-nf—14378
ALLIGATOR RIVERS ANALOGUE PROJECT
ISBN 0-642-59929-7
DOE/HMIP/RR/92/073
SKI TR 92:20-3
FINAL REPORT VOLUME 3
GEOMORPHOLOGY AND PALEOCLIMATIC HISTORY
1992
AN OECD/NEA INTERNATIONAL PROJECT MANAGED BY: AUSTRALIAN NUCLEAR SCIENCE AND TECHNOLOGY ORGANISATION ALLIGATORS RIVER ANALOGUE PROJECT FINAL REPORT
VOLUME 3
GEOMORPHOLOGY AND PALAEOCLIMATIC HISTORY
K.-H. Wyrwoll
Department of Geography The University of Western Australia Nedlands WA 6009, Australia PREFACE
The Koongarra uranium ore deposit is located in the Alligator Rivers Region of the Northern Territory of Australia. Many of the processes that have controlled the development of this natural system are relevant to the performance assessment of radioactive waste repositories. An Agreement was reached in 1987 by a number of agencies concerned with radioactive waste disposal, to set up the International Alligator Rivers Analogue Project (ARAP) to study relevant aspects of the hydrological and geochemical evolution of the site. The Project ran for five years.
The work was undertaken by ARAP through an Agreement sponsored by the OECD Nuclear Energy Agency (NEA). The Agreement was signed by the following organisations: the Australian Nuclear Science and Technology Organisation (ANSTO); the Japan Atomic Energy Research Institute (JAERI); the Power Reactor and Nuclear Fuel Development Corporation of Japan (PNC); the Swedish Nuclear Power Inspectorate (SKI); the UK Department of the Environment (UKDoE); and the US Nuclear Regulatory Commission (USNRC). ANSTO was the managing participant.
This report is one of a series of 16 describing the work of the Project; these are listed below:
No. Title Lead Author(s)
Summary of Findings P Duerden, D A Lever, D A Sverjensky and L R Townley
2 Geologic Setting A A Snelling
3 Geomorphology and Paleoclimatic History K-H Wyrwoll
4 Geophysics, Petrophysics and Structure D W Emerson
5 Hydrogeological Field Studies S N Davis
6 Hydrogeological Modelling L R Townley
7 Groundwater Chemistry T E Payne
8 Chemistry and Mineralogy of Rocks and Soils R Edis
9 Weathering and its Effects on Uranium Redistribution T Murakami
10 Geochemical Data Bases D G Bennett and D Read
11 Geochemical Modelling of Secondary Uranium Ore D A Sverjensky Formation
12 Geochemical Modelling of Present-day Groundwaters D A Sverjensky
13 Uranium Sorption TDWaite
14 Radionuclide Transport C Golian and D A Lever,
15 Geochemistry of 239Pu, 129I, "Tc and 36CI J T Fabryka-Martin
16 Scenarios K Skagius and S Wingefors CONTENTS
EXECUTIVE SUMMARY
ACKNOWLEDGMENTS
1 INTRODUCTION 1
2 THE ENVIRONMENTAL SETTING OF THE KOONGARRA REGION 1
2.1 The Regional Geomorphological Context of the Koongarra Area 1 2.1.1 Geomorphology and geomorphological evolution of Koongarra 11 2.1.2 The timing and importance of Cenozoic deep weathering events 22 2.2 The Present Climatic Regime of the Area 24 2.2.1 The synoptic and larger scale circulation controls of the regional climate 27 2.2.2 Precipitation controls 33
3 PALAEOCLIMATIC HISTORY AND PALAEOHYDROLOGY 35
3.1 Tertiary Climates 35 3.2 The Quaternary Climate Record of Northern Australia 36 3.3 Glacial and Interglacial Climates and their Controls 42 3.3.1 Full glacial stages 42 3.3.2 Interglacial stages 45 3.3.3 Summary of the palaeoclimatic findings 49
4 DENUDATIONAL REGIME AND RATES OF EROSIONAL PROCESSES 50
4.1 The Koongarra Escarpment 50 4.1.1 Escarpment denudational processes 51 4.1.2 Foot slope deposits 52 4.1.3 Rates of escarpment retreat 54 4.1.4 The characteristics of escarpment retreat 57 4.2 Denudation Rates and the Erosion History of Koongarra Valley 57 4.2.1 The nature of the denudation regime and regional denudation rates 57 4.2.2 Denudation rates at Koongarra 70
5 CONCLUDING COMMENTS 79
6 REFERENCES 80 EXECUTIVE SUMMARY
The aim of this volume is to discuss the likely influence of geomorphological and palaeoclimatic controls on the development of the secondary dispersion fan at Koongarra. The geomorphological evolution of the area has a long history and needs to be seen in the context of a Phanerozoic timescale. For the Koongarra area the Phanerozoic was a time of tectonic stability and predominantly subaerial denudation. The structural geology of the region facilitated the erosion of the Kombolgie Formation, setting in train the development of Koongarra Valley. With the removal of the Kombolgie cover the surface of the Cahill Formation could then be eroded. Due to our incomplete understanding of the structural geology, it is not possible to determine the actual amount of lowering of the Cahill Surface at Koongarra.
The geochemical controls on the development of the secondary dispersion fan require the orebody to be located in an oxidising weathering environment. Under the present weathering regimes it seems that this implies that the orebody is located at a depth of less than c. 30 m. From estimates of the present regional denudation rates of the area and wider geomorphological considerations, it is concluded that the top of the orebody would have reached such a depth at some time in the last c. 1-6 million years.
In giving this estimate it is difficult to convincingly incorporate the details of the varying climates of the Late Cenozoic and the denudational response that these may have invoked. The climates of the Late Quaternary provide some guide to Pleistocene climatic events. It seems likely that at least since the Middle Pleistocene, interglacials, lasting for some 10 Ka years, would have been characterised by climates similar to those that prevail in the region today. The climates of glacial stages, lasting for some 100 Ka, are likely to have been drier than those of today. The most intense aridity coincided with times of global glacial maxima. There is also evidence that in the Late Cenozoic there were times of elevated rates of chemical weathering. However, the timing, nature and duration of such events is unclear.
There are a great number of unknowns which remain in understanding the geomorphological and palaeoclimatic controls on the development of the secondary dispersion fan at Koongarra. The study has attempted to identify those factors which are relevant to achieving such an aim. In doing this it has also aimed to provide an estimate of the timing of events which is sufficiently robust to accommodate our incomplete understanding of the events and processes involved.
in ACKNOWLEDGMENTS
I thank Peter Jolly for his help with various aspects of the work. I am grateful to Chris Uren, Bruce Gardner and Steve Riley for discussion and loan of equipment.
IV 1 INTRODUCTION
The general aim of this part of the study was to identify the changes that have taken place in the geomorphology of the immediate Koongarra area during the Late Cenozoic (a geological timescale is given in Table 1.1) and evaluate these within the context of the climatic history over this time. Reduced to their essentials, the specific problems considered focus on -
(a) the nature and rate of erosional and weathering processes,
(b) the possible role of the escarpment and
(c) the climatic history of the area.
In previous discussions of the timing and evolution of secondary uranium dispersion at Koongarra reference has been made to possible Late Cenozoic geomorphological and climatic controls. So far these factors have been considered more as adjuncts to the broader issues of geology, geochemistry and hydrology of the area, and have not been attributed the detailed analysis they require.
2 THE ENVIRONMENTAL SETTING OF THE KOONGARRA REGION
This section discusses the geomorphology and climate of the regional setting of the area. The processes that have controlled the geomorphological evolution of the Koongarra area are strongly linked to the climatic changes that have taken place over the last few million years. Climate fluctuations have directly regulated the intensity of weathering and erosion processes, through surface and subsurface hydrological controls. The global sealevel changes which have accompanied the glacial-interglacial cycles of the Quaternary have had an important affect on the geomorphology of the region.
2.1 The Regional Geomorphological Context of the Koongarra Area
The Koongarra region is located at the margin of the Arnhem Land Plateau (Figure 2.1). It is bordered to the west by the Brockman Massif, which is an outlier of Kombolgie Sandstone separated from the Arnhem Land Plateau by the intervening Koongarra Valley. Koongarra Valley has developed in the Nourlangie and Cahill formations (Figure 2.2 and Figure 2.3). A schematic summary of the relationship between the surficial stratigraphic components of the geomorphology and the solid geology is given in Figure 2.4.
The Arnhem Land Plateau and Brockman Massif are bounded by well defined escarpments, reaching broken heights of up to 200 m. The escarpments separate the plateau terrain of the Kombolgie Sandstone from the low relief, seasonally flooded wetlands and northern lateritic plains which make up much of the land surface of this part of northern Australia (Christian and Stewart, 1953). The plains rise from heights of a few metres above sea level to reach elevations of up to 60 m at the edge of the Arnhem Land Plateau. These plains are often sandy and are associated with deeply weathered soil profiles.
1 The drainage network that has developed in the region is well defined, with alluvial plains bordering the streams. The region has a pronounced seasonality of climate and surface hydrology. The presence of extensive wetlands for part of the year is a characteristic feature of the region. Stream flow is much reduced during the dry season and many streams cease to flow by the end of winter. Most of the Koongarra region forms part of the South Alligator Catchment. Only the northern part of Koongarra Valley drains into the adjacent East Alligator Catchment via Magela Creek.
Our present knowledge of the Late Mesozoic-Cenozoic denudational history of the region has been summarised by a number of authors (Williams, 1969, 1991; Galloway, 1976; Woodroffe et al., 1986). The absence of reliable age control and questions of interpretation of the field evidence makes it necessary to adopt some caution in accepting these interpretations; this is especially the case for the earlier geomorphological events suggested. In their sequence of gemorphological events (Figure 2.5), these authors propose:
(i) The region experienced a long period of subaerial weathering since the Late Proterozoic. Consequently, even prior to the marine transgression of the Aptian (110 Ma) the region was one of little relief. In places this sub-Cretaceous surface is now being exhumed. It has the characteristics of a gently undulating surface which cuts across the Precambrian rocks and decreases in height towards the north;
(ii) Formation of the older weathered land surface which has led to minor bevelling of the Cretaceous sediments and older rocks. The resulting land surface is intensively weathered and has been named the Bradshaw Surface by Wright (1963). This denudation surface is thought to be represented today by the summits of the dissected foothills and the Arnhem Plateau (Williams, 1991). It is thought that the Bradshaw Surface is older than mid-Miocene;
(iii) Slow uplift during the Miocene leading to a dissection of the older weathered surface and formation of the weathered land surface, which extends over much of the plateau hinterland;
(iv) Widespread low-angle alluvial fan deposition and stream alluviation, especially along the lower courses of the rivers. Much of this has continued into the Late Quaternary and Holocene alluvial and marine units are strong elements in the geomorphology of the region (Figure 2.6). The Late Cenozoic geomorphological events find their expression in the Koolpinyah Surface (Christian and Stewart, 1953; Hays, 1967; Williams, 1969, 1991). The Koolpinyah Surface (Figure 2.6) is a complex depositional and erosional feature, with only limited relief variation, and deeply weathered profiles which have undergone variable degrees of erosion. TABLE 1.1
GEOLOGICAL TIMESCALE
Eon Era Period Epoch Duration Years before (years) present
Holocene Quaternary (Recent) 10.000 10 Ka Pleistocene 1.5-1.8 million 1.5-1.8 Ma
Pliocene 3-5 million Neogene 5-7 Ma
Miocene 19 million Cenozoic 26 Ma
Tertiary Oligocene 11-12 million
Palaeogene Eocene 16 million
Paleocene 10 million 64-65 Ma
Maastrichian Campanian Senonian Santonian 35 million Coniacian Phanerozoic Turonian Late Cenomanian Oetaceous 100 Ma
Albian Aptian Mesozoic Barremian Early Neocomian 36 million 136 Ma
Jurassic 54-59 million 190-195 Ma
Triassic 33 million 225 Ma
Permian 55 million 260 Ma
Carboniferous 65 million 345 Ma
Devonian 50 million Palaeozoic 395 Ma
Silurian 35-45 million 430-440 Ma
Ordovician 60-70 million 500 Ma
Cambrian 70 million 570 Ma Pre-Cambrian Adelaidaan 230-530 million 800-1.100 Ma
Undifferentiated 250-550 million Proterozoic 1.350 Ma Carpentarian 450 million 1.800 Ma Early 700 million 2.500 Ma Archaean 2,100 million 4.600 Ma ARNHEM LAND PLATEAU S\J SOUTHERN HILLS AND BASINS
FIGURE 2.1 General geomorphological divisions of the Alligator Rivers region (after Australian National Parks and Wildlife Service, 1980). FIGURE 2.2 The geology of the Koongarra area (after Needham, 1982).
KEY TO FIGURE 2.2
KOMBOLGIE FORMATION
Phk Quartz sandstone, massive to flaggy; minor siltstone, tuffaceous siltstone, breccia-conglomerate, hematitic and brown ferruginous sandstone; cross-bedded, ripple-marked, conglomerate lenses.
CAHILL FORMATION
PC Undivided schist, gneiss and quartzite.
PC Feldspathic quartzite and hematitic quartz-mica schist; commonly garnetiferous, magnetic and chloritised in places; pyritic carbonaceous schist, massive crystalline carbonate, chlorite-dolomite schist, calc-silicate gneiss, amphibole schist and amphibolite.
PC Feldspathic quartzite, feldspar-quartz schist, feldspathic schist, quartzite, -#?J '•<,: i ,--rVv; garnet-mica schist, mica-quartz schist and minor conglomerate; commonly 'I "r~ "~~" f-7' R--.-.7rv''' l» magnetitic, chloritised in places. f*"11 j^J ^-MC^r '^^Si^sSSB NOURLANGIE SCHIST
Po Quartz-mica schist, quartz schist, mica schist, commonly garnetiferous; staurolite, kyanite and magnetite in places. OENPELLI DOLERITE 3? SS'^RS^' Pdo Porphyritic olivine dolerite, ophitic dolerite, granophyric dolerite and ^;,A]Kf^ differentiates, serpentinite. / -,-...,']-«{ ^-F=^^ SURFICIAL UNITS
Czt Sandstone, quartzite and schist rubble, sand : talus and scree.
Czs Unconsolidated sand : ferruginous and clayey sand.
Czl Nodular, concretionary, pisolitic and vermicular mottled laterite : in situ and reworked remnants of standard laterite profile.
Qa Silt, sand, clay : locally consolidated grey sandy siltstone along some drainage courses : creek and river alluvium.
Qs Unconsolidated sand : outwash and colluvium.
Of Black and brown humic soil and clay deposit. ¥**& JgSBr _-\,f«S
Kombolgie Formalion Nourlangie Schist
Nanambu Complex Oenpelli Dolerite
Cahill Formation
FIGURE 2.3 Geological cross-section of the Koongarra region (after Needham, 1982). INUNDATED COASTAL ARNHEM LAND NORTHERN PLAINS AND ARAFURA FALL AND ESTUARfNE PLAINS PLATEAU Czs. Czl- 'Basement' Mesozoic Czl Qa,
'Basement' Kombolgie Formation Qcm
'Basement' Mesozoic 'Basement' granite, gneiss, metamorphics 'Basement* granite, gneiss, metamorphics
Fades change
Qcm Mud, silt: intertidal mangrove swamp oo Oca Silt, mud: coastal alluvium Qcp Clay, silt, mud: coastal mud pans Qcr Sand, shelly sand, coquina: coastal sand ridges, minor beach rock Qa Silt, sand, clay: locally consolidated grey sandy siltstone along some drainage courses: creek and river alluvium Qs Unconsolidated sand: outwash and coDuvium Qas Silt, clay: abandoned channel deposit Qal Silt, clayey silt: levee deposit Qf Black and brown humic soil and clay deposit .-.•-. Soil, rubble, sand: regolith obscuring bedrock Czw Silt and clayey sand, colluvium on deeply weathered Kla and Kum Czt Sandstone, quartzite and schist rubble, sand: talus and scree Cza Winnowed sand, clay: partially stripped Czs Czs Unconsolidated sand: ferruginous and clayey sand Czl Nodular, concretionary, pisolitfe and vermicular mottted laterite in situ and reworked remnants of standard laterite profile
FIGURE 2.4 Schematic summary of the morphostratigraphy of the surffcial deposits (after Needham, 1982). metres Pre-Cretaceous surface 400 -i Post-Creteceous surface. CO Upper Tertiary surface 200 - ri=. Kombolgie Fm. ~~ 17777?. Older
NW Quaternary alluvium
FIGURE 2.5 Denudation surfaces of the Alligator Rivers area (after Williams, 1969b and Galloway, 1976). Eroded Koolpmyah Surface
Alluvial Plains
•... •'. Koolpmvah Surface
FIGURE 2.6 Extent of the Koolpinyah Surface and associated geomorphological units (after Wiiliams, 1991) 2.1.1 Geomorphology and geomorphological evolution of Koongarra
A general summary of the geomorphological characteristics of the Koongarra site is given in Figure 2.7. The area is dominated by the upland region of the Brockman Massif which is a distinctive area of Kombolgie Sandstone terrain. The upland is bounded by a pronounced escarpment, the characteristics of which are variable and are controlled by the structural geology and lithology of the Kombolgie Formation. The massif is deeply dissected, with rugged and castellated terrain. Bare rock surfaces are widespread and patches of sand and surface lags of weathered clasts are prominent. More developed lithosols occur in locations where a more substantial vegetation cover has established itself, and are especially prominent along the small water courses. Drainage network development exploits structural controls and leads ?o the development of a series of small valleys which follow fault structures. Foot slope deposits are conspicuous at the base of the escarpment and consist of talus, taluvium and sand apron accumulations.
