M HCl etching of ?sepiocrete? Kners/4. EDX anal- ysis semi-quantitative. Mg/Si ratios are molar ratios. Continued in Figure 1.13. 32 (a) After etching area B. Scale bar 2 ?m. (b) Magnification of (a). Scale bar 1 ?m. (c) After etching area A. Scale bar 2 ?m. (d) After etching area C. Scale bar 10 ?m. Figure 1.13: Continued from Figure 1.12 - SEM images: 1M HCl etching of ?sepiocrete? Kners/4. 1.3.4 Genesis Dorbank (petroduric) Dorbank (petroduric) genesis may be explained by the high pH of the upper horizons (Table 1.3) which may increase the solubility and mobility of silica, while the sodic environment may contribute to clay dispersion (Ellis, 1988; Ellis and Lambrechts, 1994), and facilitate the hydrolytic exchange of Na+, with an associated rise in pH when the soils are submerged in water (McBride, 1994, p.91-95). Under the prevailing low rainfall conditions, the soluble silica would not leach through the profile but rather accumulate in the subsoil. The loci of silica deposition in duric soils depends on both the depth of water movement and the amount of adsorption surface area (Chadwick et al., 1987), and it seems likely that the abundance of illuviated clay in the pore spaces provided a large adsorption area for silica, and is consistent with the thin section observations that clay illuviation occurred before silica precipitation. The dorbank horizons have 33 (a) Amorphous silica (arrowed); ?: sepiolite/silica. Scale bar 0.1 mm, ppl. Way-up is top of page. (b) Same view as (a), xpl. (c) Thin section, SEM image of laminar silica (Si). Scale bar 5 ?m. (d) Thin section, SEM image of laminar silica (Si), interstitial sepiolite (white arrows) and calcite (white, prolate). Scale bar 20 ?m. Figure 1.14: Optical and SEM images of silica in ?sepiocrete? Kners/4. 34 Table 1.3: Chemistry of selected horizons described in Table 1.1 (n/a: not anal- ysed). pH EC* pH EC* HCO 3- Cl- SO42- Ca2+ Mg2+ Na+ K + Si Mg/Si SAR # Silicic Duricrust: dorbank; Profile KV196T: /A 0-0.2 m Orthic A 8.6 0.75 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /B 0.2->0.45 m Red apedal B 10.2 0.16 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /1 ?1-1.3 m Red apedal B 10.1 0.09 9.3 0.78 6.6 2.7 0.66 0.17 0.11 5.7 0.81 0.71 0.16 15 /2 1.3-1.45 m Dorbank 8.8 0.84 8.0 7.8 2.3 80 11 9.5 14 100 1.5 1.4 9.7 29 /3 1.45-1.6 m Dorbank 8.7 0.03 7.8 19 0.58 223 28 25 40 174 1.9 0.67 59 31 /4 1.6-1.8 m Calcareous dorbank 8.6 0.06 8.0 38 1.8 418 79 0.87 88 429 4.4 0.29 300 64 /5 1.8-2 m Hardpan/Soft carbonate 8.4 0.03 8.2 7.2 2.24 45 78 60 15 51 2.2 0.75 20 8 /6 2 - ? m Soft dorbank 8.5 0.09 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /7 ? - 3.5 m Soft dorbank 9.0 1.1 8.2 9.0 3.0 90 11 6.6 8.4 123 1.8 1.3 6.5 45 /8 ? 3.5-?4.5 m Dorbank 8.5 0.02 8.0 11 0.16 56 91 25 14 96 1.4 0.42 34 22 /9 ?4.5- ?6 m Dorbank 8.2 0.06 7.7 38 1.6 478 68 53 84 588 6.1 1.5 57 71 /10 ? 6- ? 10 m Marine deposit 7.6 0.07 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a Sepiolitic Duricrust: 'se piocrete'; Profile SK11: /A 0-0.2 m Orthic A 9.1 0.14 8.5 1.4 7.3 8.2 1.9 5.8 2.9 16 0.94 0.55 5.3 8 /B 0.2-0.6 m Yellow-brown apedal B 9.2 0.06 8.6 0.60 4.6 2.6 0.39 2.2 1.2 6.3 0.68 0.35 3.4 5 /C 0.6-0.65 m Gleyed sand 9.1 0.08 8.2 0.72 3.7 4.4 0.50 1.6 1.0 8.8 1.0 0.39 2.7 8 /D 0.65-0.7 m Neocutanic 8.7 1.3 7.9 9.4 4.1 88 4.3 3.7 6.2 78 2.3 0.60 10 35 /E 0.7- m 'Sepiocrete' 8.4 0.03 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /1 0.01-0.45 m 'Sepiocrete' 8.7 0.04 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /6 0.45-0.9 m Blocky calc. sed. 9.1 0.02 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /3 0.9-1.1 m Hardpan carbonate 9.2 0.02 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /9 0.9-1.1 m Hardpan carbonate 8.8 3.6 8.0 19 0.23 290 29 21 30 177 2.7 1.3 24 35 /4 1.1-1.15 m Neocarbonate 8.9 0.04 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /4b 1.15-1.5 m Neocarbonate 8.7 2.5 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /5 0.8-1.5 m Termite nest 8.9 2.6 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a /7 1.5- >2.2 m Sedimentary 8.4 6.3 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a *EC: mS/cm; cations, anions: mmolc/L; Si: mmol/L. # SAR: Sodium adsor ption ratio Na/ ((Ca + M g)/2 ) 0.5 1: 2.5 Saturated paste extract * a low water permeability (Table 1.1). It also provides the slow silica-saturated pore water flow rates (Summerfield, 1983b), which are needed for the shift toward equilibrium and subsequent silica precipitation. Although most of the clay in the dorbank profile KV196T is strongly oriented and illuvial in origin, there appears to be a neoformed Mg-rich smectite in hori- zon KV196T/3 (Figure 1.10(f)). Borchardt (1989) noted that saponites are seldom found in soils, and appear to be inherited from the parent material. The Namaqualand coastal plain, however, is a Mg-rich environment (Table 1.3). Ingle`s et al. (1998) found diagenetic Mg-rich smectites in the Ebro Basin, an en- vironment of high alkalinity, silica, magnesium and water stagnation favourable to the formation of these minerals. Hay et al. (1986) found them in the Armagosa desert, and noted that pre-existing sheet silicates such as montmorillonite can form a template for the precipitation of Mg-smectite, or it may precipitate di- rectly from solution. Jones and Galan (1988) noted that trioctahedral smectite precipitation is favoured over sepiolite or palygorskite in an environment with the highest pH (> 9.5), a high (Mg+Fe)/Si ratio (rather than high (Mg+Si)/Al), and highest the alkali salinity. 35 A broken sponge spicule is evident in horizon KV196T/3 (Figure 1.15). Sponge spicules often occur in ponded or poorly drained environments (Drees et al., 1989). Although the dryness and red colour of the dorbank profile KV196T (Table 1.1) suggest a well aerated environment, Francis et al. (2007) showed how textural discontinuities in typical Aridisols in Namaqualand can subtly modify water infiltration and enhance water storage in the profile. The change from a nodular to a massive structure around 2 m depth (Table 1.1) is accompanied by an EC maximum (Table 1.3). This, together with the presence of the sponge spicule, suggests a periodically water-saturated environment, the composition of which (Table 1.3) is favourable to Mg smectite clay neoformation. Figure 1.15: SEM: broken sponge spicule in dorbank (petroduric) fragments. ?Sepiocrete? Clay dispersion and migration are not as prominent as in the dorbank horizons, perhaps due to their lower SAR (Table 1.3). The pH of sepiolite-bearing horizons is consistent with Jones and Galan (1988), in that sepiolite formation is favoured over palygorskite and trioctahedral smectite in the 8-9.5 pH range. Rodas et al. (1994) defined ?palycrete? as generally having more than 65 % pa- lygorskite. It may be that when an horizon is enriched to such a degree with fibrous clay minerals they have the potential to cause a greater induration than when they occur as an accessory mineral (in calcrete for example), particularly where they display a morphology of intergrown mats and bundles of fibres (Fig- ures 1.11(d), 1.13(b)). A possible mechanism for preferential sepiolite enrichment could be decalcification (Vanden Heuvel, 1964), since sepiolite is closely associ- ated with calcite in Namaqualand soils (M.L. Francis, unpub. data), as in New 36 Mexico (Vanden Heuvel, 1964), and together with palygorskite is a common mineral in the clay fraction of calcretes (for example Netterberg, 1969; Van- den Heuvel, 1964; Singer and Norrish, 1974; Yaalon and Wieder, 1976; Elprince et al., 1979; Hay and Wiggins, 1980; Watts, 1980; Singer and Galan, 1984; Singer, 1989; Blank and Fosberg, 1991; Monger and Daugherty, 1991; Verrecchia and Le Coustumer, 1996; Singer, 2002; Neaman and Singer, 2004; Owliaie et al., 2006). Netterberg (pers. comm., 2003) suggested that some ?sepiocrete? may have orig- inated as decalcified calcrete. Evidence in support of this hypothesis are the presence of calcic pendants and cappings from the Kners profile (Table 1.2), indicating mobilisation of calcite; and in profile SK11 the calcic horizons occur below the sepiolite-rich and ?sepiocrete? horizons (Table 1.1), which may sug- gest preferential leaching of calcite to the lower part of the profile. Whether the calcite was originally present as a calcic or a petrocalcic horizon is not resolved. In the Kners profile it appears the amorphous silica is overlying the sepiolite fibres (Figure 1.13(c)), which suggests that sepiolite precipitated first, followed by silica. This is not so clear from Figure 1.14, where the silica seems to be associated with mineral grains. The timing and precise mechanism of silica pre- cipitation in the ?sepiocrete? is difficult to resolve: in the case of Kners, it may have occurred during the phase that formed the dorbank cutans and overlying dorbank, although there is little micromorphological evidence of illuviation into the ?sepiocrete?, and this cannot be ubiquitously applied to all ?sepiocrete? con- taining profiles, since not all are overlain by dorbank. Figures 1.13(c)-(d) do not clearly resolve whether sepiolite or silica precipitated first, although Figure 1.13(c) could be interpreted as the amorphous silica over- lying the sepiolite fibres. The association of sepiolite and silica has been noted before: Blank and Fosberg (1991) observed sepiolite crystals radiating outward from laminae of opaline silica in duripans in Idaho, and suggested that some sepiolite forms via the addition of Mg to opal, and Hay and Wiggins (1980) observed opal replacing sepiolite (and calcite) in calcrete of the southwestern United States. A possibility may be a process similar to the pseudomorphic re- placement of sepiolite by opal in deposits in the Madrid Basin, Spain (Bustillo and Alonso-Zarza, 2007), which preserved the fibrous microstructure and stri- ated birefringence of the sepiolite. Bustillo and Bustillo (2000) suggested that silicification could occur as a function of the sepiolite precipitation, as initial precipitation of sepiolite depleted the solution in Mg. However, Table 1.3 shows 37 that the Mg is present in excess of the amount of Si required to form sepiolite (using an Mg/Si ratio of 0.67), and thus the Mg will not be depleted relative to Si during sepiolite precipitation (Eugster and Jones, 1979). Several examples of etched quartz grains are present in both ?sepiocrete? profiles in SK11/1 and Kners/4 (not shown), which suggests the localised dissolution and redistribution of silica within the profile. Boettinger and Southard (1991) argued that weathered crystalline silicates from a granitic saprolite are the main source of pedogenic silica for Aridisol pedons in the western Mojave Desert. Reheis et al. (1992) suggested that the silica released by replacement of alumi- nosilicates may be locally precipitated as amorphous or opaline silica and (or) incorporated into newly formed palygorskite and sepiolite. Amorphous silica can be precipitated either by evaporative concentration, as the solution becomes increasing supersaturated, and/or it can be precipitated by a pH change. The buffering effect of sepiolite at pH 8.6 (Mart??nez-Ram??rez et al., 1996), as in the slaking tests, has not been quantified in soil. Since the sepiolite-containing horizons are generally also calcareous, the pH of these soils should be low enough to promote the precipitation of silica in solutions entering these horizons. The rise in pH associated with the hydrolytic exchange of Na+ as the soils are submerged in water (McBride, 1994, p.92) may be the source of the dissolved silica. This could be compounded by the high water-holding capacity of sepiolite, removing water and thereby creating conditions of silica saturation on a local scale. 1.3.5 Classification Dorbank (petroduric) A silcrete was defined as having (an arbitrary) lower limit of 85 wt % SiO2 by Summerfield (1983b), and was used by Nash and Shaw (1998) to assist in the classification of calcrete-silcrete intergrades in the Kalahari, although they noted the influence of detrital quartz and clay minerals on the value. By this definition, the dorbank has too little SiO2 (Table 1.4)to be classified a ?silcrete?; this is consistent with the ?silcrete? and ?dorbank? morphological differences reviewed in Section 1.1.1, although both are cemented by silica. Southern African ?silcrete? 38 does not meet the slaking requirements of dorbank, duripan (Soil Survey Staff, 1999) or petroduric (World Reference Base, 1998) horizons. Since the dorbank horizons are clearly pedogenic (colour, texture, and parent material continuity with surface horizons, Table 1.1), and also meet the requirements of duripan and petroduric horizons (with the exception of the minimum thickness for the cutan Kners/1), classifying them as one of these seems more appropriate than as ?silcrete?, despite the presence of the ?silica-cement? which corresponds to that inherent in the term ?sil-crete?. Table 1.4: XRF analyses of selected duricrusts described in Table 1.1. BULK *1 <2 ? m *2 <2 ? m *3 <2 ? m *4 BULK *5 BULK *6 BULK *1 Al2O3 7.37 12.2 10.6 3.74 6.87 2.22 2.44 - CaO 0.21 4.67 2.85 2.55 13.1 13.3 1.46 - Fe2O3 3.73 5.31 5.49 2.27 1.92 1.00 1.61 - K 2O 1.33 2.20 2.32 2.02 2.32 0.71 0.62 - MgO 0.04 4.73 4.45 21.9 9.8 14.0 6.35 24.89 MnO 0.04 0.07 0.05 0.01 - Na2O 0.06 2.52 1.17 0.59 1.63 1.38 0.16 - P 2O5 0.08 0.15 0.11 0.07 0.11 0.19 0.10 - SiO2 76.9 56.6 64.6 56.9 54.1 57.0 76.5 55.65 TiO2 0.47 0.46 0.59 0.16 0.42 0.24 0.18 - H2O- 2.02 2.40 - LOI 9.20 11.8 8.61 10.7 4.06 4.37 11.1 19.46 Total 99.4 100.8 100.8 101.0 96.4 96.8 100.5 - molar Mg/Si 0.57 0.27 0.37 0.12 0.67 'Sepiocrete'Dorbank Sepiolite 'Ideal' *7 *1: Ebenhaeser, 31 o35' 50''S, 18o16'45''E, road to Papendorp (Ellis, unpub. data). 2: KNERS/5, milled, Mg-saturated. 3: KV196T/2, crushed, Mg-saturated. 4: SK11/E, crushed, K-saturated. 5: SK11/1. 6: KNERS/4. 7: Stoessel (1988) ?Sepiocrete? As the morphology and micromorphology shows, the ?sepiocretes? are sepiolite- cemented soil horizons (with or without accessory silica). Sepiolitic horizons do not form a diagnostic horizon in the World Reference Base (1998), Soil Taxonomy (Soil Survey Staff, 1999), or the South African system (Soil Classification Work- ing Group, 1991). The ?sepiocrete? horizons do not meet the slaking requirements of dorbank/duripan/petroduric horizons. They contain less SiO2 than typical dorbank (Table 1.4), and too little to be classified a ?silcrete? after Summerfield (1983b) and Nash and Shaw (1998). Both silica and sepiolite are present in 39 the matrix, although the degree to which silica and sepiolite dominate seems to vary even within the same horizon. It seems most probable that both contribute to the structural properties of the horizon. Nash and Shaw (1998) outlined the practical difficulties in classifying calcrete-silcrete intergrades in Botswana. They concluded that while an assessment of the matrix chemistry was the ideal, for practical purposes the end members of the calcrete-silcrete spectrum could be defined by bulk chemical analysis (?silcrete? if more than 85 % SiO2 and ?cal- crete? if more than 50% CaCO3), and if a duricrust falls between these values, then to examine the micromorphology to identify any evidence of displacement or replacement. They acknowledged, however, the difficulty in classifying mate- rial where the material has had a complex evolutionary history, or been almost completely replaced. Thus the terms ?sil-sepiocrete? (for a sepiocrete replaced and now dominated by silica) or ?sepio-silcrete? could be employed in the sense used by Nash and Shaw (1998). While the ?-crete? terminology provides a useful expression of the cemented nature of the horizon, to fit the existing soil classification schemes the terms ?sepiolitic? and ?petrosepiolitic? (in the same sense as ?calcic? and ?petrocalcic?) would be appropriate. The terms ?sepiolitic? and ?petrosepiolitic? have the advantage over the ?-crete? terminology that they can be more easily be applied as adjectives to other hardpans where sepiolite is significant but not necessarily cementing, such as ?sepiolitic petrocalcic?. This is particularly relevant since sepiolite is closely associated with calcite in Namaqualand soils (M.L. Francis, unpub. data), and is a common mineral in the clay fraction of calcretes (for example Netterberg, 1969; Hay and Wiggins, 1980; Watts, 1980). The term ?sepiolitic? should be used for horizons which: (a) contain sepiolite in amounts great enough for it to be detected by XRD in the bulk soil; (b) have peds (a fractured surface and not just the cutan) that cling strongly to the wetted tongue; and (c) cause methyl orange to turn from orange to purple-pink over most of a fragmented surface. 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? (of Summerfield, 1983a; Partridge and Maud, 1987), a silica-cemented horizon which does not fit the definition of petroduric. 40 Sepiolite has an influence on the soil properties, and it is suggested that the term ?sepiolitic? be used where the sepiolite content is greater than an accessory phase in the clay fraction. Sepiolite is abundant in coastal Namaqualand soils (Francis et al., 2004) and palygorskite is also present (Singer et al., 1995). Neaman and Singer (2004) outlined the effects of palygorskite, including ?palycrete?, on soil properties. Sepiolite has a high water holding capacity (Jones and Galan, 1988). When added to lime mortars, this causes a decrease in the rate of the carbonation process by provoking a decrease in the free water content in the porous system of the mortar, which impedes CO2 dissolution, a rate-controlling step of the carbonation process (Mart??nez-Ram??rez et al., 1995). Further work is needed to establish this effect on soils, it may be similar to the effect of evapotranspiration on calcite precipitation: Netterberg (1969) suggested the decrease in pore water pressure as a result of increased suction pressure will bring about CO2 loss and favour calcite precipitation. Francis et al. (2007) speculated on the effect this water-holding capacity would have on plants growing in sepiolite-rich soils, and thus its effect on the ecosystem of arid regions. It has a high plasticity (Bain, 1971), which would affect the geotechnical usefulness of the soil. 1.4 Summary and conclusions 1.4.1 Dorbank (petroduric) The low water permeability (indicated by the amount of time taken for a water drop to penetrate during field tests) of dorbank provides the slow silica-saturated pore water flow rates needed for the shift toward equilibrium and subsequent sil- ica precipitation, and allows the deposition of strongly oriented clay. Illuviation of clay in the dorbank has been a dominant soil forming process: there is a strong orientation of clay parallel to the grains, crescent coatings on the lower part of voids, and in some cases voids are blocked by clay-coated grains. Amorphous material was observed coating some clay grains. There also appears to be a neo- formed Mg-rich smectite in horizon KV196T/3. Although the dry state and red colour of the dorbank profile suggests it is well-drained, a change to a massive structure at this depth is accompanied by an EC maximum. Together with the presence of a sponge spicule, this suggests a periodically water-saturated environ- ment, the composition of which is favourable to Mg smectite clay neoformation. 41 The high pH of the upper horizons could have increased the solubility and mo- bility of silica, and the sodic environment have contributed to clay dispersion and rise in pH due to the hydrolytic exchange of Na+. Under the low rainfall conditions, the soluble silica will not leach through the profile but accumulate in the subsoil. The abundance of illuviated clay in the pore spaces may have provided a large adsorption area for silica, since the silica is generally (although not exclusively) concentrated on the void side of the clay coatings. 1.4.2 ?Sepiocrete? The micromorphology of typical ?sepiocrete? is distinctly different from dorbank horizons, the most obvious being the granular microstructure, the micrite-rich matrix, and the lack of crescent clay coatings. It is sometimes difficult to distin- guish sepiolite from silica with certainty under the optical microscope, given the low birefringence of sepiolite. Sepiolite appears to form the matrix areas, whereas the silica is localised. The interlocking, subparallel morphology of sepiolite over much of the sample may contribute to the induration. In contrast to the dorbank samples which slaked in NaOH even if calcareous, neither HCl nor NaOH treatments resulted in slaking of the ?sepiocrete? samples. Partial slaking was observed when alkali/acid treatments were alternated. This may suggest that the cement in ?sepiocrete? is not solely amorphous silica, or that the presence of sepiolite could have cause a reduction in pH sufficient to reduce the solubility of silica. A weathering profile silcrete also did not slake in concentrated NaOH under the same experimental conditions, which suggests that in addition to sepiolite content, the degree of silica polymerisation and/or impregnation and/or total amount of secondary silica also plays a role in dictating whether slaking of a particular sample will occur. The ?sepiocrete? horizons therefore do not meet the slaking requirements of dorbank/duripan/petroduric horizons. They contain less SiO2 than typical dorbank, and too little to be classified a ?silcrete?. 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 sepiolite and silica contribute to the structural properties of the horizon. 42 The amorphous silica appears in some places to be overlying the sepiolite fibres which suggests that sepiolite precipitated first. In other areas the silica seems to be associated with mineral grains. The timing and precise mechanism of sil- ica precipitation in the ?sepiocrete? is difficult to resolve. It seems unlikely that silica precipitation was caused by the initial precipitation of sepiolite depleting the solution in Mg, since Mg is present in excess of the amount of silica required to form sepiolite (using an Mg/Si ratio of 0.67), and thus will not be depleted first during sepiolite precipitation. Etched quartz grains in both ?sepiocrete? pro- files suggests the local dissolution and redistribution of silica within the profile, perhaps prompted by a (local) rise in pH due to the hydrolytic exchange of Na+. Clay dispersion and migration are not as prominent as in the dorbank horizons, possibly a function of their differing profile chemistry. The pH of sepiolite-bearing horizons is consistent with that for sepiolite formation. The calcic pendants and cappings in the ?sepiocrete? from the Kners profile indicate the mobilisation of calcite, and suggest the profile has a complex evolutionary history. It could be that the ?sepiocrete? resulted from the (partial) silicification of what was origi- nally a sepiolite-rich calcic/petrocalcic horizon. In profile SK11 the calcic hori- zons occur below the sepiolite-rich and ?sepiocrete? horizons, suggesting leaching of calcite to the lower part of the profile, but the horizon no longer fits into a ?cal- crete? category, and can neither be classified as a dorbank/duripan/petroduric, nor a silcrete horizon. Given the physical properties of sepiolite, and its effect on physical properties of a soil, its presence should be recorded during soil classification. 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. 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 slaking requirements of a petroduric horizon. 43 Chapter 2 Soil formation in sediments of the Namaqualand coastal plain 2.1 Introduction There has been a surge of interest in arid regions, particularly their biodiversity and ecology. Namaqualand itself has been the focus of much research recently, ranging from publications such as Cowling and Pierce (1999) to a Special Issue of the Journal of Arid Environments. Although there is eighty or so years of sedimentological and paleontological research that has been conducted in the di- amond mining areas along the west coast of southern Africa (Pickford and Senut (2000) and Pether et al. (2000), for example), there is very little information available on the soils. This is becoming increasingly relevant since mining activi- ties in Namaqualand result in large ?overburden dumps?, formed by the inversion and mixing of the usual regolith profile (commonly tens of metres deep). Soil forming processes have begun anew on the surfaces of these dumps, as previ- ously deep and inaccessible layers of the regolith are now becoming important for land use. This has prompted further studies (including Schmidt, 2002; Ma- hood, 2003; Blood, 2006; Ndeinoma, 2006), in particular the work of Prinsloo (2005). More details on the effects of the mining processes on the soils and veg- etation, and the challenges of the rehabilitation processes are given by Carrick and Kru?ger (2007). 44 Information on coastal Namaqualand soils is mostly limited to Land Type Surveys (Land Type Survey Staff, 1987) and unpublished studies done for mining com- panies, focused on the saline-sodic nature of the soils and the associated re- habilitation difficulties, irrigation reports, and part of a study of Karoo soils by Ellis (1988). An exception is a study by Singer et al. (1995) that detected the fibrous clay minerals sepiolite and palygorskite in the clay fraction. Recently, Francis et al. (2007) summarised the available information and key characteris- tics of Namaqualand soils, and Chapter 1 described silica- and sepiolite-cemented horizons in Namaqualand. The aim of this Chapter is to describe and characterise the general soil form- ing features within a regolith similar to that described by Pether (1994), from the northern part of the Namaqualand coastal plain (Alexander Bay Formation (SACS, 1980)). The mining excavations tens of metres deep down to bedrock have provided an opportunity to examine soil formation in a regolith that has been subaerially exposed in some parts for much of the Neogene. Conventionally soil pits are dug and described in South Africa to a depth of about 1.5m, below which the importance of a soil for most agricultural purposes diminishes sharply (Soil Classification Working Group, 1991). However, Fey et al. (2002) suggested that the characterisation of the entire regolith before disturbance, assists in understanding the new pedogenic processes forming on the surfaces of mining dumps. Stone and Comerford (1994) reviewed biological activity far below the 1.5m soil depth; and Moody and Graham (1994), in deep sands on a marine terrace in California, argued that there is a zone of active pedogenesis in the basal unit of the regolith just above the bedrock, expressed by illuviation and precipitation of clays, iron oxides and silica. Excavations of the regolith to bedrock during mining on the Namaqualand coastal plain provided many excellent exposures of these processes, both near-surface and deeper in the regolith. Studying the deeper layers of the regolith have helped to distinguish paleosols (Nettleton et al., 2000), which have served as the ?record keepers? of the changing conditions at the soil surface. 45 2.2 Site description Sampling was done in two of the mining areas belonging to De Beers Consolidated Mines Ltd - Namaqualand Mines. These are the Buffels Marine Complex (BMC) in the Kleinzee-Port Nolloth area, and the Koingnaas Complex (KNC) further south in the Koingnaas-Hondeklip Bay area (Figure 2.1). Figure 2.1: Location of field area (grey). BMC: Buffels Marine Complex; KNC: Koingnaas Complex. The escarpment forms the eastern edge of the coastal plain. Modified from Ellis (1988). 2.2.1 Climate The key influences of climate in soil formation are moisture and temperature. Temperature affects the rate of chemical reactions in the decomposition and synthesis of minerals and in biological processes (White, 1979). 46 Temperature In Namaqualand, the prevailing winds are very strong southerly (occasionally gale force), with occasional berg (east) wind conditions, especially during win- ter. Temperatures increase markedly during berg wind conditions, which may persist for longer than a week. As a result, the average maximum temperature from February to October (late summer to spring) is over 30o C, with April and October experiencing temperatures of 38o C. In contrast, the average maximum temperature from November to January is 25o C (late spring to summer). Mini- mum temperatures from March to October are around 6o C, with July dropping to 3o C. From November to April the minimum is around 10o C (D.B.C.M., 2000, data collected at Kleinzee, July 1995 to March 1997). Moisture White (1979) noted that the effectiveness of moisture in soil forming processes is dependent on i) rainfall (the intensity, duration and seasonal distribution), ii) the evaporation rate from the surface and from within the soil, iii) the topography of the land surface, and iv) the soil and rock mass permeability. Rainfall The Namaqualand coastal plain receives less than 150 mm winter rainfall an- nually, with the figure increasing southwards. Crucially for both plant growth and water movement through the soil, the climate is characterised by highly reliable rainfall when compared to other arid regions with similar mean annual precipitation (Desmet, 2007). It typically arrives as widespread, gentle show- ers (average rainfall event 6 mm, P. Carrick unpublished data, cited by Desmet (2007)). Drought conditions are rare, and rainfall is higher than average about once every 10 years, causing ephemeral rivers to flow. Flood events occurred in the Buffels River, for example, in 1945, 1961, 1962, 1963, 1976, 1980, 1986, 1996 and 1997 (D.B.C.M., 2000). Coastal fog adds significantly to the total precipitation. It can be observed to extend inland to the rise of the escarp- ment, occurs frequently along the coastal plain at elevations below the 200 m contour line (Olivier, 2002). Data from the Pilot Fog Study at Kleinzee for the 47 period July to October 1995 showed that the highest precipitation due to fog occurred during August. The total fog recorded during this period was 1001 mm (D.B.C.M., 2000). Tinely (1985; in D.B.C.M., 2000) estimated that the plants may trap up to 300 mm of fog per annum. Paleoclimates of the northern Namaqualand region are not that well constrained, but the consensus for the greater area seems to be that is was wetter during the Last Glacial (Van Zinderen Bakker, 1976; Tankard and Rogers, 1978; Parkington et al., 2000). Evaporation That evapotranspiration greatly exceeds precipitation in Namaqualand is re- flected by a vegetation cover (Sandveld bioregion (Desmet, 2007)) of approxi- mately 60%, as well as the presence of saline pans and ephemeral rivers, and areas with salt efflorescence and crusts at the soil surface. The evaporation rate is of the order of 2 m per year. The evaporation rate is higher than expected for a coastal area due to the wind regime, but is reduced by the regular occurrence of coastal fog (D.B.C.M., 2000). Topography The coastal plain is a flat to very gently undulating pediplain truncated by the Great Escarpment (approximately 1500 m a.m.s.l.). There is a small escarp- ment on the eastern edge of the mining area, with a maximum altitude of 213 m a.m.s.l.There are two rivers in the BMC: the Buffels River with subsurface flow (Heinecken, 1981) which forms the southern border of the BMC; and the ephemeral Kwakanaab River, which ends in the Dreyer?s Pan area. Dreyer?s Pan is one of the large saline pans in the BMC, others are Bloupan, Karaspan, and Soutpan. The ephemeral Swartlintjies River runs through the KNC. The locations of these features are shown in Figure 2.2. Microrelief ?Heuweltjies? (hillocks) occur throughout Namaqualand and along the western 48 (a) Buffels Marine Complex (b) Koingnaas Complex Figure 2.2: Figure 2.1 magnified to show location of sample sites (circles) within Buffels Marine Complex (BMC) and Koingnaas Complex (KNC) at De Beers Namaqualand Mines. Farm boundaries and position of rivers and pans indicated. and southern Cape coasts (Picker et al., 2007). They are distinguishable in the field and on aerial photographs as circular features (diameter usually 10 - 20 m) showing a different vegetation pattern and a slightly raised (1 - 2.5 m) surface, and they cover most of the field area. They are thought to be termi- taria of the harvester termite Microhodotermes viator (Moore and Picker, 1991), with debated contribution from mole rats (Lovegrove and Siegfried, 1986; Cox et al., 1987; Lovegrove and Siegfried, 1989; Laurie, 2002; Midgley and Hoff- man, 1991) and/or another species of grassland termite in older (30 000 years B.P.) heuweltjies (Midgley et al., 2002) such as Hodotermes mossambicus (Moore and Picker, 1991). Soils within the heuweltjie generally tend to be more alkaline (sodic and calcare- ous) and enriched with silica (including silica- and calcite-cemented lenses) than intervening soil (Ellis, 2002). This suggests they are drier than the intermound area and is consistent with their raised topography (Francis et al., 2007). 49 Soil Permeability Hydrophobicity and crusting at the surface of Namaqualand soils is common. These phenomena affect the soil water balance by enhancing preferential flow (Francis et al., 2007). Soil permeability is also affected by the presence of tex- tural barriers and cemented horizons at depth, strongly influencing subsurface water storage and its provision to plants (Francis et al., 2007). 2.2.2 Regolith The regolith pertaining to this study comprises successive late Tertiary marine packages, each deposited during sea-level regression (Pether, 1994). Some rest on kaolinized paleochannel sediments, described in detail by Pether (1994), who con- cluded that they were laid down as a quartzo-feldspathic sediment in a fluvial environment, and subsequently deeply weathered, with extensive alteration of feldspar to kaolinite. These sediments are described in greater detail by Tankard (1966), Tankard (1975), Pether (1994), and Pether et al. (2000). De Villiers and Cadman (2001) favoured an early Tertiary age for the kaolinized paleochannel north of the Swartlintjies River, although it may also contain reworked Creta- ceous material (Pether et al., 2000; De Villiers and Cadman, 2001). The marine packages are arranged en echelon down the bedrock gradient, from oldest and highest inland to youngest and lowest at the coast. Each package is named after the elevation of its transgressive maximum as represented in the Hondeklip Bay area (Pether, 1994; Pether et al., 2000). The 90 m package is ca. 18 - 16 Ma, the 50 m package early Pliocene, and the 30 m package not well constrained, but ca. 3.3 Ma or younger (Pether et al., 2000). The 30 m package is transgressed by younger littoral deposits up to about 10 m a.m.s.l. (Pether et al., 2000). The sand fraction is dominated by quartz and feldspar, lesser glau- conite and phosphatic shell fragments, and variable amounts of heavy minerals (garnet, magnetite, ilmenite, biotite, sphene, amphibole, epidote, kyanite (rare) and zircon (rare)) (Pether, 1994). The marine packages are capped by recent aeolian deposits. Coastal dunefields originate at the mouth of the rivers and extend inland in a northerly direction. In places, older soil profiles are overlain by aeolian sands. Soil formation has taken place in aeolian, ?sheetwash?, and marine-deposited sediments. 50 2.2.3 Bedrock geology and morphology The bedrock geology in the vicinity of Kleinzee (Buffels Marine Complex, Figure 2.1) consists of Stinkfontein Formation (Gariep Belt) metasediments, mainly quartzite, phyllite, schist, arkose and minor intercalated volcanics. Amphibolite bands, quartz veins and small pegmatite intrusions are fairly common (Hartnady and Von Veh, 1990; Rogers et al., 1990). Bedrock morphology is characterised by a number of emergent wave-cut terraces (Rogers et al., 1990), on top of which are the marine sediments. The oldest cliff-line is at 95 m a.m.s.l., cut into silcrete- capped, deeply kaolinised bedrock ?pallid zone? (Partridge and Maud, 1987). The Koingnaas Complex (KNC) near Hondeklip Bay (Figure 2.1) consists of Namaqua Basement gneiss and granite-gneiss. The absence of linear terraces in the KNC is attributed to the complex bedrock morphology of the Basement gneiss (Rogers et al., 1990). Palaeochannels have incised the local bedrock, and are filled with kaolinized fluvial sediment (Pether, 1994). Silcrete and the ?pallid zone? sediments also occur in the Hondeklip-Koingnaas region. Pether (1994) found the silcretes near Hondeklip Bay to be petrographically and geochemically similar to the high-TiO2 weathering profile silcretes of Summerfield (1983c), which have been interpreted as markers of the ?African Surface? by Partridge and Maud (1987; 2000). 2.3 Methods Fieldwork took place in February 2002. Fifteen profiles to bedrock were exam- ined and sampled from south of Hondeklip Bay to north of Kleinzee (details of the dorbank (petroduric) profile KV196T are presented in Chapter 1). The sam- ple sites are shown in Figure 2.2. Samples were named according to the mining excavation to which they were closest, the prefix being the farm name. Sam- ple sites were selected to be representative of each terrace and marine package. These are the 50 m and 30 m packages in the Koingnaas area, and the 30 m, 50 m, 65 m and 90 m terraces in the Kleinzee area. Samples (approximately 2 kg) of each horizon were taken from the sidewalls of the mining excavations to bedrock, and from soil pits dug in the undisturbed areas nearby. The soil morphology was described in the field. A 10% solution of HCl was used to test for carbonate, and 51 30% H2O2 for manganese. Colour was described using a Munsell colour chart (Munsell Color Company, 1975). Root content was estimated visually. Horizons falling within 1.5 m of the surface were treated as diagnostic horizons for classifi- cation purposes. Those occurring deeper in the profile were described, but treated as non-diagnostic and no consideration given to their position in the profile for classification purposes. Eluvial, bleached horizons occurring in non-diagnostic positions were referred to as ?gleyed sand?, cemented carbonate horizons were referred to as ?hardpan carbonate? when in the diagnostic position, otherwise as ?calcrete?. Terrain units were described according to the definition in Soil Clas- sification Working Group (1991). Most of the profiles fall into the ?footslope? or ?valley bottom? units, which in this landscape is the coastal pediplain. Horizons were assigned to either aeolian, upper shoreface or lower shoreface facies parent material. Near-surface aeolian facies were further subdivided where there was a prominent lithological discontinuity. Fluvial material may have been present, but field evidence was not definitive. Munsell colours were manipulated to arrive at a single value that expressed the redness of each sample for easy comparison. A redness rating (Rr) was selected from Alexander (1985) and Arduino (1985): Rr = (K ?H)CV ?1 (2.1) with K set to 25 (Arduino, 1985), C the Munsell chroma, and V the Munsell value. H is the hue index (Munsell Color Company, 1975) obtained by adding 0 to the number preceding R, 10 to that preceding YR and 20 to that preceding Y in the Munsell colour notation. With K set to 25, all samples redder than 5Y have a redness rating greater than zero. All samples (i.e. each horizon in every profile) were air-dried, crushed, passed through a 2 mm sieve and weighed. The >2 mm cement was sorted from the >2 mm pebbles and recrushed, to minimise shattering of pebbles/gravel. The <2 mm fraction was used for chemical analyses. The pH (in water) and electrical conductivity (EC) were measured in a 20 g to 50 mL soil-water mixture. Thirty percent H2O2 was used to oxidise selected marine sands (6 samples), to check whether alternating colour layers within the unit were a result of oxidation. The uppermost horizon of eleven profiles was selected for organic carbon and nitrogen analysis, milled to a fine powder and analysed in a Eurovector Elemental Anal- 52 yser. Slaking tests were devised to distinguish silica- from carbonate-cemented hardpan horizons. These tests were based on the WRB (1998) definition of petro- duric, petrocalcic and fragic horizons. Each sample was submerged in water, 5M HCl or 6M NaOH and gently heated on a waterbath for four days. Samples which did not slake were then subjected to alternating 5M HCl and 6M NaOH treatments until slaking occurred. 2.4 Results and discussion Depositional processes on the Namaqualand coastal plain have resulted in the sediment being arranged in horizontal strata. ?The most fundamental and, possibly, the only real difference be- tween soil and other unconsolidated geological material is that, in the case of soils, the materials have been organized by natural, non- depositional processes into horizons.? (van der Eyk et al., 1969, p43) In some cases the sedimentary features have been completely obscured by pe- dogenesis, and it was difficult to determine in the field whether the horizontal layers were depositional or pedogenic in origin. The regolith characteristics are discussed below from the surface to bedrock. A generalized regolith profile is shown in Figure 2.3. This is for a calcareous soil profile. All the horizon sequence permutations and their descriptions are in Table 2.1. The location, composition, cement, and genesis of profiles SK11 (petrosepiolitic) and KV196T (petroduric, dorbank) in the BMC (Figure 2.2) are discussed in Chapter 1. Profile photographs are presented in Appendix A. Unpublished reports on rehabilitation difficulties at De Beers Namaqualand Mines have noted the highly saline-sodic nature of the base of the soil profiles, com- pared to the upper horizons and suggested this was caused by the gradual leach- ing of salts through the regolith over time (Roux and Odendaal, 1992; Scott and Johnston, 1994; Scott et al., 1995). This was mostly based on a studies of the disturbed ?overburden dumps? which form a regolith pile that is inverted from its usual horizon sequence (i.e. deepest layers on top). Table 2.1 shows this to be generally true (AK1, AK61H), but also that the individual horizon charac- 53 ::::::::::::: ? ? ?. ??? ???. ??? :::::::::::: ? ? ? ? ???????. ???.??? Regic sand (calcareous) Unconformity; deflation surface Neocarbonate B (reddened) Stoneline Calcrete; dissolution feature ?Blocky calcretized sediment? Upper shoreface facies (non-calcareous) Lower shoreface facies (non-calcareous) Bedrock ? ? ?. ??????.. .. .. ....???????? ? .. ..:::::::::::::???? ~1m ~3m ~6m ~9m Figure 2.3: Sketch of generalized regolith profile teristics are more important than absolute depth in determining the EC. The dorbank, calcrete and clay-rich horizons show elevated EC values relative to the horizons above and below. This gives an EC rise, usually around the middle of the profile, where these horizons occur in the profile (KV220PRN, LCK1-5, SN30, SNT60, TP231L. TP266Q). 2.4.1 A horizons Where the regic sands are less than 500 mm thick and lie over a classifiable buried soil, this soil has been classified. Where the regic sand is more than 0.5 m deep, it has been classified as the Namib soil form (Soil Classification Working Group, 1991). Soil profiles commonly have a crust on the surface, up to 3 mm thick. The crust appears to consist of an algal mat and deflocculated clay particles cementing the upper part of the A horizon. These were described by Francis et al. (2007). A horizons are darkened by organic matter, which also imparts a strongly hy- drophobic character to the horizon, some taking in excess of 20 seconds for a drop of water to penetrate (Table 2.1). Orthic A horizons are present at the surface where there are no regic sands, and are redder and darker than the regic sands, and are generally 0.2 to 0.3 m thick. Some show evidence of a bleaching in the A horizon (DL88, Table 2.1), a feature noted to occur commonly in South African arid soils (Ellis, 1988). 54 T abl e 2.1 : P rofil e descripti on s (S oi l Classificati on W or kin g G roup , 19 91 ). R oma n numer al s indicat e diff er en t sedime nt ar y units . Profile s picture d in Ap p endi x A . Ab breviat ions : an g.-angu lar ; c. -coarse ; calc.- calcareous ; co nt.- co nti nu ous ; diam.-d ia m eter ; disc. -d is co nti nu it y; f. - fin e; lit h.-lit hological ; med.-m ed ium ; m od .- m oderate ; oc c.- occa sio nal ; qt z-q uartz ; pr is. -p ri smatic ; se d. - sed im en t; se c.-s econ ds ; sl .- slig ht ; v.- ver y. ?- ? ?no t anal ys ed ? ex cep t fo r te xtur e an d consist en ce : no t ap pl ic ab le . DepthFaciesMoist-Field>2 mm-StructureReactHClNaOHH 2 OCementationWaterRootsTransitionOther Features pHEC (m)ureDryMoistTextureFragmentsDryMoistWetHClabsorptionmS/cm 6869: Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). KNC, 30 metre package, 41m amsl (GPS). Terrain unit: footslope. /i0-A1Regic sand (Orthic A) Aeolian duneDry10YR 7/1 10YR 6/1 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky Mod.---None1sCommonAbrupt, smooth, lith. disc. 9.7- /ii0.45-A2Regic sandAeolian deflation Dry10YR 7/1 10YR 5/1 Med. sand Few, rounded gravel ApedalLooseLooseNon-plastic, non-sticky Strong---None1sCommonAbrupt, smooth, lith. disc. 9.8- /iii0.5-II B1Neocarbonate BObscuredV. sl. moist 10YR 6/4 10YR 5/2 Med. sand NoneApedalLooseLooseV. sl.plastic, v. sl.sticky Strong---None1sCommonAbrupt, wavy9.2- /iv0.6- 0.8 II B2Neocarbonate B over hardpan carbonate ObscuredMoist10YR 5/6 10YR 5/6 Loamy sand Rounded discoid cobbles (stoneline) Weak, c., subang. blocky Sl. hard Sl. firmPlastic, sl. sticky Strong---Sl., vesicular, carbonate 2sCommonAbrupt, wavy, stoneline C., distinct mottles, 7.5YR 6/8 dry, 10YR 5/6 moist; few clay, carbonate cutans 8.8- 6869:New profile, mining excavation, upper material removed. Askham Form (Orthic A - Yellow-brown apedal B - Hardpan carbonate). /4?-0.32III BYellow-brown apedal B ObscuredDry10YR 5/6 10YR 5/4 Loamy sand V. few, c. gravelWeak to mod., c., ang. blocky Sl. hard LooseSl. plastic, sl. sticky None---None3sFewAbrupt, smoothCommon, v. f. macropores8.61.1 /50.32-IV C1Hardpan carbonate Upper shoreface Drywhite10YR 8/2 Loamy c. sand Many, rounded gravel Mod.-strong ang. blocky Hard- v.hard Loose- v. firm Sl. plastic, sl. sticky Strong544Mod.,laminar- blocky calcite 3sNot detected Diffuse, tonguing9.10.45 /61-IV C2Marine depositUpper shoreface Dry7.5YR 5/6 7.5YR 5/6 C. sandMany, rounded gravel ApedalLooseLooseNon-plastic, non-sticky None---None1sNot detected Abrupt, smooth8.10.11 6m-9m unit consists of alternate coarser, redder & finer, yellower sedimentary layers. A coarser layer (/8) is paler & richer in clay (/8a). Lowest finer olive yellow layer (/7) contains thin iron pan ~3mm thick (/7a). Numerous vertical termite burrows. /76-9V C3Marine depositLower shoreface Dry2.5Y 6/6 10YR 5/6 Med. Sand V. few gravelApedalLooseLooseNon-plastic, non-sticky None---None1sNot detected Abrupt, smoothCommon, f., prominent black mottles 7.70.15 /7a6-9V C3Marine deposit + iron pan Lower shoreface Dry10YR 6/6 10YR 4/6 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky None---None1sNot detected Abrupt, smooth8.00.52 /86-9V C3Marine depositLower shoreface Dry5YR 5/8 2.5YR 4/6 C. sandMany, c. gravelApedalLooseLooseNon-plastic, non-sticky None---None1sNot detected Abrupt, smooth9.20.05 /8a6-9V C3Marine deposit + illuvial? clay Lower shoreface Dry5YR 8/2 10YR 7/4 Loamy c. sand Many, c. gravelApedalLooseLooseNon-plastic, v. sl. sticky V. sl.---None1sNot detected Abrupt, smooth7.80.37 9-18m more homogenous texture & colour than 6m-9m. Textural layering not associated with prominent colour contrast. Iron pan ~3mm thick (/12) runs horizontally between overlying coarse /13) & underlying fine layer (/11). Abun.vertical termite burrows (/10). /139-18VI C4Marine depositLower shoreface Dry10YR 6/4 10YR 5/4 C. sandNot detectedApedalLooseLooseNon- plastic, non- sticky None---None1sNot detected Abrupt, smooth7.90.05 /119-18VI C4Marine depositLower shoreface Dry10YR 8/1 10YR 4/6 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky None---None1sNot detected Abrupt, smooth8.80.05 /129-18VI C4Iron panLower shoreface Dry10YR 6/8 10YR 6/8 Med. sand Not detectedNone---None1sNot detected Abrupt, smoothNon-calc. shell fragment in iron pan 8.60.06 /109-18VI C4Termite burrowsLower shoreface Dry10YR 6/8 7.5YR 4/4 C. sandNot detectedApedalLooseNon-plastic, non-sticky None---None1.5sNot detected Termite burrow7.50.44 Clay samples not collected in situ. Occ. roots. Two types of channel clay distinguished: coarse channel clay (6869/CCa), fine channel clay (6869/CCb). There is water dripping out of the clay at approx 2-3m above the bedrock (~30m deep). /cc18-36V C5Gleyed clay (kaolinite) Wet5Y 8/1 2.5Y 7/2 Sandy clay Not detectedC., ang. blockyHardFri- able Plastic, sticky None---None3sNot detected Occ. Roots8.90.42 /ccb18-36V C5Gleyed clay (kaolinite) Wet10YR 7/1 ClayNot detectedC., ang. blockyHard- v.hard FirmPlastic, v. sticky None---None3sNot detected Slickensides; adheres to wetted tongue; occ. roots 8.60.55 AK1/trig: Oakleaf Form (Orthic A - Neocutanic B - Unspecified). Trig beacon near AK1. BMC, 90m terrace, near pallid zone, 119m amsl (GPS). Terrain unit: crest. /10-A1Orthic A, bio crust AeolianDry7.5YR 5/5 7.5YR 3/4 Loamy sand Not detectedLooseLooseSl. plastic, non-sticky Sl.---None6sCommonClear, wavyBiological crust8.91.4 /20.05-A2Orthic AAeolianDry7.5YR 5/5 7.5YR 3/4 Loamy sand Not detectedMod., blockyLooseLooseSl. plastic, non-sticky None---None4sCommonClear, wavy8.41.5 /30.3- 0.6 BNeocarbonate BAeolianDry7.5YR 5/4 7.5YR 3/4 Loamy sand Not detectedWeak med. blocky LooseLooseNon-plastic, non-sticky Occ.