Loess and Floods: Late Pleistocene Fine-Grained Valley-Fill Deposits in The Flinders Ranges, South Australia (PHD Thesis)

Loess and floods: late Pleistocene fine-grained valley-fill deposits in the Flinders Ranges, South Australia
(excerpt from Hans Heysen 1929: Foothill of the Flinders, Morgan Thomas Bequest Fund 1939)
This thesis is submitted in fulfilment of the requirements for the degree of Doctor of Philosophy in the Faculty of Science, University of Adelaide
Geology and Geophysics School of Earth and Environmental Sciences The University of Adelaide
David Haberlah
August 2009
(Place holder)
Contents
Contents ..................................................................................................................... ........................................................................................................ ........................................................................................................ I IV V VII X 1 3
Thesis abstract Thesis declaration
Co-author contributions declaration .............................................................................. Acknowledgments 1. Introduction ........................................................................................................ ........................................................................................................ ..........................
1.1 A review of the last deglacial response in South Australia
1.2 Depositional models of late Pleistocene fine-grained valley-fill formations in the Finders Ranges, South Australia 1.3 A call for Australian loess ................................................................. 31 41 51 55 57 73 83 87 101
..............................................................................
1.4 Response to Smalleys discussion of A call for Australian loess .......................... 2. Methods ..................................................................................................................... .................................................... ....................................... .......................................
2.1 Optically Stimulated Luminescence dating
2.2 Accelerator Mass Spectrometry radiocarbon dating 2.3 Particle-Size Analysis 3. Results and Discussion Quantifying particle aggregation in fluvial sediments
........................................................................................... ...........................................................................................
3.1 Of droughts and flooding rains*: an alluvial loess record from central South Australia spanning the last glacial cycle .................................................... 103
3.2 Loess and floods: high-resolution multi-proxy data of Last Glacial Maximum (LGM) slackwater deposition in the Flinders Ranges, semi-arid South Australia 4. Conclusions and Outlook ........................................................................................... 151 201 207 208 209 211 221 223 225 227 229
5. Appendices ..................................................................................................................... 5.1 Instructions ........................................................................................................ 5.2 Methods ........................................................................................................ .............
5.2.1 Dose Rate, Equivalent Dose and OSL Age calculation Table 5.2.1.1 Cosmic Ray Dose Rate contribution Table 5.2.1.3 Thick Source Alpha Counting (TSAC) Protocol 5.2.1.4 XRF sample preparation Protocol
....................................... ..........................
5.2.1.2 Dose Rate calculation with AGE99 software Protocol .......................... .................................................... .......................................
5.2.1.5 U and Th contents TSAC conversion Table
II
5.2.1.6 XRF results Table .............................................................................. 231
5.2.1.7 Manual De-value selection using Analyst 3.22 and Radial Plot 1.3 Documentation .............................................................................. .......................... ....................................... 233 243 245 249 253 273 277 281 301 5.2.1.8 Equivalent Dose sample preparation Protocol 5.2.1.9 Finite Mixture Model OSL data Table 5.2.2 Flinders Silts Age Sheet 5.2.3.1 Particle-Sizing Protocol
................................................................. ............. ................................................................. ....................................... ..........................
5.2.3 Haberlah and McTainsh (manuscript) Supplementary Data 5.2.3.2 Component Population Analysis Protocol
5.3 Haberlah et al. 2009 (QSR manuscript) Supplementary Data 5.4 Conference Abstracts
..............................................................................
5.4.1 A terminal Last Glacial Maximum (LGM) loess-derived palaeoflood record from South Australia? .............................................................................. 5.4.2 The Flinders Silts: a last glacial alluvial loess record from South Australia 5.4.3 Dust fingerprinting in regolith: an integrated high-resolution parametric particle-size analysis quantitative spectral mineralogy approach ............. 5.4.4 Last glacial dust cycles in South Australia: employing advanced Automated Mineralogy and sediment-size analyses in the study of provenance, transport and depositional palaeo-environments ............. 309 307 303 305
III
IV
Thesis abstract
Terrace remnants of late Pleistocene fine-grained valley-fills, at present eroded by ephemeral traction-load streams, are reported from many semi-arid and arid parts of the world. While they present promising palaeo-environmental archives for recent geological times such as the Last Glacial Maximum (LGM) for which few other terrestrial depositional records exist, their poorly understood nature has limited their significance. This study examines the fine-grained valley-fill deposits from the Flinders Ranges in South Australia, here called Flinders Silts. It establishes the timing, mode and environmental controls of deposition as opposed to their advancing erosion under the current climate. A regional chronostratigraphy based on 124 numerical dates is discussed, of which 43 radiocarbon and 22 luminescence ages were obtained from 12 sections across three major catchments within the scope of this thesis. Regionally significant intervals of rapid aggradation, relative surface stability and erosion are established. Regional climatic controls are differentiated from intrinsic catchment- and site-specific effects on the system. Further, individual age proxies and age models are critically assessed in how far they reflect depositional events. The final aggradational interval bracketing the extended LGM is discussed in detail on a continuous layered to laminated stratigraphic sequence. The provenance question of the fine-grained sediments and the depositional environment of the Flinders Silts are further addressed by high-resolution particle-size analysis. In order to study the subtle variations within the fine-grained partially-aggregated material, an original parametric sediment-sizing approach is employed. Finally, a range of traditional and emerging analytical techniques are applied to improve our understanding of palaeo-environments promoting aggradation. In conclusion, arid intervals throughout the last glacial cycle resulted in significant quantities of proximal dust being deposited as loess mantles within the catchments of the Flinders Ranges, acting as a near-longitudinal dust trap in the centre of the late Pleistocene dust bowl. The fine-grained aeolian accessions were repeatedly eroded by low-frequency high-magnitude precipitation events and redistributed as loess-derived alluvium, congesting narrow gorges and raising the base level for tributaries. Locally, backflooding resulted in the aggradation of layered to laminated slackwater deposits, the most continuous recording at least 12 large and numerous smaller flood events between 24 ka and 18 ka. The synchronous termination of the Flinders Silts coincides with early Deglacial climatic amelioration. The re-establishment of a perennial plant cover stabilising both dune fields and slope mantes is discussed as a potential scenario that would have discontinued dust supply to the fluvial system, in turn promoting incision and erosion. The studied aeolian-fluvial interplay of loess and floods has large implications for our understanding of landscape evolution in semi-arid Australia.
VI
HaberlahD. 2008.ResponsetoSmalley'sdiscussionof'AcallforAustralianloess'.Area40.1,135 136. HaberlahD.2007:AcallforAustralianloess.Area39.2,224229. HaberlahD.2007:DepositionalmodelsoflatePleistocenefinegrainedvalleyfillformationsinthe Flinders Ranges, South Australia. In: Fitzpatrick R.W., & Shand P. (eds) 2007, Regolith 2006 Consolidationanddispersionofideas.CooperativeResearchCentreforLandscapeEnvironmentsand MineralExploration,Perth,122126.
Conferenceabstracts
HaberlahD.,WilliamsM.A.J.,HillS.M.,HalversonG.,SutoA.,GlasbyP.,ButcherA.R.,HrstkaT.2009. The Flinders Silts: a last glacial alluvial loess record from South Australia (oral presentation at 7th InternationalConferenceonGeomorphology2009) HaberlahD.,HillS.M.,StrongC.,ButcherA.R.,McTainshG.H.,HrstkaT.2009.Dustfingerprintingin regolith: an integrated highresolution parametric particlesize analysis quantitative spectral mineralogyapproach(posterpresentationat7thInternationalConferenceonGeomorphology2009) Haberlah D., Williams M.A.J., Hill S.M., Halverson G., Glasby P. 2007. A terminal Last Glacial Maximum (LGM) loessderived palaeoflood record from South Australia? Quaternary International 167168,15.
IX
X Acknowledgments
The teacher who is indeed wise does not bid you to enter the house of his wisdom but rather leads you to the threshold of your mind Khalil Gibran Ideas, support and supervision are best acknowledged in form of co-authorship on papers that arise from the collaboration. However, this would not rightfully acknowledge the important role that my supervisors and mentors played over the past three years guiding me through my PhD candidature. A role that goes beyond teaching me how to read geological records, conduct laboratory studies, write papers, apply for grants and present at conferences. It is the inspiration to think big questions while remaining aware of practical limitations, to enjoy what our field of science has to offer without overindulging. This important lesson required leading examples and a fine balance of patience and impatience to keep me on track like the trains of old. In this sense, I have been blessed with experienced and inspiring supervisors. I would like to thank my co-supervisors Steven Hill, Victor Gostin, John Prescott, Frances Williams, Grant McTainsh, Galen Halverson, Andreas Schmidt-Mumm and Margaret Cargill for guiding me through crucial times of my candidature. In particular, my gratitude goes to my principle supervisor Martin Williams. I feel very privileged for the time and training he invested and shall always think of him in the German term as my Doktorvater. Further, I would like to thank Peter Glasby for his good nature and patience. I am well aware that sharing a field area with me must have been at times testing. My sincere thanks for engaging in countless discussions and enduring numerous re-interpretations. I would also like to thank my examiners Robert Wasson from Charles Darwin University and David Thomas from the University of Oxford for their prompt and thorough reviews of this thesis. Finally, I am indebted to the University of Adelaide for awarding me the Endeavour International Postgraduate Research Scholarship (EIPRS) scholarship, which not only covered all fees, but provided me with generous means that allowed me to fully concentrate on my research.
XI
1. Introduction
Lo ess an d Flo o d s: rising. Muhammed Iqbal
Ch ap t er 1 (In t r o d u ct io n )
It is true that we are made of dust. And the world is also made of dust. But the dust has motes
This study is presented as a PhD thesis by publication and consists largely of material submitted for publication in the form of journal papers and book chapters. At the same time, it attempts to comply with the basic structure of a conventional thesis by including a chapter each on introduction, methods, results and discussion, and conclusions and outlook. The original papers have been to some extent adapted and are put into the larger context at the beginning of each chapter, to improve the coherence and flow of reading of the thesis. Detailed methodological reviews not suitable for publication complement the journal papers in the methodology chapter. The final chapter briefly summarises overall conclusions and outlines potential future research. Background information, laboratory protocols, detailed data tables and published conference abstracts referred to in the text are listed at the end as appendices. References are kept with the original manuscripts, or are otherwise listed at the end of individual chapters. In the introduction, the fine-grained valley-fill remnants of the Flinders Ranges, hereafter referred to as the Flinders Silts, are placed in a regional, continental and global context in the form of three separate literature reviews. The palaeo-environmental literature on the late Pleistocene of the Flinders Ranges and surrounding areas is reviewed and discussed within the chronostratigraphic framework of the Brachina valley-fills as known at the outset of this study in 2006 (Cock et al. 1999; Williams et al. 2001). This manuscript (section 1.1) did not get published and in light of the present data would need to be fully revised. However, it presents a useful literature review for future studies attempting to establish the link between rapid aggradation of the lacustrine sediments of flanking terminal playa lakes and the erosion of the Flinders Silts. In the second paper (section 1.2), depositional models inferred elsewhere for similar late Pleistocene valley-fill formations are reviewed and discussed as potential scenarios. This manuscript is written from the vantage point of a chronostratigraphic study and was consequently used to apply for funding for the regional dating campaign. The third paper (sections 1.3 and 1.4) investigates our present understanding of loessderived sediments in Australia. While the impact of Quaternary dust on Australian alluvium has been the subject of ongoing research since the Canberra INQUA-commissioned workshop in 1980 (Wasson 1982), proximal dust
Lo ess an d Flo o d s:
Ch ap t er 1 (In t r o d u ct io n )
has only recently been inferred as a significant source for the Flinders Silts (Williams et al. 2001; Williams and Nitschke 2005). The same applies to other well-published global occurrences of similar fine-grained valley-fill formations (e.g. Rgner et al. 2004).
References
Cock, B.J., Williams, M.A.J., Adamson, D.A. 1999. Pleistocene Lake Brachina: a preliminary stratigraphy and chronology of lacustrine sediments from the central Flinders Ranges, South Australia. Australian Journal of Earth Sciences 46, 61-69. Rgner, K., Knabe, K., Roscher, B., Smykatz-Kloss, W., Zller, L. 2004. Alluvial loess in the Central Sinai: Occurrence, origin, and palaeoclimatological consideration. Lecture Notes in Earth Sciences 102, 79-100. Wasson, R.J. (Ed.) 1982, Quaternary Dust Mantles of China, New Zealand and Australia. Proceedings of a Workshop. Australian National University, Canberra, 253. Williams, M.A.J., Prescott, J.R., Chappell, J., Adamson, D., Cock, B., Walker, K., Gell, P. 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International 83-85, 129-144. Williams, M.A.J., Nitschke, N. 2005. Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. South Australian Geographical Journal 104, 25-36.
A review of the last deglacial response in South Australia

Lo ess an d Flo o d s: Ch ap t er 1.1 (In t r o d u ct io n ) David Haberlah 1, 2, Victor Gostin 1, 2, Steven M. Hill 1, 2, Martin A.J. Williams 3, Peter Glasby 3
1 2
Geology & Geophysics, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia (david.haberlah@adelaide.edu.au) Cooperative Research Centre for Landscape Environments and Mineral Exploration
Geographical & Environmental Studies, School of Social Sciences, University of Adelaide, Adelaide, SA 5005, Australia
Abstract
Late Pleistocene fine-grained sediments in lake cores collected from Lake Frome, a playa lake in presently arid South Australia, had been interpreted as indicative of high lake levels. Evidence from well-dated valley-fill deposits in the nearby Flinders Ranges and from less well dated piedmont sediments and associated palaeosols, suggests an alternative interpretation. According to our model, barren loess mantles blanketing the ranges during the Last Glacial Maximum (LGM) were eroded by Late Pleistocene frontal winter rainfall as early as ~20.2 ka. The suspended load rapidly filled and levelled out pre-existing rock-cut relief and accumulated in the flanking terminal depocentres. Closely spaced gypsum bands point to a shallow lake, prone to rapid evaporation. Deposition of loess-derived fluviatile and lacustrine sediments ceased by ~16 ka, when rising temperatures and monsoonal incursions from the north promoted colonisation of the landscape by grassland communities. Subsequent sandy deposition in Lake Frome is interrupted just before ~5 ka by a 'clayey band' linked to the arid transition from dominant summer to winter precipitation. Terminal LGM frontal rainfall may explain other enigmatic 'pluvial' observations in southern Australia and be a global phenomenon, as similar occurrences of loess-derived alluvium in Namibia and Egypt suggest.
Keywords: Australia, Flinders Ranges, Lake Frome, Last Glacial Maximum, Loess
Ch ap t er 1.1 (In t r o d u ct io n )
Introduction
Fine-grained sediments associated with the time following LGM peak aridity are widespread across southern Australia. These are typically associated with aeolian deposition but also fluvial and lacustrine sedimentary systems, and provide important records of the environmental conditions at this time. Ironically, the precise palaeo-environmental significance of these deposits is poorly understood. For instance, within the sedimentary records of Lake Frome, the onset of fine-grained sedimentation has been interpreted as expressing a deepening of the lake associated with higher water levels (Bowler et al., 1986; Ullman and Collerson, 1994). However, this interpretation is not well supported by preservation of elevated shorelines or other proxy records expressed in the range of landscape components across the region. This manuscript collates and combines a wide array of palaeo-environmental data to assess the last deglacial landscape response to widespread LGM deposition of these fine-grained sedimentary mantles in South Australia. Changes in sediment provenance within the catchment, in particular the fluviatile reworking of widespread aeolian accessions of silts and fine sands referred to as loess mantles, are demonstrated to account for the rapid deposition of fine-grained valley-fill formations in the Flinders Ranges and the lacustrine facies within Lake Frome. The Flinders Ranges and its flanking terminal depocentres provide an ideal setting for a critical synthesis of wide-ranging palaeo-environmental data sets. Inter-related fluvial, aeolian, lacustrine and pedologic records, many of which have been the focus of previous climate research, form wellpreserved, significant occurrences within the study region (Fig. 1). The Flinders Ranges is an approximately 400km long, near-meridional trending mountain belt, running from the Southern Ocean to the arid interior of the continent. Frequently rising to heights above 900m ASL, the central ranges receive around twice the amount of orographically enhanced precipitation than the surrounding low-lying plains. The climate of the region is at present arid to semi-arid. Flanking salinas Lake Torrens and Lake Frome receive a spatially variable annual precipitation amounting to ~200 mm, with frontal winter rainfall associated with westerly anticyclonic air masses decreasing to the north, which is increasingly dominated by sporadic but intense summer thunderstorms associated with tropical monsoon incursions (Schwerdtfeger and Curran, 1996). Potential evaporation is twenty times that of the actual annual evaporation rate of ~160 mm, estimated at the lacustrine core sites from depth profiles of deuterium delta values (Allison and Barnes, 1985).
Fig. 1) ETM+ mosaic (modified data from http://glcf.umiacs.umd.edu/ [1 November 2006]) displaying the location of the Flinders Ranges between Lake Torrens and Lake Frome. Sites of sedimentary sequences discussed in the text are indicated in black; 1) Lacustrine cores LF82/1-3, 2) Suite of formations on the Lake Torrens piedmont plain, 3) Balcoracana Creek, 4) Hookina Silts 4) Brachina Silts. Other Australian sites mentioned in the text are indicated in red on the outline map; 1) Lake Gairdner, 2) Lake Eyre Basin, 3) Yanco Palaeochannel on the Riverine Plain, 4) Naracoorte Caves, 5) Kosciuszko Massif in the Snowy Mountains, 6) Lake George.
Lake Frome physiography

