Contrasted Perceptions of Uluru

Uluru is an inselberg shaped in arkosic sandstone located in the arid plains of central Australia. The indigenous people believed it rose out of a sand hill and has since remained unchanged. The various minor landforms represented mythological animals, people, and events of the Dreamtime. Later investigators interpreted the residual as remaining after long-distance scarp retreat, or as due to scouring by the wind or by the waters of a huge lake. The inselberg and its decorations have been construed in terms of climatic variations. Uluru also has been interpreted as a compressed and resistant compartment that was exposed as a low rise in the latest Mesozoic. The initial rise shed runoff. The steep flanks were shaped in the Eocene by deep subsurface weathering followed by stripping of the regolith and exposure of bedrock forms. Large tafoni and breaks of slope were formed on the southern side where permeable sediments abutted the residual. Following their exposure, basal flares and footcaves were shaped during a later period of subsurface weathering. The inselberg has grown as a relief feature not by uplift, but by the lowering of the surrounding plains. The morphology of Uluru is an expression of episodic exposure.     

FLARED SLOPES REVISITED

Flared slopes are smooth concavities caused by subsurface moisture-generated weathering in the scarp-foot zone of hillslopes or boulders. They are well represented in granitic terrains but also developed in other massive materials such as limestone, sandstone, dacite, rhyolite, and basalt, as well as other plutonic rocks. Notches, cliff-foot caves, and swamp slots are congeners of flared slopes. Though a few bedrock flares are conceivably caused by nivation or by a combination of coastal processes, most are two-stage or etch forms. Appreciation of the origin of these forms has permitted their use in the identification and measurement of recent soil erosion and an explanation of natural bridges. Their mode of development is also germane to the origin of the host inselberg or bornhardt and, indeed, to general theories of landscape evolution. But certain discrepancies have been noted concerning the distribution and detailed morphology of flared slopes. Such anomalies are a result of structural factors (sensu lato), of variations in size of catchment and in degree of exposure, and of several protective factors. Notwithstanding, the original explanation of flared slopes stands, as do their wider implications. [Key words: flared slope, inselberg, soil erosion, weathering, fractures.]     

The gawler ranges,South Australia: an unusual volcanic massif

The volcanic residuals of the Gawler Ranges together form an extensive massif that in its gross morphology differs markedly from most exposures of silicic volcanic rocks. The upland developed in two stages, the first involving differential fracture‐controlled subsurface weathering, the second the stripping of the regolith. As a result, an irregular weathering front was exposed, with domical projections prominent. These bornhardts are etch forms, and they are of considerable antiquity.

The differential weathering of the rock mass reflects the exploitation of various fracture systems by shallow groundwaters. Orthogonal fracture systems at various scales, sheet fractures and columnar joints control the morphology of the bornhardts in gross and in detail.

The exploitation of the structural base, which was established in the Middle Protero‐zoic, probably took place throughout the Late Proterozoic and the Palaeozoic, though only minor remnants of the Proterozoic land surface remain. The major landscape features developed during the Mesozoic. The weathering which initiated the bornhardts occurred in the Jurassic or earlier Mesozoic, and the landforms were exposed in Late Cretaceous to Early Tertiary times.

Though structural forms dominate the present landscape, some major and some minor landforms are best explained in terms of climatic changes of the later Cainozoic. The palaeodrainage system, established under humid conditions by the Early Tertiary, was alluviated during the Cainozoic arid phases, and salinas were formed. The sand dunes of the region also reflect this aridity.     

广州莲花山的人工丹霞地貌和旅游开发

通过实地考察,概述广州番禺莲花山与粤北丹霞山地质地貌的异同,从丹霞地貌角度阐述这座古石矿场“人工无意夺天工”形成的陡崖,岩洞,石墙,石巷,岩湖,石柱,石门,石桥等人工丹霞奇观,并对这座以全国罕有古石矿场为主景的莲花山旅游区的进一步开发,提出了讨论意见。     

Influence of rock strength on the valley morphometry of Big Creek, central Idaho, USA