---None3sNot detected Abrupt, wavy, to bedrock 8.11.9 AK1: Mining excavation, upper material removed. Trawel Form (Orthic A - Neocarbonate B - Dorbank). Terrain unit: crest. /10- II BNeocarbonate BObscuredDry7.5YR 5/4 7.5YR 4/4 Loamy sand Not detectedMod., med. to c., ang. blocky SoftLooseSl. plastic, sl. sticky Mod.345None3sFewAbrupt, smooth8.93.5 /20.3- ?0.8 III C1Calc. dorbankObscuredDry10YR 8/2 7.5YR 6/4 -Nodular- Laminar Hard- v.hard Friable- sl.firm Sl. plastic, non-sticky Strong02-Mod. carbonate, Si 5sV. occ. ?Gradual, smoothLocalised calcrete dissolution; tunnels 8.68.9 /3?0.45- 0.6 thin III Soft dorbankObscuredDry7.5YR 5/5 5YR 4/3 -NodularHardFirmNone-v. sl. 24-Sl. Si6sNot detected Clear, wavyWhite spots (gypsum ?)8.84.9 /40.8-III C2DorbankObscuredDry7.5YR 5/6 5YR 4/4 -MassiveHardFirmOcc.03-4-Mod. cont. Si6sNot detected Abrupt, smoothTunnels9.04.1 /51.2-III C2DorbankObscuredDry7.5YR 6/4 7.5YR 4/4 -MassiveHard- v.hard Firm-v. firm None02-3-Mod. cont. Si8sNot detected Clear, wavyWhite spots (gypsum?); Mn spots 8.74.5 /61.6-III C3Calc. dorbankObscuredDry7.5YR 8/1 7.5YR 5/4 -MassiveHard- v.hard Firm-v. firm Strong02-Mod. cont. Si, carbonate 8sNot detected Clear, wavyWhite spots (gypsum?); Mn spots 8.76.7 /7?2-2.310YR 8/2 10YR 8/2 Not detectedV. hardV. firmNon-sticky, non-plastic Strong---Srtong cont. Si, carbonate -Not detected Occ. Mn Fluvial paleo- channel 30cm diam.calc. nest in dorbank HorizonMunsellConsistence Slaking (5=full) 55 T abl e 2.1 : Co nt in ued . DepthFaciesMoist-Field>2 mm-StructureReactHClNaOHH 2 OCementationWaterRootsTransitionOther Features pHEC (m)ureDryMoistTextureFragmentsDryMoistWetHClabsorptionmS/cm HorizonMunsellConsistence Slaking (5=full) AK1:New profile below truncated II B (like AK1/1). Mining excavation, upper material removed. Trawel Form (Orthic A - Neocarbonate B - Dorbank). Terrain unit: crest. /80.3-III C1DorbankObscuredDry7.5YR 6/4 7.5YR 4/4 Not detectedMassiveV. hardV. firmNon-sticky, non-plastic None03-Cont. Si10sNot detected Abrupt, smooth (darker fissure) 8.11.1 /90.9-III C1DorbankObscuredDry7.5YR 5/4 5YR 4/4 Few Fe- concretions MassiveV. hardV. firmNon-sticky, non-plastic None03-Cont. Si>10sFewClear, wavy8.51.8 /101.5-III C2DorbankObscuredDry7.5YR 6/4 7.5YR 4/4 -V. hardV. firmNon-sticky, non-plastic Sl., loc.01-Cont. Si, carbonate >10sFewWhite spots (gypsum ?)8.73.1 /111.8-III C3V. soft dorbank?ObscuredDry10YR 6/6 7.5YR 5/4 -VesicularSl. hard Fri- able Sl. sticky, sl. plastic None---Si?5sNot detected Tunnels7.72.3 AK61H: Coega Form (Orthic A - Hardpan carbonate). BMC, 30m terrace, 28m amsl (GPS). Terrain unit: footslope. /40-AOrthic AAeolianDry10YR 6/3 10YR 4/3 Med. to f. sand Not detectedApedalLooseLooseNon- plastic, non sticky Mod.---None>10s (hyd.)CommonAbrupt, smoothBiological crust8.50.45 /10.35-II C1Hardpan carbonate ObscuredDry10YR 8/1 10YR 6/3 -Laminar-blockyLoose- v. hard Loose- v. firm Non- sticky, non- to sl. plastic Strong40-1-Cont. mod. carbonate 3sFewClear smoothMn present, well developed pseudo-anticline 8.64.9 /21.05-II C3Blocky calc- retized sediment MarineDry7.5YR 6/4 7.5YR 5/4 Not detectedStrong c., ang. blocky Sl. hard LooseNon- plastic, non sticky Cutan---Discont. carbonate 2sNot detected Clear smoothCarbonate cutans8.36.0 /31.55- ?2.1 II C4Marine depositMarineDry10YR 6/4 10YR 5/4 Not detectedStrong, med. blocky Sl. hard Friable- sl. firm Sl. plastic, sl. sticky None---None3sNot detected Abrupt, wavy, to bedrock Salt efflorescence on sidewall of mine excavation, salty taste 8.012 DL88: Augrabies Form (Orthic A - Neocarbonate B). BMC, 65m terrace, 134m amsl (GPS). Terrain unit: midslope. /10-AEBleached A + biological crust AeolianDry7.5YR 5/4 7.5YR 3/4 Loamy sand V. few, c. gravelWeak, f. blocky LooseLooseSl. plastic, sl. sticky None---None3sCommonDiffuse, wavyBio. crust; f. droppings; v. narrow tunnels; macropores 8.70.22 /20.2-BNeocarbonate B AeolianV. sl. moist 7.5YR 5/4 7.5YR 4/4 Loamy sand Not detectedWeak, f. blocky LooseLooseSl. plastic, sl. sticky Mod.---None3sFewAbruptV. narrow tunnels; many macropores; few cutans 8.72.1 /31.5-CDorbank (soft)ObscuredDry7.5YR 6/4 7.5YR 4/4 Loamy sand Few gravelWeak, blockyHardFri- able Sl. plastic, sl. sticky None3-555None10sNot detected V. narrow tunnels; many macropores 10.10.22 KV220PRN: Mining excavation. Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). BMC, 65m terrace, 91m amsl (GPS). Terrain unit: crest. /10-A1Regic sand (Orthic A) AeolianSl. moist 10YR 6/4 10YR 5/4 Med. sand Not detectedApedalLooseLooseNon-sticky, non-plastic Strong---None1sFewAbrupt, smooth, lith. disc. Aeolian stratification9.70.08 /20.18-A2Regic sand AeolianSl. moist 10YR 5/4 10YR 4/4 Med. sand Not detectedApedalLooseLooseNon-sticky, non-plastic Sl.---None10sFewAbrupt, smooth, lith. disc. Aeolian stratification; Root tunnels 9.50.20 /30.4-B1Neocarbonate BAeolianSl. moist 10YR 5/4 10YR 3/3 Med. sand Not detectedApedalLooseLooseNon-sticky, non-plastic V. sl.---None30sCommon (old) Gradual, smoothFine droppings, partially decomposed OM, darker 8.80.60 /40.7-B2Neocarbonate BAeolianSl. moist 10YR 6/4 10YR 5/4 Med. sand V. hard (silica?) nodules ApedalLooseLooseNon-sticky, non-plastic Strong---None1sFewAbrupt9.00.44 /5?1.2-B3Hardpan carbonate AeolianDry10YR 7/3 10YR 6/4 Med. sand Not detectedLaminarV. hardV. firmStrong50-1-Discont.,mod. carbonate 8sNot detected Clear, smooth8.93.3 /6?1.4- 2.3 B4NeocarbonateAeolianDry10YR 6/4 10YR 5/6 Med. sand Not detectedSingle grain apedal LooseLooseNon-sticky, non-plastic Strong---None1sNot detected ObscuredAbun. terrestrial bioturbation, calc. tunnels 8.91.1 /7?1.4-B4NeocarbonateAeolianDry10YR 6/4 10YR 4/6 Med. sand Not detectedSingle grain apedal LooseLooseNon-sticky, non-plastic Strong---None1sNot detected ObscuredAbun. terrestrial bioturbation9.90.35 /92.3-II B5NeocarbonateObscuredMoist10YR 6/4 10YR 5/4 Suspended stoneline Single grain apedal LooseLooseNon-sticky, non-plastic Strong---None1sNot detected Abrupt,smooth stone line (ang.qtz.) Snail shells; contains non-calc. termite nest (sample /8) with calc. tunnels 9.90.19 /102.8-III CCalcrete grading to calc.blocky sed. ObscuredDry10YR 7/4 10YR 6/4 Common stones, conc. upper boundary LaminarHardV. firmStrong---Mod., cont. carbonate 4sNot detected Clear, smoothCalc. tunnel8.52.6 /113.8-III C2Marine depositUpper shoreface Dry10YR 5/6 10YR 4/6 Med-c. sand Not detectedMod., c., ang. blocky SoftLooseNon-plastic, sl. sticky None---None1sNot detected Abrupt, smooth (erosional) Tabular cross-bedding; abun. narrow tunnels; U- & vertical calc. burrows 8.11.0 /125-III C3Marine depositUpper shoreface Dry10YR 6/4 10YR 5/6 Med-c. sand Not detectedMod., c., ang. blocky Loose- soft LooseNon-plastic, sl. sticky None---None1sNot detected Abrupt, wavy, sedimentary Tabular cross-bedding not as clear. Lower part paler from mixing with /13? Calc. burrow 8.50.6 /136.2-IV C4Marine depositAssumed marine Dry5YR 8/1 5Y 7/2 Loamy sand Not detectedMod., c., ang. blocky Sl. hard Fri- able Sl. plastic, sl. sticky None---None1sNot detected Not reachedMn spots; termite tunnels continue from /12 8.41.8 LCK1-5: Mining excavation. Namib Form (Orthic A - Regic sand). KNC, 50 metre package, 30m amsl (GPS). Terrain unit: footslope-valley bottom. /A0-ARegic sand (Orthic A) AeolianDry10YR 5/4 10YR 3/3 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky None---None>10sMany c.Clear, smoothLots of OM (bark)6.90.21 /B0.3-BRegic sand AeolianDry10YR 6/4 10YR 5/4 Med. sand Not detectedWeak, c., ang. blocky LooseLooseNon-plastic, non-sticky None---None1sFew c.Abrupt, smoothV. occ. burrows8.90.14 /11-II B2Neocutanic BObscuredDry10YR 5/6 10YR 5/4 Med. sand Not detectedWeak- mod., c., ang. blocky LooseLooseSl. plastic, sl. sticky None---None1sNot detected Clear, smooth7.31.6 /31.3-II B3Neocutanic, luvicObscuredDry10YR 5/8 10YR 5/6 Med. sand Not detectedMod., c., ang. blocky Sl. hard Fri- able Sticky, plastic None---None3sNot detected Clear, smoothSalty taste; few cutans8.53.6 56 T abl e 2.1 : Co nt in ued . *SK1 1 sla kin g te st s: 2- 3 alterna tin g H C l/ Na O H DepthFaciesMoist-Field>2 mm-StructureReactHClNaOHH 2 OCementationWaterRootsTransitionOther Features pHEC (m)ureDryMoistTextureFragmentsDryMoistWetHClabsorptionmS/cm HorizonMunsellConsistence Slaking (5=full) /41.9-III C1Gleyed sand?ObscuredDry2.5Y 6/4 2.5Y 5/4 Med. sand Not detectedMod., to strong, c., ang. SoftFri- able Sticky, plastic to v. plastic None---None2sNot detected Abrupt, smoothSalty taste; few slickensides6.66.6 /52.2-IV C2Sedimentary? Localised ObscuredDry2.5Y 6/3 10YR 5/4 Med. sand Not detectedStrong-v. strong HardSl. firmSticky, plasticNone---None2sNot detected Abrupt, wavy6.83.1 /62.4-V C3Sedimentary?ObscuredDry10YR 5/6 10YR 5/4 Med. sand Not detectedMod., c., ang. blocky Sl. hard LooseSticky, plasticNone---None1.5sNot detected Diffuse, smoothSalty taste6.56.0 /73.4-VI C4Marine depositUpper shoreface Dry10YR 5/6 10YR 5/6 C. sandMany gravelWeak to mod., c., ang. blocky LooseLooseNon-sticky, non- plastic None---None1sNot detected Inter-beddedCross bedding6.91.7 Pale sand (LKC1 5/8), interlayered (wavy, abrupt transitions) with more orange sand (LKC1 5/9).--- /85.4- ~12 VII C5Marine depositLower shoreface Dry5Y 7/4 5Y 4/4 Med. sand Not detectedWeak, c., ang. blocky LooseLooseNon-sticky, non-plastic None---None1sNot detected 7.11.6 /95.4- ~12 VII C6Marine depositLower shoreface Dry2.5Y 6/6 2.5Y 4/4 Med. sand Not detectedWeak, c., ang. blocky LooseLooseNon-sticky, non-plastic None---None1sNot detected Abrupt, to bedrock 7.71.8 OBT: Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). BMC, 95m terrace, 99m amsl (GPS). Terrain unit: crest. /10-rsAeolian deflation surface AeolianDry10YR 6/4 10YR 4/3 Med. sand Few, c. gravelApedalLooseLooseNon-sticky, non-plastic Strong---None10sNoneAbrupt, smoothBiological crust8.70.02 /20.03-AOrthic AAeolianDry10YR 6/3 10YR 4/3 Med. sand Few gravelApedalLooseLooseNon-sticky, non-plastic Strong---None15sManyGradual, smooth8.90.13 /30.25- 0.45 BNeocarbonate BAeolianDry10YR 6/4 10YR 4/4 Med. sand Few gravel, carbonate Weak, f. gran- ular to blocky LooseLooseNon-sticky, non-plastic Strong---None3sNot detected Abrupt, smooth9.00.13 OBT:New profile, mining excavation, upper material removed. --- /40.45-II C1Hardpan carbonate to calc.blocky sed. ObscuredDry10YR 6/3 Not detectedLaminarHard-v. hard Sl.-v. firm Strong---Discont. carbonate 3sNot detected Clear, tonguing8.90.63 /7Within II C1Dissolution or nest? Dry7.5YR 5/6 10YR 5/4 Loamy sand Not detectedMod., med., ang. blocky Sl. hard LooseV. sticky, sl. plastic Local- ized, sl. ---None3sNot detected Abun. tunnels ~3mm diam. & macropores. Dendritic black material (OM filled tunnels?), no reaction in 30%H 2 O 2 (cold) 7.92.0 /51-II C2Sedimentary?ObscuredDry10YR 7/4 10YR 6/4 Loamy sand Not detectedStrong, c., blocky Hard-v. hard Firm-v. firm Occ.---None3sNot detected Mn spots on peds; sepiolite cutans 8.70.01 /61.5- ~2.5 III C3Sedimentary?ObscuredDry10YR 6/4 2.5Y 6/4 Loamy sand Stonelines (sub- ang. qtz, granite- gneiss, silcrete) Strong, c., blocky Sl. hard- hard Fri- able Sl. sticky, non- to sl. plastic None---None4sNot detected Not reached9.20.18 SK11: Pinedene Form (Orthic A - Yellow-brown apedal B - Signs of Wetness). BMC, 65m terrace, 60m amsl (GPS). Terrain unit: footslope. Adjacent to mine /A0-AOrthic AAeolianDry7.5YR 5/5 7.5YR 3/4 Med.-f. sand Not detectedApedalLooseLooseNon- plastic, non sticky None---None4ManyAbrupt, smoothBiological crust, OM, occ. droppings 9.10.14 /B0.2-BYellow-brown apedal B AeolianDry7.5YR 6/6 7.5YR 4/6 Med. to f. sand Few gravelWeak, f. blocky LooseLooseNon- plastic, non sticky None---None2FewClear, smooth9.20.06 /C0.6-EGleyed sandAeolianDry7.5YR 6/4 7.5YR 5/4 Med. sand Not detectedApedalLooseLooseNon- plastic, non sticky None---None1Not detected Abrupt, smooth9.10.08 /D0.65-B2NeocutanicAeolianMoist7.5YR 5/6 7.5YR 4/6 Loamy sand Not detectedWeak, c., ang. blocky Sl. hard LooseSl. plastic, stickyNone---None3FewAbrupt, smoothFew cutans8.71.3 /E0.7-II C1'Sepiocrete'ObscuredNone2-32-32-3NoneNot detected 8.40.03 SK11:New profile, upper material removed by mining. Mining excavation /10.01-II C1'Sepiocrete'ObscuredDry10YR 7/3 10YR 6/4 Not detectedLaminar calcrete V. hardV. firmStrong0 *0-1 *-Si, carbonate, sepiolite >20Not detected Gradual, wavyDiffuse Mn coating some peds8.70.04 /60.45-II C2Blocky calcretized sediment ObscuredDry10YR 7/4 7.5YR 5/6 Loamy sand Not detectedStrong, c., blocky to prismatic Hard-v. hard Firm-v. firm V. strong ---Discont. carbonate 12Not detected Gradual, wavyCarbonate cutans, Mn spots9.10.02 /30.9- 1.1 II C3Hardpan carbonate ObscuredDry10YR 8/3 7.5YR 7/6 Not detectedLaminarV. hardFirm-v. firm Strong2-334Cont. carbonate 12Not detected Clear, wavyMn spots on some peds9.20.02 /90.9- 1.1 III C3Hardpan carbonate ObscuredDry10YR 8/3 7.5YR 7/6 Not detectedLaminarV. hardFirm-v. firm Strong32-32Cont. carbonate 12Not detected Clear, wavyMn spots on some peds8.83.6 /41.1- 1.15 III C4NeocarbonateObscuredDry7.5YR 6/4 7.5YR 5/6 Med. sand Not detectedWeak, ang. blocky LooseLooseNon- plastic, non sticky Strong---None1Not detected 8.90.04 /4b1.15- 1.5 III C5NeocarbonateObscuredDry10YR 7/4 10YR 6/4 Loamy sand Not detectedWeak., c., blocky LooseLooseSl. plastic, sl. sticky Mod.---None1Not detected Clear, wavy, to sample 7 8.72.5 /50.8- 1.5 Termite nestTermite nestDry10YR 6/4 10YR 5/6 Loamy sand Not detectedMod., c., blocky FirmLooseSl. plastic, sl. sticky Local---None3Not detected DiffuseCarbonate cutans, occ. Mn spots; chert pieces? 8.92.6 /71.5- 2.2 IV C6Sedimentary?Dry5Y 7/4 10YR 6/3 Sandy loam Locally abun. gravel Mod., c., blocky FirmLooseSl. plastic, sl. sticky Local---None3Not detected 8.46.3 /102.3- 2.4 IV C7Iron panDry10YR 6/6 10YR 5/6 C. loamy sand Abun. gravelMod., c., blocky FirmLooseSl. plastic, sl. sticky None---None1Not detected Clear, smoothDarker yellow-red areas - iron pan? 7.92.5 57 T abl e 2.1 : Co nt in ued . DepthFaciesMoist-Field>2 mm-StructureReactHClNaOHH 2 OCementationWaterRootsTransitionOther Features pHEC (m)ureDryMoistTextureFragmentsDryMoistWetHClabsorptionmS/cm HorizonMunsellConsistence Slaking (5=full) /11~2.1- 2.6 IV C7Sedimentary?Dry5Y 7/3 2.5Y 6/4 Clay loam Not detectedC., strong blocky HardFriableV. plastic, v. sticky Occ.---None4Not detected AbruptOcc. Mn spots8.55.4 SK11/8 cut into sample 7 & 11, shape suggestive of a channel. Horizontal white layers continuously cut across it and sample 7.---NoneNot detected /8~1.5- 2.1 Channel or biological? ?Channel? or nest? Dry10YR 6/4 10YR 6/6 C. to f. sand Suspended gravel layers Mod., c., blocky LooseLooseNon- plastic, non sticky None---None1Not detected Abrupt, concave up Abun. 5mm diam. tunnels8.21.5 /122.6-VIII CSedimentary?Dry5Y 6/6 5YR 5/4 F. sandNot detectedMod., c., blocky Sl. firmLooseNon- plastic, non sticky None---None1Not detected Abun. 5mm diam. tunnels w. OM 7.93.5 SL4-1: Pinedene Form (Orthic A - Yellow-brown apedal B - Signs of Wetness). KNC, 30 metre package, 46m amsl (GPS). Terrain unit: footslope. /10-0.3A1Orthic AAeolianDry7.5YR 5/6 7.5YR 4/4 Med. sand Not detectedApedalLooseLooseNon-sticky, non-plastic None---None1sCommonGradual, smoothV occ. droppings8.70.04 /20.3-B1Yellow-brown apedal B AeolianDry7.5YR 5/6 7.5YR 4/4 Med. sand Not detectedWeak, f., blocky LooseLooseNon-sticky, non-plastic None---None5sFewGradual, sl. tonguing Few cutans7.70.15 /30.5-EGleyed sandAeolianDry7.5YR 6/4 7.5YR 4/4 Med. sand Not detectedWeak, f., blocky LooseLooseNon-sticky, non-plastic None---None3sFewAbrupt smooth8.40.16 /40.63- ~0.78 B2NeocutanicObscuredDry7.5YR 5/4 7.5YR 3/4 Med. sand Not detectedMod., c., blocky Sl. hard Fri- able V.sl. plastic, sl. sticky None---None3sV. fewAbrupt smoothFew cutans6.70.73 SL4-1: New profile, mining excavation, continuous with horizon B2. Upper material removed. --- /50.8-II B3Slightly Gleyed sand? ObscuredDry10YR 6/4 10YR 4/4 Med. sand Not detectedMod., c., blocky Sl. hard Loose- friable Sl. sticky, sl. plastic None---None5sNot detected Diffuse, wavyCommon macropores; few cutans 7.95.6 /60.95-III B4PedocutanicObscuredDry10YR 6/5 10YR 4/3 Loamy sand Not detectedStrong, c., blocky to pris. Firm- hard Fri- able Sl. plastic, sticky None---None3sNot detected Diffuse wavyFew cutans8.24.5 /70.97-III B5Pedocutanic +calcite ObscuredDry10YR 6/4 10YR 4/4 Not detectedStrong, c., blocky to pris. HardV. firmMod.---None6sNot detected Diffuse wavyMn spots; few cutans8.83.6 /81.22-III B6PedocutanicObscuredDry10YR 7/4 10YR 5/6 Loamy sand Not detectedStrong, c., blocky to pris. HardSl. firmNon-plastic, sl. sticky None---None3sNot detected Diffuse wavyMn spots; few cutans8.84.2 /92.45-IV C1Marine depositUpper shoreface Dry10YR 6/6 10YR 5/4 Loamy sand Many gravelWeak to mod., c., blocky Loose- soft LooseSl. plastic, sl. sticky None---None1sNot detected Abrupt, smooth8.34.0 Abundant vertical termite burrows, some extending through the whole of V, into IV, which appear to have been conduits for water movement. Sediment filling them is browner oranger colour. Burrows look identical to those in 6869. /113.05-9V C2Marine depositLower shoreface Dry10YR 7/1 10YR 7/1 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky Mod.---None1sNot detected Abrupt,smooth to bedrock Local ?groundwater silcrete? ; marine shells 9.21.4 SN30: Mining excavation. Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). KNC, 50 metre package, 68.7m amsl (GPS). Terrain unit: footslope. /A0-ARegic sand AeolianDry10YR 6/3 10YR 5/3 Med. sand Few, rounded gravel ApedalLooseLooseNon- plastic, non- sticky Mod.---None4sFewAbrupt, smooth9.00.13 /B0.3-II B1Neocarbonate BObscuredDry10YR 5/4 10YR 5/6 Med. sand Few, rounded gravel Weak, med., ang. blocky LooseLooseNon-plastic, sl. sticky Mod. strong ---None4sV. fewAbrupt,smooth to tonguing 9.50.17 /20.6-III C2Hardpan carbonate Upper shoreface Dry2.5Y 8/1 10YR 8/2 Not detectedLaminarHardFirmStrong32-33-4Cont., carbonate 5sNot detected Diffuse, tonguing8.91.0 /31.1-III C3Marine depositUpper shoreface Dry10YR 4/4 10YR 5/6 C. sandMany, rounded gravel Weak, c., ang. blocky HardFri- able Sl. plastic, sticky None---None3sNot detected Clear, smooth9.30.19 /5?2-2.8IV C5Marine depositLower shoreface, rip channel? Dry10YR 7/6 10YR 6/4 Med.-f. sand Not detectedStrong, c., ang. blocky HardSl. firmNon- v. sl. plastic, non- v. sl. sticky None---None2sNot detected Abrupt, tonguing to wavy Occ. Mn spots; occ. tunnels9.10.16 /42.1- 2.3 IV C4Marine depositMarine (Local)Dry2.5Y 7/3 2.5Y 6/3 Med.-f. sand Not detectedStrong, c., ang. blocky LooseLooseNon-plastic, non- v. sl. sticky None---None1sNot detected Abrupt, smoothV. occ. tunnels; ? 8cm diam. tunnel on lower contact 9.00.04 /6?2.5- ?4.4 IV C6Marine depositLower shoreface Dry5Y 7/2 5Y 6/4 Med.-f. sand Not detectedStrong, c., ang. blocky LooseLooseNon-plastic, non-sticky Strong---None1sNot detected Abrupt, to bedrock Mn spots; marine shells & bed- ding; abun. tunnels; 1m termite tunnels filled with orange sand 9.90.09 SNT60: Mining excavation. Namib Form (Orthic A - Regic sand). KNC, 50 metre package, 43m amsl (topo map). Terrain unit: footslope. /10-AOrthic A + Regic sand AeolianDry10YR 7/2 10YR 6/2 Med. sand 2 stonelines (buried deflation surfaces) ApedalLooseLooseNon-plastic, non-sticky Mod.---None1sCommon c. Abrupt, smooth, stoneline Aeolian stratification9.70.06 /21.3-II B1Neocarbonate BAeolianDry10YR 6/3 10YR 5/3 Med. sand V. few, rounded, gravel ApedalLooseLooseNon-plastic, non-sticky Mod.---None1sCommon f. Clear, smooth9.50.07 /31.5- 1.77 III B2Neocarbonate B karst centre AeolianDry10YR 5/4 10YR 5/4 Loamy sand V. few, rounded, gravel Weak, med., ang. blocky Sl. hard LooseSl. plastic, sl. sticky Sl.---None1sCommon f. Clear, smoothFew, med., distinct, grey, yellow mottles 9.40.31 /41.77- 1.8 III B2Neocarbonate B karst edge AeolianDry10YR 5/4 10YR 5/4 Loamy sand V. few rounded, gravel, calcite Weak, med., ang. blocky Sl. hard LooseSl. plastic, sl. sticky Strong---None1sCommon f. Clear, smooth9.00.84 /5a~1.5- 1.8 IV C1 Hardpan carbon- ate next to karst ObscuredDry7.5YR 8/2 7.5YR 8/2 Not detectedLaminarHard-v. hard Firm-v. firm V. strong 50---NoneWavy upper boundary shows regular karst features 9.20.53 /5b~1.8-IV C2 Blocky calc. sed. below karst ObscuredDry7.5YR 8/2 7.5YR 8/2 Not detectedMod., c., blocky;laminar Hard-v. hard Firm-v. firm V. strong 4-555Cont. carbonate 5sNot detected Carbonate cutans8.91.1 /5c1.95-IV C3Blocky calc. sediment ObscuredDry7.5YR 8/2 10YR 6/3 Not detectedStrong, c., blocky to pris. Hard-v. hard Firm-v. firm V. strong ---Discont. carbonate 5sNot detected Mn spots on peds8.90.94 58 T abl e 2.1 : Co nt in ued . DepthFaciesMoist-Field>2 mm-StructureReactHClNaOHH 2 OCementationWaterRootsTransitionOther Features pHEC (m)ureDryMoistTextureFragmentsDryMoistWetHClabsorptionmS/cm HorizonMunsellConsistence Slaking (5=full) /5d~2.1- IV C4Blocky calc. sediment ObscuredDry10YR 7/3 10YR 6/4 Not detectedMod., c.blocky to laminar HardFri- able V. sl. sticky, non-plastic V. strong ---Discont. carbonate 5sNot detected Mn spots on peds8.81.1 /7~2.3- 2.8 V C5Blocky calc. sediment ObscuredDry2.5Y 7/3 10YR 5/4 Loamy sand Few, rounded, gravel Strong, c., ang. blocky to columnar HardFirm- friable Sl. plastic, sl. sticky Only cutan ---Discont. carbonate 4sOcc. btw peds Gradual, tonguing Mn spots on carbonate; ped adheres to tongue; sepiolite under calcite cutans 9.11.0 /6Channel into SNT60/7Local channel?Dry10YR 6/3 10YR 5/3 Loamy sand Few, rounded, gravel Mod., c., ang. blocky Sl. hard- hard Loose- friable Sl. plastic, sl. sticky Only cutan ---None3sRoot channels Mn spots on carbonate; bone (terrestrial); 8mm diam. tunnel filled with f. roots 9.20.62 /102.8-VI C6Sedimentary??Dry2.5YR 8/2 5Y 7/3 Loamy sand Occ. pebbleMod., c., ang. blocky? Sl. hard Loose- firm Sl. plastic, non-sticky None---None2sOcc.Gradual, wavyDistinct orange mottles; Mn stained small bone/egg shell 9.60.24 /113.4-VII C7Sedimentary??Dry2.5Y 7/4 2.5Y 6/4 Loamy sand Not detectedApedalLooseLooseNon-plastic, non-sticky None---None2sNot detected Gradual, wavyIron pan; few bones?; Mn staining 9.30.16 /123.7-VIII C8Sedimentary??Dry5Y 8/2 5Y 6/2 Loamy sand Not detectedMod., c., blocky to pris. Sl. hard LooseNon-plastic, non-sticky None---None1sNot detected Abrupt smooth9.90.33 SPNT: Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). BMC, 65m terrace, 57m amsl (GPS). Terrain unit: footslope. /30-A1Regic sand AeolianDry10YR 7/2 10YR 6/2 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky Mod.---None3sFewAbrupt, smoothDroppings9.40.09 /10.1-A2Buried Orthic AAeolianDry10YR 7/2 10YR 5/2 Med. sand V. few gravelApedalLooseLooseNon-plastic, non-sticky Mod.---None10sCommonClear, wavyLots of organic matter, faecal pellets 9.60.10 /20.5- 0.9 II BNeocarbonate BAeolianMoist10YR 6/3 10YR 5/4 Med. sand V. few gravelWeak, mod., blocky LooseLooseNon-plastic, non-sticky Mod.---None3sNot detected Abrupt, smoothOcc. mottles; few cutans9.60.11 SPNT: New profile, mining excavation, upper material removed. /40-III CHardpan carbonate ObscuredDrywhiteBimodal (1&5mm) Not detectedMassiveV. hardV. firmStrong40-Clinging of hammer Not detected 9.20.50 /50.3- 0.85 IV CHardpan carbonate ObscuredDry10YR 8/1 10YR 6/3 VesicularStrong2-50-Carbonate, ?Mn Not detected Dissolution by high salt content (powdery) 9.14.6 TP231L: Kinkelbos Form (Orthic A - E - Neocarbonate B). BMC, 50m terrace, 66m amsl (GPS). Terrain unit: crest. /A0-AOrthic A AeolianDry10YR 6/3 10YR 5/2 Med. f. sand Not detectedApedal to v. weak, med., ang. blocky LooseLooseNon-plastic, non-sticky Sl. to mod. ---None>20sManyClear to gradual, smooth V. thin biological crust9.10.23 /E0.3-EEAeolianDry10YR 7/3 10YR 6/3 Med. sand Not detectedApedal to v. weak, med., ang. blocky LooseLooseNon-plastic, non- v. sl. sticky Strong---None1sFewClear, smooth9.60.12 /B0.5-?BNeocarbonate BAeolianDry10YR 7/3 10YR 6/3 Loamy sand Not detectedWeak, med., ang. blocky LooseLooseNon-plastic, v. sl. sticky Strong---None1sFewNot reached9.60.11 TP231L: New profile, mining excavation, upper material removed. ---None /10-RSRegic sand AeolianDry10YR 6/4 10YR 4/3 Med. sand Not detectedApedalLooseLooseNon-plastic, non-sticky Sl.---None>20sFewAbrupt, smooth9.00.29 /20.6-II B1Neocarbonate BObscuredDry10YR 5/6 10YR 5/4 Loamy sand Not detectedApedal to weak, c., ang. LooseLooseNon- v. sl. plastic, sl.sticky Strong---None1sFewAbrupt, smooth9.70.35 /30.8-III C1Hardpan carbonate ObscuredDry7.5YR 8/1 10YR 7/4 Subrounded qtz gravel LaminarV. hardV. firmStrong3-40-Laminar carbonate 2sNot detected Gradual, tonguing Mn9.21.1 Samples 4 & 5 from same lithology. Sample 4 from upper part and is more intensely bioturbated than Sample 5. This gives strength to the structure. Transition to lower non-bioturbated part is obscured. /42-III C2 Marine depositMarineDry10YR 6/6 10YR 5/6 Sub-rounded gravel Mod. to strong, c., ang. blocky LooseLooseNon-v. sl. sticky, non-plastic Some tunnels ---None2sNot detected ObscuredAlveolar texture - root activity or burrows? 8.11.4 /5at ?15III C3Marine depositMarineDry10YR 5/6 10YR 5/6 Sub-rounded gravel Strong, c., ang. blocky LooseLooseNon-v. sl. sticky, non-plastic Occ. nodules ---None2sNot detected Abrupt, to bedrock Weak bedding/wind erosion?9.60.28 TP266Q: Namib Form (Orthic A - Regic sand). BMC, 95m terrace, 109m amsl. Terrain unit: crest. Mining excavation /50-A1Orthic A + Regic sand AeolianDry10YR 6/3 10YR 5/3 Med. sand Not detectedApedalLooseLooseNon- plastic, non sticky Strong---None>10CommonAbrupt, smoothOcc. fine droppings9.10.14 /41-II B1Neocarbonate BObscuredDry7.5YR 5/4 7.5YR 5/4 Med. sand Ang. qtz pebblesWeak, c., ang. blocky LooseLooseSl. plastic to plastic, sl. sticky Strong---None2V. fewAbrupt, smooth, stoneline (ang. qtz pebbles) 9.90.28 /11.2- 1.7 III C1Hardpan carbonate ObscuredDry2.5Y 8/1 10YR 8/2 Not detectedLaminar-blockyHardFriablePlastic, stickyStrong31-Cont.,laminar, mod.,calcite 5Not detected Gradual, tonguing Mn spots8.34.8 /21.7- 2.5 III C2NeocutanicPossibly marine Dry10YR 6/6 10YR 6/4 Med. sand Not detectedWeak to mod., c., ang. blocky LooseLoosePlastic, stickycutans, burrows ---None3Occ. root channels Abrupt, to bedrock Mn spots, salty taste, carb. cutans 8.42.2 59 Regic sands are young aeolian dunes, which may be vegetated. In the profiles studied they range in thickness from 0.3 to 3.2 m. They are generally medium sand, single-grain apedal, loose, non-plastic and non-sticky. Most are calcareous. pH in water ranges from 8.7 to 9.8 and EC (in the same 1: 2.5 suspension) is 0.02 to 0.29 mS cm?1 (Table 2.1). Average organic carbon in the surface of regic sands is 0.98 %, and is 1.05 % in orthic A horizons. Average nitrogen is 0.03 % (regic sands), and 0.05 % (orthic A). Humification at the surface of these regic sands is the only visible sign of pedogenesis. In some cases there is a lithological discontinuity at the base of the regic sand, manifest either as a heavy-mineral enriched layer (such as profile 6869, Table 2.1) or stoneline (such as SNT60, Table 2.1), interpreted to represent a buried aeolian deflation surface. Aeolian deflation processes in the modern interdune surface areas (profile OBT) are evident from heavy-mineral enriched surface sands and gravel-lag deposits with the same particle-size as the buried stonelines. When compared to profiles with no regic sands, it becomes apparent that the A horizon that developed with the neocutanic or neocarbonate B horizon on which the regic sands lie has been eroded: in areas with no regic sands there is colour continuity between the orthic A horizon and its associated B horizon (profiles AK61H, SK11, SL4-1, SPNT, OBT, TP231L, Table 2.1), whereas there is a marked colour discontinuity between the regic sand and the underlying B horizon (profiles LKC1-5, SN30, SNT60, TP266Q), and in the case of SNT60, also a lithological discontinuity. Southerly winds prevail along the coastal plain, and are responsible for the forma- tion of coastal dunefields (Harmse and Swanevelder, 1987). The calcareous dunes supply high pH calcareous material to the surface of the soil profile. Continuing burial by young aeolian dunes is evident from regic sands with buried horizons rich in organic matter (KV220PRN/3, and a buried hearth and hand axes and associated rock fragments (Figure 2.4). The age of these dunes is unclear, they form part of a dune corridor north of the Buffels River mouth, and may be anal- ogous to the one north of the Swartlintjies river (Tankard and Rogers, 1978). 60 Figure 2.4: Archaeological site. Hearth and hand axe (arrowed) buried by regic sands, approximately 1.5m deep. Near AK61H. 2.4.2 B horizons The most common are neocarbonate B horizons (generally 0.1 to 1.3 m thick), underlain in most cases by hardpan carbonate (Prieska soil form) and less com- monly by calcareous dorbank (Trawal soil form). In many cases, there is a stoneline (up to cobble-size) between the hardpan carbonate and the overlying neocarbonate horizon (such as 6869/iv, KV220PRN/9, TP266Q in Table 2.1). The neocarbonate horizons generally have a weak, medium, angular blocky struc- ture; are loose, non- to very slightly plastic and non- to very slightly sticky; and have a medium sand texture. The pH in water is 8.7 to 9.9, and EC (in the same 1 to 2.5 suspension) is 0.04 to 3.5 mS cm?1 (Table 2.1). Neocutanic and yellow-brown apedal B horizons (0.05 to 0.4 m) are also present, in one case over hardpan carbonate (Ashkam soil form, profile 6869), but more commonly over an eluvial horizon, underlain in turn by another neocutanic hori- zon with more silt and clay (SK11, SL4-1, Table 2.1). The pH is lower than 61 the neocarbonate horizons, (7.9 - 9.2) and EC is 0.1 to 2.2 mS cm?1 (Table 2.1). These horizons have a slightly stronger structure than the neocarbonate horizons, and are more sticky and plastic generally. They are loose to friable. Profile KV196T (furthest inland, Figure 2.2), has a 0.3 m thick red apedal B horizon over a 3 m dorbank horizon (Garies Form). The red apedal B horizon is loose, weak coarse angular blocky, non-plastic and non-sticky, and has a pH in water (1: 2.5 suspension) of 10.1. The EC (of the same suspension) is 0.1 mS cm?1. This profile was discussed in more detail in Chapter 1. Yaalon (1981) suggested that parent materials of different original carbonate content may be responsible for the final development of either red sands or ae- olianites on the KwaZulu-Natal coast. Highly calcareous littoral sands (calcium carbonate content higher than 30 %) become cemented to form the aeolianites, whereas sands containing less than about 25 % CaCO3 are fixed by vegetation, with soil-forming processes producing clayey red sands. Deposits with a CaCO3 content close to the threshold values separating the two diagenetic pathways can exhibit both decalcification at the surface and preservation of the CaCO3 below it (Yaalon, 1981). Pether (1994) noted that the soluble carbonate content of the coastal plain calcretes (acid digestion) is only 3%. The data of Singer et al. (1995) for the same area showed a maximum of 20% CaCO3 in a calcrete horizon, and most of the A and B horizons had less than 5%. Some profiles like LKC1-5 (Table 2.1) had non-calcareous regic sands. This suggests the relatively low CaCO3 in the aeolian sands dictates the pedogenic pathway in these deposits, and reddened neocarbonate and neocutanic horizons develop as the aeolian deposits age. Rubification was used as an index of soil development by Harden (1982) and Bockheim et al. (1992). In the Sonoran desert of Baja California, a progressive development of red pigment in Pliocene to Recent sediments has occurred by in situ alteration of iron bearing minerals through an intermediate stage of brown ferric oxide, in areas with a low water table, alkaline groundwater and interstitial oxidising conditions (Walker, 1967; Walker and Honea, 1969; Van Houten, 1973). Soils sampled in this study occur in well drained, oxidised environments. When sorted by their hue, samples fall into the field where goethite is predominant (7.5YR to 2.5YR). The red apedal samples fall into the hematite field (10R to 5YR) (Schwertmann and Taylor, 1989; Schwertmann, 1993). 62 Age Sediment supply to the aeolian sand sheet is delivered to the coastline via rivers, washed onto the beaches north of the river mouths, and then blown inland by the predominantly southerly winds (Harmse and Swanevelder, 1987). Dunefields are common on the north side of river mouths on the West Coast, and can be dis- tinguished easily on Google Earth satellite images. Three profiles (6869, SNT60 and SN30) were investigated in the ?parabolic dunefield? north of the Swartlintjies River mouth, and one profile (SL4-1) in the older material over which it accreted (Figure 2.5). This ?parabolic dunefield? was delineated by Tankard and Rogers (1978, p. 328-330), who proposed that it accreted following movement of the Swartlintjies River mouth sediment source 5 km to the west and 110 m lower Figure 2.5: Interpretation of dune plume north of Swartlintjies River, from Tankard and Rogers (1978). ?Vegetated parabolic dune plume? (grey; SN30, SNT60, 6869) accreted following the movement of the Swartlintjies River mouth sediment-source during Last Glacial (arrowed). It accreted over older ?degraded parabolic dunes? (dots; SL4-1). Modern unvegetated barchanoid dunes discontinuously override the plume (clear). Position of profiles SN30, SNT60, 6869 and SL4-1 was determined using their grid co-ordinates on the equivalently-scaled orthophoto sheets 3017AD1, AB21, AA25, AB16, AA20 (1st Edition, 1990). 63 than present, during the Last Glacial sea-level lowering. Tankard and Rogers (1978) interpreted this ?parabolic dunefield? to have formed as barchan dunes which were deposited during the Last Glacial and subsequently vegetated. Mod- ern unvegetated barchanoid dunes override the plume. Following this interpretation, horizons 6869/i,ii; SNT60/1; and SN30/A which are classified as ?regic sands? in Table 2.1 (where the only sign of pedogenesis in the profile is darkening of the upper horizon by organic material), are more specifically the vegetated parabolic dunes that were deposited during the Last Glacial Maximum. Profile SL4-1, occurs within the zone of the older ?degraded parabolic dunes? of Tankard and Rogers (1978). Its redder colour and more clay-rich nature compared to the pale regic sands lend support to its greater age. The colour and lithological discontinuities between the Last Glacial Maximum regic sands and their underlying redder, more clay-rich B horizons horizons (6869/4, SN30/B, SNT60/2, Table 2.1) suggest that their accretion took place after significant pedogenesis had already occurred. The resemblance of these underlying horizons to the B horizons of profile SL4-1 (i.e. the older ?degraded parabolic dunes? of Tankard and Rogers (1978)) suggests a pre-Last Glacial Max- imum age. This is consistent with the ?14C ages of 25 000 to 30 000 years B.P. for calcrete associated with heuweltjies (although further south) in the Clanwilliam and Elands Bay areas (Midgley et al., 2002). An attempt was made to extend this to the rest of the profiles: the degree of reddening and clay enrichment of profiles AK1/trig, DL88, KV196T, and SK11 (Group Aii in Table 2.2) are of the same magnitude as the pre-Last Glacial Maximum profile SL4-1, and the truncated B horizons (6869/4, SNT60/1 and SN30/A) (Group B). This suggests they are of an equivalent age, and older than profiles OBT, SPNT, TP231L (Group Ai). Together with the data from Midgley et al. (2002), this information shows evidence of pedogenesis from before the Last Glacial for both the northern and southern parts of Namaqualand. 2.4.3 Eluvial horizons A 0.2 m thick eluvial horizon in the diagnostic position (Soil Classification Work- ing Group, 1991) is present in a calcareous profile TP231L, where it has formed above a neocarbonate B on calcrete (Table 2.1). Bleached horizons also occur 64 Table 2.2: Redness rating (Arduino, 1985) for each soil profile, calculated from equation 2.1. The higher the redness rating, the redder the soil. /.../ indicates horizon interpreted to be missing or eroded. Abbreviations: db-dorbank; gs-gleyed sand (eluvial); hk-hard carbonate; nc-neocarbonate; ne-neocutanic; ot-orthic; rs-regic sand (may be darkened by organic matter); re-red apedal; rs-regic sand; unsp-unspecified; ye- yellow-brown apedal. Profile Soil Form Horizon Sequence Redness Rating (dry) A) A-horizon genetically related to underlying horizons i) Low redness rating OBT Prieska Form ot / nc / hk 3 / 3 / 0 SPNT Prieska Form rs / ot / nc / hk 1 / 1 / 3 / 1 TP231L Kinkelbos Form ot / E / nc 3 / 2 / 2 ii) High redness rating AK1/trig Oakleaf Form ot / ne / rock 8 / 6 / .. DL88 Augrabies Form ot / nc / unsp 6 / 6 / 5 KV196T Garies Form ot / re / db 15 / 12 / 16 SK11 Pinedene Form ot / ye / gs / ne 8 / 8 / 5 / 9 SL4-1 Pinedene Form ot / ye / gs / ne 9 / 9 / 5 / 6 B) Regic sands unconformably rest on B-horizon 6869 Prieska Form rs / .. / nc / nc / hk 1 / .. / 3 / 6 / 0 LKC1-5 Namib Form rs / .. / ne 3 / .. / 6 SN30 Prieska Form rs /.. / nc / hk 3 / .. / 4 / 0 SNT60 Namib Form rs /.. / nc / hk 1 / .. / 4 / 2 TP266Q Namib Form rs / .. / nc / hk 3 / .. / 6 / 0 third in a sequence from the surface, not in a diagnostic position as an E horizon and thus have been classified as ?gleyed sand? (?gs?) according to the definition of the Soil Classification Working Group (1991, p. 243). They are overlain by a B horizon and underlain by an horizon with a much higher clay content and EC (such as SK11 and SL4-1 (Table 2.1). Both SK11 (0.05 m thick) and SL4-1 (0.13 m thick) were non-calcareous. Profile SL4-1 is discussed in greater detail by Francis et al. (2007). The significance of these eluvial horizons for water storage in Namaqualand ecosystems has been discussed by Francis et al. (2007). Evidence for leaching below the average depth of water storage is often observed in arid-region soils (Buol et al., 1997). Buol et al. (1997) noted that while this is often attributed to a more humid paleoclimate, given the typically erratic nature of the rainfall in arid regions these soil features may simply reflect the rainfall of the extreme, rather than the average, years. An important point in considering the development of these E horizons is the reliability of the rainfall in coastal Namaqualand: although it is a ?arid? region with a high evaporation rate (2 m per year (D.B.C.M., 2000)) and less than 150 mm annual rainfall, the rainfall is reliant and occurs as frequent but low volume winter rainfall events (Desmet, 2007), Namaqualand presently receives 65 flood events about every 10 years which cause the ephemeral rivers to flow, although drought years do occur (D.B.C.M., 2000). The example in profile SL4-1 (Table 2.1) occurs in a profile in which vertical tunnels that may be attributed to termites (J. Pether, pers. comm., February 2002) have become conduits for water, extending to bedrock about 12 m deep (Figure 2.6). See Section 2.4.6 for a more detailed discussion of these burrows. The bedrock in this profile shows signs of wetness. This suggests that both lateral and vertical water movement through the regolith is extensive, and occurs along strongly preferential flow paths. Many of the mine excavations have water present, in some profiles water flows out of the deeper parts of the regolith (as in 6869, also with abundant vertical termite burrows (J. Pether, pers. comm., February 2002) which appear to have become conduits for water). The lower part of some B horizons were slightly moist to moist, especially when above a cemented horizon (for example 6869, DL88, KV220PRN, see Table 2.1), which is consistent with the suggestion of Francis et al. (2007) that these act as a barrier to water movement. Although the Namaqualand coast may have been wetter during the Last Glacial (Van Zinderen Bakker, 1976; Tankard and Rogers, 1978; Parkington et al., 2000), and many profiles are older (Section 2.4.2), Namaqualand presently receives flood events about every 10 years which cause the ephemeral rivers to flow, in addition to reliant but low volume winter rainfall events. Together with the water evident flowing out of the deeper part of the regolith in many profiles, this suggests these subsurface bleached horizons presently function as aquifers during profile wetting, and although some may have existed in a wetter paleoclimate, they continue to be active in the present. 2.4.4 Cemented horizons Calcite-cemented horizons Horizons of calcite accumulation are common. The distribution of carbonates can vary from dispersed to highly cemented (calcic to petrocalcic (Soil Survey Staff, 1999); soft carbonate to hardpan carbonate (Soil Classification Working Group, 1991)). In the diagnostic position these horizons were classified as ?hardpan 66 Figure 2.6: Possible vertical termite burrows (arrowed) that have become con- duits for preferential flow, extending through upper shoreface (red) and lower shoreface (grey) to bedrock, which shows evidence of wetness (lower right). Mine dumps obscure surface. Benches 3 m high. Profile SL4-1. carbonate?, otherwise as ?calcrete? in Table 2.1. There thickness ranges from 0.2 m to over 1 m. Petrocalcic horizons slake in acid (Soil Survey Staff, 1999). Some samples classified as hardpan carbonate had areas which slaked in water to a fine sediment (Table 2.1). This is consistent with the observations of Pether (1994) that the soluble carbonate content (acid digestion) is only 3% and the host material is a muddy sand. The average pH in water (1: 2.5 suspension) is 8.3 - 9.6, and EC ranges from 0.02 to 4.93 mS/cm (Table 2.1). Higher salt contents are often associated with dissolution of the horizon (AK1/2, SPNT/5, Table 2.1). In locations with particularly thick regic sand cover (such as profile KV220PRN (Table 2.1), and the buried hearth site in Figure 2.4) a calcrete horizon is present near the base of the calcareous regic sand (1.2 to 1.4 m deep in profile KV220PRN). This is interpreted to be a modern pedogenic process. In profile KV220PRN, a second calcrete horizon occurs deeper in the profile (2.8 to 3.8 m deep), the upper boundary of which shows dissolution features. The calcrete 67 lies under a redder neocarbonate horizon, both of which are identical to the cal- crete and reddened neocarbonate B horizons found elsewhere (Group B in Table 2.2). This suggests the horizons in Group B are the buried remains of partially eroded paleosols, and that soil forming processes have begun anew in the over- lying cover sands, via a process of aggradation as observed by (Knox, 1977) for calcrete profiles near Saldanha Bay. Silica-cemented horizons Silica-cemented horizons (dorbank (Soil Classification Working Group, 1991), equivalent to the ?duripan? of Soil Taxonomy (Soil Survey Staff, 1999) and ?petro- duric? of the World Reference Base (1998) are present. These horizons slake in concentrated alkali, or if calcareous, alternating acid and alkali treatments. Horizons that slaked only in concentrated NaOH were interpreted to be silica- cemented and classified as ?dorbank?. If they effervesced in HCl but slaked only in NaOH, they were classified as ?calcareous dorbank?. Their composition, micromorphology and genesis was discussed in more detail in Chapter 1. They can be 3 m and more thick, are generally very hard and very firm, have an average pH in water of 8.5, and an EC (of the same 1: 2.5 suspension) that ranges from 0.02 to 8.9 mS cm?1 (Table 2.1). Horizons with high EC?s show localized dissolution. The dorbank horizons are generally brown (7.5YR 5/4) to yellowish-red (5YR 5/6) when dry and reddish-brown (5YR 4/4) to yellowish-red (5YR 5/6) when moist. The calcareous dorbank horizons are paler than their non-calcareous counterparts. The dorbank horizons in both profile AK1 (Table 2.1)and KV196T (Chapter 1) contain gypsum in cracks and in voids. In KV196T the gypsum was identified by SEM-EDX, also occurs as well crystalline nodules (pictured in Appendix A). Sepiocrete and sil-sepiocrete intergrades In this study, samples that effervesced in HCl but slaked only in NaOH were classified as calcareous dorbank. One sample (SK11/1), although locally effer- vescing vigorously in HCl, only slaked partially after alternating HCl and NaOH treatments. I initially classified it as siliceous calcrete (such as Watts (1980) and Nash and Shaw (1998)), because its colour and macroscopic appearance 68 are closer to the other calcrete samples than the other dorbank samples, and it occurred above a calcrete. Subsequent SEM studies (Chapter 1) showed that calcite does not play a role in the cement, and that the interlocking of fibres is dominant, suggesting sepiolite contributes to the cement, with localised silica. These ?sepiocretes? were discussed in more detail in Chapter 1. 2.4.5 Marine sediments The marine sediments were broadly divided into an upper shoreface and lower shoreface facies, based on their bedding and particle size characteristics. The upper shoreface facies extends seaward from the low-tide level, comprising the breaker and surf zones. The onshore transport generates three-dimensional, lu- nate bedforms (megaripples), and larger bedforms. The coarsest sediment in the shallow-marine sequence is concentrated here (Pether, 1994). The lower shoreface facies is deposited seaward of the breaker zone and its as- sociated larger bedforms. It is subjected to oscillatory wave currents during fair-weather periods. Benthic organisms colonize the less routinely mobile sedi- ment. Erosion and deposition in the lower shoreface take place mainly during and after storms. Hummocky cross-laminated, fine-sandy beds consisting of low an- gle, undulatory lamination in broad overlapping troughs and low hummocks are recognised as products of storm deposition at lower shoreface to offshore depths. The lower shoreface contains more bioturbated beds and a greater diversity of trace fossils than the upper shoreface (Pether, 1994). The upper shoreface is on top of the lower shoreface facies on the Namaqualand coastal plain. This is consistent with deposition during sea level regression (Pether, 1994; Pether et al., 2000). Upper shoreface deposits are generally coarser than lower shoreface deposits, and have a redder hue. Vertical burrows filled with coarser sediment also have a slightly redder hue than surrounding sands. The coarser sediment may provide an opportunity for preferential flow, with a greater amount of interstitial space and more oxidation during dry periods. Nearby, iron pans (3 mm thick) in the finer sediment are also testament to the mobilization of iron. Probable hematite halos (comparable to those observed by Walker (1967)) are evident in pale yellow lower shoreface sands (Figure 2.7). 69 Figure 2.7: Probable hematite ?haloes? in lower shoreface sands, approximately 9m deep. Portion of hammer shown is 20cm long. Upper shoreface sediments These are generally yellowish-brown when dry (10YR 5/6), medium- to coarse sand with many gravel fragments and a weak- to moderate, coarse, angular- blocky structure. They are generally loose, non-plastic and non-sticky. The average pH in water is 8.35, and EC (of the same 1: 2.5 suspension) is 1.15 mS/cm (range 0.11 to 3.96 mS/cm, Table 2.1). The marine sands are non- calcareous, and often show large-scale bedforms. The upper shoreface ranges from 0.7 to 2 m thick in the profiles studied, but varies depending on the pack- age (Pether, 1994; Pether et al., 2000). Lower shoreface sediments These are generally light-grey (10YR 7/1), pale-yellow (5Y 7/4) or light-yellowish- brown (10YR 6/4) when dry. Oxidation of glauconite-rich sediments resulted in some layers becoming a more orange colour (such as in profiles LKC1-5 and SNT60, Table 2.1). This was determined by oxidising green lower shoreface sed- iments with H2O2, which then became a distinctly more orange colour. The average pH in water of lower shoreface sediments is 8.24, and the average EC (of the same 1: 2.5 suspension) is 0.97 mS/cm (range 0.05 to 2.87 mS/cm, Table 2.1). They are non-plastic and non-sticky, single-grain apedal, loose, non-calcareous medium sand which shows cross-lamination in places, and can contain abun- dant marine shells. In the profiles studied they are generally 12 m thick, but, 70 like the upper shoreface facies, the thickness varies depending on the package (Pether, 1994; Pether et al., 2000). In one place, there appears to be illuviation (or precipitation) of white clay at the base of a much coarser layer within the fine pale sands. This coarser layer is much redder than the over- and underlying pale fine sands. The X- ray diffraction pattern of this coarser layer shows the presence of apatite in addition to smectite, kaolinite and mica. At the base of some profiles there appear to be silica-cemented layers. Some are also calcareous. The silica and Fe-oxide precipitation, and clay illuviation observed in the lower shoreface facies on the Namaqualand coastal plain correspond with Moody and Graham?s (1994) observations in deep sands on marine terrace in central California. Often, the marine deposits rest directly on bedrock. In some profiles the bedrock is highly weathered, deeply kaolinised ?pallid zone? (Partridge and Maud, 1987) on the order of 5 m thickness, which in some cases is associated with silcrete boulders. A silcrete-capping is associated with pallid zones in southern Africa (Partridge and Maud, 1987), these boulders are therefore likely to be remnants of the original silcrete capping. It occurs in situ near Kleinzee and Hondeklip Bay. Other profiles rest on kaolinized paleochannel sediments (18 m thick in profile 6869, see Table 2.1). These sediments were described in detail by Pether (1994), who concluded that they were laid down as a quartzo-feldspathic sediment in a fluvial environment, and subsequently deeply weathered, with extensive alter- ation of feldspar to kaolinite. 2.4.6 Biological activity Evidence of biological activity is present at all levels in the regolith profile (Table 2.1), from the surface horizons with tunnels a few millimetres in diameter in the A and B horizons of to the deep marine sediments just above bedrock. It is difficult to determine whether burrows in the marine sediments were formed by marine fauna or by burrowing terrestrial organisms. In some cases the burrows have been calcretized, although the surrounding sediment is generally non-calcareous. 