Lake Frome derives its water primarily from intermittent surface runoff and direct precipitation. Most of the runoff is from ephemeral streams from the eastern Flinders Ranges and occasionally from the Olary Ranges to the south (Ker, 1966), (Fig. 1). Inflow of Great Artesian Basin ground water from isolated mound springs on the eastern side of the lake amounts to less than several litres per hour (Draper and Jensen, 1976). The western lake margin consists of coalescing delta fans from
larger streams. There is shallow subsurface flow through the coarse talus and alluvium that extends from the range front towards the playa lake (Ker, 1966). East of Lake Frome the landscape is made up of dunefields interspersed saltpans of the southern Strzelecki Desert. Historical observations describe the playa lake surface as dry (Frome, 1899; Madigan, 1929). However, years of abnormally high precipitation, last recorded for 1972-1974, result in rapid inundation of large areas of the lake floor (Draper and Jensen, 1976; Callen, 1984). At present, the floor of Lake Frome is flooded episodically for short intervals (Bowler et al., 1986) during which surface water are concentrated around small islands in the southeast and at the mouths of some creeks (ETM+ imagery).
Islands of Lake Frome

An archipelago of ~15 groups of islands rises out of the deepest part of Lake Frome. These represent former dunes aligned roughly NE-SW. At present, they are cliffed, terraced and in some cases segmented by wave erosion. Their sedimentary sequences consist of aeolian clay pellets and gypsiferous quartz sands resting on the former laminated lacustrine bed. Horizontal bedding planes with low-angle cross-bedded structure are indicative of the former dunes (Callen, 1984). The islands present the best-preserved sedimentary section recording the terminal desiccation and subsequent deflation of the playa lake floor during LGM peak aridity. The gradually desiccating playa floor supplied sediments that were blown downwind into longitudinal dunes which now overly and preserve remnants of the final pre-LGM lacustrine phase. The dune remnants consist of a sequence of beds composed of pelletal clays and gypsum fragments, with the fore-set dips indicating westerly winds. The aeolian clays at the base pass into relatively pure sand-sized gypsum crystals. Prolonged sediment supply requires repeated cycles of inundation and desiccation to allow for salt shattering and clays to dry and crack (Bowler, 1976; Pye, 1995). Hence, the higher strata of the dunes (Callen, 1984, Fig.5) are likely to represent later stages of lake-floor deflation. On the other hand, the basal unit should coincide broadly with the initial dune-building phase during the LGM. The closest age estimate is based on one conventional
14
C-date, derived from inorganic carbonate partly
precipitated around Clara oogonids blown in from the lake floor and sampled from the centre of the 3m high basal aeolian clay pellet bed. Its age of ~23.33 0.46 ka 1 (Bowler, 1976; Callen et al., 1983; Callen, 1984) gives a minimum age for the onset of deflation of the lowest part of the playa lake floor and the consequent inception of gypsum dune formation. The formation of dunes on the wettest part of the lake-bed could be coeval with the formation of the clastic sediment mound
All presented C-dates are calibrated using the calibration software 'CalPal May 2006' (www.calpal.de/ [1 November 2006]) and the calibration curve of Fairbanks et al. (2005)
14
springs, trapping aeolian sediment from the lake floor (Draper and Jensen, 1976) and with the gypsum mounds that flank the eastern side of Lake Frome (Callen et al., 1983).
The lacustrine chronostratigraphy

The sediments of the playa lake have issued from surface runoff entering the lake on all but the eastern margins, mainly in the form of delta fans of streams flowing from the Flinders Ranges. The three stratigraphic lacustrine units recognised by Draper and Jensen (1976) were analysed in greater detail by Bowler et al. (1986), Singh and Luly (1991) and Ullman and Collerson (1994), (Fig. 2). The Lower Sands (Units C and D, Bowler et al., 1986) are found at progressively greater depth towards the centre of the playa lake, where they unconformably overlie pedogenic horizons of the fluviatile Willawortina Formation (Fig. 3). This basal unit is composed of mostly well-sorted unconsolidated sand to sandy mud with subangular to rounded grains of quartz, feldspar and mica. The Lower Sands contain abundant freshwater ostracods and oogonia of freshwater algae. The Upper Sands (Unit A, Bowler et al., 1986) are distributed over much of the present surface of the embayment proper. In lithology and fossil assemblages they resemble the Lower Sands, apart from the relative abundance of detrital carbonate (Draper and Jensen, 1976). The unit is still forming today, but is restricted to the margins of the lake and immediate areas around the islands. The present playa floor is covered by a veneer of evaporite minerals, mainly halite, which forms a solid crust up to 30cm thickness towards the centre, and reddish-brown gypseous sandy clays rising a couple of decimetres from the water table. Both sandy units encompass a distinct fine-grained lithofacies, (Unit B, Draper and Jensen, 1976; Bowler et al., 1986), consisting of silty clays laminated towards the base (Singh and Luly, 1991) with a high content of post-depositional displacive gypsum (Bowler et al., 1986; Ullman and McLeod, 1986). The fine-grained unit is exposed at the surface in the central lake area, wedging out laterally and becoming covered by the Upper Sands towards the margins. The clays are composed of kaolinite, montmorillonite and illite forming thin discontinuous laminae along bedding planes, which in places are disrupted by diagenetic gypsum crystals. Fossils are rare and restricted to ostracods and oogonia, partly reworked and concentrated along the planes. Propagation of the fine-grained lithofacies over the Lower Sands is interpreted to indicate a westerly movement of the shoreline as a result of rising lake levels (Draper and Jensen, 1986; Bowler et al., 1986). A group of three nearby cores (LF82/1-3) were recovered by hollow-flight auger coring (Chivas and Bowler, 1986) from a site ~6km east of the western shore just beyond the line where fan sediments merge with deeper lake sediments (Draper and Jensen, 1976; ETM+ imagery), (Fig. 1). The cores
reveal ~580cm of lacustrine sediment unconformably overlying the fluviatile Willawortina Formation. The chronostratigraphy of this lacustrine sequence is now based on ten 14C-dates from organic carbon samples. Eight ages were determined by conventional dating of very fine disseminated organic carbon, requiring the treatment of up to 30cm core length as single samples (Bowler et al., 1986). Two subsequent 14C-dates where obtained by accelerator mass spectrometry (AMS), confirming the general validity of the conventional radiocarbon ages (Luly and Jacobsen, 2000). The fine-grained lithofacies (Unit B) accumulated on top of a sedimentary discontinuity at 346cm, interpreted to indicate the LGM-interval of lake-floor deflation. The inference of a substantial time break at this depth, "identified by sharp lithological changes" (Bowler et al., 1986, 256), was not questioned by subsequent studies. However, similar fine-grained sediments grade down to a sandy facies at a depth of ~460cm (Bowler et al., 1986, Fig.7), possibly representing the Lower Sands of Draper and Jensen (1976). A potential lower onset of initial post-LGM sedimentation would be supported by the 14C-date of 19.92 0.63 ka from 377-360cm (Fig. 2). However, very low organiccarbon contents (<0.5%) make sediments below 346cm "very unreliable for age determination" (Bowler et al., 1986, 256). Conventional 14C-dates suggest that the initial post-LGM flooding of the lake floor took place shortly before ~20.23 0.36 ka (325-310cm). The first 6cm above the unconformity are slightly reddened, implying an oxidising, relatively dry environment. Between 340120cm, the fine-grained unit consists of grey gypseous silty clays. While the lower part of the lithofacies is laminated, a marked reduction in those laminae is evident between 185-120cm (Singh and Luly, 1991). The fine-grained lithofacies terminates at this depth, with sandy clays and sand setting in, likely representing the Upper Sands of Draper and Jensen (1976). The timing of the termination of the fine-grained sedimentation regime can be estimated by taking the two 14C-dates 16.11 0.32 ka (140-125cm) and 12.69 0.4 ka (90-80cm) as maximum and minimum ages, respectively. The entire fine-grained unit is marked by sedimentation rates far in excess of those of the coarser facies above and below (Bowler et al., 1986), (Fig. 2). The sedimentation regime of the upper sandy facies was interrupted by the deposition of a conspicuous thin grey 'clayey band' associated with secondary gypsum precipitation between 45cm and 34cm (Singh and Luly, 1991). The age of the fine-grained band is best estimated by an AMS sample from 38-35cm, dated to 5.06 0.15 ka (Luly and Jacobsen, 2000). The top 34cm of the sequence consist of reddened sandy clays (Singh and Luly, 1991). An earlier independent radiocarbon chronology for the fine-grained lithofacies (Unit B) is in good agreement with the ages presented above (Draper and Jensen, 1976). According to it the ages
between the two sandy units range from 20 0.7 ka near the base to 15.8 0.5 ka near the top of the fine-grained unit.
Fig. 2) Calibrated depth-age plots of Lake Frome core LF82/2 (blue, thick line for fine-grained lacustrine facies, dates from Bowler et al. 1986; Luly and Jacobsen 2000) and the Brachina Silts (red for the 'Slippery Dip' type section, black for 'Section G', dates from Cock et al. 1999; Williams et al. 2001). Correlated lithology, location of gypsum bands, radiocarbon date sample intervals (modified from Bowler et al., 1986; Luly 2001) and Callitris and Gramineae pollen and carbonised particle concentrations (Singh and Luly, 1991; Luly, 2001) are indicated.
Geochemical record from Lake Frome

Chlorine/Bromine ratios of salts of the brine and halite crust suggest that they are derived either from second cycle brines or weathering of saline rocks exposed in the Flinders Ranges. Similar bromide deficiencies in the brines of Lake Torrens and Lake Gairdner to the west (Johns, 1968), (Fig. 1), and the lack of substantial evaporite deposits in the catchment suggest an allochthonous aeolian provenance. Indirect evidence of lake-level fluctuations and desiccation events may be derived from the interpretation of evaporite minerals. Well-formed gypsum crystals, up to several centimetres in diameter, have formed diagenetically within the lacustrine sequence disrupting laminations of
10
organic-rich clays. Due to the presence of unoxidised pyrite and lack of evidence for solution of detrital carbonate, gypsum precipitation is interpreted to result from saturation reached by evaporation (Draper and Jensen, 1976; Ullman and McLeod, 1986). By modern analogue, it can be assumed that the gypsum crystals formed within the top 20cm of the sediment surface (Ullman and Collerson, 1994). Gypsum bands of variable thickness occur primarily in the lower section of the finegrained lithofacies and in the 'clayey band' between 45-34 (Bowler et al., 1986, Fig.7), (Fig. 2). Small crystal sizes and multiple phases of crystal growth in larger crystals as indicated by inclusions, suggest rapid evaporation rates (Draper and Jensen, 1976). Changes in strontium isotope ratios, recorded in authigenic gypsum precipitating from the brines, are assumed to reflect local hydrological changes related to climate variations (Ullman and Collerson, 1994). Drier intervals are expected to result in higher 87Sr/86Sr ratios, due to a greater flux of radiogenic continental dust and an increased residence time of surface groundwater allowing for a greater degree of reaction with more radiogenic sedimentary particles. Accordingly, above a depth of 330cm, a sharp drop in the strontium isotope ratios indicates an abrupt transition from arid to humid conditions which is reversed at ~140cm. Thereafter, a drying trend culminating in the present set in, reversed only by possibly deeper lake levels recorded by two gypsum crystals at 39-38cm and 36cm, within the 'clayey band'. Minor element concentrations were analysed in 38 gypsum samples from cores LF82/1-3 with the aim of establishing a palaeosalinity record for Lake Frome (Ullman and McLeod, 1986). Displacive gypsum crystals are formed post-depositionally and their analytical ratio of Na+/Ca2+ averages gypsum-precipitating intervals of unknown length. Therefore, the interpolated absolute 14C-ages can only be taken as approximations of minimum ages. However, the relative stratigraphy remains intact and reequilibration and recrystallisation were dismissed due to large variations in adjacent samples (Ullman and McLeod, 1986). The Na+/Ca2+ ratio was measured to estimate the chemistry and salinity of the brines since the LGM. From the dataset, a sudden initial drop and steep subsequent rise from hypersalinity to lower salinity and back, followed by a similar event of lower magnitude, is evident (Ullman and McLeod, 1986, Table 1). The first event is recorded by six gypsum crystals between 337317cm. Salinity concentrations rapidly dropped after the initial onset of post-LGM sedimentation with low-salinity conditions lasting until 20.58 0.37 ka (325-310cm, for carbonate). The second low-salinity interval is recorded by three gypsum crystals between 196cm and 160cm, corresponding with the two 14C-ages 16.17 0.42 ka (199-169cm, for carbonate) and 16.16 0.19 ka (180-165cm). Higher salinity values are recorded for the overlying ~30cm by another four crystals, prior to a return to two lower salinity values and a long hiatus until 39cm that is interrupted by two crystals recording
11
low-salinity concentrations at 93cm and 87cm. From these results, Ullman and McLeod (1986) suggest a freshwater phase with salinity levels below that of gypsum precipitation that can be bracketed by the two ages 16.16 0.19 ka (180-165cm) and 7.14 0.14 ka (64-44cm, for carbonate).
Pollen record
A pollen sequence, derived from the dated cores, presents an important and independent line of palaeo-environmental evidence, establishing a vegetation history since the 346cm unconformity associated with LGM deflation (Singh and Luly, 1991). Lake Frome is uniquely positioned within the shifting ecological boundary dividing winter and summer rainfall zones. Hence, the palynological record registers both the influence of monsoonal incursions from the north and frontal precipitation associated with the westerlies from the south. Distinctive although overlapping plant communities reflect the differing seasonal rainfall regimes. Chenopods register winter rainfalls, and hummock and tussock grassland communities are indicative of dominantly monsoonal summer precipitation. Most probably due to oxidation, no workable amounts of pollen were preserved below the disconformity at 346cm (Luly 2001). Above 346cm, the pollen from tree populations rapidly attained high values during initial post-LGM sedimentation (Fig. 2). Callitris dominates over Eucalyptus, with a total pollen count many times above the present. High rainfall events are important in encouraging mass recruitment of Callitris glaucophylla (Lacey, 1972), the currently dominant species in the Flinders Ranges (Gell and Bickford, 1996). The palynological evidence implies that precipitation associated with the lower laminated unit between 346-185cm must have been largely derived from westerly anticyclonic air-masses. A sharp decrease of Cupressaceae pollen is offset by an increase in Eucalyptus pollen at 185cm. This is interpreted to indicate a reduction in winter rainfall while summer rainfall was still low (Singh and Luly, 1991). At present, Callitris-dominated woodlands are best developed in semi-arid eastern Australia with annual rainfall averaging between 300-650mm (Luly, 2001). The closest ages for this transitional period are 16.17 0.42 ka (199-169cm, for carbonate) and 16.16 0.19 ka (180-165cm, Bowler et al., 1986, Table 2). Winter rainfall must have been plentiful enough to allow for a more abundant growth of trees than today or after ~16 ka. The decline in Cupressaceae pollen is closely followed by a dramatic increase in carbonised particles (Luly, 2001, Fig.6), (Fig. 2). Extremely low pollen counts of fern spores below a depth of 120cm suggest that during sedimentation of the fine-grained lithofacies humidity was low and summer months have been as dry as or even drier than at present (Singh and Luly, 1991). The sporadic presence of pollen from taxa typical of temperate areas further to the south indicates that
12
temperatures were lower, an interpretation later supported by independent studies further north (Miller et al., 1997). A substantial rise in Gramineae, accompanied by the first appearance of subtropical to tropical taxa from the summer rainfall zone, sets in at 120cm, just 20-5cm below an organic carbon sample dated to 16.11 0.32 ka (Bowler et al., 1986), (Fig. 2). This suggests that the onset of summer monsoon incursions to latitudes of ~31S took place ~16 ka. Although relative vegetation stability is recorded until a depth of 34cm, grass values constantly decline until ~45cm, where they suddenly recover. From 34cm to the top of the sequence, a rapid decline of Gramineae coincides with a rise in pollen from halophytes and ephemerals and the onset of present playa conditions, suggesting a reduction in the frequency of tropical rainfall incursions. The transition at 34cm postdates 5.06 0.15 ka, as indicated by an AMS sample from 38-35cm (Luly and Jacobsen, 2000). The palynological evidence from Lake Frome therefore suggests that the Flinders Ranges have been within the winter rainfall zone, except for the interval between ~16-5 ka.
The piedmont plain and palaeosol record