A number of small streams drain out of the Brockman Massif and join up to form Koongarra Creek. The suggestion has been made that the development of Koongarra Creek is a relatively recent event. The proposal is that prior to its development the predominant surface drainage was via small streams to the southeast, connecting to Namorrgan Creek. The advent of Koongarra Creek was thought to mark an increase of erosion rates over the site of the orebody (Prowse, 1990; Snelling, 1990). However, a detailed examination of the drainage network (summarised in Figure 2.7) shows no evidence of a significant rearrangement of the drainage network through capture by Koongarra Creek.
The location of Koongarra Creek is related to the variations in relief which coincide with differences in lithology. Quartzite outcrops form sub-parallel, low strike- ridges. The pelitic lithologies are more easily weathered and give rise to the depression followed by Koongarra Creek. It seems likely that these relief differences were also a feature of the original Proterozoic land surface before burial by the Kombolgie Formation. The influence of lithological differences on drainage network development would only be masked through a thick mantle of surficial material which could have "filled" the relief contrast.
During the wet season Koongarra Creek transports significant volumes of bedload. Substantial volumes of alluvial deposits are associated with the creek. The alluvial sediments consist almost totally of medium to coarse sand. Weak, organic rich palaeosols are exposed in some stream bank sections. Bleaching of formerly iron-stained sands is evident, attesting to iron mobility in the alluvial sediments. Ferricretes have developed in the alluvial deposits and can attain significant thicknesses. For much of the year the bed of the stream is close to the water table. Stagnant pools of water are still present at the end of the dry season. The
11 of the Koongarra area. The water in these pools exhibited characteristics consistent with high iron-organic content. This observation is consistent with the development of ferricretes in the alluvial build-up.
The Koongarra Valley forms part of the Koolpinyah Surface. The valley has only limited relief variation, generally reflecting differences in lithology. The bedrock is extensively weathered to depths of up to circa 20 m. Shallow alluvial deposits are found along the more defined stream courses. The dominant soils in the catchments of the Koolpinyah Surface are uniform to gradational, shallow , sandy to gravelly red to yellow earths, earthy sands and siliceous sands (Duggan, 1988). Ths A1 horizon of the soils, is shallow, humus stained and weakly to moderately acid (Aldrick, 1976). The soils are massive with an earthy fabric (Wells, 1979) with a loose to friable consistency when moist. The profiles are generally strongly leached with a low clay content in the upper horizons and increasing clay with depth. In places siliceous sand to loamy sand makes up the surface or is present at shallow depths and overlies a ferricrete. Stone horizons are evident in some profiles and gravel lags of ferruginous pisolites and quartzite/quartz clasts are prominent in some areas. Williams (1968) established that in this region termite activity is a strong influence on soil characteristics.
The development of the regional-scale geomorphological setting of the Koongarra deposits is relatively straightforward, being related to the "opening" of the Komboigie Sandstone and through this the formation of Koongarra Valley. Given the structural geology of the area, the development of this geomorphological setting was inevitable. In the most general sense the original Kombolgie Sandstone "cover" was clearly divided by major fault structures, lineaments and major regional-scale joint sets (Figure 2.8). These could be readily exploited by denudational processes. This stage of geomorphological development is mirrored over parts of the present Arnhem Land Plateau where major fault lines (e.g. Sawcut Fault) or large joints have been "opened" through weathering processes.
The timing of these geomorphological events and the long- term erosional history of the area remains unclear. There have been suggestions that the initial impetus for a removal of the Kombolgie Sandstone cover in the Koongarra Valley may have been linked to regional-scale tectonic "adjustments" during the Cretaceous or Early Tertiary. A claim for post- Cretaceous differential uplift, "warping" and minor faulting in the wider region has been made by a number of authors (e.g. most recently, Needham, 1988; Mulder and Whitehead, 1988). Regional scale tectonic deformation would have led to an opening and relaxation of the rock mass of the Kombolgie Sandstone, facilitating weathering along the major structural features. Once a rock mass has been opened and an escarpment has been established, the rock mass behind the scarp face can "relax" - this facilitates weathering and mass failure processes which perpetuate escarpment retreat.
12 CO FIGURE 2.7 Overview of the geomorphological characteristics X !\ ^-^ "XK of the Koongarra area. w__<^ ^ - /;;^/—~ >-,,\^ --A' : ""x b X' ^ ,% J x-' s -^S\'-' v SECTION 2
.«
Komoolgie Formation
V**.* I Major talus accumulations ».*..T.t..*:
Escarpment
•10° Spot elevation
Well defined channel
^^ Areas subject to inundation Pools Lake SECTION1 FIGURE 2.8 Major joint and fault systems of the wider Koongarra region. SECTION 2 The major structural elements which partition the rock mass of the Kombolgie Formation are likely to date from the Proterozoic. Alignment of the 1370 Madolerite along some of the major fault sets in the Kombolgie Formation suggests that most post-Kombolgie Formation faulting probably occurred before, or at the same time as, the dolerite intrusion (Needham, 1988). Since the early Late Proterozoic the area is thought to have been tectonically stable. In past discussions within the ARAP, both Prowse (1990) and Snelling (1990) have placed some emphasis on the likely role of a Cretaceous transgression in the geomorphologicai development of the area. There is clear evidence of a widespread marine transgression in this area at that time (Needham, 1988; Mulder and Whitehead, 1988). However, the interpretation of the stratigraphic - sedimentological evidence for this transgression is, in places, questionable (e.g. Needham, 1988:45). The discussion of the likely relevance of this transgression to the specifics of the geomorphological development of the Koongarra area has been speculative. No convincing stratigraphic evidence is available to which such inferences can be tied. It is clear from estimates of present denudational rates and rates of escarpment retreat (to be discussed in a later section) that the overall regional-scale erosional history of the Kombolgie Sandstone terrain needs to be seen in the context of much longer timescales than those of the Late Phanerozoic. It should not be overlooked that the area has undergone subaerial weathering over most of the period of the last 1 Ga. With the relatively low denudation rates envisaged (the evidence for this will be discussed in subsequent sections) and the known geological history of the area, it is more appropriate to interpret the regional-scale geomorphological development as integrating events and processes which were active over a long period of time. Of importance in the geomorphology and geomorphological development of the Koongarra area is the exhumation of what is essentially a Middle Proterozoic land surface underlying the Kombolgie Formation. In the study area the Koongarra Valley has developed on the unconformity between the Kombolgie Formation and the underlying Cahill and Nourlangie formations and now forms part of the Koolpinyah Surface. Needham (1988) has suggested that prior to the deposition of the Kombolgie Sandstone, the palaeotopograhy of the eastern part of the Pine Creek Geosyncline underwent intense chemical weathering. He suggested that this weathering event is marked by a palaeo-saprolite profile, commonly over 50 m thick. The presence of such a deep weathered profile has direct implications for the rate of geomorphological development, for with the presence of such a weathered profile, the initial rate of surface lowering after the loss of the Kombolgie cover, can be expected to be high. Furthermore, there has been the suggestion that some of the deep weathering that characterises the upper part of the Cahill and Nourlangie formations may date back to the Proterozoic. 17 A well defined contact of the Kombolgie Formation and Cahill Formation is exposed 4 km north of Koongarra. in this section there is no indication of a weathered profile marking the unconformity. The schists of the Cahill Formation are essentially unweathered (Figure 2.9). In addition, the chemical data presented by Needham (1988:37) for a proposed saprolitic palaeoweathered profile developed in rocks of the Nanambu Complex at Granite Hill do not have the characteristics indicative of intensive chemical weathering. The Proterozoic land surface which marks the unconformity of the Kombolgie Formation and the underlying Cahill and Nourlangie formations has considerable relief. Along the western margin of the Brockman Massif the contact varies in height from c. 30 m to 161 m (Figure 2.10). Along the Arnhem Land Escarpment the contact between the Nourlangie Formation and the Kombolgie Formation ranges from <40 m - 163 m. The height range gives some indication of maximum land surface lowering that has occurred over the Cahill and Nourlangie formations. There is no apparent stratigraphic evidence which can be used to establish the rate at which the land surface was lowered. It is not surprising that in an area which has so little relative relief and is so close to sealevel, that base-level controls have been suggested as being an important consideration in establishing long term regional denudation history. While the history of Pleistocene base-level events in this area can be inferred from the deep sea oxygen-isotope record shown in Figure 2.11. This figure acts as a proxy for sealevel changes for the last c. 700 Ka (a change of 618O of 0.1 per mil corresponds approximately to a sealeve! of 10 m). Little is known of Tertiary base-level events. The topographic position of the Cretaceous sediments suggests that some uplift must have taken place since their deposition. Veevers (1991) suggests this area has undergone relatively little net vertical motion since the Albian (98 Ma). 18 FIGURE 2.9 Photograph showing the characteristics of the Cahill/Kombolgie unconformity. 19 132"3D' irao- ^O- Kombolgie Formation 76 Observation point I elevation in metres above mean sea level) Dip of unconformity surface *• > Dip, probably depositions! origin »• Dip. origin strongly influenced by faulting Dip within Kombolgie Formation -"- Dip. observed Anticline — —" Fault, inferred from barometer survey • Local drape structure ;..;:.•'•,. . -:*~---r' -f ^ggi:;|/ 13°00' FIGURE 2.10 Barometric levels of the base of the Kombolgie Formation (after Needham, 1988). 20 STANDARD OXYGEN ISOTOPE STRATIGRAPHY 6"0 -1 -2 910 FIGURE 2.11 Standard oxygen isotope stratigraphy to 750 Ka BP (after Prell et al., 1986). 21 2.1.2 The timing and importance of Cenozoic deep weathering events A regional-scale geomorphological consideration which has received some attention in the past, is the age of deep weathering - laterisation in the area.ln discussions within the ARAP, substantial consideration has been given to residual areas of "laterite" which occur in Koongarra Valley. These have been seen by both as representing the remnant of a once continuous cover which essentially "sealed" the land surface from erosion (Prowse, 1990; Snelling, 1990) and it was postulated that only when this "seal" was lost was it possible for erosional processes to lower the land surface sufficiently to allow the development of the secondary dispersion fan. The discussion of deep weathering and laterite remnants has been handicapped by an absence of numerical dates and stratigraphic detail, and by an apparent confusion between presumed past "deep" weathering and laterisation events and more routine and on going pedogenic processes. The history of deep weathering in the wider Koongarra region is unclear. In early work, Williams (1969) recorded the existence of a series of "laterites" - of these, detrital laterite, pisolitic laterite, mottled-zone laterite and concretionary laterite were recognised by Needham (1982) in the wider Koongarra region. The more recent literature has essentially reiterated this early work (e.g. Needham, 1988). In earlier discussions of the Koongarra area, these "laterites" have caused problems because they have been taken to be indicative of Tertiary deep weathering events. A clear distinction has to be drawn between well developed ferricretes and laterite duricrusts associated with deep weathering. In deep and intense weathering environments minerals weather into kaolinite and amorphous or crystallised ferruginous oxyhydroxides, and through a series of pathways can result in the development of strong ferruginous duricrusts (see e.g. Nahon, 1986). Such duricrusts are distinct from the development of an iron indurated crust - ferricrete - which develops in a soil profile as the result of absolute enrichment in iron. The iron moves in the soil as soluble metallo-organic chelating complexes. Fulvic acids are a common chelating agent in soils. In the latter case the ferricrete is clearly not associated with a deeply weathered profile. In the Koongarra region ferricretes are found at the surface and at shallow depths in areas adjacent to the escarpment. At depth they mark the contact between the top of the weathered Cahill schists and the overlying sand apron. Ferricretes are also found overlying weathered Cahill metasediments along Koongarra Creek, and are associated with sand-rich soil profiles in Koongarra Valley. They have also developed over weathered Cahill Formation lithologies in Koongarra Valley and are found in shallow alluvial deposits associated with the low-order streams of Koongarra Valley. In the alluvial deposits secondary re-cementation of transported pisolites is evident, which then form ferricretes. 22 The age of the ferricretes in the area is clearly diachronous. And from the occurrences of ferricretes in evidently "young" alluvial deposits, it is clear that the formation of ferricretes is, in places, a relatively recent (possibly even Holocene ?) feature. The field evidence makes it clear that the "laterites" present in the Koongarra area need not be the result of Tertiary deep weathering events. In fact, many are a relatively minor pedogenic feature found throughout much of the area in a variety of soil and geomorphological settings and are forming under present weathering conditions. However, that is not to say that strong ferruginous duricrust resultant on intense chemical weathering are not found in the area. A conspicuous example is located in the area of Koongarra airport. The characteristics of the site offer the opportunity to date the profile using palaeomagnetic methods, (e.g. Idnurm, 1986; Idnurm and Senior, 1979). From the present knowledge of the area it seems clear that the emphasis that has been previously placed on the stripping of a Tertiary "laterite" cover of the Koongarra Valley is likely to have been misplaced. The ferricretes that have been observed are frequently not part of a deeply weathered profile and many are likely to be Quaternary in age. The only absolute ages of ferricretes presently available come from Short et al., (1989), who undertook thorium-uranium disequilibrium dating on ferruginous concretions and rinds from the upper part of a soil profile near the centre of the mined Ranger One, No. 3 uranium ore body. For the pisolitic rinds, ages ranging from 202 -135 Ka BP were obtained. The pisolith cores were found to be at or near secular equilibrium and are therefore considerably older. The most informative work on the weathering history of northern Australia comes from studies on the Weipa bauxite deposits. The "Weipa Beds" of the Cape York area are thought to be of Late Cretaceous to Tertiary age. These beds have been intensively weathered and exhibit a classic laterite profile (Schaap, 1990). In general terms, it is thought that bauxitisation commenced in the mid-Tertiary on the Aurukun Surface of Grimes (1979). Indications from the 6D and 6180 composition of gibbsite from these deposits make it clear that a more intense monsoonal regime than that of today prevailed in the region during the mid-Tertiary period of bauxitisation (Bird et al., 1989). However, Schaap (1990) considers that the aluminium content of the bauxite was upgraded at least twice, once in the Pliocene and once in the Pleistocene. It has also been suggested that bauxitisation may be continuing to the present day under the influence of the present monsoonal climate. These conclusions support the early studies which imply that for northern Australia there was an intense weathering event during the mid-Tertiary, but it also extends the early results in recognising subsequent events. In addition, it is clear that the importance of weathering under present climatic conditions should not be underestimated. This will be discussed in a subsequent section of this report. 23 2.2 The Present Climatic Regime of the Area Climate is a determinant of both weathering processes and the erosional regime. The denudational processes of the area and their rate of operation have been strongly influenced by the climatic variations of the Late Cenozoic. At present the area has a climate which has a Koeppen classification of A^, - hot moist climate with a marked dry season in winter. All indications are that past climates of the Late Cenozoic have differed significantly from the climate regime presently prevailing in the region. The regional climatic setting has been discussed by Christian and Stewart (1953), Anon (1961), McAlpine (1969, 1976) and McDonald and McAlpine (1991). The climatic statistics available for Koongarra are summarised in Figure 2.12. Because of the limited length of the Koongarra record the discussion which follows draws on the climate data from Oenpelli which has a much longer record. The region has two quite distinct seasons - a dry warm winter and a hot, wet summer. The mean annual rainfall for the region is shown in Figure 2.13. The rainfall is almost totally confined to the summer period of late October to early April. Rainfall variability is low (Table 2.1). The coefficient of variation (standard deviation of the rainfall over a given period of years divided by the mean) for Oenpelli is 0.21. TABLE 2.1 RAINFALL VARIABILITY AT OENPELLI (mm) Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec Lowest 135 113 49 0 0 0 0 0 0 0 3 55 10% 213 117 101 3 0 0 0 0 0 0 40 122 50% 310 272 274 37 1 0 0 0 0 0 92 202 90% 449 498 473 143 24 3 2 1 17 92 200 362 Highest 777 655 594 414 194 35 62 15 36 168 276 583 (After McDonald and McAlpine, 1991) 24 (a) Koongarra: CAMP (1973-1989) 400 O Rainfall ••— Pan Evaporation Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Koongarra: CAMP (1972-1987) 3000 2000- E E 1000- Year FIGURE 2.12 Summary of rainfall and evaporation at Koongarra. 