Analysis of valley morphometry and bedrock strength along Big Creek, central Idaho, shows that valley floor width is strongly controlled by bedrock. We performed statistical analysis of Schmidt hammer rock strength as a function of lithology and aspect and of valley morphometry as a function of rock strength. Rock strength is significantly greater on the south side of the valley and in Eocene granodiorites. Rock strength is weakest in Eocene volcanic tuffs. Valley floor width depends negatively on weakest valley-side rock strength, and hillslope gradient on the north side of the valley depends positively on rock strength. Stream gradient does not depend on rock strength. Valley floor width appears to be controlled by bedrock strength on the weaker side of the valley, which was generally the north (south-facing) side. We speculate that a higher degree of weathering via freeze–thaw cycles contributes to lower strength on the north side. The positive dependence of hillslope gradient on rock strength on the north side provides evidence that differential weathering across lithologies determines the gradient that can be maintained as lateral migration of the stream erodes valley walls. These results suggest that in situ rock strength exerts strong influences on some measures of valley morphometry by modulating hillslope mass wasting processes and limiting lateral erosion.     

Differential weathering and erosion in an inselberg landscape in southern Zimbabwe: A morphometric study and some notes on factors influencing the long-term development of inselbergs

A morphometric analysis of inselberg systems in southern Zimbabwe indicated that there are two types of inselberg systems that differ in the mode of development. Inselberg systems characterized by backwearing tend to occur on granitoid rocks that are foliated, highly fractured, and prone to grusification. The inselbergs are mostly of the koppie-type and are covered to a great extent with regolith or debris. The second type of inselbergs consists of sparsely fractured granitoid rocks, and areas underlain by these rocks are characterized by clusters of closely spaced rock domes. In the case of these inselbergs, backwearing is not indicated. They appear to be the result of the exposure of structurally pre-designed, highly resistant rock compartments.The hypothesis that differences in lithology and structure may influence the rate and style of inselberg development was tested by means of a simulation model. The simulations are run under the same boundary conditions with differing assumptions concerning role of divergent weathering. The results show that both modes of inselberg development may occur alongside one another as long as a critical relief is maintained. The modelling provided also indicates that, although the inselberg systems must have integrated several sequences of environmental changes during their development, these changes were not able to divert the general trend of the process–response system away from its steady course. Lithological and structural factors, which are responsible for the differences in the mobility of material on the inselberg sideslopes, are thus considered to be the primary source of influence of the different modes of inselberg development. It is suggested that the compositional and structural diversity in granitoid rock terrains is what enables inselbergs to pass through different development modes and sequences as rocks of differing resistance become subject to weathering during the development of the landscape.     

Palaeomagnetism of middle Proterozoic (c. 1.25 Ga) dykes from central North Greenland

From a nunatak in central North Greenland (81.5°N, 44.7°W) nine sites of Middle Proterozoic basic dykes, cutting Archaean basement, were palaeomagnetically investigated. After AF and thermal cleaning the nine dyke sites and three adjacently baked gneiss sites give a stable characteristic remanent mean direction of D = 265°, I = 21.5° ( N = 12, α ₉₅= 5.6°), the direction being confirmed by a detailed and positive baked contact test.
The polarity of the dykes in the nunatak area is opposite to that of the Zig-Zag Dal Basalts and the Midsommersø Dolerites in eastern North Greenland some 200–300 km away, the volcanics of which are assumed to be of similar age (about 1.25 Ga). The remanent directions of the two sets of data are antiparallel within the 95 per cent significance level of confidence.
When rotating Greenland 18° clockwise back to North America by the 'Bullard fit', the pole of the central North Greenland dolerites (NDL) falls at (14.3°N, 144.3°W). The reversed pole (14.3°S, 35.7°E) fits well on to the loop between 1.2 and 1.4 Ma on the apparent polar wander swath of Berger & York for cratonic North America.
The palaeomagnetic results from the Middle Proterozoic basic dykes from central North Greenland thus strengthen previous palaeomagnetic results from the Midsommersø Dolerites and Zig-Zag Dal Basalts from the Peary Land Region in eastern North Greenland, suggesting that Greenland was part of the North American craton at least for the period between c . 1.3 and 1 Ma (and probably up to the end of Cretaceous time). The major geographical meridian of Greenland was orientated approximately E–W, and the palaeo-latitude of Greenland was about 10°–15°.     