The Lake Frome record may be compared with the record of alluvial and aeolian deposition on the piedmont plains and within the ranges. The widespread occurrences of distinct pedogenic carbonate segregations separate the depositional records into genetic sequences that are ideal for comparison. The carbonate segregations occur as a series of morphologically distinct horizons at or just below sedimentary discontinuities. A subhumid climate with annual rainfall between 200-600 mm (Goudie, 1983), high rates of evaporation or evapotranspiration (Wright and Tucker, 1991), and slow sedimentation rates relative to rates of pedogenesis (Wright, 1990; Kraus, 1999) are conducive to pedogenic carbonate precipitation. Hence, these fossil soils register intervals of landscape stability between erosional and depositional episodes. In contrast to the depositional records from playa lakes, alluvial fans and valley-fill formations, the record from pedogenic carbonates provides information about moist intervals of minimal deposition. The piedmont plains around the Flinders Ranges constitute a link between valley-fill formations within the ranges and the flanking terminal depocentres of the salinas Lake Torrens and Lake Frome (Fig. 1). The Lake Torrens piedmont plain is mantled by some of the best-developed alluvial fans and pediments in Australia (Williams, 1973; Bourne and Twidale, 1998). General stratigraphic observations suggest that a prolonged interval of fan aggradation was succeeded by aeolian accession and shorter intervals of fan dissection, all separated by periods of calcareous soil-
13
formation. Fan dissection prevails today and is indicative of overall aridity and low stream-discharge (Williams, 1973). Three morphologically distinct pedogenetic phases are recorded within the alluvial and aeolian deposits (Williams and Polach, 1971; Williams 1973). Their common diagnostic attributes consist of nodules of authigenic carbonate, concentrated within distinct horizons. On the Lake Torrens piedmont plain, the chronology is mostly based on conventional radiocarbon dating of these carbonate concretions (Williams and Polach, 1971; Williams, 1973). In a similar study, the ages of longitudinal dunes and alluvium along the eastern side of Lake Frome were estimated (Callen et al., 1983). Pedogenic carbonates are often unreliable in terms of ages of the calcrete segregation processes. However, by excluding whole-soil samples and samples from truncated buried soils, and by sampling only fine-grained carbonate (<20m) of the lower Bc-horizons from locations above present flood levels, chronological sequences can be obtained that correlate remarkably well with the other palaeo-environmental evidence. Whenever possible, pedogenic age estimations are paired with 14Cdates from organic carbon samples (Williams, 1973; Lampert and Hughes, 1988; Cock et al., 1999) and thermoluminescence (TL) dating (Gardner et al., 1987; Williams et al., 2001). A major period of alluviation and fan building in the piedmont plains is dated by two radiocarbon ages for detrital wood to 39.33 2.51 and >42.54 ka, extending back beyond the range of radiocarbon dating. The alluvial fan deposits, with a maximum exposed thickness of ~15m, consist of mudstones with lenticular pebble- to boulder-conglomerates in a red matrix of clayey silts and sands. Towards the range-front, the so-called Pooraka Formation passes without a break in surface slope into talus breccias with a similar matrix and into mottled grey and yellow fine-grained valleyfills within the gorges (Twidale, 1966; GE Williams, 1973; DLG Williams, 1982; MAJ Williams et al., 2001). Along the eastern side of Lake Frome, apparently coeval fluviatile sands of the Eurinilla Formation were deposited in channel, overbank and deltaic environments (Callen et al., 1983). Within the fluviatile parent material of the Pooraka Formation, a distinct 1-2m thick pedogenic carbonate horizon is developed impregnating the fine-grained matrix and coating the larger clasts. Locally, the calcareous bands grade laterally into massive laminated groundwater calcrete. Excluding whole-soil carbonate samples, eight 14C-dates from the so-called Wilkatana Palaeosol were obtained from various stratigraphic settings. The radiocarbon dates range from ~39.81 2.11 to 24.39 0.46 ka, grouping around a median of 33.75 1.08 ka (Williams, 1973, Table 2). East of Lake Frome, coeval calcareous segregations form several individual layers of hard calcareous nodules and cement
14
the fluviatile sediments of the Eurinilla Formation. Collectively referred to as the Pinpa Palaeosol, they were dated by more than 17 samples. The 14C-dates span an interval between 44.26 2.52 and 26.2 0.28 ka. The considerable spread of the dates may reflect an extended period of pedogenesis consisting of multiple calcrete forming events which resulted in composite profiles (Candy et al., 2003). It is likely that pedogenesis coincided with ongoing and laterally shifting fan building episodes of the Pooraka Formation. As with the overall fan aggradation, the pedogenetic processes could have been initiated earlier than indicated by the
14
C-dates dates. Taken at face value, the
radiocarbon ages for carbonates nodules within the fluviatile Pooraka and Eurinilla Formations suggest three major intervals of soil formation towards ~39, ~31 and ~27 ka. Subsequent to a decline in pedogenetic processes ~27 ka, an undated interval of dune building and fan dissection is stratigraphically recorded in the Lake Torrens piedmont plain. A "red clayey horizon as much as 1m thick" (Williams and Polach, 1971, 3073) mantles the upper reaches of the fans. Towards the base, the aeolian deposits develop into 2-5m thick longitudinal dunes of the Lake Torrens Formation, erosively overlying the Wilkatana Palaeosol. On the eastern side of Lake Frome, aeolian deposits form partly consolidated cores of longitudinal and lunette dunes. The aeolian unit, which unconformably overlies the Pinpa Palaeosol and is locally covered by a veil of modern loose drift sands, is referred to as the Coonarbine Formation (Callen et al., 1983). Within these finegrained aeolian formations, horizons of carbonate concretions occur, with 14C-dates for the so-called Motpena Palaeosol from the Lake Torrens piedmont plain ranging from 19.3 0.24 to 14.01 0.16 ka (Williams, 1973, Table 2). Within the Coonarbine Formation, two episodes of carbonate segregation are distinguished. Based on their close morphological resemblance and similar scope of
14
C-ages, the stratigraphically lower and therefore older calcareous horizons of the Moko Palaeosol
can tentatively be linked with the Motpena Palaeosol from the western piedmont plain. Excluding two spurious samples, all remaining five 14C-dates fall within the range from 19.36 0.18 to 17.32 0.51 ka, with two distinct probability peaks at ~19.4 and ~18.2 ka. Holocene alluviation is recorded by several conglomeratic terraces, deposited within the main channels of the alluvial fans, and by abandoned fan segments on the western piedmont plain. Collectively termed the Eyre Gravel Formation (Williams and Polach, 1971; Williams, 1973), their lithology of the pebble- to boulder-gravels with an overall light reddish brown colour has provided a
14
C-date of 5.88 0.12 ka for a charcoal sample collected 40cm below the top of the oldest and
highest terrace of the complex. Laterally equivalent sandy deposits, flooring the distal interdune corridors of the Lake Torrens Formation and truncating the Wilkatana and Motpena palaeosols, are collectively called the Thompson Creek Formation. The latter consists of two up to 2m thick
15
members which are separated by a vesicular layer of a desert loam and the pedogenetic horizon of the so-called Nacoona Palaeosol consisting of soft calcareous tubules and patches of "earthy carbonate" (Williams and Polach, 1971, 3075). Charcoal from the parent material is dated to 6.86 0.13 ka. At present, the youngest pedogenic formation is capped by unweathered gravels and sands which are the product of ongoing fan dissection. Further to the N, the Motpena Palaeosol is buried by red, moderately consolidated aeolian quartz-sands. The youngest interval of pedogenic carbonate segregation in the aeolian Coonarbine Formation remains unnamed. While two dates obtained at the northern tip of the Lake Frome are close to ~12.35 ka, samples from morphologically similar soft small calcareous rhizomorphs, crust and patches close to the southern tip present ages of 9.09 0.19 and 8.00 0.13 ka. Stratigraphic accounts from the eastern piedmont plain of the Flinders Ranges are rare. A channel bank section of the Balcoracana Creek discharging into the south-western end of Lake Frome has been logged and dated (Gardner et al., 1987; Lampert and Hughes, 1980, 1988), (Fig. 1). Shell samples from a pedogenic horizon described as a "probably sheetwash layer of blocky fine to medium sand rich in land snail shells" (Gardner et al., 1987, 350) ~210cm below top were radiocarbon dated to 16.34 0.9 ka and 14.78 0.15 ka. The former date corresponds well with a carbonate sample from the underlying palaeosol dated to 16.36 0.36 ka (Gardner et al., 1987). These ages lie in-between those inferred for the Moko Palaeosol to the east and the Motpena Palaeosol to the W. Below modern loose drift sands, soft carbonate nodules were concentrated in the youngest pedogenic horizon which, based on the TL-age of 5.5 1.6 ka from similar depth (Gardner et al., 1987), is likely to correlate closely with the unnamed youngest soil-forming interval with the dunefields east of Lake Frome (Callen et al., 1983).
Fine-grained valley-fill formations of the Flinders Ranges

In many cases, coalescing fan deposits from the piedmont plains extend upstream, breaching the scarp of the ranges where they form a distinctive and recurring sequence of valley-fill deposits.
Hookina Creek
Laminated fine-grained valley-fill formations in the Flinders Ranges were first described from Hookina Creek (Twidale, 1966), (Fig. 1), and later correlated with fan deposits of the Pooraka Formation from the adjacent piedmont plain further downstream (Williams, 1973) on the basis of a single radiocarbon date. The charcoal sample, dated to 38.36 2.18 ka, was derived from 9-8m below the top of the valley-fill formation, further referred to as the Hookina Silts. The modern creek
16
meanders within deeply incised bedrock channels, eroding the unconsolidated fine-grained sediments and exposing their near-horizontal bedding planes. Three main facies were differentiated from a sedimentary sequence exposed in a ~15m vertical river bank section (Williams, 1982). Basal conglomerates rest unconformably upon bedrock. Two overlying units of fine-grained, unconsolidated valley-fill alluvium form the bulk of the Hookina Silts. The lower unit consists of clayey to sandy laminae of variable colour that can be traced for tens of metres laterally. The sediments are poorly sorted and display localised current-bedded structures. The upper unit consists of uniform red-brown silt and sand with occasionally inset sandy or gravelly channel-fill deposits. It is characterised by calcareous laminations. Aquatic mollusca described from the lower unit become increasingly rare in the upper unit and are replaced by fossil shells of land snails. Within the entrenched modern creek channel, low gravel terraces abut alternate sides. A single charcoal sample, radiocarbon dated to 2.16 0.14 ka (Williams, 1982), suggests that the terraces are late Holocene phenomena, possibly deposited coevally with the upper members of the Eyre Gravel and Thompson Creek Formations in the adjacent foreland (Williams, 1973). Based on the observation of fresh-water mollusca (Corbiculina angasi, Pettancylus sp. and Physastra sp.), the Hookina Silts were initially interpreted as "accretion-gley deposited in a marshy area of local high, fluctuating water table" (Williams, 1973). In a subsequent study, an assemblage of 2270 individuals of aquatic and terrestrial species was described from a sandy channel-fill deposit ~5m below top (Williams, 1982). The most common aquatic mollusc is the Hydrobiidae Potomopyrgus sp (85%), followed by the Planorbidae Gyraulus sp. (9%), both living in a habitat of still to flowing water and tolerant of low to moderate salinity levels. The former still occurs within the modern creek bed, where it is found clinging to the walls of seasonally flooded rock pools.
Brachina Creek
Comparable sequences occur throughout the Flinders Ranges, forming extensive gently sloping valley-fills upstream of resistant ridges and within narrow rock-cut gorges. The best dated sequence is that of the Brachina Creek ~45km north of Hookina Creek, here termed the Brachina Silts (Fig. 1). At the 'Slippery Dip' type section, a sedimentary sequence consisting of three distinct units is exposed in a ~9m high near-vertical cliff face (Williams et al., 2001, Fig.7). The basal Lower Unit consists of largely uniform silty clay with discrete lenses of gravel. Shells of the aquatic freshwater species Glyptophysa sp. are abundant but decrease towards the distinctively brown top of the unit that is interpreted to represent a palaeosol (Cock et al., 1999). Current microfossil studies describe an assemblage of free-swimming and benthic freshwater mollusca and ostracoda, testifying a diversity of habitats. This is followed by ~5m of alternating light- and dark-coloured bands that form
17
the Upper Unit, grading towards a more sandy, coarsely bedded 'red layer' forming the top of the sequence and the modern near-level valley floor While the total number of ostracods and mollusca rapidly increases and then fluctuates towards the top of the sequence, the Upper Unit is characterised by a marked decline in diversity of the assemblage and by a prevalence of broken shell fragments indicative of high-energy transport. The lithology of the Lower and Upper Units consists of similar silty clay, with occasional thin discontinuous lenses of local scree structuring the more massive bedding planes. While several depositional models are possible (Haberlah, 2006), the pattern of laminations consisting of massive layers alternating with calcareous bands and veneers of organic mud is indicative of slackwater couplets. They could be explained by back-flooding of the tributary mouth just upstream of the confining gorge (Zawada, 1997). Initially, the Brachina Silts were believed to represent a Pleistocene lake (Callen and Reid, 1994; Cock et al., 1999). A later theodolite survey established that the surface of the Silts has a sloping gradient parallel to the modern bedrock channel and higher terraces, and therefore was re-interpreted as a former wetland (Williams et al., 2001). The possibility that the upper sequence represents a terminal Pleistocene palaeoflood record is currently under study. While aggradation of the Lower Unit terminates ~33 ka, as inferred from AMS 14C-dates 33.94 0.54 ka for charcoal and 32.61 1.6 ka for shell, supported by the minimum OSL age 32.8 2.8 (Williams et al., 2001), the deposition of the bulk of fine-grained sediments above the 'palaeosol' seems to reflect from two or three groups of episodes of rapid deposition (Fig. 3). The calibrated dates for thirteen AMS shell and charcoal samples for the Silts of the Upper Unit from 'Slippery Dip' and 'Sections G' group tightly around probability peaks ~20.2 ka, 19.4 ka and 18.2 ka (Williams et al., 2001, Table 1), (Fig. 3). The chronostratigraphy is supported by three OSL ages ranging from 20.3 1.0 ka to 18.0 1.2 ka (Williams et al., 2001, Table 3). Rapid sedimentation of the Brachina Silts is analogous and contemporaneous with the deposition of the fine-grained lacustrine facies in Lake Frome (Fig. 2). Williams and Nitschke (2005) demonstrated that loess mantles accumulating within the catchment in the period between the alluviation of the Lower and Upper Units contributed the bulk of the finegrained sediments (Fig. 3), thereby explaining the marked difference to the contemporary fluvial regime characterised by coarse bedload transported during erratic floods.
18
Fig. 3) 2-D dispersion of calibrated
14
C-dates for the Brachina Silts (modified diagram of
14
C-
calibration program 'CalPal May 2006', www.calpal.de/ [1 November 2006]) using the calibration curve 'Fairbanks Aug 2005' (Fairbanks et al., 2005). Indicated depositional units are collated with 'Vostok ice core 1999' temperature and dust content records (Petit et al., 1999). Samples of the 'Slippery Dip' type section are listed in black and those of 'Section G' in red (dates from Williams et al., 2001).
Discussion
Deglacial environmental responses in southern Australia are complex. In order to develop a sound understanding of palaeoclimatic changes following LGM peak aridity, the range of widely occurring terminal Pleistocene and Holocene aeolian, fluviatile and lacustrine fine-grained sediments needs to be explained. The Lake Frome data are based on a limited number of cores largely from one location,
19
with a low-resolution conventional radiocarbon chronology. However, clearer patterns emerge when the Lake Frome data are compared with the records from the dunes, piedmont deposits and valleyfills in and around the Flinders Ranges
Interpretation of silty lacustrine sequence of Lake Frome