25 FIGURE 2.13 Regional rainfall characteristics (McDonald and McAlpine, 1991). 26 At Oenpelli the mean monthly temperature varies from 24.6°C in July to 30.2°C in November. The extremes of temperature are up to 10°C higher and more than 10°C lower than the corresponding means. A summary of the water balance for the area is given in Figure 2.14, where a strong excess of rainfall over evaporation during the wet season. A strong negative water balance prevails over the winter spring months. September is the month with the highest water deficit of -240 mm. 2.2.1 The synoptic and larger scale circulation controls of the regional climate The controls of the dynamic climatology of the region are essentially the northern Australian monsoon which determines the summer climates, and the Subtropical High Pressure Cell which largely influences the winter situation. The Australian monsoon is part of the planetary-scale monsoon regime, which is essentially a manifestation of the seasonal migrations of the equatorial trough (or Intertropical Convergence Zone) (Gadgil, 1988). When the equatorial trough lies close to the equator it is a region of confluence between the two trade wind regimes of the opposing hemispheres. If the trough moves from the equator, it is a zone of low pressure separating trade wind easterly flow on its higher-latitude side (summer hemisphere) from low-level westerly flow between it and the equator. This zone is known as the "monsoon" trough and the westerly winds are referred to as monsoon westerlies (McBride, 1987). The zone of westerlies is characteristically a zone of high summer rainfall. The Australian monsoon forms part of a monsoon regime which extends from the equator to about 15°S and from west of 100°E to about 155°E (Hendon et al., 1989). The tropospheric circulation of the Australian monsoon is characterised by a lower tropospheric trough with low level convergent westerlies to the north and a broad region of upper level divergent easterlies (Troup, 1961). An overall review of the dynamic climatology of the Australian monsoon has been recently undertaken by McBride (1987), with important additional discussion provided by Holland (1986) and Hendon et al. (1989). The overall dynamic climatology of the Australian monsoon is complex and only partially understood. A comprehensive discussion of monsoon dynamics is out of place in this report and can be obtained from the references cited. A convenient summary of the broad scale circulation pattern and life cycle of the Australian monsoon can be obtained from a series of mean monthly 900 m streamline analyses (Bureau of Meteorology, 1977). The results of these analyses are given in Figure 2.15. They provide the broad scale circulation pattern over the Australasian region during the selected months of October, December, January and March. In addition the figures show the relationship of the Australian monsoon regime to the dynamics of the transequatorial flow and the more local Australian controls. 27 (a) mm 200 Mean monthly P — E 100 0 -100 -200 (b) Soil moisture -100mm • Soil - moisture -500 - M M N FIGURE 2.14 (a) Mean monthly water balance at Oenpelli and (b) mean monthly soil moisture storage at Oenpelli and'cumulative P-E curve (after McAlpine, 1976). 28 HO' ^ iso* ieo* ^^.tm March December FIGURE 2.15 Mean monthly streamline/isotach (km/h) analyses at 900m (gradient level) for selected months (after Australian Bureau of Meteorology, 1978). 29 October is adopted as the starting point of the monsoon cycle, for it is the month which is the transition period between the cessation of the Asian summer monsoon and the Australian winter trades, and the Australian - Indonesian summer monsoon and the Asian winter monsoon. The October 900 m mean streamline analysis shows the remnants of the summer monsoon over Malaysia and the beginning of the Asian winter monsoon over the South China Sea. In the southern hemisphere, a weak easterly flow exists in the near equatorial region but the southeast trades are still dominant over the Coral Sea and the Indian Ocean. By December, the Asian winter monsoon has intensified but there is still no evidence of its crossing the equator. A fairly general westerly flow has developed over near-equatorial parts of the southern hemisphere, where the monsoon shear line has shifted southwards to about 10°S. (The monsoon shear line has been defined by McBride and Keenan (1982) as the line of separation between trade easterlies and the equatorial westerlies). Westerlies have now extended over all of Papua New Guinea but the monsoon has barely penetrated the Australian continent. The western Australian heat low has become more extensive and the Indian Ocean Trades have begun to weaken. In January, the northeast flow of the northern hemisphere curves as it crosses the equator and becomes a well developed monsoon over the near-equatorial latitudes of the southern hemisphere. The monsoon shear line has penetrated over the extreme north of the Northern Territory and lies across the northern Coral Sea. A monsoon low is located just south of Darwin while the heat low is still developing over northwestern Australia. A similar pattern prevails during February. But during this month, the monsoonal shear line reaches its maximum poleward extent, and extends well into northwestern Australia. By March, the monsoon shear line has retreated to about 10°S and the monsoons have essentially ceased over the Australian continent. The Coral Sea Trades have begun to strengthen but a monsoon is still located in the North Coral Sea, just to the south of the Solomons. In the near-equatorial zone, there is still a marked cross-equator low, although the Asian monsoon is weakening. There is still a heat trough over Western Australia but the heat low has disappeared. By May, the seasonal reversal is complete. The heat low over Western Australia and the monsoon shear line over the southern hemisphere have disappeared. The trades have strengthened over the Coral Sea and Indian Ocean and a southeast to east flow prevails over the near-equatorial latitudes. The summer monsoon has commenced over Malaysia and is moving into southeast Asia. The winter synoptic situation is much simpler than that of the summer. During these months the equatorial trough lies woll to the north of Australia and the Sub Tropical High Pressure Cell dominates the climate of northern Australia. This gives rise to subsidence, a stable troposphere, clear skies and high incident radiation. 30 The subtropical high pressure cell exerts its influence well into the summer months. The summer heat low is frequently overridden by southeasterlies emanating from the anticyclones to the south. Given strong southeasterlies overriding the surface heat low (or the monsoon shear line), a synoptic situation arises where surface humidities to the north (and within the shear zone) may be quite high, but any deep convection is still suppressed. Hendon et al. (1989) were able to demonstrate this from a consideration of moist static energy (H) and saturated moist static energy profiles (Hs). Their study (December 1986-January 1987) showed that prior to the onset of the monsoon dry southeasterlies with large scale subsidence persisted across northern Australia. Before the onset of the monsoon H and Hs profiles were such that a parcel of air at 950 mb would not become buoyant unless lifted well above 800 mb. During the active monsoon, southeasterlies were no longer present, a monsoonal trough had developed and strong westerlies and rising motion had developed over northern Australia. In these circumstances a parcel of air lifted from 950 mb will become buoyant below 900 mb and deep instability is realised. The northern Australian monsoon regime fluctuates between a strong and weak mode during which quite different circulation patterns prevail (Figure 2.16). There has been much discussion of the relationship of the Australian monsoon to Southern Oscillation/El Nino (ENSO) events (Allan, 1988). It has been known for many years (Bjerknes, 1969) that Darwin pressure is highly correlated to the strength of the Walker circulation. Hence there are good reasons to suppose that the intensity of the monsoon should be related to monsoon activity. 31 40 100E 100 SOW 20 40 100E 120 140 160E 180 160W 140 120 100 SOW FIGURE 2.16 Schematic representation of (a) strong and (b) weak summer monsoon in northern Australia (after McDonald and McAlpine, 1991). 32 Southern oscillation indices (SOI) are a measure of the difference in sealevel pressure between the Indonesian-Australasian region and the southeastern Pacific. They are positive (negative) when surface pressure is low (high) over the Australian region. A large number of statistical analyses are now available which evaluate the relationship of ENSO events to Australian tropical climate. Allan (1983, 1985, 1988) has made a strong case for the association of anti-ENSO years with the development of strong monsoonal conditions which extend well into the continent. More recent seasonal synoptic comparisons also demonstrate this. The southern summer of 1987-1988 was, in contrast to 1988-1989, preceded by a significant warm El Nino event. In this year the SOI reached a minimum of -22 in April 1987. During this year the monsoon in the southern hemisphere was weaker than normal, resulting in below average rainfall over northern Australia. For 1988-1989 most stations normally affected by the monsoon received above average rainfall. The marked contrast between these two years is demonstrated in the distribution of convection. The classical eastward displacement of the pattern of convection toward the dateline during an El Nino event (1987-1988), compared with the pattern during a positive excursion of the SOI (1988-1989), is pronounced (Keith et al., 1991). It is clear that any consideration of the climatic history of the region must bear in mind the likely importance of ENSO controls. 2.2.2 Precipitation controls Summer precipitation over northern Australia is directly related to monsoon activity. Precipitation is primarily the result of storms induced by shear line convection, but tropical cyclones, embedded in the monsoon flow, also make a significant contribution, as do more local convective systems. Southern (1966) recognised five discrete categories of rain producing systems (Figure 2.17): local convection, organised convection, monsoon troughs, tropical cyclones and easterly disturbances. At the height of the wet season, December to March, monsoonal trough and tropical cyclones are the dominant rain producing systems in northern Australia. 33 Local convection Equatorial or monsoonal troughs Organised convection Tropical cyclones Easterly disturbances FIGURE 2.17 Annual rainfall at Oenpelli and the five categories of rainfall producing systems (after Southern, 1966). 34 3 PALAEOCLIMATIC HISTORY AND PALAEOHYDROLOGY 3.1 Tertiary Climates The Cenozoic climatic history of the region and northern Australia generally is poorly known (Bowler, 1982; Kemp, 1978; Kershaw, 1988). The timing of the onset of the present monsoon climates of the region is not clear. But the work that is available suggests that this is unlikely to have occurred before the late Miocene. At present the best indications of the Tertiary climates of northern Australia comes from northeastern Queensland (see the summary discussion of Kershaw, 1988). In the Early Miocene complex rainforest existed under high precipitation levels in the coastal areas of Central Queensland. The Mid-to-Late Miocene saw the development of progressively drier conditions and the development of extensive drier rainforest types with open herbaceous and sclerophyll woody communities (also see Stein and Robert, 1985). The Atherton Tablelands, further inland, supported well developed rainforest similar to today, indicating high precipitation but with lower mean annual temperatures from those of today. In the southern Kimberleys of Western Australia lacustrine sediments are found associated with diamond-bearing lamproitic diatremes dated to c. 20 Ma - 22 Ma (Jaques et al., 1986). The sediments are clearly indicative of high lake levels, strong fringing vegetation and regional climates significantly wetter than those of today (Wyrwoll, unpublished). It is still not possible to ascertain the timing of the onset of the present climate of northern Australia. There may be some relationship between the onset of aridity in the Lake Amadeus region of central Australia and the general development of the climate regime which now prevails in the Koongarra region. Lake Amadeus is a groundwater discharge playa located some 100 km northwest of Ayers Rock. The work of Chen and Barton (1991) on the magnetostratigraphy of the sediments of the playa, indicates that full aridity in central Australia developed by either 0.9 Ma or about 1.6 Ma. Before this period the sedimentology of the lake deposits suggest a fluvial-lacustrine environment which dried seasonally. The sparsity of the record makes it difficult to discuss the details of the Tertiary and Pleistocene climate events in the area. The palaeoclimate record shows that there have been periods in the Middle-Late Tertiary when climate conditions were suitable for deep weathering; this corresponds with the evidence of bauxite formation and/or enrichment at Weipa. It would seem likely that once the full glacial-interglacial fluctuations of the Pleistocene were established, the climate regime of northern Australia would have approximated that of the Late Quaternary. 35 3.2 The Quaternary Climate Record of Northern Australia The most comprehensive Quaternary climate record of northern Australia comes from Lynch's Crater in northern Queensland (Figure 3.1) which is thought to cover the last two glacial/interglacial cycles. The palynological record obtained from this core indicates significant climate changes, and especially precipitation variations, during this period. Very high values for rainforest angiosperms in zones G, E and A indicate a dominance of complex rainforest existing in high rainfall conditions. Some variation in effective rainfall is also evident in the zones. A marked reduction in effective rainfall is especially indicated by the pollen assemblages at the beginning of zones F and D, where marked increases in sclerophyll taxa are evident. The overall summary of likely precipitation changes shows very low precipitation values at the height of the last glacial maximum (LGM). A similar trend is evident for the previous glacial - interglacial stages. It should be noted that numerical dates are only available for the upper part of the core, while the older "dates" are inferred. More recently van der Kaars (1991) has reported the palynological record for a series of deep sea cores from northern Australia and eastern Indonesia. The chronological base for many parts of these cores is limited, but nevertheless, they yield important Quaternary climatic data for northern Australia. The two most informative cores were taken - (i) 170 km southwest of Sumba and 850 km northwest of the Dampier Peninsular (Figure 3.2) and (ii) 115 km northwest of Timor (Figure 3.3). For northern Australia the pollen record of the cores shows a major change of eucalypt forest to grassland during the last glacial period. The vegetation changes indicate significant decreases in precipitation over northern Australia at that time. It is claimed that the longer record of G 6-4 reveals expansions and contractions of the arid zone over the last 300 Ka. From the inferred ages it is suggested that these repetitions were not mirror images of the last glacial period. If the inferred dates are accepted the pollen record appears to indicate that the overall climate of northern Australia before circa 190 Ka BP was persistently much wetter with a greater proportion of eucalypt woodland in the vegetation. There is no other evidence to support this claim and it is not reflected in the Lake Amadeus record (Chen and Barton, 1991; Chen et al., 1991), where it should have been registered. Considerable care has to be exercised in accepting the validity of the inferred ages. The oxygen- isotope stratigraphy proposed for this core and on which the chronological inferences are largely based is not without problems (Biekart, 1989). The correlation of G 6-4 with the Lynch's Crater stratigraphy (Figure 3.2) is problematic. Some unpublished preliminary findings of Late Quaternary climates come from northwestern Australia where a limited number of thermoluminescence dates on desert sand dunes suggest that more arid conditions prevailed in this area for much of the Late Pleistocene (Wyrwoll, unpublished). To some extent this claim is echoed in cores G 6-4 and G 6-2 of van der Kaars (1991). 36 FIGURE 3.1 Pollen record and climatic reconstruction from Lynch's Crater, northeast Queensland (after Kershaw,-1986). Pollen diagram from Lynch's Crater. Lombok Ridge NE Queensland (after Kershaw. 1966 in sisif * — 130 — —,90 us G2 GRASSLAND TAX) — 2*4 MANGROVE TAXA MONTANE FOREST TAXA t^l> TROPICAL (COASTAL) TAXA WOODLAND TAXA l»1000yrBP> FIGURE 3.2 Tentative correlation of the palynological record from Lynch's Crater with the record obtained from the Lombok Ridge, piston-core G6-4 (after van der Kaars, 1991). 