Unravelling the widening of the earliest Andean northern orogen: Maastrichtian to early Eocene intra‐basinal deformation in the northern Eastern Cordillera of Colombia

The onset of deformation in the northern Andes is overprinted by subsequent stages of basin deformation, complicating the examination of competing models illustrating potential location of earliest synorogenic basins and uplifts. To establish the width of the earliest northern Andean orogen, we carried out field mapping, palynological dating, sedimentary, stratigraphic and provenance analyses in Campanian to lower Eocene units exposed in the northern Eastern Cordillera of Colombia (Cocuy region) and compare the results with coeval succession in adjacent basins. The onset of deformation is recorded in earliest Maastrichtian time, as terrigenous detritus arrived into the basin marking the end of chemical precipitation and the onset of clastic deposition produced by the uplift of a western source area dominated by shaly Cretaceous rocks. Disconformable contacts within the upper Maastrichtian to middle Palaeocene succession document increasing supply of quartzose sandy detritus from Cretaceous quartzose rocks exposed in eastern source areas. The continued unroofing of both source areas produced a rapid shift in depositional environments from shallow marine in Maastrichtian to fluvial‐lacustrine systems during the Palaeocene‐early Eocene. Supply of immature Jurassic sandstones from nearby western uplifts, together with localized plutonic and volcanic Cretaceous rocks, caused a shift in Palaeocene sandstones composition from quartzarenites to litharenites. Supply of detrital sandy fragments, unstable heavy minerals and Cretaceous to Ordovician detrital zircons, were derived from nearby uplifted blocks and from SW fluvial systems within the synorogenic basin, instead of distal basement rocks. The presence of volcanic rock fragments and 51–59 Ma volcanic zircons constrain magmatism within the basin. The Maastrichtian–Palaeocene sequence studied here documents crustal deformation that correlates with coeval deformation farther south in Ecuador and Peru. Slab flattening of the subducting Caribbean plate produced a wider orogen (>400 km) with a continental magmatic arc and intra‐basinal deformation and magmatism.     

OLDLANDS: CHARACTERISTICS AND IMPLICATIONS BASED ON THE AUSTRALIAN EXPERIENCE

Oldlands are complex surfaces of low relief preserved on Precambrian shields and cratons and Paleozoic massifs. Interpretation of their character and age is difficult, but as a consequence of its particular location and of conceptual developments, much of the Australian Craton is now susceptible to analysis in terms of exhumation, etching, and multistage development. Exhumed surfaces of many ages are recorded. Long periods of weathering and erosion generated low relief, although recurrent block tectonics produced a differentiated topography and also resulted in regolithic veneers, some of them with mineral concentrations that later became duricrusts. The associated landforms are of various ages, but are mostly of Early and Middle Tertiary ages. Cretaceous and Early to Middle Tertiary etch forms are widely developed and preserved. Earlier Mesozoic (Jurassic, Triassic) surfaces are also represented or implied. Many cratonic landforms are related to the subsurface weathering and subsequent erosion to which oldlands have been subjected, to the exploitation of fractures in the basement rocks, to underprinting from fracture zones in the basement, and to deep erosion, causing rivers to breach alien structures. Multistage as well as two-stage forms are common, and pre-weathered detritus derived from regoliths was contributed to adjacent basins. [Key words: oldland, etch surface, underprinting, duricrust, paleosurface.]     

Sediment provenance and routing evolution in the Late Cretaceous–Eocene Ager Basin,south-central Pyrenees,Spain

This study constrains the sediment provenance for the Late Cretaceous–Eocene strata of the Ager Basin, Spain, and reconstructs the interplay between foreland basin subsidence and sediment routing within the south-central Pyrenean foreland basin during the early phases of crustal shortening using detrital zircon (DZ) U-Pb-He double dating. Here we present and interpret 837 new DZ U-Pb ages, 113 of which are new DZ (U-Th)/He double-dated zircons. U-Pb-He double dating results allow for a clear differentiation between different foreland and hinterland sources of Variscan zircons (280–350 Ma) by leveraging the contrasting thermal histories of the Ebro Massif and Pyrenean orogen, recorded by the zircon (U-Th)/He (ZHe) ages, despite their indistinguishable U-Pb age signatures. Cretaceous–Paleocene sedimentary rocks, dominated by Variscan DZ U-Pb age components with Permian–Triassic (200–300 Ma) ZHe cooling ages, were sourced from the Ebro Massif south of the Ager Basin. A provenance shift occurred at the base of the Early Eocene Baronia Formation (ca. 53 Ma) to an eastern Pyrenean source (north-east of the Ager Basin) as evidenced by an abrupt change in paleocurrents, a change in DZ U-Pb signatures to age distributions dominated by Cambro-Silurian (420–520 Ma), Cadomian (520–700 Ma), and Proterozoic–Archean (>700 Ma) age components, and the prominent emergence of Cretaceous–Paleogene (<90 Ma) ZHe cooling ages. The Eocene Corçà Formation (ca. 50 Ma), characterized by the arrival of fully reset ZHe ages with very short lag times, signals the accumulation of sediment derived from the rapidly exhuming Pyrenean thrust sheets. While ZHe ages from the Corçà Formation are fully reset, zircon fission track (ZFT) ages preserve older inherited cooling ages, bracketing the exhumation level within the thrust sheets to ca. 6–8 km in the Early Eocene. These DZ ZHe ages yield exhumation rate estimates of ca. 0.03 km/Myr during the Late Cretaceous–Paleocene for the Ebro Massif and ca. 0.2–0.4 km/Myr during the Eocene for the eastern Pyrenees.     