The Lake Frome cores shows a change from generally sandy to fine-grained sedimentation between ~20 ka and ~16 ka and possibly around ~7 ka. Previous workers interpreted these changes in terms of fluctuating lake-levels, with the intervals of fine-grained sedimentation considered to reflect deep-water conditions (Bowler, 1986; Ullman and Collerson, 1994). However, the only direct indicator of higher lake levels is an elevated beach ridge on the southern margin of Lake Frome, with an inferred minimum age from recrystallised shells of 41.56 1.49 ka (Gardner et al., 1987). The beach ridge is probably contemporary with the Lower Sands of the lacustrine sequence and coincides with an interval of widespread alluvial deposition and high lake-levels in southeast Australia, before giving way to increasing aridity (Bowler and Wasson, 1984). The LGM was a time of widespread aeolian activity and deflation evident in the Lake Eyre basin (Magee, 1998) to the north (Fig. 1), and in the gypsum dune formations dated to ~23 ka, that form the present islands of Lake Frome. The high silt content of the fine-grained lithofacies prompted Bowler et al. (1986) to propose aeolian activity coeval with the resumption of lacustrine deposition. Cores from the Tasman Sea show high rates of aeolian dust flux during the LGM (Hesse, 1994), but the Vostok dust record reveals a sharp and progressive decline in dust content after ~21 ka (Fig. 3). Ker (1966) and Draper and Jensen (1976) noted the role of present-day extreme rainfall and runoff from the eastern Flinders Ranges in providing water to Lake Frome. Sedimentary evidence shows that the lake sediments were primarily derived from the Flinders Ranges. Luly (2001) considered that the high quantities of charcoal characteristic of the fine-grained deposits were probably washed in with the sediment. The chronostratigraphy of Bowler et al. (1986) showed far higher sedimentation rates for the fine-grained lithofacies than for the coarser Holocene sediments. Since the playa lake occupies much of the basin and relief in the immediate catchment is subdued, we would expect low sedimentation rates. This contradiction prompts us to question the prevailing assumption that the fine sediments represent a distal lacustrine facies and hence deep water conditions. An alternative interpretation of the fine-grained lacustrine facies and its rapid sedimentation rates is based on the assumption that these sediments were derived from the erosion of loess mantles in and around the ranges. Such erosion would lead to an influx of very high suspension loads to the basin over short intervals. Bromide deficiencies in the brines of all major playa lakes in the region (Draper and Jensen, 1976; Johns, 1968), the lack of saline bedrock outcrops and Kers (1966)
20
observations of steadily increasing salinity in shallow aquifers towards the Lake suggest that the salts are derived from second cycle brines. It is proposed that they were transported intermittently along the southeast dust path (Bowler, 1976) during glacial intervals from the exposed continental shelf and desiccated playa lakes (Crocker, 1946). LGM deflation of gypseous sediments flooring Lake Torrens is recorded in source-bordering dunes of gypsum sands deposited on the eastern playa lake margin (Williams, 1973; Schmid, 1985). Diagenetic gypsum bands in the laminated section of the lower part of the fine-grained lithofacies, extending from ~120cm beyond the inferred disconformity at 346cm, are likely to reflect rapid evaporation events (Draper and Jensen, 1976). This inference is consistent with the dry summer months interpreted from the pollen evidence for this sedimentation interval (Singh and Luly, 1991). Both the salinity and the strontium isotope records discussed earlier support this model. According to the gypsum chemistry, the first transition from hypersalinity to low salinity was a rapid but brief event, indicative of a sudden concentration of surface waters reaching Lake Frome between ~21-20 ka. This was followed by a second period of low-salinity towards 16 ka. Apart from two gypsum crystals at 97cm and 93cm, indicating a low-salinity brine chemistry, no further gypsum precipitated until the top of the Holocene 'clayey band'. Na+/Ca2+ ratios of these two gypsum crystals suggest again a low-salinity precipitation environment (Ullman and McLeod, 1986, Table 1), which is in accord with the pollen record indicating the resumption of summer rainfall (Singh and Luly, 1991). Ullman and McLeod (1986) interpreted the general absence of gypsum bands between and after these times in terms of a freshwater phase with salinity levels below that of gypsum precipitation, with a subsequent loss of gypsum crystals above the 'clayey band' through deflation. However, the findings could equally reflect the strong influence exerted by sediment type on the occurrence and spatial distribution of minor elements (Draper and Jensen, 1976). Accordingly, during intervals in which an allochthonous loess-derived suspension load was rapidly accumulating in the playa lake, the brine chemistry was altered by the addition of Ca2+-ions reducing the excess of SO42--ions, established throughout the whole salinity range in the present groundwaters and brines of Lake Frome (Ullman and McLeod, 1986). If so, the gypsum-based palaeosalinity record would register brief intervals of fluctuations in the brine chemistry caused by inundations with runoff waters introducing a Ca2+-rich loess-derived suspension load to the lake system. Assuming that the 'clayey band' interposed between the sandy units represents eroded loess mantles, the gradual decline and termination of this particular depositional regime towards 16 ka can be explained by environmental changes in the catchment area and by a shift of the dominant precipitation regime. Lower sedimentation rates in the upper 65cm of the fine-grained lithofacies
21
coincide with a marked reduction in the occurrence of laminations and gypsum bands. While Singh and Luly (1991), in line with previous workers, interpret this finding to indicate shallower depositional conditions, they could equally result from a more restricted supply of allochthonous fines. The return to a coarser-grained shore facies, consisting of autochthonous weathering products, could reflect two interdependent trends: a gradual depletion of easily erodible loess mantles and stabilisation of its remnants by vegetation. The pollen record is consistent with this inference. Dramatic changes in the vegetation cover occurred concomitantly with the termination of the fine-grained lithofacies. During this transitional period, a reduction in intensity and magnitude of frontal precipitation events induced stress on the woody elements, witnessed by the sharp decline of Callitris and a subsequent increase in charcoal. The collapse of Callitris woodlands, as opposed to the synchronous increase in Eucalyptus pollen, may indeed reflect Aboriginal burning (Luly, 2001), but evidence of human presence in the region is sparse and not well dated. Thereafter, rising temperatures, more humid conditions and a southward shift of monsoonal incursions led to the colonisation of the region by subtropical grassland communities from the north (Fig. 2). While Callitris woodlands had probably only supported a sparse understorey (Singh and Luly, 1991), which would have little affected runoff, Gramineae exert a stabilising effect on unconsolidated sediment. By increasing surface roughness and infiltration rates, they are likely to have reduced runoff flow velocities and erosion rates and possibly favoured linear incision. Hence, whatever remained of the loess mantles was at this point stabilised and retained within the catchment. The deterioration of the stabilising grass cover with the gradual retreat of summer monsoon incursions was interrupted by a short-lived recovery in the pollen count of Gramineae towards ~5 ka. Apparently simultaneously, a second flux of fine-grained sediments is recorded in the Lake Frome record as the 'clayey band' (Fig. 2), and reflected in the parent material of the youngest palaeosols on the piedmont plain. It is possible that this inferred short phase of increased precipitation (Singh and Luly, 1991) led to renewed incision of the valley-fills in the Flinders Ranges, again providing fine sediments to the depocentres.
Interpretation of silty sediments within the drainage lines and piedmont plains of the Flinders Ranges
Subsequent to a more than 10 ka long interval of minimal deposition, the fluvial system of the Brachina Creek was suddenly accumulating a massive sequence of loess-derived alluvium. The bulk of the Brachina Silts appears to have been deposited within ~2 ka, possibly within two or three groups of episodes of rapid deposition ~20.2 ka, 19.4 and 18.2 ka (Fig. 3). Rapid sedimentation rates
22
(Fig. 2), and the morphologic context of the bedding planes levelling out pre-existing relief upstream of confined gorges suggests palaeoflood as a likely depositional model (Haberlah, 2006). The pollen evidence from Lake Frome indicates that these terminal LGM floodwaters may reflect frontal winter precipitation events. At the same time, the dunes east of Lake Frome were probably stabilised during the soil-forming interval of the Moko Palaeosol (Callen et al., 1983), with peaks at ~19.4 and ~18.2 ka. This period of widespread pedogenesis further coincides with the rapid sedimentation rates established for the fine-grained lacustrine facies. As soil-forming intervals are indicative of surface stability, a coeval direct input from wind-blown dust (Bowler et al., 1986) to Lake Frome seems unlikely. The youngest dates for the Moko and Motpena Palaeosols coincide with colonisation of the region by subtropical grasslands as a consequence of increasing temperatures and the onset of summer monsoon incursions to these latitudes. Consequent stabilisation of the loess remnants and dunes at the base of the alluvial fans prevented further large-scale erosion. Gradually increasing aridification caused a decline in Gramineae before their sudden albeit shortlived recovery prior to ~5 ka. As the vegetation cover became sparser, aeolian sands were mobilised again as witnessed by the capping of the Motpena palaeosol on the northern Lake Torrens piedmont plain (Williams, 1973). Coinciding with the deposition of the lacustrine 'clayey band', weak soil formation in the distal reaches of the alluvial fans of the western piedmont plain took place, with the Nacoona Palaeosol postdating ~7 ka. The fluviatile fine-grained parent material of the youngest pedogenic interval, and the conspicuous 'clayey band' interrupting the coarser lacustrine sedimentation in Lake Frome, could reflect the erosive effect that the final revival of monsoonal incursions had on the once more largely barren landscape. The source of the pedogenic carbonate segregations is as unresolved as the provenance of the gypsum in the valley-fills and playa lake deposits. Callen et al. (1983) suggest fluctuating groundwater tables as the most likely cause of carbonate precipitation in the palaeosols, but the carbonate horizons of the Lake Torrens piedmont plain developed in well-drained locations (Williams, 1973). We consider that the carbonate is of allochthonous aeolian origin, blown in and deposited as loess accessions (Williams and Nitschke, 2005). Strontium isotope studies (Dart et al., 2005; Lintern et al., 2006) support early suggestions of a marine provenance and aeolian introduction of carbonate constituents within the southern Australian regolith (Crocker, 1946; Fairbridge and Teichert, 1953).
23
Conclusion
Facies changes in Australian playa lakes have been used to infer lake-level fluctuations. We propose an alternative interpretation of such changes in lacustrine records as a result of variations in sediment load brought in by local streams. The lacustrine sedimentary sequence of Lake Frome correlates well with intervals of erosion and stabilisation of loess mantles inferred from valley-fill formations within the Flinders Ranges and intervals of calcareous soil formation on the slopes of adjacent piedmont plains. Erosion is likely to have been triggered by intense rainfall on mostly bare surfaces, a scenario supported by the pollen data. An onset of terminal Pleistocene frontal winter precipitation as early as ~20.2 ka stripped the loess mantles from the slopes of the Flinders Ranges and adjacent piedmont plains. The eroded finegrained sediment was rapidly washed into valley bottoms and structural depressions, choking the gorges and in places causing backflooding and prolonged ponding. Similar deposits were washed from the piedmont plains into the terminal playa lakes. According to our interpretation, these fine sediments in the lake floor do not represent a distal lacustrine facies indicative of deep lake-level environments, nor are they suggestive of prolonged humid conditions. On the contrary, the finegrained lacustrine depositional regime was prone to rapid evaporation as reflected by the closely spaced displacive gypsum bands. The rapidly accumulating fine-grained lithofacies rather record the unique effect of terminal Pleistocene floods on a highly erosive loess-mantled environment. From ~16 ka onwards, a rise in temperatures and regular monsoonal incursions resulted in the colonisation of the loess mantle remnants by grassland. A dramatic reduction in suspension load is likely to have caused the termination of the aggradational regime of the fine-grained valley-fill formations within the ranges. At the same time, the deposition of a sandy lithofacies began in Lake Frome, characterised by an absence of gypsum precipitation and slow sedimentation rates. This coarser lacustrine sedimentation regime was only interrupted once over a short-lived interval terminating ~5 ka reflected in the intercalated 'clayey band'. Both pollen and palaeosol records suggest that it is linked to the return of a dominantly frontal precipitation regime to the region that was preceded by increasing aridification and a demise in stabilising plant cover. Fine-grained deposits of similar age and appearance occur as remnants within the valleys of the Great Escarpment of Namibia and the Sinai Peninsula of Egypt and appear to have been laid down during discrete flood events (Srivastava et al., 2006; Rgner et al., 2004). Loess was established as the allochthonous source for both locations (Yaalon and Dan, 1974; Eitel et al., 2001).
24
Evidence for sporadic but intense precipitation events during the terminal Pleistocene include TL dated river deposits in the Riverine Plains (Page et al., 1996), (Fig. 1). Just to the south, a highresolution 500 ka record of relative water excess necessary for speleothem formation was established using
230
Th/234U dating. Accordingly, "greater effective precipitation levels for the
southeastern interior of Australia" (Ayliffe et al., 1998, 149) only occurred during transitional periods between glacial maxima and warm interstadial and interglacial climate regimes. Barrows et al. (2001) have dated the three most recent glacial advances in the uplands of southeast Australia to 32 2.5 ka, 19.1 1.6 ka and 16.8 1.4 ka, using the cosmogenic isotope
10
Be, (Fig. 1). Apart from low
temperatures, advances of glaciers require substantial amounts of precipitation. Inferred periods of increased snowfall seem to coincide with the timing of fluvial aggradation of the Lower and Upper Units of the Brachina Silts in the Flinders Ranges, and the timing of their subsequent incision. Still within the Eastern Highlands, lake level fluctuations are preserved in the elevated beach ridges of Lake George to the north, (Fig. 1). Although the chronostratigraphy is complex (Coventry and Walker, 1977; Singh et al., 1981; Pietsch, 2006), it seems that the lake attained its highest still-stand at ~33 ka, followed by a number of high but progressively lower water levels after ~20 ka. The beach ridges of Lake George prompted Galloway (1965) to propose a minevaporal model to explain pluvial features in an otherwise arid glacial Australian landscape. We suggest that the fluctuating lake levels could reflect the onset of deglacial rainfall events in a largely unvegetated cold landscape. Seasurface temperatures around Australia were up to 9-7C lower than today until 20.5 1.4 ka (Barrows and Juggins, 2005). The marine oxygen isotope records show that maximum global ice volumes were only attained some 2 ka later. Temperature amelioration inferred from the Vostok ice core (Petit et al., 1999), (Fig. 1), and from temperature-dependent amino-acid racemization in 14Cdated emu eggshells for low-latitudes of Australia did not begin until ~17 ka. The 3.5 ka interval inbetween is likely to have been characterised by moisture transport towards the continent.
25
References
Allison GB, Barnes CJ. 1985. Estimation of evaporation from the normally "dry" Lake Frome in South Australia. Journal of Hydrology 78: 229-242. Ayliffe LK, Marianelli PC, Moriarty KC, Wells RT, McCulloch MT, Mortimer GE, Hellstrom JC. 1998. 500 ka precipitation record from southeastern Australia: evidence for interglacial relative aridity. Geology 26: 147150. Barrows, TT, Stone JO, Fifield LK, Cresswell RG. 2001. Late Pleistocene Glaciation of the Kosciuszko Massif, Snowy Mountains, Australia. Quaternary Research 55: 179-189. Barrows TT Juggins S. 2005. Sea-surface temperatures around the Australian margin and Indian Ocean during the Last Glacial Maximum. Quaternary Science Reviews 24: 10171047. Bourne JA, Twidale CR. 1998. Pediments and alluvial fans: genesis and relationships in the western piedmont of the Flinders Ranges, South Australia. Australian Journal of Earth Sciences 45: 123135. Bowler JM. 1976. Aridity in Australia: Age, origins and expression in aeolian landforms and sediments. Earth-Science Reviews 12: 279-310. Bowler JM, Wasson RJ. 1984. Glacial age environments of inland Australia. In Late Cainozoic palaeoclimates of the southern hemisphere, Vogel JC. (ed). Proceedings of an International Symposium held by the South African society for Quaternary research, Swaziland 29 August - 2 September 1983; 183-208. Bowler JM, Huang Q, Chen K, Head J, Yuan B. 1986. Radiocarbon dating of playa-lake hydrologic changes: Examples from northwestern China and central Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 54: 241-260. Callen, RA. 1984. The islands of Lake Frome. Geological Survey of South Australia Quarterly Geological Notes 88: 2-8. Callen RA, Wasson RJ, Gillespie R. 1983. Reliability of radiocarbon dating of pedogenic carbonate in the Australian arid zone. Sedimentary Geology 35: 1-14. Callen RA, Reid PW. (comp) 1994. Geology of the Flinders Ranges National Park. South Australian Geological Survey: Special Map 1:75,000.
26
Candy I, Black S, Sellwood BW, Rowan JS. 2003. Calcrete profile development in Quaternary alluvial sequences, southeast Spain: implications for using calcretes as a basis for landform chronologies. Earth Surface Processes and Landforms 28: 169-185. Chivas AR, Bowler JM. 1986. Introduction Palaeoclimatology, Palaeoecology 54: 3-6. Cock BJ, Williams MAJ, Adamson DA. 1999. Pleistocene Lake Brachina: a preliminary stratigraphy and chronology of lacustrine sediments from the central Flinders Ranges, South Australia. Australian Journal of Earth Sciences 46: 61-69. Coventry RJ, Walker D. 1977. Geomorphological significance of Late Quaternary deposits of the Lake George area, N.S.W. Australian Geographer 13: 369-376. Crocker, RL 1946. Post-Miocene climatic and geologic history and its significance in relation to the genesis of major soil types of South Australia. CSIR Bulletin 193: 56. Dart RC, Barovich KM, Chittleborough D. 2005 Pedogenic carbonates, strontium isotopes and their relationship with Australian dust processes. In Regolith 2005 - Ten years of CRC LEME, Roach IC (ed). Proceedings of the CRC LEME regional regolith symposia 2005, Adelaide and Canberra November 2005; 64-66. Draper JJ, Jensen AR. 1976. The geochemistry of Lake Frome, a playa lake in South Australia. Bureau of Mineral Resources, Journal of Australian Geology and Geophysics 1. 83-104. Eitel B, Blmel WD, Hser K, Mauz B. 2001. Dust and loessic alluvial deposits in northwestern Namibia (Damaraland, Kaokoveld): sedimentology and palaeoclimatic evidence based on luminescence data. Quaternary International 76/77. 57-65. Fairbanks RG, Mortlock RA, Chiu T-C, Cao L, Kaplan A, Guilderson TP, Fairbanks TW, Bloom AL, Grootes PM, Nadeau M-J. 2005. Radiocarbon calibration curve spanning 10,000 to 50,000 years BP based on paired 230Th/234U/238U and 14C dates on pristine corals. Quaternary Science Reviews 25: 1781-1796. Fairbridge RW, Teichert C. 1953. Soil horizons and marine bands in the coastal limestones of Western Australia. Royal Society of New South Wales, Journal and Proceedings 86: 6887 Frome EC. 1899. Extract from the government gazettes of 1843 and 1844. Proceedings of the Royal Geographical Society of South Australia 4: 132. The SLEADS project. Palaeogeography,
27
Galloway RW. 1965. Late Quaternary climates in Australia. Journal of Geology 73: 603-618. Gardner GJ, Mortlock AJ, Price DM, Readhead ML, Wasson RJ. 1987. Thermoluminescence and radiocarbon dating of Australian desert dunes. Australian Journal of Earth Sciences 34: 343357. Gell PA, Bickford S. 1996. Vegetation. In Natural history of the Flinders Ranges, Davies M, Twidale CR, Tyler MS. (eds). Royal Society of South Australia: Adelaide; 86-101. Goudie AS. 1983. Calcrete. In Chemical sediments and geomorphology: precipitates and residua in the near-surface environment, Goudie AS, Pye K (eds). Academic Press: London; 93-131. Haberlah D. 2006. Depositional models of late Pleistocene fine-grained valley-fill formations in the Flinders Ranges, South Australia. In Regolith 2006 Consolidation and Dispersion of Ideas, Fitzpatrick RW, Shand P. (eds). Proceedings of the CRC LEME Regolith Symposium, Hahndorf November 2006; 122-127. Hesse PP. 1994. The record of continental dust from Australia in Tasman Sea sediments. Quaternary Science Reviews 13: 257-272. Johns RK. 1968. Investigation of Lake Torrens and Gairdner. Geological Survey of South Australia Report of Investigations 31: 90. Ker DS. 1966. The hydrology of the Frome Embayment, South Australia. Geological Survey of South Australia Report of Investigations 27: 98. Kraus MJ. 1999. Paleosols in clastic sedimentary rocks: their geologic applications. Earth-Science Reviews 47: 41-70. Lacey CJ. 1972. Factors influencing occurrence of cypress pine regeneration in New South Wales. Technical Paper, Forestry Commission of New South Wales 21: Sydney; 20. Lampert RJ, Hughes PJ. 1980. Pleistocene archaeology in the Flinders Range: research prospects. Australian Archaeology 10: 11-20. Lampert RJ, Hughes PJ. 1988. Early human occupation of the Flinders Ranges. Records of the South Australian Museum 22: 139-168. Lintern MJ, Sheard MJ, Chivas AR. 2006. The source of pedogenic carbonate associated with goldcalcrete anomalies in the western Gawler Craton, South Australia. Chemical Geology 235: 299324.
28
Luly JG. 2001. On the equivocal fate of Late Pleistocene Callitris Vent. (Cupressaceae) woodlands in arid South Australia. Quaternary International 83-85: 155-168. Luly JG, Jacobsen G. 2000. Two new AMS dates from Lake Frome, arid South Australia. Quaternary Australasia 18: 29-33. Madigan CT. 1929. An aerial reconnaissance int the S.E. part of Central Australia. Proceedings of the Royal Geographical Society of South Australia 30: 83. Magee JM. 1998. Late Quaternary environments and palaeohydrology of Lake Eyre, arid central Australia. Unpublished PhD thesis: Australian National University. Miller GH, Magee JW, Jull AJT. 1997. Low-latitude glacial cooling in the Southern Hemisphere from amino-acid racemization in emu eggshells. Nature 385: 241-244. Page KJ, Nanson GC, Price DM, 1996. Chronology of Murrumbidgee River palaeochannels on the Riverine Plain, southeastern Australia. Journal Quaternary Science 11: 311-326. Petit JR, Jouzel J, Raynaud D, Barkov NI, Barnola JM, Basile I, Bender M, Chappellaz J, Davis M, Delaygue G, Delmotte M, Kotlyakov VM, Legrand M, Lipenkov VY, Lorius C, Pepin L, Ritz C, Saltzmann E, Stievenard M. 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399: 429-436. Pietsch TJ. 2005. Fluvial geomorphology and late Quaternary geochronology of the Gwydir fan-plain. Unpublished PhD thesis: University of Wollongong. Pye K. 1995. The nature, origin and accumulation of loess. Quaternary Science Reviews 14: 653-677. Rgner K, Knabe K, Roscher B, Smykatz-Kloss W, Zller L. 2004. Alluvial loess in the Central Sinai: Occurrence, origin, and palaeoclimatological consideration. Lecture Notes in Earth Sciences 102: 79-100. Schmid RM. 1985. Lake Torrens, South Australia: sedimentation and hydrology. Unpublished PhD thesis: Flinders University of South Australia. Schwerdtfeger P, Curran E. 1996. Climate of the Flinders Ranges. In Natural history of the Flinders Ranges, Davies M, Twidale CR, Tyler MS. (eds). Royal Society of South Australia, Adelaide; 6375.
29
Singh G, Opdyke ND, Bowler JM. 1981. Late Cainozoic stratigraphy, palaeomagnetic chronology and vegetational history from Lake George, N.S.W. Journal of the Geological Society of Australia 28: 435-452. Singh G, Luly J. 1991. Changes in vegetation and seasonal climate since the last full glacial at Lake Frome, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 84: 75-86. Srivastava P, Brook GA, Marais E, Morthekai P, Singhvi AK. 2006. Depositional environment and OSL chronology of the Homeb silt deposits, Kuiseb River, Namibia. Quaternary Research 65: 478491. Twidale CR. 1966. Chronology of denudation in the southern Flinders Ranges, South Australia. Transactions of the Royal Society of South Australia 90: 3-32. Ullman WJ, McLeod LC. 1986. The late-Quaternary salinity record of Lake Frome, South Australia: Evidence from Na+ in stratigraphically-preserved gypsum. Palaeogeography,
Palaeoclimatology, Palaeoecology 54: 153-169. Ullman WJ, Collerson KD. 1994. The Sr-isotope record of late Quaternary hydrologic changes around Lake Frome, South Australia. Australian Journal of Earth Sciences 41: 37-45. Williams DLG. 1982. Late Pleistocene vertebrates and palaeo-environments of the Flinders and Mount Lofty Ranges. Unpublished PhD thesis: Flinders University of South Australia. Williams GE. 1973. Late Quaternary piedmont sedimentation, soil formation and paleoclimates in arid South Australia. Zeitschrift fr Geomorphologie N.F. 17: 102-125. Williams GE, Polach HA. 1971. Radiocarbon dating of arid-zone calcareous palaeosols. Geological Society of America Bulletin 82: 3069-3085. Williams MAJ, Prescott JR, Chappell J, Adamson D, Cock B, Walker K, Gell P. 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International 8385: 129-144. Williams MAJ, Nitschke N. 2005. Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. South Australian Geographical Journal 104: 25-36. Wright VP. 1990. Estimating rates of calcrete formation and sediment accretion in ancient alluvial deposits. Geological Magazine 127(3): 273-276.
30
Wright VP, Tucker ME. 1991. Calcretes: an introduction. In Calcretes, Wright VP, Tucker ME. (eds). Blackwell Scientific Publications; 1-24. Yaalon D, Dan G. 1974. Accumulation and distribution of loess-derived deposits in the semi-desert and desert fringe areas of Israel. Zeitschrift fr Geomorphologie N.F. 20: 91105. Zawada PK. 1997. Palaeoflood hydrology: method and application in flood-prone southern Africa. South African Journal of Science 93(3): 111-132.
Depositional models of late Pleistocene fine-grained valley-fill formations in the Finders Ranges, South Australia
Lo ess an d Flo o d s: Ch ap t er 1.2 (In t r o d u ct io n ) David Haberlah 1, 2
1
31
(similar version published in: Fitzpatrick, R.W., Shand, P. (eds.) 2007, Regolith 2006Consolidation and dispersion of ideas. Cooperative Research Centre for Landscape Environments and Mineral Exploration, Perth, 122-126)
School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia Cooperative Research Centre for Landscape Environments and Mineral Exploration
32
Introduction
Although the general climatic development of Australia during the late Quaternary is well established (Chappell & Grindrod, 1983; Kershaw & Nanson, 1993; Veevers, 2001), detailed regional palaeo-environmental knowledge remains limited. This is particularly the case for the semi-arid landscapes located between the continents arid interior and the humid coastal areas (Harrison, 1993; Hesse et al., 2004). Semi-arid environments are very susceptible to hydrologic changes related to variation in precipitation, temperature, vegetation and associated evapo-transpiration. They can therefore be considered key regions in reconstructing the environmental impact of global warming and rising sea levels during the last deglacial termination. In the Australian context, fundamental questions such as the latitudinal shift of the westerly winds and the trade wind-dominated subtropical zone during the last glacial interval remain to be established (Harrison, 1993; Shulmeister et al., 2004). Recent studies date the onset and southward incursion of the summer monsoon to as early as ~15 ka cal. BP (Miller et al., 1997; Wyrwoll & Miller, 2001), decoupling it from more gradual changes in low-latitude summer insolation forcing, a conclusion that needs to be tested by independent palaeo-environmental data. So far, the main obstacle for regional-scale Quaternary palaeo-environmental reconstructions in Australia has been the limited availability of continuous sedimentary records from semi-arid and arid regions. This is especially the case for the period leading towards and culminating in the Last Glacial Maximum (LGM), during which terminal playa lakes occupying large structural basins dried out and were actively deflated (Schmid, 1985; Harrison, 1993; Magee et al., 1995). Consequent dust entrainment can be expected to result in loess accumulations (i.e. terrestrial deposits of dominantly silt-sized lithogenic wind-blown dust) in down-wind desert margins (Smith, et al., 2002). Loess sequences are globally recognised as valuable terrestrial proxies for reconstructing palaeoclimatic changes (Liu, 1987). One problem with Australian loess deposits is their generally limited thickness and irregular and discontinuous distribution. However, in many cases they still need to be identified and dated. Moreover, the nature of their deposition remains ambiguous and ill-defined (Hesse & McTainsh, 2003).
Fine-grained valley-fill formations in the Finders Ranges

