The author thanks Dr Margaret Parmée for translation of various sections into French; the Editor, Gilles Arnaud-Fassetta, for necessary amendments; and two anonymous referees for their positive recommendations.
1The landforms known as pediments were so called by Gilbert (1890, p. 183; see also McGee, 1897, p. 92) as a metaphor for features typical of classical Greek architecture. They have been defined in various ways and degrees of complexity, but there was a broad understanding that they are smooth, gently inclined, erosional surfaces located between a backing scarp and the alluviated valley or basin axis. Many carry a complete or partial detrital veneer. They are well developed in arid and semiarid mid-latitude lands. They were regarded as epigene in origin, having been shaped at the surface by running water.
2Unfortunately, many aspects of such characterisations are called into question by the field evidence. Three types of pediment – covered, mantled, and rock (fig. 1 and fig. 2) – have been identified (Twidale, 1981). Though all the types theoretically conform to the same idealised morphology, they differ in their constitution and likely origin. Thus, the bedrock surface of covered pediments carries a veneer of introduced or allochthonous coarse alluvium, whereas mantled forms take their name from an autochthonous weathered veneer in situ that rests on a surface etched in the country rock. Bare rock pediments are smooth, gently inclined, and extend from the base of the backing scarp, whereas platforms are separated from uplands and stand in isolation surrounded by mantled pediments. Problems confronting the conventional view of pediments are first discussed, and an alternative model of development is then proposed.
Fig. 1 – A: Apron of covered pediments developed in western piedmont of Flinders Ranges, South Australia. The formations exposed in the upland rampart dip west towards the viewer. B: Mantled pediment in granite, Erongo Mountains, Namibia. C: Rock pediment, developed in western scarp foot of The Humps, southwestern Yilgarn Craton, Western Australia.
Fig. 1 – A : Apron de pédiments couverts développé sur le piedmont occidental des Flinders Ranges, Australie du Sud. Les formations exposées sur le « rampart » des hautes terres s’inclinent vers l’ouest en direction de l’observateur). B : Pédiment couvert (autochtone) granitique, Montagnes Erongo, Namibia (photo : J.W. Gevers). C : Pédiment rocheux développé sur le piedmont occidental de l’escarpement des Humps, Yilgarn Craton, Australie Occidentale.
Fig. 1A: photo, C.R. Twidale; fig. 1B: photo: J.W. Gevers: fig. 1C: photo: C.R. Twidale.
Fig. 1A : photo, C.R. Twidale ; fig. 1B : photo : J.W. Gevers : fig. 1C : photo : C.R. Twidale.
Fig. 2 – Location maps of part of southern Australia.
Fig. 2 – Cartes de repérage d’une partie de l’Australie du Sud.
3Several of those who studied pediments in the American South West and elsewhere assumed that retreating escarpments were essential to pediment formation. A.C. Lawson (1915, p. 37), for example, concluded that fault scarps had been worn back, leading to the formation of a bedrock surface between the base of the retreating scarp and the valley or basin floor. Lawson called it the sub-alluvial bench (but surely ‘platform’, for a bench is planate structural feature?) though others referred to it as a pediment. He left open the question of how the ‘sub-alluvial bench’ had been shaped, but he entertained the possibility that streams could dissect or strip the detrital cover and expose the underlying bedrock surface.
4The evolution of scarps and hillslopes had been studied long before Lawson cited their recession as a fundamental aspect of pediment development. Their sequential development was deduced on theoretical grounds and later supported by astute field observations and inferences drawn from them (e.g., Fisher, 1866, 1872; Jutson, 1914; Holmes, 1918, p. 93; Lehmann, 1933; King, 1949, 1953, 1957). Two structural settings call for consideration.
5First, that scarps are worn back is suggested by the observation that dissected plateaus are reduced to mesas and buttes or even pillars or columns. The residuals concerned are developed in capped flat-lying sedimentary or volcanic sequences, but regardless of degree of dissection, the usually faceted bounding slopes, which develop from the base upwards (e.g., Jack, 1912, p. 12; Penck, 1924) are closely similar in both morphology and inclination (King, 1942, 1953). This has been confirmed by slope measurements (Fair, 1947, 1948) that indicate variations only within close limits during recession. Scarps are indeed worn back parallel with themselves, though at a rate that probably decreases as the area of the upland catchment, represented by the plateau surface, is reduced to mesas and buttes, and the volume of runoff passing over the scarp is thereby diminished (Twidale, 1978).