38 CO 01 o o> 0. a >» "55 a> TIMOR TRENCH G5-6-149P2 calcareous clay MONTANE FOREST GROUP TROPICAL (COASTAL) LOWLAND GROUP MANGROVE GROUP GRASSLAND GROUP WOODLAND GROUP piSt n C re (after vln E ° ' ° 39 It is evident from the data that the most extreme reductions in precipitation coincided with glacial maxima. In western Australia at the LGM strong arid conditions prevailed as far south as Carnarvon (Wyrwoll, unpublished). The clear indications are that at glacial maxima, monsoon activity is significantly reduced and widespread arid conditions characterise much of northern Australia (Figure 3.4). A problem which needs to be resolved in the context of the Late Quaternary climates of northern Australia, is the development of Lake Carpentaria in what is now the Gulf of Carpentaria during the time period > 40 to 13 Ka BP (Torgersen et al., 1985). The development of this lake occurred at a time when northern Australia is supposed to have been significantly drier than today. The suggestion by Bowler (1983) that there was a massive expansion of lakes in northern Australia at c. 32 Ka BP is now no longer thought valid. The stratigraphic record of both Lake Woods (133° 30'E, 170° 45'S) and Lake Gregory (127° 20'E, 20° 15'S) show evidence of massive expansion during the Quaternary but the 14C dates are problematic and new thermoluminescence dates are now being obtained. Whatever the age of these phases of lake expansion, they clearly indicate a significantly more intense monsoon regime. The working assumption of present studies is that the last major lake expansion dates from the Last Interglacial (c. 120 Ka BP). While the Holocene climate record of northern Australia is generally better known problems of interpretation still remain. Wasson and Donnelly (1991) have recently reviewed the relevant Australian literature and conclude that over the last c. 13 Ka the Australian monsoon has essentially "tracked" the behaviour of its northern hemisphere counterparts. In their reconstruction monsoon activity was already significant by 13 Ka BP, was more intense at 9 ka BP and fell to present levels after 4 Ka BP. However, the data on which this conclusion is based remains limited. For northwestern Australia Wyrwoll et al., (1986 and 1992) have suggested that monsoon activity was subdued until well into the Holocene. However, more recent, and as yet unpublished findings, may require this view to be revised. Strong evidence for an onset of monsoon conditions over northern Australia by 13 Ka BP comes from variations in ^Sr/^Sr ratios in ostracod valves from the Carpentarian Basin (McCulloch et al., 1989). The highest Sr present in the stratigraphic record occurred at c. 13 Ka BP and this is attributed to the southward migration of the monsoonal rainfall pattern at that time. From the evidence now available it is reasonable to conclude that precipitation levels similar to that of today has prevailed in the Koongarra area for the last c. 10 Ka. In tropical Australia there may have been times during the middle Holocene when mean annual precipitations were somewhat higher than those of today. Furthermore, from evidence in the Magela Creek Wasson et al., (1990) propose a period of reduced streamflow between 1.9 ± 0.2 Ka -1.4 ± 0.3 Ka. 40 18.000 yr BP II <250 1500-2000 11 250-500 2000-2500 500-1000 >2500 1000-1500 G5-6-U9P2 120 lOOOtn FIGURE 3.4 Alternative mean annual precipitation reconstructions for northern Australia at 18 Ka BP: (a) van der Kaars, 1991. (b) Hypothesised summer isohyets in mm Shoreline at 18.000 yrBP FIGURE 3.4 Alternative mean annual precipitation reconstructions for northern Australia at 18 Ka BP: (b) Woodroffe et al., 1986. 42 3.3 Glacial and Interglacial Climates and their Controls The climates of full glacial and interglacial stages represent the two end members of the range of likely climate changes that this area experienced during the Quaternary. It may prove instructive to use their climatic characteristics - dry glacial and wet interglacia! - as climate scenarios which provide the upper and lower bound estimates for palaeohydrological reconstructions. 3.3.1 Full glacial stages From a consideration of the controls of the regional climates of northern Australia there are good reasons to expect the climates of glacial stages to be dry. This can be best demonstrated with reference to the LGM. At the height of the LGM, c. 18 Ka BP , much of the presently shallow shelf areas in the Indonesia - North Australia region became land (Figure 3.5) as a result of the lowered sealevel (c. -130 m) of that time. This must have been a factor in the existence of the widespread aridity which prevailed over northern Australia at that time. Webster and Streten (1978) attempted a general circulation interpretation for the Australasian sector during the LGM. They postulated a pattern characterised by only small annual change, an enhanced Indian Ocean trough, marked ridging over eastern Australia, and another trough further east. They considered such a basic flow pattern consistent with reduced rainfall in southeastern Australia and frequent cold outbreaks favouring glaciation in New Guinea. Prell et al. (1980) suggested that the cold surface waters of the adjacent Indian Ocean associated with the intensified West Australian Current may, through reduced evaporation, have contributed to the aridity of tropical Australia. The likely characteristics of the Walker circulation at the LGM and the possible ENSO influence on the precipitation regime of northern Australia remain unknown. Quinn (1971) suggested the prevalence of conditions characterised by a strongly positive mode of the Southern Oscillation, a vigorous Walker Circulation and pronounced upwelling in the Eastern Equatorial Pacific and a westward extension of the Equatorial Pacific Dry Zone. DeVries (1987) has reviewed the geological evidence for the occurrence of ancient El Nino events in Peru. He showed that the evidence - both terrestrial and marine - for El Nino activity during the Late Pleistocene is inconclusive. Similarly, Salinger (1984) questioned the validity of the inferences drawn by Quinn (1971). 43 18.000 y««r»BP »—• 10.000 yem BP FIGURE 3.5 Approximate location of shoreline at 18 Ka BP and 10 Ka BP (after Woodroffe, etal., 1986). 44 Another factor which can be identified as contributing to the existence of widespread aridity in northern Australia during the LGM is reduced tropical cyclone genesis. The general controls of tropical cyclone genesis have been summarised by Gray (1975, 1988) in his seasonal Tropical Cyclone Genesis Parameter. This is controlled by the Coriolis parameter, low-level relative vorticity, the magnitude of the vertical shear of the horizontal wind between the lower and upper troposphere, sea surface temperature, the vertical gradient of equivalent potential temperature between the surface and 500 mb and middle troposphere relative humidity. Even from general considerations, it is clear that at the LGM many of these parameters would have changed so as to reduce the incidence of tropical cyclones (Wyrwoll and Milton, 1976; Hobgood and Cerveny, 1968). Another consideration is the effect that the more extensive land area would have had on cyclone formation and path. The controls of regional climates and the stratigraphic evidence, all point to northern Australia being drier during full glacial stages. However, the estimates of mean annual precipitation at the LGM given in Figure 3.4 must be treated with caution. The general conclusion of drier conditions prevailing in the Australian tropics during the LGM, corresponds with our present understanding of global tropical climates during full glacial stages. At the LGM the tropics generally saw a significant reduction in mean annual precipitation (see Crowley and North, 1991). The other factor of significance to water balance consideration and palaeohydrology in general, is temperature. Unfortunately, it is more difficult to obtain indications/estimates of temperature differences for northern Australia at the LGM. It is obvious that mean annual temperatures over northern Australia at the time of the LGM can be expected to be lower than present. Evidence from the New Guinea Highlands (Bowler et al., 1976) suggests that at the LGM mean annual temperatures were some 6°C lower than present, while estimates of sea surface temperatures show differences of little more than 2°C. (CLIMAP Project Members, 1976). A discussion of the implications of these estimates is provided by Webster and Streten (1978, 1981) Prell et al. (1980), and Prell (1982). Webster and Streten (1978, 1981) considered frequent and short-lived cold air incursions into the New Guinea highlands as a partial explanation for the differences in the estimates. Hastenrath (1988) noted that in the light of plausibly different lapse rates in dry as compared to moist climatic regions, it appears difficult to meaningfully compare temperature around the 5000 m level with sea surface conditions. Moreover, in the light of the widely spaced coring stations and other factors, the sea surface temperature reconstructions from deep-sea cores may be open to revision. This leaves the problem of estimates of temperature changes in the Koongarra region open to doubt. Using evidence from other parts of the tropics, a reduction in the mean annual temperature of 3°C - 5°C is possible (Crowley and North, 1991). However, it still remains difficult to reconcile the observation that during the LGM the snowline of the tropical mountains fell by some 950 m, while tropical sea surface temperatures fell only 1°C - 2°C (the most recent discussion of this is provided by Betts and Ridgway, 1992). 45 Because of the inherent problem of resolution and accuracy the available numerical simulations of the atmospheric circulation during the LGM are only of limited help in establishing possible temperature changes. The results of two simulation experiments - one general circulation model (Lautenschlager, 1991) and one energy balance model (Hyde et a!., 1989) - are shown in Figures 3.6 and 3.7. In control runs the Lautenschlager model produces tropical temperatures which are 2°C-3°C too cold. Nevertheless, the relative depression of the LGM temperature estimates given by this model are as good as any others and give convenient "working" estimates. It is concluded that the climates of the Koongarra area during full glacial stages were characterised by significant decreases in precipitation and lower temperatures. Estimates of likely relative reduction in the mean annual precipitation of the Koongarra area over the last c. 100 Ka can be gauged from a comparison with the data provided by Kershaw (Figure 3.1). 3.3.2 Interglacial stages The synoptic-scale and wider characteristics and controls of the present interglacial climate provide the background against which the dynamics of past interglacials can be evaluated. It follows that interglacial climates need to be seen in the context of the development of monsoon activity, which is much stronger than during glacial stages. However, the Holocene does not provide a complete climatic analogue for interglacial stages. It is clear from the evidence of lacustrine environments in northern Australia, that at times during the Pleistocene, monsoon activity was considerably more intense than at any time in the Holocene. In recent years a great deal of attention has been given to the relationship between Quaternary insolation values of low latitudes and the intensity of the monsoon regime (Kutzbach, 1981; Kutzbach and Otto-Bliesner, 1982; Kutzbach and Guetter, 1986; Short and Mengel, 1986; Prell and Kutzbach, 1987; Kutzbach and Gallimore, 1988, Prell et al., 1991; Wyrwoll et al., 1992). The degree of success that has been achieved by the various computer simulation studies is well documented and the general model of a direct forcing of the monsoon circulation/intensity by insolation values resultant on the Milankovitch mechanism has now been widely accepted (Kutzbach and Street-Perrott, 1985; Kutzbach and Guetter, 1986; Prell and Kutzbach, 1987; Mitchell et al., 1988). However, it has now also become apparent that monsoon dynamics are more complex than being simply insolation driven (Clemens et al., 1991; Wyrwoll et al., 1992). 46 (a) 1BOW 150W 120W 90W 60W SOW 0 30E 60E 90E 120E 150E 180E 60N - O- 60N 30N - - SON 0 - - 0 30S - - 303 60S - -v - 60S i • T ii—i—|—i—i—|—i—i—|—i—i—|—i—i I I—i—|—i—i—|—i—i—|—r—i—|—i—i—|—i—r 1BOW 150W 120W 90W 60W 30W 0 30E 60E 90E 120E 150E 180E (b) 1BOW 150W 120W 90W 60W 30W 30E 60E 90E 120E 150E 180E 60N - •?- 60N 30N - SON 0 - - 0 30S - - 303 60S - ""-"-'-—^^-w^^-_;-.v^^^^^vv;- - -1S_ 60S _....-----••» •-•-.*"ji^^ - "~< 1 ' I ' ' 1BOW 150W 120W 90W 60W 30W 0 30E 60E 90E 120E 150E 1BOE FIGURE 3.6 18 Ka BP - Present for (a) January and (b) July temperature differences (after Lautenschlager, 1991) 47 (b) -...... : ...... 1.0 0 • ..-7.00 2^^^---...... -*- w...... " £ FIGURE 3.7 18 Ka BP - Present for (a) July and (b) January temperature differences (after Hyde et al.f 1989). 48 So far, the simulation work has emphasised increased monsoon activity in the northern hemisphere forced by the high insolation values received by low latitude areas at 9 Ka BP. A variety of field evidence was used to "verify" the results of the computer simulation studies (Street and Grove, 1979; Street-Perrott and Roberts, 1983; Rossignol-Strick, 1983; Prell, 1984; Fc.ntugne and Duplessy, 1986). The response of the southern hemisphere monsoon regime to insolation variations is seen as more equivocal. It is not clear whether more intense monsoon activity over northern Australia can be directly related to periods of high summer insolation over the low latitudes of the southern hemisphere, but some guide can be obtained from the Holocene record. At 9 Ka BP, the time when the low latitudes of the northern hemisphere saw an increase in insolation, the austral summer low latitudes experienced a significant decrease in the insolation received. Simulation runs by Kutzbach and Guetter (1986) and Mitchell et al. (1988) have highlighted these differences and their results suggest a decrease in southern hemisphere monsoon intensity at that time. However, it has been proposed that a simple forcing of the southern hemisphere monsoon regime by variations in low-latitude insolation variations has to be ruled out "...because the smaller size of the southern continents especially Australia, muted this effect" (COHMAP, 1988:1049). Despite this, some claims have been made (Wyrwoll et al., 1986 and 1992; Markgraf, 1989) that the southern hemisphere monsoon regime may have been less intense during the Early Holocene. For the northern Australian monsoon, this inference is based on stratigraphic findings from an area of northwestern Australia which is located at the margin of monsoon influence (Although, in the light of recent, as yet incomplete, field findings from this area the claim of Wyrwoll et al. (1986) may have to be revised). As already noted, a counter argument exists (Singh and Luly, 1991; Wasson and Donnelly, 1991) which sees the northern Australian monsoon essentially mimicking the behaviour of the northern hemisphere monsoon regime. It has to be realised that the controls of monsoon activity can be quite subtle. Some of the controls on monsoon activity could come from more local-scale effect leading to perturbation of the monsoon flow (Wyrwoll, et al., 1992). This is proposed with the results of McSride and Nicholls (1983) in mind, which showed that although the onset of the monsoon tended to be controlled by large scale atmospheric events, the interannual variability after the "break" appears mostly in higher principal components; i.e. due to smaller scale, local effects. In understanding the possible palaeoclimatic controls on monsoon activity we have to turn to a multitude of factors, in addition to the possible direct role that insolation variations can play. Most of these factors are poorly understood. For instance, the influence of outflow from the Asian landmass is a possible factor which could explain possible increased/reduced monsoon activity over northern Australia during glacial and interglacial stages. 49 The results of Mitchell et al.(1988) show for the Early Holocene, reduced temperature gradients in the extra- tropical regions of the northern hemisphere in winter. Consequently, one might expect fewer and/or less intense cold surges which form the "northern end" of the monsoon. Other candidates for large-scale controls are the ENSO mechanism and interaction with higher latitude synoptic systems but these likely controls are poorly understood even for the present day situation (Davidson et al., 1983; Holland, 1986; McBride, 1987; Hendon et al., 1989) and it is premature to cast these into a palaeoclimatic context. From this discussion, it is clear that the problems of identifying the details of the processes which force monsoon intensity are formidable. Despite this, a comparison of the Holocene and LGM circulation makes it clear that interglacials can be expected to be characterised by higher levels of monsoon activity than glacial stages. A conservative assumption is that past interglacial stages can be expected to have seen precipitation levels of similar magnitude to those of the present. 3.3.3 Summary of the palaeoclimatic findings An understanding of Tertiary climate events is very incomplete. From present knowledge it is clear that there were climates during the middle and late Tertiary which would have led to more intense chemical weathering than at present; but such weathering events may also have taken place during the Pleistocene. In reconstructing the climates of the last 700 Ka the most parsimonious approach that can be taken is to use the Middle-Late Quaternary record as a guide. Following this, and using the Lynch's Crater record, a "working" palaeoclimate framework for this time period consists of- (i) drier glacial periods lasting for the best part of 100 Ka, with extreme dryness prevailing at glacial maxima; and (ii) interglacials lasting of the order of 10 Ka during which precipitation levels were similar to those which prevail today. The sequence of palaeoclimatic events for the last 700 Ka is an idealised scheme based on the direct hindcasting of the Late Quaternary record. The problems and inconsistencies inherent in this should be obvious from the earlier discussions. Nevertheless, it is a scheme which has merit for palaeohydrological modelling. The sequence can be used to set the limits of the climatic boundary conditions for the last c. 1 Ma. The two climate states can be set as end members and the effect of variable durations and intensities of these climate scenarios could be modelled in a random variate context. For such an exercise to have any physical meaning the sensitivity of hydrological and geochemical processes to climate variations has to be understood. 50 4 DENUDATIONAL REGIME AND RATES OF EROSIONAL PROCESSES Given the rates of geomorphological processes in the area (see below), the explanation of the timing of the development of the secondary dispersion fan must be linked to the events acting over a timescale of 10 Ma. Whatever the sequence of long-term geomorphological events (timescales of the order of 100 Ma), the potential for the formation of a secondary dispersion fan, necessitates that the orebody was located in an oxidising weathering environment. Hence the immediate problem of identifying the geomorphological controls on the development of the secondary dispersion fan, has to be seen in the context of Late Cenozoic erosion-weathering events. In this context, the possible role of the Brockman Massif escarpment must also be considered. Any discussion of the details of Late Cenozoic geomorphological events and evolution of the area is constrained by the currently limited work on weathering and erosion processes, and poor stratigraphic control. Recent work from adjacent areas has, however, now become available and this allows some inference to be drawn on rates of denudation. 