The semi-arid Flinders Ranges are seasonally influenced by winter precipitation associated with frontal systems embedded in the Southern Westerlies, and summer monsoonal incursions from the north (Gentilli, 1972; Schwerdtfeger & Curran, 1996). The tilted, uplifted and dissected Precambrian and Palaeozoic sedimentary rocks form a sequence of weathering-resistant ridges with a wide
33
distribution of peaks rising up to 1,170 m above the surrounding plains (Preiss, 1987; Drexel & Parker, 1993). They act as an orographic barrier that can be expected to amplify palaeoenvironmental changes related to changes in the precipitation regime. The central Flinders Ranges hosts one of Australias best preserved loess-derived sedimentary sequences: the late Pleistocene fine-grained valley-fill formations occupying the Brachina and Wilkawillina drainage systems (Callen & Reid, 1994; Cock et al., 1999; Williams et al., 2001; Chor et al., 2003; Williams & Nitschke, 2005). Radiocarbon and luminescence dating of the stratigraphic section Slippery Dip established a continuous depositional record from ~33 to ~17 ka cal. BP, followed by incision and erosion that exhumed the underlying bedrock topography (Williams et al., 2001). This interval embraces the intensification and peak glacial conditions of the last glacial cycle and extends into the Deglacial, associated with the most significant and rapid climate changes in the past 120 ka (Petit et al., 1999).
Depositional models and palaeo-environmental implications