6Second, though scarps are worn back relatively rapidly where they are breached by major rivers, the resultant gaps may be superimposed on an escarpment induced and maintained by seepage, undermining, and collapse, followed by gully gravure (Bryan, 1940; Twidale, 2000). Given a reduced rate of backwearing, there has been time for the initiation of indents, depressions, and false cuestas (e.g., Twidale and Milnes, 1983) by shallow groundwaters concentrated in the scarp foot zone. Similar retardation is inherent in other structural settings. The scarp backing the Brachina pediments in the western piedmont of the central Flinders Ranges, South Australia, is a dip slope in moderately inclined (45-60º W) crystalline limestone (fig. 1A). That the rampart is being degraded is indicated by solutional forms and by limestone debris scattered over the immediate piedmont, but the process is so slow that the scarp can be regarded as stable, as indicated by the excavation of a complex depression developed during the formation of a flight of covered pediments (Twidale, 1981).
7Many types of rock, including granite, are relatively stable when dry, but are rapidly altered or broken down when in contact with water (e.g., Alexander, 1959). However, though upper slopes shed water, moisture accumulated in scarp-foot zones causes weathering, followed by erosion by regressive piedmont streams, and the shaping of steep basal slopes and scarp-foot depressions. The formation of the latter is possibly aided by flushing of fines and volume decrease and subsidence consequent on chemical weathering (Ruxton, 1958; Trendall, 1962). In addition to the formation of the assemblage of minor piedmont forms, slope steepening has generated an abrupt transition from hillslope to plain in the piedmont angle (Twidale, 1967). Such weathering and erosion imply development over time in an essentially static scarp-foot zone. As H. Bremer (1965) commented regarding the base of Uluru, the arkosic inselberg standing some 300 km southwest of Alice Springs in the deserts of central Australia, such landforms as footcaves and flares slopes indicate that there has been little or no recent backwearing at the site.
8Accordingly, and unlike plateaus terrains, the field evidence suggests that in granitic and similar terrains the bounding scarps of residuals are effectively stable. Flared slopes and scarp-foot depressions, for example, occur not only at the margins of isolated boulders and hills (fig. 3) but also at the borders of massifs like the Kamiesberge of Namaqualand, in the Northern Cape Province, Republic of South Africa. They are developed, albeit faintly, in the highly resistant silicic volcanic rocks of the southern Gawler Ranges, in the arid interior of South Australia.
Fig. 3 – A: Flared flank of rhyolite block with narrow platform and scarp foot depression, City of Rocks, southern New Mexico. B: Scarp-foot depression at base of granitic nubbin or knoll, Mojave Desert, southern California.
Fig. 3 – A : Flanc concave d’un bloc de rhyolite avec plateforme étroite et dépression de piedmont, City of Rocks, Nouveau-Mexique méridional. B : Dépression de piedmont à la base d’un « nubbin » ou « knoll » granitique, Désert Mojave, Californie du Sud.
Fig. 3A: (photo : J.E. Mueller; fig. 3B: photo : T. Oberlander.
Fig. 3A: (photo : J.E. Mueller; fig. 3B: photo : T. Oberlander.
9Though various processes have been suggested in explanation of pediments, whether covered, mantled or bedrock, all invoke the work of water. For example K. Bryan (1922) concluded that the reasonably planate rock surfaces left behind by scarp retreat were surfaces of stream transportation, the common gently concave-upwards profiles being attributed to the downslope decrease in the calibre of detritus washed from the adjacent uplands. King also emphasised concavity as a characteristic geometric attribute of pediments and indeed referred to pediplains, that he considered were produced by the coalescence of pediments, as multi-concave forms (e.g., King, 1953). Some pediments, however, like the covered forms in the Brachina piedmont of the western Flinders Ranges (fig. 1A), and the mantled forms exposed in the Granite Hills of southeastern California are essentially rectilinear in profile and where stream rejuvenation has induced regradation of distal sectors, mantled granite pediments display convex-upward profiles, as in parts of the Einasleigh Uplands of north Queensland (Twidale, 1966, p. 46).
10W.J. McGee (1897 p. 100 et seq.) famously witnessed a sheetflood in the Sonoran region of Arizona and was so impressed that he attributed the planed granite and schist surfaces – or pediments – he had so graphically described, to such events. Subsequently, sheetfloods and the rock pediments they supposedly shaped became part of the received geomorphological lore, or what Bryan called “a favourite American doctrine” (Kirk Bryan in Brock, 1977, p. 11).
11The occasional, perhaps rare, occurrence of sheetfloods is beyond question and they achieve spectacular landscape changes (e.g., Bosworth, 1922). Yet McGee’s hypothesis involved floodwaters rampaging over a pre-existing and unexplained valley floor (e.g., Keyes, 1908; Cooke et al., 1993, p. 199). Moreover, though they deposited considerable amounts of detritus, the erosive capacity of the floodwaters was questioned, presumably because the sediment load was perceived as derived from a pre-existing regolithic cover. Prompted by Bryan (Brock, 1977, p. 11), J.T. Jutson expressed the view of many workers when he stated that floods, including sheetfloods or ‘sheet-flows’ (Jutson, 1934, p. 257-258), are agents of transport and deposition rather than erosion.