Late Cretaceous‐Palaeogene stratigraphic and basin evolution in the Zhepure Mountain of southern Tibet: implications for the timing of India‐Asia initial collision

This article presents combined stratigraphic, sedimentological, subsidence and provenance data for the Cretaceous–Palaeogene succession from the Zhepure Mountain of southern Tibet. This region records the northernmost sedimentation of the Tethyan passive margin of India, and this time interval represents the transition into continental collision with Asia. The uppermost Cretaceous Zhepure Shanpo and Jidula formations record the transition from pelagic into upper slope to delta‐plain environments. The Palaeocene–lower Eocene Zongpu Formation records a carbonate ramp that is overlain by the deep‐water Enba Formation (lower Eocene). The upper part of the Enba Formation records shallowing into a storm‐influenced, outer shelf environment. Detrital zircon U–Pb and Hf isotopic data indicate that the terrigenous strata of the Enba Formation were sourced from the Lhasa terrane. Unconformably overlying the Enba Formation is the Zhaguo Formation comprising fluvial deposits with evidence of recycling from the underlying successions. Backstripped subsidence analysis indicates shallowing during latest Cretaceous‐earliest Palaeocene time (Zhepure Shanpo and Jidula formations) driven by basement uplift, followed by stability (Zongpu Formation) until early Eocene time (Enba Formation) when accelerated subsidence occurred. The provenance, subsidence and stratigraphy suggest that the Enba and Zhaguo formations record foredeep and wedge‐top sedimentation respectively within the early Himalayan foreland basin. The underlying Zongpu Formation is interpreted to record the accumulation of a carbonate ramp at the margin of a submarine forebulge. The precursor tectonic uplift during latest Cretaceous time could either record surface uplift over a mantle plume related to the Réunion hotspot, or an early signal of lithospheric flexure related to oceanic subduction, continental collision or ophiolite obduction. The results indicate that the collision of India with Asia occurred before late Danian (ca. 62 Ma) time.     

The Evolution of the Morphological Framework of the Central Namib Desert, Namibia, Since the Early Cretaceous

The Central Namib Desert in Namibia is a hyper-arid area which was greatly affected by tectonic changes in the Early Cretaceous, associated with the opening up of the South Atlantic Ocean, continental fragmentation of West Gondwanaland and the movement of a major mantle plume (the Tristan Plume). These events led to the formation of a range of subvolcanic complexes – the so-called Damaraland Complexes – and to the deposition of flood basalts – the Etendeka Lavas. The Damaraland Complexes include some striking inselberg features of great size, including Erongo, Brandberg and Spitzkoppe. The Great Escarpment, which bounds the Central Namib to landward, is of uncertain age, but it appears to have experienced a substantial degree of erosion by the Late Cretaceous. The feature is rather less well developed and persistent in the Central Namib than elsewhere in southern Africa. It is probable that the Namib has been dry for much of the last 130 Ma, and there is evidence for aridity in the early Cretaceous and in the mid-Tertiary.     