The depositional nature of the late Pleistocene Brachina valley-fill formation, unconformably overlying Neoproterozoic shales and effectively filling in a deeply dissected palaeo-relief, has been repeatedly reinterpreted over the past years. These interpretations range from Pleistocene lake deposits (Callen & Reid, 1994; Cock et al., 1999), wetland deposits (Williams et al., 2001), to "floodplain deposits in perennially wet grassy meadows" (Williams & Nitschke, 2005; 25). The latest study emphasises the importance of aeolian contributions as the source of the fine-grained sediments. However, the mode of their deposition, the causes behind the tabular sub-horizontal bedding planes and laminations, and the reason for their postglacial degradation remain to be unequivocally identified. This formation and similar terrace deposits along drainage systems throughout the Flinders Ranges record important geomorphological events at an exceptionally high temporal resolution and continuity for terrestrial records embracing the LGM. Yet they cannot be used for palaeo-environmental interpretations unless the processes involved in their formation and degradation are fully resolved. Comparable gently sloping terrace remnants of fine-grained valley-fill formations have been reported from other parts in the world such as the Sahara (e.g. Linstdter & Krpelin, 2004) and the Sinai Peninsula (Issar & Bruins, 1983). However none display such striking geomorphologic and lithostratigraphic similarities as the Namib Silts in the catchments of the Great Escarpment which delimits the Namib Desert along a near-longitudinal line between 32S and 14S. The age range between ~30 and ~8 ka cal. BP for these "loessic alluvial deposits" (Eitel et al., 2001; 57) is similar to that of the valley-fills in the Brachina catchment of the Flinders Ranges (Cock et al., 1999; Williams et al., 2001). Ephemeral streams are eroding the deposits under present arid climatic conditions.
34
Terrace remnants of up to 25 m thickness are preserved in protected localities such as the mouths of tributary valleys (Ward, 1987). Over the past 30 years, studies on the Namib Silts resulted in various competitive depositional models that can be divided into three groups. Lake deposits: Initial observers proposed a lacustrine origin caused by dune damming (Goudie, 1972; Scholz, 1972; Rust & Wieneke, 1974; 1980), similar to the first interpretation of the late Pleistocene Brachina formation (Callen & Reid, 1994; Cock et al., 1999). Since, gently sloping gradients of the tabular finegrained deposits were established for multiple catchments in the Great Escarpment (Hvermann, 1978) and in the Flinders Ranges (Williams et al., 2001), ruling out lacustrine emplacement. High-energy flood deposits: Various studies assume that the silt formations were deposited during flash-flood events with rapid sedimentation from heavily laden high-energy floodwaters, partly backflooding tributary side valleys (Ollier, 1977; Heine, 1987; Ward, 1987; Smith et al., 1993; Heine & Heine, 2002; Srivastava et al., 2005). According to this scenario, occurrences of laminated sequences upstream of confluences (e.g. Slippery Dip) would represent slackwater deposits (Kochel & Barker, 1982; Zawada, 1997). Low-energy floodout deposits: An alternative explanation is offered by attributing the fine-grained formations to low energy accumulations of river endpoints (Marker, 1977; Marker & Mller, 1978; Vogel, 1982; Eitel & Zller, 1995; Rust, 1999; Eitel et al., 2005). Continuous headward retreat of endoreic streams is suggested to result in successive upstream deposition of their suspension load, and to account for the apparently unchannelled nature of the aggradational surfaces. Aeolian loess accessions: An alternative working hypothesis presented here takes into account the recently established accumulation of loess mantles in the Flinders Ranges from the deflated bed of Lake Torrens (Williams & Nitschke, 2005), and in the Great Escarpment from the western Kalahari (Eitel et al., 2001). Accordingly, the silty unchannelled aggradational surfaces could reflect aeolian loess accessions. Loess can absorb gentle precipitation and runoff (Yair, 1987; 1994) and could therefore absorb weak intermittent rainfalls associated with cold fronts embedded in prevailing westerly winds. A moist surface could sustain perennial grass vegetation and successfully entrap more loess, with grasses growing upwards as dust accretion proceeds (Pye, 1995). Net deposition would further be enhanced by diurnal mesoscale wind circulation systems such as sea- and salt lake breezes, anabatic valley winds and nocturnal katabatic drainage flows (Tapper, 1991; Schwerdtfeger, 1996;
35
Sturman & Tapper, 2006). In this scenario, laminated successions such as Slippery Dip could perhaps be interpreted as areas of seepage that were able to sustain prolonged growth of cryptogamic crusts (Verrecchia et al., 1995). Each of these morphodynamic sedimentation models stipulates a fundamentally different palaeoenvironment. High-energy flood deposits require intense precipitation events, with layered to laminated sedimentary sequences reflecting episodic large-magnitude flood events. Low-energy floodouts or river-end deposits would indicate a shortening of the river course by decreasing runoff and reflect a successive aridification of the upper catchment. Aeolian loess accessions suggest a stable low rainfall regime, drawing attention to the influence of wind-blown material on the catchment hydrology. Hence, in order to infer any palaeoclimatic, palaeohydrologic and palaeoenvironmental conclusions from these exceptional terrestrial archives, the processes responsible for their aggradation and subsequent erosion need to be resolved first.
Chronostratigraphic approach
So far, existing sedimentological studies describing lithology, sedimentary structures, particle-size distribution, mineral composition, sediment colour, carbonate and microfossil content have failed to unambiguously establish the depositional environment of the fine-grained formations. This may be due to the allochthonous, uniform aeolian nature of the sediments (Eitel et al., 2001; Williams & Nitschke, 2005), or simply reflect post-depositional bioturbation (Smith et al., 1993). In the absence of clear sedimentological and lithological evidence, an alternative way to determine the morphodynamic modus operandi responsible for the aggradation of the valley-fills is suggested here. By establishing a chronostratigraphic transect along the thalweg of the terrace remnants, new detail concerning their sedimentation will be revealed. The discussed hypothetical morphodynamic models can be expected to result in unique temporalspatial deposition patterns. Combining detailed mapping of lithofacies associations with radiocarbon and luminescence dating provides a means to distinguish and to test each model. High-energy flood deposition is characterised by synchronous deposition of thick sedimentary sequences over great expanses along the main channel and into the mouth of tributaries (Zawada, 1997). Given that the depositional mode is event-driven, discrete sequential tabular bedding planes should produce similar ages across their lateral and vertical expanse. Low-energy floodout deposits on the other hand are characterised by a gradual upstream migration of the sedimentation focus. Therefore, ages of aggradational surfaces should decrease along the thalweg in an upstream direction. In contrast, loess deposits blanket the landscape. Their net deposition rate varies according to the local
36
topography, surface properties and the path of dust storm events. However, overall deposition can be expected to be gradual and largely synchronous across the thalweg and across various catchments.
Conclusion
Continuous high-resolution terrestrial records covering the culmination and termination of the last glacial cycle are scarce for the semi-arid mid-latitudes of Australia. In the past years, loess-derived valley-fill formations in the central Flinders Ranges were identified and dated to cover this climatically important interval. These formations are here linked to similar well-studied fine-grained terrace remnants in Namibia. However, their former depositional environment remains controversial and requires to be more firmly established before any palaeoclimatic and palaeohydrologic conclusions can be inferred. Sedimentological studies so far could not resolve their genesis, possibly due to their uniform nature based on the aeolian provenance of the material. By applying a chronostratigraphic approach in reconstructing the former temporal-spatial pattern of deposition along transects of thalwegs across multiple catchments, more conclusive evidence for any of three discussed potential depositional environment can be obtained. Current research by the author has the aim to resolve the ongoing international debate on the nature of these conspicuous late Pleistocene fine-grained valley-fill formations or Silts, presently eroding under semi-arid to arid climatic conditions.
37
References
CALLEN, R.A., REID, P.W. (Compilers) 1994. Geology of the Flinders Ranges National Park. South Australian Geological Survey, Special Map 1:75,000. CHAPPELL, J.M.A., GRINDROD, A. (eds) 1983. Proceedings of the first CLIMANZ conference, held at Howmans Gap, Victoria Australia, February 8-23: a symposium of results and discussions concerned with late quaternary climatic history of Australia, New Zealand and surrounding seas. Australian National University, Canberra. CHOR, C., NITSCHKE, N., WILLIAMS, M. 2003. Ice, wind and water: Late Quaternary valley-fills and aeolian dust deposits in arid South Australia. In: ROACH, I.C. (ed.), Advances in regolith. Cooperative Research Centre for Landscape Environments and Mineral Exploration, Perth, pp. 70-73. COCK, B.J., WILLIAMS, M.A.J., ADAMSON, D.A. 1999. Pleistocene Lake Brachina: a preliminary stratigraphy and chronology of lacustrine sediments from the central Flinders Ranges, South Australia. Australian Journal of Earth Sciences 46, 61-69. DREXEL, J.F., PREISS, W.V. (eds) 1993. The geology of South Australia. Vol. 1, The Precambrian. South Australia Geological Survey, Bulletin 54, South Australia. EITEL, B., ZLLER, L. 1995. Die Beckensedimente von Dieprivier und Uitskot (NW-Namibia): Ein Beitrag zu ihrer paloklimatischen Interpretation auf der Basis von Thermolumineszenzdatierungen. Mitteilungen der sterreichischen Geographischen Gesellschaft 137, 245-254. EITEL, B., BLMEL, W.D., HSER, K., MAUZ, B. 2001. Dust and loessic alluvial deposits in northwestern Namibia (Damaraland, Kaokoveld): sedimentology and palaeoclimatic evidence based on luminescence data. Quaternary International 76/77, 57-65. EITEL, B., KADEREIT, A., BLMEL, W.D., HSER, K., KROMER, B. 2005. The Amspoort Silts, northern Namib desert (Namibia): formation, age and palaeoclimatic evidence of river-end deposits. Geomorphology 64, 299-314. GENTILLI, J. 1972. Australian climate patterns. Nelson, Melbourne. GOUDIE, A., 1972. Climate weathering, crust formation, dunes and fluvial features of the Central Namib Desert, near Gobabeb, South West Africa. Madoqua II 1 54-64, 15-31. HARRISON, S.P. 1993. Late Quaternary lake-level changes and climates of Australia. Quaternary Science Reviews 12, 211-231. HEINE, K. 1987. Jungquartre fluviale Geomorphodynamik in der Namib, Sdwestafrika/Namibia. Zeitschrift fr Geomorphologie N.F. Supplementband 66, 113-134. HEINE, K., HEINE, J.T. 2002. A paleohydrologic reinterpretation of the Homeb Silts, Kuiseb River, central Namib Desert (Namibia) and paleoclimatic implications. Catena 48, 107-130.
38
HESSE, P.P., MCTAINSH, G.H. 2003. Australian dust deposits: modern processes and the Quaternary record. Quaternary Science Reviews 22, 2007-2035. HESSE, P.P., MAGEE, J.W., VAN DER KAARS, S. 2004. Late Quaternary climates of the Australian arid zone: a review. Quaternary International 118/119, 87-102. HVERMANN, J. 1978. Formen und Formung in der Prnamib (Flchen-Namib). Zeitschrift fr Geomorphologie N.F. Supplementband 30, 55-73. ISSAR, A., BRUINS, J. 1983. Special Climatological Conditions in the Deserts of Sinai and Negev during the Late Pleistocene. Palaeogeography, Palaeoclimatology, Palaeoecology 43, 63-72. KERSHAW, A.P., NANSON, G.C. 1993. The last full glacial cycle in the Australian region. Global and Planetary Change 7, 1-9. KOCHEL, R.C., BAKER, V.R. 1982. Paleoflood hydrology. Science 215, 353-361. LINSTDTER, J., KRPELIN, S. 2004. Wadi Bakht revisited: Holocene climate change and prehistoric occupation in the Gilf Kebir region of the Eastern Sahara, SW Egypt. Geoarchaeology 19(8), 753-778. LIU, T. (ed) 1987. Aspects of loess research. China Ocean Press, Beijing. MAGEE, J.W., BOWLER, J.M., MILLER, G.H. WILLIAMS, D.L.G. 1995. Stratigraphy, sedimentology, chronology and palaeohydrology of Quaternary lacustrine deposits at Madigan Gulf, Lake Eyre, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 1, 3-42. MARKER, M.E. 1977. Aspects of the geomorphology of the Kuiseb River valley, South West Africa. Madoqua 10(3), 199-206. MARKER, M.E., MLLER, D. 1978. Relict vlei silts of the middle Kuiseb River valley, South West Africa. Madoqua 11(2), 151-162. MILLER, G.H., MAGEE, J.W., JULL, A.J.T. 1997. Low-latitude glacial cooling in the Southern Hemisphere from amino-acid racemization in emu eggshells. Nature 385, 241-244. OLLIER, C.D. 1977. Outline geological and geomorphological history of the Central Namib Desert. Madoqua 10(3), 207-212. PETIT, J.R., JOUZEL, J., RAYNAUD, D., BARKOV, N.I., BARNOLA, J.M., BASILE, I., BENDER, M., CHAPPELLAZ, J., DAVIS, M., DELAYGUE, G., DELMOTTE, M., KOTLYAKOV, V.M., LEGRAND, M., LIPENKOV, V.Y., LORIUS, C., PEPIN, L., RITZ, C., SALTZMANN, E., STIEVENARD, M. 1999. Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399, 429-436. PREISS, W.V. (compiler) 1987. The Adelaide Geosyncline: Late Proterozoic stratigraphy, sedimentation, palaeontology and tectonics. Geological Survey of South Australia, Bulletin 53, South Australia. PYE, K. 1995. The nature, origin and accumulation of loess. Quaternary Science Reviews 14, 653-677.
39
RUST, U., WIENEKE, F. 1974. Studies on gramadulla formation in the middle part of the Kuiseb River, South West Africa. Madoqua II 3, 3-15. RUST, U. WIENEKE, F. 1980. A reinvestigation of some aspects of the evolution of the Kuiseb River valley upstream of Gobabeb, South West Africa. Madoqua 12(3), 163-173. RUST, U. 1999. River-end deposits along the Hoanib River, northern Namib: archives of Late Holocene climatic variations on a subregional scale. South African Journal of Science 95, 205208. SCHMID, R.M. 1985. Lake Torrens, South Australia: sedimentation and hydrology. Unpublished PhD thesis, School of Earth Sciences, Flinders University of South Australia. SCHOLZ, H. 1972. The soils of the central Namib Desert with special reference to the soils in the vicinity of Gobabeb. Madoqua 11(1), 33-51. SCHWERDTFEGER, P., CURRAN, E. 1996. Climate of the Flinders Ranges. In: Davies, M., Twidale, C.R., Tyler, M.J. (eds.), Natural history of the Flinders Ranges, pp. 63-75. Royal Society of South Australia Occasional Publication. SHULMEISTER, J., GOODWIN, I., RENWICK, J., HARLE, K., ARMAND, L., MCGLONE, M.S., COOK, E., DODSON, J.R., HESSE, P.P., MAYEWSKI, P., CURRAN, M. 2004. The Southern Hemisphere westerlies in the Australasian sector over the last glacial cycle: a synthesis. Quaternary International 118-119, 23-53. SMITH, B.J., WRIGHT, J.S, WHALLEY, W.B. 2002. Sources of non-glacial, loess-size quartz silt and the origins of "desert loess". Earth Science Reviews 59, 1-26. SMITH, R.M.H., MASON, T.R., WARD, J.D. 1993. Flash-flood sediments and ichnofacies of the Late Pleistocene Homeb Silts, Kuiseb River, Namibia. Sedimentary Geology 85, 579-599. SRIVASTAVA, P., BROOK, G.A., MARAIS, E. 2005. Depositional environment and luminescence chronology of the Hoarusib River Clay Castles sediments, northern Namib Desert, Namibia. Catena 59, 187-204. STURMAN, A., TAPPER, N.J. 2006. The weather and climate of Australia and New Zealand, 2nd ed. Oxford University Press TAPPER, N.J. 1991. Evidence for a mesoscale thermal circulation over dry salt lakes. Palaeogeography, Palaeoclimatology, Palaeoecology 84, 259-269. VEEVERS, J.J. 2001. Atlas of billion-year earth history of Australia and neighbours in Gondwanaland. GEMOC Press, Sydney. VERRECCHIA, E., YAIR, A., KIDRON, G.J., VERRECCHIA, K. 1995. Physical properties of the psammophile cryptogamic crust and their consequences to the water regime of sandy soils, north-western Negev Desert, Israel. Journal of Arid Environments 29, 427-437. VOGEL, J.C. 1982. The age of the Kuiseb River silt terrace at Homeb. Palaeocecology of Africa 15, 201-209.
40
WARD, J.D. 1987. The Cenozoic succession in the Kuiseb Valley, central Namib Desert. Geological Survey of South West Africa/ Namibia Memoir 9, 1-124. WILLIAMS, G.E. 1973. Late Quaternary piedmont sedimentation, soil formation and paleoclimates in arid South Australia. Zeitschrift fr Geomorphologie N.F. 17(1), 102-125. WILLIAMS, M.A.J., PRESCOTT, J.R., CHAPPELL, J., ADAMSON, D., COCK, B., WALKER, K., GELL, P. 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International 83-85, 129-144. WILLIAMS, M.A.J., NITSCHKE, N. 2005. Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. South Australian Geographical Journal 104, 25-36. WYRWOLL, K.H., MILLER, G.H. 2001. Initiation of the Australian summer monsoon 14,000 years ago. Quaternary International 83-85, 119-128. YAIR, A. 1987. Environmental effects of loess penetration into the northern Negev Desert. Journal of Arid Environments 13, 9-24. YAIR, A. 1994. The ambiguous impact of climate change at a desert fringe: Northern Negev, Israel. In: Millington, A.C., Pye, K. (eds.), Environmental change in drylands: biogeographical and geomorphological perspectives, pp. 199-227. Wiley, Chichester. ZAWADA, P.K. 1997. Palaeoflood hydrology: method and application in flood-prone southern Africa. South African Journal of Science 93(3), 111-132.
A call for Australian loess

Lo ess an d Flo o d s: (similar version published in Area (2007) 39.2, 224229) David Haberlah 1, 2
1 2
41
School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia Cooperative Research Centre for Landscape Environments and Mineral Exploration
Abstract
The term loess for silty terrestrial deposits of aeolian origin is widely avoided in the Australian context. This seems to be linked to a prevailing notion among Australian geoscientists that loess is an inherently periglacial late Pleistocene sediment and hence negligible on the mainland. Addressing this conception, loess is presented here as a product of both cold and hot semi-arid environments and therefore a widespread feature in Australia. The adoption of a non-prescriptive definition of loess will align the variety of local descriptions with overseas terminology. More importantly, it will relate hitherto only vaguely-defined wind-blown dust occurrences to a broader palaeoenvironmental concept. Keywords Australia, loess, dust, wind-blown, parna
42
Introduction: the Australian liaison with loess

Australian Earth Science writers generally avoid the use of the term loess to denote aeolian silt deposits. In the reference list of Hesse and McTainshs (2003) recent review on Australian dust deposits the word loess does not appear in the continental context. Instead a variety of descriptive and local terms are used, including: parna (Butler 1956), aerosolic quartz (Alloway et al. 1992), continental dust (Hesse 1994), wind transported dust (Kiefert 1995), aeolian dust (Hesse and McTainsh 1999), windblown dust (Gatehouse et al. 2001), lithogenic fluxes (Kawahata 2002) and allochthonous, aeolian mineral grains (Stanley and De Deckker 2002). Related terrestrial deposits are referred to as aeolian dust mantles (Butler 1982), aeolian accessions (Chartres et al. 1988), aeolian mantles (Hesse et al. 1998), aeolian dust deposits (Chen 2001) or simply dust deposits (Hesse and McTainsh 2003). It is puzzling why the term dust, which is only vaguely defined, indicative of long-term suspension and embracing a variety of non-lithogenic aerosols (Pye 1987), is preferred to loess, a well-established geomorphologic concept with an extensive international literature. Two reasons have been reiterated in conversations with the local geoscientific community: one that loess is inherently linked to glaciation and therefore has had a limited impact on mainland Australia during the Late Pleistocene (Barrows 2001); and, the second that Australian loess is a unique local variation distinct from other global loess occurrences. Whereas the Australian pioneers of regolith geology such as Hills (1939), Crocker (1946) and Gill (1953) readily identified fine-grained terrestrial sediments as loess, Butler (1956) for the first time discriminated them from the then known global loess occurrences. He deemed the high clay content of the local deposits to be incompatible with the existing definitions of loess and coined the term parna (Butler 1956, 1974). In addition, at a time when deserts were still believed to contract during pluvial glacial intervals, he looked in vain for studies identifying loess deposits in the present-day deserts of Australia and concluded that "hot loess [ie. desert loess] [is] more hypothetical than real" (1956:147). Butlers approach towards Australian loess and his concept of parna have been most influential for the reason that they were institutionally adopted during the active CSIRO soil surveys of the 1950s and 1960s (Hesse and McTainsh 2003). However, since the 1960s Quaternary research has established that the arid Australian continental interior expanded and attained its greatest extent during glacial maxima (Bowler and Wasson 1984, Williams 2000). Deserts are now recognized as source areas of loess entrainment, even for the classic Chinese glacial loess deposits (Derbyshire 1983). Furthermore, subsand-sized aeolian deposits are embraced in the present understanding of loess, regardless of their clay content, a characteristic regarded as primarily
43
source-dependent and not the effect of dissimilar geomorphic processes (Dare-Edwards 1984). The absolute clay content of Australian dust deposits has yet to be established for a representative number of well-controlled sites, despite remaining the main reason to differentiate parna from other international desert loess occurrences (Hesse and McTainsh 2003). It is therefore the intention of this paper to challenge the prevailing belief among many Australian geoscientists that loess is primarily a periglacial phenomenon. The argument put forward here is that loess-generating environments are essentially semi-arid, a climatic condition characterising vast tracts of the Australian continent during the late Pleistocene. To begin with, a definition of loess to include that of the continental context is suggested.
Loess a non-prescriptive definition

Following the application of the German word L by Karl Caesar von Leonhard (1824) to describe silty deposits along the Upper Rhine Valley and the English adoption of the term loess by Lyell (1833), many authors have proposed their own definitions of the material. These have depended on the various processes deemed to be essential for its generation and were limited by the scope of their field experience. Of those, who have made wide-ranging international observations, Pye (1995) provided a practical generic definition which facilitates comparative studies from various climatic regimes and landscapes: "[Loess is] a terrestrial clastic sediment, composed predominantly of siltsize particles, which is formed essentially by the accumulation of wind-blown dust" (Pye 1995:654). Pye further stresses that loess deposits are often characterised by syn- or post-depositional reworking caused by weathering, bioturbation and pedogenic processes. These result in variable calcium carbonate and clay contents, but contrary to Butlers conception (Butler 1956, 1974) he suggests that neither their presence nor absence should be used as diagnostic criteria. Obviously, a non-prescriptive definition by itself lacks conviction if processes that have led to the accumulation of the classic loess sequences in Central Asia and China are assumed to be glacial and periglacial and therefore need to be ruled out in the Australian context. Any satisfactory geomorphologic term must indeed relate to specific modes of genesis of the material and landforms. Loess therefore has to refer to more than just aeolian silt (Pcsi 1990). In order to demonstrate that loess is primarily the product of semi-arid environments, the essential conditions for loess formation, i.e. silt generation and surficial concentration of erodible deposits, are discussed. Further, the fundamental process-orientated similarities between classical glacial loess and Australian desert loess are examined in more detail.
44
Discussion: linking loess-generating processes in periglacial and hot desert environments