12Erosion by laterally migrating streams was advocated and championed by D.W Johnson (1932; see also Gilbert, 1877; Howard, 1942) as responsible not only for the planation of pediments but also for the development of the piedmont angle. Some rivers debouching on to the piedmont are diverted a short distance along the mountain front but this is local and limited for on leaving the uplands most rivers flow directly toward the valley axis. However, and as adduced by R.P. Sharp (1940) and others, lateral stream planation has produced planate surfaces in unlithified materials like alluvia or other regolithic materials. Flat flood plains are reasonably attributed to such processes (Crickmay, 1933; Wolman and Leopold, 1957). But whether laterally migrating distributary streams, even in flood, are capable of planation in cohesive resistant rocks save in particular circumstances remains to be demonstrated. Thus, E. Blackwelder (1931) suggested that divaricating streams armed with coarse debris, and debouching from uplands on to piedmonts underlain by relatively unresistant rocks, have produced gently sloping covered pediments comprising a cut bedrock surface with a protective veneer of autochthonous material (see also Twidale, 1981).
13‘Coarse’, however, is a relative term as is illustrated by covered forms shaped in deformed and weathered argillite at Zabriskie Point, Death Valley, California by episodic streams and protected against the potential ravages of an hyper-arid climate by transported fragments of shale. Gravel affords protection to G.K. Gilbert’s (1877, p. 130-131) ‘hills of planation’ in the Henry Mountain of Utah, and in parts of the Flinders Ranges and the Cape Fold Belt of southern Africa, quartzite has weathered to coarse blocks, some rounded to cobble size during stream transport, to form effective tools of erosion and subsequently, after deposition in point bars and channel fills, a protective cover (Twidale, 1981; Bourne and Twidale, 1998; Twidale and Van Zyl, 1981).
14Lester King considered laminar flow or sheet wash to be responsible for shaping and smoothing pediments, by contrast with the turbulent streams that erode the backing scarps (King, 1949). He may, however, have confused cause and effect, for ephemeral wash may be developed on relatively smooth gently inclined surfaces and turbulent flow on rough, rocky, scarps. Observations of wash generated by heavy rains on a mantled sedimentary pediment backed by a quartzite scarp in the Chace Range of the central Flinders Ranges suggest that the runoff was disturbed by projecting sand grains and tiny rock fragments, causing brief phases of smooth flow to become turbulent. This generated innumerable small shallow gullies that, however, were filled with sand transported from upslope almost as soon as they had been scored in the surface. The smoothness of the pediment surface was maintained by innumerable small-scale alternations of cut-and-fill. Similar observations have been made on mantled pediments in granite on Eyre Peninsula and on the southwestern Yilgarn Craton of Western Australia. In detail, deposition by surface wash contributes to the regularity of mantled pediments, just as streams in flood wash and smooth the surfaces of covered pediments.
15Whether invoking lateral corrasion, sheetwash or sheetfloods, most early students of pediments assumed that running water was capable of degrading coherent rocks such as fresh granite, and accepted the potholes, fluting and scalloping cut in granitic rocks exposed in riverbeds as evidence of erosion. Certainly the capacity of concentrated or channeled streams far exceeds that of diffuse flows. Also and clearly, events like the Late Pleistocene Spokane Flood of the northwestern United States (Baker, 1973) sustain the view that high velocity flows generate channel erosion in cohesive rocks. But with more usual levels of discharge, stream erosion appears to cease or at least to markedly diminish, once cohesive rock is encountered. Thus, A.J. Noldhart and J.D. Wyatt (1962, p. 46) working in the Pilbara region of the northwest of Western Australia, noted that: “Where stream channels have cut through the overlying rocks down to, or below the granite basement, as little as two feet of degradation has occurred in rivers that are at grade. These observations suggest that the granitic floors have undergone very little vertical erosion since exposure to the present cycle of erosion …”.
16A similar contrast in rates of erosion is illustrated in the piedmont of the Willunga fault scarp, south of Adelaide, South Australia. The escarpment is fronted by an apron of reddish-brown, Late Pleistocene fanglomerates. The fans have been subject to phases of cut and fill. They are incised by steep-sided gullies up to four metres deep. Where the streams responsible for the most recent incision emerge from the backing upland they encountered kaolinised Cambrian argillites, but whereas the streams have readily cut through the fanglomerate, producing narrow but deep gullies, erosion virtually ceases where the streams have penetrated through the base of the unlithified fanglomerate and into the underlying mudstone.
17Thus, V.R. Baker’s (1988, p. 81) comment that “… intact granite on bare slopes affords immense resistance to fluvial erosion”, even rivers in flood, is especially noteworthy in the context of the formation of mantled and rock pediments in particular, and pediplanation in general.