Structural evolution and the partitioning of deformation during basin growth and inversion: A case study from the Mizen Basin Celtic Sea,offshore Ireland

The Celtic Sea basins lie on the continental shelf between Ireland and northwest France and consist of a series of ENE–WSW trending elongate basins that extend from St George’s Channel Basin in the east to the Fastnet Basin in the west. The basins, which contain Triassic to Neogene stratigraphic sequences, evolved through a complex geological history that includes multiple Mesozoic rift stages and later Cenozoic inversion. The Mizen Basin represents the NW termination of the Celtic Sea basins and consists of two NE–SW-trending half-grabens developed as a result of the reactivation of Palaeozoic (Caledonian, Lower Carboniferous and Variscan) faults. The faults bounding the Mizen Basin were active as normal faults from Early Triassic to Late Cretaceous times. Most of the fault displacement took place during Berriasian to Hauterivian (Early Cretaceous) times, with a NW–SE direction of extension. A later phase of Aptian to Cenomanian (Early to Late Cretaceous) N–S-oriented extension gave rise to E–W-striking minor normal faults and reactivation of the pre-existing basin bounding faults that propagated upwards as left-stepping arrays of segmented normal faults. In common with most of the Celtic Sea basins, the Mizen Basin experienced a period of major erosion, attributed to tectonic uplift, during the Paleocene. Approximately N–S Alpine regional compression-causing basin inversion is dated as Middle Eocene to Miocene by a well-preserved syn-inversion stratigraphy. Reverse reactivation of the basin bounding faults was broadly synchronous with the formation of a set of near-orthogonal NW–SE dextral strike-slip faults so that compression was partitioned onto two fault sets, the geometrical configuration of which is partly inherited from Palaeozoic basement structure. The segmented character of the fault forming the southern boundary of the Mizen Basin was preserved during Alpine inversion so that Cenozoic reverse displacement distribution on syn-inversion horizons mirrors the earlier extensional displacements. Segmentation of normal faults therefore controls the geometry and location of inversion structures, including inversion anticlines and the back rotation of earlier relay ramps.     

阿尔泰山晚新生代以来的古风化壳及其形成时代

阿尔泰山古风化壳研究结果表明，晚新生以来，至少存在两个风化期，形成了不同类型的风化壳，本文阐述这两个风化期的特征及相应形成的风化壳的类型及特点。     

Post-Palaeozoic evolution of weathered landsurfaces in Uganda by tectonically controlled deep weathering and stripping

A model for the evolution of weathered landsurfaces in Uganda is developed using available geotectonic, climatic, sedimentological and chronological data. The model demonstrates the pivotal role of tectonic uplift in inducing cycles of stripping, and tectonic quiescence for cycles of deep weathering. It is able to account for the development of key landforms, such as inselbergs and duricrust-capped plateaux, which previous hypotheses of landscape evolution that are based on climatic or eustatic controls are unable to explain. Development of the Ugandan landscape is traced back to the Permian. Following late Palaeozoic glaciation, a trend towards warmer and more humid climates through the Mesozoic enabled deep weathering of the Jurassic/mid-Cretaceous surface in Uganda during a period of prolonged tectonic quiescence. Uplift associated with the opening South Atlantic Ocean terminated this cycle and instigated a cycle of stripping between the mid-Cretaceous and early Miocene. Deep weathering on the succeeding Miocene to recent (African) surface has occurred from Miocene to present but has been interrupted in the areas adjacent to the western rift where development of a new drainage base level has prompted cycles of stripping in the Miocene and Pleistocene.     

Mélange development in the Neyriz region of Zagros Orogen,Iran: Record of convergence and collision in the Neotethyan Realm

Mélanges are formed by sedimentary, tectonic and diapiric processes and are generally found in collisional belts. The Zagros Orogeny provides an intriguing geological laboratory for the study of mélange-forming processes during the progressive tectonic evolution of the Neotethys Ocean. Different types of tectonic and sedimentary mélanges occur in specific structural positions within the Zagros orogenic belt in the Neyriz Region (Iran). Based on their block-in-matrix fabrics, and tectonostratigraphic positions, we differentiated 14 different mélange types, which mark different episodes of the tectonic evolution of the Neyriz Region from the Cretaceous subduction to the Miocene collision. The Cretaceous subduction stage is recorded by volcanic-sedimentary mélanges (Mv). Sedimentary mélanges characterized by megabreccia from the Cretaceous limestone (Ms1) and Eocene polymictic megabreccia (Ms2) represent epi-nappe mélanges formed during the Palaeocene–Eocene in wedge-top basins. The ophiolite emplacement in the Oligocene resulted in local extensional tectonics in the upper part of the ophiolitic nappe, and deposition of a polymictic megabreccia (Ms3, Ms4). As the final production of the Neotethys Ocean closure and the Eurasian-Arabian collision, the sedimentary mélanges characterized by different types of chaotic rock units (Ms5, Ms6, Ms7 and Ms8 facies) were developed in front of the Cretaceous–Eocene nappes due to growth of the orogenic wedge in the Miocene. Our findings indicate that the recognition and distinction of different types of mélange may provide additional constraints for a better understanding of the tectono-sedimentary evolution of the Neotethyan region.     