4.1 The Koongarra Escarpment The escarpment of the Brockman Massif occurs immediately adjacent to the orebody. The juxtaposition of the two has, in previous discussion of the geomorphological controls on the development of the secondary dispersion fan, prompted the suggestion that the denudational history of the escarpment may have played a role in triggering the development of the fan (Roberts, 1989; Snelling, 1989). In addition the character of the footslope deposits fringing the escarpment has been seen to indicate escarpment stability since the Late Cretaceous. The morphological characteristics of the escarpment bounding the Brockman Massif are quite variable, reflecting differences in rock mass geometry and wider aspects of the structural geology. In the Koongarra area the most pronounced difference is between the low bounding escarpment of the northern sections (the area approaching Koongarra Saddle) , which rises into the massif via a series of structurally controlled "steps" and the escarpment towards the south (towards Nourlangie Rock), where the structural and lithological controls have led to a well defined high, near vertical scarp face. Of note is the fact that the morphology and erosional character of the Koongarra escarpment is significantly different from that of the Arnhem Land escarpment, a difference which again reflects the different regional-scale structural geology and more local rock mass characteristics. 51 4.1.1 Escarpment denudational processes Escarpment erosion is largely controlled by mass failure processes at a variety of scales, while chemical weathering is only of secondary importance. The lithology of the Kombolgie Sandstone making up the escarpment is variable. The lower parts of the escarpment have developed in a well- sorted, medium to coarse sandstone, which in places is very friable. The sandstone is well bedded, with ripple marks widely evident on bedding planes. In parts of the upper escarpment the sandstone is customarily more strongly cemented, more reminiscent of an orthoquarzite and more massively bedded. A similar distinction was made by Mills (1991) who recognised a lower apron "soft" sandstone and the upper part of the escarpment consisting of "hard" sandstone. The degree of chemical weathering is variable, ranging from essentially fresh rock with no indications of weathering, to material which has undergone almost total granular disintegration. Chemical weathering processes have been more active on the rocks of the lower parts of the escarpment. Case hardening is widely evident on rock surfaces. The differences in lithology and degree of chemical weathering between the rocks of the upper and lower parts of the escarpment are reflected in the intact strength of the rocks. Intact rock strength was established from field tests using a Schmidt rebound hammer. Ten readings were taken at each site. The mean values of the 10 readings obtained at the various sites (total number of sites = 9) ranged from 63 to 18. The higher rebound numbers were obtained on the more strongly cemented, massive, pale coloured members of the upper parts of the escarpment. The lowest rebound numbers were obtained on weathered, friable, ferruginous sandstone, which is extensively fractured and associated with shear zones. The Schmidt Hammer rebound values are indicative of the unconfined compressive strength of the rock (Augustinus, 1991) although some care has to be taken in their interpretation (Johnson and DeGraff, 1988). Nevertheless, they are a good general indicator of rock strength and degree of chemical weathering. Rebound values of 10-35 correspond to very weak rocks which are extensively weathered, while values of 50-60 are indicative of strong, competent rocks (Selby, 1982:65). In terms of the BGD classification (Int. Soc. for Rock Mech., 1978) the two weathering classes correspond to categories W3 (moderately weathered) and W1 (fresh). The rock mass of the escarpment is strongly jointed and "relaxed" with prominent open joints and other open fracture planes. The details of the joint systems vary along the Koongarra Escarpment and it is these differences which control the character and erosional processes of the escarpment. In the area adjacent to the orebody, bedding joints dip out of the slope at low angles (6°-10^. This joint system taken in conjunction with near vertical cross joints, divide the rock mass into well defined joint blocks. In fact, in places the rock mass is so :'open " that it is now no more than an amalgam of loose, stacked joint blocks. This is especially the case in the lower parts of the escarpment, where the low angle bedding plane 52 joints have a spacing of 30 cm to 50 cm, with the near vertical cross-joints having a spacing of 80 cm - 100 cm. The spacing of these joint sets, which characterise the base of the escarpment, can vary significantly, producing both larger and smaller blocks. For instance, in some areas of the lower part of the escarpment, mean cross-joint spacing is less than 30 cm. In areas located near fault zones, the joint and fracture pattern is complex, irregular, closely spaced and the rock is highly weathered. In these areas weathering of the rock mass yields decimetre sized clasts. The rock mass of the upper part of the escarpment is characterised by larger joint-rock blocks, some reaching dimensions of up to 6 m to 8 m; these are separated by very open joints and fracture planes, which have clearly been enlarged through weathering. The joint surfaces are not lined with weathered infills, although smooth silica films lining the joint surfaces indicate that these planes act as major joint water conduits. In places the joint pattern and fracture pattern of the upper part of the escarpment have resulted in wedge failure - where large blocks have daylighted out of the slope. There are also indications in the upper part of the escarpment of plane failures along isolated failure surfaces which dip out of the slope at quite high angles. Toppling failure (Goodman and Bray, 1977) of large blocks is generally prevalent along the escarpment of the Brockman Massif. The trigger for rock slope failure has to be seen in the context of :'aging" and general relaxation of the rock mass, coupled with external triggers such as the development of high joint water pressures. More irregular and smaller scale rock and debris failure is widespread, and on parts of the escarpment, they are the dominant mode of rock mass failure. In terms of the Geomorphic Rock Mass Strength Classification and Ratings developed by Selby (1980) and based on earlier classifications (e.g. Bieniawski, 1973), the rock mass of the Koongarra escarpment has a rating range of the order of 81 (strong) - 43 (weak). This criterion gives an indication of the nature of the slope profile and the susceptibility of a rock slope to mass failure processes and weathering. 4.1.2 Foot slope deposits The nature of the footslope deposits that have accumulated at the base of the escarpment is a direct reflection of failure mode, rock mass characteristics, weathering and the geometry of the escarpment. Depending on these factors, footslope deposits in the area can either take the form of talus (consisting of a range of block sizes) or taluvial-colluvial accumulations. The sand apron which has developed at the base of the escarpment (Figure 4.1) is of significance to the geomorphological history of the orebody. The formation of this apron is linked to the chemical weathering and granular disintegration of the Kombolgie Sandstone. The supply of sediments to the sand apron is through the direct transport by wash processes of sand-textured material weathered out of the Kombolgie Sandstone of the Brockman Massif. 53 2S 5000E 857 7000N 857 3000N CO 265000E 7857 3000N Sand apron and Cahill Formation alluvial unds Amphlbolite Kombolgie Formation Schist. Quartz Mica Schist Massive quartz Biotite Schist, Chlorite Schist sandstone, conglomerate FIGURE 4.1 Extent of the sand apron at Koongarra (after McMahon, Burgess and Yeates, 1981) 54 A second source of sand is from the weathering and granular disintegration of talus clasts. Because of the morphology, topography and the nature of the mass failure processes of the escarpment, the run-out distance of material is restricted to the foot of the escarpment. Consequently, there is a significant volume of material, deposited in a restricted area at the base of the escarpment, from which sand-sized sediments can develop. 4.1.3 Rates of escarpment retreat Indications of the mechanics of scarp failure, the manner and variation of failure mode and weathering characteristics can be ascertained in a relatively straight forward manner. However, it is not possible to convert the presently available data into estimates of rates of scarp retreat. Rates of escarpment erosion are clearly variable and are controlled by geomorphological position and rock mass characteristics. This is clearly reflected in the position of the escarpment with respect to the fault-zone (Figure 4.2). Galloway (1976) proposed the extraordinarily high rate of escarpment retreat of 1 m/Ka. This estimate was arrived at without any consideration of failure processes or the denudational history of the escarpment. Even from general geomorphological considerations, it is clear that the proposed rate of escarpment retreat is much too high, possibly by orders of magnitude. The most direct field indication that a rate of 1 m/Ka is unlikely is given by the presence of dated archaeological sites at the foot of the Kombolgie Escarpment (Roberts et al., 1990) and the widespread occurrence of aboriginal paintings on the rock faces of the Koongarra Escarpment. An estimate of escarpment retreat of 1m/Ka is comparable with rates in alpine upland areas - which have some of the highest established rates of cliff face recession, with common values being around 0.7 m/Ka (Caine, 1974; Barsch, 1977). In areas of less active erosion, rates of escarpment retreat can be much lower. Young and McDougall (1985), for instance, have obtained rates of scarp retreat of between 12 m and 28 m/Ma for areas in the Eastern Highlands of Australia. Useful estimates of escarpment retreat come from the work of Roberts et al. (1991) who obtained overall weathering rates/sand apron formation consistent with a rate of scarp retreat ranging from 20 to 200 m/Ma. This range is indicative of variations in rockmass characteristics, lithology and groundwater-joint water pressure conditions. Such differences in the rates of escarpment retreat are clearly evident in the Koongarra area (Figure 4.2). A series of 10Be ages were obtained on joint-planes on the escarpment behind Koongarra (Klein, pers. comm.). The ages obtained are consistent with a maximum rate of rock surface weathering of 2.2 mm/Ka. The average exposure age for the surfaces is of the order of 220 Ka BP. These data are consistent with the claim of very low rates of escarpment denudation. 4.1.4 The characteristics of escarpment retreat The question remaining is the role of escarpment retreat in the development of the secondary dispersion fan. The Koongarra Escarpment is a fault-line escarpment. As the Koongarra Fault is a reverse fault, the escarpment is not maintained simply by retreat (Rgure 4.3). From the work of Roberts et al. (1991), it is clear that the differences in the mean denudation rates of the Kombolgie Formation and the Cahill Formation are significant (see below). These differences in the rate of surface lowering result in the fault-zone essentially "forcing" the position of the escarpment - i.e. forming a fault-line escarpment. This results from the fact that, given the structural geology and the differences in the rate of surface lowering, rather than simply retreating, the escarpment is being "exhumed" along the fault (Figure 4.3). This is quite different from other structural settings, such as the opposing Arnhem Land Escarpment, where the escarpment once formed, simply retreated into the rock mass (Figure 4.3). It is clear that the escarpment at Koongarra has not simply retreated "into" the Brockman Massif and through this, placing the orebody into a weathering environment. If the escarpment has affected the development of the secondary dispersion fan, it would have been more indirectly through the hydrological influences that the Kombolgie terrain may have as a "feeder" of surface and shallow subsurface water. 4.2 Denudation Rates and the Erosion History of Koongarra Valley The most direct geomorphological controls on the development of the secondary dispersion fan are the site-specific denudation rates. These will control the timing of the "arrival" of the orebody in a geochemical environment suitable for radionuclide mobilisation. A guide to the depth below the surface that has to be attained before a secondary dispersion fan can develop is given by the depth of the weathered zone of No. 1 orebody. For No. 1 orebody, the depth of weathering and radionuclide dispersion extends to c. 25 m. (Snelling, Volume 2 of this series). This suggests that under the climate-hydrological conditions of the Late Quaternary an orebody has to be situated at a depth of c. 30 m before a secondary dispersion fan can develop. 4.2.1 The nature of the denudation regime and regional denudation rates A convenient way of discussing the denudational regimes of an area is to see these as being either weathering or transport limited (Carson and Kirkby, 1971; Slaymaker, 1988). In the case of weathering-limited regimes the rate of potential material removal is well in excess of material supply through weathering. These regimes often prevail in areas where soils are thin and relief differences significant. In contrast, in transport-limited regimes soils are thick, relief subdued, erosion rates are low, and soils and solid weathering products are cation poor (Stallard 1985, 1988). 59 Kombolgie Sandstone Kombolgie Sandstone Fault Cahill Formation Sands a Gravels FIGURE 4.3 Schematic representation of the Kombolgie Sandstone escarpments indicating how the differences in the structural geology determine the nature of escarpment evolution: (a) typifies the escarpment setting along the western margins of the Brockman Massif, (b) applies to the escarpment at Koongarra (modified from Williams, 1991). The general geomorphological expression of the Koongarra area is one which indicates an environment in which erosion rates are low. In outcrop, the Cahill Schists are generally well weathered. The weathered profile has depths in excess of 20 m. In surface exposures, the degree of weathering of the schists is variable. It ranges from alteration of the rock to yellow coloured, soft, saprolite in which relict schistosity can be still clearly distinguished, to more advanced stages of weathering where the original schist has weathered into a pale, plastic clay containing small flakes of mica and isolated larger mica "book" remnants. A selection of representative shallow profiles describing the characteristics of the regolith is given in Figure 4.4. Tne intensity of weathering in the weathering profiles is evident in the clay mineral assemblage present. Koppi et al. (1992) and Gray (1986) have shown the dominance throughout much of the weathered profile of the low-activity clay mineral kaolinite. Kaolinite is the product of the weathering of muscovite, biotite and chlorite and is present as pseudomorphs of these minerals. The intensity of weathering of the schists is also reflected in the geotechnical characteristics of the regolith materials. The Atterberg Limits of a selection of weathered schist samples are summarised in Figure 4.5. The shear strength of the weathered Cahill Schists similarly indicate a high degree of chemical weathering and a high susceptibility to surface erosion processes. The results of triaxial tests undertaken on weathered schist of the Cahill Formation are given in Figure 4.6. It should be noted that these tests were undertaken on remoulded samples. Over the range of normal stress used the upper bound for these data are q = 2c' tan(45 + The general field characteristics and the mineralogical and geotechnical characteristics of the regolith clearly demonstrate the existence of a transport-limited denudation regime on the Koolpinyah Surface. This inference is consistent with general field indications. It is also reflected in the major ions of the solute transport load (Hart et al., 1987a) and the mineralogical and elemental composition of the suspended load in Magela Creek (Hart and Beckett, 1986). The explanation for the transport-limited denudational regime of the region lies in the restricted relief variation/erosion potential of the Koolpinyah Surface, possible hydraulic controls of surface erosional processes (Roberts et al., 1991), the density 61 .c Soil or Rock Type, Structure 8- Soil or Rock Type, Structure 0 IS o S or o 5? I GRAVELLY SILTY SAND, grey brown, organic matter. GRAVELLY SAND, grey brown, fine to medium roots, sand fine to medium. Loose. in grained, dry to moist. Loose. As above, yellow brown, with roots, moist. GRAVELLY SILTY SAND, yellow brown, fine to •'. .,•*". medium grained, plant roots, moist. Medium dense. Medium dense. :o ' GRAVELLY CLAYEY SAND, yellow brown, red brown „ mottled, medium grained. Gravel is rounded laterile pisolites to 10mm and angular quartz to 20mm. Roots, CLAYEY SAND, red brown yellow mottled, trace fine to moist. Medium dense. medium gravel (laterised mica schist fragments to v.-V 5mm.) Sand fino grained, some gravel as shown. CLAYEY GRAVEL, red brown, medium to coarse Dense. X; 1.0 1.0 • grained quartz gravel to 25mm. Clay is high plasticity / * interstitial, cavities, voids and occasional plant roots. 0/0 Medium dense to loose. •o ° y CLAY, red brown, high plasticity, grading into extremely weathered amphibolite (SILT of high liquid limit • grey green), moist. Occasional quartz pebbles. Undulose polished joints and relict joint structures are present In tO)o Transitional amphibolite. Very stiff to friable. 2.0 SILTY CLAYEY SAND, red brown and white mottled, fine grained, trace of fine gravel (laterite pisoliths) and 2.0- quartz gravel to 20mm., micaceous, moist. Relict schistose structure, plant roots with leached zones. Dense to very dense. 1 F\\ *» 2.8m. Back hoe digging slow!/ in Extremely Weathered 3.0- 3.0. AMPHIBOLITE. 