18Covered pediments are associated with fold mountain belts that frequently supply coarse debris. Where catchments are predominantly argillaceous the sediment load of rivers debouching on to the piedmont consists of fines that are deposited in alluvial fans (e.g., Bourne and Twidale, 1998). Though the most common type of pediment and best developed in arid and semiarid mid-latitude lands, mantled forms have been reported also from tropical monsoon lands such as Japan and northern Australia (Twidale, 1956; Akagi, 1972), and from seasonally cold climates (e.g., Van Horn, 1976). Rock pediments are developed adjacent to uplands, typically inselbergs, and in relatively massive rocks such as granite and arenaceous and rudaceous rocks.
19No satisfactory general account of pediment development emerged from early studies of the forms. They were linked to scarp retreat, which is limited in many areas where pediments are well represented. They were regarded as epigene and shaped by running water, which in reality appears to be ineffective save in limited circumstances. Only covered pediments, which have a restricted distribution, appear to be compatible, or mainly so, with the traditional criteria for the recognition of pediments.
20Whereas most workers perceived pediments as epigene forms, Paige (1912), and Tuan (1959) suggested that pediments are exhumed forms, which implies formation, burial and re-exposure, but neglected to explain how the bedrock surface had been shaped prior to interment. R.U. Cooke and P.F. Mason (1973) referred to the bedrock surfaces of the pediments they studied in southeastern California as ‘exhumed’ and attributed them to subsurface weathering beneath alluvium. Notwithstanding, though the exhumation concept simply deferred the question of the shaping of pediments in the American Southwest, it may have redirected attention from the surface to the shallow subsurface. Accordingly, the long-established impacts of soil moisture weathering were considered in relation to mantled pediments and associated bedrock forms.
21The two-stage development of landforms involving the weathering of the country rock by moisture retained in the regolith and subsequent stripping of the friable weathered mantle to expose the sculpted bedrock forms, has long been recorded in the European and American literature (Hassenfratz, 1791; MacCulloch, 1814; Pumpelly, 1879; Branner, 1896; Falconer, 1911). Such developments, to which the term ‘etching’ was later applied (Wayland, 1934; Willis, 1936), accounted for several landforms including corestone boulders and bornhardts as well as a suite of minor decorations commonly developed in granitic rocks (Twidale, 2002). Forms resulting from partial etching also were identified (Twidale, 1993; Twidale and Campbell, 1998). In parallel with these developments concerning mainly granitic terrains, M. Eckert (1902) and H. Lindner (1930) invoked etching in the karst context. Just as mantled and stripped granitic forms were recognised, so covered and bare karst, cutaneous and subcutaneous, forms were distinguished (Zwittkovits, 1966).
22Many landforms that at one time were attributed to epigene or subaerial processes, are now considered to have been initiated below the land surface, at the weathering front or base of the regolith (Mabbutt, 1961). This was demonstrated by their occurrence on freshly exposed rock surfaces and in particular in quarries and other artificial excavations (e.g., Boyé and Fritsch, 1973; Twidale, 1962; Twidale and Bourne, 1975a, 1976). Etching involves weathering in the shallow subsurface by groundwaters armed with chemicals and biota. They interact with the country rock and cause disintegration and alteration in what J. MacCulloch (1814, p. 72) called a ‘gangrenous process’, a description that graphically evokes the notion of water eating into and rotting the country rock (see also e.g., Larsen, 1948; Hutton et al., 1977).
23Etching was not invoked in respect of the widely developed mantled and rock pediments until the second half of the last century though the basic conditions necessary for its operation were noted much earlier in the literature. For instance, J.R. Logan (1851, p. 326) who worked on granite exposed on Singapore Island, observed that even in the humid tropics, outcrops are dried in the sun from time to time, but in the subsurface the rocks are in constant contact with moisture. He wrote: “When an exposed rock is attacked, the decomposing portion is washed off, and the decomposition is arrested for the time. Under-ground decomposition tends to spread unchecked on all sides”.
24D.C. Barton (1916) described similar discrepancies between the weathering of exposed and covered rocks in Egypt, and K. Bryan (1927) noted the occurrence and superior effectiveness of subsurface moisture in the weathering of bedrock. Later, F.E.S. Alexander (1959, p. 128-129) referred to subsurface solution weathering of granite on Singapore Island, and C.R. Twidale (1962, p. 62-66) invoked soil moisture weathering in explanation of the pronounced scarp-foot weathering that has shaped basal concavities or flared slopes on granite inselbergs on Eyre Peninsula. C. Wahrhaftig (1965, p. 1176-1177) attributed the weathering of granitic rocks in the Californian Sierra Nevada to subsurface moisture: and so on – several workers came to appreciate the destruction wrought by water in the shallow subsurface. These observations and surmises were brought together by J.A. Mabbutt (1966), who, based on field work at a site near Kulgera, about 250 kms south of Alice Springs in central Australia, ascribed the mantled pediments he excavated to what he called ‘ground level trimming’ or ‘mantle-controlled planation’ based in subsurface weathering or etching. He also noted the minor but significant contribution of deposition by wash and wind to the smoothing of the pediment surface. In these terms, mantled pediments are first-stage etch forms, i.e. weathering fronts shaped by subsurface moisture attack.