71 Tunnels Calcrete horizons are extensively tunnelled, the tunnels are approximately 6mm- 8mm diameter and a few centimeters long. In some cases the calcrete and deeper marine sediment contains larger tunnels (diameter approximately 10-15 cm). The dorbank horizons often contain 6 mm diameter borings and gypsum-filled voids (tunnels?). Roots Klappa (1980) distinguishes rhizoliths from faunal borings and burrows. They are abundant in calcretes and aeolianites of Quaternary age. The calcretized tunnels in sediment below a calcrete horizon in TP231L (Table 2.1) resemble roots and may be rhizoliths. Roots were identifiable in many profiles, up to 2 - 3 m deep (Table 2.1). Termites On the Namaqualand coast the heuweltjies are distinguishable on aerial pho- tographs, and in some places may be buried by recent sands. There was also abundant evidence of termite activity in deep profiles. The age of the activity was not always clear, except in places where secondary silica and calcite had clearly replaced the nests or tunnels. A particularly well preserved example of a nest was present in the calcretized sidewall of SK11 approximately 1.7 m deep (Table 2.1; Figure 2.8). In a mine excavation adjacent to an area with abundant heuweltjies (visible using Google Earth, 29? 30? 45?? S, 17? 05? E), a strongly-cemented calcretized nest was present about 2 m deep within a silica cemented, locally calcareous dorbank profile (profile AK1, Table 2.1). It is nearly identical to the fossilised nests presented by Coaton (1981). Silica cementation of the soil surrounding the nest continues from about 30cm from the surface to below the hive (Figure 2.9). Unfortunately AK1 was truncated by mining processes. The clay content of the surface horizon (AK1/1 Table 2.1) suggests it is a B horizon (neocarbonate). This horizon has a redness rating (dry) of 6. This is in line with Group Aii in Table 2.2, 72 Figure 2.8: Profile SK11 (described in Table 2.1) with termite nest, 1.7 m deep below calcrete. which suggests a pre-Last Glacial Maximum age for this neocarbonate B horizon. While the underlying dorbank and nest must be older than or contemporaneous with the neocarbonate B horizon, field relationships do not resolve the age of the nest relative to the dorbank. However, given the massive nature and the thickness of the silica cementation in the dorbank overlying the nest, it seems likely that the nest predates the silica cementation. This age is broadly consistent with the ?14C ages of 25 000 to 30 000 years B.P. for calcrete associated with heuweltjies (further south) in the Clanwilliam and Elands Bay areas (Midgley et al., 2002), and the 32 100 ? 720 years B.P. for the nearly identical fossilised nests from Clanwilliam presented by Coaton (1981). As discussed in Section 2.4.3, what appear to be vertical termite burrows (J. Pether, pers. comm., February 2002) are observable for 3 m or more vertically in the marine sediments in sediments up to 12 m deep. They are distinguishable from the surrounding sediment by texture and colour differences (Figure 2.10) and appear in some cases to have acted as conduits for preferential flow from the surface (as in Figure 2.6). Stone and Comerford (1994) note that some African and central Asian desert and dryland termite species extend galleries to remarkable depths to reach free water, citing depths of 10 m to 70 m. As noted in the discussion on eluvial horizons, there were signs of wetness on the bedrock of SL4-1, and water coming out the regolith towards the base of 6869. 73 (a) (b) Figure 2.9: Profile AK1 (described in Table 2.1) with strongly-cemented, cal- careous hive 2 m deep within dorbank. Hammer 30 cm long. Figure 2.10: Possible vertical termite burrows (arrowed), now conduits for pref- erential flow in lower shoreface sands. 74 2.5 Conclusions Soil formation on the Namaqualand coastal plain seems to be linked to periodic and episodic profile saturation that intrude the background of aridity. That evap- otranspiration greatly exceeds precipitation is reflected accumulations of leach- able salts such as halite and gypsum, and accumulations of calcite and silica. 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. Evidence for this includes moist horizons above cemented barriers and water was observed flowing out of the regolith. This is con- sistent with the regular winter rainfall, and every 10 year flood events. Seems likely that during the periodic and episodic wet events that the formation of eluvial horizons takes place, and leachable salts are moved through the profile. These features remain metastable in the soil profile during intervening periods of aridity. The surface soil horizons have generally formed from an aeolian parent material, and seem to follow either a calcareous or non-calcareous pedogenic pathway, perhaps depending on whether the original parent material was calcareous. The non-calcareous pathway leads to more clay-rich, redder apedal horizons that gradually show a stronger structure with depth, and generally rest directly on marine sands via a subtle discontinuity (as observed by Pether (1994)) that suggests pedogenesis continues through the profile into the marine facies. The clays appear to form predominantly in situ, since cutans are not obvious in the field. The lack of cutans may also be a result of the fairly low clay contents to begin with, however, and redistribution of clays within the profile is likely given the water movement through the profile. Further work on the micromorphology of the B horizons would help to confirm this. The calcareous pathway leads 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. The relatively low CaCO3 in the aeolian sands dictates the pedogenic pathway in these deposits, and reddened neocarbonate and neocutanic horizons develop as the aeolian deposits age. Deeper in the profile, there is generally a regular alteration of sedimentary units, such that the upper shoreface facies sediments show a degree of reddening, the lower shoreface sands have remained pale. This seems to be a function of the grain 75 size, since the upper shoreface materials are coarser due to their higher energy depositional environment, and the redder parts of the lower shoreface are also associated with slightly coarser sands. In some places oxidation of glauconite-rich sediments resulted in some layers becoming a more orange colour. 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. The water-holding capacity of Namaqualand soils is surprising given that they are Aridisols (Francis et al., 2007), and suggests that a case could be made for their endemism (Bockheim, 2005). Some of the features which deserve further study are regular calcrete layers and fossils (marine, terrestrial) which have the potential for dating; the abundance of pedogenic carbonates which could lend themselves to techniques to reveal whether they have played a role in atmospheric CO2 sequestration (Monger and Gallegos, 2000); evidence of deep/relict termite activity; dunefields overlying buried and truncated soils such as the well characterised Last Glacial dunefield north of the Swartlintjies. The identification and classification of the paleosols in the detail suggested by Nettleton et al. (2000) to reveal subtleties in past soil forming factors, could possibly offer some constraints on the climate conditions under which they formed. 76 Chapter 3 ?Pseudo-anticlines? and biogenic features in sepiolite-containing calcrete in Namaqualand 3.1 Introduction The Namaqualand coastal plain, with its extensive areas of calcrete development, is almost a textbook setting for calcrete development by inorganic processes: it is an area where the evaporation greatly exceeds precipitation; experiences inputs of marine aerosols due to strong onshore winds and the aggradation of calcareous regic sands (Chapter 2) similar to that described by Knox (1977) at Saldanha Bay. It also contains abundant ?pseudo-anticline? structures (pictured in (Francis et al., 2007)) which could be attributed to displacive growth of calcite and/or expandable clay minerals (Watts, 1977). There is an abundance of literature on calcrete, with many studies focussing on calcrete micromorphology (Allen, 1985; Monger et al., 1991a; Monger and Adams, 1996, to mention only a few) and it has been reviewed by authors such as Wright and Tucker (1991) and Alonso- Zarza (2003) and references therein. Notable work in southern Africa (Kalahari) includes Netterberg (1969) and Watts (1980), and recently work such as Nash and McLaren (2003). Biogenic processes, are increasingly being seen as important to calcrete genesis: Phillips et al. (1987) suggested that just as fungal hyphae can stabilize macro- 77 aggregates in soils, the mineralized filaments may produce permanently stabilized macro-aggregates that provide the locus for further carbonate precipitation, lead- ing to eventual induration. Monger et al. (1991) found that soil microorganisms were involved in calcite precipitation in desert soil near Las Cruces, New Mexico, and that where soil columns were irrigated with Ca-rich solutions, calcite only formed in soils containing soil microorganisms. Folk (1993) suggested that bacteria and ?nannobacteria? (sic) play an important role in catalyzing the precipitation of carbonate minerals. Loisy et al. (1999) noted that the originally organic micro-rods were transformed to calcite via a biomineralization process, which increased the CaCO3 content of the primary matrix. Diagenetic evolution lead to their recrystallization to microsparite, con- tributing to the hardening of soil layers as the micro- and nanoporosity were progressively filled. Wright and Tucker (1991) reviewed the formation and characteristics of biotic and abiotic calcretes, and a review of the biogenic process in calcrete development is given by (Goudie, 1996). This Chapter aims to describe the biogenic contribution to calcrete formation, following Knox?s (1977) suggestion that a study of the semi-indurated parts of the profile could bring to light biogenic structures. In calcretes near Saldanha Bay on the west coast of South Africa, Knox (1977) observed that needle fibre calcite (?acicular networks?) and filaments (?tubules?) were associated with the semi-indurated zones in the profile. The strongly indurated zones consisted of cryptocrystalline calcite, and the beta-fabric features were no longer recognizable. THis study particularly aimed to distinguish the biogenically acicular calcite from the similar-shaped clay minerals sepiolite and palygorskite using HCl-etching and EDAX analyses, since Singer et al. (1995, p. 65) questioned whether the biogenic calcite fibres described by Knox (1977) in calcretes near Saldanha Bay were indeed calcite, rather than sepiolite/palygorskite fibres. This study also aimed to determine whether the pseudo-anticline structures could be attributed to displacive growth of calcite and/or expandable clay minerals, as speculated by Francis et al. (2007). 78 3.2 Materials and methods A subset of samples (calcrete and calcareous nodules) was selected from those described in (Chapter 2). The calcrete samples are SNT60/5A; SPNT/4; OBT/4; TP266Q/1 and calcite nodules (calcite-cemented sand) approximately 3 cm in diameter from neocarbonate B horizons 6869/iii and 6869/iv. The locations and profile descriptions are shown in Chapter 2, Figure 2.2 2 and Table 2.1 respectively. The profiles are pictured in Appendix A. Namaqualand, particularly the coastal part that was the focus of this study, is a region that experiences an extremely high evaporation rate (about 2 m per year) and less than 150 mm annual winter rainfall, supplemented by regular fog occurrences (Eckardt and Schemenauer, 1998; D.B.C.M., 2000; Olivier, 2002). Crucially for both plant growth and water movement through the soil, the climate is characterised by highly reliable rainfall when compared to other arid regions with similar mean annual precipitation (Desmet, 2007). It typically arrives as widespread, gentle showers (average rainfall event 6 mm, P. Carrick unpublished data, cited by Desmet (2007)). Drought conditions are rare, and rainfall is higher than average about once every 10 years, causing ephemeral rivers to flow. Paleoclimates of the northern Namaqualand region are not that well constrained, but the consensus for the greater area seems to be that is was wetter during the Last Glacial (Van Zinderen Bakker, 1976; Tankard and Rogers, 1978; Parkington et al., 2000). A more detailed summary of the landscape, (paleo)climate and vegetation is given by Desmet (2007) and in Chapters 1 and 2. Uncoated fragments and thin sections were observed before and after etching in 1M HCl using low vacuum SEM-EDX with a Philips Xl30 ESEM. High vacuum SEM was done on Au-coated fragments using a Leo 1430VP SEM-EDX system. Thin sections for optical microscopy were impregnated with an epoxy resin, and ground without water. For XRD analysis of the carbonates and cutans, an air-dried fragment was ground in an agate pestle and mortar and analysed as a powder. For XRD analysis of the clay fraction, the bulk samples were air- dried, crushed and passed through a 2 mm sieve. To minimize shattering of pebbles/gravel, the >2 mm cement was sorted from the >2 mm pebbles and recrushed. Selected samples were milled and treated with a pH 5 sodium ac- etate buffer for a short time (following the findings of Vanden Heuvel, 1964) until effervescence ceased, and then concentrated by centrifugation. This was 79 repeated with fresh solution until there was no more effervescence. The <2 ?m fraction was separated by dispersion (shaking briefly by hand, raising the pH to approximately 10 with Na2CO3) and settling. The clay suspension was floc- culated by the addition of MgCl2, after lowering the pH to 5 to 7 with HCl to prevent the precipitation of brucite and/or clay destruction. The clay suspension was then Mg- or K-saturated, concentrated by centrifugation, and sedimented (or smeared, many of the seopiolite-rich samples developed ?mudcracks?) onto a glass slide. XRD analyses were done using a stepsize of 0.05 degrees and step- time of 40 seconds, using a Bruker D8 Advance Powder Diffractometer with a graphite monochromator, 40 kV and 40 mA. Ethylene glycol was sprayed lightly onto the surface of the Mg-saturated sample slides. 3.3 Results and discussion 3.3.1 Composition Carbonates The dominant carbonate mineral in the calcretes is calcite (Figure 3.1). Dolomite appears to be absent. Calcite in the clay fraction shows a very small shift towards magnesian calcite (Figure 3.2). Both SPNT and SNT60/5A have the calcite peak at 0.3023 nm. This yields a 0.037 mole fraction MgCO3 in the calcite following the calculation Y = 0.3033 - 0.0271 X, where Y is the d(211) in nanometres (Doner and Lynn, 1989) . The 0.326 nm peak (well-developed in SPNT/4 and to a lesser extent in SNT60/5a) is somewhat of an enigma. It could be attributed to aragonite (0.327 nm, in- tensity = 52, JCPDS card 05-0453 from Doner and Lynn, 1989), with the 0.34 nm aragonite peak (intensity = 100) overlapping that of quartz (JCPDS card 05-0490 Drees et al., 1989), but the 0.198 nm (intensity = 65) and other aragonite peaks are absent. The 0.326 nm peak is absent from the clay fraction in both of these samples, although calcite is present (Figure 3.2). If authigenic aragonite were present, it should occur together with calcite in the clay fraction too. Since the material used to determine the nature of the carbonates and to extract the clay fraction were obtained from different peds, a possibility is that an aragonitic 80 Q 0.426 n m Q 0.334 n m C 0.304 n m SPNT/ 4 SNT60/5 a D 0.289 n m Mg - C 0.297 n m ? A 0.34 n m ? A 0.198 n m ? A 0.327 n m ? A 0.27 n m OBT/ 4 C 0.229 n m C 0.210 n m C 0.25 n m C 0.191 n m C 0.386 n m C 0.188 n m Q 0.246 n m Q 0.228; Q 0.182 n m Q 0.224 n m D 0.219 n m D 0.18 n m D 0.179 / D 0.178 n m 20 25 30 35 40 45 50 55 Cu K ? ? 2 ? Figure 3.1: Carbonates in calcrete: XRD traces of crushed, bulk samples. Mg-C: Mg-carbonate. Carbonate peaks: Doner & Lynn (1989); quartz: Drees et al. (1989). A: aragonite (JCPDS card 05-0453); C: calcite (JCPDS card 05-0586); D: dolomite (JCPDS card 11-0078); Q: quartz (JCPDS card 05-0490). shell fragment was present in the ped crushed for bulk XRD analysis, but not the ped from which clays were extracted. Other contenders for the 0.326 nm peak are titanate (0.325nm, JCPDS card no. 73-2066) and zircon (0.328 nm, JCPDS card no. 83-1379), both present in the heavy mineral fraction of the marine sands (Pether, 1994). A possible reason for their presence in the bulk trace but not the clay fraction, could be that their greater specific gravity caused them to settle faster than the clay particles and so not be extracted with the clay fraction. Clays The clay fraction of coastal Namaqualand calcretes is dominated by sepiolite and contains varying amounts of kaolinite, palygorskite and mica (Figure 3.2). Mica is likely to be detrital, given the abundance of it in the lower shoreface facies (see Chapter 2) and that it is a common component of the bedrock in the region. The decrease of the 1.2 nm peak in TP266Q/1 and the increase of the 1.04 nm peak on heating is consistent with sepiolite (Hayashi et al., 1969). These 81 Sep iolit e 1. 21 n m Pal ygo rski te 1 .048 nm Sm ecti te 1 .5 n m Sm ecti te E G 1 .7 n m Mic a 1. 0 nm Pal ygo rski te 0 .418 nm Pal ygo rski te 0 .323 nm Sep iolit e; p aly. 0.4 49 nm Sep iolit e; m ica 0.33 4 nm Ca lcite 0.3 04 n m sep iolit e (di ff. ) 0.5 04 n m milled; <2 ?m; Mg milled; <2 ?m; Mg; EG milled; acetate; <2 ?m; Mg SPNT/4 milled; acetate; <2 ?m; Mg; EG milled; <2 ?m; Mg milled; <2 ?m; Mg; EG SNT60/5A milled; acetate; <2 ?m; Mg milled; acetate; <2 ?m; Mg; EG Kao linit e 0. 716 nm Kao linit e 0. 357 nm Pal ygo rski te 0 .644 nm Mic a; <2 ?m; Mg TP266Q/1 <2 ?m; Mg; EG <2 ?m; K; 300?C 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 3.2: XRD traces of the clay fraction from 3 calcretes. Calcite (JCPDS card 05-0586, Doner & Lynn, 1989); kaolinite (Dixon, 1989); mica (Fanning et al., 1989); palygorskite (Bradley, 1940; Christ 1969); sepiolite (Brindley, 1959). Diff.: diffuse; EG: ethylene glycol; paly.: palygorskite. 82 peak positions overlap with palygorskite and mica, but the heated trace does show a small increase in the 0.94 nm region (Van Scoyoc et al., 1979), sug- gesting some palygorskite is present even though the 0.644 nm palygorskite peak is absent. This is consistent with the clay mineralogy of other southern African calcretes (Netterberg, 1969; Watts, 1980). Sepiolite and palygorskite are present in Namaqualand soils, with sepiolite dominant in the coastal areas (Singer et al., 1995). Hay and Wiggins (1980) found sepiolite to be the most common clay mineral associated with calcretes in the southwestern USA, and palygorskite was notably absent. Although the <2 mm samples were milled before acetate treatment to liberate calcite-cemented clays, the clay peaks are substantially enhanced after the ac- etate treatment (Figure 3.2). The difference is most noticeable with palygorskite, and consistent with Vanden Heuvel?s (1964, p. 201) observation that palygorskite showed ?greater immobilization? by the calcite aggregates than sepiolite. Many of the calcretes show sepiolite cutans around the blocky peds, evident in the field by their non-effervescence with HCl and by turning methyl orange pink (Mifsud et al., 1979). An XRD trace of a typical cutan (Figure 3.3) shows sepiolite, quartz and calcite. The sepiolite peak is broad, and it extends into the palygorskite peak zone of 1.048 nm, but the presence of a small amount of palygorskite is suggested by small 0.644 and 0.323 nm peaks. Watts (1980) noted an inverse relationship between the mol % MgCO3 in calcite and the sepiolite/palygorskite content, which suggested that these Mg-rich clay minerals were the main Mg-bearing phase in Botswana calcretes. The apparent absence of Mg-calcite and dolomite, and abundance of sepiolite in the calcretes of coastal Namaqualand suggests a similar partitioning has occurred. 3.3.2 Morphology and micromorphology ?Pseudo-anticlines? Some profiles show ?pseudo-anticline? features (Watts, 1977), pictured in Francis et al. (2007). These may be attributed to both the displacive growth of cal- cite and expansion of clay minerals (Watts, 1977; Wright and Tucker, 1991). A blocky structure and accommodating planes in calcrete sample TP266Q/1 (Fig- 83 S 1.21n m P 1.048 n m S 0.75 n m P 0.644 n m P 0.5 42 n m S 0.43 n m Q 0.426 n m S 0.374 n m P 0.369 n m P 0.35 n m P 0.323 n m S ; P 0.449 n m Q ; S 0.334 n m C 0.304 n m C 0.386 n m 0 5 10 15 20 25 30 Cu K ? ? 2 ? Figure 3.3: XRD trace of sepiolite cutan on calcrete (OBT/4), powdered bulk sample. C: calcite (JCPDS card 05-0586, Doner & Lynn (1989)); P: palygorskite (Bradley, 1940; Christ 1969); S: sepiolite (Brindley, 1959) Q: quartz (JCPDS card 05-0490, Drees et al., 1989). ures 3.4(a), (b)) suggest shrink/swell behaviour has occurred, and is consistent with presence of smectite in the clay fraction of TP266Q/1 (Figure 3.2), although the ?diapiric anticlines? that Watts (1977) attributed to the upward injection of swelling clays were not present. Volumetrically, the displacive calcite appears to be more significant than the shrink/swell behaviour represented by the accom- modating planes. Clay neoformation/illuviation may also contribute to displacement and folding via an increase in volume in a similar manner to displacive calcite, rather than via a shrink/swell mechanism. Clay coatings almost completely fill the intergrain spaces after the calcite is removed, including spaces between a fractured quartz grain (Figure 3.4(c)). This is consistent with Pether?s (1994) observations that the soluble carbonate content (acid digestion) is only 3% and the host material is a muddy sand. The oriented nature of these coatings suggests they are illuvial in origin, although according to Ranson and Bidwell (1990, in Wright and Tucker, 1991), clay coatings in calcareous soils may have formed through stress rather than illuviation. 84 (a) Accommodating planes (arrowed). Scale bar 0.1 mm, ppl. (b) Same as (a). Micrite displacing fractured quartz grains along fractures. Xpl, ? plate, fast direction indicated. (c) HCl-etched. Orientated clay matrix; fractured quartz grain (1-3). Scale bar 0.1 mm, xpl, ? plate, fast direction indicated. (d) Calcite displacing feldspar grain along cleavage planes. Scale bar 0.1 mm, xpl. (e) Calcite replacing ferro-magnesian mineral (pyroxene); embayed quartz grain. Scale bar 0.1 mm, xpl. (f) Needle fibre calcite and mineralized filaments (arrowed, cross-section) with radial length-fast (?)calcite/oxalate, lining void. xpl, ? plate, fast direction indicated. Scale bar 0.02 mm. Figure 3.4: Displacive, replacive and needle fibre calcite; clay textures in calcrete TP266Q/1. 85 Both displacive and replacive calcite are present. Displacive calcite is dominant (Figures 3.4(b), (d)), resulting in a ?floating fabric? (Watts, 1978), represented in this sample by a close porphyric c/f-related distribution (Stoops, 2003). As noted by Watts (1978), replacement appears to be of a ferro-magnesian mineral, the shape of which suggests pyroxene (Figure 3.4(e)). Both the embayed quartz grain and the calcite pseudomorph (Figure 3.4(e)) suggest silicate dissolution has also occurred in the profile, although to a lesser extent. Reheis et al. (1992) found that pedogenic calcite both displaced and replaced grains in sediments of the Kyle Canyon alluvial fan, southern Nevada. In Buczynski and Chafetz?s (1987) model, calcite crystals in the vadose zone pre- cipitate from a supersaturated solution that is not being constantly replenished, forming rapidly at grain contacts as the last fluid dries. This crystallization is too rapid to be accommodated by pressure solution and the interpenetra- tion of quartz grains. Consequently, the ?rapid? build-up in stress is released by grain fracturing. This is consistent with much of sample TP266Q/1, with the abundant evidence of grain shattering and the micrite suggesting rapid crystal- lization. Chadwick and Nettleton (1990, in Wright and Tucker, 1991) attributed displacive fabric features in calcretes to calcite crystals preferentially forming co- hesive bonds with other calcite crystals, thereby displacing non-carbonate grains. Biogenic fabric The typical abiotic alpha-fabric features (Wright and Tucker, 1991) dominate the mature calcrete horizon TP266Q/1 (Figure 3.4). Biogenic beta-fabric features (needle fibre calcite, mineralized filaments) are present, but to a lesser degree. They are, however, dominant in calcite-cemented nodules from a calcic B horizon (6869/iii and 6869/iv) above calcrete (Figure 3.5). Figure 3.4(f) shows the cross-section through mineralized filaments in a void in a mature calcrete horizon (TP266Q/1). They appear to comprise a radially- oriented, length slow mineral with high birefringence, and are closely associated with needle fibre calcite. They are consistent with the morphology of fungal filaments found by Phillips et al. (1987) and Verrecchia et al. (1993). Mineralized filaments dominate the fabric of calcite-cemented nodules from a calcic B horizon above a calcrete (Figure 3.5). 86 Verrecchia et al. (1993) found that filament coatings were predominantly calcium oxalate. Their analyses showed oxalate to have a C/Ca ratio of about 1, with calcite having a much smaller C/Ca ratio. The C/Ca ratio of filaments in Fig- ure 3.5(a) (analyses 2-4, Figure 3.5(b)) suggests calcium oxalate in the coating, whereas the non-filament matrix (analyses 1 and 5) are micrite. Non-mineralized filaments (Phillips et al., 1987) and nanobacteria (Folk, 1993; Folk, 1999) are also present (Figures 3.5(c), (d)). Verrecchia et al. (1993) attributed the fila- ments in a calcrete from Israel to fungi, since they occurred in an environment without light, eliminating the involvement of photosynthetic microorganisms like cyanobacteria and algae. This also seems a reasonable assumption to make for the Namaqualand samples. Folk (1993) argued convincingly against modern contamination of the sample, and provided many examples for nanobacteria in calcretes (Folk, 1999). Needle fibre calcite (Figure 3.5(e), (f)) is present as the MA(2), MAB, and MB rod forms of Verrecchia and Verrecchia (1994), who concluded that MA rods are of biological origin which can be attributed to calcite precipitation in the interior of fungal mycelian bundles, which are then released and redistributed in the soil following the decay of the organic matter. MB rods form after the release of MA rods; their serrated plates are a secondary syntactic growth of CaCO3 of physico-chemical origin on an MA skeleton. The monocrystalline M micro-rods are much smaller than MA rods (0.5 - 1 ?m), and have an ambiguous origin related to either physico-chemical nuclei or calcified bacteria; they are commonly associated with gels or organic material and do not appear to evolve toward larger or more complex morphologies (Verrecchia and Verrecchia, 1994). Loisy et al. (1999) recognised the micro-rods as bacilliform and threadlike bacteria, which mineralize into calcite, forming micro-rod mats and can form clusters and micritic platelets. In Figure 3.5(d) M micro-rods are seen covering needle fibre calcite; this relationship is also present in calcrete in Israel (Verrecchia and Verrecchia, 1994, p. 657) and a calcrete paleosol in France (Loisy et al., 1999, p. 198). The analysis of the fibrous mat under the nanobacteria (Figure 3.5(d)) suggests that it is not calcite. The morphology of the M micro-rods is very similar to that of the fibrous clay minerals sepiolite and palygorskite, which are abundant in calcretes on the Namaqualand coastal plain (Figures 3.2 to 3.3). The <1 ?m clay fraction of this sample is dominated by sepiolite, with less intense palygorskite, smectite, kaolinite and possibly illite (Chapter 4), therefore it seems likely that 87 (a) Micrite and Ca-oxalate covered filaments. Scale bar 2 ?m. Au-coated. Wt % 1 2 3 4 4 5 C 14 21 19 19 19 17 Mg 1.4 0.7 0.8 Ca 34 16 20 22 20 28 O 51 63 60 59 60 55 Total 100 100 100 100 100 100 C/Ca 0.4 1.3 1.0 0.9 1.0 0.6 (b) EDAX analyses of areas indicated in (a) (c) Collapsed, non-mineralized filament. Scale bar 10 ?m. Au-coated. (d) Nannobacteria (left) on fibrous clay mineral mat (analysis arrowed); M micro-rods covering MA needle-fibre calcite (top right). Scale bar 1 ?m. Au-coated. (e) Needle fibre calcite: MAB rod. Scale bar 5 ?m. Uncoated. (f) Mineralized filaments (F); needle fibre calcite: MA rods (MA2 rod (bottom)), MB rod (arrowed). Scale bar 5 ?m. Uncoated. Figure 3.5: SEM images: Biogenic calcite from calcite nodules in 6869/iii and /iv, calcic B above a calcrete. Needle fibre calcite classified according to Verrec- chia and Verrecchia (1994, further details in text). Images (a)-(d) collected at University of Stellenbosch, (e), (f) at University of Kwazulu-Natal. 88 the mat is in fact interwoven (predominantly fibrous) clay minerals. Further evidence for the association of sepiolite/palygorskite with the organic structures is in Figure 3.6. The samples with mineralized filaments and needle fibre calcite from Figure 3.5 were etched in 1M HCl. The micrite, needle fibre calcite and mineralized filaments dissolved (Figure 3.6(a)), and revealed the fibrous clay minerals (Figure 3.6(b)). The composition of the area is consistent with the HCl-etched calcite/oxalate and the clay fraction described above. The large proportion of Mg and Si is consistent with the predominant sepiolite. (a) Mineralized filament, partially dissolved. Scale bar 5 ?m. (b) Magnification of area indicated in (a). Scale bar 1 ?m. Figure 3.6: 6869/iii, from Figure 3.5 after etching for 15 seconds with 1M HCl. 3.4 Conclusions The clay fraction of coastal Namaqualand calcretes is dominated by sepiolite, with varying kaolinite, palygorskite and detrital mica contents. The clay fraction is closely associated with calcite: although the <2 mm samples were milled before acetate treatment to liberate calcite-cemented clays, all the clay peaks are substantially enhanced after the acetate treatment. The difference is most noticeable with palygorskite, which is in some cases only detectable after acetate treatment. The apparent absence of Mg-calcite and dolomite, and abundance of sepiolite in the calcretes of coastal Namaqualand suggests that this Mg-rich clay mineral is 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 89 behaviour is present in the form of accommodating planes, it does not appear to be as volumetrically significant as displacive calcite. Clay minerals are closely as- sociated with the displacing calcite, suggesting that clay neoformation/illuviation may contribute to displacement and folding, but calcite crystallization seems to be mainly responsible for opening fractures. Both biotic and abiotic factors contributed to calcrete formation, although abi- otic alpha-fabric seems dominant in the mature calcrete horizon, and beta-fabric in calcareous nodules in a calcic B horizon above calcrete. There is evidence for the association of sepiolite/palygorskite with the organic calcite structures. Since the micromorphology of acicular calcite can sometimes resemble the acicular clay minerals sepiolite or palygorskite, EDAX analyses and HCl-etching is necessary to distinguish the two. In this system, the distinction is made easier by the fact that sepiolite (and to a lesser extent palygorskite) are the main Mg-bearing phases, since the XRD analysis revealed the main carbonate phase in the calcretes is low-Mg calcite. 90 Chapter 4 Clay mineral occurrences in Aridisols on the west coast of South Africa (Namaqualand) 4.1 Introduction Namaqualand is an arid region of low agricultural and habitation density, and so information on the soils and their clay mineralogy is sparse in comparison to other areas in South Africa. The Namaqualand coastal plain experiences an extremely high evaporation rate (about 2 m per year) and less than 150 mm an- nual winter rainfall that is supplemented by regular fog occurrences (Eckardt and Schemenauer, 1998; D.B.C.M., 2000; Olivier, 2002). Namaqualand has recently become the focus of an increasing amount of biodiversity and ecological research (ranging from publications such as Cowling and Pierce, 1999; to a Special Is- sue of the Journal of Arid Environments, 2007). Rehabilitation efforts after the hundred or so years of mining, coupled with the increasing interest in ecology and biodiversity, has prompted a recognition of the soils as a substrate for plant growth and has motivated studies such as those by Prinsloo (2005), Francis et al. (2007) and Chapter 2. It is this arid climate which has allowed the formation of large amounts of the relatively rare soil clay minerals sepiolite and palygorskite. Singer et al. (1995) studied the mineralogy of Namaqualand soils from the Land Type survey (Land 91 Type Survey Staff, 1987), and found that palygorskite dominated the soils inland of the escarpment whereas sepiolite was more prominent in the clay fraction of soils from coastal areas due to marine-influenced elevation of magnesium levels. Chapter 1 reported occurrences of ?sepiocrete? in Namaqualand, horizons which appear to be cemented by sepiolite. Sepiolite and palygorskite are often reported from arid region soils and are commonly found in calcretes (Vanden Heuvel, 1964; Singer and Norrish, 1974; Yaalon and Wieder, 1976; Elprince et al., 1979; Hay and Wiggins, 1980; Singer and Galan, 1984; Singer, 1989; Blank and Fosberg, 1991; Monger and Daugherty, 1991; Verrecchia and Le Coustumer, 1996; Singer, 2002; Neaman and Singer, 2004; Owliaie et al., 2006), southern African examples of which are described by Netterberg (1969) and Watts (1980). Sepiolite is considered to be rarer in the pedogenic environment than palygorskite (Singer and Galan, 1984; Singer, 1989; Singer, 2002), although Hay and Wiggins (1980) found sepiolite to be the most common clay mineral associated with calcretes in the southwestern USA with palygorskite notably absent. Kerolite is a hydrated and disordered talc-like mineral (Mg3Si4O10(OH)2.nH2O) with a basal spacing of 0.96 nm, giving a broad diffraction peak at 1.0 nm (Brindley et al., 1977) that is sometimes found with sepiolite and other Mg-rich clay phases. Kerolite and randomly interstratified kerolite-stevensite have been reported to occur with sepiolite (Eberl et al., 1982; Khoury et al., 1982; Hay et al., 1986; Martin De Vidales et al., 1991, among others). Both Stoessell (1988) and Le?veille? et al. (2002) suggested that the inherent difficulties in its identification (poor crystallinity and XRD peak overlap with other phases) in mixed mineral assemblages means it could be overlooked. In spring-related deposits in the Amboseli Lake Basin on the Kenya-Tanzania border, kerolite appeared often to have had a sepiolite precursor, and may have formed as weathering product during a lowering of pH (Stoessell and Hay, 1978; Hay and Stoessell, 1984). Its stability field is close to that of sepiolite (Stoessell, 1988). Since sepiolite is so abundant in Namaqualand soils, there is a possibility of kerolite occurring too, particularly in the fine clay fraction due to the fine, disordered nature of kerolite (Brindley et al., 1977; Pozo and Casas, 1999). It has not yet been recorded from a pedogenic environment. The aims of this Chapter were to evaluate the effect of pretreatments on the se- piolite and palygorskite peaks; concentrate kerolite if present; examine the clay mineral associations throughout the profile including the marine parent materi- 92 als to establish which phases are inherited; and to examine the morphology and modes of occurrence of the fibrous clay minerals in Namaqualand soils, especially for evidence of disintegration or transformation into sheet silicates, since Singer et al. (1995) concluded that the sepiolite and palygorskite in their study were in a state of disintegration or transformation into sheet silicates. This study is relevant to the growing body of literature on sepiolite and palygorskite oc- currences, particularly the issue of neoformation and transformation, that has appeared since the findings of Singer et al. (1995) were published (Torres-Ru??z et al., 1994; Sa?nchez and Gala?n, 1995; Lopez-Galindo et al., 1996; Jamoussi et al., 2003; Zaaboub et al., 2005). 4.2 Materials and methods For analysis of the clay fraction, a subset of 43 samples was selected from those described in Chapter 2. These were selected to be representative of the hori- zon types encountered. Sampling was done on the Namaqualand coastal plain, in two of the mining areas belonging to De Beers Consolidated Mines Ltd - Namaqualand Mines. These are the Buffels Marine Complex (BMC) in the Kleinzee-Port Nolloth area, and the Koingnaas Complex (KNC) further south in the Koingaas-Hondeklip Bay area The location of the study area overlaps with the north-western coastal area of the region studied by Singer et al. (1995), an area where they found sepiolite to be dominant. These soils can be classified into the Aridisol and Entisol orders of Soil Taxonomy (Soil Survey Staff, 1999). The clay mineralogy of the dorbank (petroduric) and ?sepiocrete? horizons was pre- sented in Chapter 1, and the clay mineralogy of calcretes on the Namaqualand coastal plain was presented in Chapter 3. Profile descriptions and sampling de- tails of the samples used in this Chapter are in Chapter 2. A summary of the landscape, climate and vegetation is in Desmet (2007) and Chapter 2. Profile photographs are presented in Appendix A. The parent material comprises successive late Tertiary marine packages, each deposited during sea-level regression (Pether, 1994). The marine packages are arranged en echelon down the bedrock gradient, from oldest and highest inland to youngest and lowest at the coast. Each package is named after the elevation of its transgressive maximum as represented in the Hondeklip Bay area (Pether, 93 1994; Pether et al., 2000). The 90 m package is ca. 18 - 16 Ma, the 50 m package early Pliocene, and the 30 m package not well constrained, but ca. 3.3 Ma or younger (Pether et al., 2000). The 30 m package is transgressed by younger littoral deposits up to about 10 m a.m.s.l. (Pether et al., 2000). The sand fraction is dominated by quartz and feldspar, lesser glauconite and phosphatic shell fragments, and variable amounts of heavy minerals (garnet, magnetite, ilmenite, biotite, sphene, amphibole, epidote, kyanite (rare) and zircon (rare)) (Pether, 1994). The marine packages are capped by recent aeolian deposits. Coastal dunefields originate at the mouth of the rivers and extend inland in a northerly direction. In places, older soil profiles are overlain by aeolian sands. Soil formation has taken place in aeolian, ?sheetwash?, and marine-deposited sediments. Some of the marine sediments rest on kaolinized paleochannel sediments, de- scribed in detail by Pether (1994), who concluded that they were laid down as a quartzo-feldspathic sediment in a fluvial environment, and subsequently deeply weathered, with extensive alteration of feldspar to kaolinite. These channel sedi- ments are described in greater detail by Tankard (1966), Tankard (1975), Pether (1994), and Pether et al. (2000). De Villiers and Cadman (2001) favoured an early Tertiary age for this channel north of the Swartlintjies River, although it may also contain reworked Cretaceous material (Pether et al., 2000; De Villiers and Cadman, 2001). Uncoated fragments and thin sections were observed before and after etching in 1M HCl using low vacuum SEM-EDX with a Philips Xl30 ESEM. High vacuum SEM was done on Au-coated fragments using a Leo 1430VP SEM-EDX system. For XRD analysis of the clay fraction, the bulk samples were air-dried, crushed and passed through a 2 mm sieve. To minimize shattering of pebbles/gravel, the >2 mm cement was sorted from the >2 mm pebbles and recrushed. To compare the effect of milling and acetate pretreatments, selected samples were also milled for 5 minutes using a stainless steel ball mill, and then treated with a pH 5 sodium acetate buffer for a short time (following the findings of Vanden Heuvel, 1964) until effervescence ceased, and then concentrated by centrifugation. This was repeated with a fresh pH 5 sodium acetate buffer solution until no further effervescence was observed in the sample. The <2 ?m fractions of the (a) sieved, (b) milled, and (c) milled and acetate-treated subsamples were then separated by dispersion (shaking briefly by hand, raising the pH to approximately 10 with 94 Na2CO3) and settling. The 1 to 0.08 ?m and <0.08 ?m fractions were separated by centrifugation from the <2 ?m fraction. The clay suspension was flocculated by addition of MgCl2 after lowering the pH to 5 to 7 with HCl to prevent the precipitation of brucite and/or clay destruction. The clay suspension was then Mg-, Ca- or K-saturated, concentrated by centrifugation, and sedimented (or smeared, many of the sepiolite-rich samples developed ?mudcracks?) onto a glass slide. XRD analyses were done with a stepsize of 0.05 degrees and steptime of 40 seconds, using a Bruker D8 Advance Powder Diffractometer with a graphite monochromator, 40 kV and 40 mA. Ethylene glycol was sprayed lightly onto the surface of the Mg-saturated sample slides. After analysis, the ethylene glycollated 1 to 0.08 ?m and <0.08 ?m samples were allowed to remain under ethylene glycol for 10 weeks, and then re-analysed to allow comparison to the findings of Brindley et al. (1977), Hay et al. (1995) and Le?veille? et al. (2002) for kerolite. For samples where glycerol was used, it was by gaseous diffusion. NaN3 was added to the 1 to 0.08 ?m and <0.08 ?m suspensions as a bacteriocide during storage. TEM-EDX analyses of clay fractions were done using a Philips CM120 Biotwin equipped with EDX detector. Total chemical analysis was done on the Mg- and K-saturated clay fractions, using fused glass beads in a Philips 1404 Wavelength Dispersive XRF spectrometer, fitted with a Rh tube. 4.3 Results and discussion 4.3.1 Effect of milling and acetate pretreatments on clay fraction Milling before separating the clay fraction substantially reduced the sharpness of the peaks. Although the < 2 mm samples were milled before acetate treatment in order to liberate calcite-cemented clays, the clay peaks were substantially enhanced after the acetate treatment (Figures 4.1(a), 4.2). This suggests that the clays are aggregated with calcite. The calcite aggregates binding the clays appear to be finer than the size fraction which can be liberated with ball milling, because it is only in the <0.08 ?m clay fraction that acetate treatment appeared not to liberate any different clays (compare the <0.08 ?m with the 1 to 0.08 ?m and <2 ?m traces in Figure 4.2). 95 This was also noted in Chapter 3, where the difference was most noticeable with palygorskite and is consistent with Vanden Heuvel?s (1964, p. 201) observation that palygorskite showed ?greater immobilization? by the calcite aggregates than sepiolite. The <0.08 ?m is the only size fraction where palygorskite could be detected before acetate treatment (Figure 4.2). It may be that palygorskite appeared to be more closely associated (aggregated) with calcite because it is finer than the other clay minerals, and finer than the size fraction that can be liberated by ball milling. 4.3.2 Clay mineral variations in soil profiles Table 4.1 shows the dominant clay phases in the <2 mm fraction for representa- tive soil profiles. The corresponding XRD traces are presented alphabetically in Figures 4.1 to 4.6. Sepiolite occurs across all packages and elevations up to 99 m a.m.s.l., the farthest sampled from the sea (about 30 km). Neither sepiolite nor palygorskite are present in the marine sediments, which contain only smectite, mica, and kaolinite in varying proportions. Although the appearance of sepiolite and palygorskite seems to be limited to subaerially-reworked material, not all aeolian material contains sepiolite/palygorskite (profiles LKC1-5 and SL4-1 in Figures 4.4 and 4.6, for example). It is therefore unlikely that the sepiolite is inherited from either the marine or aeolian parent materials. The appearance of sepiolite and palygorskite are most consistently associated with the presence of calcite in the soil profile. The presence of high angled peaks in 6869/4 and /iv (Figures 4.1(a), (b)) suggests either that some interlayers are present, or that these may be reflections of sepiolite. Visual evaluation of the peak intensities in Figures 4.1 - 4.6 suggests that sepio- lite is more abundant than palygorskite. The relative abundances of sepiolite and palygorskite in Table 4.1 were estimated using the 0.644 nm peak of palygorskite in addition to the 12.1/1.048 nm peak ratio, since the 0.644 nm peak does not overlap with other phases, and its intensity is medium strong (Bradley, 1940), increasing if the sample is oriented (Nathan et al., 1970, in Singer, 1989). How- ever, in a mixture of phases the quantity of sepiolite could be overestimated and that of palygorskite underestimated using peak height or area of the main reflec- tion line, since even in low concentrations sepiolite might show high diffraction intensity (Ouhadi and Yong, 2003). Despite this, the dominance of sepiolite over 96 Table 4.1: Clay mineral variation down profiles. Abundances were visually es- timated from peak sizes: ?xxx? dominant peak; ?x? trace or only in <0.08 ?m fraction; ?-? not detected. Traces shown in Figures 4.1 to 4.6. Depth Horizon ? Facies ? Estimated React Smectite 1.4 nm Sepio- Palygor- Mica/ Kaolinite Inter- (m) Texture ? HCl lite skite Illite stratified 6869: Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). KNC, 30 m package, 41 m a.m.s.l. (GPS), footslope. 0- / i A1 Regic sand (Orthic A) Aeolian dune Med. sand Calc. 0.45- / ii A2 Regic sand Aeo lian deflation Med. sand Calc. 0.5- / iii II B1 Neocarbonate B Aeolian? Med. sand Calc. - - XXX X - X - 0.6- / i v II B2 Neocarbonate B Aeolian? Lo. sand Calc. X - XXX XX X X ? 0.8 over hardpan carbonate New profile, mining excavation, upper material removed. Askham Form (Orthic A - Yellow-brown apedal B - Hardpan carbonate). ?-0.32 / 4 III B Yellow-brown apedal B Obscured Lo. sand <1?m X XXX X XX X ? 0.32- / 5 IV C1 Hardpan carbonate U. shoreface Lo. c. sand Calc. X X X X - 1- / 6 IV C2 Marine deposit U. shoreface C. sand - 6m-9m unit consists of alternate coarser, redder & finer, yellower sedimentary layers. 6-9 / 7 V C3 Marine deposit L. shoreface Med. sand - XXX - - - X XX - 6-9 / 8 V C3 Marine deposit L. shoreface C. sand - XXX - - - X XX - 9-18m more homogenous texture & colour than 6m-9m. Textural layering not associated with prominent colour contrast. 9-18 / 13 VI C4 Marine deposit L. shoreface C. sand - 9-18 / 11 VI C4 Marine deposit L. shoreface Med. sand - XXX - - - X X - 18-36 / cc V C5 Gleyed clay Fluvial paleo- (Sa.) clay - - - - - - XXX - Water dripping out 2-3m above bedrock. channel AK1/TRIG: Oakleaf Form (Orthic A - Neocutanic B - Unspecified). BMC, 90m terrace, 119 m a.m.s.l. (GPS). Terrain unit: crest. 0- / 1 A1 Orthic A, bio crust Aeolian Lo. sand Calc. 0.05- / 2 A2 Orthic A Aeolian Lo. sand - 0.3- / 3 B Neocarbonate B Aeolian Lo. sand Calc. XBr. - XXX X - Bedrock from 0.6 m. AK61H: Coega Form (Orthic A - Hardpan carbonate). BMC, 30m terrace, 28 m a.m.s.l. (GPS), footslope. 0- / 4 A Orthic A Aeolian F. sand Calc. Very little clay in horizon - 0.35- / 1 II C1 Hardpan carbonate Obscured Calc. XX XXX XX XXX X - 1.05- / 2 II C3 Blocky calcretized sed. Marine Calc. cutan 1.55- / 3 II C4 Marine deposit Marine - Bedrock from 2.1 m. LCK1-5: Namib Form (Orthic A - Regic sand). KNC, 50 m package, 30 m a.m.s.l. (GPS), footslope-valley bottom. 0- /A A Regic sand (Orthic A) Aeolian Med. sand - Very little clay in horizon X - - 0.3- / B B Regic sand Aeolian Med. sand - - - - - XX - - 1- / 1 II B2 Neocutanic B Obscured Med. sand - XX - - - XXX XX - 1.3- / 3 II B3 Neocutanic, luvic Obscured Med. sand - 1.9- / 4 III C1 Gleyed sand? Obscured Med. sand - 2.4- / 6 V C3 Sedimentary? Obscured Med. sand - 3.4- / 7 VI C4 Marine deposit U. shoreface C. sand - Pale sand (LKC1-5/8) interlayered with more orange sand (/9). 5.4- / 8 VII C 5 Marine deposit L. shoreface Med. sand - XXX - - - - - - 5.4- / 9 VII C 6 Marine deposit L. shoreface Med. sand - XX - - - - - - Bedrock from 12 m. OBT: Prieska Form (Orthic A - Neocarbonate B - Hardpan carbonate). BMC, 95 m terrace, 99 m a.m.s.l. (GPS), crest. 0- / 1 rs Aeolian deflation surface Aeolian Med. sand - 0.03- / 2 A Orthic A Aeolian Med. sand - 0.25- / 3 B Neocarbonate B Aeolian Med. sand - - - XX Br. - - - - 0.45- / 4 II C1 Hardpan carbonate Obscured Calc. - - XXX X X - - 1- / 5 II C2 Sedimentary? Obscured Lo. sand Occ. calc. Bedrock not reached. SL4-1: Pinedene Form (Orthic A-Yellow-brown apedal B-Signs of Wetness). KNC, 30 m package, 46 m a.m.s.l. (GPS), footslope. 0-0.3 / 1 A1 Orthic A Aeolian Med. sand - - - - - XX X - 0.3- / 2 B1 Yellow-brown apedal B Aeolian Med. sand - - - - - XXX X - 0.5- / 3 E Gleyed sand Aeolian Med. sand - - - - - XX X - 0.63- / 4 B2 Neocutanic Obscured Med. sand - X - - - XXX XX - 0.8- / 5 II B3 Slightly gleyed sand? Obscured Med. sand - - - - - XX X - 0.95- / 6 III B4 Pedocutanic Obscured Lo. sand - X - - - XXX XX - 0.97- / 7 III B5 Pedocutanic Obscured Calc. 1.22- / 8 III B6 Pedocutanic Obscured Lo. sand - 2.45- / 9 IV C1 Marine deposit U. shoreface Lo. sand - 3.05- / 11 V C2 Marine deposit L. shoreface Med. sand Calc. Bedrock from 9 m, evidence of wetness. ? Classified according to the Soil Classification Working Group (1991). Roman numerals indicate different sedimentary units. ? ? XXX: dominant peak; X: trace/only <0.08 ?m fraction; -: not detected. Br.-broad; c.-coarse; calc.-calcareous; f.-fine; med.-medium; l.-lower; lo.-loamy; occ.-occasional; sed.- sediment; sl.-slight; u.- upper; v.-very. --------------Br.-------------- No clay No clay -------Br.------- No clay No clay -------Br.------- 97 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; P aly. 0.4 49 nm Q u artz ; Se pio lite; Mi ca 0.3 34 nm Ca lcite 0.3 04 nm Mic a; s epi olit e ( d iff.) 0.5 04 nm Sep iolit e 0 .37 4 n m <2 ?m; Mg <2 ?m; Mg; EG Silt; Mg Silt; Mg; EG Milled; acetate; <2 ?m; Mg Milled; <2 ?m; Mg 2.4 nm 6869/iv Neocarbonate 0.6-0.8 m depth 1-0.08 ?m; Ca 1-0.08 ?m; Ca; EG <0.08 ?m; Ca <0.08 ?m; Ca; EG 1-0.08 ?m; Ca; EG 10 weeks <0.08 ?m; Ca; EG 10 weeks -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.1: (a): XRD traces of clay fraction from profile 6868. Bold lines show kerolite peak positions. Diff.- diffuse; EG: ethlyene glycol. ?Milled EG? and ?milled acetate EG? traces identical to ?