3.6m. Maximum reach of backhoe. FIGURE 4.4 Summaries of logs describing the characteristics of the reoolith at Koonoarra /attar MnMahnn 10 ATTERBERG LIMITS Plastic Limit WP% C D Mean- 27 Liquid Limit WP% sample s S«6i s»6 Mean -49 s-22 a> CO 2 4-j 3 z 13 samples y////\ 40 60 so too Moisture Content % FIGURE 4.5 Atterberg Limits of weathered Cahill Schist samples (after McMahon, Burgess and Yeates, 1981) Upper bound <>'- 36.5° c' - 24.9 KPa 500 c' - 3.5 KPa Sample Identification O T30 2.7 - 3.9 + T63 13 Lower bound 100 200 300 400 500 600 700 800 900 1000 MEAN EFFECTIVE NORMAL STRESS P'f (kPa) FIGURE 4.6 Triaxial strength characteristics of weathered Cahill Schists (after McMahon, Burgess and Yeates, 1981). and character of the vegetation cover, regolith characteristics and regional base-level controls. At present there is an absence of data which would allow a systematic evaluation of the relative importance of these factors. The contemporary regional denudational rates can be calculated from sediment and solute yields. This approach follows standard procedures which are discussed extensively in the literature. The available sediment and solute transport data have been compiled by Roberts et al. (1991) for the adjacent Magela Creek catchment and their estimates are given in Table 4.1. The work of these authors is the most recent and comprehensive discussion of regional denudation rates in this area. The authors recognise that the load estimates are associated with significant errors. These are claimed to be of the order of +100%/-50%, but these error estimates are themselves open to doubt. The assumptions made by Roberts et al. (1991) in the interpretations of the solute and sediment transport data are that all washload and solutes are derived from the metasediments and soils of the Koolpinyah Surface. Bedload sediments are all attributed to the Kombolgie Formation. Field work in Koongarra Valley has demonstrated that streams draining small catchments, which are exclusively in the Cahill Formation, carry little bedload material. The assumption that the bedload sediments are derived from the Kombolgie Formation seems realistic. Furthermore, the mineralogical composition of the suspended load is dominated by kaolinite and quartz (Hart and Beckett, 1986). While it is not possible to establish the relative amounts, it is possible that the Kombolgie Formation may be the source of some of the suspended sediments (see Gray, 1986 - his Table, 3.2). However, this is likely to be of minor significance. Similarly, the contribution that the Kombolgie Formation could make to the solute load is likely to be a minor consideration given the overall imprecision inherent in the solute load estimates. For the Koolpinyah Surface the washload estimate amounted to a mean annual loss of 31.76 tonnes/km2/a. The washload estimates were obtained using sediment rating-curve methods. These methods are far from ideal (Rieger and Olive, 1988) and in an environment in which there is no stream flow for a significant part of the year, can lead to significant error. In a more limited study Hart et al., (1987b) used a "storm-loading" method to estimate washload. The differences in the estimates of washload obtained by the two techniques varied. For 1982-83 the two estimates were virtually identical, while for 1978-1979 the estimate of Roberts et al. (1991) was 26% (780 tonnes) higher. No other data are available for comparison. Solute load estimates must be evaluated in the context of differences in climate, geology, geological history and topography of the catchment. Therefore, any comparison between catchments/regions is often misleading. The estimate of solute transport of c. 9 tonnes/km2/a for the Koolpinyah Surface seems low when compared with some data from other "comparable" tropical regions (e.g. Dunn, 1978). The average world chemical denudation rate, after corrections for 65 TABLE 4.1 ANNUAL WASHLOAD, SOLUTE AND TOTAL TERRIGENOUS YIELDS Wet Washload Yield Solute Yield Probable Material Total Season Bed Yield Terrigenous Yield (tonnes) (%) (tonnes) (%) (tonnes) (%) (tonnes) 1971-72 3367 40 1164 14 3938 46 8469 1972-73 4024 45 1183 13 3796 42 9003 1973-74 8795 41 2536 12 9888 47 21219 1974-75 8279 47 2087 12 7236 41 17602 1975-76 13401 46 3160 11 12394 43 28955 1976-77 4582 40 1559 14 5280 46 11421 1977-78 5629 42 1725 13 5990 45 13344 1978-79 3723 39 1303 14 4540 47 9566 1979-80 7529 49 1769 11 6253 40 15551 1980-81 10040 53 2024 11 6747 36 18811 1981-82 3739 40 1309 14 4260 43 9308 1982-83 2468 42 836 14 2563 44 5867 1983-84 9471 50 2012 11 7421 39 18904 1984-85 4174 45 1242 14 3760 41 9176 1985-86 1879 40 710 15 2073 45 4662 1986-87 2110 39 687 13 2551 48 5348 1987-88 417 40 250 24 378 36 1045 1988-89 3581 41 1201 14 3882 45 8664 Total 97207 45 26758 12 92951 43 216915 Mean 5400 45 1487 12 5164 43 12051 (After Roberts et al., 1991) 66 atmospheric additions, amounts to c. 21 tonnes/km2/a (Berner and Berner, 1987) but variations are high. The global ratio of mechanical to chemical denudation is about 6:1. For the Koolpinyah Surface this ratio is at present c. 3.5:1. The estimates of solute loss for the Magela Catchment can be used as a crude guide to the rate of advance of the weathering front (the weathering front is defined simply as being the approximate boundary between saprolite (bulk density = 1600 kg/m3) and fresh rock (bulk density = 2600 kg/m3)). The assumption is made that 0.8 of the total solute load comes from bedrock-saprolite transformation (see the discussion in Dethier, 1983). From these idealised assumptions a rate of advance of the weathering front of c. 7 m/Ma is obtained. This rate can be compared with the results summarised by Nahon (1986) of work on rates of weathering of granitic rocks under lateritic conditions in the Ivory Coast (5-50 m/Ma and 14 m/Ma) and Chad (13.5 m/Ma). The weathering of ultramafic rocks in New Caledonia reached rates of 29-47 m/Ma, depending on the topographic position of the rock in the landscape. Working in the Alligator region, Airey et al. (1982) used an open system model to account for uranium series disequilibria associated with three ore deposits (Ranger One, Jabiluka One and Nabarlek) and from this propose a rate of advance of the weathering front which have a range of 10 m - 300 m/Ma. It would seem that the assumptions made by the study need to be evaluated. It is not claimed that the estimate of 7 m/Ma is an accurate rate of weathering front advance; it is a very simplified first estimate (see Velbel, 1983; Dethier, 1983) . But in light of the emphasis that has been placed in past studies on the deep weathering events of the Tertiary, and the implied absence of significant chemical weathering at present, the calculations illustrate that even given the relatively low rates of chemical weathering prevailing today, significant modifications in the weathered zone will have taken place since deep weathering in the middle Tertiary. This conclusion is further reinforced by the fact that the area is likely to have seen periods during the Pliocene and Pleistocene in which chemical weathering rates were significantly higher than during the present interglacial. Unfortunately, the numerical age and the duration of the Late Cenozoic events is not known. All this leads to the suggestion that much of the evidence for chemical weathering in this area need not be seen solely in the context of an intense, deep weathering event in the mid-Tertiary. Instead, it needs to be seen in the context of a transport-limited denudation regime which is likely to have prevailed in this area for much of the Late Cenozoic, and during which rates of chemical weathering have varied. Roberts et al. (1991) conclude thai the combined sediment and solute yield estimates convert to a rate of lowering of the Koolpinyah Surface of the order 30 mm/Ka (assuming a bulk density of c. 1450 kg/m3 for the Koolpinyah soils). This compares with a mean rate of surface lowering of the Kombolgie Formation of 4.5 mm/Ka. However, these estimates need to be revised. In obtaining the results, the assumption was made that the solute load can be directly related to surface 67 volume loss. It is more realistic to see solute load in terms of isovolumetric weathering and saprolite formation. If 0.8 of the total solute load is attributed to this process the rate of surface lowering of the Koolpinyah Surface is reduced to 22 mm/Ka. In addition the density changes that take place with solum formation have to be incorporated within estimated rates of surface lowering - this can be significant (Dethier, 1983) but the details are complex (e.g., Brimhall et al., 1991). Combined, it is likely that the sediment and solute loads convert to rates of surface lowering for the Koolpinyah Surface of the order of 20 mm/Ka. In the original data presented by Roberts et al. (1991), the washload transport rates alone convert to a rate of surface lowering of 22 mm/Ka. From suspended sediment and solute yield estimates Duggan (>]988) calculated denudation rates specifically for the Koongarra Creek catchment. Mean rates of surface erosion of 25-30 mm/Ka were calculated from very limited data obtained over two wet seasons (1981-1982 and 1982-1983). More generally Duggan estimated a mean denudation rate of 22 mm/Ka for the Koolpinyah Surface. The estimates of Hart et al. (1987b) for the Magela Catchment were significantly lower, amounting to 12 mm/Ka. However, this estimate needs to be doubled because the data were obtained during "abnormal" years (Roberts et al., 1991). Some of the other data available for this region will be discussed in Section 4.2.2. The likely range of the denudation rates of the region is comparable with the denudation rates calculated for other Australian catchments (Rieger and Olive, 1988). The lowland areas of the Amazon also have similar denudation rates (Stallard, 1985). In comparison, the average rates of reduction of continental elevation, based on combined mechanical and chemical denudation rates, is about 43 ml Ma (Summerfield, 1991). Table 4.2 gives an indication of the controls that climate and relief exert on denudation rates. From this table there are clear indications that the estimates obtained for the Koongarra region are comparable to those that can be expected for this type of geographical setting. In fully evaluating the denudational significance of the washload estimates, the sources of the material need to be established. It is not clear to what extent these load estimates have been influenced by anthropogenic disturbances. The area is now regularly burned and there may be other landuse changes which could be of importance. For instance, Williams (1984) noted significant disturbance of the Melaleuca stand of the Magela Creek floodplain. It is difficult to establish how rates of denudation have changed in this area over the Late Cenozoic. At present the only long term estimates of regional erosion rates come from Chappell (1985), who from stratigraphic data calculated a long term surface lowering rate of 55 ± 10 mm/Ka. 68 TABLE 4.2 ESTIMATED DENUDATION RATES FOR THE WORLD'S THIRTY-FIVE LARGEST DRAINAGE BASINS BASED ON SOLID AND SOLUTE TRANSPORT RATES DRAINAGE RUNOFF TOTAL MECHANICAL CHEMICAL* CHEMICAL AREA (mm ka. ') DENUDATION DENUDATION DENUDATION DENUDATION (13°kiri*) (mm ka.1) (mm ka.'1) mm ka. ") AS % OF TOTAL Amazon 6.15 1024 70 57 13 18 Zaire (Congo) 3.82 324 7 4 3 42 Mississippi 3.27 177 44 35 9 20 Nile 2.96 30 15 13 2 10 Parana (La Plata) 2.S3 166 19 14 5 28 Yenisei 2.58 217 9 2 7 80 Ob 2.50 154 7 2 5 70 Lena 2.43 206 11 2 9 81 Chiang Jiang (Yangtze) 1.94 464 133 96 37 28 Amur 1.85 175 13 10 3 22 Mackenzie 1.81 169 30 20 10 33 Volga 1.35 196 20 7 13 64 Niger 1.21 159 24 13 11 47 Zambezi 1.20 186 31 28 3 11 Nelson 1.15 96 - - - - Murray 1.06 21 13 11 2 18 St Lawrence 1.03 434 13 1 12 89 Orange 1.02 89 58 55 3 5 Orinoco 0.99 1111 91 78 13 14 Ganges 0.98 373 271 249 22 8 Indus 0.97 245 124 108 16 13 Tocantins 0.90 385 - - - - Chari 0.88 69 3 2 1 29 Yukon 0.84 232 37 27 10 28 Danube 0.81 254 47 31 16 35 Mekong 0.79 595 95 75 20 21 Huang He (Yellow) 0.77 63 529 518 11 2 Shaft -el-Arab 0.75 61 104 93 11 11 Rio Grande 0.67 5 9 6 3 38 Columbia 0.67 375 29 16 13 46 Kolyma 0.64 111 5 3 2 31 Colorado 0.64 31 84 78 6 7 Sao Francisco 0.60 151 - - - - Brahmaputra 0.58 1049 677 643 34 5 Dnepr 0.50 104 6 1 5 88 Allowance made for non-denudational component of solute loads. (After Meybeck, 1976) 69 During the Late Cenozoic denudation rates would have changed in response to climatic and other environmental controls. The relationship between climate and denudation rates is extremely complex. Figure 4.7 shows various estimates of the relationship between erosion and mean annual precipitation. In this diagram only two curves come "close" to the erosion estimate of 20 mm/Ka obtained for the Koolpinyah Surface. The sediment yield response to more arid conditions during the glacial stages of the Pleistocene is difficult to gauge. The only dated deposits which have some relevance as possible indicators of sediment yield processes during glacial stages are the sand aprons, however, they give no indication of any significant changes. A decrease in vegetation cover may have increased rates of erosion (Rodgers and Schumm, 1991). But there are no stratigraphic data available to test this inference. More certain is the fact that during drier phases, rates of chemical denudation would certainly have been reduced. Similarly, a more intense monsoon regime during past interglacials would have enhanced rates of chemical weathering, while possibly lowering erosion rates in response to a more dense vegetation cover. In past o.scussions some emphasis has been placed on basin response to the repeated base-level changes of the glacial stages of the Pleistocene and the resultant stream network response. The importance of base-level controls is emphasised by the relatively limited lowering of the landsurface that has taken place below the Cahill/Nourlangie-Kombolgie unconformity since the removal of the Kombolgie cover. A possible regional fluvial response to the lowering of baselevels during Pleistocene low sealevel stands is indicated by the "flushing" of channel sediments in Magela Creek (Nanson et al., 1990; Roberts et al., 1991). However, this was a relatively minor event, amounting to a scour depth of some 8-10 m. Similarly, for the Alligator River valley Wcodroffe et al. (1986) show a pre-Mid Holocene river valley which had a flat floor about 10-12 m below AMD (Australian Height Datum). The Late Pleistocene low sealevel was of the order of - 130 m (Fairbanks, 1989). With the prevailing shelf geometry off northern Australia, this means that the coastline was located some 500 km from the Koongarra region (Figure 3.5), yielding a regional channel slope of only 0.0026 - notwithstanding more local baselevel controls. Changes in base-level may well have provoked an alluvial response which could have been transmitted through the stream network. But changes in sediment discharge/water discharge relationships may have equally resulted in channel degradation. With these complexities in mind, and in the absence of stratigraphic control, it is simply not possible to advance any precise estimates of long term catchment denudation rates. An estimate of long-term erosion rates on the Koolpinyah Surface could be obtained if the actual rate of escarpment retreat could be determined at a siie where the height of the Cahill/Kombolgie unconformity was 70 1000 500 Wilson 200 E100 E £ 50 .9 \ Langbein and Schumm •o i 20 01 Ohmori O 10 500 1000 1500 2000 Mean Annual Precipitation (mm) FIGURE 4.7 The relationship between denudation rates and mean annual precipitation (after Ohmori, 1983). 71 known. At present there is simply no choice but to use the mean catchment estimates of contemporary denudation rates as guides to the events of the last few million years - the problems inherent in this are obvious. 4.2.2 Denudation rates at Koongarra Mean catchment erosion rates have to be seen in the context of the sediment sources and the delivery ratio which characterise a basin. Erosion rates can vary considerably within a catchment depending of the specifics of the geomorphology, hydrology and geology of a site. So that the g«3omorphological details of a specific part of the basin may result in rates which are quite different from the mean catchment erosion rate. In addition controls will change through time, so that, for instance, a sediment store under one set of environmental conditions may become a source as conditions change. The scale relations inherent in sediment yield estimates are clearly indicated in the summary of the regional sediment data given in Figure 4.8. The data shown in this figure are largely for suspended load but depending on the methods used contain some proportion of bedload (Wasson et al., 1990). The large error bars associated with these estimates should be noted. Using these data to establish a predictive relationship between basin size and sediment yield estimates would be misguided. The data of Williams (1973) are the result of work on erosion plots in the Brock Creek area over a limited 2-3 year period. Williams was concerned with establishing the relative importance of soil creep and wash processes in surface erosion. The results suggest mean rates of surface lowering of hillslopes of 54 ± 40 mm/Ka and 56 ± 30 mm/Ka by wash processes and 18 ± 12 mm/Ka and 11 ± 9 mm/Ka by soil creep. However, the range of these data need to be recognised. Wash processes erosion rates ranged 12 mm/Ka to 169 mm/Ka. Soil creep rates ranged from 0 to 16 cm3/cm/a. Other site specific erosion estimates come from Airey et al., (1982-1983). From their modelling work these authors provide estimates of rates of surface erosion which range from 20 mm/Ka to 900 mm/Ka. The estimates are thought to apply to the last c. 60 Ka. Both Wasson et al. (1990) and Roberts et al. (1991) summarise the data of Airey et al. (1982) and from these propose that the data suggests a mean rate of erosion of c. 40 mm/Ka. In fact, the statistics of these data do not allow such an interpretation. The site specific erosion estimates are high and do not correspond to a landscape which is characterised by a transport-limited denudation regime. The only indicator of long term erosion rates that is available is the height of the Kombolgie-Cahill/Nourlangie unconformity (Figure 2.10). This, linked with the proposed rates of escarpment retreat, make it clear that, over geologically significant periods of time, the landscape of the Koongarra area could not sustain the high rates of erosion suggested by much of the sediment yield/erosion data. 72 -i60 50 40 1 10, 11 30 g 0) 10 10-5 10"4 10'3 10'2 10'1 10° 101 102 103 104 Catchment area (km2) FIGURE 4.8 The relationship between catchment area and sediment yield in the Alligator Rivers region (after Wasson et al., 1990). 