25Such mantled pediments are the most common and extensive of the three types recognised. They dominate many granitic terrains both in shield and orogenic settings. They are also prominent in dissected plateaux based on flat-lying sedimentary sequences and in fold mountain belts. The mantled form comprises a bedrock surface overlain by autochthonous debris, though with a veneer introduced by sheet wash. That the mantles of pediments formed in granite have not been transported is demonstrated by a lack of sorting, stratification, or cross-bedding. Exotic debris does not occur save in some sedimentary settings, such as the Arcoona Plateau, including the southern Tent Hills section, and the Flinders Ranges, where many of the pediments carry, and are protected by, a scatter of lag deposits residual from quartzitic strata that previously overlay the sites. On northwestern Eyre Peninsula, calcrete derived from wind blown calcareous dust (the parna of Australian workers such as B.E. Butler and J.T. Hutton, 1956) occurs patchily but particularly in the scarp-foot zone of mantled pediments, where it forms a protective and stabilising carapace.
26Mantled pediments underlain at shallow depth by granite have been observed in the Granite Hills of southeastern California and are demonstrated by augering, at various sites in the Isa Highlands of northwest Queensland (Twidale, 1956). K. Bryan (1925) reported that in the Papago country of southern Arizona the mantle was of the order of 50 cm thick, and by implication was consistently of that order. Elsewhere, however, the thickness of the mantle is greater and variable. For instance, regoliths up to 30 m thickness are recorded from northern Arizona (Moss, 1980) and around Ucontitchie Hill on northwestern Eyre Peninsula (fig. 4), the mantle is of the order of 6-10 m thick (R. Schmucker, pers. comm. 2012). Thus the topography of the bedrock surface beneath the regolith is in some instances more irregular than had been supposed but is negligible when considered in the context of pediments that extend several kilometres from apex to toe (fig. 4 and fig. 5A).
Fig. 4 – Mantled pediment fringing Ucontitchie Hill, a granite bornhardt located on northwestern Eyre Peninsula, SA and here seen from the east.
Fig. 4 – Pédiment couvert (autochtone) bordant Ucontitchie Hill, un « bornhardt » granitique situé au nord-ouest de l’Eyre Peninsula, SA, vu ici du côté est.
Fig. 5 – A: Map extract, northwestern Eyre Peninsula showing low inclination conical mantled pediments fringing minor granite residuals. B: Intricately dimpled crest of Parla Peak, some 42 km west of Minnipa, northwestern Eyre Peninsula, South Australia, looking over a smooth gently inclined mantled pediment that leads down to depositional plains.
Fig. 5 – A : Extrait de carte, nord-ouest de l’Eyre Peninsula, montrant des pédiments couverts (autochtones) de forme conique et faiblement inclinés, bordant de petits débris résiduels de granite. B : Complexe de crêtes ridées de Parla Peak , à environ 42 km à l’ouest de Minnipa, au nord-ouest de l’Eyre Peninsula, Australie du Sud, donnant sur un pédiment couvert (autochtone) lisse et très légèrement incliné vers des plaines d’accumulation.
Fig. 5A: 5931-4 Pordia, 1:50,000 Topographic Series, DENR, South Australia; fig. 5B: photo: C.R. Twidale.
Fig. 5A: 5931-4 Pordia, 1:50,000 Topographic Series, DENR, South Australia; fig. 5B: photo: C.R. Twidale.
27Although some granitic mantled pediments slope up to the foot of linear uplands, many, as illustrated and discussed by K. Bryan (1922), A.D. Howard (1942), Y.F. Tuan (1959), L.C. King (1962, p. 160), R.U. Cooke and P.F. Mason (1973), and others, form low angle cones encircling either an inselberg, a disintegrated remnant, a low dome, or where weathering has reduced the residual mass, merely a dimpled rock platform or a scatter of blocks and boulders (fig. 5A and B). They can be described as conoplains in the sense of I.H. Ogilvie (1905). The crestal outcrops have not been worn back, for whether in the form of bornhardts or blocks and boulders, features indicative of prolonged scarp foot weathering such as flared slopes, are present and in places prominent. Rather have they been worn down by vertical weathering with densely fractured remnants like Pordia and Cocata hills more readily reduced than more massive outcrops such as the Ucontitchie and Kolballa hills. The disintegration of the residuals may have been initiated in the subsurface while they were projections on the weathering front (Falconer, 1911). Regional evidence suggests that on Eyre Peninsula the weathering may have commenced in the Mesozoic (Twidale and Bourne, 1975b; Campbell and Twidale, 1991).