<2 ?m EG?. 98 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; P aly. 0.4 49 nm Q u artz ; Se pio lite; Mi ca 0.3 34 nm Ca lcite 0.3 04 nm Mic a; S epi olite ( di ff.) 0.5 04 nm Sep iolit e 0 .37 4 n m <2 ?m; Mg 6869/5 Hardpan carbonate 0.32-1 m depth <2 ?m; Mg; EG <2 ?m; Mg 6869/4 Yellow-brown apedal B - 0.32 m depth 1.4 2 n m <2 ?m; Mg; EG 1-0.08 ?m; Ca 1-0.08 ?m; Ca; EG <0.08 ?m; Ca <0.08 ?m; Ca; EG <0.08 ?m; Ca; EG 10 weeks 1-0.08 ?m; Ca; EG 10 weeks -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.1: (b): Continued - XRD traces of clay fraction from profile 6868. 99 Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite 1.7 8 n m Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Pal ygo rsk ite 0.4 49 nm Q u artz ; M ica 0.3 34 nm Mic a 0 .5 n m <2 ?m; Mg 6869/8 Lower shoreface marine sediment (coarse sand) 6-9 m depth <2 ?m; Mg; Glycerol <2 ?m; Mg 6869/11 Lower shoreface marine sediment (medium sand) 9-11 m depth <2 ?m; Mg; Glycerol <2 ?m; Mg <2 ?m; Mg; Glycerol 6869/7 Lower shoreface marine sediment (medium sand) 6-9 m depth -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.1: (c): Continued - XRD traces of clay fraction from profile 6868. palygorskite on the coastal plain is consistent with the results of Singer et al. (1995, p. 45), who found that sepiolite-containing soils were concentrated toward the coast, and palygorskite-containing soils farther inland. The XRD traces from two of the horizons (bulk soil) overlying sepiolite-containing calcretes (AK61H, Figure 4.2; OBT/3, Figure 4.5) and a neocarbonate B hori- zon over hard rock (AK1/TRIG/3, Figure 4.3) are typified by very broad 1.21 nm and 0.4 to 0.374 nm peaks, suggesting a very fine or poorly crystalline se- piolite phase. TEM analysis of OBT/3 (Figure 4.7(a)) and 6869/iv (Figure 4.10) showed an amorphous phase in addition to crystalline sepiolite and mica. The broad rise in the 0.4 to 0.374 nm region also occurred in traces with a sharper sepiolite peak, such as the calcrete horizon OBT/4. This suggests there is an amorphous silica component in addition to sepiolite in the clay fraction. In AK1/TRIG/3 (Figure 4.7(b)), acicular crystals are apparently developing from an amorphous phase, although well crystallized sepiolite is also present (Figure 100 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; P aly. 0.4 49 nm Q u artz ; Se pio lite; Mi ca 0.3 34 nm Ca lcite 0.3 04 nm Mic a; S epi olite ( di ff.) 0.5 04 nm Sep iolit e 0 .37 4 n m Milled; acetate; <2 ?m; Mg Milled; <2 ?m; Mg Milled; acetate; <2 ?m; Mg; EG AK61H/1 Hardpan carbonate 0.35-1.05 m depth Milled; acetate; 1-0.08 ?m; Ca Milled; acetate; 1-0.08 ?m; Ca; EG Milled; acetate; 1-0.08 ?m; Ca; EG 10 weeks Milled; acetate; <0.08 ?m; Ca Milled; acetate; <0.08 ?m; Ca; EG Milled; acetate; <0.08 ?m; Ca; EG 10 weeks <2 ?m; Mg AK61H/4 Orthic A 0-0.35 m depth <2 ?m; Mg; EG 1.4 2 n m Milled; 1-0.08 ?m; Ca Milled; <0.08 ?m; Ca <2 ?m; Mg <2 ?m; Mg; Glycerol -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.2: XRD traces of clay fraction from profile AK61H. Bold lines show kerolite peak positions. Diff.-diffuse; EG: ethlyene glycol. ?Milled EG? traces identical to ?milled acetate EG?. 101 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; P aly. 0.4 49 nm Q u artz ; Se pio lite; Mi ca 0.3 34 nm Ca lcite 0.3 04 nm Mic a; S epi olite ( di ff.) 0.5 04 nm Sep iolit e 0 .37 4 n m <2 ?m; Mg <2 ?m; Mg; EG AK1/TRIG/3 Neocarbonate 0.3-0.6m depth Op al A 0.4 03 nm 1-0.08 ?m; Ca 1-0.08 ?m; Ca; EG 1.4 2 n m 1-0.08 ?m; Ca; EG 10 weeks -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.3: XRD traces of clay fraction from profile AK1/TRIG. Bold lines show kerolite peak positions. Diff.-diffuse; EG: ethlyene glycol. 4.7(c)). The amorphous silica may be available for sepiolite formation via a pro- cess similar to that described by D??az Torres et al. (1990), in which sepiolite and palygorskite fibres develop from silica gel, or as suggested by Blank and Fosberg (1991), by which sepiolite may form via the addition of Mg to opal. The hardpan carbonate horizon 6869/5 (Figure 4.1(b)) differs from the other calcrete samples (Figures 4.2, 4.5 and in Chapter 3) in that there is a broad 1.4 nm non-expanding phase dominant and a less dominant 1.21 nm sepio- lite peak. Chlorite occurred in the palygorskite-containing soils examined by Singer et al. (1995). Hay and Wiggins (1980), however, found some samples in the southwestern USA that showed a delicate fibrous structure under TEM, but which gave a broad peak from 1.26 to 1.4 nm that was unchanged with ethylene glycol and heating to 300 ?C, and only slightly reduced by heating to 500 ?C. There is no discernible trend in mica abundance with depth (Table 4.1). In the Namaqualand soils, the mica is likely to be detrital. Mica is a common component of the bedrock in the region and is also detectable macroscopically in many of the marine deposits, especially the lower shoreface facies. The aeolian deposits are 102 Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite 1.7 8 n m Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Pal ygo rsk ite 0.4 49 nm Q u artz ; M ica 0.3 34 nm Ca lcite 0.3 04 nm Mic a 0 .5 n m 1.4 2 n m LKC1-5/9 Lower shoreface marine sediment 5.2-12 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol LKC1-5/8 Lower shoreface marine sediment 5.2-12 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol LKC1-5/1 Neocutanic B 1-1.3 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol LCK1-5/B Mg Regic Sand 0.3-1 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol <2 ?m; Mg LCK1-5/A Mg Regic Sand 0-0.3 m depth -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.4: XRD traces of clay fraction from profile LKC1-5. Bold lines show kerolite peak positions. generally derived from the mechanical reworking of this marine material (Tankard and Rogers, 1978; Harmse and Swanevelder, 1987; Pether, 1994). Bu?hmann et al. (2004) found for soils in the Western Cape (Mediterranean climate) that mica was generally inherited from the parent material, and decreased in proportion to present day mean annual rainfall. Mica neoformation, however, can also take place under 550 mm annual rainfall, provided soils have a pool of exchangeable K and swelling clays as a precursor phase (Bu?hmann et al., 2004). Smectite generally increases downwards (Table 4.1) dominating the clay fraction of the marine sands at the base of the profiles. In the case of LKC1-5, a non- calcareous, non-sepiolite/palygorskite bearing profile, it appears to increase at the expense of mica/illite (Figure 4.4). 103 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Sep iolit e; P aly. 0.4 49 nm Q u artz ; Se pio lite; Mi ca 0.3 34 nm Ca lcite 0.3 04 nm Mic a; S epi olite ( di ff.) 0.5 04 nm Sep iolit e 0 .37 4 n m <2 ?m; Mg <2 ?m; Mg; EG OBT/4 Hardpan carbonate with sepiolite cutans 0.45-1 m depth Op al A 0.4 03 nm Milled; <2 ?m; Mg Milled; <2 ?m; Mg Milled; <2 ?m; Mg; EG OBT/3 Neocarbonate B 0.25-0.45 m depth 1.4 2 n m -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.5: XRD traces of clay fraction from profile OBT. Bold lines show kerolite peak positions. Diff.-diffuse; EG: ethlyene glycol. There is a slight increase in the kaolinite content down the profile. The basal parts of the regolith profile are marine sands, which are older than the subaerially reworked upper parts. The basal parts are also wetter, with many profiles having water running out just above or on bedrock. The leaching process in the basal part of the regolith has therefore been more intense than in the upper parts. This is particularly evident in profile 6869, where the lowest part of the profile consists of much older kaolinized fluvial sediments in a bedrock channel underlying the marine sands. Pether (1994) concluded that the channels were laid down as a quartzo-feldspathic sediment in a fluvial environment, and subsequently deeply weathered, with extensive alteration of feldspar to kaolinite. De Villiers and Cadman (2001) favoured an early Tertiary age for this channel north of the Swartlintjies River, although it may also contain reworked Cretaceous material. The overlying marine sediments of profile 6869 are from the 30 m package, the age of which is not well constrained, being estimated at ca. 3.3 Ma or younger (Pether et al., 2000). Bu?hmann et al. (2004) found for soils in the Western Cape that the transformation of chlorite to kaolinite was the most significant pedogenic 104 Pal ygo rsk ite 1.0 48 nm Sm ect ite 1.5 nm Sm ect ite 1.7 8 n m Mic a; K ero lite 1.0 nm Kao linit e 0 .71 6 n m Pal ygo rsk ite 0.6 44 nm Pal ygo rsk ite 0.5 42 nm Q u artz 0.4 26 nm Pal ygo rsk ite 0.4 18 nm Kao linit e 0 .35 7 n m Pal ygo rsk ite 0.3 5 n m Pal ygo rsk ite 0.3 23 nm Pal ygo rsk ite 0.4 49 nm Q u artz ; M ica 0.3 34 nm Ca lcite 0.3 04 nm Mic a 0 .5 n m 1.4 2 n m SL4-1/5 Gleyed sand? 0.8-0.95 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol SL4-1/3 Gleyed sand 0.5-0.63 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol <2 ?m; Mg SL4-1/2 Yellow-brown apedal B 0.3-0.5 m depth <2 ?m; Mg; Glycerol SL4-1/1 Orthic A 0-0.3 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol SL4-1/6 Pedocutanic 0.95-0.97 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol SL4-1/4 Neocutanic 0.63-0.8 m depth <2 ?m; Mg <2 ?m; Mg; Glycerol -5 0 5 10 15 20 25 30 35 40 Cu K? ? 2 ? Figure 4.6: XRD traces of clay fraction from profile SL4-1. Bold lines show kerolite peak positions. 105 (a) <2 ?m (milled; acetate-treated) fraction from OBT/3 (Figure 4.5): mica (platy); sepiolite (acicular); amorphous phase (center); aggregated silica(?) spheres. Scale bar 500 nm. (b) 1 to 0.08 ?m fraction from AK1/TRIG/3 (Figure 4.3). Scale bar 200 nm. (c) Same as (b): mica (platy); sepiolite (acicular); aggregated silica (?) spheres. Scale bar 500 nm. Figure 4.7: TEM images of clay fractions. 106 process in soils derived from metamorphic rocks, and that the kaolinite content was poorly related to the present day mean annual precipitation, and more likely associated with the chlorite content in the parent rock. Although the soils in the region of the present study on the Namaqualand coast were not derived from metamorphic rock directly, the bedrock underlying the marine deposits on which the soils formed is metamorphic (see references in Chapter 2), and so this theory of kaolinite formation may also apply here. 4.3.3 Kerolite Kerolite has broad asymmetric XRD peaks at approximately 1.0 and 0.45 nm, and a less intense 0.32 nm peak (Brindley et al., 1977; Le?veille? et al., 2002). These peaks overlap with other phases present in the samples: the palygorskite 1.048 nm and mica 1.0 nm peaks; the sepiolite 060 and palygorskite 040 peaks at 0.449 nm (Bradley, 1940; Brindley, 1959); and the palygorskite 400 peak at 0.323 (Bradley, 1940) or 0.317 nm (Christ et al., 1969). As a result, all samples with a mix of these phases show peaks in the correct position for kerolite (Figures 4.1(a) and (b), 4.2, 4.3, and 4.5). These patterns show no change in the 1.0 nm peak after ethylene glycol treatment for ten weeks, which is similar to the findings of Le?veille? et al. (2002), but unlike that of Brindley et al. (1977) and Hay et al. (1995). It may be hypothesized that due to the fine, disordered nature of kerolite (Brindley et al., 1977; Pozo and Casas, 1999), its presence would be better resolved by a study of the fine clay fraction. The concentration of palygorskite in the fine clay fraction, however, means that this was not the case. TEM data were also inconclusive, with both Figures 4.7(a) and (b) (corresponding to the XRD traces of 4.3 4.7(a)), for example, showing poorly crystalline material, although Figures 4.10(a) and (b) do resemble the ?colloform aggregates? of Pozo and Casas (1999, p. 402). 4.3.4 Chemical composition of clay fraction Trends in MgO, Al2O3 and SiO2 (Figure 4.8(a)) show that Namaqualand soil clay fractions lie on a mixing line between sepiolite and mica end-members, with a contribution from smectite. This is consistent with the XRD and TEM results. The soils have a similar chemical composition to the Namaqualand sepiolite- bearing soils studied by Singer et al. (1995). 107 The influence of mica/illite on the clay fraction chemistry is reflected by the positive correlation between K2O and Al2O3 (Figure 4.8(b)). The trend is also observed in the K-saturated soils where it is displaced to higher K2O values, and it is very similar to the trend displayed by the data of Singer et al. (1995) for sepiolite-containing soils from nearly the same area. The palygorskite-containing soils of Singer et al. (1995) (from an area more inland of the present study area) are displaced to higher Al2O3 values in comparison to both the soils from this study, and the sepiolite-containing soils from their study. This displacement is more pronounced at lower K2O values, and is consistent with the elevated Al2O3 contents of pure palygorskite compared to pure sepiolite (Figure 4.8(b)). The K2O content of the Mg-saturated clay fractions ranges from 0.7 to 3.57 %. This is equivalent to 10 to 50 % illite, following the method of Singer et al. (1995) in using 7.1 % K2O for average illite and assuming that all the K2O is attributed to mica/illite, since the XRD traces display no prominent feldspar peaks. If, however, the main potassium-bearing phase is mica rather than a relatively K- depleted illite, the maximum calculated mica content is 38 % (using 9.35 % K2O as the average mica from the mean K2O content of the mica analyses listed in Figure 4.8). The present data show a good correlation between Fe2O3 and TiO2 (Figures 4.8(c) and (d)). This can be attributed to the presence of mica in the soils, since the soil clays fall on the mica Fe2O3-TiO2 trendline (Figure 4.8(d)) and show similar Fe2O3/TiO2 ratios to mica (compare the trendline in Figure 4.8(c) for the sepiolitic and non-sepiolitic soils with the trendline for single mineral micas shown in Figure 4.8(d)). In contrast, pure kerolite, sepiolite, palygorskite (Figure 4.8(c)) and smectite (Figure 4.8(d)) show a poor correlation between Fe2O3 and TiO2. The present data suggest the Fe2O3 in Namaqualand soils is due to the mica content. This is in contrast to the conclusion of Singer et al. (1995), that Namaqualand sepiolite is a ferric sepiolite. Their conclusion was based on XRF data from whole clay fractions that were CBD-pretreated to remove free iron oxides. Although illite was noted to be present, Singer et al. (1995) did not publish the TiO2 content of their soils. Also, the interpretation that Fe2O3 in Namaqualand soils is due to the mica content is more consistent with the results of Galan and Carretero (1999), who found sepiolite to be a true triocta- 108 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 K 2 O wt% Al 2O 3 wt% This study: sepiolitic/palygorskitic soil Kerolite This study: no sepiolite/palygorskite Sepiolite Singer et al (1995): Namaqualand sepiolitic soil Palygorskite Singer et al (1995): Namaqualand palygorskitic soil Smectite K-sat. 40 60 80 100 0 10 20 30 40 (a) Al 2 O 3 wt% Mg O+SiO 2 wt% Kerolite or Mica? Mg-sat HK-sep-?silica K-sat Brindley et al 1977 Kerolite Eberl et al 1982 Kerolite Stoessell and Hay 1978 Kerolite Stoessel 1988 Kerolite L?veill? et al 2002 Kerolite EMP Stoessel 1988 Sepiolite Stoessell and Hay 1978 Sepiolite Singer p834 Sep Singer p834 Paly Hay et al 1995 Ke/St minor Sep SKB 1995 NM decalc, defer, Mg-sat sep SKB 1995 NM decalc, defer, Mg-sat paly sepiolite non mica Series21 smectite Mica Kerolite Sepiolite Palygorskite Saponite Nontronite Montmorillonite Beidellite R 2 = 0.9754 0 5 10 15 20 0 1 2 3 4 (b) K 2 O wt% Al 2O 3 wt% This study: sepiolitic/palygorskitic soil This study: no sepiolite/palygorskite Kerolite Sepiolite Palygorskite Singer et al (1995): sepiolitic soil Singer et al (1995): palygorskitic soil K-sat. R 2 = 0.7609 0 2 4 6 8 10 0.0 0.2 0.4 0.6 0.8 1.0 (c) TiO 2 wt% Fe 2O 3 (Total) wt% Palygorskite Sepiolite Kerolite Trendline mica: R 2 = 0.7636 0 5 10 15 20 25 30 0.0 0.5 1.0 1.5 2.0 2.5 3.0 (d) TiO 2 wt% Fe 2O 3 (Total) wt% Nontronite Figure 4.8: Trends in total element composition of Mg- and K-saturated clay fractions from Namaqualand soils compared to pure minerals. (a) to (c): R2 trendlines based on sepiolitic and non-sepiolitic soils from the present study; (d) R2 trendline based all the mica data. Note scale differences in (c) and (d), where (c) shows data from this study plotted with kerolite, sepiolite, palygorskite, and (d) plotted with mica and smectite Sepiolite (Stoessell and Hay, 1978; Stoes- sell, 1988; Singer, 1989), kerolite (Brindley et al., 1977; Stoessell and Hay, 1978; Eberl et al., 1982; Stoessell, 1988; Le?veille? et al., 2002), palygorskite (Singer, 1989), and mica and smectite compiled by Deer et al. (1992) - rose muscovite , muscovite, glauconite, phlogopite, biotite, lepidolite, beidellite, saponite, montmorillonite and nontronite. hedral mineral, very pure (near end-member) with negligible structural substitu- tion and eight octahedral positions filled with magnesium, close to Brauner and Preisinger?s (1956) theoretical formula Si12Mg8O30(OH)4(OH2)4.8H2O. Galan and Carretero?s (1999) EDX analyses of sepiolite showed no structural Al and Fe, and an octahedral occupancy close to 8. They attributed the presence of some Al, Fe, K or Ca in some of the determinations to minor impurities. 109 4.3.5 Sepiolite/palygorskite modes of occurrence Chapter 3 showed that fibrous clay minerals in Namaqualand calcretes are associ- ated with mineralized filaments and needle fibre calcite, and also occur as frayed mats composed of woven fibres. ?Fraying? is a very common texture in sepio- lite/palygoskite occurrences. It is either reported from a sheet that is composed of fibres (Khoury et al., 1982; Jamoussi et al., 2003), or a sheet which appears to be solid (such as Monger and Daugherty, 1991; Pozo and Casas, 1999). Suarez et al. (1994) showed the fraying observed under TEM to be illite particles evolv- ing laterally into palygorskite fibres. SEM analyses show that much of the sepiolite/palygorskite in this study occurs as fringed sheets (Figure 4.9(a), (b)). These are found coating (rather than evolving from) mica/illite particles (Figures 4.9(c)-(d)), as free-standing mats such as those shown in Chapter 3, and are common on the grain-side of cutans (for example Figures 4.9(e)-(f)). While the textures in Figures 4.9(b)-(f) suggest illuviation of the fibrous clay minerals, another explanation may be that sites such as that immediately adja- cent to silicate grains have the highest concentration of silica for their formation. Many examples of etched and embayed silicate grains have been observed in this study. Monger and Daugherty (1991) concluded that silica released from silicate grains by pressure solution during calcite crystallization was required for paly- gorskite formation in New Mexico aridisols. While Figure 4.9(d) shows distinctly that mica is not evolving laterally to a fibrous mineral, Figure 4.9(a) shows fi- bres developing from a wavy sheet-like mineral. This resembles the SEM images of mixed-layer kerolite/smectite associated with sepiolite from the Amargosa Desert, Nevada (Eberl et al., 1982, p. 324) and the Amboseli Basin, Tanzania- Kenya (Hay et al., 1995, p. 461). 4.3.6 Palygorskite/sepiolite transformation In any discussion of sepiolite or palygorskite, particularly palygorskite and its origins, the issue of transformation to/from a sheet silicate becomes relevant: se- piolite/palygorskite being transformed to smectite (Bigham et al., 1980; Golden et al., 1985), palygorskite fibres developing from an illite (Suarez et al., 1994), 110 (a) Fringed sheets, etched 15 sec. in 1M HCl. Scale bar 2 ?m. (b) Fringed texture. Scale bar 2?m. (c) Magnification of (b), showing fibres coating mica grains and fringed texture. Scale bar 1?m. (d) Fibres on mica grain (EDX analysis not shown), but no fringing. Scale bar 2?m. (e) Sepiolite/palygorskite cutan. Scale bar 100 ?m. Fe from sample-holder. (f) Magnification of cutan in (e). Scale bar 1 ?m. Figure 4.9: SEM images of sepiolite/palygorskite occurrences in calcareous nod- ules from 6869/iv (XRD traces shown in Figure 4.1(a)). 111 or no evidence of a solid state smectite-palygorskite transition (Verrecchia and Le Coustumer, 1996; Akbulut and Kadir, 2003). Jones and Galan (1988, p. 654 to 657) reviewed the issue of transformation vs. direct precipitation in se- piolite/palygorskite genesis. Recent studies using trace elements and isotopes favoured a direct precipitation of sepiolite, and a transformation or dissolution- reprecipitation of illite and/or smectite as a mechanism for palygorskite forma- tion (Torres-Ru??z et al., 1994; Sa?nchez and Gala?n, 1995; Lopez-Galindo et al., 1996; Jamoussi et al., 2003; Zaaboub et al., 2005). This study has found no TEM evidence to favour a hypothesis for fibrous mineral degradation to sheet silicates, as was suggested Singer et al. (1995). Figure 4.7(b) shows what seems to be the formation of sepiolite, and Figures 4.7(c), 4.10(a) and 4.10(b) show what seems to be well crystalline sepiolite/palygorskite, the only exception being where a fibre bundle was burned by the electron beam (Figure 4.10(c)). Singer (1989) noted that the variation in morphology of sepiolite and palygorskite crystals under TEM is greater within than between the species, and suggested that may be due to different pretreatments used, or to transport in the case of detrital clays. Owliaie et al. (2006) found three morphological forms of palygorskite in relation to the degree of weathering: i) sheaves or bundles repre- senting the least weathering; ii) ?split? bundles and sharp-pointed crystals, some of which may have been transformed to other clay minerals such as smectite; and iii) individual fibres with rounded tips and no sharp edges, indicating maximum weathering. The TEM images (Figures 4.7 and 4.10) often show sepiolite/palygorskite to occur as flat aggregates giving the appearance of frayed edges, as well as sin- gle fibres: they resemble the palygorskite ?nanofibres? and ?fibre bundles? of Esquivel et al. (2005), the palygorskite textures found by Verrecchia and Le Coustumer (1996) and Khademi and Mermut (1999). None of these authors attributed the textures to transformation into a sheet silicate. The large fibre bundle in Figure 4.10(c) resembles the TEM image of sepiolite precipitated by Siffert and Wey (1962). The fibre bundles are not only evident in the images of the mechanically and chemically treated clay extracts: an in situ fibre bundle in an etched calcareous nodule is shown in Figure 4.10(d). 112 (a) TEM: 1 to 0.08 ?m fraction. Scale bar 200 nm. (b) TEM: <0.08 ?m fraction. Scale bar 200 nm. (c) TEM: <0.08 ?m fraction; beam damage arrowed. Scale bar 500 nm. (d) SEM: palygorskite/sepiolite bundle in calcareous nodule etched 15 sec. in 1M HCl. Scale bar 1 ?m. Figure 4.10: TEM images of sepiolite/palygorskite fibres and fibre bundles from 6869/iv (XRD traces shown in Figure 4.1(a)). 4.4 Conclusions Milling before separating the clay fraction reduced the sharpness of all peaks. Acetate treatment after milling substantially enhanced the clay peaks, which suggests that the clays are aggregated with calcite. The calcite aggregates bind- ing the clays appear to be finer than the size fraction that can be liberated with ball milling, because it is only in the <0.08 ?m fraction that acetate treatment appeared not to liberate any further clays. The <0.08 ?m fraction is the only size fraction in which palygorskite could be detected before acetate treatment. It may be that it is finer than the other clay 113 minerals, and finer than the size fraction which can be liberated with ball milling. Sepiolite was difficult to distinguish from palygorskite in both TEM and SEM images and so this could not be visually confirmed. Sepiolite is more prominent than palygorskite in the XRD traces. It is unlikely that these fibrous clay minerals are inherited from either the marine or aeolian parent materials, because neither sepiolite nor palygorskite are present in the marine sediments (which contain only smectite, mica, and kaolinite in varying proportions), and although the appearance of sepiolite and palygorskite seems to be limited to subaerially-reworked material, not all aeolian material contains sepiolite/palygorskite. Sepiolite and palygorskite are most consistently associ- ated with the presence of calcite in the soil profile. Sepiolite occurs across all packages and elevations up to 99 m a.m.s.l., the farthest sampled from the sea being about 30 km inland. There is little trend in mica abundance with depth. In the Namaqualand soils, the mica is likely to be detrital - it is a common component of the bedrock in the region, and also detectable macroscopically in many of the marine de- posits, especially the lower shoreface facies. Smectite content generally increases downwards, and dominates the clay fractions of the marine sands at the base of the profiles. In the case of LKC1-5, a non-calcareous, non-sepiolite/palygorskite bearing profile, it appears to increase at the expense of mica/illite. There is a slight increase in the content of kaolinite down the profile. The basal portion of the regolith profile comprises marine sands which are older than the subaerially reworked upper parts. The lower parts are also wetter, with many profiles hav- ing water seepage above the bedrock where the leaching process has been more intense and protracted than in the upper parts. Some SEM images show fibres developing from a wavy sheet-like mineral. These resemble the SEM images of mixed-layer kerolite/smectite associated with sepi- olite from the Amargosa Desert (Nevada) and the Amboseli Basin (Tanzania- Kenya). This study was not able to find conclusive evidence for or against the presence of pedogenic kerolite in the clay fraction, although it does not appear to be a dominant phase in the less than 2 ?m fraction. The positive correlation between K2O and Al2O3 is a reflection of the ubiquitous mica in the sediments. The K2O content of the Mg-saturated clay fractions ranges from 0.7 to 3.57 %, equivalent to 10 to 50 % illite or up to 38 % mica in the clay fractions. 114 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. This is consistent with the XRD and TEM results. There is a strong correlation between Fe2O3 and TiO2. This can be attributed to the presence of mica in the soils, since the soil clays fall on the mica Fe2O3-TiO2 trendline with similar Fe2O3/TiO2 ratios to mica, rather than a ferric sepiolite. 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 magnifica- tion often reveals 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. One TEM image shows what was interpreted to be acicular crystals developing from an amorphous phase, although well crystalline sepiolite was also present. 115 Chapter 5 Brine evolution and geochemical barriers: mineral formation in Namaqualand soils 5.1 Introduction Pedogenic sepiolite dominates the clay fraction of calcareous soils on the Namaqua- land coastal plain, sometimes accompanied by palygorskite (Singer et al., 1995, and Chapter 4). The occurrence of sepiolite in soils, especially as a dominant mineral, is generally reported to be rare except for some calcic and petrocalcic horizons. It is more common in arid-region alkaline lake sediments such as in the Tertiary Madrid Basin (Torres-Ru??z et al., 1994; Sa?nchez and Gala?n, 1995; Lopez- Galindo et al., 1996) and Tunisian sequences (Jamoussi et al., 2003; Zaaboub et al., 2005). This is an environment which in some cases is difficult to distin- guish from the pedogenic (Singer, 1989). The genesis of both sepiolite and palygorskite requires solutions with high pH, soluble Mg and Si. Palygorskite requires additional Al. Sepiolite requires a slightly higher pH and higher (Mg+Si)/Al than palygorskite (Jones and Galan, 1988, p. 664). This Al for palygorskite formation may be obtained through a pre-existing sheet silicate, possibly through a dissolution-reprecipitation mech- anism (Elprince et al., 1979; Galan and Ferrero, 1982; Singer, 1989; Jones and Galan, 1988; Lopez-Galindo et al., 1996). Jones and Galan (1988, p. 663) noted 116 that the water-accommodating structure of sepiolite is more suited to direct precipitation from solution than that of palygorskite. Some authors, however, have favoured the idea of a neoformation of palygorskite: Singer and Norrish (1974), for example, found no precursor minerals such as smectite, and Mon- ger and Daugherty (1991) favoured neoformation of palygorskite based on its morphology. Recent trace element and isotopic data, however, have tended to show that palygorskite is transformed whereas sepiolite is neoformed. Sa?nchez and Gala?n (1995) found that palygorskite was formed by transformation of il- lite and smectite involving reaction with Mg, and sepiolite was a neoformation product favoured by evaporation of silica- and Mg-bearing waters. Zaaboub et al. (2005) showed that sepiolite would have precipitated directly in lacustrine, playa-lake or sebka environments under alkaline conditions, high Si, high Mg and low Al activity; whereas palygorskite would have formed by transformation of already existing illite and/or smectite type aluminosilicates in solutions in equi- librium with isotopically heavier, and therefore more evaporated, solutions than the sepiolite. In the pedogenic setting, Bouza et al. (2007) concluded that the association of calcite, fluorite, sepiolite, and possibly opal-CT in Argentinian Aridisols indi- cated a successive precipitation of these minerals under alkaline conditions during evaporation. Singer and Norrish (1974), considering occurrences of palygorskite in Australian soils, suggested that a soil solution having a similar composition to that of lakes from arid regions where palygorskite typically occurs, would itself constitute a suitable medium for palygorskite formation. Mineral stability dia- grams have subsequently been applied to the genesis of sepiolite and palygorskite in soil environments (Singer and Norrish, 1974; Elprince et al., 1979; Fernandez- Marcos et al., 1979; Monger and Daugherty, 1991; Ducloux et al., 1995; Bouza et al., 2007). Coastal Namaqualand soils experience greater evaporation than precipitation, and exhibit macroscopically many of the typical evaporite minerals (calcite, gyp- sum, halite). The presence of eluvial and bleached subsurface horizons, and wetness above layers of reduced permeability, suggest the presence of (tempo- rary) subsurface water (Chapter 2 and Francis et al. (2007)). Also, many of the profiles have a calcrete horizon (with abundant sepiolite in the clay frac- tion, and accessory palygorskite) at the transition from the upper, finer aeolian- deposited sediments, to their underlying coarser upper shoreface marine sedi- 117 ments (Chapter 2). This fine-coarse transition has been identified by authors such as Yaalon and Wieder (1976), Singer (1989), Singer (2002) and Bouza et al. (2007) as the position at which sepiolite and palygorskite precipitation takes place, due to temporary waterlogging resulting from the accumulation of water at the boundary, until it attains the higher fluid pressure needed to fill the larger pores of the horizon below (Rode, 1969, cited by Yaalon and Wieder (1976)). The soil mantle in an arid environment is analogous to a closed basin in that it loses more water by evaporation than it loses by hydraulic flow. Although Chapter 2 showed the presence of preferential vertical water movement through the regolith to bedrock, this loss of water through evaporation as the dominant mechanism of mineral formation is evident in the accumulation of evaporite min- erals (ranging in the Namaqualand soils from halite and gypsum to calcite and sepiolite). The upward movement of water (vapour) through the soil due to the nocturnal chilling of the surface was demonstrated by (Prinsloo, 2005) and is outlined by Francis et al. (2007). The presence of these evaporite minerals will influence the composition of the soil solution (appropriately simulated through making an extract of a saturated paste) to a degree dependent on both the con- centration of the more soluble minerals (e.g. halite) and the solubility of the less soluble minerals (e.g. calcite, sepiolite). This Chapter hypothesized, therefore, that the permeable upper horizons in the 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 main objective was to apply the principles of brine evolution, particu- larly those of the Hardie-Eugster model (Hardie and Eugster, 1970; Eugster and Hardie, 1978; Eugster and Jones, 1979), to sepiolite genesis in arid soils of the Namaqualand coastal plain, assuming that in equilibrated soil solutions the pre- cipitating mineral controls the activity of one of its common ions (Kittrick, 1971). The Hardie-Eugster model was conceptualized for the evolution of closed-basin brines, and has since been applied to a diverse range of environments, from saline soil seeps (Timpson et al., 1986) to coastal pans (Smith and Compton, 2004) and a model of calcrete genesis (Wang et al., 1993). Evaporite minerals including calcite, gypsum, halite and sepiolite were consid- 118 ered by Garrels and Mackenzie (1967), Hardie and Eugster (1970) and Eugster and Jones (1979) to precipitate from solution during evaporation and evolution of closed-basin brines. The precipitation of these minerals affects the composi- tion of the remaining brine in a manner determined by the initial ratio of the component ions in the solution. Mineral precipitation was considered to con- stitute an important chemical divide in the evolving composition of the brine, but it is not the only mechanism which governs chemical fractionation in brines. Other mechanisms of solute depletion include sorption or exchange reactions on active surfaces, degassing, and bacterial reduction (Eugster and Jones, 1979). Conservative solutes, unaffected by such processes, can be used to measure the concentration effects. Depending on the removal mechanisms involved and their relative importance, each solute follows a different evolutionary path. Paths for a specific solute may differ from basin to basin, with the result that solute be- haviour can be used to delineate the important processes for a particular basin and aid in unravelling its chemical evolution. The objective was thus to use the variation in solute composition of equilibrated saturated paste extracts as a means of understanding the pedochemical history of mineral precipitation in Namaqualand soils. The present study was limited to those phases which were observed to be present and have been shown to be neoformed by precipitating from water: gypsum, halite, amorphous silica, calcite, dolomite (to represent Mg-rich calcite, even though no dolomite was detected in the calcretes analysed in Chapter 3), sepiolite and, hypothetically, kerolite, since its environment of formation is very similar to that of sepiolite and it was not possible either to confirm or refute its presence (Chapter 4). The Al-bearing phases that were detected with XRD can all be adequately explained by mineral transformation or dissolution re-precipitation: palygorskite (Jones and Galan, 1988; Sa?nchez and Gala?n, 1995; Zaaboub et al., 2005), kaolinite from feldspar weathering (Chapter 2), and smectite from illite/mica (Chapter 4). Although these transformations can only be achieved in a medium of suitable composition (Faure, 1998), for the purposes of this thesis only those phases that are neoformed were considered. Sepiolite was much more prominant on the XRD traces (Chapter 4) than palygorskite. Soil solution chemistry is used in this Chapter to refine our understanding of the origin of sepiolite, and to contrast this with the genesis of other minerals that formed in the process of evaporation of the soil solution on the Namaqualand 119 coastal plain. This Chapter focuses particularly on the genesis of sepiolite, be- cause of its relative rarity in the soil environment (although reports of it are increasing) and the uncertainties that sometimes surround its genesis. 5.2 Materials and methods Fieldwork took place in February 2002. Fourteen soil profiles to bedrock were examined and sampled, from south of Hondeklip Bay to north of Kleinzee, ap- proximately 150 km south of the Orange River on the west coast of South Africa. Samples were taken from the sidewalls of the mining excavations and from soil pits dug in the undisturbed areas nearby (described in Chapter 2). A summary of the landscape, climate and vegetation is given by Desmet (2007) and Chap- ter 2. Profile photographs are presented in Appendix A. The sepiolite on the Namaqualand coastal plain is clearly pedogenic in origin, since neither sepiolite nor palygorskite are present in the marine sediments (they contain only smec- tite, mica, and kaolinite in varying proportions), and although the appearance of sepiolite and palygorskite seems to be limited to subaerially-reworked material, not all aeolian material contains sepiolite/palygorskite (Chapter 4). Sepiolite and palygorskite are most consistently associated with the presence of calcite in the soil profile. Sepiolite occurs across all packages and elevations up to 99 m a.m.s.l., the farthest sampled from the sea (about 30 km). Saturated paste extracts were prepared for every sample that was analysed by X-ray diffraction (Chapters 1, 3, 4). The horizon types and their clay mineralogy are described in detail in those Chapters. Saturated pastes were prepared with 300g of < 2 mm material, to which water (approximately 100g, depending on the clay content of the soil) was added to form a saturated paste. A subset of samples was duplicated and the saturated pastes allowed to stand for 6 and 60 hours. There was very little difference in composition (and resulting saturation indices) after 24 hours of equilibration and so 24 hours was used for all subsequent samples. Samples were equilibrated at ambient temperatures and open to gas exchange with the atmosphere, and so effectively equilibrated with a pCO2 of -3.5 atm. The pH of the saturated paste was measured directly. The water from the saturated paste was extracted under vacuum and the following variables analysed: electrical conductivity (EC), using 120 a calibrated conductivity meter; alkalinity, by potentiometric titration to pH 4.5 with 0.01M HCl; major cations (Ca2+, Mg2+, K+, Na+) by atomic absorption spectroscopy; anions (Cl?, F?, NO?2 , NO ? 3 , Br ?, and SO2?4 ) by ion chromatog- raphy; Si using the blue silicimolybdous acid procedure of Weaver et al. (1968) (described by Hallmark et al., 1982), and P using the ascorbic acid-molybdenum blue method of Murphy and Riley (1962). Square brackets [ ] denote ion activities throughout this Chapter. The saturation index (SI) is the log(IAP/Keq) for a dissolution reaction (Drever, 1997, p. 25). When the SI = 0, the ion activity product (IAP) = K eq (the solubility product) which implies equilibrium with respect to the phase in question. A negative SI means that the IAP is less than K eq which implies undersaturation of the solution with respect to the phase, and conversely a positive SI implies supersaturation. The saturation index (SI) was used in preference to stability diagrams, because it is a numerical index and allows the variation of mineral stability to be plotted against an independent salinity variable. Stability diagrams also have the dis- advantage of requiring assumptions to be made about the activities of accessory solutes. Speciation and saturation index calculations were performed with the program PHREEQC (Interactive Version 2.8.0.0; Parkhurst and Appelo, 1999), run with the database file phreeq.dat. The solubility equations and equilibrium constants in phreeq.dat are: Halite: NaCl = Na+ + Cl? logK = 1.582 (5.1) Gypsum: CaSO4.2H2O = Ca 2+ + SO2?4 + 2H2O logK = ?4.580 (5.2) Quartz: SiO2 + 2H2O = H4SiO4 logK = ?3.980 (5.3) Amorphous silica: SiO2 + 2H2O = H4SiO4 logK = ?2.710 (5.4) 121 Calcite: CaCO3 = Ca +2 + CO2?3 logK = ?8.480 (5.5) Dolomite: CaMg(CO3)2 = Ca 2+ + Mg2+ + 2CO2?3 logK = ?17.090 (5.6) Sepiolite: Mg2Si3O7.5(OH).3H2O + 4H + + 0.5H2O = 2Mg 2+ + 3H4SiO4 logK = 15.760 (5.7) Two equilibruim constants for sepiolite are included in phreeq.dat : ?sepiolite? which uses log K = 15.760 from Stoessell (1988), and ?sepiolite (d)? which uses log K = 18.78 from Wollast et al. (1968). The former was used in this Chapter because it was calculated from samples equilibrated for nearly 10 years at 25o C and 1 atm and is in good agreement with that extrapolated by Christ et al. (1973) from higher temperatures. The log K determined by Wollast et al. (1968) was for a poorly crystalline, non-equilibrated sample (Christ et al., 1973; Stoessell, 1988). Kerolite was added to the phreeq.dat database file using the log K from Stoessell (1988): Mg3Si4O10(OH)2.H2O + 6H + + 3H2O = 3Mg 2+ + 4H4SiO4 logK = 25.79 (5.8) PHREEQC is a computer program written in the C programming language that is designed to perform a wide variety of low-temperature aqueous geochemical calculations (Parkhurst and Appelo, 1999) with capabilities for speciation and saturation-index calculations. PHREEQC uses ion-association and Debye Hu?ckel expressions to account for the non-ideality of aqueous solutions. This type of aqueous model is adequate at low ionic strength but may break down at higher ionic strengths (in the range of seawater and above) (Parkhurst and Appelo, 1999). The authors of PHREEQC attempted to extend the range of applicability of the aqueous model through the use of ionic strength terms in the Debye Hu?ckel expressions. These terms were fit for the major ions using chloride mean-salt activity-coefficient data. Thus in sodium chloride dominated solutions (as with the saturated paste extracts from Namaqualand Coastal Plain, see Section 5.3, Table 5.1) the model may be reliable at higher ionic strengths. For high ionic 122 strength waters, the specific interaction approach to thermodynamic properties of aqueous solutions should be used (Pitzer, 1973; Pitzer, 1979; Harvie and Weare, 1980; Harvie et al., 1984), but this approach is not incorporated in PHREEQC 2.8.0.0 (Parkhurst and Appelo, 1999). PHREEQC Version 2.12 uses the Pitzer activity coefficient model and so replaces PHRQPITZ (Plummer et al., 1988) for brine speciation calculations, using the pitzer.dat database. However, at the time of writing the pitzer.dat database included for download with PHREEQC Version 2.13.2 is of limited usefulness to soil applications since the database contains no data for silica and pedogenic clay minerals. The database file phreeq.dat consequently remains the most suitable for this system. 5.3 Results and discussion Seawater has an ionic strength of 0.67 according to the PHREEQC calculation of the seawater data from Nordstrom et al. (1979) (included with the PHREEQC distribution as Example 1 of the Help File). The speciation calculations showed the ionic strength of the samples to be in range of 0.004 to 0.76, with most of the samples falling well below the concentration of seawater (Figure 5.1). Given the discussion in Section 5.2, it suggests that PHREEQC is an adequate tool with which to perform speciation calculations in these soils. Watts (1980) noted an inverse relationship between the mol % MgCO3 in calcite and the sepiolite/palygorskite content, which suggested that these Mg-rich clay minerals are the main Mg-bearing phase in Botswana calcretes. The apparent absence of Mg-calcite and dolomite, and abundance of sepiolite in the calcretes of coastal Namaqualand suggests that a similar partitioning has occurred (Chapter 3) and so it has been assumed in this Chapter that sepiolite controls the Mg levels in the saturated paste extracts. The saturated paste extract data (Table 5.1) were divided into sepiolitic and non-sepiolitic categories based on XRD findings (Chapters 1, 3, 4) and then PHREEQC (Parkhurst and Appelo, 1999) was used to calculate the ion activities and mineral saturation indices. A few samples showed detectable Br?, NO3?, F? and NO2?, but these could not be accurately determined because the high salinity of the samples required a large dilution, resulting in the less abundant ions being diluted to below the detection limit of ion chromatography. 123 -5 -4 -3 -2 -1 0 -3 -2 -1 0log[Cl - ] log [Na + ] no sep Seaw Serie s -5 -4 -3 -2 -1 0 -3 -2 -1 0log[Cl - ] log [K + ] no sep K+ Series2 -5 -4 -3 -2 -1 0 -3 -2 -1 0log[Cl - ] log [C a 2+ ] no sep Ca+2 Series2 -5 -4 -3 -2 -1 0 -3 -2 -1 0log[Cl - ] log [M g 2+ ] no sep Seawat e Series2 4 5 6 7 8 9 10 -3 -2 -1 0 log[Cl - ] pH Series4 no sep Seawater Series2 (e) -5 -4 -3 -2 -1 0 -3 -2 -1 0 log[Cl - ] log [HC O 3- ] no sep Seaw ater Serie s2 -5 -4 -3 -2 -1 0 -3 -2 -1 0log[Cl - ] log [SO 42- ] no sep Seaw ater Serie s2 -5 -4 -3 -2 -1 0 -3 -2 -1 0 log[Cl - ] log [H 4SiO 4] no sep Seawater Series2 Figure 5.1: Variation of ion activities with Cl? in saturated paste extracts. Solid circles represent soils with sepiolite (some with accessory palygorskite) in the clay fraction; open circles represent no detectable sepiolite or palygorskite; ? - no clay present. Na-Cl represents the conservative slope. Grey square with ? is seawater from Nordstrom et al. (1979) which is included in the PHREEQC User Manual (Parkhurst and Appelo, 1999) as Example 1. Dashed line: seawater composition projected back to zero. 124 T abl e 5.1 : Com p ositi on of sa tura te d past e extr acts . SampleDepthHorizonClassification ?pHECAlkalinity ? Na + K + Mg 2+ Ca 2+ Cl - SO 4 2- P ?SiSAR ? mmS/cm 6869/i 0-A1Regic sand (Orthic A) CalcareousNo clay 8.950.3802.091.910.3210.5841.161.500.2200.0080.2072 6869/ii 0.45-A2Regic sand CalcareousNo clay 9.191.002.806.180.3661.041.245.410.822n.a.0.2146 6869/iii 0.5-II B1Neocarbonate B CalcareousSepiolite 8.315.832.9650.91.018.435.6047.7-n.a.0.18519 6869/iv 0.6-0.8II B2Neocarbonate B CalcareousSepiolite 8.107.191.7467.11.038.985.0759.43.71n.a.0.21225 6869/5 0.32-IV C1Hardpan carbonate Calcareous?8.524.703.5230.50.8376.097.8039.23.58 n.a. 0.56 12 6869/7 1-IV C2Marine deposit Non-calcareousNo sepiolite 8.021.090.2135.930.2491.321.358.850.2050.0180.4225 6869/8 6-9V C3Marine deposit Non-calcareousNo sepiolite 8.380.6900.1503.390.1850.6010.6365.160.162n.a.0.3824 6869/11 9-18VI C4Marine deposit Non-calcareousNo sepiolite 8.390.4960.2033.520.1150.2520.2863.280.2620.1510.4277 AK1/10 1.5-III C2Dorbank CalcareousSepiolite 7.5627.10.1802524.5448.713.634239.40.0150.62445 AK1/6 1.6-III C3Dorbank CalcareousSepiolite 7.9648.00.1503966.2998.269.671556.00.0080.22743 AK61H/4 0-AOrthic A CalcareousSepiolite8.153.199.2420.62.2812.818.220.39.11 n.a. 0.51 5 AK61H/1 0.35-II C1Hardpan carbonate CalcareousSepiolite 7.9619.70.5001673.0243.126.937030.10.0080.27528 KV196T/1 ?1-B2Red apedal B Non-calcareous?9.270.7796.625.670.8150.1140.1652.660.656 n.a. 0.71 15 KV196T/2 1.3-C1Dorbank Non-calcareous?8.007.812.251001.5314.09.5179.911.5 n.a. 1.44 29 KV196T/3 1.45-C2Dorbank Non-calcareousNo sepiolite 7.7918.90.5831741.9239.724.522327.80.0130.66831 KV196T/4 1.6-C3Dorbank Calcareous?7.9937.91.784294.4188.10.87441878.6 n.a. 0.29 64 KV196T/5 1.8-C4Hardpan carbonate Calcareous?8.227.242.2451.02.2414.860.044.778.2 n.a. 0.75 8 KV196T/7 ? - 3.5C6Soft dorbank Non-calcareousNo sepiolite8.228.963.051231.858.456.5689.610.7 n.a. 1.30 45 KV196T/8 ? 3.5-C7Dorbank Non-calcareousNo sepiolite 8.0310.90.16096.01.3914.325.456.091.00.0070.42522 KV196T/9 ?4.5-C8Dorbank Non-calcareousNo sepiolite7.7337.61.645886.0884.252.747868.3 n.a. 1.48 71 LKC1-5/A 0-ARegic sand (Orthic A) Non-calcareousNo clay 5.991.510.7607.261.133.204.767.823.760.1140.3264 LKC1-5/B 0.3-BRegic sand Non-calcareous? 8.211.340.3756.090.9312.502.5010.02.520.0180.1854 LKC1-5/1 1-II B2Neocutanic B Non-calcareousNo sepiolite 7.1112.70.2681191.5115.45.6916016.70.0100.27937 LKC1-5/3 1.3-II B3Neocutanic, luvic Non-calcareousNo sepiolite 7.9424.40.3752192.5944.716.135136.80.0090.26540 LKC1-5/9 5.4-~12VII C6Marine deposit Non-calcareousNo sepiolite 7.7611.90.0601191.2212.13.3812913.20.0080.33543 OBT/5 1-II C2Sedimentary? Calcareous?8.186.323.7180.21.9314.614.659.716.7 n.a. 0.81 21 SK11/A 0-AOrthic A Non-calcareousSepiolite8.491.437.2616.30.9392.945.778.171.93 n.a. 0.55 8 SK11/B 0.2-BYellow-brown apedal B Non-calcareousSepiolite8.560.6034.626.340.6831.202.212.570.39 n.a. 0.35 5 SK11/C 0.6-EGleyed sand Non-calcareousSepiolite8.200.7213.718.811.031.051.614.370.50 n.a. 0.39 8 SK11/D 0.65-B2Neocutanic Non-calcareousSepiolite7.909.424.1377.62.336.203.7087.64.28 n.a. 0.60 35 SK11/6 0.45-II C2Blocky calcretized sediment CalcareousSepiolite8.3020.26.401713.3016.09.8816130.5 n.a. 0.52 48 SK11/9 0.9-1.1III C3Hardpan carbonate CalcareousSepiolite 8.0318.60.2301772.7329.920.729028.90.5031.26035 SK11/5 0.8-1.5Termite nest CalcareousSepiolite8.1910.92.631282.5820.90.14693.224.3 n.a. 0.32 39 SL4-1/1 0-0.3A1Orthic A Non-calcareousNo sepiolite 7.470.4750.4401.340.2941.311.761.870.3130.0110.3461 SL4-1/2 0.3-B1Yellow-brown apedal B Non-calcareousNo sepiolite 5.031.670.0589.190.9493.361.0914.61.380.0230.5276 SL4-1/3 0.5-EGleyed sand Non-calcareousNo sepiolite 5.231.590.05710.30.7491.600.61513.70.9230.0090.36810 SL4-1/4 0.63-B2Neocutanic Non-calcareousNo sepiolite 4.607.540.02372.01.576.671.8177.37.16n.a.0.98035 SL4-1/5 0.8-II B3Gleyed sand? Non-calcareousNo sepiolite 6.5736.10.1433654.1111133.553842.20.0080.92243 SL4-1/6 0.95-III B4Pedocutanic Non-calcareousNo sepiolite 7.8025.10.4302533.3841.015.531434.50.0110.43648 SNT60/5~1.5-2.3 IV C1 Hardpan carbonate CalcareousSepiolite8.762.813.6335.71.664.183.7822.34.16 na 0.50 18 SNT60/6 Local channel? Calcareous-8.383.353.1823.61.194.042.3724.82.66 na 0.38 13 SNT60/7 ~2.3-2.8V C5Blocky calc. sediment CalcareousSepiolite 8.403.630.34528.71.314.512.6638.12.550.0080.33215 SNT60/10 2.8-VI C6Sedimentary? Non-calcareousNo sepiolite 8.461.520.56312.70.3270.3940.4129.831.580.0110.45320 SNT60/11 3.4-VII C7Sedimentary? Non-calcareousNo sepiolite 8.511.150.3079.060.2400.1790.1747.720.7700.0080.41722 SNT60/12 3.7-VIII C8Sedimentary? Non-calcareousNo sepiolite 8.942.060.54017.40.4130.5130.48012.50.6600.0080.31925 TP231/4 2-III C2 Marine deposit CalcareousSepiolite8.118.081.9595.61.7716.719.579.119.1 n.a. 0.79 22 TP266Q/1 1.2-1.7III C1Hardpan carbonate CalcareousSepiolite 7.5825.20.1801672.9453.871.742430.50.0080.39221 mmolc/Lmmol/L ? Clas si fie d ac cor din g to th e Soi l Clas si ficatio n W ork in g Gro up (1991) . Roma n nu m era ls in dicat e di ff ere nt se dim en ta ry un its . ? Al kal ini ty : C O 3 2 ? +HC O 3 ? , si nc e C O 3 2 ? w as on ly de te ct ed in L K C1 -5 /B an d SNT60/1 0 an d ? 