73 As an example, a catchment erosion rate of 28 m3/km2/a, as proposed for the Koolpinyah Surface, would result in the erosion of the cross-section of Koongarra Valley shown in Figure 4.9 in about 2 Ma. This does not correspond to our present general understanding of the geomorphological history of the region. It is especially difficult to reconcile the 10Be data which give low rates of escarpment retreat, and hence indications of the time required for the formation of Koongarra Valley, and the Koolpinyah Surface erosion rates suggested by the sediment yield information. The indications are clearly there that the long term erosion rates may be significantly lower than those inferred from the sediment discharge data. The orebody's position adjacent to the Koongarra Escarpment leads to the partial taluvium/sand apron cover which is a factor specific to the site. This must influence the rates of lowering of the Cahill surface. In addition recharge of the groundwater from runoff from the Brockman Massif creates a setting conducive to chemical weathering. It is to be expected that both factors exert a control on the denudation rates of the site. Although, the relatively small elevation differences between the area adjacent to Koongarra Creek (20 m), and the area of the orebody (24 m- 30 m) suggests that the presence of the sand apron does not greatly inhibit rates of surface lowering. The sand apron is both extensive and, in places, deep (Figure 4.10). But over the area of the orebody the surficial sand and taluvial apron is relatively thin, being only up to 3 m thick and for a significant part of the area as little as 0.5 m. The presence of this surficial cover clearly inhibits the removal of the weathered profile that has developed on the underlying Cahill Formation. Erosion is further influenced by ferricretes which have developed at the contact between the sand and the weathered profile of the Cahill lithologies, and by other ferricretes developed within the sand apron. The age structure of the sand apron at Koongarra is not known. The only information of the ages of sand aprons in the region comes from Roberts et al. (1991 - Figure 4.11 this volume) who have dated a number of sand aprons in the upper part of the Magela Creek catchment, using radiocarbon and thermoluminescence (TL) techniques. From the dates obtained it seems that sand aprons are active for long periods of time. But it is not clear how their origin and development relates to general erosional-geomorphological controls. The processes responsible for controlling the dynamics of sand apron development through time are not understood. The suggestion has been made that the contemporary sand aprons are being mainly eroded by rain-impacted sheetflow (Roberts et al., 1991), a mechanism which has been adopted from the work of Coventry et al. (1988) on low-angle slopes in northern Queensland. The applicability of this model to the sand apron of the Arnhem Land escarpment has yet to be demonstrated. At Koongarra, the variations in the thickness and geometry of the sand apron, its high permeability, the relatively low hydraulic conductivity of the top of the underlying weathered profile and high recharge volumes make saturated overland 74 Arnhem Land Escarpment Mount Brockman Massif ^o> •^x-"*» FIGURE 4.9 The height of the Cahill Surface across Koongarra Valley. The height of the Cahill Surface along the Arnhem Land Escarpment is based on the barometric data given by Needham (1988) - see Figure 2.11 of this volume. 75 flow on the apron possible. As a result surface wash erosive processes will act for part of the wet season. Gullies are evident on parts of the sand apron, but are largely associated with small streams emerging out of the Kombolgie escarpment. There are no estimates of how much material is translocated over the sand apron in a given period of time. The erosive mechanisms and transport controls which have been inferred for sand aprons are difficult to reconcile with the established TL age-structure of the sand aprons obtained by Roberts et al. (1991). At its most general, TL dating utilises the luminescence of radiation defects in a crystal lattice (Aitken, 1985). In sand aprons quartz grains are utilised and it is assumed that the "grain" has been effectively zeroed through exposure to sunlight. At Koongarra the sand apron develops both from the direct transport of sand to the base of the escarpment and from the granular disintegration of talus material. The latter material will not have been "zeroed" and any TL date obtained from the sand apron has the potential to be spurious. Similarly, any grain transported downslope by surface wash processes would be exposed to sunlight, and some bleaching of the TL signal could occur. These reservations pose difficulties when trying to use the TL age structure of the sand aprons to gain an insight into the dynamics which control their development. A lack of understanding of the erosive controls on the sand apron and the absence of data on rates of sediment input and loss from the sand apron make it difficult to attempt to model the conditions under which the sand sheet can persist. However, it is clear that Koongarra Creek, by regulating the extent and volume of the sand apron, is an important control on sand apron persistence. The creek transports a large amount of bedload, as is evident from channel bedforms and the accumulation of sediments behind channel obstructions; significant volumes of alluvial sediments are also stored in the flood plain of the creek. The sedimentological work that has been carried out on the lower parts of the sand apron and the general geomorphology of the area make it clear that the alluvial deposits are made up of fluvially reworked sand apron sediments. The general presence or absence of the sand apron can be described in terms of simple mass balance considerations:. dS/dt = I-Q The change in storage (S) through time is controlled by the addition of sediment (I) through the weathering of talus clasts and the transport of sand from the massif and the losses (Q) which are determined by the transport rate down the slope of the sand apron and the rate of removal by the creek. Most simply the sand apron will be lost through (i) reduction in material supply to the sand apron and (ii) higher rate of removal/lower residence time of the material incorporated into the sand apron. The details of the geomorphological processes which could achieve these states are complex and there is a great deal of 76 8S76SOON (576600X SECT I OKI B57C400N 77 FIGURE 4.10 Thickness of the sand apron at Koongarra (after McMahon, Burgess and Yeates, 1981). Sand apron thickness (metres) ~i I SECTION 2 equ'ifinality in the argument; there are also other more local controls which are of importance. These are emphasised by the fact that a sand apron is absent from the areas to the northeast of No. 2 orebody (Figure 4.11) and the very thin sand apron cover over No. 1 orebody. At this stage in our understanding of the geomorphology of the orebody, there are insufficient data to establish specific erosion rates of the site. It would be possible to determine these by a long term, detailed study of erosional processes coupled to a very comprehensive dating program of the weathered profile, the Quaternary depositional sequences and soil forming/erosion rates. AH this can now be achieved in a relatively routine manner. In the absence of such information, the option exists to specify a conservative range of erosion rates based on known catchment rates and general geomorphological considerations and evaluate the development of the dispersion fan within this context. Given the available catchment sediment data, the erosion data and the wider geomorphological considerations outlined, an estimate of surface erosion of 30 m/Ma over the Koongarra area is clearly high. A rate of surface lowering of 5 m/Ma over the area of ths orebody is possibly a conservative figure. These estimates give a range of 1-6 million years in which a lowering of the landsurface over the orebody of 30 m could have been achieved. This tentative approach to the reconstruction of denudational events emphasises the fact that the development of the secondary dispersion fan can be adequately explained within the context of the Late Quaternary geomorphological regime of the area. There is no need to take recourse to some triggering event, such as a removal of a laterite "cover". The very nature of the denudational regime at Koongarra required that the orebody would "arrive" in a shallow surface location some time within the last few million years. The relationship between the rate of surface lowering and changes in the position and rate of advance of the weathering front has so far not been considered. The fact that the geomorphology of the region is characterised by a transport-limited denudation regime clearly indicates that the rates of weathering front advance were (are?) in excess of the rate of surface lowering. A partial explanation for this may simply be that estimates of erosion of the Koolpinyah Surface are too high. However, as already noted, there are indications of higher rates of chemical weathering even since the Pliocene. The depth of the present weathered profile is simply an integral of these events. If this view is accepted it would imply that the secondary dispersion fan could have been activated during times of deeper weathering and exhumed with surface lowering; even given this argument the fan can still date from the last few million years. Given the denudational setting of the orebody and even with erosion rates as low as 5 m/Ma it is unlikely that the orebody would have been in a geochemical environment suitable for development of a secondary dispersion fan before that time. 79 56 52 - Sand Apron 2 Sand Apron 3 48- 44 - 2.7 ± 2.0 125.5 ± 23.6 40- 157.2 ± 29.2 o 209.7 t 39.8 •c 232.0 i 43.0 53 36- 19.4 i 4.6 32- 663 ± 12.6 Magcla Crcck 64.2 1 12.4 oo cross-section 7 o 220.6 ± 42.0 129.8 ± 25.1 28- 34,4 ± 125.8 1 25.0 24 - Sandstone bedrock Sandstone bedrock 3.6 ± 0.2- 20 - TL ages are expressed in kyr and the I4C age (marked *) is expressed in kyr DP 200 400 600 800 1000 1200 1000 800 600 400 200 DiMajicc from ilic escarpment foot (m) FIGURE 4.11 Age structure of a sand apron deposit in the Magela Creek catchment (after Roberts et al., 1991). 5 CONCLUDING COMMENTS This volume has attempted to outline the current understanding of the geomorphological and palaeoclimatic controls which are likely to have played a role in the development of the secondary dispersion fan. The geomorphological and palaeoclimatic data which are presently available are tentative and only provide an incomplete outline of events. A major problem which remains is obtaining precise measures of rates of erosion and weathering - those proposed in this report are poorly constrained and can be treated only as "first estimates". The subaerial denudation history of Koongarra spans the Phanerozoic. From the structural geology it is clear that in that time a significant cover of Kombolgie Sandstone had to be removed to expose the Cahill Surface. The spatial arrangement of large scale structural lineaments and the regional joint and fault patterns clearly facilitated the denudation of the rock mass, over what is now Koongarra Valley. The outcome of this was the separation of the Brockman Massif from the Arnhem Land Escarpment. Once the Kombolgie Sandstone cover was removed the exposed Cahill Surface was lowered by erosion. The present topographic position of the Kombolgie/Cahill unconformity along the western margins demonstrates that the Cahill Surface must have been lowered by more than 100 m. The actual amount of lowering of the Cahill Surface at Koongarra cannot be established because the vertical displacement of the reverse fault is not known. Through this lowering of the Cahill Surface the orebody was "placed" into a weathering environment which could lead to the development of the secondary dispersion fan. From the possible range of surface lowering/erosion rates that was established, it seems likely that the arrival of the orebody at a depth suitable for the development of the dispersion fan occurred some time in the last 1 -6 Ma. It is stressed that this is a very tentative estimate. The estimate is derived from sediment yield data and more general geomorphological considerations. The lack of accuracy is compounded by the incomplete understanding of how rates of weathering front advance and erosion could have varied through time. Most important in this respect is our tentative knowledge of the timing and character of Late Cenozoic deep weathering events. During the time in which the orebody was located in a "weathering environment" climates and hydrology would have varied significantly. Again the details of Late Cenozoic climatic states remain largely unknown. But some guide for much of the Pleistocene can be obtained by using the climatic conditions of the Late Quaternary interglacials and glacials. This leads to the suggestion of dry glacial stages lasting for some 100 Ka, and shorter wet interglacials lasting for some 10 Ka. 81 In final summary - given the estimate of the range of denudation rates over the last few millions years and the wider geomorphological considerations discussed, it is likely that it was only in the last 1-6 Ma that the orebody came into a geochemical environment suitable for the development of a secondary dispersion fan. During this time the climate of the area would have changed profoundly. The regional hydrological changes which resulted must be incorporated into an understanding of the processes which controlled the development of the secondary dispersion fan. The suggested range of development of 1 -6 Ma is considered a "safe" range given the present data. With better stratigraphic control and a more thorough understanding of the geomorphological and weathering processes operating in the area today this range could be considerably refined. 6 REFERENCES Airey, P.L (1982) Radionuclide migration around uranium ore bodies - Analogue of radioactive waste repositories. Australian Atomic Energy Commission Research Establishment, Lucas Heights Research Laboratories. Annual Report 1981 -82. Aitken, M.J. (1985) Thermoluminescence dating. Academic Press, London. Aldrick, J.M. (1976) Description of soil families. In Story, R., Galloway, R.W., McAlpine, J.R., Aldrick, J.M. and Williams, M.A.J. (eds.) Lands of the Alligator Rivers Area, Northern Territory. CSIRO Land Research Series, No.38, 140-164. Allan, R.J. (1983) Monsoon and teleconnection variability over Australasia during the southern hemisphere summers of 1973-1977. Monthly Weather Review, 111, 113-142. Allan, R.J. (1985) The Australasian summer monsoon, teleconnections and flooding in the Lake Eyre basin. South Australian Geographical Papers No. 2, Royal Geographical Society of Australasia (South Australian Branch). Allan, R.J. (1988) El Nino Southern Oscillation influences in the Australasian region. Progress in Physical Geography, 12, 313-348. Anon. (1961) Climatology survey. Region 1: Darwin-Katherine, Northern Territory. Bureau of Meteorology, Melbourne. Augustinus, P. (1991) Rock resistance to erosion: some further considerations. Earth Surface Processes and Landforms, 16,563-569. Australian Bureau of Meteorology (1978) Australian Tropical Cyclone Forecasting Manual (Preliminary Edition). Department of Science, Australia. Australian National Parks and Wildlife Service (1980) Kakadu National Park plan of management. Australian Government Publishing Service, Canberra. 82 Barsch, D. (1977) Eine Abschaetzung von Schuttproduktion und Schutttransport im Bereich aktiver Blockgletscher der Schweizer Alpen. Zeitschrift fuer Geomorphologie, Supplementband, 28,148-160. Bemer, E.K. and Berner, R.A. (1987) The global water cycle: Geochemistry and environment. Prentice-Hall. Betts, A.K. and Ridgway, W. (1992) Tropical boundary layer equilibrium in the Last Ice Age. Journal of Geophysical Research, 97, 2529-2534. Biekart, J.W. (1989) The distribution of calcareous nannoplankton in Late Quaternary sediments collected by the Snellius II expedition in some southeast Indonesian basins. Proc. K. Ned. Akad. Wet., Ser. B, 92, 77-141. Bieniawski, Z.T. (1973) Engineering classification of jointed rock masses. Transaction of the South African Institution of Civil Engineers, 15, 335-344. Bird, M.I., Chivas, A.R. and Andrew, A.S. (1989) A stable isotope study of lateritic bauxites. Geochimica et Cosmochimica Acta, 53, 1411-1420. Bjerknes, J. (1969) Atmospheric teleconnections from the equatorial Pacific. Monthly Weather Review, 97, 163-72. Bowler, J.M. (1982) Aridity in the late Tertiary and Quaternary of Australia. In Barker, W.R. and Greenslade, P.J.M. (eds.) Evolution of the flora and fauna of arid Australia. Peacock Publications, 35-46. Bowler, J.M. (1983) 32+/- 5 KA: Northern Australia: Hydrologic evidence. In Proceedings of the First CLIMANZ Conference, February 1981. Department of Biogeography and Geomorphology, Research School of Pacific Studies, Australian National University. Bowler, J.M., Hope, G.S., Jennings, J.N., Singh, G. and Walker, D. (1976) Late Quaternary climates of Australia and New Guinea. Quaternary Research, 6, 359-394. Brimhall, G.H., Chadwick, O.A., Lewis, C.J. Compston, W., Williams, I.S., Danti, K.J., Dietrich, W.E., Power, M.E., Hendricks, D. and Bratt, J. (1991) Deformational mass transport and invasive processes in soil evolution. Science, 255, 695-702. Caine, N. (1974) The geomorphological processes of the alpine environment. In Iver, J.D. and Barry, R.G. (eds.) Arctic and Alpine Environments, 721-748. Carson, M.A. and Kirkby, M.J. (1971) Hillslope form and process. Cambridge University Press. 83 Chappe!!, J. (1985) Denudation and sedimentation in some Northern Territory river basins. In Loughran, R.J. (ed.) Drainage basin erosion and sedimentation. University of Newcastle Conference and Review Papers, 2, 69-76. Chen, X.Y. and Barton, C.E. (1991) Onset of aridity and dune building in central Australia: Sedimentological and magnetostratigraphic evidence from Lake Amadeus. Palaeogeography, Palaeoclimatology, Palaeoecology, 84, 55- 73. Chen, X.Y, Bowler, J.M. and Magee, J.W. (1991) Aeolian landscapes in central Australia: Gypsiferous and quartz dune environments from Lake Amadeus. Sedimentology, 38, 519-518. Christian, C.S. and Stewait, G.A. (1953) General report on survey of the Katherine-Darwin region, 1946. CSIRO Australian Land Research Series No. 1. Clemens, S.C., Prell, W.L, Murray, D.W., Shimfield, G.B. and Weedon, G.P. (1991) Forcing mechanisms of the Indian Ocean monsoon. Nature, 353, 720-725. COHMAP Project Members (1988) Climatic changes of the last 18 000 years: Observations and model simulations. Science, 241, 1043-1052. Coventry, R.J., Moss, A.J. and Verster.E. (1988) Thin- surface soil layer attributable to rain-flow transportation on low-angle slopes: an example from semi-arid tropical Queensland. Earth Surfaces and Landforms, 13, 421-430. Crowley, T.J. and North, G.R. (1991) Paleoclimatology. Oxford Monographs on Geology and Geophysics, No. 16. Oxford University Press. Davidson, N.E., McBride, J.L and McAvaney, B.J. (1983) The onset of the Australian monsoon during winter MONEX: synoptic aspects. Monthly Weather Review, 111,496-516. Dethier, D.P. (1986) Weathering rates and the chemical flux from catchments in the Pacific northwest, U.S.A. In Coleman, S.M. and Dethier, D.P. (eds.) Rates of chemical weathering of rocks and minerals. Academic Press, 503-528. DeVries, T.J. (1987) A review of geological evidence for ancient El Nino activity in Peru. Journal of Geophysical Research (92), 14471-14479. Duggan, K. (1988) Mining and erosion in the Alligator Rivers Region of northern Australia. Ph.D. Thesis, School of Earth Sciences, Macquarie University. Dunne, T. (1978) Rates of chemical denudation of silicate rocks in tropical catchments. Nature, 274, 244-246. Fairbanks, R.G. (1989) A 17000-year glacio-eustatic sea level record: Influence of glacial melting rates on the Younger Dryas event and deep ocean circulation. Nature, 342, 637-642. 