28Few streams drain from such minor crestal catchments, effectively ruling out surface runoff as the agency responsible for the shaping of the surrounding pediments. They are, however, readily understood in terms of etching, that is mantle planation facilitated mainly by direct rainfall but aided by seepage along the weathering front. The gradual reduction in the area of available catchment does not imply that the surrounding mantled pediments became relic forms. Precipitation falling directly on the surfaces would ensure continued activity at the weathering front. Apart from the immediate scarp-foot zone that is the site of maximum and locally deep weathering, runoff along the front ought to cause thickening of the mantle downslope and though scarce, the evidence from the pediment around Ucontitchie Hill is not inconsistent with this suggestion. However, those mantled pediments associated with backing uplands (and hence scarps) received runoff that became seepage from the outcrops, presumably causing them to develop more rapidly than those standing in isolation like those of Mt Damper.
29Mantled pediments are also developed in sedimentary terrains. Two environments are discussed – flat-lying and folded. The Arcoona Plateau, located northwest and west of Port Augusta, is developed in a gently tilted sequence of Neoproterozoic sediments that include several thin quartzite members that serve as caprocks to plateaux and domed plateaux (King, 1968).
30The Tent Hills region is the markedly dissected southernmost part of the Plateau. It consists of numerous mesas plus a few lesser remnants such as Sugarloaf Hill separated by extensive pedimented plains that consist of a thin regolith overlying kaolinised sediments. The regolith carries a scatter of quartzitic lag. In places the scarps bounding the mesas display structural benches, some displaced in rotational landslides. The scarps have developed from below (e.g., Twidale and Milnes, 1983) and the resulting forms suggest they have been worn back. However, there has been time for scarp-foot weathering and the formation of numerous well-developed depressions and false cuestas, the latter preserved by gravel but also in places by silcrete (fig. 6).
31Similar mantled forms are found also in the Flinders Ranges, which is a fold mountain belt developed in folded Proterozoic and Cambrian strata. Located in the central part of the upland, Rawnsley Bluff is an outward-facing southeastern scarp of the basinal Wilpena Pound. It is also the limb of a distorted dome structure. The Bluff is stepped with structural benches coincident with various quartzites and sandstones, and with a pronounced scarp-foot depression separating the Bluff from a valley floor graded to the Arkaba Creek. The valley floor is a mantled pediment with a thin regolith cover, including lag gravel, overlying an etch surface (Z in fig. 7). Linear and low strike outcrops interrupt the regularity of the surface, the character of which is demonstrated by its merging with the weathering front preserved beneath kaolinised remnants capped by a layer of coarse quartzitic blocks (e.g., Y in fig. 7, see also Twidale, 2007, p. 50). These isolated deposits are interpreted as rock falls that have occurred spasmodically, probably as a result of earth tremors. For instance, in April 1972 a magnitude 5.3 earthquake occurred in the Flinders Ranges. It caused a rockfall on the outward facing eastern scarp of Wilpena Pound. Friction caused sparks and a small bushfire.
32A higher and earlier pediment with associated scarp-foot remnants (X in fig. 7) is preserved on the slope of Rawnsley Bluff. Like the younger assemblage it is separated from the adjacent pedimented valley floor by the scarp-foot depression.
Fig. 6 – Plateau, false cuestas and scarp-foot depression eroded in flat-lying shales, Tent Hills region, South Australia.
Fig. 6 – Plateau, fausses cuestas et dépression de piedmont érodée dans des schistes argileux, région de Tent Hills, Australie du Sud.
Mantled pediments occupy all of the foreground.
Des pédiments couverts (autochtones) occupent le tout premier plan.
Photo: C.R. Twidale.
Photo : C.R. Twidale.
Fig. 7 – Remnants of mantled pediments and associated scarp-foot assemblages located adjacent to Rawnsley Bluff, central Flinders Ranges South Australia.
Fig. 7 – Restes de pédiments couverts (autochtones) et de formations de piedmonts situés à proximité de Rawnsley Bluff, Flinders Ranges central, Australie du Sud.
X: remnant of older mantled pediment; Y: debris-capped residual; Z: mantled pediment related to present local base level).
X : restes de pédiments couverts (autochtones) anciens ; Y : relief résiduel couronné de débris ; Z : pédiment couvert (autochtone) associé à l’actuel niveau de base local.
Photo: C.R. Twidale.
Photo : C.R. Twidale.
33Stripping of the mantle has exposed the erstwhile weathering fronts in rock pediments that are second-stage etch forms located adjacent to and contiguous with the basal slopes of inselbergs (Twidale, 1978, p. 1165; fig. 1C). Rock pediments are identified as of etch origin for some extend from the flared hill base for several tens of metres. Others, however, plunge to depth beneath a weathered mantle, in granitic terrains typically with corestones in place, within a few metres of the hill base (Jack, 1912). Similar associations of flared backwalls and rock pediments are found in arkose, dacite and rhyolite as well as granite (e.g., Twidale 1980; Mueller and Twidale 1988; fig. 3B).