0.0 1 m m ol c /L . ? N.a . - no t an alys ed . Si nc e P w as ver y lo w an d tog et he r wit h F mad e no diff er en ce to th e sat uratio n in de x ca lc ulatio ns it w as no t m ea su re d in al l sam pl es . ? SAR = so di um adsor ptio n ratio : N a/ ? (C a + Mg )/ 2. 125 Soluble Al was not measured since it was assumed to be low at the pH range encountered, and the Al-bearing phases that were detected with XRD could all be adequately explained by mineral transformation or dissolution re-precipitation (see Section 5.1). The phosphate and fluoride were very minor made no difference to the saturation index calculations. The system was therefore simplified to the components Na+, K+, Ca2+, Mg2+, Si, pH, Cl?, SO2?4 , and HCO ? 3 . Most of the soils for which there was a positive sepiolite identification (1.2 nm XRD peak, Chapters 1, 3, 4) showed a positive sepiolite saturation index (Figure 5.2), suggesting that the saturated paste extracts were in equilibrium with sepi- olite. This is in direct contrast to the findings of Singer et al. (1995), which were that sepiolitic and palygorskitic soils in the same region were undersaturated with respect to both these phases when using a 1:1 soil-water solution. Their interpretation, therefore, that these minerals were pedogenically formed in the past and are currently in a state of alteration towards mixed layer or smectite clay due to a reduction in alkalinity and/or Mg supply induced by climatic shifts during the Late Pleistocene and Holocene, therefore requires revision. Follow- ing Kittrick?s (1971, p. 453) suggestion that ?where the soil solution has had appreciable contact with precipitating minerals, it is likely to be in equilibrium with them?, the data from the current study showing sepiolite equilibrium in sepiolitic soils suggest that sepiolite can be considered a ?precipitating mineral? and thus forming in the present day. Furthermore, in Chapter 4 there was no TEM evidence to favour a hypothesis for fibrous mineral degradation to sheet silicates. 5.3.1 Chloride as a proxy for evaporative concentration Chloride is a reliable measure of the degree of evaporative concentration up to halite saturation because chloride is usually conserved over the widest concen- tration range since anion exchange is minor and Cl? remains in solution until halite reaches saturation (Eugster and Jones, 1979). Eugster and Jones (1979) explained that this does not mean the solute is not involved in the formation of efflorescent crusts (which were observed in some profiles in this study, see Chapter 2), but that it will be redissolved quantitatively by subsequent runoff. Halite remains undersaturated at all concentrations in the saturated paste ex- tracts (Figure 5.2), confirming the suitability of Cl? being used as a proxy for evaporative concentration in this system. 126 -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Halite SI Se n o s e Se -20 -15 -10 -5 0 5 10 -3 -2 -1 0log[Cl - ] Gypsum SI Series 4 no sep Series 1 -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Quartz SI S n o s e S -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Amorphous silica SI S n s e S -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Calcite SI no sep Series1 -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Dolomite SI S n o s e S -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Sepiolite SI Se no se p Se -20 -15 -10 -5 0 5 10 -3 -2 -1 0 log[Cl - ] Kerolite SI Ser no se p Ser Figure 5.2: Variation of mineral saturation indices (SI) with Cl? activity in sat- urated paste extracts. SI > 0 indicates supersaturation, SI<0 undersaturation, and SI = 0 equilibrium with respect to a phase. Symbols as for Figure 5.1. 127 The variation in activity of all the ions with Cl? is consistent with a seawater dilution line (Figure 5.1). The ions K+, Ca2+ and Mg2+ increase with increasing evaporation but at a slope less than that for the conservative solutes Na+ and Cl? (Figure 5.1). According to Eugster and Jones (1979), this is typical for removal mechanisms such as mineral precipitation, sorption, or degassing, which operate over the full range of evaporative concentration so that an essentially linear relationship results, but with a slope less than that for conservative solutes. Sulfate increases until log[Cl?] reaches -1 and then levels off (Figure 5.1) as gypsum reaches saturation (Figure 5.2: gypsum SI vs log[Cl?]). H4SiO4 activity remains unchanged for all levels of evaporation (Figure 5.1). This is consistent with the behaviour of an uncharged solute which has reached satu- ration with respect to the corresponding solid phase (Eugster and Jones, 1979). In the Namaqualand soil system the most likely neoformed minerals control- ling the H4SiO4 solubility are sepiolite and amorphous silica/quartz. Many of the sepiolite-containing samples show an amorphous silica component in their XRD trace (Chapter 4), and silica-cemented horizons are common (Chapter 1), ranging from silcrete to petroduric (WRB, 1998). There are many examples of etched quartz grains suggesting remobilisation of silica. Figure 5.2: sepiolite SI vs log[Cl?] shows that the sepiolite-bearing soils are saturated or supersaturated with respect to sepiolite through the entire concentration range. All the soils analysed are close to saturation with respect to both quartz and amorphous silica (Figure 5.2: quartz and amorphous silica SI vs log[Cl?]). Both quartz and amorphous silica saturation indices are independent of pH (Fig- ure 5.3) and directly dependent on silica concentration (Figure 5.4). This is because the solubility of both quartz and amorphous silica is expressed only in terms of one aqueous species, H4SiO4 (equations 5.3 and 5.4). As a result, in the amorphous silica and quartz SI vs log[H4SiO4] plots (Figure 5.4), the SI reaches zero where log[H4SiO4] is equal to the log K used to calculate the saturation index of the respective phases, in this case -2.710 for amorphous silica and - 3.980 for quartz. In the H4SiO4 vs Cl? activity plot, however, (Figure 5.1), the X- and Y-axes were calculated from independent variables. Therefore, in the Namaqualand coastal plain soils the best explanation for the unchanged H4SiO4 activity for all levels of evaporation (Figure 5.1) is that it is consistent with the behaviour of an uncharged solute which has reached saturation with respect to the corresponding solid phase (Eugster and Jones, 1979), i.e. quartz, amorphous silica and/or sepiolite. 128 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Halite SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Gypsum SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Quartz SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Amorphous silica SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Calcite SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Dolomite SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Sepiolite SI Series4 ? no sep Series1 -20 -15 -10 -5 0 5 10 4 6 8 10 pH Kerolite SI Series4 ? no sep Series1 Figure 5.3: Variation of mineral saturation indices (SI) with pH; saturated paste extracts. SI> 0 indicates supersaturation, SI<0 undersaturation, and SI= 0 equilibrium with respect the phase in question. Symbols as for Figure 5.1. 129 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[Na + ] Ha lite SI Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[Ca 2+ ] Gyp su m SI Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[SO 42-] Gyp su m SI Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[H 4SiO4] Quartz SI Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[H 4SiO4] Am orp ho us si lic a S I Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[Ca 2+ ] Ca lci te SI Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[HCO 3-] Ca lci te SI Series4 no sep Series1-10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[Ca 2+ ] Do lom ite SI Series4 no sep Series1 Figure 5.4: Variation of mineral saturation indices (SI) with activity of com- ponent ion; saturated paste extracts. SI> 0 indicates supersaturation, SI<0 undersaturation, and SI= 0 equilibrium with respect to a phase. Symbols as for Figure 5.1. 130 Alkalinity was only present as bicarbonate. The HCO?3 activity vs Cl ? activity plot (Figure 5.1) most effectively distinguishes the sepiolitic vs non-sepiolitic soils, with the sepiolitic soils generally having higher HCO?3 activities consistent with observations that only the calcareous soils are sepiolitic. The HCO?3 activity of both the sepiolitic and non-sepiolitic soils decreases slightly with increasing Cl?. The decrease in HCO?3 activity coupled with a simultaneous increase in Ca2+ is consistent with calcite precipitation at the first ?chemical divide? in the Hardie-Eugster model of brine evolution (Hardie and Eugster, 1970; Eugster and Hardie, 1978; Eugster and Jones, 1979). All the Namaqualand soils are close to saturation with respect to calcite (Figure 5.2: calcite SI vs log[Cl?]). Calcite, usually the first mineral to precipitate on evaporative concentration, represents a critical branching point in the evolutionary path of the evaporating water (Eugster and Jones, 1979). The path then taken by the solution depends on whether the calcium concentration in equivalents is greater or less than the carbonate alkalinity in equivalents, that is, whether 2mCa2+ is greater or less than (mHCO?3 + 2mCO2?3 ) where m is molal units (Drever, 1997, p. 330). This is because two conditions must be fulfilled simultaneously during the equilibrium precipitation of calcite: 1) the ion activity product [Ca2+]?[CO2?3 ] must remain constant, and 2) Ca2+ and CO2?3 are removed in solution in equal molar pro- portion (Eugster and Jones, 1979). Condition 1 specifies that the concentrations of Ca2+ and CO2?3 must vary antithetically, whereas condition 2 requires that the Ca2+/CO2?3 molar ratio will change unless it was precisely one at the outset (Eugster and Jones, 1979). The net effect is that waters in which HCO?3 is dom- inant over Ca2+ initially will become enriched in HCO?3 and depleted in Ca 2+, and vice versa (Eugster and Jones, 1979). In these coastal Namaqualand soil solutions it appears that calcite precipitation results in a solution in which the Ca2+ increases and HCO?3 decreases (Figure 5.1). According to the conditions outlined above, this is because the initial Ca2+/HCO?3 ratio is greater than one. Given the proximity of the sea (maximum sampling distance 30 km from coast) and strong onshore winds, these soils would be expected to reflect maritime ionic ratios. Ca2+/HCO?3 ratio is greater than one for seawater, according to the data from Nordstrom et al. (1979) (plotted in Figure 5.1): the molality of Ca2+ in seawater is 9.504e-03, and of HCO?3 is 1.514e-03. CO2?3 is negligible in comparison at 3.821e-052. This is consistent with trends in Figure 5.1 and suggests that initial maritime ratios control the 131 evaporative evolution of the soil solution by providing an initial Ca2+/HCO?3 ratio greater than one. 5.3.2 Marine-influenced initial ratios and subsequent evap- oration trends The increasing Mg with increasing Cl activity (evaporation) (Figure 5.1), is in contrast to that which is expected to occur, given that sepiolite is expected to precipitate with increasing evaporation, as was shown in the classic model of Garrels and Mackenzie (1967). Since both sepiolitic and non sepiolitic soils show the increasing Mg with Cl trend, it cannot be attributed to an increase in actual sepiolite content. The Ca-HCO?3 -Cl trend is also opposite to that of Garrels and Mackenzie (1967). It was established in Section 5.3.1 that the increasing Ca and decreasing HCO?3 with Cl trend is a result of evaporating a solution with a Ca2+/HCO?3 ratio greater than one, and that the initial ratio is marine-influenced. Singer et al. (1995) considered sea-spray enriched precipitation as one source of Mg for sepi- olite precipitation. The saturated paste extracts have similar Mg/Cl ratios to seawater (Figure 5.1). It can be hypothesized that the trend of increasing (rather than decreasing) Mg with Cl may also be marine-influenced. Mg-Cl trends from a variety of literature sources are tabulated in Table 5.2. Only Mg-silicates and Ca-Mg carbonates were included in the table because the focus was on the marine-influenced increasing Mg-Cl and Ca-HCO?3 -Cl trends. Where the original authors? data were not in mmolc/L, it was replotted in these units and on a log-log scale for comparison with data from the present study. Theoretical evaporation trends for two soil solutions in the present study were also included: one from a non-calcareous, non-sepiolitic horizon and the other from a calcareous, sepiolitic horizon. Both horizons are otherwise similar in terms of their environments of formation (clay-enriched sand over a barrier, see Chapter 2). Table 5.2 shows that the maritime and continental areas show distinctly op- posite trends. The marine-influenced areas, as in this study, generally show a trend of decreasing pH, increasing Ca and increasing Mg with increasing evap- oration. This can be explained by their initial solute proportions Ca>HCO?3 , 132 T abl e 5. 2: Io n- co nc en tra tio n tr end s in maritim e an d co nt ine nta l w ate r systems . Wher e trend s w er e est ablishe d by a nu m b er of sample s th e initia lr atio s w er e calcula te d as an av era ge of al lth e sa mple s. A bbrevia tions : Co nc .-conc en tratio n; const.- co ns ta nt ; de cr.- de crea se ; I-ion ic st re ngth ; incr.-i nc reas e; max.-maxi m um ; n. d.-n o da ta ; sl.-slig ht ; v.- ve ry . Mg-silicate/ Dolomite/ Calcite present ? Max. conc. Cl and/or ionic strength ScalePlotted against pHCa vs Cl trend HCO 3 - vs Cl trend Mg vs Cl trend Si vs Cl trend Initial Ca/HCO 3 - (mmol c /L) Initial Mg/Si (mmol c /L) Initial Ca/Mg (mmol c /L) Initial Mg/HCO 3 - (mmol c /L) Plotted self Marine-influenced This study sample SL4-1/2 Namaqualand coastal plainTheoretical evaporation single sample 30x conc.? None present, so none specified 411 mmol/L I = 0.52 Log-logConc. factor Decr.Incr.Incr.Incr.Incr. then const. Ca > HCO 3 - Mg > SiCa sl. > Mg Mg > HCO 3 - This study sample 6869/iv Namaqualand coastal plainTheoretical evaporation single sample 30x conc.? Calcite, sepiolite, less palygorskite 1780 mmol/L I = 2.27 Log-logConc. factor Decr.Incr.Decr.Incr.Const. (sl. incr.-sl.decr.) Ca > HCO 3 - Mg > SiCa < Mg Mg > HCO 3 - This studyNamaqualand coastal plainSaturated paste extractsCalcite, sepiolite, less palygorskite 715 mmol/L. I = 0.76 Log-logCl activityDecr.Incr.Decr.Incr.Const. Ca > HCO 3 - Mg > SiCa < Mg Mg > HCO 3 - Compton (2007)Namibian coastal plainWater samplesCalcite5143 mmol/kg I = 5.8 Log-logCl mmol c /kg Decr.Incr. (then decr. ?) Decr.Incr.n.d. Ca > HCO 3 - n.d.Ca < Mg Mg > HCO 3 - Yes Smith & Compton (2004) South Africa, Western Cape, Brak River Theoretical evaporation single sample 100x conc. Calcite, dolomite~7750 mmol/kgLog-logConc. factor Const.Incr. (then decr. ?) Decr.Incr.n.d. Ca > HCO 3 - n.d.Ca < Mg Mg > HCO 3 - No Smith & Compton (2004) Sea water (composition from Faure (1992)) Theoretical evaporation single sample 20x conc. Calcite, dolomite~7750 mmol/kgLog-logConc. factor Const.Incr. (then decr. ?) Decr.Incr.n.d. Ca > HCO 3 - n.d.Ca < Mg Mg > HCO 3 - No Smith & Compton (2004) Average coastal rain water (composition from Faure (1992)) Theoretical evaporation single sample 100000x conc. Calcite, dolomite~7750 mmol/kgLog-logConc. factor Const.Incr. (then decr. ?) Const. then incr. Incr.n.d.Ca sl. < HCO 3 - n.d.Ca < MgMg sl. > HCO 3 - No Smith & Compton (2004) South Africa, Western Cape coastal pans Water samplesLow- and high-Mg calcite; dolomite 7400 mmol/kgLog-logCl mmol c /kg Decr.Incr. (then decr. ?) Decr. then incr. Incr.n.d. Ca > HCO 3 - n.d.Ca < Mg Mg > HCO 3 - Yes Smith & Compton (2004) South Africa, Western Cape coastal pans Saturated paste extractsLow- and high-Mg calcite; dolomite 4660 mmol/kgLog-logCl mmol c /kg Decr.Incr. (then decr. ?) n.d.Incr.n.d.n.d.n.d.Ca < Mgn.d.Yes Singer et al. (2003)Dead Sea, Mediterranean- influenced Soluble salts from dust collected in Dead Sea Calcite, minor palygorskite ~160 mmol/LLinear- linear Cl mmol c /L n.d.Incr.n.d.Incr.n.d.n.d.n.d.n.d.n.d.No Jones et al. (1994)Marine-influenced Murray Basin, Australia: Piangil Pore waters, boreholesn.d.~987 mmol/LLinear- linear Cl mg/Ln.d.Incr.n.d.Incr.n.d.n.d.n.d.Ca < Mgn.d.No Macumber (1983), in Jones et al. (1994) Marine-influenced Murray Basin, Australia: Lake Tyrrell Surface, groundwatersn.d.~5077 mmol/LLinear- linear Cl mg/Ln.d.Const. ?n.d.Incr.n.d.n.d.n.d.Ca < Mgn.d.No Continental Garrels & Mackenzie (1967) Spring water, Sierra Nevada Theoretical evaporation single sample 1000x conc. Theorical: sepiolite, calcite precipitate 15 mmol/kgLog-logConc. factor Incr.Decr.Incr.Sl. incr. then decr. Incr. then const. Ca < HCO3-Mg < SiCa > Mg Mg < HCO 3 - Yes- verified plot Gac et al. (1977)Chari River, Lake ChadExperimental evaporationCalcite, Mg- montmorillonite 1.023 mmol/L I = 0.01 Log-logConc. factor Incr.Decr.Incr.Incr then decr. Incr. Ca < HCO 3 - Mg < SiCa > Mg Mg < HCO 3 - No Thellier et al. (1988)Chari River, Lake ChadTheoretical evaporation single sample Palygorskite~0.7 mmol/kgLog-logConc. factor Incr.Incr then decr Incr.Decr.Incr. Ca < HCO 3 - Mg < SiCa > Mg Mg < HCO 3 - No Von Damm & Edmond (1984) Closed basin lakes, Ethiopian Rift Valley Water SamplesCalcite, smectite92 mmol/LLog-logCl mmol c /kg Incr.Decr.Incr.Decr.Const. Ca < HCO 3 - Mg < SiCa > Mg Mg < HCO 3 - Yes Stoessel & Hay (1978) Surface waters Amboseli Basin Water samplesSepiolite, dolomite, calcite 27 mmol/LLog-logCl mmol c /kg Incr.Decr.Incr.Const.Const. Ca < HCO 3 - Mg ~ SiCa sl. < than Mg Mg < HCO 3 - Yes Banks et al. (2004)Groundwater southern central Siberia Water samplesCalcite1589 mmol/LLog-logCl mmol c /kg Const.Const.; 1 v. high Const.Incr.Decr. Ca ~ HCO 3 - Mg > SiCa sl. < than Mg Mg sl. > HCO 3 - Yes Banks et al. (2004)Rivers, lakes southern central Siberia Water samplesCalcite216 mmol/LLog-logCl mmol c /kg Incr.Decr.Incr.Incr.Decr. Ca < HCO 3 - Mg > SiCa < MgMg sl. > HCO 3 - Yes Deocampo (2007)Lake Eyasi and Ngorongoro Crater, Tanzania Water samplesCalcite; Mg-clays221 mmol/LLog-logCl mmol c /kg Incr.Decr.Incr.Sl. incr. then decr. Incr. then decr. Ca < HCO 3 - Mg > SiCa < Mg Mg < HCO 3 - Yes- verified ? In addition to other phases that may be present, such as gypsum or other clay minerals. ? Settings: pCO2 -3.5 because equilibrated in atmospheric conditions. Dolomite, calcite, sepiolite, kerolite, amorphous silica to precipitate, removed from solution at saturation in sample 6869/iv, and only gypsum and amorphous silica from SL4-1/2. ? Authors attributed this to Ca removal during gypsum precipitation. 133 Mg>HCO?3 , Mg>Si and Mg>Ca, consistent with the ?chemical divides? of the Hardie-Eugster model of brine evolution (Hardie and Eugster, 1970; Eugster and Hardie, 1978; Eugster and Jones, 1979). In contrast, continental areas such as that studied by Garrels and Mackenzie (1967) show decreasing Ca, decreasing Mg and increasing HCO?3 with evaporation, due to their initial solute proportions CaMg (Table 5.2). The trends do not depend on such factors as the maximum concentration or ionic strength, scale, using Cl as the concentration factor, modelling the evaporation from a single sample or plotting all the samples against Cl for a system. According to the Hardie-Eugster scheme presented in Drever (1997, p. 331), sepiolite precipitation affects the Mg/HCO?3 ratio. Where Mg was present in excess initially, it is the HCO?3 that is depleted (as Table 5.2 shows to have occurred in maritime areas). Where the HCO?3 is present in excess initially, it is the Mg that is depleted - as in the model by Garrels and Mackenzie (1967) and other continental areas (Table 5.2). This maritime-continental model is simplistic in that factors other than the marine influence could affect initial Mg ratios, such as the parent soil or rock type, which have not been accounted for in the compilation of the data in Table 5.2. The data collected in Table 5.2 serves to show, however, that the changes in ionic ratios in the saturated paste extracts are consistent with other marine-influenced surface waters, and differ antithetically to surface waters where the maritime influence is absent. The data by Banks et al. (2004) is an exception to the maritime-continental evolution trends, but they are from a cold arid environment, in contrast to the other continental locations in Table 5.2. Additionally, they are from a diversity of hydrological, geological and topographical environments which were sparsely sampled, and the immature highland groundwaters are dominated by Ca, Mg and HCO?3 ions resulting from carbonate and silicate weathering reactions (Banks et al., 2004). The maritime influence on the Mg-Ca-HCO?3 ratio in coastal areas is also indi- cated by the different composition of fog water collected at the coast compared to fog water collected further inland. The west coast fog, which can be observed to extend inland to the rise of the escarpment, occurs frequently along the coastal plain at elevations below the 200 m contour line (Olivier, 2002). Data from Eckardt and Schemenauer (1998) showed that fog collected at Gobabeb (Desert Ecological Research Unit) in Namibia, 60 km from the coast, had an average com- position showing Ca>HCO?3 >Mg, after rinsing to remove previously built-up 134 aerosols. These aerosols had a much higher Mg content than the fog. In contrast, fog collected in the same study from ?the gravel plain near the coastal town of Swakopmund? had a composition showing Mg>Ca>HCO?3 , a considerably more maritime signature according to the data in Table 5.2. This suggests that fog deposition nearer the coast is accompanied by a greater proportion of marine aerosols than fog which has travelled further inland. Further evidence of the elevated Mg from marine aerosols in coastal fog is found in the work of Olivier (2002), who found Mg>Ca>HCO?3 in a windblown sea-salt contaminated fog collector at Cape Columbine on the West Coast of South Africa. After washing down the collector screens prior to collecting the second sample, a more accurate reflection of the fog water composition was obtained: HCO?3 >Ca>Mg. 5.3.3 pH control of sepiolite equilibrium Although sepiolitic soils are more abundant on the higher side of the Mg-Cl trend (Figure 5.1), the degree of evaporative concentration shown by the saturated paste extracts does not distinguish the sepiolitic from non-sepiolitic soils (Figure 5.2). In contrast to halite and gypsum, which clearly approach saturation as the Cl? activity increases, the phases sepiolite, calcite, dolomite and kerolite are consistently close to saturation for all levels of Cl? activity (Figure 5.2). On first consideration, this appears to be a problem related to the kinetics of precipitating a complicated Mg-silicate such as sepiolite, compared to a relatively simple salt like gypsum or halite. However, unlike evaporating water bodies that may be supersaturated with respect to a phase for which the kinetics of precipitation may be slower than the rate of evaporation, the solutions used in the present study are from equilibrated saturated paste extracts. The likelihood of equilibration with respect to sepiolite is indicated by the fact that the samples in which sepiolite was detected in the clay fraction have extracts that are close to saturation with respect to sepiolite. In the equilibrated saturated paste extracts it would therefore be expected that higher Mg levels would be correlated with a higher soil sepiolite content. Of all the species plotted against the chloride activity in Figure 5.1, it is the HCO?3 activity that most obviously distinguishes the sepiolitic- from non-sepiolitic soils. This is reflected in the saturation index trends with pH: rather than increas- 135 ing with increasing chloride activity, the saturation indices of sepiolite, calcite, dolomite and kerolite are controlled by increasing pH (Figure 5.3). In contrast, gypsum and halite, the two minerals which do show increasing saturation with increasing chloride activity, display no change in saturation with pH. Furthermore, halite and gypsum saturation are equally controlled by the concen- tration of each of their component ions in solution, but this is not the case for calcite, dolomite, sepiolite and kerolite (Figures 5.4 and 5.5). Sepiolite and kero- lite saturation are independent of Mg2+ and H4SiO4 activities and depend only on H+ activity at the magnesium and silica levels measured on the Namaqualand coastal plain. This is consistent with the work of Deocampo (2005) concerning the stabilities of sepiolite and kerolite during evaporation in waters of Tanzania. Deocampo (2005) noted that ?...no simple evaporatively-controlled trend is ap- parent. In fact, pH variability is responsible for most of the observed change in saturation state...?. As with sepiolite and kerolite, saturation with respect to calcite and dolomite in Namaqualand soils is dependent neither on Ca2+ nor the Mg2+ activity, but only the HCO?3 activity over all concentrations of these elements encountered in the soils studied. The pH-dependence of the sepiolite and kerolite saturation indices is to be ex- pected, since the H+ term is present in their solubility equations (equations 5.7 and 5.8, Section 5.2). The H+ term is absent in halite and gypsum solubility equations (equations 5.1 and 5.2). Although the solubility equations of calcite and dolomite do not contain an H+ term (equations 5.5 and 5.6), the pH de- pendence of calcite and dolomite is due to the PHREEQC speciation equation CO2?3 + H + = HCO?3 , log K = 10.329, which uses the pH and HCO ? 3 concentra- tions input by the user to calculate the CO2?3 concentration which is then used in the solubility equations of calcite and dolomite (equations 5.5 and 5.6). The implication for sepiolite genesis on the Namaqualand coastal plain is that sepiolite precipitation is more likely to be triggered when a solution encounters a pH barrier than by the concentration of ions through evaporation, since even the lowest Mg-Si levels are almost sufficient for sepiolite saturation (Figures 5.1, 5.2). This is similar in principle to the ?geochemical barriers? discussed by Sauer and Stahr (2004), and implies that lateral subsurface flow with no evaporative concentration along its route could result in sepiolite precipitation. In Namaqualand, there is evidence of lateral subsurface water movement in the 136 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[HCO 3-] Do lom ite SI Series4 no sep Series1 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 log[Mg 2+ ] Do lom ite SI Series4 no sep Series1 -15 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[Mg 2+ ] Se pio lite SI Series4 no sep Series1 -15 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[H 4SiO4] Se pio lite SI Series4 no sep Series1 -20 -15 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[Mg 2+ ] Kero lite SI Series4 no sep Series1 -20 -15 -10 -5 0 5 10 -6 -5 -4 -3 -2 -1 0 log[H 4SiO4] Kero lite SI Series4 sep no Series1 Figure 5.5: Continued from Figure 5.4 - Variation of mineral saturation indices (SI) with activity of component ion; saturated paste extracts. SI> 0 indicates supersaturation, SI<0 undersaturation, and SI= 0 equilibrium with respect to a phase. Symbols as for Figure 5.1. form of extensive, prominent E horizons (described in Francis et al. (2007) and Chapter 2). Evidence of pH control of sepiolite is that sepiolite is only associated with calcareous horizons (Table 5.1 and Chapter 4) suggesting that the carbonate buffers the pH at a sufficiently high level for sepiolite to precipitate. An additional mechanism of pH control on the Namaqualand coastal plain may be changing pCO2. Loss of CO2 has been cited as a factor in calcite precipitation (Netterberg, 1969; Watts, 1980; Goudie, 1983; Wright and Tucker, 1991; Rodas et al., 1994, for example). PHREEQC was used to calculate mineral saturation indices with the pCO2 set at values ranging from -1 to -5. This was achieved by setting the log partial pressure 137 for CO2(g) (equivalent to the target saturation index) from -1 to -5 in successive simulations in the EQUILIBRIUM PHASES datablock of PHREEQC and using the default initial amount of 10 moles CO2(g). None of the mineral phases were allowed to precipitate when they reached equilibrium so that the variation in saturation indices was only due to changing pCO2 and not changing ion ratios due to removal from solution during mineral precipitation. The results in Figure 5.6 confirm the pCO2/pH dependence of calcite (and dolomite), and show that both sepiolite and kerolite saturation are even more dependent on the pH/pCO2 than calcite. Figure 5.6 shows that increasing pCO2 causes a decreasing pH in the soil solution, with a resulting undersaturation in the pH-dependent phases kerolite, sepiolite, dolomite and calcite. Neither gypsum nor amorphous silica are affected. This is consistent with Jones and Galan (1988), who noted (p. 664) that both sepiolite and palygorskite require a low sediment-water pCO2, and is identical to the trend of decreasing kerolite and sepiolite saturation indices with increasing pCO2 found by Deocampo (2005) for waters in Tanzania. Palygorskite is not well confined on the diagram in Figure 5.6, as it contains Fe and Al terms that were not analysed in this study. Instead, PHREEQC was specified to calculate Al from equilibrium with K-mica, and Fe from equilib- rium with goethite from the phreeq.dat database distributed with the program (Parkhurst and Appelo, 1999). The palygorskite dissolution reaction and log K were obtained from Singer and Norrish (1974). The undersaturated position of this mineral relative to sepiolite, however, is consistent with the XRD finding that palygorskite is a much less prominent constituent than sepiolite. The data in Figure 5.6 suggest that if the soil solution became equilibrated at a higher than atmospheric pCO2 (as is typical in the soil environment) and was subsequently exposed to a lower pCO2, equilibrium with sepiolite would be approached more rapidly than with calcite. One of the mechanisms for kerolite precipitation in caves in Hawaii proposed by Le?veille? et al. (2000) involves CO2- degassing. 5.3.4 Coastal sepiolite and inland palygorskite in Namaqualand Singer et al. (1995) studied the mineralogy of Namaqualand soils from the Land Type Survey (Land Type Survey Staff, 1987). They found that palygorskite 138 Page 1 of 1 TP266Q/1 A 60hrs -8 -6 -4 -2 0 2 4 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 pCO2 Mi ne ral Sa tur ati on Ind ice s -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 pH Dolomite Sepiolite Calcite Kerolite Am. Silica Gypsum Palygorskite pH SI = 0 nosolidsCO2_sep Figure 5.6: Variations in saturation indices with pCO2 in the saturated paste extract of calcrete TP266Q/1. Palygorskite is poorly constrained. Equilibrated at atmospheric pCO2 (-3.5), PHREEQC calculated a pCO2 of -3.7 from the alkalinity input data. dominated the soils inland of the escarpment whereas sepiolite dominated the clay fraction of soils from the lower-lying, coastal areas. Singer et al. (1995) sug- gested a dissolution-reprecipitation mechanism for the formation of palygorskite, and an additional marine Mg source for sepiolite formation. This is consistent with Chapters 1, 3, and 4, which reported that sepiolite is more prominent than palygorskite in the XRD traces from Namaqualand coastal plain soils. The marine-influenced elevated Mg is a likely explanation for sepiolite-dominance at the coast, and palygorskite-dominance inland. Sepiolite is favoured to form in an environment with a higher (Mg+Si)/Al ratio than palygorskite (Jones and Galan, 1988). The Mg concentration in the Texas High Plains appears to 139 have determined whether sepiolite-dolomite or palygorskite-calcite were formed (McDaniel et al., 1992). In Section 5.3.2 it was established that the coastal areas had elevated Mg com- positions relative to continental areas, which resulted in an increasing Mg trend with evaporation during the precipitation of sepiolite according to the Hardie- Eugster scheme since Mg>HCO?3 initially. Figures 5.1, 5.2 and 5.5 suggest that even nearly the lowest Mg2+ and H4SiO4 activity is sufficient for sepiolite satura- tion on the Namaqualand coastal plain, given an appropriate environmental pH. The result is that after sepiolite precipitation is initiated (by an geochemical pH barrier, see section 5.3.3), Mg levels will rise causing the increasing (Mg+Si)/Al ratio to continue to favour sepiolite precipitation. This suggests that once sepio- lite has begun to precipitate, the subsequent salinity with its accompanying Mg increase makes substantial palygorskite formation unlikely to follow. The sequence of formation of accessory amounts of palygorskite relative to sepi- olite on the Namaqualand coastal plain remains speculative. Jones and Galan (1988, p. 664) noted that in lacustrine closed basins or marginal continental basins transitional to the marine, the occurrence of palygorskite appears to in- dicate brackish water and the presence of sepiolite reflects salinity increases. Zaaboub et al. (2005), however, showed that sepiolite would have precipitated directly in lacustrine, playa-lake or sebka environments under alkaline conditions, high Si and Mg and low Al activity, and arid to semiarid climate, whereas pa- lygorskite would have formed by transformation of already existing illite and/or smectite type aluminosilicates in solutions in equilibrium with isotopically heav- ier and, therefore, more evaporated solutions than the sepiolite. 5.3.5 Summary of mineral genesis on the Namaqualand coastal plain The results presented in this Chapter allow the following general summary to be made of secondary mineral formation in the study area. 140 Halite and gypsum Halite and gypsum approach saturation as the concentration of their compo- nent ions in solution increases. Halite remains undersaturated at all sodium and chloride concentrations in the saturated paste extracts. At higher chloride concentrations, gypsum reaches saturation, and begins to remove sulfate from solution. Calcite All the soils are close to saturation with respect to calcite. The HCO?3 activity decreases slightly with increasing Cl?. The decrease in HCO?3 activity coupled with a simultaneous increase in Ca2+ is consistent with calcite precipitation at the first ?chemical divide? in the Hardie-Eugster model of brine evolution, and consistent with the maritime Ca2+ >HCO?3 initial ratio. Calcite remains consistently close to saturation for all levels of Cl? activity (Fig- ure 5.2). The calcite saturation index is not dependent on the Ca2+ activity, only on the HCO?3 activity and pH for the range of concentrations encountered for these parameters. Sepiolite-palygorskite Although the sepiolitic soils are more abundant on the higher side of the Mg-Cl trend, the degree of evaporative concentration (as reflected by Cl concentration) in the saturated paste extracts does not distinguish sepiolitic from non-sepiolitic soils. Even the lowest Mg2+ and H4SiO4 activity is sufficient for sepiolite sat- uration given an appropriate environmental pH. Sepiolite remains consistently close to saturation for all levels of Cl? activity. The sepiolite saturation index is independent of Mg2+ and H4SiO4 and depends only on the H+ activity at the magnesium and silica levels measured. The HCO?3 activity vs Cl ? activity plot most effectively distinguishes the sepi- olitic vs non-sepiolitic soils, with the sepiolitic soils generally having higher HCO?3 activities. This is reflected in the saturation index trends with pH: rather than increasing with increasing chloride activity, the saturation indices of sepiolite is 141 controlled by increasing pH. Evidence of pH control on sepiolite precipitation is that sepiolite is only associated with calcareous horizons suggesting that the carbonate buffers the pH at a sufficiently high level for sepiolite to precipitate. An additional mechanism of pH control on the Namaqualand coastal plain may be changing pCO2, similar in principle to the loss of CO2 that has been cited as a factor in calcite precipitation. Sepiolite precipitation is therefore more likely to be triggered when a solution encounters a pH barrier than by the concentration of ions as a result of evaporation, since even the lowest Mg-Si levels are almost sufficient for sepiolite saturation. The marine-influenced high Mg level coupled with the Hardie-Eugster model of brine evolution provides 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. Even the lowest Mg2+ and H4SiO4 activity is sufficient for sepiolite saturation on the Namaqualand coastal plain, given an appropriate environmental pH. 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. Most of the soils for which there was a positive sepiolite identification showed a positive sepiolite saturation index, suggesting that the saturated paste extracts were in equilibrium with sepiolite. This is in direct contrast to the findings of Singer et al. (1995), which were that sepiolitic and palygorskitic soils in the same region were undersaturated with respect to both these phases when using a 1:1 soil-water solution. Their interpretation, therefore, that that these minerals were formed pedogenically in the past, but are in a state of alteration towards a mixed layer or smectite clay as a consequence of a reduction in alkalinity and/or Mg supply induced by climatic shifts during the Late Pleistocene and Holocene requires revision. The data showing sepiolite equilibrium in sepiolitic soils suggest that sepiolite can be considered a ?precipitating mineral? and thus as forming in the present day. Furthermore, in Chapter 4 there was no TEM evidence to favour a hypothesis for fibrous mineral degradation to sheet silicates. 142 Silica In the Namaqualand soil system the most likely minerals controlling the H4SiO4 solubility are sepiolite and amorphous silica/quartz. H4SiO4 activity remains unchanged for all levels of evaporation, consistent with the behaviour of an un- charged solute which has reached saturation with respect to the corresponding solid phase. pH variation and its potential to affect the genesis of calcic, sepiolitic and duric soils The data in Figure 5.6 suggest that decreasing the pH by 0.5 units decreases the calcite saturation index by 0.45 units and the sepiolite saturation index by 1.9 units, while leaving the amorphous silica saturation index unchanged. The effect of a pH change on the sepiolite saturation index is therefore > 4 times that of the effect on calcite. The effect of pH fluctuations on calcite and silica solubilities is important in the genesis of calcrete-silcrete intergrades in the Kalahari, Botswana (Nash and Shaw, 1998; Kampunzu et al., 2007). The data in Figure 5.6 show that pH affects the solubility of sepiolite even more than it does calcite. Since sepiolite is often a significant part of arid region soils where calcite and amorphous silica are common, the mineral solubility-pH relationships may assist in explaining the relationships between calcic, sepiolitic and duric soils and their intergrades. 5.4 Conclusions PHREEQC is an adequate tool with which to perform speciation calculations for Na-Cl dominated soils of the Namaqualand Coastal Plain system, since the ionic strength of the samples falls in the range of 0.004 to 0.76, with most of the samples being well below the concentration of seawater. The variation in activity of all the ions with Cl? is consistent with a seawater dilution line. The equilibrated saturated paste extracts were ranked in order of their Cl? con- tent as a proxy for degree of evaporative concentration in the field. The data trends when all the ?closed system? saturated pastes were considered together are consistent with the predictions of the Hardie-Eugster model for closed-system 143 evaporating brines, as well as with experimental and simulated evaporations of solutions with similar initial ratios as the saturated paste extracts. The Hardie- Eugster model of brine evolution in closed basin brines can therefore be applied generally to the soil system in Namaqualand. The Namaqualand coastal plain, like other maritime areas, generally shows a trend of decreasing pH, increasing Ca and increasing Mg with increasing evap- oration. This can be explained by their marine-influenced initial ratios and is consistent with the ?chemical divides? of the Hardie-Eugster model of brine evo- lution. K+, Ca2+ and Mg2+ increase with increasing evaporation but at a slope less than that for the conservative solutes Na+ and Cl?, and suggests removal mechanisms such as mineral precipitation, sorption, or degassing which operate over the full range of evaporative concentration. Considering the formation of minerals in the arid soil environment in the same terms as the geochemical evolution of brines has helped to refine the model of mineral formation on the Namaqualand coastal plain. Halite remains undersatu- rated at all concentrations in the saturated paste extracts. At higher concentra- tions, gypsum reaches saturation, and sulfate is removed from solution. H4SiO4 activity remains unchanged for all levels of evaporation or 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. Sepiolite precipitation is therefore more likely to be triggered when a solution encounters a pH barrier than by the concentration of ions by evaporation. Evidence of the pH control on sepiolite saturation is that sepiolite is commonly associated with calcareous horizons. The effect of a pH change on the sepiolite saturation index is much greater than its effect on calcite. The marine-influenced high Mg level 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 144 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. 145 Chapter 6 Micromorphology, mineralogy and genesis of soils associated with a Namaqualand ?heuweltjie? 6.1 Introduction ?Heuweltjies? (Afrikaans for small hills) occur throughout Namaqualand and along the western and southern Cape coasts (Picker et al., 2007). They are thought to be termitaria of the harvester termite Microhodotermes viator (Moore and Picker, 1991), with a debated contribution from mole rats (Lovegrove and Siegfried, 1986; Cox et al., 1987; Lovegrove and Siegfried, 1989; Laurie, 2002; Midgley and Hoffman, 1991) and other animals (Milton and Dean, 1990). Midg- ley et al. (2002) obtained ?14C ages of 25 000 to 30 000 years B.P. for calcrete associated with heuweltjies in the Clanwilliam and Elands Bay areas, similar the 32 100 ? 720 years B.P. for the nearly identical fossilised nests from Clanwilliam presented by Coaton (1981). Heuweltjies are distinguishable in the field as circular features (diameter usually 10 to 20 m) showing a different vegetation pattern and a slightly raised (1 to 2.5 m) surface, presenting what Laurie (2002) referred to as an ?ostrich leather? tex- ture on aerial photographs. When using Google Earth (http://earth.google.com/) to observe the study area, between Papendorp and Strandfontein, the heuweltjies become discernable as green circular features at an eye altitude of 2 to 5 km. 146 They represent zones of denser vegetation because of the nutrient cycling into the heuweltjie (Midgley and Musil, 1990; Midgley and Hoffman, 1991) and higher base status of on-mound compared to off-mound soils (Ellis, 2002). A good ex- ample of heuweltjie-modified soils influencing vegetation patterns is evident at the location 31? 44? 10?? S, 18? 14? 20?? E, just north of Strandfontein and a little south of the heuweltjie in the present study, where the greener vegetation pat- tern of heuweltjies in this region has persisted through subsequent ploughing and cropping. Similar examples can been seen throughout their zone of occurrence along the western and southern Cape coasts. Ellis (2002) noted that there seems to be a direct relationship between hardpan occurrence and termite activity: where the rainfall is at its lowest, termite ac- tivity is presently relict and mounds are characterized by a central petrocalcic to petroduric hardpan, and a petroduric over petrocalcic horizon towards the outer edge and in the surrounding inter-mound areas. In the higher rainfall zone (250 to 450 mm) the general tendency is that limited termite activity is present though only in the centre of most mounds but absent on certain, apparently the oldest, mounds. Where termites are still active, the areas of activity have hypocalcic to hypercalcic horizons with petroduric or petrocalcic horizons on the perimeter of the mound, even when no hardpans occur in surrounding soils. With increasing rainfall, age and relict termite activity, a broken hardpan and/or more base-rich soil material is found on the mounds. In these latter cases the soils of the inter-mound areas have moderate to low base status without any of the above-mentioned hardpans. Since the publication of the work by Ellis (2002), it was noticed that the fibrous clay minerals sepiolite and/or palygorskite are commonly associated with the cal- crete in the heuweltjies. Palygorskite tends to occur more in heuweltjies in inland areas, consistent with the sepiolite-palygorskite distribution in Namaqualand (Singer et al. (1995) and Chapter 5). Sepiolite and palygorskite are often re- ported from arid region soils and are commonly found in calcretes (Vanden Heuvel, 1964; Singer and Norrish, 1974; Yaalon and Wieder, 1976; Elprince et al., 1979; Hay and Wiggins, 1980; Watts, 1980; Singer and Galan, 1984; Singer, 1989; Blank and Fosberg, 1991; Monger and Daugherty, 1991; Verrecchia and Le Coustumer, 1996; Singer, 2002; Neaman and Singer, 2004; Owliaie et al., 2006). The hardpan horizon of the heuweltjie commonly grades from a petrocalcic in the centre through a ?sepiolitic?/?petrosepiolitic? horizon (as defined in Chapter 1), 147 to the petroduric horizon on the edges. In the heuweltjie soils, the petrocal- cic horizon is commonly ?sepiolitic? as defined by Chapter 1. Usually there are abundant sepiolite cutans (up to a few millimetres thick) around the nodules of calcrete. These cutans are generally white, non-calcareous, adhere strongly to the wetted tongue and react with methyl orange to change it from orange to pink/purple, indicating sepiolite (Mifsud et al., 1979). Personal observations showed that both pure calcite, and calcrete in which sepiolite is only detectable when the clay fraction is concentrated and examined by XRD, react negatively to the methyl orange test (Chapter 1). In contrast, most of the calcrete in the centre of the heuweltjies react positively to the methyl orange test, suggesting substantial concentrations of sepiolite. In an area like Worcester, for example, these horizons are only associated with the heuweltjies and are not present in the inter-heuweltjie areas. Heuweltjies are sometimes described as Mima-like mounds (for example Cox et al., 1987; Lovegrove and Siegfried, 1986; Milton and Dean, 1990; Lovegrove and Siegfried, 1989). Horwath and Johnson (2006), (2007) summarised the char- acteristics of Mima and Mima-like mounds. Horwath and Johnson (2006) noted that the term ?Mima mound? comes from Mima Prairie, a tract in the Puget Sound lowlands south of Olympia, Washington. Mima-like mounds elsewhere, although they may bear resemblances to the type locality, differ in height, diam- eter, texture, internal composition and structure, and in the elevations at which they occur. Numerous theories for the origins of Mima-mounds have been pro- posed (Washburn, 1988; Reider et al., 1996; Horwath and Johnson, 2006, and references therein). These can be grouped into five main genetic categories: ero- sional, depositional, fossorial (burrowing) animals (usually rodents), periglacial, and seismic origins (Horwath and Johnson, 2006). The heuweltjie soils show some similarities to Mima(-like) mounds, particularly their raised surfaces, silica and/or calcium carbonate hardpans and clay-accumulation horizons (Spackman and Munn, 1984; Hobsan and Dahlgren, 1998), as well as their organic mat- ter (nitrogen) enrichment (Midgley and Musil, 1990; Litaor et al., 1996; Hobsan and Dahlgren, 1998). They differ in that the clay mineralogy of the Mima(- like) mounds does not contain sepiolite or palygorskite, only smectite, kaolinite (Hobsan and Dahlgren, 1998) and illite (Spackman and Munn, 1984), and we have not observed in heuweltjies the plug or central pipe that was noted in Mima-like mounds by Spackman and Munn (1984) and Reider et al. (1996). 148 Most of the previous heuweltjie work has focused on biogenic aspects, such as their spacing, origin and age, but even though the heuweltjie is in fact a soil feature there have been few published studies done on the soil forming processes within the heuweltjie (Van der Merwe, 1940; Slabber, 1945; Ten Cate, 1966; Burgers, 1975; Ellis et al., 2001; Ellis, 2002). There has been some work done on the soil properties (such as Watson, 1962) and micromorphology of termite mounds (for example Stoops, 1964; Mermut et al., 1984; Lobry de Bruyn and Conacher, 1990), but these did not include heuweltjies. We therefore aimed to use soil micromorphology and the principles of brine evolution from Chapter 5 to help explain the soil formation within the heuweltjie, particularly the relationship between the petrocalcic, ?(petro)sepiolitic? (Chapter 1) and petroduric horizons, and to refine Ellis? (2002) model of hardpan formation in heuweltjies in the light of the new data. Since this is the first published account of thin section descriptions through a heuweltjie, the detailed descriptions, photographs and terminology that were used are included in full to allow for easier comparison in future studies. 