84 Fontugne, MR. and Duplessy, J.C. (1986) Variations of the monsoon regime during the upper Quaternary: Evidence from carbon isotopic record of organic matter in North Indian Ocean sediment cores. Palaeogeography, Palaeoclimatology, Palaeoecology, 56, 69-88. Gadgil, S. (1988) Recent advances in monsoon research with potential reference to the Indian monsoon. Australian Meteorological Magazine, 36, 193-204. Galloway, R.W. (1976) Geomorphology of the Alligator River area. In Story, R., Galloway, R.W., McAIpine, J.R., Aldrick, J.M. and Williams, M.A.J. Lands of the Alligator River area, Northern Territory. CSIRO Land Research Series No.38, 52-70. Goodman, R.E. (1989) Introduction to rock mechanics. Second edition. John Wiley and Sons. Goodman, R.E. and Bray, J.W. (1977) Toppling of rock slopes. Proceedings, Speciality Conference on Rock Engineering for Foundations and Slopes, ASCE, 2, 201-234. Grimes, K.G. (1979) The stratigraphic sequence of old land surfaces in northern Queensland. Bureau of Mineral Resources, Journal of Australian Geology and Geophysics, 4, 33-46. Gray, W.M. (1975) Tropical cyclone genesis. Department of Atmospheric Sciences Paper No. 234, Colorado State University. Gray, W.M. (1986) Tropical cyclone global and regional climatology. Proceedings of the WMO International Workshop on Tropical Cyclones (IWTC) Report, WMO Technical Document No. 83. Gray, W.M. (1988) Environmental influences on tropical cyclones. Australian Meteorological Magazine, 36, 127-139. Hart, B.T. and Beckett, R. (1986) The composition of suspended particulate matter from the Magela Creek system, northern Australia. Environmental Techonolgy Letters, 7, 613-624. Hart, B.T., Ottaway, E.M. and Noller (1987a) Magela Creek system, northern Australia. I. 1982-83 wet season water quality. Australian Journal of Marine and Freshwater Research, 38, 261 -288. Hart, B.T., Ottaway, E.M. and Noller, (1987b) Magela Creek system, northern Australia. II. Material budget for the floodplain. Australian Journal of Marine and Freshwater Research, 38, 861 -876. 85 Hasienrath, S. (19S8) Climate and circulation of the tropics. Second edition, D. Riedel Publishing Company. Hays, J. (1967) Land surfaces and laterites in the north of the Northern Territory. In Jennings, J.N. and Mabbutt, J.A. (eds.) Landform Studies from Australia and New Guinea. Australian National University, 181-210. Hendon, H.H., Davidson, N.E. and Gunn, B. (1989) Australian summer monsoon onset during AMEX 1987. Monthly Weather Review, 117, 370-390. Hobgood, J.S. and Cerveny, R.S. (1988) Ice-age hurricanes and tropical storms. Nature, 333, 243-245. Hoek, E. and Bray, J.W. (1977) Rock slope engineering. Institute of Mining and Metallurgy, London. Holland, G.J. (1986) Interannual variability of the Australian summer monsoon at Darwin. Monthly Weather Review, 114, 594-604. Hyde, W.T., Crowley, T.J., Kwang-Yul, K. and North, G.R. (1989) Comparison of General Climatic Model and Energy Balance Model simulations of seasonal temperature changes over the past 18 000 years. Journal of Climate, 2, 864-887. Idnurm, M. (1986) 12th International Sedimentological Congress, Canberra. Abstract 150. Idnurm, M. and Senior, B.R. (1978) Palaeomagnetic ages of Late Cretaceous and Tertiary weathering profiles in the Eromanga Basin, Queensland. Palaeogeography, Palaeoclimatology, Palaeoecology, 24, 263-277. International Society for Rock Mechanics (1978) Methods for the quantitative description of rock masses and discontinuities. International Journal of Rock Mechanics Mineral Science and Geomechanics Abstracts, 15, 319-368. Jaques, A.L., Lewis, J.D. and Smith, C.B. (1986) The kimberlites and lamproites of Western Australia. Geological Survey of Western Australia, Bulletin 132. Johnson, R.B. and DeGraff, J.V. (1988) Principles of engineering geology. John Wiley & Sons. Keith, R., Bate, P.W., Cheang, B.K. and Sankaran, P. (1991) The tropical circulation in the Australian/Asian region November 1988 to April 1989. Australian Meteorological Magazine, 39, 37-46. Kemp, E.M. (1978) Tertiary climatic evolution and vegetation history in the southeast Indian Ocean region. Palaeogeography, Palaeoclimatology, Palaeoecology, 24, 169- 208. 86 Keishaw, A.P. (1986) Climalic change and Aboriginal burning in north-east Australia during the last two glacial/inter- glacial cycles. Nature, 322, 47-49. Kershaw, A.P. (1988) Australasia. In Huntley, B. and Webb, T. Ill (eds.) Vegetation history. Kluwer, 237-306. Koppi, A.J., Klessa, D.A. and Conoley, C. (1991) Particle size analysis and mineral weathering in M1, M2 and M3. Alligator Rivers Analogue Project, Progress Report March- August 1991. ANSTO, 225-233. Kutzbach, J.E. (1981) Monsoon climate of the Early Holocene: Climate experiments with the earth's orbital parameters for 9000 years ago. Science, 274, 59-67, Kutzbach, J.E. and Gallimore, R.G. (1988) Sensitivity of a coupled atmosphere/mixed layer ocean model to changes in orbital forcing at 9000 years BP. Journal of Geophysical Research, 93, 803-827. Kutzbach, J.E. and Guetter, PJ. (1986) The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years. Journal of Atmospheric Science, 43, 1726-1759. Kutzbach, J.E. and Otto-Bliesner, B. (1982) The sensitivity of the African-Asian monsoonal climate to orbital parameter changes for 9000 years BP in a low-resolution general circulation model. Journal of Atmospheric Science, 39, 1117-1188. Kutzbach, J.E. and Street-Perrott, F.A. (1985) Milankovitch forcing of fluctuations in the level of tropical lakes from 18 to OkyrB.P. Nature, 317, 130-134. Lautenschlager, M. (1991) Simulation of the ice age atmosphere - January and July means. Geologische Rundschau, 80, 513-534. Markgraf, V. (1989) Palaeoclimates in Central and South America since 18,000 B.P. based on pollen and lake level records. Quaternary Science Reviews, 8, 1-24. McAlpine, J.R. (1969) Climate of the Adelaide-Alligator area. In Story, R., Williams, M.A.J., Hooper, A.D.L, O'Ferrall, R.E. and McAlpine, J.R. Lands of the Adelaide- Alligator area, Northern Territory. CSIRO Land Research Series, 25, 49-55. McAlpine, J.R. (1976) Climate and water balance. In Story, R., Galloway, R.W., McAlpine, J.R., Aldrick, J.M. and Williams, M.A.J. Lands of the Alligator Rivers area, Northern Territory. CSIRO Land Research Series, 38, 35-49. McBride, J.L (1987) The Australian summer monsoon. In Chang, C.-P. and Krishnamurti, T. (eds.) Monsoon meteorology. Oxford University Press, 203-230. 87 McBride, J.L. and Keenan, T.D. (1982) Climatology of tropical cyclone genesis in the Australian region. Journal of Climate, 2, 13-33. McBride, J.L and Nicholls, N. (1983) Seasonal relationships between Australian rainfall and the Southern Oscillation. Monthly Weather Review, 111, 1998-2004. McCulloch, M.T., De Deckker, P. and Chivas, A.R. (1989) Strontium isotope variations in single ostracod valves from the Gulf of Carpentaria, Australia: A palaeoenvironmental indicator. Geochimica et Cosmochimica Acta, 53, 1703-1710. McDonald, N.S. and McAlpine, J. (1991) Floods and droughts: The northern climate. In Haynes, C.D., Ridpath, M.G. and Williams, M.A.J. (eds.) Monsoonal Australia: Landscape, ecology and man in the northern lowlands. A.A. Balkema, 19-29. McMahon, Burgess and Yeates (1981) Koongarra project report on site investigation for water management structures. Meybeck, M. (1976) Total annual dissolved transport by world major rivers. Hydrological Sciences Bulletin, 21, 265-289. Mills, K. (1991) Report on geological field inspection and investigations in the Koongarra area, August 7-9, 1990. Alligator Rivers Analogue Project. Progress Report, 1 September 1990-30 November 1990. ANSTO, 93-102. Mitchell, J.F.B., Grahame, N.S. and Needham, K.J. (1988) Climate simulation for 9000 years before present: Seasonal variations and effect of the Laurentide ice sheet. Journal of Geophysical Research, 93, 8283-8303. Morris, J.D. (1991) Application of cosmogenic 10-Be to problems in the earth sciences. Annual Review in Earth and Planetary Sciences, 19,313-350. Mulder, C.A. and Whitehead, B.R. (1988) Geoicgy of the Katherine Gorge National Park. Department of Mines and Energy, Northern Territory Geological Survey, Report 3. Nahon, D.B. (1986) Evolution of iron crusts on tropical landscapes. In Colman, S.M. and Dethier, D.P. (eds.) Rates of chemical weathering of rocks and minerals. Academic Press, 169-191. Nanson, G.C., East, T.J., Roberts, R.G., Clark, R.L and Murray, A.S. (1990) Quaternary evolution and landform stability of Magela Creek catchment near the Tanger uranium Mine, Northern Australia. Supervising Scientist for the Alligator Rivers Region, Open Record File 63. Needham, R.S. (1982) Cahill, Northern Territory. 1:100 000 Geological Map and Commentary. Bureau of Mineral Resources, Geology and Geophysics. Canberra, Australia. 88 Needham, RS. (1988) Geology of the Alligator Rivers uranium field. Northern Territory. Bureau of Mineral Resources Bulletin, 224. Ohmori, H. (1983) Erosion rates and their relation to vegetation from the viewpoint of world-wide distribution. Bulletin of the Department of Geography No. 15, University of Tokyo. Prell, W.L (1982) Reply to comments by P.J. Webster and N.A. Streten regarding "Surface circulation of the Indian Ocean during the last glacial maximum, approximately 18 000 YBP". Quaternary Research, 17, 128-131. Prell, W.L (1984) Monsoonal climate of the Arabian Sea during the Late Quaternary: A response to changing solar radiation. In Berger, A., Imbrie, J., Hays, J., Kukla, G. and Saltzman, B. (eds.) Milankovitch and Climate. D.Reidel, 349-366. Prell, W.L, Hutson, W.H., Williams, D.F., Be, A.W.H., Geitzenauar, K. and Molfino, B. (1980) Surface circulation of the Indian Ocean during the Last Glacial Maximum, approximately 78,000 yr BP. Quaternary Research, 74, 309- 336. Prell, W.L., Imbrie, J., Martinson, D.G., Morley, J.J., Pisias, N.G., Shackleton, N.J. and Steeter, H.F. (1986) Graphic correlation of oxygen isotope application to the Late Quaternary. Paleoceanography, 1, 137-162. Prell, W.L and Kutzbach, J.E. (1987) Monsoon variability over the past 150 000 years. Journal of Geophysical Research, 92, 8411 -8425. Prowse, G. (1990) Private Communication. Quinn, W.H. (1971) Late Quaternary meteorological and oceanographic developments in the equatorial Pacific. Nature, 229, 330-331. Rieger, W.A. and Olive, LJ. (1988) Channel sediment loads: Comparisons and estimation. In Warner, R.F. (ed.) Fluvial geomorphology of Australia. Academic Press,69-83. Roberts, R.G. (1989) Geomorphological field observations. Alligator Rivers Analogue Project. Progress Report 1 September 1989 - 30 November 1989, 9-11. Roberts, R.G., Jones, R. and Smith, M.A. (1990) Thermoluminescence dating of a 50,000 year-old human occupation site in northern Australia. Nature, 345, 153-156. Roberts, R.G., Nanson, G.C. and East, T.J. (1991) Sediment budgets and Quaternary history of the Macjela Creek catchment, tropical northern Australia. Supervising Scientist for the Alligator Rivers Region, Open File Record 80. Rodgers, R.D. and Schumm, S.A. (1991) The effect of sparse vegetative cover on erosion and sediment yield. Journal of Hydrology, 123, 19-24. 89 RossJgnol-Strick, M. (19S3) African monsoons, an immediate response to orbital insolation. Nature, 304, 46-49. Salinger, M.J. (1984) New Zealand climate: The last 5 million years. In Vogel, J.C. (ed.) Late Cainozoic palaeoclimates of the Southern hemisphere. Proceedings of symposium by South African Society for Quaternary Research, Swaziland August-September 1983. Balkema, 131-150. Schaap, A.D. (1990) Weipa kaolin and bauxite deposits. In Hughes, F.E. (ed.) Geology of the Mineral Deposits of Australia and Papua New Guinea. The Australian Institute of Mining and Metallurgy, Melbourne, 1669-1673. Selby, M. J. (1980) A rock mass strength classification for geomorphic purposes: With tests from Antarctica and New Zealand. Zeitschrift fur Geomorphologie, 24, 31-51. Selby, M.J. (1982) Hillslope materials and process. Oxford University press. Short, D.A. and Mengel, J.G. (1986) Tropical climatic phase lags and Earth's precession cycle. Nature, 323, 48-50. Short, S.A., Lowson, R.T., Ellis, J. and Price, D.M. (1989) Thorium-uranium disequilibrium dating of Late Quaternary ferruginous concretions and rinds. Geochimica et Cosmochimica Acta, 53, 1379-1389. Singh, G. and Luly, J. (1991) Changes in vegetation and seasonal climate since the last full glacial stage at Lake Frame, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology, 84, 75-86. Slaymaker, O. (1988) Slope erosion and mass movement in relation to weathering in geochemical cycles. In Lerman, A. and Meybeck, M. (eds.). Physical and chemical weathering in geochemical cycles. Kluwer Academic Publishers, 83-111. Snelling, A. (1989) Private Communication Snelling, A. (1990) Geological development of the Koongarra uranium deposit Alligator Rivers Analogue Project. Progress Report, 1 September-30 November 1990, 7-44. Southern, R.L. (1966) A review of weather disturbances controlling the distribution of rainfall in the Darwin-Katherine region, N.T. Australian Bureau of Meteorology, Working Paper 65/3203. Stallard, R.F. (1985) River chemistry, geology, geomorphology and soils of the Amazon and Orinoco basin. In Drever, J.I. (ed.) The chemistry of weathering, D. Reidel, 293-316. 90 Stallard, R.F. (1988) Weathering and erosion in the humid tropics. In Lerman, A. and Meybeck, M. (eds.) Physical and chemical weathering in geochemical cycles. D. Reidel. Stein, R. and Robert, C. (1985) Siliciclastic sediments at sites 588, 590, and 591: Neogene and Paleogene evolution in the southwest Pacific and Australian climate. In Kennett, J.P. and von der Borch, C.C. Initial reports of the Deep Sea Drilling Project, 90, 1437-1455. Street, F.A. and Grove, AT. (1979) Global maps of lake- level fluctuations since 30 000 Yr BP. Quaternary Research, 12, 83-118. Street-Perrott, F.A. and Roberts, N. (1983) Fluctuations in closed-basin lakes as an indicator of past atmospheric circulation patterns. In Street-Perrott, F.A., Beran, M. and Radcliffe, R.A.S. (eds.) Variations in global water budget. D. Reidel, 331-345. Summerfield, M.A. (1991) Global geomorphology. Longman, London. Torgersen, T., Stephens, M.R., Searle, D.E. and Ulman, W.J. (1985) Late Quaternary hydrological changes in the Gulf of Carpentaria. Nature, 313, 785-787. Troup, A.J. (1961) Variations in upper tropospheric flow associated with the onset of the Australian summer monsoon. Indian Journal of Meteorology and Geophysics, 12, 217-230. Van der Kaars, W.A. (1991) Palynology of Eastern Indonesian marine piston-cores: A Late Quaternary vegetational and climatic record for Australasia. Palaeogeography, Palaeoclimatology, Palaeoecology, 85, 239-302. Veevers, J.J. (1991) Mid-Cretaceous tectonic climax, Late Cretaceous recovery, and Cainozoic relaxation in the Australian region. In Williams, M.A.J., De Deckker, P. and Kershaw, A.P. (eds.) The Cainozoic in Australia: A re- appraisal of the evidence. Geological Society of Australia Special Publication No. 18, 1-14. Velbel, M.A. (1983) The mathematical basis for determining rates of geochemical and geomorphic processes in small watersheds by mass balance: Examples and implications. In Coleman, S.M. and Dethier, D.P. (eds.) Rates of chemical weathering of rocks and minerals. Academic Press, 439-451. Wasson, R.J., Clark, R.L, East, T.J., Guppy, J., Hancock, G., Johnston, A., Leaney, F.R., Mahon, D.A., Marten, R., Martin, P., McBride, P., Murray, A.S., Noller, B., Pfitzner, J., Warner, R.F. and Willett, I.R. (1990) Modern sedimentation and late Quaternary evolution of the Magela Creek plain. Report to the Supervising Scientist for the Alligator Rivers Region, Sydney. Wasson, R.J. and Donnelly, T.H. (1991) Palaeoclimate reconstructions for the last 30 000 years in Australia- A contribution to prediction of future climate. CSIRO Division cf Water Resources. Technical Memorandum 91/3. 91 Webster, P.J. and Streten, N.A. (1978) Late Quaternary ice age climates of tropical Australasia: Interpretations and reconstructions. Quaternary Research, 10, 279-309. Webster, P.J. and Streten, N.A. (1981) Comments on: "Surface circulation of the Indian Ocean during the last glacial maximum, approximately 18 000 yr. B.P.". Quaternary Research, 16, 421 -423. Wells, M.R. (1979) Soil studies in the Magela Creek catchment, Part 1. Land Conservation Unit, Northern Territory Parks and Wildlife Commission Report. Williams, A.R. (1984) Changes in Melaleuca forest density in the Magela floodplain, Northern Territory, between 1950- 1975. Australian Journal of Ecology, 9,199-202. Williams, M.A.J. (1968) Termites and soil development near Brocks Creek, Northern Territory. Australian Journal of Science, 31, 153-154. Williams, M.A.J. (1969a) Geology of the Adelaide-Alligator area. In Story, R., Williams, M.A.J., Hooper, A.D.L, O'Ferrall, R.E. and McAlpine, J.R. (eds.) Lands of the Adelaide-Alligator area, Northern Territory. CSIRO Land Research Series No. 25, 56-70. Williams, M.A.J. (1969b) 1:500 000 geomorphology map of Adelaide-Alligator area, Northern Territory. In Story, R., Williams, M.A.J., Hooper, A.D.L., O'Ferrall, R.E. and McAlpine, J.R. (eds.) Lands of the Adelaide-Alligator area, Northern Territory. CSIRO Land Research Series No. 25, 71- 94. Williams, M.A.J. (1973) The efficacy of creep and slopewash in tropical and temperate Australia. Australian Geographical Studies, 11, 62-78. Williams, M.A.J. (1991) Evolution of the landscape. In Haynes, C.D., Ridpath, M.G. and Williams, M.A.J. (eds.) Monsoonal Australia: landscape, ecology and man in the northern lowlands. A.A. Balkema, 5-18. Woodroffe, C.D., Chappell, J.M.A., Thorn, 3.G. and Wallensky, E. (1986) Geomorphological dynamics and evolution of the south Alligator tidal river and plains, Northern Territory. Australian National University North Australian Research Unit, Mangrove Monograph No. 3. Wright, R.L (1963) Deep weathering and erosion surfaces in the Daly River basin, Northern Territory. Journal of the Geological Society of Australia, 10, 151-164. Wyrwoll, K.-H. and Milton, D. (1976) Widespread Late Quaternary aridity in Western Australia. Nature, 264, 429-430. 92 Wyrwoll, K.-H., Hopwood, J. and McKenzie, N.L (1992) The Holocene paleohydrology and climatic history of the northern Great Sandy Desert-Fitzroy Trough: With special reference to the history of the northwest Australian monsoon. Climatic Change (in press). Wyrwoil, K.-H., McKenzie, N.L, Pederson, B.J. and Tapley, I.J. (1986) The Great Sandy Desert of northwestern Australia: The last 7000 years. Search, 17, 208-211. Young, R.W. and McDougall, I. (1985) The age, extent and geomorphological significance of the Sassafras basalt, southeastern New South Wales. Australian Journal of Earth Sciences, 32, 323-331. 93