34The subsurface provenance of rock pediments shaped in sandstone in the southern Drakensberg of South Africa, is suggested by their association not only with backwall concavities but also with other features that are strongly suspected of being of subsurface origin, such as rock doughnuts, that is, rims of rock that border rock basins, and that may be a result of differential etching associated with thin regolithic veneers (Twidale, 1980, 1993; Twidale and Campbell, 1998).
35The dimpled and grooved morphology typical of many platforms identifies the similarly decorated bevelled crests of low inselbergs like Pildappa and Turtle rocks in the Minnipa and Wudinna districts, respectively, of northwestern Eyre Peninsula and Hyden and MacDermid rocks on the southern Yilgarn Craton, as pediment platforms that now stand high in the local relief as a result of the stripping of the regolith and the lowering of the surrounding plains. On Hyden Rock this interpretation is reinforced by the occurrence of numerous rock basins with associated doughnuts.
36In addition to rock pediments contiguous with backing uplands, platforms standing in isolation either flush with the surface of flat plains, or exposed as flats or low domes on the crests of rounded rises, call for consideration (fig. 8). The term pediment platform serves to distinguish the features from other erosional flats such as shore platforms (the adjective ‘etch ‘ would not suffice, as many shore platforms are in part of etch character (Molina Ballesteros et al., 1995; Twidale et al., 2005). It also anticipates their postulated eventual development, for the lowering of the regolith implies their possible merging with and incorporation in rock pediments. Such platforms do not necessarily represent earlier phases of stripping for as mentioned previously scarp foot excavations and borehole data suggest that the weathering front is deeper in the immediate piedmont than a few tens of metres downslope.
37Some platforms, like those exposed in a road cutting at Midrand, Gautang Province RSA, and in the Elkington Quarry, some 12 km northwest of Minnipa on northwestern Eyre Peninsula are the crests of nascent bornhardts (Boyé and Fritsch, 1973; Twidale and Vidal Romani, 2005, p. 131 et seq.). They are typically dimpled and grooved owing to the development of rock basins or gnammas, gutters, and Kluftkarren. Like rock pediments they are etch forms but are isolated from any upland, in some instances owing to the formation of a scarp-foot moat of weathered country rock.
Fig. 8 – Part of the faintly dimpled and grooved granite platform located on the summit of Waddikee Rocks, northwestern Eyre Peninsula, South Australia.
Fig. 8 – Partie du plateforme de granite légèrement ridée et striée, située au sommet de Waddikee Rocks, au nord-ouest de l’Eyre Peninsula, Australie du Sud.
Photo: C.R. Twidale.
Photo : C.R. Twidale.
38Mantle planation accounts for the development of mantled and rock pediments as well as associated platforms in a range of climatic settings. At Corrobinnie Hill on northwestern Eyre Peninsula stripping of the regolith has caused extension of a rock pediment enough for it to merge with previously isolated platforms, which thus have become part of a wider rock pediment. But rock pediments and pediment platforms display stark morphological contrasts. The weathering fronts exposed in rock pediments are essentially smooth. Minor features attributable to local differential weathering and erosion occur here and there but overall the bedrock surfaces slope smoothly and gently away from the basal slope of the residual. On the other hand, whether the platforms are flat, or slightly convex-upward, they are intricately decorated (fig. 5B and fig. 8).
39Rock pediments are shaped by run-off and seepage from the adjacent hill slope. This has encouraged plant growth and colonisation by small fauna, so that the soil waters are charged with organic chemicals. In the context of rock pediments chemical attack is so effective that any potential projections are eliminated and runoff from the adjacent outcrop ensures that the products of weathering are evacuated.
40By contrast it is speculated that a few metres from the hill base (distance depending on inclination of the surface, and the volume and velocity of water spreading from the hill) the granular mantle disperses the wash such that it is insufficient to flush away the grus produced by weathering. Detritus accumulates in clefts and depressions. Moisture is retained, and colonisation by organisms encouraged. The basins and clefts characteristic of platforms proliferate and are enlarged. This contrasted development of rock pediments and platform is compatible with and reinforces the suggestion that backed pediments have evolved more rapidly than those lacking such a source of water.
41Most definitions of pediments include reference to a backing scarp. For instance, W.D. Thornbury (1954, p. 280) described pediments as abutting against mountain fronts, and R.L. Bates and J.A. Jackson (1987, p. 487) refer to pediments as occurring at the base of “an abrupt and receding mountain front or plateau escarpment…”. This association of pediment and backing scarp is in keeping with the realities of the field in many areas. Yet architectural pediments are not associated with a backing wall and as is apparent in the foregoing discussion a related upland does not play an essential role in the formation either of the widespread mantled variety or the less extensive platforms. ‘Pediment’ has come to be a portmanteau term embracing superficially similar forms, that are, however, of different origins.