6.2 Materials and methods Heuweltjies were sampled from Papendorp?Strandfontein (1 heuweltjie), Stellen- bosch (1 heuweltjie), Oudtshoorn (1 heuweltjie), and Worcester (2 heuweltjies) The locations are shown in Figure 6.1. Figure 6.1: Location of Papendorp?Strandfontein, Stellenbosch, Oudtshoorn, and Worcester, where heuweltjies were sampled for this study. Modified from Pether et al. (2000). 149 Seven samples of the calcrete horizons and the associated sepiolite/palygorskite cutans were sampled in the Oudtshoorn and Worcester heuweltjies to confirm the presence of palygorskite and/or sepiolite, after field tests with methyl orange were positive for sepiolite (methyl orange shows a purple-pink colour, Mifsud et al., 1979). Bulk samples were ground with an agate pestle and mortar with distilled water and sedimented onto glass slides for XRD analysis (machine setup as for the Papendorp heuweltjie detailed below). The main focus of this study was the Papendorp heuweltjie. It is located at 31? 42? 32?? S, 18? 13? 32?? E, elevation 60 m a.m.s.l., on the south side of the Olifant?s River mouth, on a road-cut along the R362 between the towns of Papendorp and Strandfontein (Figure 6.1), on an undulating coastal plain, sea-facing (west) in natural (grazed) veld. Mean annual rainfall for the Papendorp/Strandfontein area is < 150 mm. Samples were taken of the hardpan horizons and their over- lying loose horizons through a cross-section of the heuweltjie from the centre to the edge (Figure 6.2). Approximately 2 kg bulk sample was taken of each horizon of the Papendorp heuweltjie, as well undisturbed, vertically orientated material for the preparation of thin sections. Section dimensions are given in Section 6.3.4 (Table 6.2). Thin sections for optical microscopy from the Papendorp heuweltjie were impregnated with a polyester resin and ground without water, and left uncovered on one side to allow for etching and SEM-EDX work. Fine grained calcite was identified by effervescence in 1M HCl under the optical microscope. Calcite was distinguished from calcium oxalate by its insolubility in 2M acetic acid, since both calcium carbonate and calcium phosphate are soluble (Post, 1985). Slaking tests were based on the WRB (1998) definition of petroduric, petrocalcic and fragic hori- zons. An intact fragment a few centimetres in diameter was submerged in water, 5M HCl or 6M NaOH and gently heated on a waterbath for four days. Uncoated fragments and thin sections were observed before and after etching in 1M HCl using low vacuum SEM-EDX with a Philips Xl30 ESEM. High vacuum SEM was done on Au-coated fragments using a Leo 1430VP SEM-EDX system. For XRD analysis of the clay fraction, the bulk samples were air-dried, crushed and passed through a 2 mm sieve. The <2 ?m fraction was separated from the bulk samples by dispersion (shaking briefly by hand, raising the pH to approximately 10 with Na2CO3) and settling. The clay suspension was flocculated by addition of MgCl2 after lowering the pH to 5 to 7 with HCl to prevent the precipitation of brucite 150 and/or clay destruction. The clay suspension was then Mg-saturated, concen- trated by centrifugation, and sedimented (or smeared, many of the sepiolite-rich samples developed ?mudcracks?) onto a glass slide. XRD analyses were done with a stepsize of 0.05 degrees and steptime of 40 seconds, using a Bruker DD8 Advance Powder Diffractometer with a graphite monochromator, 40 kV and 40 mA. Ethylene glycol was sprayed lightly onto the surface of the Mg-saturated sample slides. TEM-EDAX analyses were performed on the clay fractions using a Philips CM120 Biotwin equipped with EDAX detector. Thin sections were made through exit towers and frass from the surface of an active heuweltjie in Stellenbosch, to assist in the recognition of the fossil termite features in the Papendorp heuweltjie. The excrement samples were collected loose and randomly oriented on the section (3 x sections, 20 x 40 mm diameter; thickness 30 ?m). Two 2 termite exit towers were sampled (approximately 10 Ce ntr e Ed ge 14 cm Or thi c (oc hri c) A, sa mp le 2.1 ~5 0 c m de ep Ne oc uta nic (ca mb ic) B, sa mp le 2.2 10 cm 15 cm ~1 4 m Do rba nk (pe tro du ric ), s am ple 1. 3 ?Se pio litic ?pe do cu tan ic(a rgi c), sa mp le 2D Ca lcr ete sa mp les 2A -2C 4 c m Ha rdp an / s oft ca rbo na te (pe tro /hy pe rca lcic ) Figure 6.2: Photographs of sample locations within heuweltjie. White cutans in the calcrete and sepiolitic? pedocutanic (argic) horizons react positively with methyl orange in the field, suggesting sepiolite. 151 cm high, diameter 8 to 10 mm), both of which were orientated in cross-section on the thin section (20 x 40 mm diameter; thickness 30 ?m). Thin sections were impregnated with an epoxy resin under vacuum and ground without wa- ter. The Stellenbosch heuweltjie is located on the south side of Papegaaiberg (Stellenbosch) above the graveyard in a pine plantation. Parent material is non- calcareous, yellow-brown apedal soil formed on granite. Mean annual rainfall for the Stellenbosch area is >650 mm. 6.3 Results and discussion 6.3.1 Profile descriptions The heuweltjie (pictured in Figure 6.2 with the sample locations) is typical of those in the arid western parts of South Africa described by Ellis (2002). Termite activity is presently low, increasing after rainfall. It has a petrocalcic horizon in the centre of the mound and a petroduric horizon at the edges. Between the two is a ?sepiolitic? argic horizon. The landscape is otherwise free of calcrete. Profile descriptions through a cross-section of the heuweltjie are given in Table 6.1. The calcrete horizon is ?sepiolitic? (as defined by Chapter 1), similar to heuweltjies from widely spread localities such as the Knersvlakte, Oudtshoorn and Worcester (Figure 6.3). It contains sepiolite cutans that are generally white, non-calcareous, adhere strongly to the wetted tongue and react with methyl orange to change it from orange to pink/purple (Mifsud et al., 1979). Towards the centre of the heuweltjie the calcrete is more laminar, and then grades outwards and downwards into a more nodular texture. The sepiolite cutans are much more abundant on the nodular peds. Between the petrocalcic at the centre and the dorbank (petroduric) at the edge of the heuweltjie is a hard but friable horizon that is composed of brown non-calcareous nodular, clay-rich peds that are covered by white sepiolite cutans and sepiolite coating cracks in the peds (Figure 6.2). Based on these cutans I classified this pedocutanic (argic) horizon (sample 2D) as ?sepiolitic? (Chapter 1). The cutans are not present in the dorbank at the edge of the heuweltjie (Figure 6.2). 152 T abl e 6.1 : P rofil e descripti on s fro m P ap endo rp he uw eltjie , pictu re d in Figur e 6. 2 (Sectio n 6. 2) . Blan k fie ld s: no t de te ct ed , ex cep t sl aki ng te st s no t tes te d. Ab breviat ions : co nt .- co nti nuo us ; m od .- m od er at e; qt z -qua rt z; s -s econ ds ; sl.-slig ht ; v. -v er y. SampleDepthDiagnosticpH sat-Est.CoarseStructureEfferv-MethylSlake (0=no, 5=full)Cement-WaterRootsTrans-Other NameHorizon ?urated paste DryMoistField Texture Frag.DryMoistWetescen- ce HCl OrangeHClNaOHWaterationAbsor- ption itionFeatures Profile "2": Centre of heuweltjie: Gamoep Form (Orthic A - Neocutanic B - Hardpan carbonate). 2.1 0-0.2 mOrthic (ochric) A; bleached; crust 7.6Sl. moist 5YR 4/4 5YR 4/6 SandFew sepiolite; calcrete; shale Apedal, single- grain LooseLooseNon- plastic, non-sticky NoneNo reaction None3sCom- mon Abrupt, smooth Tunnel ~3 cm diameter, few macropores 2.2 0.2- 0.55 m Neocutanic (cambic) B 6.9Sl. moist 5YR 5/4 5YR 4/6 SandFew sepiolite; qtz Apedal, single- grain LooseLooseNon- plastic, non-sticky NoneNo reaction None1sFewAbrupt, wavy to tonguing Few macropores 2A 0.55 m -Soft carbonate ((hyper-)calcic) 8.3Dry5YR 8/2 5YR 7/3 NodularHardFriableSl. plastic, sl. sticky StrongPurple- pink ? 55-Cont.sl. calcite 3sFewNot reached White sepiolitic cutans C horizon grades laterally in sequence 2A-2B-2C-2D towards edge of heuweltjie 2B (2 m from 2A) 1 mHardpan carbonate (petrocalcic) Dry5YR 8/2 7.5YR 7/4 MassiveHardHard- Friable Sl. plastic, sl. sticky StrongPurple- pink ? 42-Cont. sl. calcite >20sFewNot reached Tunnels ~3 mm wide, 3 cm long; white sepiolitic cutans 2C (3.2 m from 2A) 1 mSoft carbonate ((hyper-)calcic) Dry5YR 8/2 5YR 8/3 NodularHardFriableSl. plastic, sl. sticky StrongPurple- pink ? 454Cont. sl. calcite 3sFewNot reached Few cracks; white sepiolitic cutans 2D (7.5 m from 2A) 1 m'Sepiolitic' pedocutanic (argic) Dry7.5YR 4/2 7.5YR 4/4 Medium blocky HardFriableSl. plastic, sl. sticky NonePurple- pink ? 2-344Cont. sepiolite? Sl. silica? 5s ?FewNot reached Tunnels ~3 mm wide, 3cm long; abund. cracks; white sepiolitic cutans Profile "1": Edge of heuweltjie ~8 m from 2D: Oudshoorn Form (Orthic A - Neocutanic B - Dorbank) 1.1 0- 0.04 m Orthic (ochric) A; bleached; crust 8.7Sl. moist 5YR 4.5/6 5YR 3/4 SandMod. qtz gravel, shale Apedal, single- grain LooseLooseNon- plastic, non-sticky V. sl.No reaction None3sCom- mon Abrupt, smooth Few macropores 1.2 0.04- 0.15 m Neocutanic (cambic) B 8.5Sl. moist 5YR 4.5/6 5YR 4/6 SandV. few qtz Apedal, single- grain LooseLooseNon- plastic, v. sl. sticky NoneNo reaction None1sFewAbrupt, smooth Few macropores 1.3 0.15 m -Dorbank ( petroduric) 9.1Dry7.5YR 5/6 7.5YR 3.5/4 MassiveVery hard FirmNoneNo reaction 03-4-Cont. mod. silica 5sNoneNot reached Few fine cracks ? Classified according to the Soil Classification Working Group (1991), in brackets is the WRB (1988) equivalent. ? Changes to purple in the presence of sepiolite or Mg-smectite (Mifsud et al, 1979). ? Water from dropper spreads instantly away along very fine cracks (dendritic pattern). Munsell Consistency 153 worcester.xls ? ? Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48n m Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a 1 .0n m Pal ygo rsk ite 0.6 44n m Pal ygo rsk ite 0.5 42n m Q u artz 0.4 26n m Pal ygo rsk ite 0.4 18n m Op al C T/A 0.4 1nm Pal ygo rsk ite 0.3 5nm Pal ygo rsk ite 0.3 23n m Sep iolit e; p aly. 0.4 49n m Q u artz ; se pio lite 0.3 34n m Ca lcite 0.3 04n m Sep iolit e ( d iff.) 0.5 04n m Sep iolit e 0 .37 4nm Worcester Worcester Op al A 0.4 03n m Worcester Worcester 1.4 2 n m Worcester Oudtshoorn Oudtshoorn 0 5 10 15 20 25 30Cu K? ? 2 ? 1.28 nm 1.3 nm 1.29 nm 1.29 nm 3.84 nm 3.84 nm 3.84 nm 3.84 nm 3.84 nm 3.84 nm Figure 6.3: XRD traces of white palygorskite/sepiolite cutans and associated calcrete from the centre of heuweltjies from Worcester and Oudtshoorn. Bulk samples ground with agate pestle and mortar and sedimented onto glass slides. Diff.: diffuse; paly.: palygorskite. Quartz and calcite peaks cropped. 6.3.2 Clay mineralogy Calcrete in centre of heuweltjie: samples 2A, 2B, 2C XRD analyses of the clay faction (Figure 6.4), shows a strong 1.28 or 1.33 nm peak for the calcrete (samples 2A to 2C) in the centre of the heuweltjie. Hay and Wiggins (1980) found a 1.26 nm peak for sepiolite in some calcretes of the 154 southwestern United States. SEM images of a calcrete fragment from sample 2A showed a fibrous clay mineral resembling sepiolite or palygorskite (Figure 6.13(f)). Sepiolite is associated with ?non-heuweltjie? calcrete from Namaqualand (Chapters 3, 4) and other southern African calcretes (Netterberg, 1969; Watts, 1980). Palygorskite was not evident in the heuweltjie calcrete, consistent with previous findings for the Namaqualand coastal region (Singer et al. (1995) and Chapters 1, 3, 4). With the exception of mica, the peaks of the overlying neocu- tanic (cambic) B horizon are not as well defined. ?Sepiolitic? pedocutanic (argic) 7.5 m from centre of heuweltjie: sample 2D The clay mineralogy of this sample 2D is less clear: it adhered to the wetted tongue, but the darker colour (Table 6.1) meant that the methyl orange test was only clear on the white, non-effervescent cutans (Figure 6.2), where it turned purple suggesting sepiolite. It did not show as clear a sepiolite peak as the calcrete samples 2A to 2C in the centre of the heuweltjie, and the ?sepiocrete? described in Chapter 1. Figure 6.4 shows the major clay peak of 2D to be broader than in the calcrete samples 2A to 2C, and centred around 1.4 to 1.5 nm. The 1.28 nm peak became more apparent on glycolation. Hay and Wiggins (1980) also found a broad 1.26 to 1.4 nm peak for clay minerals in sepiolite-rich calcretes of the southwestern United States. TEM images (Figure 6.5) of the clay fraction from sample 2D showed acicular minerals to be present, confirming the presence of sepiolite and possibly palygorskite. Acicular minerals were not the dominant morphology over the entire prepared TEM grid, however. Dorbank (petroduric) at edge of heuweltjie The rise in the background of the dorbank trace at around 0.4 nm (6.4) is con- sistent with the silica cement suggested by full slaking only in NaOH (Table 6.1). The dorbank horizon (sample 1.3) and the overlying neocutanic (cambic) B horizon (sample 1.2) show poorly defined clay peaks. These broad clay peaks were also observed in some dorbank horizons from other parts of Namaqualand (Chapter 1), and could be a function of disordered siliceous coatings masking the clay minerals in dorbank horizons (Flach et al., 1969; Blank and Fosberg, 1991). 155 Sep iolit e 1 .21 nm Pal ygo rsk ite 1.0 48n m Sm ect ite 1.5 nm Sm ect ite EG 1.7 nm Mic a 1 .0n m Pal ygo rsk ite 0.6 44n m Pal ygo rsk ite 0.5 42n m Q u artz 0.4 26n m Pal ygo rsk ite 0.4 18n m Op al C T/A 0.4 1nm Pal ygo rsk ite 0.3 5nm Pal ygo rsk ite 0.3 23n m Sep iolit e; p aly. 0.4 49n m Q u artz ; se pio lite 0.3 34n m Ca lcite 0.3 04n m Sample 2D, 1m deep, 'Sepiolitic' pedocutanic (argic), 7.5 m from heuweltjie centre. Sep iolit e ( d iff.) 0.5 04n m Sep iolit e 0 .37 4nm Op al A 0.4 03n m Sample 2B, 1m deep, Hardpan carbonate (petrocalcic), 2 m from heuweltjie centre. Sample 2C, 1m deep, Soft carbonate ((hyper-)calcic), 3.2 m from heuweltjie centre. Sample 1.2, 0.04-0.15 m deep, Neocutanic (cambic) B, Heuweltjie edge; 8 m from sample 2D. Sample 1.3, 0.15 m - , Dorbank (petroduric), Heuweltjie edge; 8 m from sample 2D. 1.4 2 n m -5 0 5 10 15 20 25 30Cu K? ? 2 ? EG EG EG EG EG EG EG Sample 2.2, 0.2-0.55 m deep, Neocutanic (cambic) B, Heuweltjie centre. Sample 2A, 1m deep, Soft carbonate ((hyper-)calcic), Heuweltjie centre. 1.28 nm 1.33 nm 1.28 nm 1.28 nm Figure 6.4: XRD traces of Mg-saturated, <2 ?m fraction from milled samples from Papendorp heuweltjie. Diff.: diffuse; EG: ethlyene glycol; paly.: paly- gorskite. 156 (a) (b) Figure 6.5: TEM image of the same clay fractions from the XRD traces in Figure 6.4. (a) Fibrous clay in < 0.08 ?m fraction from ?sepiolitic? pedocutanic (argic) sample 2D. Scale bar 500 nm. (b) < 2 mm fraction from dorbank (petroduric) sample 1.3, fibrous minerals possibly developing from mica (arrowed). Scale bar 1 ?m. EDAX analysis showed the large platy crystals to have a biotite composition. In one instance (Figure 6.5(b)) the morphology suggests the development of fi- brous clay minerals from a mica-like structure, similar to the palygorskite fibres observed developing from an illite by Suarez et al. (1994). 6.3.3 Micromorphology of excrement and exit towers: Stel- lenbosch heuweltjie In addition to examining thin sections through the heuweltjie near Papendorp, material was sampled from the surface of an active, non-calcareous heuweltjie in Stellenbosch, to assist in the recognition of the fossil termite features in the Papendorp heuweltjie. Thin sections of loose, fresh termite excrement and a cross-section through a termite exit tower (approximately 10 cm high and 1 cm diameter) are pictured in Figures 6.6 to 6.7. The excrement samples were collected loose and randomly oriented on the sec- tion. They are non-calcareous ellipsoids (0.8 x 0.4 mm) or spherical (0.3 to 0.4 mm diameter). Each has a massive microstructure, a c/f5?m ratio of 1:20, which is considerably finer than any of the samples from the heuweltjie near Papendorp (Table 6.2), and an open porphyric c/f-related distribution. The 157 (a) (b) (c) (d) (e) (f) Figure 6.6: Photomicrographs of termite excrement composed of very fine quartz and organic matter from the surface of an active, non-calcareous heuweltjie from Papegaaiberg (Stellenbosch). (a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW- SE: termite excrement (collected loose, random orientation) set in resin. (d)? ppl, (e)?xpl, (f)??-plate, fast-direction NW-SE: magnification of (a)?(c) showing spherulite with pseudo-negative interference figure in termite excrement. Note the perpendicular arrangement of acicular grains in (e) and (f). 158 (a) (b) (c) (d) (e) (f) Figure 6.7: Photomicrographs of cross-section through termite exit towers, com- posed of subrounded quartz particles in a clay-organic matrix. Collected on the surface of an active heuweltjie on Papegaaiberg (Stellenbosch). Inward side of exit towers indicated. Termite excrement pellet (arrowed) within wall of exit tower. (a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE, (d)?ppl, (e)?xpl, (f)??-plate, fast-direction NW-SE. 159 coarse material mostly comprises < 30 ?m, angular, acicular-prolate fragments likely to be quartz. There are occasional 4 mm spheroidal quartz particles. The micromass appears to be an organic-clay mixture, and has an undifferentiated to stipple-speckled b-fabric. The excrement contains spherulites 8 to 15 ?m diameter, non-effervescent with cold 10% HCl, with pseudo-negative uniaxial interference colours (Figure 6.6). The termite exit towers (Figure 6.7) have a 6 to 7 mm internal diameter. The wall is 3 to 5 mm thick and has a massive microstructure. They contain consid- erably more coarse material (c/f5?m ratio 3:1 and a close porphyric c/f-related distribution). The coarse material comprises angular, predominantly spheroidal quartz fragments 0.3 mm to 20 ?m diameter. There is no apparent change in composition (clay, organic material) on either inner- or outer wall surfaces. They are non-calcareous. The micromass appears to be organic matter- clay mixture with an undifferentiated b-fabric. They resemble the material shown by Jungerius et al. (1999, p. 355). The wall includes one ellipsoid excrement (0.85 x 0.3 mm), identical to the material in the excrement sample. The only spherulite (with pseudo-negative uniaxial interference colours) observed in both of the towers sectioned occurred within this excrement. The spherulites are similar to the spherulites in herbivore dung described by Canti (1998). According to Canti (1997), spherulites in herbivore dung are 5-15 ?m diameter, compared to the 8 to 15 ?m diameter for those observed in the Stellenbosch heuweltjie frass (Figure 6.6). The spherulites did not effervesce in cold 10% HCl. No reaction with HCl was observed by Brochier et al. (1992) either. However, Canti (1997) noted that the gas bubbles (but usually only one per spherulite) were clearly evolved, appearing almost immediately on wetting, and so could only be observed by allowing a front of HCl to advance onto dry spherulites in the field of view at high power magnification. This method was employed in this study but produced no effervescence. 6.3.4 Micromorphology of the heuweltjie at Papendorp Full thin sections descriptions through a cross-section of the Papendorp heuweltjie are presented in Table 6.2, with the corresponding pedofeatures pictured in Fig- ures 6.13 to 6.23. These are the first micromorphological data presented for South African heuweltjie soils. 160 Table 6.2: Thin section descriptions of samples from Table 6.1. Abbreviations: c/f: coarse/fine; diam.: diameter; incl.: including; max.: maximum; occ.: oc- casional; OIL: oblique incident light; PPL: plane polarised light; XPL: cross polarised light. Heuweltjie centre; Orthic (Ochric) A (sample 2.1, 20 cm deep). Vertical section; size 40 x 100 mm; thickness 30 - 40 ?m. See Figure 6.8(a). Micro- structure: Bimodal ped size distribution: 1) peds 2 to 3 cm diam.: moderately separated suban- gular blocky to subrounded microstructure; 20% of section; 2) highly separated sub- rounded spheroidal crumb microstructure. Interped areas (40% voids) not as densely packed as peds. Groundmass c/f limit: 5 ?m c/f ratio: 9:1 c/f-related distribution: Coarse monic to fine chitonic (better expressed in OIL, grains coated by amorphous Fe-oxides incl. hematite (red in OIL)). Coarse Material: Mostly subrounded-subangular quartz, coated with fine hematite (red OIL). Bimodal size distribution: 0.4?0.8 mm diam. (15%) and 0.05?0.1 mm (70% in peds; 30% between). Also orthopyroxene, garnet, zircon, opaques (metallic lustre in OIL). Some altered organic material. Micromass: Very little fine material. Mostly amorphous Fe-oxides, incl. hematite (red in OIL). Some amorphous organic material, in one case linearly concentrated along the top of a ped. Undifferentiated b-fabric. Pedofeatures: Nodules Few moderately impregnated Fe (hydr)oxides. Coatings Sand and silt particles coated with red (OIL) Fe oxides. Channels Some vertical channels which may have originated from water movement, and channels around peds. Heuweltjie centre; hardpan carbonate (petrocalcic) sample 2A. 3 x vertical sections; size 40 x 55 mm; thickness 30?35 ?m. See Figures 6.8?6.11. Micro- structure: Weak to moderately separated granular microstructure dominant. Granules (?peloids?; ?ooids?) 0.3?0.03 mm diam. Channels, chambers (together 10%). Some (5%) thin planar voids (0.6 mm x 3 cm), some of which contain fine roots in handspecimen and define larger scale moderately separated subangular blocky microstructure (2 cm diam. peds). Groundmass c/f limit: 5 ?m c/f ratio: 1:8 c/f-related distribution: Double-spaced to open porphyric, chitonic. Coarse Material: Mostly subrounded-subangular very fractured quartz. Max. diam. 0.5 mm, common 0.4 mm, and also 0.04 mm diam. Opaques, generally rounded, garnet, orthopyroxene, occ. zircon, and also mica(?). Tissue residue. Micromass: Micrite interspersed with clay. PPL: Yellow to brown, speckled. XPL: crystallitic b-fabric dominant (micrite), also granostriated (clay; micrite). continued... 161 Table continued from previous page Pedofeatures: Nodules i) Clay-quartz compound fragments (Figures 6.8(e)-(h)): rounded, generally spherical 1 cm to 0.7 mm diam. clay aggregates, distinguishable also in handspecimen. Less calcite than matrix, darker in PPL, more Fe (hydr)oxides, and have a granostriated b-fabric; not crystallitic like matrix . These clay nodules contain quartz grains (coated with oriented clay coatings) and cross-cutting micrite veins. The entire nodule often has a micrite hypocoating (banded, speckled). ii) Limpid yellow clay nodules (1 to 0.04 mm) (Figures 6.9, 6.10(a)?(e), 6.11): suban- gular, very low 1st order interference colours, often concentric-striated b-fabric, also cross-striated. Some coated with micrite hypocoatings giving the appearance of an ?ooid? (Figure 6.9(a)), larger ones have Mn(?), and Fe accumulations. Some have concentric dark rings and core, possibly oxides. Some nodules have an internal radial and concentric orientation (Figure 6.10(c)). The limpid cores often show pseudo- negative uniaxial interference figures, and one seems to be overgrown from the centre by micrite (Figures 6.10(d)?(e)). iii) Micrite nodules (granules, resembling ?peloids?) Figure 6.9, likely mixed with clay, 0.35 mm diam. One shows a pseudo-positive interference figure, some seem to be combined radial/concentric orientation, but generally the orientation not clear, pos- sible due to the small size of the particles. These seem related to ii) above, it may be that the section does not pass through the core. iv) Opaque Localised moderate to strongly impregnated Fe (hydr)oxide nodules (dark in PPL, yellowish-green in OIL), and accessory dendritic and dispersed Mn oxides (10% section, effervesces in cold H2O2) . Coatings i) Quartz and orthopyroxene grains coated with red (in OIL) Fe (hydr)oxides (as in sample 2.2) (Figures 6.8(c)?(f), 6.9(a)) many of which are in turn coated by micrite (mixed with clay) hypocoatings (?ooids?), and some by acicular calcite (radial orien- tation). The micrite hypocoatings are spherical, even when coating an elongate grain (Figure 6.8(e)). In some cases grain is coated by a limpid orange oriented (length- slow grain-parallel and perpendicular) clay coating (Figures 6.10(h)?(j)), which can in turn coated by a micrite hypocoating. Sometimes fibro-radial orientation can be discerned. ii) Limpid clay nodules coated with micrite (?ooids?), but not coated with hematite as the quartz and orthopyroxene grains (as in i) above). Instead there is generally a dark boundary (oxides?) between the nodule and the micrite coating (Figures 6.9), 6.10(d), (e), 6.11). Sometimes fibro-radial orientation can be discerned. Quite common is pseudo-negative limpid core zoning out to pseudo-positive micrite (Figure 6.9(c)), and in one case a second generation of micrite causes the ?ooid? to zone back to pseudo-negative (Figure 6.10(e)). iii) Vughs coated by micrite. iv) Some micrite (+clay) seems to form laminar coatings of larger clay aggregates, resembles pendents but not restricted to the lower part (Figures 6.8(e)?(h). Channels i) Filled with limpid yellowish clay with cross-striated b-fabric, some channels are mixed with micrite and have a crystallitic b-fabric. ii) Coated by dark amorphous organic matter (Figures 6.8(b)?(d)). Some are contin- uously filled with more loosely packed matrix material, possibly ellipsoid mineralo- organic excrements, some contain tissue residue. Phytoliths i) 0.6 x <0.1 mm with some square cells (0.1 x 0.05 mm) which are colourless in PPL, have 4th order green/pink interference colours, and are composed of calcite not Ca- oxalate (dissolve in 2N acetic acid) (Figures 6.10(f)-(g)). Covered by matrix micrite in some places. ii) Silica phytolith (isotropic). continued... 162 Table continued from previous page Excrement Ellipsoidal, dark, speckled, 0.02 mm diam. (Figures 6.8(c)-(d)) HCl etching: Limpid yellow clay nodules become clearer, orientation easier to see. They appear to be oriented clay. b-Fabric is concentric striated. Quartz fractures show less relief (Figure 6.11). 2m from heuweltjie centre; hardpan carbonate (petrocalcic) sample 2B. 3 x vertical sections; size 40 x 55 mm; thickness 30?35 ?m. See Figures 6.14?6.17. Micro- structure: Weakly separated granular microstructure. Granules (?peloids?; ?ooids?) 0.3 - 0.03 mm diam. Groundmass c/f limit: 5 ?m c/f ratio: 1:8 c/f-related distribution: Double-spaced to open porphyric; chitonic. Coarse Material: Mostly subrounded-subangular very fractured quartz, 0.5 - 0.04 mm diam., commonly 0.4 mm. Opaques, generally rounded, garnet, cpx, occ. zircon. Micromass: Micrite (less abundant than 2A) interspersed with clay and oxides. PPL: Yellow to brown, speckled, with Fe (hydr)oxides (dark PPL, yellowish-green OIL), and accessory dendritic or dispersed Mn oxides (effervesces cold H22O2) in 40% section - more abundant than 2A. XPL: speckled b-fabric dominant, also granostriated. Pedofeatures: Nodules i) Clay-quartz compound fragments (Figures 6.14): Rounded, generally spherical 0.7 mm diam. clay aggregates, distinguishable also in handspecimen. Very similar to the matrix, but have less calcite, and have a grano-circular-striated b-fabric, not crystallitic. They contain quartz grains (coated as the matrix), (granostriated) clay, opaque material, and what appears to be altered organic remains. One clay aggregate (Figure 6.17(b)?(d)) has low birefringence (sepiolite?) and contains quartz grains and needles. The clay forms spherical coatings on the angular quartz grains and needles that are aggregated to form the nodule. ii) Opaque Localised moderate to strongly impregnated Fe (hydr)oxide nodules (dark in PPL, yellowish-green in OIL), and accessory dendritic or dispersed Mn oxides (effervesces cold H2O2) in 40% section - more abundant than 2A. The nodules appear to have less calcite than the matrix, some have cross-cutting micrite veins. iii) Limpid yellowish clay nodules (300 - 70 ?m) (Figures 6.15, 6.16): subangular to rounded, with a cross-striated b-fabric and very low (first order) interference colours, occasionally will show pseudo-negative uniaxial interference colours, may zone out- wards to pseudo-positive (commonly in the micrite coating); sometimes fibro-radial orientation can be discerned. Some have concentric dark narrow rings which may be oxides, or a dark core. iv) Micrite nodules (granules, resembling ?peloids?) (e.g. Figure 6.15) likely mixed with clay, 0.35 mm diam. One shows a pseudo-positive interference figure, some seem to be combined radial/concentric, but generally the orientation not clear, possible due to the small size of the particles. v) Digitate calcite nodule 0.2 mm diam. Coatings i) Quartz and orthopyroxene grains coated by red (OIL) Fe (hydr)oxides (as in samples 2.2, 2A), in turn coated by micrite (mixed with clay) and sepiolite (Figure 6.15). ii) Limpid clay nodules occasionally coated with hematite and/or other oxides (Figure 6.16), but most commonly only by micrite. iii) Some micrite and sepiolite coating larger clay aggregates (Figures 6.14, 6.17(b)? (d)), micrite coating is thicker around the angular parts of the nodule to give a spherical ooid-like structure. continued... 163 Table continued from previous page Altered grains Generally elongated, 3 x 0.6 mm, speckled orange-brown, have a circular striated b-fabric (Figure 6.16). Pendents and cappings Large scale (cm?s) micrite laminations (mixed with clay(?)) (for example Figures 6.17(e)-(f)). Excrement Rounded, 0.05 mm diam., ellipsoidal, dark, micritic, appears to be contained in plant tissue residue (Figure 6.17(a)). A lot of tissue residue seems altered to micrite, and is associated with more clay rich fragments. Channels i) Filled with limpid yellowish clay with cross-striated b-fabric and very low interfer- ence colours (sepiolite), some channels are mixed with micrite and have a crystallitic b-fabric. 3.2 m from heuweltjie centre; hardpan carbonate (petrocalcic) sample 2C. 3 x vertical sections; size 40 x 25 mm; thickness 30?35 ?m. See Figures 6.18?6.19. Micro- structure: Weakly separated granular to massive microstructure. Granules (?peloids?; ?ooids?) 0.3 - 0.03 mm diam. Groundmass c/f limit: 5 ?m c/f ratio: 1:8 c/f-related distribution: Double-spaced to open porphyric; chitonic. Coarse Material: Mostly subrounded-subangular very fractured quartz. 0.5?0.04 mm diam., commonly 0.4 mm. Opaques, generally rounded, garnet, orthopyroxene, occ. zircon. Micromass: Micrite interspersed with clay. PPL: Yellow to brown, speckled. XPL: speckled b- fabric dominant, also granostriated. Pedofeatures: Nodules i) Clay-quartz compound fragments (Figures 6.18(a), (b): Rounded, generally spheri- cal (also angular) 0.7 mm diameter clay aggregates, distinguishable also in handspec- imen. Very similar to matrix, but have less calcite. Micrite veins cross cutting, and micrite hypocoatings (banded, speckled). They are darker in PPL, contain more Fe (hydr)oxides than the matrix, and have a granostriated b-fabric, not crystallitic. They contain quartz grains (Fe-, clay coated) , (granostriated) clay, and opaque material. ii) Limpid yellow clay nodules (Figures 6.19(a)?(c): subangular (300 - 70 ?m) which have a cross-striated b-fabric and low first order interference colours, occasionally show pseudo-negative uniaxial interference figures, and may be coated with micrite and acicular calcite (radial orientation). Some have concentric dark narrow rings, which could be oxides and is suggestive of multistage growth. One has a light-coloured core (quartz?) (Figure (d)), in contrast to the dark cores usually observed in samples 2A?2B. Coatings i) Quartz and orthopyroxene particles coated with red (OIL) Fe (hydr)oxides (as in samples 2.2, 2A, 2B), in turn coated by acicular calcite (radial orientation, Figures 6.19(f), (h), easier to see than in samples 2A-B), and also by grain-parallel oriented clay that is revealed on etching in HCl (Figures 6.19(g), (i)). ii) Limpid yellow clay nodules have oxide coatings, covered by radially oriented calcite (Figures 6.19(a), 6.19(d)). Channels; excrement Filled with 0.05 mm diam. spheres (?peloids?), which appear to be excrement and have the same b-fabric (speckled, micrite) as matrix (Figures 6.18(e), (f)). Matrix and excrement both dissolved during HCl-etching. Some channels have chambers and most have fine, dark (hypo)coatings that appear to be organic matter. continued... 164 Table continued from previous page Pendents Large scale (cm?s) micrite laminations (Figures 6.18(a)?(d)). HCl etching: Figures 6.19(d)-(i) Limpid yellow clay nodules (spherulites and concentric nodules) become clearer, orientation easier to see, and concentric dark rings are less prominent. Appear to be clay with low interference colours. Fractures in silicate grains show less relief. Micrite matrix (?peloids?) dissolved. 7.5 m from heuweltjie centre; ?sepiolitic? pedocutanic (argic) (sample 2D). 2 x vertical sections; size 20 x 15 mm; thickness 20 to 25 ?m. See Figures 6.20, 6.21, 6.22. Micro- structure: Weakly separated vughy to massive microstructure. Groundmass c/f limit: 5 ?m c/f ratio: 1:2 c/f-related distribution: Single-spaced porphyric; chitonic. Coarse Material: Mostly subrounded to subangular very fractured quartz. 0.5 to 0.04 mm diam., com- monly 0.4 mm. Opaques, generally rounded, occ. zircon, feldspar. Micromass: PPL: Pale yellow-brown, mostly limpid clay (and silica?). XPL: Grano- and circular striated b-fabric. Pedofeatures: Nodules i) Clay-quartz compound fragments: Angular compound fragments comprising either 1) quartz grains and clay nodules with granostriated b-fabric (caused by oriented clay coatings) set in a darker, more speckled matrix than the surrounding matrix, and the entire fragment is coated with clay oriented parallel to the edges. Along one side of one fragment there is a zone of clay with very low interference colours, consisting of interwoven domains of oppositely oriented fibres (perpendicular and parallel to the nodule) which appear to be sepiolite (Figure 6.21), or 2) comprising quartz grains set in a fibrous matrix (Figure6.22) that is consistent with sepiolite. Coatings Quartz particles coated with red (OIL) Fe (hydr)oxides (as in samples 2.2, 2A to C), which is in turn coated by oriented clay coatings (Figure 6.20). Clay nodules not coated with hematite. Heuweltjie edge (8 m from 2D); dorbank (petroduric), sample 1.3. 1 x vertical section; size 45 x 90 mm; thickness 35 to 40 ?m. See Figure 6.23. Micro- structure: Vughy microstructure Groundmass c/f limit: 5 ?m c/f ratio: 2:1 c/f-related distribution: Chito-gefuric to close porphyric. Coarse Material: Mostly subrounded-subangular very fractured quartz grains 0.4 to 0.03 mm diam. coated with fine hematite (red in OIL). Some larger grains have embayments (0.02 mm) filled with hematite (Figure 6.23(a)). Occ. orthopyroxene, garnet, opaques, zircon. Micromass: Silica and clay and Fe oxides. PPL: Orange-brown, mostly speckled. XPL: Granos- triated b-fabric. Pedofeatures: Nodules: i) Spherulites, some of which show pseudo-negative uniaxial interference figures, some appear to have an isotropic core. Others show concentric banding. continued... 165 Table continued from previous page ii) Aggregates (Figures 6.23(a)?(c)) containing quartz grains with red Fe coatings, cemented by silica (mixed with clay and Fe to give a limpid orange colour). The aggregates contain more interstitial silica and fewer quartz grains than the matrix. Coatings: i) Quartz particles coated with red (OIL) Fe (hydr)oxides (as in 2.2, 2A to D). Some larger grains have embayments (0.02 mm) filled with hematite (Figures 6.23(a), (g) to (f)). ii) Laminated (in places) orange-brown, generally limpid, compound clay and silica coating a) spherical, rounded peds 0.5 to 1 cm diam., b) an elongated channel(?) 3 x 0.5 cm, and c) large irregular (5 cm) peds. In one instance the amorphous silica is clearly on the outer part (Figure 6.23(d) to (f)). iii) Silica and Fe-stained clay compound coatings around channels and vughs. The silica is in places transparent in PPL but in other places seems to be coloured by iron and mixed with clay. In some cases the silica is clear, and occurs on the void- side of the coating (6.23(g) to (h)). In one case it can be discerned as edge-parallel length-slow (Figure 6.23(i)). 166 (a) (b) (c) (d) (e) (f) (g) (h) Figure 6.8: Photomicrographs from centre of heuweltjie. Way-up is top of page.(a) Orthic A horizon (sample 2.1, 20 cm deep), v: void, ppl. (b)-(g) Calcrete (sample 2A) underlying sample 2.2. (b) Tunnel lined with organic matter, ppl. (c) Chamber and tunnel lined with organic matter; excrement in nodule; quartz grain coated with red Fe, ppl. (d) xpl. (e)?ppl, (f)?xpl, (g)??-plate, fast-direction NW-SE: clay nodule with micrite hypocoating, internal ooid-like structures form granostriated b-fabric. Note thinning of micrite coating around an angular hematite-coated quartz grain to form spherical ooid (arrowed in (e).)(h) Clay nodule fractured by micrite, ppl. 167 (a) (b) (c) (d) (e) (f) Figure 6.9: Photomicrographs from calcrete (2A) in the centre of heuweltjie. Way-up is top of page.(a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: ori- ented granules (arrowed top-left), unoriented granules/?peloids?(arrowed centre), limpid oriented core with an oriented micrite coating (arrowed right); quartz grains coated with red Fe. (d)?ppl, (e)?xpl, (f)??-plate, fast-direction NW-SE: limpid oriented nodule with thin unoriented micrite coating (arrowed top), ori- ented micrite nodule (arrowed centre), quartz grain with oriented micrite coating (arrowed bottom-left), oriented limpid nodule undergoing ?micritization? from the centre (arrowed bottom-right). 168 (a) (b) (c) (d) (e) (f) (g) (h) (i) (j) Figure 6.10: Photomicrographs from calcrete (2A) in centre of heuweltjie. Way-up is top of page. (a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: radial and concen- tric orientation in layers of nodule.(d)?ppl, (e)??-plate, fast-direction NW-SE: limpid pseudo-negative core (C) surrounded by pseudo-positive (M1) and pseudo-negative (M2) micrite.(f)?ppl, (g)?xpl: calcite (which dissolved in 2N acetic acid) phytolith.(h)?ppl, (i)? xpl, (j)??-plate, fast-direction NW-SE: Fe, clay coatings: note perpendicular orientations. 169 (a) Scale 0.1 mm, ppl. Qtz: quartz. (b) Same view as (a); etched in 1M HCl. (c) Same view as (a); xpl. (d) Same view as (b); etched in 1M HCl. (e) Same view as (a); ?-plate, fast-direction NW-SE. (f) Same view as (b); ?-plate, note that fast direction is opposite to (e). Figure 6.11: Photomicrographs: HCl etching of peloids and a coated limpid nodule with pseudo-negative uniaxial interference figure (sample 2A, calcrete in centre of heuweltjie). Way-up is top of page. 170 (a) Scale 0.1 mm, ppl, magnified in (d) to (i). (b) Same as (a), xpl. (c) Same as (a), ?-plate, fast-direction NW-SE. (d) Magnification of (a), scale bar 100 ?m. (e) Magnification of (a), scale bar 50 ?m. (f) Magnification of (e), sepiolite (grey mesh), radial calcite (white). Scale 5 ?m. (g) Magnification of (a), scale bar 50 ?m. (h) Magnification of (g), radial orientation of needle fibres. Scale 5 ?m. (i) Magnification of (g), radial orientation of needle fibres. Scale 5 ?m. Figure 6.12: Photomicrographs and corresponding ESEM images from calcrete (2A) in centre of heuweltjie. Calcite is white; grey colours are lower atomic numbers such as Mg or Si. Way-up is to the right. 171 (a) ?Ooid? composed of radial acicular calcite. EDAX suggests sepiolite present in lesser amounts. Scale 10 ?m. (b) Radial, acicular calcite coating silicate grain. EDAX suggests sepiolite present in lesser amounts. Scale 5 ?m. (c) Fragment: micrite nodules (ooids), note hollow centres, and fungal filament. Scale 2 ?m. (d) Fragment showing radial orientation of acicular calcite in nodules. Area magnified in (e) arrowed. Scale 10 ?m. (e) Magnification of (d), MB rod arrowed (after Verrecchia and Verrecchia (1994)). Scale 2 ?m. (f) Sepiolite fibres in the same fragment. Scale 2 ?m. Figure 6.13: ESEM images of fragments and thin sections from calcrete (2A) in centre of heuweltjie. 172 (a) (b) (c) Figure 6.14: Photomicrographs from calcrete (2B), 2 m from centre of heuweltjie. Way-up is top of page. (a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: quartz-clay aggregate with sepiolite (S) coating covered by micrite (M) coating. Note thickening of micrite coating to give spherical ooid-like structure (arrowed in (a)). This pattern is also present around the quartz grain on the left. Dispersed opaques in nodule and matrix are Mn and Fe (hydr)oxides. 173 (a) (b) Figure 6.15: Continued - photomicrographs from calcrete (2B), 2 m from centre of heuweltjie. Way-up is top of page. (a)?ppl, (b)?xpl: sepiolite nodule and layered calcite-sepiolite grain coatings in a micrite-clay matrix. 174 (a) (b) (c) Figure 6.16: Continued - photomicrographs from calcrete (2B), 2 m from centre of heuweltjie. Way-up is top of page. (a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: limpid clay nodules (pseudo-negative and unoriented), with oriented (pseudo-positive) micrite coatings. 175 (a) (b) (c) (d) (e) (f) Figure 6.17: Continued - photomicrographs from calcrete (2B), 2 m from centre of heuweltjie. Way-up is top of page. (a)?ppl: calcified, excrement-filled or- gan residue. (b)?ppl, (c)?xpl, (d)??-plate, fast-direction NW-SE: sepiolite clay- quartz nodule containing organ residue (needle, arrowed) in oxide-rich matrix. Quartz grains coated by red Fe oxides. Needle and grain above it are coated by clay, giving the appearance of spherical ?ooids?. Lower part of the nodule shows grain-parallel clay orientation. (e)?ppl, (f)?xpl: micrite capping ped (arrowed), in a more oxide-rich matrix (m). 176 (a) (b) (c) (d) (e) (f) Figure 6.18: Photomicrographs from calcrete (2C), 3.2 m from centre of heuweltjie. Way-up is top of page. (a)?ppl, (b)?xpl: calcite pendent below clay- quartz compound fragment. (c)?ppl, (d)?xpl: calcite pendent associated with an organic matter-lined tunnel. (e)?ppl, (f)?xpl: micrite-containing excrement in tunnel. 177 (a) (b) (c) (d) (e) (f) (g) (h) (i) Figure 6.19: Continued - photomicrographs from calcrete (2C), 3.2 m from centre of heuweltjie. Way-up is top of page. (a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: fibro-radial calcite coating a limpid clay nodule. (d)?ppl, (e)?ppl, HCl- etched, (f)?xpl, (g)?xpl, HCl-etched, (h)??-plate, fast-direction NW-SE, (i)??- plate, HCl-etched: fibro-radial calcite and grain-parallel clay coating a quartz grain; 2-generations of clay oriented around quartz fragment forming an ooid- like structure with a pseudo-negative interference figure. 178 (a) (b) (c) (d) Figure 6.20: Non-calcareous ?sepiolitic? pedocutanic (argic) (2D), 7.5 m from centre of heuweltjie.(a) SEM image of fragment, showing smooth coatings of grains. Scale 10 ?m. (b)?ppl, (c)?xpl, (d)??-plate, fast-direction NW-SE, way-up is top of page: photomicrographs of grain-parallel, aligned clay and dominating granostriated b-fabric. Section is thinner than 2A to 2C. 179 (a) (b) (c) (d) (e) (f) Figure 6.21: Continued - photomicrographs of non-calcareous ?sepiolitic? pedocu- tanic (argic) (2D), 7.5 m from centre of heuweltjie. Way-up is top of page.(a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: clay capping on a (rotated) quartz- clay aggregate. (d)?ppl, (e)?xpl, (f)??-plate, fast-direction NW-SE: magnifica- tion of clay capping showing interwoven fibre domains. 180 (a) (b) (c) (d) (e) (f) Figure 6.22: Continued - photomicrographs of non-calcareous ?sepiolitic? pedo- cutanic (argic) (2D), 7.5 m from centre of heuweltjie. Way-up is top of page.(a)? ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: Aggregate with sepiolite matrix. (d)?ppl, (e)?xpl, (f)??-plate, fast-direction NW-SE: magnification of fibrous se- piolite matrix showing fibre domains. 181 (a) (b) (c) (d) (e) (f) (g) (h) (i) Figure 6.23: Photomicrographs of non-calcareous dorbank (petroduric, sample 1.3), 8 m from sample 2D. Way- up is top of page.(a)?ppl, (b)?xpl, (c)??-plate, fast-direction NW-SE: aggregate (top-left) in granostriated matrix of Fe-coated quartz grains cemented by limpid-orange silica mixed with clay and Fe. (d)?ppl, (e)?xpl, (f)??-plate, fast-direction NW-SE: silica-rich (Si) over clay-rich (Cl) capping on ped. Voids in ped contain silica. (g)?ppl, (h)?xpl, (i)??-plate, fast-direction NW-SE: Fe coating quartz grain (Q); silica (length-slow arrowed in (i)) lining voids (v). 182 Biological activity Evidence of biological activity could only be found in the calcrete in the centre of the heuweltjie (samples 2A to 2C). Neither the ?sepiolitic? pedocutanic (argic; sample 2D) halfway to the edge, nor the dorbank (petroduric; sample 1.3) at the edge exhibited such evidence. This is consistent with observations that the centre of the heuweltjie is where the termite activity is concentrated. Tunnels and chambers were lined with organic matter (Figures 6.8(b) to (d); 6.18(c) to (d)). The dark ellipsoid (0.8? 0.4 mm) just above the tunnel in Figures 6.18(c) to (d) and in the ped in Figures 6.17(e) to (f) appear to be termite excrement, based on the morphology of the termite excrement from the active heuweltjie in Stellenbosch (Figures 6.6 to 6.7). Smaller excrement particles in the form of 0.1 to 0.05 mm diameter calcareous spheres are also abundant: in one case within what appears to be termite excrement (Figures 6.8(c) to (d)); in calcified plant organ residue (Figures 6.17(a)); and in a tunnel (Figures 6.18(e) to (f)). Plant remains are present, some of which appear to be fresh (such as Figure 6.8(c) to (d)), some that seem to be needles that have been coated in clay (Figure 6.17(b) to (d)), and some (most commonly) that appear to have been calcified (Figures 6.17(a), 6.10(f) to (g)). Highly birefringent phytoliths such as that pictured in Figures 6.10(f) to (g) are composed of calcite and not calcium oxalate, since they dissolved completely in 2N acetic acid. Rectangular silica phytoliths (isotropic) are also present. Horizon over calcrete in the heuweltjie centre The horizon is non-calcareous and has very little clay (Figure 6.8(a)). Interped areas and some voids surrounding the peds have much less of the fine (50 to 100 ?m) material. Some of the coarser (0.4 to 0.8 mm) material seems wedged in vertically elongated voids (planes, ped margins, channels). The bleached A horizon (Table 6.1) is also consistent with clay dispersion and leaching (Ellis, 1988; Francis et al., 2007). The lack of clay and presence of hematite coatings suggest a well-drained environment, and is consistent with its lower-pH and the tonguing nature of the transition to hardpan carbonate (petrocalcic) horizon below (Table 6.1). 183 Calcrete in the heuweltjie centre The calcrete contains much fewer sand grains (10-20% vs 60% for the overlying orthic (ochric) A horizon), and virtually no void space (compare Figure 6.8(a) with Figure 6.15, for example). The calcite is mostly micrite, possibly a result of the high clay content of the matrix (Wieder and Yaalon, 1974, 1982, in Wright and Tucker 1991). Hay and Reeder (1978) also noted that porosity appeared to control the micrite distribution for ooids formed by replacement of clay coatings. Calcic pendants and cappings (Figures 6.17(e) to (f), 6.18(a) to (d)) suggest remobilisation of calcite, consistent with the more acidic, well-drained overlying horizons and tonguing transition to the calcrete. This tonguing transition on the heuweltjie crest resembles the karstic processes that operated on the anticline crests in folded Miocene limestones in the Madrid Basin (Sanz, 1996, in Alonso- Zarza, 2003). Calcareous ooids and peloids The terms ?peloid? and ?ooid? are applied descriptively to carbonate sedimentary rocks, usually marine limestones. They were not used in Stoops? (2003) soil micromorphological description scheme; they are included instead in the concepts of ?granular microstructure? and micrite ?(hypo)coatings?. Boggs (1995, p. 199) defined ?ooid? as a general name for coated carbonate grains (0.02 to 2 mm in size) containing a nucleus (commonly a shell fragment, pellet or quartz grain) surrounded by thin coatings of fine calcite or aragonite. Most ooids display an internal structure consisting of concentric layers, but some show a radial structure. Boggs (1995, p. 201) defined ?peloid? as a nongenetic term for carbonate grains composed of micro- or cryptocrystalline calcite/aragonite that do not display distinctive internal structures, which are generally smaller than ooids (0.03 to 0.1 mm). Peloids may be: 1) Fecal pellets, the most common kind of peloid, produced by organisms that ingest calcium carbonate muds and generally contain enough organic matter to make them appear dark or opaque. 2) Formed by the micritization of small ooids or rounded skeletal fragments, due to boring activities of certain organisms (boring algae) which convert the original grains into a nearly uniform homogeneous mass of micrite. 3) Very small, well rounded intraclasts formed by reworking of semiconsolidated mud or mud aggregates (Boggs, 1995). 