Two key issues underlie loess-generating processes. The first is the question of modi capable of breaking and splitting up the generally sand-sized quartz crystals of igneous and metamorphic rocks. The second is that of mechanisms operating in their release and concentration in non-cohesive deposits in the landscape which can subsequently be deflated by wind. A classic and still well disseminated idea of silt production is that of glacial grinding (Hardcastle 1889, Smalley 1966). Since the introduction of this concept, many other more important large-scale mechanisms generating silt-sized material have been discovered and subsequent observations and laboratory studies cast doubt on its capability to produce substantial amounts of silt (Whalley 1979, Haldorsen 1981, Nahon and Trompette 1982, Derbyshire 1983, Sharp and Gomez 1986, Wright 1995, Wright et al. 1998). Therefore, regardless of the continuing popularity of this concept which is still being propagated by a number of authors (Smalley et al. 2005), glacial grinding will be neglected in the ensuing discussion of silt-generating processes. Only those that could be empirically verified in field observations supported by laboratory simulations to produce substantial amounts of silts will be considered further. Over the past decades, a variety of weathering processes and geomorphological transport processes were demonstrated to be capable of silt generation. These will now be linked to the most favourable climatic regimes in which they operate, and to subsequent silt-sorting mechanisms.
Weathering processes
A number of observations and empirical studies have focused on physical weathering processes as silt-generating mechanisms. Frost shattering was first noticed on Mount Kenya by Zeuner (1949) to generate silt-sized particles. His observations were supported by later laboratory experiments (Brockie1973, Moss et al. 1981, Lautridou and Ozouf 1982, Wright et al. 1998). The potential of frost shattering processes to generate silt depends on properties of the source material and environmental controls such as the occurrence of salt, affecting the moisture absorption and ice crystallisation processes (Pye 1987). Haloclasty alone causes weathering and the disintegration of rocks and sediments. The ability of salt weathering to generate significant quantities of silt-sized particles has been demonstrated by Goudie (1977) on archaeological brick building structures of the Harappa Culture, on glacial till in the
45
Karakoram Mountain Range of Pakistan (Goudie 1984), in the Sahara (Goudie and Watson 1984) and in laboratory experiments (Goudie and Viles 1995). Both physical weathering processes are most active and common within semi-arid climates, irrespective of the prevalent temperature regime. In contrast, chemical weathering and pedogenic processes are dependent on more temperate and humid conditions. Nahon and Trompette (1982) demonstrated that chemical weathering acts as an important loess-producing mechanism throughout all geologic epochs (see also Berg 1927). They argue that the key role of glaciations in the process of loess generation is that of erosion and entrainment of the pre-glacially weathered debris and the subsequent concentration of silts by glaciofluvial melt-waters (see also Warnke 1971, Whalley 1979, Pye 1995, Wright 2001). However, the sequence of chemical weathering, erosion, transport, sorting, concentration and final surficial re-deposition of readily erodible fine-grained material is not restricted to the periglacial parts of the world. On the contrary, it is paralleled in many hot semi-arid to arid environments. During the Mesozoic, Australias landscape was highly weathered (Hill 1999, Twidale 2000). Fines, resulting from processes such as weathering remain trapped in cohesive weathering mantles until the inherited regolith is eroded and the clay and silt fractions are sorted and concentrated into surficial deposits by the action of running water. In the Quaternary endoreic environment of central Australia, the suspended load eventually settled in inland drainage depocentres such as the tectonic basins presently occupied by playas, salinas or lakes. To be entrained by wind, the surface of these deposits has to desiccate and the fine-grained material dispersed and exposed. Silt and clay deflation therefore requires geomorphologically active surfaces, which are typical of semi-arid climates promoting a course of inundation and desiccation events. During glacial intervals, basins like those of Lake Eyre, Lake Torrens and the Murray Basin repeatedly dried out (Hesse et al. 2004) exposing vast repositories of loess-sized sediments to deflation.
Transport processes
Other studies have focused on geomorphological transport processes capable of rapid particle size reduction and the production of substantial quantities of silt. Moss (1972) conducted field experiments in the Murrumbidgee catchment area and noticed a rapid downstream disintegration of quartz due to crushing and abrasion in the bedload. Violent motion between moving boulders, pebbles and bedrock projections causes the comminution of coarse particles until silt-sized grains escape further disintegration by suspension. His observations were confirmed by subsequent laboratory experiments (Moss et al. 1973, 1981). More recent
46
independent tumbling experiments by Wright and co-authors (1993, 1998, 2001), simulating highenergy fluvial systems of mixed-sized sediments, support the capability of fluvial abrasion to generate significant amounts of silt. Aeolian systems are also capable of producing considerable quantities of silt. Attrition between saltating sand particles was first described by Knight (1924). Subsequent laboratory studies (Whalley et al. 1982, Smith et al. 1991, Wright 2001, Bullard et al. 2004) identified aeolian abrasion as the cause of the typical rounded morphology of dune sand grains. During the process of edge rounding, substantial quantities of silt-sized fines are chipped off. Apart from chipping and spalling of grains, aeolian abrasion can also remove and release clay coatings, a process likely to be particular important in the reddened dune fields of Australia (Bullard and White 2005). Although past emphasis on geomorphic transport processes capable of silt-generation has often been placed on glacial grinding, comparative quantitative laboratory studies established that the two most effective ones are fluvial and aeolian abrasion (Wright et al. 1998). Turbulent high-energy fluvial environments capable of carrying a bedload of a wide range of particle sizes do exist in periglacial environments during periodic, poorly-sorted glaciofluvial outwash events. Episodic flash floods channelled into ephemeral river beds are similarly significant. Both systems display intermittent discharge which creates active surfaces from which freshly generated silt-sized particles can be readily mobilised and entrained by wind. Aeolian abrasion is mainly active in semi-arid and arid environments with bare, sparsely vegetated surfaces and drifting sand. Therefore, abrasion will generate silt both in dune fields forming on glacial outwash plains and in active sand seas characteristic of parts of the Sahara and the interior of Australia during the glacial intervals.
Conclusion: loess not just another term but a concept

The variety of mechanisms and processes operating in generating, sorting and concentrating siltsized grains for subsequent entrainment by wind and aeolian deposition as loess operate both in periglacial and hot semi-arid environments. Rates of silt particle formation and sorting and concentration processes are considerably lower under hyper-arid and glacial conditions than semiarid and periglacial environments. Arid and glacial environments are climatically more stable and characterised by low weathering rates, relatively stable atmospheric conditions and a high proportion of land surfaces protected by crusting, gravel-armouring and glacier ice, respectively (Pye 1995). The fundamental process-based similarities between glacial loess and desert loess suggest that a comprehensive loess definition is readily applicable to Australian terrestrial occurrences of aeolian dust mantles or any of the variety of other terms listed at the beginning of this paper.
47
Ultimately, the issue is not just about a difference in terms. The power of words, especially in the scientific context, is that they embody and relate concepts. Loess sequences, even if they appear considerably reworked and pedogenically altered in their present state, are indicative of specific former conditions regarding dust source proximity, airflow regimes, climatic gradients and vegetation cover. Loess entrainment requires a geomorphologically active, regularly disturbed landscape over a period of time. High net deposition rates in turn require particular surface properties in terms of moisture, vegetation and topography. Furthermore, loess exerts various significant impacts on the hydrology, geochemistry and flora in regions where it is deposited (Yair 1994). These aspects are all relevant for Quaternary landscape reconstructions and palaeoenvironmental interpretations. Calling appropriate aeolian dust deposits loess can change not only the way we look at them, but also how we look for them as comparable occurrences in the international literature or during identification in the field.
Acknowledgments
I am grateful to Dr. Andreas Schmidt-Mumm who supported my scholarship application at a crucial time and helped me to settle down in Australia. I also thank the Graduate Centre of the University of Adelaide for offering the Integrated Bridging Program (Research) and in particular Margaret Cargill for her excellent introduction to the Australian scientific discourse. My supervisors Martin A. J. Williams, Steven M. Hill and Victor Gostin and the reviewers are thanked for inspiration and critical feedback.
48
References
Barrows T T, Stone J O, Fifield L K and Cresswell R G 2001 Late Pleistocene Glaciation of the Kosciuszko Massif, Snowy Mountains, Australia Quaternary Research 55 179-189 Berg L S 1927 Loess as a product of weathering and soil formation Pedology 2 21-37 Bowler J M and Wasson R J 1984 Glacial age environments of inland Australia in Vogel J C ed Late Cainozoic palaeoclimates of the southern hemisphere. Proceedings of an International Symposium held by the South African society for Quaternary research/ Swaziland/ 29 August 2 September 1983 183-208 Brockie W J 1973 Experimental frost-shattering. Proceedings of the Seventh New Zealand Geography Conference New Zealand Geographical Society conference series 7 177-186 Bullard J E, McTainsh G H and Pudmenzky C 2004 Aeolian abrasion and modes of fine particle production from natural red dune sands: an experimental study Sedimentology 51(5) 11031125 Bullard J E and Kevin W 2005 Dust production and the release of iron oxides resulting from the aeolian abrasion of natural dune sand Earth Surface Processes and Landforms 30 95-106 Butler B E 1956 Parna an aeolian clay Australian Journal of Science 18 145-151 Butler B E 1974 A contribution towards the better specification of parna and some other aeolian clays in Australia Zeitschrift fr Geomorphologie NF Supplement Band 20 106-116 Crocker R L 1946 Post-Miocene climatic and geologic history and its significance in relation to the genesis of the major soil types of South Australia Council for Scientific and Industrial Research Bulletin 193 5-65 Dare-Edwards A J 1984 Aeolian clay deposits of south-eastern Australia: parna or loessic clay? Transactions of the Institute of British Geographers New Series 9 337-344 Derbyshire E 1983 On the morphology, sediments, and origin of the loess plateau of Central China in Gardner R and Scoging H eds Mega-geomorphology Oxford University Press, Oxford 173-194 Gill E D 1953 Geological evidence in western Victoria relative to the antiquity of the Australian Aborigines Memoirs of the National Museum of Victoria Melbourne 18 25-92 Goudie A S 1977 Sodium sulphate weathering and the disintegration of Mohenjo Daro, Pakistan Earth Surface Processes 2 75-86 Goudie A S 1984 Salt efflorescences and salt weathering in the Hunza Valley, Karakoram Mountains, Pakistan in Miller K ed Proceedings of the International Karakoram Project, Vol. II Cambridge University Press, Cambridge 607-615 Goudie A S and Watson A 1984 Rock block monitoring of rapid salt weathering in Southern Tunisia Earth Surface Processes and Landforms 9 95-98
49
Goudie A S and Viles H A 1995 The nature and pattern of debris liberated by salt weathering: a laboratory study Earth Surface Processes and Landforms 20 437-449 Haldorsen S 1981 Grain size distribution of subglacial till and its relation to glacial crushing and abrasion Boreas 10 91-105 Hardcastle J 1889 Origin of the loess deposits of the Timaru Plateau Transactions of the New Zealand Institute 22 406-414 Hesse P P and McTainsh G H 2003 Australian dust deposits: modern processes and the Quaternary record Quaternary Science Reviews 22 2007-2035 Hesse P P, Magee J W and van der Kaars S 2004 Late Quaternary climates of the Australian arid zone: a review Quaternary International 118-119 87-102 Hill S M 1999 Mesozoic regolith and palaeolandscape features in southeastern Australia: significance for interpretations of denudation and highland evolution. Australian Journal of Earth Sciences 46 217-232 Hills E S 1939 The physiography of north-western Victoria Proceedings of the Royal Society of Victoria 51 297-322 Knight S H 1924. Eolian abrasion of quartz grains (abstract) Geological Society of America Bulletin 35 107-108 Lautridou J P and Ozouf J C 1982 Experimental frost shattering: 15 years of research at the Centre de Gomorphologie du CNRS Progress in Physical Geography 6 215-232 Moss A J 1972 Initial fluviatile fragmentation of granitic quartz Journal of Sedimentary Petrology 42(4) 905-916 Moss A J, Walker P H and Hutka J 1973 Fragmentation of granitic quartz in water Sedimentology 20 489-511 Moss A J, Green P and Hutka J 1981 Static breakage of granitic detritus by ice and water in comparison with breakage by flowing water Sedimentology 28 261-272 Nahon D and Trompette R 1982 Origin of siltstones: glacial grinding versus weathering Sedimentology 29 25-35 Pcsi M 1990. Loess is not just the accumulation of dust Quaternary International 7/8 1-21 Pye K 1987 Aeolian dust and dust deposits Academic Press Inc, London Pye K 1995 The nature, origin, and accumulation of loess Quaternary Science Reviews 14 653-667 Sharp M and Gomez B 1986 Processes of debris comminution in the glacial environment and implications for quartz sand-grain micromorphology Sedimentary Geology 46 33-47 Smalley I J 1966 The properties of glacial loess and the formation of loess deposits Journal of Sedimentary Petrology 36 669-676
50
Smalley I J, Kumar R, OHara Dhand K, Jefferson I F and Evans R D 2005 The formation of silt material for terrestrial sediments: particularly loess and dust Sedimentary Geology 179 321-328 Smith B J, Wright J S and Whalley W B 1991 Simulated aeolian abrasion of Pannonian sands and its implications for the origins of Hungarian loess Earth Surface Processes and Landforms 16 745752 Smith B J, Wright J S and Whalley W B 2002 Sources of non-glacial, loess-size quartz silt and the origins of "desert loess" Earth Science Reviews 59 1-26 Twidale C R 2000 Early Mesozoic (?Triassic) landscapes in Australia: evidence, argument, and implications The Journal of Geology 108 537-552 Von Leonhard K C 1823-1824 Charakteristik der Falsarten Engelman, Heidelberg Warnke D A 1971 The shape and surface texture of loess particles: discussion Geological Society of America Bulletin 82 2357-2360 Whalley W B 1979 Quartz silt production and sand grain surface textures from fluvial and glacial environments Scanning Electron Microscopy 1979/I 547-554 Whalley W B, Marshall J R and Smith B J 1982 Origin of desert loess from some experimental observation Nature 300 433-435 Williams M A J 2000 Quaternary Australia: extremes in the last glacial-interglacial cycle in Veevers J J ed Billion-year earth history of Australia and neighbours in Gondwanaland Gemoc Press, Sydney 55-59 Wright J S and Smith B J 1993 Fluvial comminution and the production of loess-sized quartz silt: a simulation study Geografiska Annaler A 75 25-34 Wright J S 1995 Glacial comminution of quartz sand grains and the production of loessic silt: a simulation study Quaternary Science Reviews 14 669-680 Wright J S, Smith B J and Whalley W B 1998 Mechanisms of loess-sized quartz silt production and their relative effectiveness: laboratory simulations Geomorphology 23 15-34 Wright J S 2001 Making loess-sized silt: data from laboratory simulations and implications for sediment transport pathways and the formation of desert loess deposits associated with the Sahara Quaternary International 76/77 7-19 Yair A 1994 The ambiguous impact of climate change at a desert fringe: Northern Negev, Israel in Millington A C and Pye K eds Environmental change in drylands: biogeographical and geomorphological perspectives Wiley, Chichester 201-227 Zeuner F E 1949 Frost soils on Mount Kenya Journal of Soil Science 1 20-32
David Haberlah 1, 2
1
Response to Smalleys discussion of A call for Australian loess