42The question arises whether a low mantled conoplain like Mt Damper merits the name ‘pediment’. It and adjacent features form a sequential series in which the pediment cones persist but the backing uplands gradually have been eliminated, Ucontitchie and Kolballa hills are intact, but Pordia and Cocata have been reduced to decayed nubbins or knolls, and the crest of Mt Damper marked by a platform with a scatter of blocks and a low dome which do not constitute a backing scarp (fig. 5A). It is construed as a relic of a once substantial outcrop. The question then arises whether the pediments around say Cocata Hill ought still to be regarded as pediments once the upland and backing scarp have been eliminated, not as a result of scarp recession (there is evidence of stable etched piedmont forms and notably outcrops with flared sidewalls) but by epigene weathering. Similarly with the mantled pediments, that dominate both the piedmont zones of plateau remnants and the adjacent plains in the Tent Hills.
43The scarps of mesas such as South Tent Hill are being worn back. Corraberra Hill represents a stage in which the caprock has been undermined on one flank, and the Sugarloaf Hill a yet more advanced stage of reduction with a capping of thin quartzite and quartzite debris (Twidale, 1978). When in time the last-named residual has been eliminated will the bounding pediments be known by some other name because they no longer abut an escarpment or steep hill slope? Furthermore, and identical with the conical features around Ucontitchie Hill and associated remnants (fig. 5A), there are many gentle, mantled slopes of low rises with crestal platforms found throughout northwestern Eyre Peninsula,. Bearing in mind that these crestal platforms have been exposed as a result of wash and stripping the regolith from the crest where it cannot be replaced, ought they be called pediments only if there are indications of the former existence of a crestal residual?
44Accordingly, is a backing scarp an essential aspect of the definition of a pediment when it is not a critical to the development of the etch, mantled and rock forms? It is noted that many inselberg landscapes in Australia, in southern Africa and the interior of Brazil, for instance (e.g., Branner, 1896; King, 1962; Journaux, 1978), are dominated by plains that include significant areas of mantled pediment but from which most backing uplands and scarps if they ever existed have long since been eliminated (fig. 9).
45With these considerations in mind it is suggested that covered pediments, which as necessarily associated with a backing scarp and shaped by epigene processes reasonably can continue to be referred to by their present name, e.g., the Hayward Pediment near Brachina. Accepting their etch character, the more widely distributed mantled forms can be called backed mantled pediments (as around Ucontitchie Hill) but as scarpless mantled pediments where they are not (Mt Damper). Rock pediments remain, where contiguous with backing scarps, their etch origin understood (The Humps, Ucontitchie Hill) but pediment platforms where standing in isolation their etch origin is also a given, Middle Rock for example.
46It is appreciated that many could not contemplate scarpless pediments. Yet the conoplains are similar both in morphology and genesis to the conventional backed features. Thus, it is rational to call them pediments, but as E. Kant pointed out: “There is more than one logic”.
Fig. 9 – Inselberg landscape with etch plain comprising coalesced mantled pediments, in schist, Namaqualand, Northern Cape Province, Republic of South Africa.
Fig. 9 – Paysage d’inselberg avec surface d’altération composé de pédiments couverts (autochtones) coalescents sur des schistes, Namaqualand, Province du Cap Nord, République d’Afrique du Sud.
Photo: C.R. Twidale.
Photo : C.R. Twidale.
47Pediments are convergent landforms but all types have been construed as epigene forms shaped by running water. The field evidence however shows that only covered pediments are explicable in terms of epigene processes, namely lateral corrasion and simultaneous deposition by divaricating streams debouching from certain uplands but without the intervention of scarp retreat. They are an expression of a particular set of geological and hydrological conditions found in some sectors of fold mountain belts. Mantled pediments appear to be developed in climates other than arid and semiarid but the susceptibility of regolithic veneers to dissection implies they are likely to be best represented in lower rainfall areas and in areas where a vegetation cover is preserved. Mantled and rock forms however are shaped by etching or subsurface moisture attack, which process accounts for the salient features of mantled and bedrock pediments. It explains the autochthonous nature of the mantles. By contrast with hypotheses involving runoff as an erosional agency, subsurface weathering can account for those mantled and rock forms lacking a substantial source of surface runoff.
48Etching or mantle planation operates independently of any backing scarp or upslope catchments. Subsurface soil moisture attack is ubiquitous and constant. It accounts for minor variations in thickness of mantle demonstrated in some areas, but also eradicates all but the most resistant irregularities to produce smooth bedrock surfaces. It clarifies and renders credible the formation of the extensive planation surfaces that have been referred to as pediplains as well as several aspects of long-term landscape evolution.
49Pediments abutting scarps are typical of many landscapes and are readily recognisable as such but the conclusion that many pediments are etch forms poses problems for they evolve independently of catchments, uplands and scarps. Surely pediments ought to be redefined to accommodate these scarpless features possibly by distinguishing the epigene covered forms from those shaped in the shallow subsurface.