184 Peloids and ooids are not restricted to limestones of marine origin. Wright and Tucker (1991) noted that these are a very common feature of many cal- cretes. Siesser (1973) referred to them as ?diagenetic ooids/intraclasts? to distin- guish them from their marine counterparts, and noted that Quaternary calcretes contain the only known ooids/intraclasts in any South African Cretaceous or Cenozoic rocks. Ooids and peloids have also been recorded from pedogenic cal- cretes in northern Tanzania (Hay and Reeder, 1978); in the southwestern United States (Hay and Wiggins, 1980); Tarragona, northeastern Spain (Calvet and Ju- lia, 1983); calcretes developed in the Old Red Sandstone of Scotland (Wright et al., 1993); the Kalahari valley (Nash and McLaren, 2003); and calcretes con- taining bee nests on the Eastern Canary islands (Alonso-Zarza and Silva, 2002). The granular microstructure that is most strongly developed in the center of the heuweltjie (Table 6.2) comprises 0.3 to 0.03 mm diameter granules that re- semble ?peloids?, and abundant micrite-coated grains that resemble ?ooids? (for example, Figures 6.9, 6.10(a) to (e), 6.11). They were not observed in a calcrete described in Chapter 3, which is a ?non-heuweltjie? calcrete farther north on the Namaqualand coastal plain near Kleinzee. Pellets have been described in soils associated with termites by Stoops (1964) and Mermut et al. (1984). The ooids in the Papendorp heuweltjie often have a core of limpid yellowish clay/silica with low birefringence, usually rimmed with a dark layer under the micrite (Figures 6.9(a) to (c); 6.10(d) to (e), 6.11; 6.16). Less common is a nucleus of Fe-coated quartz grains (top of Figures 6.8(c) to (d); bottom-right of Figures 6.8(e) to (f); bottom-left of Figures 6.9(d) to (f)). In some cases the core seems to be hollow (Figures 6.13(a), (c)). In most cases the coatings are composed of radially oriented acicular calcite, which also commonly forms thin coatings on much larger detrital grains (Figures 6.13(a), (b), (d), 6.19), similar in principle to the ?pseudo-ooids? of Hay and Reeder (1978). In many cases the radial acicular calcite is contained within concentric envelopes, to form successive layers of differently-oriented and/or oppositely-elongated calcite (Figures 6.10(a) to (e), 6.12(e) to (i)). Etching in HCl to remove a radial calcite coating on a detrital grain revealed a concentrically-oriented, length-slow coating that appears to be clay (Figures 6.19(d) to (i)). According to Siesser (1973), the formation of diagenetically formed ooids and intraclasts in South African coastal calcretes takes place as rain water percolates 185 through an unlithified calcareous sand deposit, dissolving carbonate along its downward path. This carbonate-laden water is eventually checked in its descent and drawn upward somewhat by capillary action. Evaporation and soil suction in the uppermost zones of the sediment body cause precipitation of a concentric coating of carbonate mud around individual grains (incipient ooids) or compos- ites of grains (incipient intraclasts). As this dissolution-precipitation cycle is repeated, the micritic coatings thicken and locally push grains apart. Davies et al. (1978) found that ooids synthesised in quiet-water conditions had a radial orientation of carbonate crystals (as observed in the heuweltjie), whereas the ooids that formed in agitated conditions has a prevalently tangential orien- tation due to the grain collisions which inhibited any crystal growth other than tangential. The most important parameters in quiet-water ooid formation were high molecular weight organic components, the presence of Mg in addition to Ca, the presence of carboxylic groups (for example solutions containing humic acids), and a capacity to participate in hydrophobic/hydrophilic interactions. These parameters were critical to membrane formation, and it was the forma- tion of the hydrophilic high molecular-weight membranes that was important in the formation of ooids. These membranes formed concentric shells which acted as growth surfaces for carbonate, and also induced periodicity in carbonate pre- cipitation. There is much literature on the possible mechanisms involved in the formation of rounded calcite aggregates, spheres, spherulites and peloids, deal- ing particularly with the question of whether biotic or abiotic processes were the cause (for example, Tracy et al., 1998a; Tracy et al., 1998b; Gonza?lez-Mun?oz et al., 2000; Raz et al., 2000; Braissant et al., 2003; Bosak et al., 2004; Ferna?ndez- D??az et al., 2006). The mechanism proposed by Davies et al. (1978) for the forma- tion of quiet-water ooids is the best explanation for the formation of ooids in the heuweltjie soils, in view of the number of parallels between the quiet-water ooid- forming processes described by Davies et al. (1978) and conditions in the calcrete of the Papendorp heuweltjie. The heuweltjie ooids are closely associated or cored with sepiolite (Figures 6.4, 6.12(e) to (f), 6.13(f)) with its strongly hydrophilic character (Alvarez, 1984), and its high Mg content. The induced periodicity in carbonate precipitation described by Davies et al. (1978) explains the successive acicular layers seen in the ooids, particularly the successive sepiolite (hydrophilic and therefore a precipitational substrate) and micrite/acicular calcite layers in the coatings (Figures 6.10(a) to (e), 6.12(g) to (i), 6.15 (arrowed), 6.16 (arrowed middle)). 186 Further support for the hydrophilic/hydrophobic membrane theory of Davies et al. (1978) is that the organic material in the Namaqualand soils impart a strongly hydrophobic character (Francis et al. (2007) and Chapter 2). In the Papendorp heuweltjie, it took longer for a drop of water to penetrate the sandy non-cemented A horizon than the cemented subsurface horizons (Table 6.1). The ooids in this heuweltjie are also closely associated with organic/biogenic material: fungal filaments (Figure 6.13(c)) and biogenic calcite (MA and MB rods of Verrecchia and Verrecchia (1994), Figure 6.13(e)). Alonso-Zarza and Silva (2002) noted organic material and organic films associated with ooids formed in calcrete containing bee nests in the Canary Islands. They concluded this favoured the precipitation of calcite and the adhesion of clays (mostly palygorskite). Limpid nodules Pseudo-negative uniaxial ?spherulites? Somewhat of an enigma are limpid yellow nodules 300 to 70 ?m in diameter, with pseudo-negative uniaxial interference figures, usually with a dark round- or dumbbell-shaped core, and a dark rim between them and the matrix calcite (Figures 6.9(a) to (c), 6.11, 6.16). Canti (1997) defined spherulites as ?crystal aggregations with an approximately circular outline and a permanent cross of ex- tinction in crossed polarized light?. Although the samples do show a permanent uniaxial extinction cross under crossed-polars, they do not show the same neo- formed radial crystallitic texture that characterises calcite spherulites commonly observed on laminar calcretes (for example Verrecchia et al., 1995; Mees, 1999). Canti (1998) acknowledged this problem while defining fecal spherulites: since true spherulitic crystallization consists essentially of constrained dendritic growth from the spherulite centre, all spherulites sensu stricto are radial. However, if that was taken as part of a definition, petrographers and micromorphologists would be unable to classify spherical bodies with an extinction cross because of the impossibility of determining the often microcrystalline layout purely mor- phologically. Canti (1998) therefore used the term ?spherulite? to describe all objects displaying the basic extinction cross. The limpid yellow nodules with pseudo-negative uniaxial interference figures pictured in Figures 6.9(a) to (c), 6.11, 6.16 superficially resemble the fecal spherulites found in herbivore dung 187 (Canti, 1997; Canti, 1998) and in fresh termite frass (Figure 6.6, described in Section 6.3.3). They are similar in their pseudo-negative interference figures, dark core and lack of effervesce in cold 10% HCl, but differ in being considerably larger than the 5 to 15 ?m calcite spherulites that commonly occur in herbi- vore dung (Canti, 1997). Seong-Joo and Golubic (1999) noted similar dumbbell- shaped, dark centres in 30 to 500 ?m spherulites formed in silicified stromatolites (carbonates) in China, and noted (p. 194) that the texture is reminiscent of that produced by the process of diagenetic granularization or condensation of organic matter as discussed by Knoll et al. (1988). This is consistent with the termite activity which concentrates organic matter in heuweltjies. In addition, some process akin to the ?granularization or condensation of organic matter? may also occur in the termite gut, producing a dark core like those seen in the termite fecal spherulites (section 6.3.3 and Figure 6.6). Possible Other Morphological Forms The limpid yellow nodules with pseudo-negative uniaxial interference figures pic- tured in Figures 6.9(a) to (c), 6.11, 6.16 appear to be the same material that also forms (a) the limpid nodules with a cross-striated b-fabric and low first or- der interference colours (top, Figures 6.9(d) to (f), Figures 6.19(a) to (c)). This material also appears to form (b) unoriented coatings (interpreted as sepiolite in Figures 6.14, 6.15) juxtaposed with micrite. Sometimes it appears to form (c) oriented coatings around grains (Figures 6.10(h) to (j)), some of which seem to have formed in more than one stage (Figures 6.19(d) to (i), left). Some of the nodules seem to have undergone ?micritization? which gives them a highly birefringent crystallitic b-fabric in the centre and which ESEM shows to be acicular calcite and micrite (Figures 6.9(d) to (f), 6.12, top-left of Figure 6.16). It is unclear whether they are the same material as the ?clay-quartz compound fragments? in Table 6.2, which occur in the ?sepiolitic? pedocutanic (argic) sam- ple 2D as well as the calcrete horizons 2A to 2C. The ?clay-quartz compound fragments? are somewhat darker in PPL but have a matrix that is strongly gra- nostriated, and the granostriations are commonly oriented in a ?pseudo-negative uniaxial? pattern (Figures 6.8(e) to (g), 6.14, 6.17(b) to (d), 6.18(a) to (b)). In the ?sepiolitic? pedocutanic (argic) horizon (sample 2D) it appears to form the fibrous capping as well as the coatings (Figures 6.21, 6.22). 188 Composition The nodules do not react with cold 10% HCl. After etching the thin section in HCl, most of the matrix micrite is removed, and the spherulites show much clearer pseudo-negative uniaxial interference figures (Figures 6.11, 6.19(d) to (i)). This suggests that while some may have been ?micritized?, they are not primarily composed of calcite. Their low Ca, high Mg, and high Si contents (SEM-EDAX analysis such as in Figure 6.12) coupled with their low birefringence argue against them being composed of chalcedony or other carbonate minerals which have a tendency to form ?true? spherulites (aragonite, dolomite and vaterite, for ex- ample). Also, in the Papendorp heuweltjie samples, the limpid yellow nodules are always pseudo-uniaxial negative under crossed-polars, whereas the calcite coatings are generally oriented pseudo-positive (Figures 6.9(a) to (c), top-left; compare to the pseudo-negative limpid nodule coated by a pseudo-positive mi- crite coating on the right), but Figures 6.10(d) to (e) show the variability. In addition to the 1.2 nm sepiolite peak in the calcrete samples 2A to 2C, they and the silica-cemented dorbank sample 1.3 all show a broad rise in the back- ground at around 0.4 nm (Figure 6.4), consistent with amorphous silica (Drees et al., 1989). However, the fibrous nature (particularly Figure 6.12(f) and its ac- companying EDAX analyses) suggests that they are composed of sepiolite rather than (amorphous) silica. The EDAX analyses in Figure 6.12(d) and 6.12(f) have molar Mg/Si ratios of 0.64 to 0-68, consistent with the molar Mg/Si of 0.667 for sepiolite (Stoessell, 1988). Also, where orientation can be discerned it seems to be concentric rather than radial (Figure 6.11), suggesting it is a length-slow mineral like sepiolite (Phillips and Griffen, 1981). Summerfield (1983b, Plate 3.1(h)) noted that length-fast chalcedony occurs predominantly as spherulites with pseudo-uniaxial extinction crosses in southern African silcretes. Summer- field (1983a), however, noted that length-slow chalcedony formed in the presence of Mg2+ and SO4? by replacement (rather than void-fill) in carbonate and evapor- ite host materials. ?Spherulites? with pseudo-negative uniaxial interference figures were also present in a dorbank horizon overlying a ?sepiocrete? (?petrosepiolitic?) horizon in the Knersvlakte (Chapter 1, Figure 1.9). Although the low birefringence of the nodules is consistent with the other ev- idence for sepiolite rather than (amorphous) silica, silica (opal-CT) has been reported to show weak anisotropy in soils. In Figure 6.23(h), for example, the 189 anisotropic void coating was interpreted as silica. Broadly similar pedofeatures (although in a massive rather than oriented form) were noted in Sardinia, Italy in >2.7 Ma buried paleosols (Usai and Dalrymple, 2003). They suggested that the ?milky pedofeatures? as they were termed, were made of different phases of silica such as inorganic opal-A, opal-C and opal-CT, but mainly composed of opal-C. They noted that slight and local birefringence in colourless white-milky pedofeatures had been previously described for opal-CT by Drees et al. (1989). Although silica was the main constituent of the ?milky pedofeatures? (reaching 67% of the total weight), the remaining constituents were around 25% Al2O3, 5% MgO, 2 to 4.6% Fe2O3 and 2% CaO, which they attributed either to im- purities chemisorbed to silica (Drees et al., 1989) or to analytical inaccuracy. Gutie?rrez-Castorena et al. (2006) described weak anisotropy in opaline coatings on channels and fissures in the strongly alkaline sediments of the former Texcoco Lake (Mexico City). They suggested that the weak anisotropism could be related to an initial crystallisation of the opal-A to opal-CT (anisotropic domains), or to the presence of fine clay particles oriented according to the flow direction of the colloids (parallel anisotropic streaks). A possible explanation for these limpid nodules may be a process similar to the pseudomorphic replacement of sepiolite by opal in deposits in the Madrid Basin, Spain (Bustillo and Bustillo, 2000; Bustillo and Alonso-Zarza, 2007). The Madrid Basin typically consists of carbonates, high magnesium clays and opaline cherts. Bustillo and Alonso-Zarza (2007) suggested that the preferential silicification of the sepiolite deposits is due to the fibrous structure and high absorption capacity of sepiolite, which helped to retain interstitial fluids, and the fact that sepiolite and opal are stable under relatively similar geochemical conditions. Bustillo and Alonso-Zarza (2007) proposed that the replacement of sepiolite by opal took place via an initial silicification, which resulted in an atypical, porous opal. In thin section this showed striated birefringence, as a consequence of the sepiolite fibrous structure. When the silicification progressed the fibres became joined and cemented together to form blades, resulting in a compact structure in which the inherited striated birefringence could still be observed. In a later phase, compact and isotropic opal may have formed due to the loss of the inherited sepiolite structure. In some cases, noted Bustillo and Alonso-Zarza (2007), it is difficult to distinguish Mg clay fibres from their imprints on the opal. 190 ?Sepiolitic? pedocutanic (argic) midway to edge of heuweltjie (sample 2D) The clay mineralogy of non-calcareous (Table 6.1) sample 2D is somewhat of an enigma (see discussion in Section 6.3.2). It appears that the cutans are sepiolite, and the matrix is dominated by a 1.4 to 1.5 nm non-expanding mineral. The micromorphology is consistent with this interpretation: the quartz grains coated by strongly oriented clay with low first-order interference colours which appears to be illuvial in origin. This clay is distinct from the interwoven fibre domains of the clay forming cappings and the matrix of nodules (Figure 6.20 to 6.22). The origin of fibrous coating on the side of the clay nodule in 2D (Figure 6.21) is unclear. It may be part of a crust that has been buried by bioturbation, since the clay aggregate fragment to which it is attached seems discordant with the surrounding material, or it may be a sepiolite cutan, many of which are evident in the field (Figure 6.2). In Figure 6.22 it appears to form the matrix of a nodule that is discordant with the rest of the matrix. The interwoven, length-slow fibres with low birefringence occur in domains similar to those described by Stahr et al. (2000) for a palygorskite-cemented hardpan in Portugal. While the illuviation of clay in the matrix is indicated by the strongly oriented grain coatings, the interwoven nature of most of the sepiolite occurrences suggest that the sepiolite is neoformed. Singer and Norrish (1974) noted neoformed palygorskite cutans in otherwise palygorskite-free soils in Australia. Sample 2D was therefore classified as a ?sepiolitic? (Chapter 1) pedocutanic (Soil Classification Working Group, 1991) or argic (WRB, 1998) horizon. Sample 2D differs from the calcrete samples 2A to 2C only by its lack of calcifica- tion. It contains the same quartz-clay aggregates, the same clay coatings oriented so strongly that it gives the appearance of a pseudo-negative interference figure, and the same limpid yellow clay nodules, some of which show pseudo-negative uniaxial interference figures. There are no signs of calcite leaching (pendants) in 2D. This suggests that the calcrete horizons 2A to 2C formed via the calcification of a pedocutanic (argic) horizon similar to 2D. 191 Dorbank (petroduric) at heuweltjie edge; sample 1.3 This dorbank is similar to others in southern Africa (see, for example, Chapter 1). There is a strong grano-, poro- and circular striated pattern to the matrix ( Figures 6.23(a) to (c)), suggesting the illuviation of clay and silica has been a prominent process in the horizons formation. The quartz grains are coated with hematite, as in the overlying neocutanic (cam- bic) B horizon, the calcrete and the sepiolitic pedocutanic (argic) horizons of the heuweltjie. The matrix does not show the same red colour, and appears to be composed of amorphous Fe (hydr)oxides mixed with amorphous silica and clay. The clay- and/or Fe-enriched coatings appear to have formed before the silica coatings: in one instance, the silica is concentrated on the outside of clay capping a ped (Figures 6.23(d) to (f)), and also silica is generally (but not exclusively) concentrated on the void side of the clay coatings (Figures 6.23(g) to (i)), similar to that described by Litchfield and Mabbutt (1962) for red-brown hardpans in Western Australia. These observations are consistent with Summerfield (1983b, p. 80), who noted that slow movement of silica-saturated pore waters is required for the shift towards supersaturation and subsequent precipitation to take place, and that silica will only precipitate where the presence of other constituents does not favour the formation of other silicates. This may explain the location of the silica on the outer (void) sides of the clay and Fe-oxides (Figures 6.23(d) to (i)). The aggregate in the top-left corner of Figures 6.23(a) to (c) contains silica nodules but their orientation is obscured by Fe. In one case the silica coating a void was discerned to be length-slow (Figure 6.23(i)). Neither limpid clay nodules nor ?spherulites? seem to be present, unlike the calcrete samples 2A to 2C and ?sepiolitic? pedocutanic (argic) sample 2D. This horizon seems to be formed by silicification of an horizon similar to the overlying neocutanic (cambic) B horizon, similar to Figure 6.8(a). While the calcrete horizons 2A to 2C appear to have formed via the calcification of a pedocutanic (argic) horizon similar to 2D, the dorbank at the edge of the heuweltjie does not. It does not contain the ?clay-quartz compound fragments? which occur in both the calcrete and the ?sepiolitic? pedocutanic (argic) 2D, and has a closer resemblance to the overlying neocutanic (cambic) horizon than the pedocutanic (argic) horizon in which the calcrete formed. 192 6.3.5 Heuweltjie formation Organic matter as a source of ions for calcrete formation The calcrete occurs only in the centre of the heuweltjie mounds, a feature which is common in many heuweltjie landscapes such as the Worcester-Robertson area. The concentration of calcite in termite mounds in otherwise calcite-free soils has been noted before: Watson (1962), for example, noted calcareous soil below a non-active termite mound in otherwise calcite-free ferrallitic soils in what is now Zimbabwe. Monger and Gallegos (2000, p. 278 to 280) suggested that termites could be mining carbonate from depth, obtaining the calcite from the plants they consumed, and/or biomineralising calcite internally. Ellis (2002) proposed that the petrocalcic horizon at the centre of the mound formed as the concentration of plant material by termite activity led to the build-up of bases (especially Ca) and silica in the heuweltjies over time. This is consistent with the nutrient cycling into heuweltjie soils noted by authors such as Midgley and Hoffman (1991) and Midgley and Musil (1990). The heuweltjie at Papendorp shows abundant mi- cromorphological evidence of the accumulation of organic matter (plant remains and calcite phytoliths, tunnels lined with organic matter, see Section 6.3.4) that is associated only with the calcrete in the centre of the heuweltjie (samples 2A to 2C). Evidence of organic matter accumulation is absent from both ?sepiolitic? pedocutanic (argic) sample 2D and dorbank sample 1.3 on the periphery. The presence of many calcite phytoliths (which dissolved completely in 2N acetic acid) in the calcrete further argues for the plant matter?calcite association. Data from Midgley and Musil (1990) showed that foliar Ca for plants associated with heuweltjies in the Worcester-Robertson valley ranges from 853.29 mmol/kg (3.42%) for on-mound Ruschia caroli to 256.48 mmol/kg (1.03%) for off-mound Euphorbia burmannii. This is higher than the 0.67 to 1.14% foliar Ca and 0.83 to 1.36% stem Ca in Brunia albiflora found by Poole (1999) in the Western Cape coastal zone (Rooi Els, Kleinmond, Grabouw). It is even twice that of Cornus florida, the calcium ?pump? known for its high foliar Ca levels, which has 1.73% foliar Ca (data from Jenkins et al., 2007; published in Holzmueller et al., 2007). This calcite mineralisation in the centre of heuweltjies in an otherwise non- calcareous landscape bears some resemblance to the calcite mineralisation in Iroko trees (Milicia excelsa) and surrounding soil in the Ivory Coast and Cameroon 193 (Braissant et al., 2004; Cailleau et al., 2005, and references therein). These trees contain significant calcium carbonate accumulations within their tissues and are associated with calcium carbonate in their surrounding soils, even though they grow on non-calcareous orthox soils which normally have a pH of 4.3 to 6.0. Braissant et al. (2004) and Cailleau et al. (2005) showed that the calcium car- bonate mineralisation in the trees and their surrounding soils is caused by soil bacteria, which oxidise the Ca oxalate that is present in the living tissues of M. excelsa, the wood-rot fungi, decaying wood and their surrounding soils. Braissant et al. (2002) studied the growth of bacteria in a medium containing potassium oxalate. They used Ralstonia eutropha (syn: Alcaligenes eutrophus) and Xanthobacter autotrophicus, which are ubiquitous and most easily found in oxalate-rich litters such as those associated with Oxalis, Rumex, Rheum and Eucalyptus (Braissant et al., 2002). Much of the plant material available to the foraging termites in Namaqualand is enriched in oxalate (A. Milewski, pers. comm., 2007), for example ?vygies? (Mesembryanthemum spp.), Oxalis spp and Rumex spp (VetPath Veterinary Pathologists, 2004). Braissant et al. (2002) found that the growth of R. eutropha and X. autotrophicus showed a rapid consumption of oxalate associated with a continuous increase in pH. The final pH after 7 days of incubation was >9.5, from an initial pH of 7, and consistent with the theoretical pH of 9.55?0.05 after total oxalate consumption (Braissant et al., 2002). Castanier et al. (1999, p. 20) noted that ?carbonate precipitation always appears to be the response of heterotrophic bacterial communities to enrichment of the environment in organic matter?. Given the high nitrogen content as a result of the accumulation of organic matter by termites and its subsequent degrada- tion (Midgley and Musil, 1990), the passive carbonatogenesis induced by several metabolic pathways of the nitrogen cycle described by Castanier et al. (1999) may also be relevant to calcite genesis within the heuweltjie. These induce production of carbonate and bicarbonate ions, and as a metabolic end-product, ammonia, which induces pH increase (Castanier et al., 1999). 194 Source of ions for sepiolite/palygorskite formation Chapter 5 showed that most of the measured Mg and Si levels on the Namaqualand coastal plain were sufficient for sepiolite saturation given an appropriate envi- ronmental pH, and concluded that sepiolite precipitation is more likely to be triggered when a solution encounters a pH barrier (such as a calcic horizon) than by the concentration of ions by evaporation. This suggests that after the calcite has precipitated in the centre of the heuweltjie, it buffers the pH sufficiently high allowing for the precipitation of sepiolite. In addition to the coastal proximity of the Papendorp heuweltjie causing high soil Mg levels (Chapter 5), data from Midgley and Musil (1990) showed that foliar Mg is high in plants associated with heuweltjies in the Worcester-Robertson valley: the foliar Mg ranges from 888.52 mmol/kg (2.16%) for on-mound Ruschia caroli to 259.98 mmol/kg (0.63%) for off-mound Euphorbia burmannii. It seems likely therefore that termite concentration of plant material, in addition to raising the Ca-levels, also raises the Mg levels to allow for sepiolite (and palygorskite) precipitation within heuweltjies. Leaching and redistribution of ions within the heuweltjie: zoning of sepiolite and dorbank (petroduric) Ellis (2002) observed that in areas of lower rainfall, CaCO3 is leached to a shallow depth to form the (petro)calcic horizon at the centre of heuweltjie mounds where termite activities have ceased. Pendents in the calcrete sample 2C confirm the remobilisation of calcite within the calcrete horizon (see section 6.3.4). Spackman and Munn (1984) used differential leaching to explain why the upper parts of a profile in a mima-like mound in Wyoming contained more Ca2+, and the lower parts more Mg2+and Na+. The gradual zoning outward from sepiolitic-calcite in the centre through (petro)- sepiolitic to silica-cement at the edges of heuweltjies (Figure 6.24) is very similar to the calcite-palygorskite sequence developed in an alluvial fan sequence in Spain (Rodas et al., 1994). Rodas et al. (1994) observed calcretes (with palygorskite in the clay fraction) in the proximal part of the alluvial fan, and non-calcareous ?palycrete? where palygorskite formed the cementing agent in the distal parts. Silicification was associated with both calcrete and ?palycrete? in the alluvial fan 195 sediments, but was interpreted by Rodas et al. (1994) to be superimposed on previous duricrust levels. 1 Dorbank (petroduric) Calcrete ?Sepiolitic? pedocutanic (argic) Dorbank (petroduric) A: pH 7.6 B: pH 6.9 pH 8.3 B: pH 8.5 A: pH 8.7 B: pH 8.5 ~20 m pH 9.1 Neocutanic/cambic ~1 m Non-calc. crust Calc. crust ?Sepiolitic? pedocutanic (argic) Calcrete Tonguing transition Dorbank (petroduric) Calcrete ?Sepiolitic? pedocutanic (argic) Dorbank (petroduric) A: pH 7.6 B: pH 6.9 pH 8.3 B: pH 8.5 A: pH 8.7 ~20 m pH 9.1 Neocutanic/cambic ~1 m Non-calcareous crust Calcareous crust * *hydrophilic** ?Sepiolitic? pedocutanic (argic) * *hydrophilic** Calcrete Tonguing transition Calcrete Figure 6.24: Distribution of minerals and inferred water movement (dashed ar- rows)through the Papendorp heuweltjie. See text for explanation. Calc.- cal- careous. Modified from (Ellis, 2002). In heuweltjies, Ellis (2002) noted that the highest silica mobility occurred in the well-drained cambic and calcic horizons, and that calcic horizons, with their higher pH, had the highest water-soluble silica concentration. Ellis (2002) sug- gested that because the calcic horizon usually occurs near the centre of the heuweltjie, this is likely to be the zone where most of the silica, together with that in the cambic horizon, is generated after a rain event. The silica then moves vertically and horizontally (Figure 6.24) to the depth where the petroduric hori- zon is located (Ellis, 2002). The position of the silica-cemented horizons on the edge of the heuweltjie is consistent with Ellis and Lambrechts?s (1994) observa- tions that the extensive dorbank horizons in southern Africa usually occur on level to slightly sloping land, and mainly on footslopes. Rodas et al. (1994) suggested that the calcretes (with palygorskite in the clay fraction) in the proximal part of the alluvial fan, and non-calcareous ?palycrete? formed during the movement of alkaline phreatic waters through the fan, as the precipitation of calcite in the proximal part of the fan shifted the composition of the water to a lower Ca/Mg-ratio and lower CO2?3 in the distal parts. This is similar in principle to the effect mineral precipitation has on the evolving com- position of an evaporating brine (for example Hardie and Eugster, 1970; Eugster and Hardie, 1978; Eugster and Jones, 1979). Chapter 5 concluded that the pre- 196 cipitation of calcite and sepiolite on the Namaqualand coastal plain resulted in increasing Ca and Mg and decreasing HCO?3 trends with evaporation, due to the initial Mg2+ >HCO?3 and Ca 2+ >HCO?3 ratios according to the Hardie-Eugster scheme. The movement of the Mg-Si enriched water downslope, coupled with the decrease in HCO?3 and increase in Mg 2+ due to sepiolite precipitation, allows for the precipitation of the ?sepiolitic? zone on the outer side of the calcrete. The presence of Fe, Al and Mg in silica-rich solutions reduces the solubility of silica, but it tends to cause clay authigenesis or the formation of other silicates rather than the precipitation of pure silica (Summerfield, 1983b, p. 79). This is consistent with conditions within the heuweltjie, where the enrichment of cations such as Ca and Mg in the heuweltjie centre caused by termite foraging results in calcite and clay authigenesis near the centre, leaving the precipitation of silica to occur on the periphery. 6.4 Conclusions Biological activity was only evident in the calcrete in the centre of the heuweltjie (samples 2A to 2C), and neither in the ?sepiolitic? pedocutanic (argic; sample 2D) halfway to the edge, nor the dorbank (petroduric; sample 1.3) at the edge. This is consistent observations that the centre of the heuweltjie is where the termite activity is concentrated. The pedogenesis of the hardpans in the heuweltjie is proposed to be as follows: en- richment 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 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 predicted decrease in HCO?3 and increase in Mg 2+ 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. Further work to confirm the extent to which the (a) passive bacterial precipitation 197 of calcium carbonate in the nitrogen cycle and (b) the bacterial decomposition of plant oxalates are likely to contribute to the formation of calcic horizons in the heuweltjie. Oxalates are particularly common in Namaqualand plant species and so this avenue of research may be a fundamental link in the organic matter-calcite relationship in heuweltjies. 198 General summary of findings and recommendations for further work Although some ?sepiocrete? horizons superficially resemble silica-cemented hori- zons in their hardness, they are different to both dorbank/duripan/petroduric and silcrete horizons. The ?sepiocrete? horizons do not meet the slaking require- ments of dorbank/duripan/petroduric horizons. They contain less total SiO2 than either a typical dorbank, or a ?silcrete?. Sepiolite appears to form the matrix areas, whereas the silica is localised. The interlocking, subparallel morphology of sepiolite may contribute to the induration, although the degree to which silica and sepiolite dominate seems to vary even within the same horizon. It seems most probable that both sepiolite and silica contribute to the structural proper- ties of the horizon. The timing and precise mechanism of silica precipitation in the ?sepiocrete? is difficult to resolve. It seems unlikely that silica precipitation was caused by the initial precipitation of sepiolite depleting the solution in Mg, since XRF analysis of the clay fractions showed Mg is present in excess of the amount of silica required to form sepiolite. This is consistent with other data from the Namaqualand coastal plain that show Mg levels to be high, possibly as a result of the marine influence. Etched quartz grains in both ?sepiocrete? profiles suggests the localised dissolution and redistribution of silica. Evidence of calcite mobilisation indicates a complex evolutionary history. It could be that the ?sepiocrete? resulted from the (partial) silicification of what was originally a sepiolite-rich calcic/petrocalcic horizon, but the horizon no longer contains enough calcite to be classified as a petrocalcic. 199 While the ?-crete? terminology provides a useful expression of the cemented nature of the horizon, to fit the existing soil classification schemes the terms ?sepiolitic? and ?petrosepiolitic? (in the same sense as ?calcic? and ?petrocalcic?) would be more appropriate. The terms ?sepiolitic? and ?petrosepiolitic? have the advantage over the ?-crete? terminology that they can be more easily be applied as adjectives to other hardpans where sepiolite is significant but not necessarily cementing, such as ?sepiolitic petrocalcic?. This is particularly relevant since sepiolite is closely associated with calcite in Namaqualand soils, and is a common mineral in the clay fraction of calcretes. The term ?sepiolitic? is appropriate 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. 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 ap- propriate. The ?sepiolitic? criteria distinguish the ?petrosepiolitic? horizon from a ?silcrete?, a silica-cemented horizon which does not fit slaking requirements of a petroduric horizon. Sepiolite has a high water holding capacity. Its high plasticity affects the geotech- nical usefulness of the soil, but many effects of large quantities of sepiolite in a soil still need to be quantified. The high water-holding capacity would directly affect the water-availability to plants growing in sepiolite-rich soils, which in turn affects the ecosystem of arid regions. It also indirectly causes a decrease in the rate of the carbonation process in lime mortars, as a decrease in the free water content in the porous system of the mortar impedes CO2 dissolution, which is a rate-controlling step of the carbonation process. This effect on soils may be similar to the effect of evapotranspiration and CO2 loss on calcite precipitation. The usefulness of a soil survey, therefore, would be increased by indicating the presence of large quantities of sepiolite in a soil using the terms ?sepiolitic? or ?petrosepiolitic?. When sepiolite is present in sufficient quantity, the bulk soil ad- heres to the wetted tongue, and methyl orange turns from orange to purple-pink. In combination, the field tests can be quite diagnostic, and only a few samples need to be verified by XRD. 200 Although the volume of literature on calcrete is now very large, there is very little detailed information on the Namaqualand coastal plain calcretes. Their examination formed only a part of the larger Namaqualand soils investigation, and so there is still a lot of detail that is unresolved. This includes the micromor- phological aspects of sepiolite/palygorskite occurrences in calcrete, particularly their relationship to other clay minerals, calcite, and biogenic features; as well as a full comparison to ?heuweltjie calcrete?. This study has, however, resolved two main points: i) 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, this does not appear to be as volumetrically significant as displacive calcite. Clay minerals are closely associated with the displacing calcite, suggesting that clay neoformation/illuviation may contribute to displacement and folding, but calcite crystallization seems to be mainly re- sponsible for opening fractures. ii) Both biotic and abiotic factors contributed to calcrete formation. Abiotic alpha-fabric seems dominant in mature calcrete horizons, whereas biogenic beta-fabric is dominant in calcareous nodules in a calcic B horizon above calcrete. This study confirmed the prominence of pedogenic sepiolite in the clay fraction of calcareous coastal plain soils. The <0.08 ?m fraction was the only size fraction where palygorskite could be detected before acetate treatment. There was no conclusive evidence for or against the presence of kerolite in the clay fraction. 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 occurred 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 were common on the grain-side of cutans. One TEM image revealed what appeared to be acicular crystals developing from an amorphous phase, although well crystalline sepiolite was also present. The abundance of sepiolite in the calcic soils of the coastal plain (in some horizons it comprises nearly 100% of the clay fraction) makes it ideal for further work on the formation of sepiolite. The sepiolite and calcite saturation indices are only dependent on pH, for the Mg, Si and pH ranges that were encountered on the Namaqualand coastal plain. In contrast, amorphous silica remained saturated regardless of the pH or con- centration. Sepiolite precipitation is therefore more likely to be triggered when 201 a solution encounters a pH barrier than by the concentration of ions by evap- oration. Evidence of the pH control on sepiolite saturation is that sepiolite is commonly associated with calcareous horizons. The effect of a pH change on the solubility of sepiolite is even greater than its effect on calcite solubility. Since sepiolite is often a significant part of arid region soils where calcite and amor- phous silica are common, the mineral solubility-pH relationships may assist in explaining the relationships between calcic, sepiolitic and duric soils and their intergrades. The marine-influenced high Mg coupled with the Hardie-Eugster model of brine evolution provides an explanation for sepiolite-dominance at the coast, and paly- gorskite-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 an 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. Most of the soils for which there was a positive sepiolite identification showed a positive sepiolite saturation index, suggesting that the saturated paste extracts were in equilibrium with sepiolite. This is a direct contrast to previous findings for Namaqualand, and previously published interpretations relating to the alteration/disintegration of sepiolite and palygorskite therefore require revision. The data from the present study showing sepiolite equilibrium in sepiolitic soils suggest that sepiolite can be considered a ?precipitating mineral?, and so forming in the present day. This study developed a model for the formation of the calcrete, ?(petro)sepiolitic?, and petroduric horizons of heuweltjies. Enrichment of cations such as Ca and Mg in the heuweltjie centre caused by termite foraging results in calcite and sepiolite authigenesis. 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 Mg 2+ 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, 202 leaving the precipitation of silica to occur on the periphery. The following mecha- nisms are likely to contribute to the formation of calcic horizons in the heuweltjie but are as yet untested: passive bacterial precipitation of calcium carbonate in the nitrogen cycle; and the bacterial decomposition of plant oxalates. Oxalates are particularly common in Namaqualand plant species and so this avenue of research may be a fundamental link in the organic matter-calcite relationship in 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. Some of the features which deserve further study are regular calcrete layers and fossils (marine, terrestrial) which have the potential for dating; the abundance of pedogenic carbonates lend themselves to techniques that could reveal whether they played a role in atmo- spheric CO2 sequestration; evidence of deep/relict termite activity; dunefields overlying buried and truncated soils such as the Last Glacial dunefield north of the Swartlintjies. Soil formation and termite activity is at least as old as the Last Interglacial, and the identification and classification of the paleosols in greater detail may reveal subtleties in past soil forming factors and possibly offer some constraints on the climatic conditions under which they formed. E horizons may have formed in a wetter Last Interglacial paleoclimate, but they are still active in the present day. 203 References Akbulut, A. and Kadir, S. (2003). The geology and origin of sepiolite, pa- lygorskite and saponite in Neogene lacustrine sediments of the Serinhisar- Acipayam Basin, Denizli, SW Turkey, Clays and Clay Minerals 51 (3): 279? 292. Alexander, E. B. (1985). Estimating relative ages from iron-oxide/total-iron ratios of soils in the western PO Valley, Italy - A discussion, Geoderma 35: 257?259. Allen, B. L. (1985). Micromorphology of Aridisols, in L. A. Douglas and M. L. Thompson (eds), Soil Micromorphology and Soil Classification, Soil Science Society of America, Madison, Wisconsin, pp. 197?216. Alonso-Zarza, A. M. (2003). Palaeoenvironmental significance of palustrine carbonates and calcretes in the geological record, Earth-Science Reviews 60: 261?298. Alonso-Zarza, A. M. and Silva, P. G. (2002). Quaternary laminar calcretes with bee nests: evidences of small-scale climatic fluctuations, Eastern Canary Islands, Spain, Palaeogeography, Palaeoclimatology, Palaeoecology 178: 119? 135. Alvarez, A. (1984). Sepiolite: properties and uses, in A. Singer and E. Galan (eds), Palygorskite-Sepiolite: Occurrences, Genesis and Uses, Developments in Sedimentology 37, Elsevier Science Publishers, Amsterdam, pp. 253?287. Arduino, E. (1985). Estimating relative ages from iron-oxide/total-iron ratios of soils in the western PO Valley, Italy - A reply, Geoderma 35: 257?259. Bain, J. A. (1971). Atterberg limits of clay, Clay Minerals 9: 1?17. 204 Banks, D., Parnachev, V. P., Frengstad, B., Holden, W., Karnachuk, O. V. and Vedernikov, A. A. (2004). The evolution of alkaline, saline ground- and surface waters in the southern Siberian steppes, Applied Geochemistry 19: 1905?1926. Bigham, J. M., Jaynes, W. F. and Allen, B. L. (1980). Pedogenic degradation of sepiolite and palygorskite on the Texas High Plains, Soil Science Society of America Journal 44: 159?167. Blank, R. R. and Fosberg, M. A. (1991). Duripans of the Owyhee Plateau Region of Idaho: genesis of opal and sepiolite, Soil Science 152: 116?133. Blood, J. R. (2006). Monitoring rehabilitation success on Namakwa Sands heavy minerals mining operation, Namaqualand, South Africa, Master?s thesis, University of Stellenbosch. Bockheim, J. G. (2005). Soil endemism and its relation to soil formation theory, Geoderma 129: 109?124. Bockheim, J. G., Kelsey, H. M. and Marshall III, J. G. (1992). Soil develop- ment, relative dating, and correlation of Late Quaternary marine terraces in southwestern Oregon, Quaternary Research 37: 60?74. Boettinger, J. L. and Southard, R. J. (1991). Silica and carbonate sources for Aridisols on a granitic pediment, western Mojave Desert, Soil Science Soci- ety of America Journal 55: 1057?1067. Boggs, S. (1995). Principles of Sedimentology and Stratigraphy, second edn, Prentice Hall, Englewood Cliffs, N.J. Borchardt, G. (1989). Smectites, in J. B. Dixon and S. B. Weed (eds), Minerals in the Soil Environment, second edn, SSSA Book Series, no. 1, Soil Science Society of America, Madison, Wisconsin, pp. 675?727. Bosak, T., Souza-Egipsy, V. and Newman, D. K. (2004). A laboratory model of abiotic peloid formation, Geobiology 2 (3): 189?198. Bouza, P. J., Simo?n, M., Aguilar, J., del Valle, H. and Rostagno, M. (2007). Fibrous-clay mineral formation and soil evolution in Aridisols of northeast- ern Patagonia, Argentina, Geoderma 139: 3850. 205 Bradley, W. F. (1940). The structural scheme of attapulgite, American Miner- alogist 25: 405?410. Braissant, O., Cailleau, G., Aragno, M. and Verrecchia, E. P. (2004). Bacterially induced mineralization in the tree Milicia exelsa (Moraceae): its causes and consequences to the environment, Geobiology 2 (1): 59?66. Braissant, O., Cailleau, G., Dupraz, C. and Verrecchia, E. P. (2003). Bacterially induced mineralization of calcium carbonate in terrestrial environments: the role of exopolysaccharides and amino acids, Journal of Sedimentary Research 73 (3): 485?490. Braissant, O., Verrecchia, E. P. and Aragno, M. (2002). Is the contribution of bacteria to terrestrial carbon budget greatly underestimated?, Naturwis- senschaften 89: 366?370. Brauner, K. and Preisinger, A. (1956). Struktur un enstehung des sepioliths, Tschermaks Mineralogische und Petrographische Mitteilungen 6: 120?140. Brindley, G. W. (1959). X-ray and electron diffraction data for sepiolite, Amer- ican Mineralogist 44: 494?500. Brindley, G. W., Bish, D. L. and Wan, H. M. (1977). The nature of kerolite, its relation to talc and sepiolite, Mineralogical Magazine 41: 443?452. Brochier, J. E., Villa, P., Giacomarra, M. and Tagliacozzo, A. (1992). Shepherds and sediments: geo-ethnoarchaeology of pastoral sites, Journal of Anthro- pological Archaeology 11: 47?102. Buczynski, C. and Chafetz, H. S. (1987). Siliciclastic grain breakage and dis- placement due to carbonate crystal growth: an example from the Lueders Formation (Permian) of north-central Texas, USA, Sedimentology 34: 837? 843. Bu?hmann, C., Nell, J. P. and Samadi, M. (2004). Clay mineral associations formed under Mediterranean-type climate in South Africa, South African Journal of Plant and Soil 21 (3): 166?170. Buol, S. W., Hole, F. D., McCracken, R. J. and Southard, R. J. (1997). Soil Genesis and Classification, fourth edn, Iowa State University Press, USA. 206 Burgers, C. (1975). Heuweltjies (kraaltjies) in die Westelike Provinsie. Unpub- lished report of the Cape Department of Nature Conservation, Cape Town. Bustillo, M. A. and Alonso-Zarza, A. M. (2007). Overlapping of pedogenesis and meteoric diagenesis in distal alluvial and shallow lacustrine deposits in the Madrid Miocene Basin, Spain, Sedimentary Geology . Article in press, doi:10.1016/j.sedgeo.2006.12.006. Bustillo, M. A. and Bustillo, M. (2000). Miocene silcretes in argillaceous playa deposits, Madrid Basin, Spain: petrological and geochemical features, Sed- imentology 47: 1023?1037. Cailleau, G., Braissant, O., Dupraz, C., Aragno, M. and Verrecchia, E. P. (2005). Biologically induced accumulations of CaCO3 in orthox soils of Biga, Ivory Coast, Catena 59 (1): 1?17. Calvet, F. and Julia, R. (1983). Pisoids in caliche profiles of Tarragona (NE Spain), in T. Peryt (ed.), Coated Grains, Springer-Verlag, Berlin, pp. 456? 473. Canti, M. G. (1997). An Investigation of Microscopic Calcareous Spherulites from Herbivore Dungs, Journal of Archaeological Science 24: 219?231. Canti, M. G. (1998). The Micromorphological Identification of Faecal Spherulites from Archaeological and Modern Materials, Journal of Archaeological Sci- ence 25: 435?444. Carrick, P. and Kru?ger, R. (2007). Restoring degraded landscapes in lowland Namaqualand: Lessons from the mining experience and from regional eco- logical dynamics, Journal of Arid Environments 70 (4): 767?781. Castanier, S., Le Me?tayer-Levrel, G. and Perthuisot, J. (1999). Ca-carbonates precipitation and limestone genesis - the microbiogeologist point of view, Sedimentary Geology 126: 9?23. Chadwick, O. A., Hendricks, D. M. and Nettleton, W. D. (1987). Silica in duric soils: I. A depositional model, Soil Science Society of America Journal 51: 975?982. Chadwick, O. A. and Nettleton, W. D. (1990). Micromorphological evidence of adhesive and cohesive forces in soil cementation, in L. A. Douglas (ed.), 207 Soil micromorphology: a basic and applied science, Developments in Soil Science, 19, Elsevier, Amsterdam, pp. 207?212. Chartres, C. J. (1985). A Preliminary investigation of hardpan horizons in north- west New South Wales, Australian Journal of Soil Research 23: 325?337. Christ, C. L., Hathaway, J. C., Hostetler, P. B. and Shepard, A. O. (1969). Palygorskite: new X-ray data, The American Mineralogist 54: 198?205. Christ, C. L., Hostetler, P. B. and Siebert, R. M. (1973). Studies in the sys- tem MgO-SiO2-CO2-H2O (III): The activity-product constant of sepiolite, American Journal of Science 273: 65?83. Coaton, W. G. H. (1981). Fossilised nests of Hodotermitidae (Isoptera) from the Clanwilliam district, Cape Province, Journal of the Entomological Society of South Africa 44 (2): 79?81. Compton, J. S. (2007). Holocene evolution of the Anichab Pan on the south-west coast of Namibia, Sedimentology 54: 55?70. Cowling, R. M., Esler, K. J. and Rundel, P. W. (1999). Namaqualand South Africa - an overview of a unique winter rainfall desert ecosystem, Plant Ecology 142: 3?21. Cowling, R. M. and Hilton-Taylor, C. (1999). Plant biogeography, endemism and diversity, in W. R. J. Dean and S. J. Milton (eds), The Karoo: Ecological Patterns and Processes, Cambridge University Press, Cambridge, England, pp. 42?56. Cowling, R. M. and Pierce, S. (1999). Namaqualand: A Succulent Desert, Fern- wood Press. Cox, G. W., Lovegrove, B. G. and Siegfried, W. R. (1987). The small stone con- tent of Mima-like mounds in the South African Cape Region: implications for mound origin, Catena 14: 165?176. Davies, P. J., Bubela, B. and Ferguson, J. (1978). The formation of ooids, Sedimentology 25: 703?730. D.B.C.M. (2000). Environmental Management Program Report (E.M.P.R.) - Namaqualand Mines. De Beers Consolidated Mines Ltd, Private Bag X01, KLEINZEE, 8282, telephone (+27) 807-1911. 208 De Villiers, S. E. and Cadman, A. (2001). An analysis of the palynomorphs ob- tained from Tertiary sediments at Koingnaas, Namaqualand, South Africa, Journal of African Earth Sciences 33 (1): 17?47. Deer, W. A., Howie, R. A. and Zussman, J. (1992). An Introduction to the Rock- Forming Minerals, second edn, Longman Scientific and Technical, Essex, England, chapter Sheet Silicates, p. 285; 354. De?ka?ny, I., Turi, L. and Nagy, A. F. J. B. (1999). The structure of acid treated sepiolites: small-angle X-ray scattering and multi MAS-NMR investigations, Applied Clay Science 14: 141?160. Deocampo, D. M. (2005). Evaporative evolution of surface waters and the role of aqueous CO2 in magnesium silicate precipitation: Lake Eyasi and Ngoron- goro Crater, northern Tanzania, South African Journal of Geology 108: 493? 