Shallow-water deltaic clinoforms and process regime

Utilizing two outcrop data sets with dip direction exposures of shallow-water (tens of meters) deltaic clinoforms, this paper quantifies sedimentary facies proportions and clinoform lengths and gradients, and links process regimes to delta clinoform dimensions. Both data sets are from foreland basins, the Cretaceous Chimney Rock Sandstone of the Rock Springs Formation from the US Western Interior, and the Eocene Brogniartfjellet Clinoform Complex 8 of the Battfjellet Formation from the Central Basin of Spitsbergen. Sedimentary facies indicate presence of both river- and wave-dominated clinothems in each data set. Facies characteristics and distribution implies that river-dominated clinothem progradation was primarily driven by deposition from weak hyperpycnal flow turbidity currents across the clinoforms, and minor slumps. Wave-dominated clinothems were constructed by wave processes rather than alongshore currents, and are also progradational subaerial clinoforms, with one exception, where the formation of a compound subaqueous clinoform set indicates erosion and sediment bypass above the wave base. Sediment distribution and lithological heterogeneity in the river-dominated clinothems is controlled by individual hyperpycnal flow events or mouth-bar collapse events, and thus by self-organization and minimal reworking that results in a heterogeneity that is difficult to predict (high entropy). The efficient reworking of river-derived sediments in wave-dominated clinothems results in predictable lithological sediment partitioning (low entropy). Clinoform dimension analyses show that although of similar sediment caliber, river-dominated clinoforms in both data sets are on average 3–4 times steeper and 3–4 times shorter than the wave-dominated clinoforms, with mean gradients of ca 4 degrees and ca 1 degree, respectively, and mean lengths of 150–230 m and 640–760 m. These results require corroboration from additional data sets, but do suggest that river- and wave-dominated delta clinoforms are likely to have distinct downdip extents (lengths) and gradients for given clinoform heights. Clinoform shape can thus be a method for differentiating ancient river- vs. wave-dominated deltaic clinoforms, in addition to their sedimentary facies, biogenic features and sandstone maturity, and helpful when incorporated into reservoir models.     

西南极乔治王岛北海岸火山岩的~（40）Ar／~（39）Ar和K-Ar年龄测定

本文采用４０Ａｒ／３９Ａｒ、Ｋ－Ａｒ和激光微区等时线方法，对乔治王岛北海岸的火山岩进行了系统的年龄测定。实验结果表明，该区的火山活动从晚白垩世延续到始新世末期，主要喷发时代为始新世。岩石的年龄由南西向北东依次变新，表明乔治王岛的火山活动中心在不断迁移，与整个南设得兰群岛火山岩的时、空分布规律相符。侵入岩比火山岩生成晚而岩石化学成分更偏酸性，说明侵入岩可能是火山岩同源岩浆分异演化的产物。这批高质量数据的获得，将为区域火山岩地层时代的厘定和构造岩浆演化过程的研究提供有力的依据。     

A new history of cave development at Bungonia,N.S.W.

Bungonia Caves are the result of three distinct phases of speleogenesis. The first, Late Cretaceous phase is characterised by the development of horizontal passages close to the plateau surface. Caves developed in the lower limestone during this phase probably drained southward to risings in Becks Gully. The second, Palaeocene phase resulted in the development of dolines and large dynamic phreatic conduits. Caves extended to depths approximating the level of the Efflux and drainage from caves in the lower limestone was captured by the caves in the middle limestone, rising above the level of the Efflux. The second phase ended when the caves were filled with laminated clays, blocking underground drainage, and the surface was buried by quartz‐rich fluvial sediments prior to the Eocene. The third, and continuing phase, which began in the Late Tertiary, is characterised by the development of vadose shafts and by the removal of sediment deposited following the second phase.