504. Desmet, P. G. (2007). NamaqualandA brief overview of the physical and floristic environment, Journal of Arid Environments 70: 570?587. D??az Torres, J. A., An?orbe Urmeneta, M., Sua?rez Barrios, M., Navarrete Lo?pez- Cozar, J. and Mart??n Pozas, J. M. (1990). Presencia y ge?nesis de los min- erales fibrosos de la arcilla en las proximidades de Villamuriel de Cerrato, Provincia de Palencia. (Abstract in English), Stvdia Geologica Salmanticen- sia 26: 7?25. Doner, H. E. and Lynn, W. C. (1989). Carbonate, halide, sulphate, and sul- phide minerals, in J. B. Dixon and S. B. Weed (eds), Minerals in the Soil Environment, second edn, SSSA Book Series, no. 1, Soil Science Society of America, Madison, Wisconsin, pp. 279?330. Drees, L. R., Wilding, L. P., Smeck, N. E. and Senkayi, A. L. (1989). Silica in soils: quartz and disordered silica polymorphs, in J. B. Dixon and S. B. Weed (eds), Minerals in the Soil Environment, second edn, SSSA Book Series, no. 1, Soil Science Society of America, Madison, Wisconsin, pp. 913? 872. Drever, J. I. (1997). The Geochemistry of Natural Waters: Surface and Ground- water Environments, third edn, Prentice Hall, Upper Saddle River, N.J. 209 Ducloux, J., Delhoume, J. P., Petit, S. and Decarreau, A. (1995). Clay differential in Aridisols of northern Mexico, Soil Science Society of America Journal 59: 269?276. Eberl, D. D., Jones, B. F. and Khoury, H. N. (1982). Mixed-layer kero- lite/stevensite from the Armagosa Desert, Nevada, Clays and Clay Minerals 30 (5): 321?326. Ece, O?. I. and C?oban, F. (1994). Geology, occurrence and genesis of Eskis?ehir sepiolites, Turkey, Clays and Clay Minerals 42 (1): 81?92. Eckardt, F. D. and Schemenauer, R. S. (1998). Fog water chemistry in the Namib Desert, Namibia, Atmospheric Environment 32: 2595?2599. Ellis, F. (1988). Die Gronde van die Karoo, PhD thesis, University of Stellen- bosch. Ellis, F. (2002). Contribution of termites to the formation of hardpans in soils of arid and semi-arid regions of South Africa. Proceedings of the 17th World Congress of Soil Science, Bangkok, Thailand, August 2002, ABSTRACT GeoBase-ref. Ellis, F., de Clercq, W. P. and Engelbrecht, H. (2001). Soils associated with microrelief features (?heuweltjies?) occurring on an ancient land surface in the lower Berg River Valley. Paper delivered at the Soil Science Society of South Africa Congress, Pretoria. Ellis, F. and Lambrechts, J. J. N. (1994). Dorbank, a reddish brown hardpan of South Africa: a proto-silcrete? Poster paper presented at the 15th World Congress of Soil Science, Acapulco, Mexico. Ellis, F. and Schloms, B. H. A. (1982). A note on the dorbanks (duripans) of South Africa, Palaeoecology of Africa 15: 149?157. Elprince, A. M., Mashhady, A. S. and Aba-Husayn, M. M. (1979). The occurrence of pedogenic palygorskite (attapulgite) in Saudi Arabia, Soil Science 128 (4): 211?218. Esquivel, E. V., Murr, L. E., Lopez, M. I. and Goodell, P. C. (2005). TEM observations of a 30 million year old mountain leather nanofiber mineral composite, Materials Characterization 54 (4-5): 458?465. 210 Eugster, H. P. and Hardie, L. A. (1978). Saline Lakes, in A. Lerman (ed.), Lakes: Chemistry, Geology, Physics, Springer-Verlag, New York, pp. 237?293. Eugster, H. P. and Jones, B. F. (1979). Behaviour of major solutes during closed- basin brine evolution, American Journal of Science 279: 609?631. Faure, G. (1992). Principles and Applications of Inorganic Geochemistry, Macmillan Publishing Company, New York. Faure, G. (1998). Principles and Applications of Geochemistry, second edn, Prentice Hall, New Jersey, chapter 12: Mineral Stability Diagrams, pp. 172? 199. Ferna?ndez-D??az, L., Astilleros, J. M. and Pina, C. M. (2006). The morphology of calcite crystals grown in a porous medium doped with divalent cations, Chemical Geology 225: 314?321. Fernandez-Marcos, M. L., Mac??as, F. and Guitia?n-Ojea, F. (1979). A contri- bution to the study of the stability of clay minerals from the soil solution composition at different pF values, Clay Minerals 14: 29?37. Fey, M. V., O?Brien, R. D. and Mills, A. J. (2002). Rehabilitation of waste rock and overburden dumps, Encyclopedia of Soil Science, Marcel Dekker, Inc., New York, pp. 1115?1117. Flach, K. W., Nettleton, W. D., Gile, L. H. and Cady, J. G. (1969). Pedocemen- tation: induration by silica, carbonates, and sesquioxides in the Quaternary, Soil Science 107 (6): 442?453. Folk, R. L. (1993). SEM imaging of bacteria and nannobacteria in carbonate sediments and rocks, Journal of Sedimentary Petrology 63 (5): 990?999. Folk, R. L. (1999). Nannobacteria and the precipitation of carbonate in unusual environments, Sedimentary Geology 126: 47?55. Francis, M. L., Fey, M. V. and Ellis, F. (2004). Sepiolite formation in soils of the Namaqualand coastal plain, South Africa, Berichte der Deutschen Mineralogischen Gesellschaft (Beihefte zum (Supplement to the) European Journal of Mineralogy) 16: 40. 211 Francis, M. L., Fey, M. V., Prinsloo, H. P., Ellis, F., Mills, A. J. and Medinski, T. V. (2007). Soils of Namaqualand: compensations for aridity, Journal of Arid Environments 70 (4): 588?603. Gac, J. Y., Droubi, A., Fritz, B. and Tardy, Y. (1988). Geochemical behaviour of silica and magnesium during the evaporation of waters in Chad, Chemical Geology 19: 215?228. Galan, E. and Carretero, M. I. (1999). A new approach to compositional limits for sepiolite and palygorskite, Clays and Clay Minerals 47 (4): 399?409. Galan, E. and Ferrero, A. (1982). Palygorskite-sepiolite clays of Lebrija, southern Spain, Clays and Clay Minerals 30 (3): 191?199. Garrels, R. M. and Mackenzie, F. T. (1967). Origin of the chemical composition of some springs and lakes, in R. F. Gould (ed.), Equilibrium Concepts in Natural Water Systems, Advances in Chemistry Series 67, American Chem- ical Society, Washington, D.C., pp. 222?242. Golden, D. C., Dixon, J. B., Shadfan, H. and Klippenberger, L. A. (1985). Paly- gorskite and sepiolite alteration to smectite under alkaline conditions, Clays and Clay Minerals 33 (1): 44?50. Gonza?lez-Mun?oz, M. T., Ben Chekroun, K., Ben Aboud, A., Arias, J. M. and Rodriguez-Gallego, M. (2000). Bacterially induced Mg-calcite formation: role of Mg2+ in development of crystal morphology , Journal of Sedimentary Research 70 (3): 559?564. Goudie, A. S. (1983). Calcrete, in A. S. Goudie and K. Pye (eds), Chemical Sed- iments and Geomorphology: Precipitates and Residua in the Near-Surface Environment, Academic Press, London, pp. 93?131. Goudie, A. S. (1996). Organic agency in calcrete development, Journal of Arid Environments 32: 103?110. Goza?lez, L., Ibarra, L. M., Rodr??guez, A., Moya, J. S. and Valle, F. J. (1984). Fi- brous silica gel obtained from sepiolite by HCl attack, Clay Minerals 19: 93? 98. Gutie?rrez-Castorena, M. d. C., Stoops, G., Ortiz-Solorio, C. A. and Sa?nchez- Guzma?n, P. (2006). Micromorphology of opaline features in soils on the 212 sediments of the ex-Lago de Texcoco, Me?xico, Geoderma 132 (1-2): 89? 104. Hallmark, C. T., Wilding, L. P. and Smeck, N. E. (1982). Silicon, in A. L. Page, R. H. Miller and D. R. Keeney (eds), Methods of Soil Analysis, Part 2. Chemical and Microbiological Properties, second edn, Agronomy Monograph no. 9, ASA-SSSA, Madison, USA, pp. 263?273. Harden, J. W. (1982). A quantitative index of soil development from field de- scriptions: examples from a chronosequence of soils in central California, Geoderma 28: 1?28. Hardie, L. A. and Eugster, H. P. (1970). The evolution of closed-basin brines, Mineralogical Society of America Special Publication 3: 273?290. Harmse, J. T. and Swanevelder, C. J. (1987). Spatial variation in the granulo- metric properties of Sandveld aeolianites, Annals of the Geological Survey of South Africa 21: 51?58. Hartnady, C. J. H. and Von Veh, M. W. (1990). Tectonostratigraphic and struc- tural history of the Late Proterozoic-Early Palaeozoic Gariep Belt, Cape Province, South Africa, Guidebook Geocongress ?90 Geological Society of South Africa PO1: 1?49. Harvie, C. E., M?ller, N. and Weare, J. H. (1984). The prediction of mineral solu- bilities in natural waters: the Na-K-Mg-Ca-H-Cl-SO4-OH-HCO3-CO3-CO2- H2O system to high ionic strengths at 25o C, Geochimica et Cosmochimica Acta 48: 723?751. Harvie, C. E. and Weare, J. H. (1980). The prediction of mineral solubilities in natural waters: the Na-K-Mg-Ca-Cl-SO4-H2O system from zero to high concentration at 25o C, Geochimica et Cosmochimica Acta 44: 981?997. Hay, R. L., Hughes, R. E., Kyser, T. K., Glass, H. D. and Liu, J. (1995). Magnesium-rich clays of the meerschaum mines in the Amboseli Basin, Tan- zania and Kenya, Clays and Clay Minerals 43 (4): 455?466. Hay, R. L., Pexton, R. E., Teague, T. T. and Kyser, T. K. (1986). Spring-related carbonate rocks, Mg clays and associated minerals in Pliocene deposits of the Armagosa Desert, Nevada and California, Geological Society of America Bulletin 97: 1488?1503. 213 Hay, R. L. and Reeder, R. J. (1978). Calcretes of Olduvai Gorge and the Ndolanya Beds of northern Tanzania, Sedimentology 25: 619?673. Hay, R. L. and Stoessell, R. K. (1984). Sepiolite in the Amboseli Basin of Kenya: A new interpretation, in A. Singer and E. Galan (eds), Palygorskite- Sepiolite: Occurrences, Genesis and Uses, Developments in Sedimentology 37, Elsevier Science Publishers, Amsterdam, pp. V?VI. Hay, R. L. and Wiggins, B. (1980). Pellets, ooids, sepiolite and silica in three calcretes of the southwestern United States, Sedimentology 27: 559?576. Hayashi, H., Otsuka, R. and Imai, N. (1969). Infrared study of sepiolite and palygorskite on heating, American Mineralogist 54: 1613?1624. Heinecken, T. J. E. (1981). Estuaries of the Cape. Part 2: Synopsis of Avail- able Information on Individual Systems. Report no.2: Buffels (CW3), CSIR Research Report 401, Stellenbosch. Hobsan, W. A. and Dahlgren, R. A. (1998). A quantitative study of pedo- genesis in California vernal pool wetlands (Chapter 7), Quantifying Soil Hydromorphy, Soil Science Society of America, Madison, Wisconsin, USA, pp. 107?127. SSSA Special Publication Number 54. Holzmueller, E. J., Jose, S. and Jenkins, M. A. (2007). Relationship between Cornus florida L. and calcium mineralization in two southern Appalachian forest types, Forest Ecology and Management 245: 110?117. Horwath, J. L. and Johnson, D. L. (2006). Mima-type mounds in southwest Missouri: Expressions of point-centered and locally thickened biomantles, Geomorphology 77: 308?319. Horwath, J. L. and Johnson, D. L. (2007). Erratum to ?Mima-type mounds in southwest Missouri: Expressions of point-centered and locally thickened biomantles? [Geomorphology 77 (2006) 308319], Geomorphology 83: 193? 194. Ingle`s, M., Salvany, J. M., Mun?oz, A. and Pe?rez, A. (1998). Relationship of min- eralogy to depositional environments in the non-marine Tertiary mudstones of the southwestern Ebro Basin (Spain), Sedimentary Geology 116: 159?176. 214 Jamoussi, F., Aboud, A. B. and Lo?pez-Galindo, A. (2003). Palygorskite genesis through silicate transformation in Tunisian continental Eocene deposits, Clay Minerals 38: 187?199. Jenkins, M. A., Jose, S. and White, P. S. (2007). Impacts of an exotic dis- ease and vegetation change on foliar calcium cycling in Appalachian forests, Ecological Applications 17: 869?881. Jones, B. F. and Galan, E. (1988). Sepiolite and palygorskite, in S. W. Bailey (ed.), Hydrous Phyllosilicates, Vol. 19 of Reviews in Mineralogy, Mineralog- ical Society of America, pp. 631?674. Jones, B. F., Hanor, J. S. and Evans, W. R. (1994). Sources of dissolved salts in the central Murray Basin, Australia, Chemical Geology 111: 135154. Jungerius, P. D., van den Ancker, J. A. M. and Mu?cher, H. J. (1999). The contribution of termites to the microgranular structure of soils on the Uasin Gishu Plateau, Kenya, Catena 34: 349?363. Kampunzu, A. B., Ringrose, S., Huntsman-Mapila, P., Harris, C., Vink, B. W. and Matheson, W. (2007). Origins and palaeo-environments of Kalahari duricrusts in the Moshaweng dry valleys (Botswana) as detected by major and trace element composition, Journal of African Earth Sciences 48: 199? 221. Kavas, T., Sabah, E. and C?elik, M. S. (2004). Structural properties of sepiolite- reinforced cement composite, Cement and Concrete Research 34: 2135?2139. Keller, W. D. (1982). Kaolin - A most diverse rock in genesis, texture, physical properties, and uses, Geological Society of America Bulletin 93: 27?36. Khademi, H. and Mermut, A. R. (1999). Submicroscopy and stable isotope geochemistry of carbonates and associated palygorskite in Iranian Aridisols, European Journal of Soil Science 50: 207?216. Khoury, H. N., Eberl, D. D. and Jones, B. F. (1982). Origin of magnesium clays from the Armagosa Desert, Nevada, Clays and Clay Minerals 30 (5): 327? 336. Kittrick, J. A. (1971). Soil solution composition and stability of clay minerals, Soil Science Society of America Proceedings 35: 450?454. 215 Klappa, C. F. (1980). Rhizoliths in terrestrial carbonates: classification, recog- nition genesis and significance, Sedimentology 27: 613?629. Knoll, A. H., Strother, P. K. and Ross, S. (1988). Distribution and diagenesis of microfossils from the Lower Proterozoic Duck Creek Dolomite, Western Australia, Precambrian Research 38: 257?279. Knox, G. J. (1977). Caliche profile formation, Saldanha Bay (South Africa), Sedimentology 24: 657?674. Land Type Survey Staff (1987). Land Types of the maps 2816 Alexander Bay; 2818 Warmbad; 2916 Springbok; 2918 Pofadder; 3017 Garies; 3018 Loeries- fontein. Memoirs on the Agricultural Natural Resources of South Africa No. 9. Laurie, H. (2002). Optimal transport in central place foraging, with an applica- tion to the overdispersion of heuweltjies , South African Journal of Science 98: 141?146. Le?veille?, R. J., Fyfe, W. S. and Longstaffe, F. J. (2000). Unusual secondary Ca-Mg-carbonate-kerolite deposits in basaltic caves, Kauai, Hawaii, The Journal of Geology 108): 613?621. Le?veille?, R. J., Longstaffe, F. J. and Fyfe, W. S. (2002). Kerolite in carbonate- rich speleothems and microbial deposits from basaltic caves, Kauai, Hawaii, Clays and Clay Minerals 50 (4): 514?524. Litaor, M. I., Mancinelli, R. and Halfpenny, J. C. (1996). The influence of pocket gophers on the status of nutrients in Alpine soils, Geoderma 70: 37?48. Litchfield, W. H. and Mabbutt, J. A. (1962). Hardpan in soils of semi-arid Western Australia, Journal of Soil Science 13: 148?159. Loisy, C., Verrecchia, E. P. and Dufour, P. (1999). Microbial origin for pedo- genic micrite associated with a carbonate paleosol (Champagne, France), Sedimentary Geology 126: 193?204. Lopez-Galindo, A., Ben Aboud, A., Fenoll Hach-Ali, P. and Casas Ruiz, J. (1996). Mineralogical and geochemical characterization of palygorskite from Gabasa (NE Spain). Evidence from a detrital precursor, Clay Minerals 31: 33?44. 216 Lovegrove, B. G. and Siegfried, W. R. (1986). Distribution and formation of Mima-like earth mounds in the western Cape Province of South Africa, South African Journal of Science 82: 432?436. Lovegrove, B. G. and Siegfried, W. R. (1989). Spacing and origin(s) of Mima-like earth mounds in the Cape Province of South Africa, South African Journal of Science 85: 108?112. Macumber, P. G. (1983). Interactions between ground water and surface systems in northern Victoria, PhD thesis, Melbourne University. Mahood, K. (2003). Strip mining rehabilitation by translocation in arid coastal Namaqualand, South Africa, Master?s thesis, University of Stellenbosch. Martin De Vidales, J. L., Pozo, M., Alia, J. M., Garcia-Navarro, F. and Rull, F. (1991). Kerolite-stevensite mixed-layers from the Madrid Basin, Central Spain, Clay Minerals 26: 329?342. Mart??nez-Ram??rez, S., Puertas, F. and Blanco Varela, M. T. (1995). Carbonation processes and properties of a new lime mortar with added sepiolite, Cement and Concrete Research 25 (1): 39?50. Mart??nez-Ram??rez, S., Puertas, F. and Blanco Varela, M. T. (1996). Stability of sepiolite in neutral and alkaline media at room temperature, Clay Minerals 31: 225?232. McBride, M. B. (1994). Environmental Chemistry of Soils, Oxford University Press. McDaniel, P. A., Bathke, G. R., Buol, S. W., Cassel, D. K. and Falen, A. L. (1992). Secondary Manganese/Iron ratios as pedochemical indicators of field-scale throughflow water movement, Soil Science of America Journal 56: 1211?1217. Mees, F. (1999). Unsuitability of calcite spherulites as indicators of subaerial exposure, Journal of Arid Environments 42: 149?154. Mehra, O. P. and Jackson, M. L. (1960). Iron oxide removal from soils and clays by a dithionite-citrate system buffered with sodium bicarbonate, Proceed- ings of the 7th National Conference on Clays and Clay Minerals, Pergamon Press, New York, pp. 317?327. 217 Mermut, A. R., Arshad, M. A. and St. Arnaud, R. J. (1984). Micropedological study of termite mounds of three species of macrotermes in Kenya, Soil Science of America Journal 48: 613?620. Midgley, G. F. and Musil, C. F. (1990). Substrate effects of zoogenic soil mounds on vegetation composition in the Worcester-Robertson valley, Cape Province, South African Journal of Botany 56 (2): 158?166. Midgley, G., Hannah, L., Roberts, R., MacDonald, D. and Allsopp, J. (2001). Have Pleistocene climate cycles influenced species richness patterns in the greater Cape Mediterranean Region?, Journal of Mediterranean Ecology 2: 137?144. Midgley, G. and Hoffman, T. (1991). Heuweltjies: nutrient factories, Veld and Flora September: 72?75. Midgley, J. J., Harris, C., Hesse, H. and Swift, A. (2002). Heuweltjie age and vegetation change based on ?13C and ?14C analyses, South African Journal of Science 98: 202?204. Mifsud, A., Huertas, F., Barahona, E., Linares, J. and Forne?s, V. (1979). Test de couleur pour la sepiolite, Clay Minerals 14: 247?248. Milton, S. J. and Dean, W. R. J. (1990). Mima-like mounds in the southern and western Cape: are the origins so mysterious?, South African Journal of Science 86: 207?208. Monger, H. C. and Adams, H. P. (1996). Micromorphology of calcite-silica de- posits, Yucca Mountain, Nevada, Soil Science Society of America Journal 60: 519?530. Monger, H. C. and Daugherty, L. A. (1991). Neoformation of palygorskite in a southern New Mexico Aridisol, Soil Science Society of America Journal 55: 1646?1650. Monger, H. C., Daugherty, L. A. and Gile, L. H. (1991a). A microscopic ex- amination of pedogenic calcite in an Aridisol of Southern New Mexico, Soil Science Society of America Special Publication 26: 37?60. Monger, H. C., Daugherty, L. A., Lindemann, W. C. and Liddell, C. M. (1991). Microbial precipitation of pedogenic carbonate, Geology 19: 997?1000. 218 Monger, H. C. and Gallegos, R. A. (2000). Biotic and abiotic processes and rates of pedogenic carbonate accumulation in the southwestern United States - relationship to atmospheric CO2 sequestration, in R. Lal, J. M. Kimble, H. Eswaran and B. A. Stewart (eds), Global Climate Change and Pedogenic Carbonates, Lewis Publishers, Boca Raton, Florida, pp. 273?289. Moody, L. E. and Graham, R. C. (1994). Pedogenic processes in thick sand deposits on a marine terrace, central California, in D. L. Cremeens, R. B. Brown and J. H. Huddleston (eds), Whole Regolith Pedology, Soil Science Society of America, Madison, Wisconsin, USA, pp. 57?74. SSSA Special Publication Number 34. Moore, J. M. and Picker, M. D. (1991). Heuweltjies (earth mounds) in the Clanwilliam district, Cape Province, South Africa: 4000-year-old termite nests, Oecologia 86: 424?432. Mosaddeghi, M. R., Hajabbasi, M. A. and Khademi, H. (2006). Tensile strength of sand, palygorskite and calcium carbonate mixtures and interpretation with the effective stress theory, Geoderma 134: 160?170. Munsell Color Company (1975). Munsell Soil Color Charts. Munsell Color, Baltimore, Maryland. Murphy, J. and Riley, H. P. (1962). A modified single solution method for the determination of phosphate in natural waters, Analytica Chimica Acta 27: 31?36. Nash, D. J. and McLaren, S. J. (2003). Kalahari valley calcretes: their nature, origins, and environmental significance, Quaternary International 111: 3? 22. Nash, D. J. and Shaw, P. A. (1998). Silica and carbonate relationships in silcrete- calcrete intergrade duricrusts from the Kalahari of Botswana and Namibia, Journal of African Earth Sciences 27(1): 11?25. Nathan, Y., Bentor, Y. K. and Wu?rzburger, U. (1970). Vein palygorskites in Isreal and Sinai, their origin and symmetry, Israel Journal of Chemistry 9: 469?476. 219 Ndeinoma, A. (2006). Mycorrhiza re-establishment on post mined rehabilitated areas of the Brand se Baai Succulent Karoo vegetation, Master?s thesis, University of Stellenbosch. Neaman, A. and Singer, A. (2004). The effects of palygorskite on chemical and physico-chemical properties of soils: a review, Geoderma 123: 297303. Netterberg, F. (1969). The Geology and Engineering Properties of South African Calcretes, PhD thesis, University of the Witwatersrand. Nettleton, W. D., Olson, C. G. and Wysocki, D. A. (2000). Paleosol classification: Problems and solutions, Catena 41: 61?92. Nordstrom, D. K., Plummer, L. N., Wigley, T. M. L., Wolery, T. J., Ball, J. W., Jenne, E. A., Bassett, R. L., Crerar, D. A., Florence, T. M., Fritz, B., Hoffman, M., Holdren, G. R. Jr., Lafon, G. M., Mattigod, S. V., McDuff, R. E., Morel, F., Reddy, M. M., Sposito, G. and Thrailkill, J. (1979). A comparison of computerized chemical models for equilibrium calculations in aqueous systems, in E. A. Jenne (ed.), Chemical modeling in aqueous systems, speciation, sorption, solubility, and kinetics, Symposium Series 93, American Chemical Society, Washington, D.C., pp. 857?892. Olivier, J. (2002). Fog-water harvesting along the West Coast of South Africa: A feasibility study, Water SA 28 (4): 349?360. Available on http://www.wrc.org.za. Ouhadi, V. R. and Yong, R. N. (2003). The role of clay fractions of marly soils on their post stabilization failure, Engineering Geology 70: 365?375. Owliaie, H. R., Abtahi, A. and Heck, R. J. (2006). Pedogenesis and clay miner- alogical investigation of soils formed on gypsiferous and calcareous materials, on a transect, southwestern Iran, Geoderma 134 (1-2): 62?81. Parkhurst, D. L. and Appelo, C. A. J. (1999). User?s guide to PHREEQC (Version 2). A computer program for speciation, batch-reaction, one- dimensional transport, and inverse geochemical calculations: U.S. Geologi- cal Survey Water-Resources Investigations Report 99-4259, Technical report. http://water.usgs.gov/software/. 220 Parkington, J., Baxter, A., Cartwright, C., Cowling, R. M. and Meadows, M. (2000). Palaeovegetation at the last glacial maximum in the western Cape, South Africa: wood charcoal and pollen evidence from Elands Bay Cave, South African Journal of Science 96: 543?546. Partridge, T. C. and Maud, R. R. (1987). Geomorphic evolution of southern Africa since the Mesozoic, South African Journal of Geology 90 (2): 179? 208. Partridge, T. C. and Maud, R. R. (2000). Macro-scale geomorphic evolution of southern Africa, in T. C. Partridge and R. R. Maud (eds), The Cenozoic of Southern Africa, Oxford University Press, pp. 3?18. Pether, J. (1994). The Sedimentology, Palaeontology and Stratigraphy of Coastal- Plain Deposits at Hondeklip Bay, Namaqualand, South Africa, Master?s the- sis, University of Cape Town. Pether, J., Roberts, D. L. and Ward, J. D. (2000). Deposits of the West Coast, in T. C. Partridge and R. R. Maud (eds), The Cenozoic of Southern Africa, Oxford University Press, pp. 33?54. Phillips, S. E., Milnes, A. R. and Foster, R. C. (1987). Calcified filaments: an example of biological influences in the formation of calcrete in South Australia, Australian Journal of Soil Research 25: 405?428. Phillips, W. R. and Griffen, D. T. (1981). Optical Mineralogy: The Nonopaque Minerals, W H Freeman & Co., San Francisco. Picker, M. D., Hoffman, M. T. and Leverton, B. (2007). Density of Microhodoter- mes viator (Hodotermitidae) mounds in southern Africa in relation to rain- fall and vegetative productivity gradients, Journal of Zoology 271: 37?44. Pickford, M. and Senut, B. (1997). Cainozoic mammals from coastal Namaqualand, South Africa, Palaeontologia Africana 34: 199?217. Pickford, M. and Senut, B. (2000). Geology and Palaeobiology of the Namib Desert, Southwestern Africa: Geology and History of Study, Vol. 1 of Mem- oir 18, Geological Survey of Namibia. Pitzer, K. S. (1973). Thermodynamics of electrolytes, I. Theoretical basis and general equations, Journal of Phsyical Chemistry 77 (2): 268?277. 221 Pitzer, K. S. (1979). Theory: ion interaction approach, in R. M. Pytkowitcz (ed.), Activity Coefficients in Electrolyte Solutions, Vol. 1, CRC Press, Inc., Boca Raton, Florida, pp. 157?208. Plummer, L. N., Parkhurst, D. L., Fleming, G. W. and Dunkle, S. A. (1988). A computer program incorporating Pitzer?s equations for calculation of geochemical reactions in brines: U.S. Geological Sur- vey Water-Resources Investigations Report 88-4153, Technical report. http://water.usgs.gov/software/. Poole, C. J. (1999). Growth and Nutrition of Bruniaceae, Master?s thesis, Uni- versity of Stellenbosch. Post, J. L. (1985). Identification of calcium oxalate crystals using alizarin red S stain, Archives of Pathology and Laboratory Medicine 109 (2): 186?189. Pozo, M. and Casas, J. (1999). Origin of kerolite and associated Mg clays in palustrine-lacustrine environments. The Esquivias deposit (Neogene Madrid Basin, Spain), Clay Minerals 34: 395?418. Prinsloo, H. (2005). Alteration of the soil mantle by strip mining in the Namaqualand Strandveld, Master?s thesis, University of Stellenbosch. Ranson, M. D. and Bidwell, O. W. (1990). Clay movement and carbonate ac- cumulation in Ustolls of central Kansas, USA, in L. A. Douglas (ed.), Soil micromorphology: a basic and applied science, Developments in Soil Science, 19, Elsevier, Amsterdam, pp. 417?423. Raz, S., Weiner, S. and Addadi, L. (2000). Formation of high-magnesian cal- cites via an amorphous precursor phase: possible biological implications, Advanced Materials 12 (1): 38?42. Reheis, M. C., Sowers, J. M., Taylor, E. M., McFadden, L. D. and Harden, J. W. (1992). Morphology and genesis of carbonate soils on the Kyle Canyon fan, Nevada, U.S.A., Geoderma 52: 303?342. Reider, R. G., Huss, J. M. and Miller, T. W. (1996). A groundwater vortex hypothesis for mima-like mounds, Laramie Basin, Wyoming, Geomorphology 16: 295?317. 222 Lobry de Bruyn, L. A. and Conacher, A. J. (1990). The role of termites and ants in soil modification: a review, Australian Journal of Soil Research 28: 55?93. Rodas, M., Luque, F. J., Mas, R. and Garzon, M. G. (1994). Calcretes, paly- cretes and silcrete in the paleogene detrital sediments of the Duero and Tajo Basins, Central Spain, Clay Minerals 29: 273?285. Rode, A. A. (1969). Theory of Soil Moisture Properties of Soils and Movement of Soil Moisture, IPST, Jerusalem. Translated from the Russian 1965 edition. Rogers, J., Pether, J., Molyneux, R., Hill, R. S., Kilham, J. L. C., Cooper, G. and Corbett, I. B. (1990). Cenozoic geology and mineral deposits along the west coast of South Africa and the Sperrgebiet, Guidebook Geocongress ?90 Geological Society of South Africa PR1: 1?111. Roux, A. L. and Odendaal, F. J. (1992). Report on the study of natural recovery of overburden dumps on D.B.C.M. properties in Namaqualand, Technical report. Unpublished report, Geology Department, De Beers Consolidated Mines Ltd, Private Bag X01, KLEINZEE, 8282, telephone (+27) 807-1911. Sa?nchez, C. and Gala?n, E. (1995). An approach to the genesis of palygorskite in a Neogene-Quaternary continental basin using principle factor analysis, Clay Minerals 30: 225?238. Sanz, M. E. (1996). Sedimentolog??a de las Formaciones Neo?genas del Sur de la Cuenca de Madrid, Minesterio de Fomento, Cedex Madrid. Sariiz, K. and Isik, I. (1995). Meerschaum from Eskis?ehir Province, Turkey, Gems & Gemology 31 (1): 42?51. Sauer, D., Glatzel, S., Ku?hn, J., Zarei, M. and Stahr, K. (submitted). Micro- morphology of duricrusts in soils of the Alentejo (Portugal), . Sauer, D. and Stahr, K. (2004). Formation of calcretes and silcretes in the Alentejo (Portugal) and consequences for the landscape, ecosystems and land use, Eurosoil Proceedings, University of Freiburg. Schmidt, A. (2002). Strip-mine rehabilitation in Namaqualand, Master?s thesis, University of Stellenbosch. 223 Schmiedel, U. and Ju?rgens, N. (1999). Community structure on unusual habitat islands: quartz-fields in the Succulent Karoo, South Africa, Plant Ecology 142: 57?69. Schwertmann, U. and Taylor, R. M. (1989). Iron Oxides, in J. B. Dixon and S. B. Weed (eds), Minerals in the Soil Environment, second edn, SSSA Book Series, no. 1, Soil Science Society of America, Madison, Wisconsin, pp. 379?438. Schwertmann, U. (1993). Relationship between iron oxides, soil color and soil formation, in J. M. Bigham and E. J. Coilkosz (eds), Soil Color, Soil Science Society of America, Madison, Wisconsin, USA, pp. 51?69. SSSA Special Publication Number 31. Scott, D. F., du Toit, B., Johnston, M. A. and Burns, M. E. R. (1995). Soil leaching trials on overburden soil from Kleinzee and Koingnaas, Technical report. Report FOR-C 283, CSIR, Division of Forest Science and Tech- nology, prepared for De Beers Consolidated Mines Ltd, Private Bag X01, KLEINZEE, 8282, telephone (+27) 807-1911. Scott, D. F. and Johnston, M. A. (1994). The effects of mining on soil properties and the ability of the soil to support vegetation after mining at Kleinzee and Koingnaas, on the west coast of South Africa., Technical report. Report FOR-C 253, CSIR, Division of Forest Science and Technology, prepared for De Beers Consolidated Mines Ltd, Private Bag X01, KLEINZEE, 8282, telephone (+27) 807-1911. Seong-Joo, L. and Golubic, S. (1999). Microfossil populations in the context of synsedimentary micrite deposition and acicular carbonate precipitation: Mesoproterozoic Gaoyuzhuang Formation, China, Precambrian Research 96: 183?208. Siesser, W. G. (1973). Diagenetically formed ooids and intraclasts in South African calcretes, Sedimentology 20 (4): 539?551. Siffert, B. and Wey, R. (1962). Synthese d?une sepiolite a temperature ordinaire, Comptes rendus de l?Acade?mie des sciences 254: 1460?1462. Singer, A. (1989). Palygorskite and sepiolite group minerals, in J. B. Dixon and S. B. Weed (eds), Minerals in the Soil Environment, second edn, SSSA 224 Book Series, no. 1, Soil Science Society of America, Madison, Wisconsin, pp. 829?872. Singer, A. (2002). Palygorskite and sepiolite, in J. B. Dixon and D. G. Schulze (eds), Soil Mineralogy with Environmental Applications, third edn, SSSA Book Series, no. 7, Soil Science Society of America, Madison, Wisconsin, pp. 555?583. Singer, A. and Galan, E. (1984). Palygorskite-Sepiolite: Occurrences, Genesis and Uses, Developments in Sedimentology 37, Elsevier Science Publishers, Amsterdam. Singer, A., Ganor, E., Dultz, S. and Fischer, W. (2003). Dust deposition over the Dead Dea, Journal of Arid Environments 53: 41?59. Singer, A., Kirsten, W. and Bu?hmann, C. (1992). Occurrence of sepiolite in the northern Transvaal, South Africa, South Africa Journal of Geology 95: 165? 170. Singer, A., Kirsten, W. and Bu?hmann, C. (1995). Fibrous clay minerals in the soils of Namaqualand, South Africa: characteristics and formation, Geo- derma 66: 43?70. Singer, A. and Norrish, K. (1974). Pedogenic palygorskite occurrences in Aus- tralia, American Mineralogist 59: 508?517. Singer, A., Stahr, K. and Zarei, M. (1998). Characteristics and origin of sepiolite (Meerschaum) from Central Somalia, Clay Minerals 33: 349?362. Slabber, M. H. (1945). ?n Grondopname in die Malmesbury-Piketberg streek, PhD thesis, University of Stellenbosch (D.Sc. (Agric.)). Smith, M. and Compton, J. S. (2004). Origin and evolution of major salts in the Darling pans, Western Cape, South Africa, Applied Geochemistry 19: 645? 664. Soil Classification Working Group (1991). Soil Classification: A Taxonomic Sys- tem for South Africa, second edn, Department of Agricultural Development, Republic of South Africa. Memoirs on the Agricultural Natural Resources of South Africa, Vol. 15. 225 Soil Survey Staff (1999). Soil Taxonomy: a basic system of classification for mak- ing and interpreting soil surveys, second edn, Natural Resources Conserva- tion Service, United States Department of Agriculture. Agric. Handbook No. 436. Spackman, L. K. and Munn, L. C. (1984). Genesis and morphology of soils asso- ciated with formation of Laramie Basin (Mima-like) mounds in Wyoming, Soil Science Society of America Journal 48: 1384?1392. Stahr, K., Ku?hn, J., Trommler, J., Papenfu?, K.-H., Zarei, M. and Singer, A. (2000). Palygorskite-cemented crusts (palycretes) in Southern Portugal, Australian Journal of Soil Research 38: 169?188. Stoessell, R. K. (1988). 25o C and 1 atm dissolution experiments of sepiolite and kerolite, Geochimica et Cosmochimica Acta 52: 365?374. Stoessell, R. K. and Hay, R. L. (1978). The geochemical origin of sepiolite and kerolite at Amboseli, Kenya, Contributions to Mineralogy and Petrology 65: 255?267. Stone, E. L. and Comerford, N. B. (1994). Plant and animal activity below the solum, in D. L. Cremeens, R. B. Brown and J. H. Huddleston (eds), Whole Regolith Pedology, Soil Science Society of America, Madison, Wisconsin, USA, pp. 57?74. SSSA Special Publication Number 34. Stoops, G. (1964). Application of some pedological methods to the analysis of termite mounds, in A. Bouillon (ed.), Etude sur les Termites africains, Lovanium, Kinshasa, pp. 379?398. Stoops, G. (2003). Guidelines for the Analysis and Description of Soil and Re- golith Thin Sections, Soil Science Society of America, Madison, Wisconsin, USA. Suarez, M., Robert, M., Elsass, F. and Martin Pozas, J. M. (1994). Evidence of a precursor in the neoformation of palygorskite - new data by analytical electron microscopy, Clay Minerals 29: 255?264. Sullivan, L. (1994). Clay coating formation on impermeable materials: deposition by suspension retention, in A. J. Ringrose-Voase and G. S. Humphreys (eds), Soil Micromorphology: Studies in Management and Genesis, Developments 226 in Soil Science 22, Elsevier, Amsterdam, pp. 373?380. Proceedings of the IX International Working Meeting on Soil Micromorphology, Townsville, Australia, July 1992. Summerfield, M. A. (1983a). Petrography and diagenesis of silcrete from the Kalahari Basin and Cape Coastal Zone, southern Africa, Journal of Sedi- mentary Petrology 53: 895?909. Summerfield, M. A. (1983b). Silcrete, in A. S. Goudie and K. Pye (eds), Chem- ical Sediments and Geomorphology: Precipitates and Residua in the Near- Surface Environment, Academic Press, London, pp. 59?91. Summerfield, M. A. (1983c). Silcrete as a palaeoclimatic indicator: evidence from southern Africa, Palaeogeography, Palaeoclimatology, Palaeoecology 41: 65? 79. Tankard, A. J. (1966). The Namaqualand Coastal Deposits with special reference to the area between the Groen and Buffels rivers. B.Sc Project, University of Natal. Tankard, A. J. (1975). The late Cenozoic History and Palaeoenvironments of the coastal margin of the southwestern Cape Province, PhD thesis, Rhodes University, Grahamstown. Tankard, A. J. and Rogers, J. (1978). Late Cenozoic palaeoenvironments on the west coast of southern Africa, Journal of Biogeography 5: 319?337. Ten Cate, H. (1966). Die gronde van die Overhex-Nuy gebied naby Worcester, Master?s thesis, University of Stellenbosch (M.Sc. (Agric)). Thellier, C., Fritz, B., Paquet, H., Gac, J. Y. and Clauer, N. (1988). Chemical and mineralogical effects of saline water movement through a soil during evaporation, Soil Science 146 (1): 22?29. Timpson, M. E., Richardson, J. L., Keller, L. P. and McCarthy, G. J. (1986). Evaporite mineralogy associated with saline seeps in southwestern North Dakota, Soil Science Society of America Journal 50: 490?493. Tinely, K. L. (1985). Coastal Dunes of South Africa. SA Nat. Scientific Pro- grammes Report no.109. FRD Council for Scientific and Industrial Research, Pretoria. 227 Torres-Ru??z, J., Lo?pez-Galindo, A., Gonza?lez-Lo?pez, J. M. and Delgado, A. (1994). Geochemistry of Spanish sepiolite-palygorskite deposits: genetic consideration based on trace elements and isotopes, Chemical Geology 112: 221?245. Tracy, S. L., Franc?ois, C. J. P. and Jennings, H. M. (1998a). The growth of calcite spherulites from solution I. Experimental design techniques, Journal of Crystal Growth 193: 374?381. Tracy, S. L., Franc?ois, C. J. P. and Jennings, H. M. (1998b). The growth of calcite spherulites from solution II. Kinetics of formation, Journal of Crystal Growth 193: 382?388. Usai, M. R. and Dalrymple, J. B. (2003). Characteristics of silica-rich pedofea- tures in a buried paleosol, Catena 54: 557?571. van der Eyk, J. J., Macvicar, C. N. and de Villiers, J. M. (1969). Soils of the Tugela Basin, Natal Town and Regional Planning Reports, Volume 15, Town and Regional Planning Commission, Natal. Van der Merwe, C. R. (1940). Soil groups and sub-groups of South Africa. Science Bulletin of the Department of Agriculture of South Africa no. 231/Chem- istry series of the Dept. of Agriculture and Forestry no. 165; Originally presented as the author?s thesis (D. Sc. University of Stellenbosch, 1940). Van Houten, F. B. (1973). Origin of red beds: a review 1961-1972, Annual Review of Earth and Planetary Sciences 1: 39?61. Van Scoyoc, G. E., Serna, C. J. and Ahlrichs, J. L. (1979). Structural changes in palygorskite during dehydration and dehydroxylation, America Mineralogist 64: 215?223. Van Zinderen Bakker, E. M. (1976). The evolution of late-Cenozoic paleoclimates of southern Africa, Palaeoecology of Africa 9: 160?202. Vanden Heuvel, R. C. (1964). The occurrence of sepiolite and attapulgite in the calcareous zone of a soil near Las Cruces, New Mexico, Thirteenth National Conference on Clays and Clay Minerals (Clays and Clay Minerals) 13): 193? 207. 228 Verrecchia, E. P., Dumont, J. and Verrecchia, K. E. (1993). Role of calcium oxalate biomineralisation by funghi in the formation of calcretes: a case study from Nazareth, Israel, Journal of Sedimentary Petrology 63 (5): 1000? 1006. Verrecchia, E. P., Freytet, P., Verrecchia, K. E. and Dumont, J. (1995). Spherulites in calcrete laminar crusts: biogenic CaCO3 precipitation as a major contributor to crust formation, Journal of Sedimentary Research A65 (4): 690?700. Verrecchia, E. P. and Le Coustumer, M. (1996). Occurrence and genesis of palygorskite and associated clay minerals in a Pleistocene calcrete complex, Sde Boqer, Negev Desert, Israel, Clay Minerals 31: 183?202. Verrecchia, E. P. and Verrecchia, K. E. (1994). Needle-fiber calcite: a critical review and a proposed classification, Journal of Sedimentary Research A64 (3): 650?664. VetPath Veterinary Pathologists (2004). Nephrotoxins in Cattle in Southern Africa. Available at http:// www.vetpath.co.za/ large 16 nephrotoxins in cattle. htm. Email: info@vetpath.co.za; modified 2004, last accessed 2007. Von Damm, K. L. and Edmond, J. M. (1964). The occurrence of sepiolite and attapulgite in the calcareous zone of a soil near Las Cruces, New Mexico, Thirteenth National Conference on Clays and Clay Minerals (Clays and Clay Minerals) 13): 193?207. Walker, T. R. (1967). Formation of red beds in modern and ancient deserts, Geological Society of America Bulletin 78: 353?368. Walker, T. R. and Honea, R. M. (1969). Iron content of modern deposits in the Sonoran Desert: a contribution to the origin of red beds, Geological Society of America Bulletin 80: 535?544. Wang, Y., Nahon, D. and Merino, E. (1993). Geochemistry and dynamics of calcrete genesis in semi-arid regions, Chemical Geology 107: 349?351. Washburn, A. L. (1988). Mima mounds: an evaluation of proposed origins with special reference to the Puget Lowland, Technical report. Washington Divi- sion of Geology and Earth Resources, Report of Investigations, vol. 29. 229 Watkeys, M. K. (1999). Soils of the arid south-western zone of Africa, in W. R. J. Dean and S. J. Milton (eds), The Karoo: ecological patterns and processes, Cambridge University Press, Cambridge, pp. 17?26. Watson, J. P. (1962). The soil below a termite mound, Journal of Soil Science 13 (1): 46?51. Watts, N. L. (1977). Pseudo-anticlines and other structures in some calcretes of Botswana and South Africa, Earth Surface Processes 2: 63?74. Watts, N. L. (1978). Displacive calcite: Evidence from recent and ancient cal- cretes, Geology 6: 699?703. Watts, N. L. (1980). Quaternary pedogenic calcretes from the Kalahari (southern Africa): mineralogy, genesis and diagenesis, Sedimentology 27: 661?686. Weather Bureau (1988). Climate of South Africa: Climate Statistics up to 1984, Technical report. Department of Environment Affairs, Pretoria. Weaver, C. E. and Beck, K. C. (1977). Miocene of the S.E. United States: a model for chemical sedimentation in a peri-marine environment, Sedimen- tary Geology 17: 1?234. Weaver, R. M., Syers, J. K. and Jackson, M. L. (1968). Determination of silica in citrate-bicarbonate-dithionite extracts of soils, Soil Science Society of America Proceedings 32: 497?501. White, J. (1979). Introduction to the Principles and Practice of Soil Science, Blackwell, Oxford. Wieder, M. and Yaalon, D. H. (1974). Effect of matrix composition on carbonate nodule crystallization, Geoderma 11: 95?121. Wieder, M. and Yaalon, D. H. (1982). Micromorphological fabrics and devel- opmental stages of carbonate nodular forms related to soil characteristics, Geoderma 28: 203?220. Wollast, R., Mackenzie, F. T. and Bricker, O. P. (1968). Experimental precip- itation and genesis of sepiolite at earth surface conditions, The American Mineralogist 53: 1645?1662. 230 World Reference Base for Soil Resources (WRB) (1998). World Soil Resources Reports no 84, Food and Agricultural Organization of the United Nations (FAO), Rome. Wright, V. P. and Tucker, M. E. (1991). Calcretes: an introduction, in V. P. Wright and M. E. Tucker (eds), Calcretes, Reprint Series Volume 2 of the International Association of Sedimentologists, Blackwell Scientific Publica- tions, Oxford, pp. 1?22. Wright, V. P., Turner, M. S., Andrews, J. E. and Spiro, B. (1993). Morphology and significance of super-mature calcretes from the Upper Old Red Sand- stone of Scotland, Journal of Geological Society of London 150: 871?883. Yaalon, D. H. (1981). On the aeolianite-red sands relationship in coastal Natal, Proceedings of the 6th Conference of S.A.S.Q.U.A. (The Southern African Society for Quaternary Research), Balkema, Pretoria, pp. 145?148. Yaalon, D. H. and Wieder, M. (1976). Pedogenic palygorskite in some arid (calciorthid) soils of Israel, Clay Minerals 11: 73?80. Yalc?in, H. and Bozkaya, O?. (1995). Sepiolite-palygorskite from the Hekimhan region (Turkey), Clays and Clay Minerals 43 (6): 705?717. Zaaboub, N., Abdeljaouad, S. and Lpez-Galindo, A. (2005). Origin of fibrous clays in Tunisian Paleogene continental deposits, Journal of African Earth Sciences 43 (5): 491?504. 231 Appendix A Photographs of soil profiles: BMC and KNC Soil profiles taken in the Buffels Marine Complex (BMC) in the Kleinzee-Port Nolloth area, and the Koingnaas Complex (KNC) further south in the Koingnaas- Hondeklip Bay area. Photographs of the Knersvlakte and Strandfontein heuweltjie are in the text. 232 Lower Shoreface (C3) Upper Shoreface Hardpan Carbonate 6869 Regic Sand (calcareous) sample 6869/i Regic Sand (calcareous) sample 6869/ii Neocarbonate sample 6869/iii Neocarbonate sample 6869/iv Hardpan Carbonate (calcretized upper shoreface) sample 6869/5 Stoneline Yellow-Brown Apedal sample 6869/4 undisturbed Mechanically truncated Upper Shoreface Marine Deposit sample 6869/6 Upper Shoreface Marine Deposit ? 6m ?100cm Marine Deposit Lower Shoreface (C3) Fine sample 6869/7 Marine Deposit Lower Shoreface (C3) Fine Marine Deposit Lower Shoreface (C3) Fine + Iron Pan sample 6869/7a ? 9m Marine Deposit Lower Shoreface (C3) Coarse sample 6869/8 White Clay, slightly calcareous sample 6869/8a sample 6869/13 Lower Shoreface (C4) sample 6869/11 Iron pan sample 6869/12?12m Lower Shoreface (C4) + Termite Burrow sample 6869/10 ?18m ?36m Pit overview ?9m ?6m Channel clay sample 6869/cc Lower Shoreface (C4) Mine Heaps Regic Sand (calcareous) sample 6869/1 Neocarbonate w mottles sample 6869/2;3 KNC: Profile 6869 233 AK1 0cm ?60cm Calcareous Biological Crust sampleAK1/trig Orthic A sampleAK2/trig Neocutanic B (occasionally calcareous) sampleAK3/trig Bedrock Dorbank sampleAK1/8 Dorbank sampleAK1/9 Dorbank (locally calcareous) sampleAK1/10 Roots Dorbank sampleAK1/11 Neocarbonate sampleAK1/1 Mechanically disturbed Neocarbonate B sampleAK1/1 Mechanically disturbed Gypsum/carbonate? ?30cm ?90cm ?30cm Calcareous Dorbank sampleAK1/2 ?90cm Dorbank (v slightly calcareous) sampleAK1/3 Dorbank (gypsum spots) sampleAK1/5 Dorbank (occasionally calcareous) sampleAK1/4 ?120cm Nest Strongly cemented, calcareous Calcareous Dorbank (gypsum spots) sampleAK1/6 TRIG beacon MINE MINE ?1500cm BMC: Profile AK1 234 AK61H 0cm 35cm ?35cm Orthic A (calcareous) sample AK61H/4 ?105cm ?145cm ?210cm Hardpan/Soft Carbonate sample AK61H/1 Blocky Calcretized Sediment sample AK61H/2 Marine deposit (salt efflorescence) Sample AK61H/3 Biological Crust BMC: Profile AK61H 235 DL88 non-calcareous Bleached calcareous A sample DL88/1 Neocarbonate B sample DL88/2 Dorbank sampleDL88/3 Surface Biological Crust BMC: Profile DL88 236 KV196T Orthic A sample KV196T/A tunnels Red Apedal B sample KV196T/B ?20cm ?100cmRed Apedal B sample KV196T/1 ?145cmDorbank sample KV196T/2 Dorbank sample KV196T/3 Calcareous Dorbank sample KV196T/4 Hardpan/Soft Carbonate sample KV196T/5 ?200cm Soft Dorbank sample KV196T/6 ( ?KV196T/7?) ?300cm Soft Dorbank sample KV196T/7 ( ?KV196T/6?) Dorbank +gypsum sample KV196T/8 ?450cm ?600cm Dorbank sample KV196T/9 Marine Deposit sample KV196T/10 Gypsum BMC: Profile KV196T 237 K V220PRN Regic Sand (calcareous) sample KV220PRN/1 organic matter ?80cm 0cm Regic Sand (calcareous ) sample KV220PRN/2 Neocarbonate sample KV220PRN/3 Neocarbonate sample KV220PRN/4 ?120cm Hardpan Carbonate sample KV220PRN/5 Neocarbonate, abundant bioturbation (casts, nest) sample KV220PRN/6 nest Neocarbonate sample KV220PRN/7( ?KV220PRN/9?) ?150cm ?230cm burrows calcretized burrows burrows ?280cm ?380cm nest (occasionally calcareous); 40cm diam sample KV220PRN/8 Hardpan Carbonate to Calcretized Blocky Sediment sample KV220PRN/10 Neocarbonate sample KV220PRN/9( ?KV220/7?) ?680cm ?980cm Marine Deposit sample KV220PRN/11 Marine Deposit sample KV220PRN/12 Marine Deposit sample KV220PRN/13 stoneline BMC: Profile KV220PRN 238 Upper Shoreface Marine Deposit sample LKC1-5/7 LKC1-5 ?100cm ?0cm ?190cm Regic Sand sample LKC1-5/A Regic Sand sample LKC1-5/B Neocutanic B sample LKC1-5/1 Neocutanic, Luvic sample LKC1-5/3 ?130cm ?130cm Gleyed sand sample LKC1-5/4 Sedimentary sample LKC1-5/5 Sedimentary sample LKC1-5/6 ?220cm Pit overview ?240cm ?540cm ?10m Lower Shoreface Marine Deposit sample LKC1-5/8 sample LKC1-5/9 ?240cm KNC: Profile LKC1-5 239 OBT ?50cm ?70cm Regic Sand (deflation) (calcareous) sample OBT/1 Orthic A (calcareous) sample OBT/2 Neocarbonate B sample OBT/3 Hardpan Carbonate ?150cm Hardpan Carbonate to Blocky Calcretized Sediment sample OBT/4 (sepiolite cutan) nest (slightly calcareous ) sample OBT/7 Blocky Calcretized Sediment sample OBT/5 Sedimentary sample OBT/6 disturbed Sepiolite cutan (pinkish) BMC: Profile OBT 240 SK11 Orthic A sample SK11/A Yellow-Brown Apedal B sample SK11/B Gleyed Sand sample SK11/C Neocutanic sample SK11/D Sepiocrete sample SK11/E Sedimentary (locally calcareous) sample SK11/7 Sepiocrete sample SK11/1 Blocky Calcretized Sediment sample SK11/6;2 Hardpan Carbonate sample SK11/3 ( ?SK11/9?)nest (locally calcareous ) (SK11/5) Neocarbonate sample SK11/4a Neocarbonate sample SK11/4b Iron pan sample SK11/10 Hardpan Carbonate sample SK11/9 ( ?SK11/3?) Channel and/or bioturbation sample SK11/8 Sedimentary sample SK11/11 Sedimentary sample SK11/12 ?70cm ?2.7m ?1.7m ?2.7m ?1.7m ?70cm disturbed 1 cm BMC: Profile SK11 241 Slightly Gleyed Sand? sample SL4-1/5 Gleyed Sand sample SL4-1/3 Neocutanic sample SL4-1/4 Yellow-Brown Apedal B sample SL4-1/2 SL4-1 Orthic A sample SL4-1/1 Yellow-Brown Apedal B sample SL4-1/2 Gleyed Sand sample SL4-1/3 Neocutanic sample SL4-1/4 ?65cm ?100cmPedocutanic sample SL4-1/6 Pedocutanic +Carbonate sample SL4-1/7 Pedocutanic sample SL4-1/8 Marine Deposit Upper Shoreface sample SL4-1/9 ?245cm ?9m ?12m Marine Deposit Upper Shoreface Marine Deposit Lower Shoreface (calcareous) sample SL4-1/10 Marine Deposit Lower Shoreface (calcareous) sample SL4-1/11 Termite Burrows ?245cm Marine Shells ? storm deposit KNC: Profile SL4-1 242 SN30 Regic Sand sample SN30/A Neocarbonate B sample SN30/B Hardpan Carbonate sample SN30/2 ?110cm Upper Shoreface Marine Deposit sample SN30/3 ?200cm sample SN30/5 sample SN30/4Lower Shoreface Marine Deposit ?290cm tunnel sample SN30/6 sample SN30/5 KNC: Profile SN30 243 Hardpan Carbonate samples SNT60/5; SNT60/8 SNT60 Regic Sands (calcareous) sample SNT60/1 ?135cm Neocarbonate B sample SNT60/2 Unconformity. Neocarbonate samples SNT60/3 (karst centre); SNT60/4 (edge) Calcretized Blocky Sediment sample SNT60/5b sample SNT60/5c sample SNT60/5d?200cm ?200cm Hardpan Carbonate samples SNT60/5-5d Calcretized Blocky Sediment sample SNT60/7 ?280cm Sedimentary sample SNT60/10 Sedimentary sample SNT60/11 ?370cm Sedimentary sample SNT60/12 ?200cm Calcretized Blocky Sediment (incl terrestrial fossil bone) sample SNT60/6 Dissolution features sepiolite (and calcite) cutan 1 cm Hardpan Carbonate SNT60/8 KNC: Profile SNT60 244 SPNT Regic sand (calcareous) sample SPNT/3 Orthic A (calcareous) sample SPNT/1 Neocarbonate B sample SPNT/2 ?80cm disturbed moist Hardpan Carbonate Hardpan Carbonate sample SPNT/4 Hardpan Carbonate sample SPNT/5 ?150cm KNC: Profile SPNT 245 TP231L Orthic A (calcareous ) sampleTP231L/A ?30cm ?50cm?80cm 0cm E (calcareous) sampleTP231L/E Neocarbonate B sampleTP231L/B ?60cm Regic Sand (calcareous) (disturbed?) sampleTP231L/1 Neocarbonate sampleTP231L/2 Hardpan Carbonate sampleTP231L/3 ?80cm Unconformity ?200cm Marine deposit - structure possibly from bioturbation Abundant tunnels; carbonate-filled tunnels sampleTP231L/4 ?15m Marine deposit sampleTP231L/3 New Profile BMC: Profile TP231L 246 TP266Q Regic Sand (calcareous) sample TPQ266/5 ?100cm Neocarbonate B sample TPQ266/4 Unconformity Hardpan Carbonate sample TPQ266/3 Hardpan Carbonate sample TPQ266/1 Possibly Marine sample TPQ266/2 ?170cm ?290cm Calcretised burrows Stoneline BMC: Profile TP266Q