Lo ess an d Flo o d s: Ch ap t er 1.4 (In t r o d u ct io n ) (similar version published in Area (2008) 40.1, 135136) School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia Cooperative Research Centre for Landscape Environments and Mineral Exploration
2
51
Abstract
A concise response to Smalleys discussion of my Call for Australian loess paper (Haberlah 2007, Area 39 2249) elaborating on past and possible future Australian contributions to the international loess debate. The suggested idea of dividing loess into three separate categories glacial loess, desert loess and mountain loess is questioned by emphasising common links. Key words: Australia, loess, dust, parna, desert loess, mountain loess
52
Right from the opening lines of the discussion section in my paper (Haberlah 2007) it becomes clear that one could not possibly wish for a more appropriate scholarly reply and discussion of my dismissive stance of an imperative glacial link to loess than from Ian Smalley (Smalley 2008). When Smalley was pioneering the idea of glacial grinding as a (the) key loess-generating process more than four decades ago (Smalley 1966), he firmly associated the formation and nature of loess to glaciation. His ideas had a profound effect on the perception of loess in areas in the world such as continental Australia, where glaciation over the Last Glacial Maximum was negligible. When now the same author who is still perceived by many as the outspoken proponent of glacial loess revisits his conceptions of the nature and formation of loess in light of an increasing number of discoveries of desert loess clearly detached from glaciation, this should not be taken lightly. It is particularly instructive to follow his historical insight on how the whole debate developed until present, and the role Australian geoscientists played in it. Smalley demonstrates the important impact Butlers early denial of a significant desert-loess relationship in Australia (Butler 1956) had on the conceptionalisation of the nature of loess and its assignment to the glacial domain a decade later. In the 1970s, it was once again an Australian CSIRO scientist who strongly advanced the loess debate by demonstrating defects in quartz particles facilitating their breakage into silt-sized detritus. The observations by Moss (Moss and Green 1975) set the ground for the recognition of fluvial generation of silt as opposed to glacial grinding. Smalley expands my argument that silt-generating environments are quintessentially semi-arid environments irrespective of prevailing temperatures by adding a spatial dimension to the debate. He highlights the fact that under both peri-glacial and semi-arid climatic conditions, high relief is crucial in providing the geo-energy required for the key silt generating mode of fluvial comminution. This could prove to be an important point made in the future of loess discussion and necessitates more detailed spatial studies of loess occurrences in relation to the drainage of major mountain ranges. I readily support the notion that high relative relief is as essential for the generation of sufficient silt-sized particles as is the intermittent nature of transportation, sorting and deflation from both glacial outwash and from wadis and episodically inundated floodplains of larger river systems. However, I question that they necessitate a separate category as proposed with mountain loess. Perhaps they are but the second neglected link between glacial and desert loess? Could loess generation require the twofold process of erosion and fluvial comminution most effectively operating in high mountain ranges to produce and sort substantial quantities of silt, and then the winnowing of these from outwashes, floodplains and terminal depocentres resulting in their aeolian redistributing most effective in semi-arid environments? Further studies and time will tell.
53
For the immediate future, Smalley in his discussion presents us with a roadmap of what it takes to put Australia on the global loess map. I can only second his call for detailed granulometric and mineralogical studies and approaching the big question on where, how and when loess was generated in the Australian landscape. Since his call for such practical steps will likely be followed by a number of researchers, I would like to conclude my response with a note of caution based on experience gained from ongoing research on loess-derived alluvium in and around the Flinders Ranges, South Australia (i.e. Haberlah et al. 2007). Since Australian loess seems to consist of coarse and very coarse silt-sized clay aggregates, the application of a well-conceived, appropriate granulometric protocol is crucial. It should best be tested in controlled settings where, for example, subtle upward-fining trends can be anticipated (i.e. couplets of loess-derived slackwater deposits). Basic field texturing is problematic due to the tendency of loess aggregates to gradually disperse on application of mechanical force. This characteristic attribute was first described for the loessic soils of the Riverina of SE Australia and termed subplasticity by Butler (1955). It has since been the topic of much research by Australian CSIRO geoscientists (i.e. McIntyre 1976). Dispersion protocols developed for laboratory studies (i.e. Walker and Hutka 1976; Chittleborough 1982) should now be called upon to develop reliable protocols that ensure that the original transport-stable aggregates and not artefacts of physical and chemical pre-treatment are measured. As importantly, an appropriate and sufficiently accurate method of particle-size analysis should be employed avoiding analytical breaks between multiple sizing techniques (McTainsh and Duhaylungsod 1989). Since the Coulter Multisizer covers the complete size range of loess and allows sizing of minimally dispersed transport-stable aggregates as well as fully dispersed particulates at submicron resolution (McTainsh et al. 1997; Leys et al. 2005), it is suggested as most suitable for a comparative granulometric study of Australian loess. A coordinated effort in this respect and the application of the same protocols in other desert loess domains of the world could perhaps be the next significant contribution out of Australia to the international loess debate?
54
References
Butler B E 1955 A system for the description of soil structure and consistence in the field Journal of the Australian Institute of Agricultural Science 1 23152 Butler B E 1956 Parna an aeolian clay Australian Journal of Science 18 14551 Chittleborough D J 1982 Effect of the method of dispersion on the yield of clay and fine clay Australian Journal of Soil Research 20 33946 Haberlah D 2007 A call for Australian loess Area 39 2249 Haberlah D, Williams M A J, Hill S M, Halverson G and Glasby P 2007 A terminal Last Glacial Maximum (LGM) loess-derived palaeoflood record from South Australia? Quaternary International 167168 150 Leys J, McTainsh G, Koen T, Mooney B and Strong C 2005 Testing a statistical curve-fitting procedure for quantifying sediment populations within multi-modal particle-size distributions Earth Surface Processes and Landforms 30 57990 McIntyre D S 1976 Subplasticity in Australian soils I. Description, occurrence and some properties Australian Journal of Soil Research 14 22736 McTainsh G H and Duhaylungsod N C 1989 Aspects of soil particle-size analysis in Australia Australian Journal of Soil Research 27 62936 McTainsh G H, Lynch A W and Hales R 1997 Particle-size analysis of aeolian dusts, soils and sediments in very small quantities using a Coulter Multisizer Earth Surface Processes and Landforms 22 120716 Moss A J and Green P 1975 Sand and silt grains: predetermination of their formation and properties by microfractures in quartz Australian Journal of Earth Sciences 22 48595 Smalley I J 1966 The properties of glacial loess and the formation of loess deposits Journal of Sedimentary Petrology 36 66976 Smalley I J 2008 A call for Australian loess: discussion and commentary Area 40.1, 131134 Walker P H and Hutka J 1976 Subplasticity in Australian soils III. Particle-size properties of soil materials of varying plasticity Australian Journal of Soil Research 14 24960
2. Methods
Ch ap t er 2 (Met h o d s)
55
There is one thing even more vital to science than intelligent methods; and that is, the sincere desire to find out the truth, whatever it may be. Charles Pierce The methods and techniques selected to establish the depositional environment of the Flinders Silts can be grouped into chronostratigraphic, geochemical and geophysical approaches. The regional chronostratigraphic approach aimed at comparing and correlating aggradational sequences from different geomorphologic settings and catchments. While marker bands such as palaeosols were used along single section faces, numerical dating proved to be fundamental to the comparison of sections typically spaced several kilometres apart from each other. It also provided the only means by which to establish the timing and rates of deposition, and the duration of erosional intervals. Previous work (Williams 1982) suggested that the limit of radiocarbon dating could locally be exceeded. Site-specific unavailability of organic material, as well as a recent study questioning the validity of radiocarbon ages from the Homeb Silts in Namibia (Bourke et al. 2003) prompted the design of a regional dating campaign involving both optically stimulated luminescence (OSL) dating and accelerator mass spectrometer (AMS) radiocarbon dating. Standard laboratory methods and protocols were employed, but additional comparisons between radiocarbon ages obtained from different organic material were made. Similarly, small aliquots and single-grains for luminescence dating were juxtaposed, and the results from multiple age models were compared and critically discussed. Luminescence dating was carried out in collaboration with Ed Rhodes at the Australian National University (ANU) luminescence dating laboratory, and Tim Pietsch at the CSIRO luminescence dating laboratory in Canberra. All samples were prepared at the luminescence dating laboratory of the University of Adelaide under the kind supervision of John Prescott and Frances Williams. The same is true for all determinations of all environmental dose rates, equivalent dose estimates, and the final OSL age calculations. Radiocarbon dating involved two weeks of training in sample pre-treatment and combustion at the ANSTO research facilities (Lucas Heights). Additional radiocarbon samples were sent to the ANU Radiocarbon Laboratory and the Waikato Dating Laboratory. The largely homogenous, fine-grained and partially-aggregated material required an integrated highresolution sediment-sizing and advanced statistical particle-size distribution analysing approach, developed over repeated visits to the sedimentological laboratory of Griffith University (Brisbane)
Ch ap t er 2 (Met h o d s)
56
under the kind supervision of Grant McTainsh. The results are presented in form of a paper in section 2.3. Finally, a range of well-established and emerging analytical techniques was applied to determine the local vegetation, hydrological and weathering history of the layered to laminated section BRA-SD (Slippery Dip). This involved quantitative mineral spectroscopy, induced magnetic susceptibility measurements and isotopic carbon geochemistry, the methods of which are summarised with the results as a manuscript in section 3.2. Over the course of this study, the importance of careful field observations and detailed lithofacies mapping was realised and given appropriate time. The main results are summarised together with the chronostratigraphies in relevant papers in sections 3.1 and 3.2 and the corresponding appendices.
Ch ap t er 2.1 (Met h o d s)
57
2.1
Optically Stimulated Luminescence dating
The sand so pale it might be grains of light . Roselle Angwin Optically Stimulated Luminescence (OSL) dating is a relatively new method in the field of radiometric dating (Huntley et al. 1985). Its underlying principle is based on the characteristic of common mineral grains such as quartz sand to accumulate stored energy in their crystal lattices which is released on exposure to light. Consequently, sediments buried by a depositional event gradually acquire an increasing charge which can be freed and measured in the laboratory by stimulation with a beam of light. From this, the time elapsed since burial, i.e. the age of the sedimentary unit, can be resolved by establishing the environment-specific average energy flux the constituent sample grains were exposed to. OSL dating is closely related to Thermoluminescence (TL) dating and Electron Spin Resonance (ESR) dating, where stimulation by heat or electromagnetic field is used to depopulate electron traps in mineral grains (Aitken 1985). The great advantage of OSL over TL and ESR for palaeo-environmental studies is based on the prevalent and comparably rapid mode for re-setting the radiometric clock or luminescence signal of mineral grains in nature: the exposure to daylight during erosion, transport and deposition. Optical stimulation only samples light-sensitive electrons at a trap depth freed by light (Huntley et al. 1985). Electrons ejected from such traps within the crystal lattice recombine and release energy, some in the form of luminescence. With the help of a photomultiplier, these photons are converted to electric pulses and measured (Aitken 1998). Since, single quartz grains, or aliquots of such, behave differently to irradiation, it is important to individually assess that sensitivity. This is done by comparing the natural luminescence signal to responses to a sequence of laboratory radiation doses (Murray & Wintle 2000). The dose required to produce a luminescence signal equal to the measured natural signal is called equivalent-dose (De), measured in grays (Gy = 1J/kg), and is the numerator of the OSL-age equation [1]. The denominator consists of the average rate of energy flux absorbed by the sample in its natural depositional environment. This is referred to as dose-rate (DR), and expressed as grays per kiloyear (Gy ka-1): [1] OSL-age (ka) = De (Gy)/ DR (Gy ka-1)
The DR absorbed by the sample mineral depends on its exposure to radioactivity from the decay processes of naturally-occurring radioactive elements within the sediment and on cosmic irradiation. These produce secondary electrons through ionisation in the crystal lattice which subsequently
58
become trapped in structural defects. Energy absorption increases proportionally over time until all available traps are filled and the mineral becomes saturated (Aitken 1998). Hence, the upper limit of the OSL dating range is a function of the natural radioactivity of a given sedimentary unit, the mineralogy of its constituent grains and the burial time. Dose response curves of quartz grains generally become non-linear at ~30 Gy and saturate before ~150 Gy (Prescott & Robertson 1997), often limiting the OSL dating to the last glacial cycle (Walker 2005). The accuracy of the OSL-age determination of a depositional event is impaired by the following factors affecting De and DR, respectively. The largest potential source of error affecting the De arises when the depositional context prevents an effective re-setting of the luminescence signal, resulting in so-called partial bleaching. While OSL dating encounters significantly less residual signals than TL dating, for light-sensitive traps of quartz grains the signals are reduced to a negligible level in only a few seconds of exposure to daylight (Murray & Olley 2002), this prerequisite is not always achieved in all transport media, durations, compositions and saturations, or at night time (Bourke et al. 2003). Variation in the resultant luminescence signal is dealt with by establishing the spread of De-values over multiple individual grains or small aliquots for a given sample, and by applying appropriate protocols and age models in the interpretation of the De frequency distribution (Olley et al. 1998; Murray & Wintle 2000; Duller 2004; Jacobs et al. 2006). The DR on the other hand is affected by variations in the water content over the burial time of the sedimentary unit, pedogenic alterations of its mineral composition, and changes in the thickness of the overburden; all of which require a site-specific model for corrections. Further, mobile nuclides in the uranium decay chain can result in both loss and augmentation of radionuclides and consequent changes in the rate of delivery of the radiation dose. Radioactive disequilibrium can be assessed by comparison of results for the uranium (and thorium) decay chain members obtained by different methods (Prescott & Hutton 1995). Finally, sources of systematic errors arise from the employed techniques, precision and calibration of instruments and data conversions, limiting the precision of OSL-age determination (Murray & Olley 2002; Jacobs et al. 2006).
Field sampling
These sources of errors need to be kept in mind while deciding on stratigraphic sections and sedimentary units to sample for OSL dating. In terms of DR evaluation, this translates to strata where post-depositional disturbances such as diagenetic pedogenic alterations, fluctuating water tables (leaching) and more general variations in the water content can be assumed minimal over the time of burial. Since gamma rays contribute from as far as 30 cm to the ionising radiation, OSL samples
59
should ideally be surrounded by the same (homogenous) strata and lie below the surface by this distance (Aitken 1998). In terms of De evaluation, bioturbated strata prone to contamination with younger material must be avoided. The depositional modes that form a given stratigraphic sequence should be assessed for resetting the radiometric clock. While aeolian accessions such as loess mantles have the best bleaching potential, this is less so for a variety of fluvial deposits. Colluvial sediments such as slope wash can prove very difficult to date, potentially reflecting multiple partially-bleached cycles of erosion and deposition resulting in large apparent OSL-age variations (Prescott & Robertson 1997). Finally, the employed sampling procedure must shield the sample from present-day light. For this study, all samples are extracted from the selected strata by hammering steel cylinders of 12.5 cm length and 6 cm diameter into vertical sections, after scraping them clean and level. A ball of aluminium foil is inserted into the cylinder first, reducing shake-up of the sediment sample in the cylinder during extraction and flattening out to an opaque seal towards the end. Both ends are sealed by black plastic lids and the sample is placed in an air-tight sample bag to minimise changes in the present-day water content for subsequent measurement in the laboratory. Wherever feasible, OSL samples are paired or bracketed by radiocarbon samples.
Environmental dose-rate estimation

The luminescence signal is proportional to the amount of ionisation which in turn is proportional to the energy absorbed from radiation from the sample environment. An accurate measurement of the environmental dose-rate (DR) and a critical assessment of possible changes over the time of burial determine half the age equation. Methods of DR determination and conversion factors are wellestablished (Aitken 1985; 1998; Adamiec & Aitken 1998) and have remained much the same over the past decade (Duller 2004). For most OSL samples, the DR is produced in roughly equal proportions by the decay of the naturally-occurring radioactive isotopes potassium-40, and the decay chains of the radionuclides of Thorium-232 and Uranium-238. The range of ionising and highly saturating alpha particles is <5 m, and is effectively removed during sample preparation. Beta radiation, i.e. electrons, penetrates a few millimetres, and the gamma radiation up to a distance of 30 cm through sediments (Aitken 1985). Hence, DR determination requires a detailed analysis of the radioactive elements in both the sample and its surroundings. In addition, a usually small but not negligible DR contribution is derived from cosmic rays. The exact amount varies with altitude, latitude and sample depth, as well as with long-term time variations in galactic primary cosmic ray intensity and near earth modulation by changes in solar conditions and the geomagnetic dipole moment (Prescott & Hutton 1994). For the geomagnetic latitude (~31), age
60
range (LGM) and altitude (<400 m) for the presented samples, shielding afforded by the overburden (e.g. ranging from 35 to 1465 cm for WL07-FP) is deemed significant and calculated employing formula [2] published in Prescott & Hutton (1994) (see appendix 5.2.1.1). In order to test for radioactive disequilibrium in the potentially problematic (wet) alluvial depositional environment of sampled stratigraphic sections, laboratory measurements of the radioactive elements by combined X-ray fluorescence spectrometry (XRS) for potassium-40 (K) and Thick Source Alpha Counting (TSAC) for thorium (Th) and uranium (U) are conducted for all samples and compared to independently obtained data from reactor-based Delayed Neutron Analysis (DNA) and in-situ field scintillometry (Hutton & Prescott 1992). No disequilibrium was established in terms of statistically significant discrepancies between results for radioactivity of the U and Th decay chain members between differing methods (Prescott & Hutton 1995). Dose-rates based on all measurements discussed below are calculated using the DOS-based AGE99 software (Grn 1994). For in-situ derived K-, U-, and Th-contents and the total gamma ray count, adjustment for water content is required, using the following equation (Aitken 1985): X(dry) = X(moist) * (1 + F * H2O-content) F = 1.14 for total gamma count, 1.10 for K and U, 1.09 for For the computation, information regarding the diameter of the grain-size fraction used for the De estimation, the thickness of the outer layer removed during sample preparations, the average water content, and the cosmic ray estimate are required. The detailed protocol is given in appendix 5.2.1.2. Scintillometer field measurements The most advantageous method of evaluating the environmental DR for a given sample is by using a field gamma spectrometer for in-situ measurements of the radioactivity of the three main emitters. The measurement time (~30 min) is fast compared to the work load involving alternative and complementary laboratory measurements and calculations described below. Further, in-situ measurements best take account of the natural heterogeneity of the sample environment. However, since the detecting crystal needs to be inserted at least 30 cm clear of all surfaces (Aitken 1985), the extraction hole needs to be deepened and widened. Thereby, a source of uncertainty is introduced by departing from the actual and immediate sample-specific irradiation environment. The measurements for the individual radioactive components assume equilibrium in the U and Th decay chains (Prescott & Hutton 1995).
61
Scintillometry is based on the characteristic of gamma photons to produce scintillation pulses in sodium iodide crystals corresponding to their full energy of emission. These are converted to individual spectral photopeaks by a photomultiplier. Potassium-40 and the Th and U decay chains emit photons of characteristic energies resulting in a composite of such photopeaks. (Aitken 1985). Within this continuum, the following photopeaks best relate to the main DR components; 1.46 MeV for potassium-40, 1.76 MeV for bismuth-214 of the U-series (with correction for interferences) and 2.62 MeV for lead-208, the last member of the Th-series (Prescott & Hutton 1995). In collaboration with J.R. Prescott and F. Williams, the Adelaide portable multichannel gamma ray spectrometer (EXPLORANIUM) housing a 76 x 76 mm area low background sodium iodide scintillation crystal was employed at some sample locations. In addition to the spectral windows for the three radionuclides, the total gamma ray activity (0.5 2.78 MeV) was recorded (Prescott & Hutton 1987). The large size of the crystal requires no correction for the rate of muons from the cosmic ray flux, with the largest contribution to the measured count located in the Th spectral window (2.46 2.77 MeV) below ~0.7 %, well within the instrumental counting uncertainty of 2.2 % (Prescott & Clay 2000). The spectrometer is introduced horizontally into the sample holes which are widened to 80 mm diameter and deepened to 60 cm. All on-site spectrometry values (converted to their dry equivalent) are reported in appendix 5.2.1. Thick Source Alpha Counting and X-Ray Fluorescence spectroscopy Alpha counting in combination with an alternative method to determine the contribution of nonalpha particle emitting potassium-40 is the preferred laboratory technique of DR estimation. The applied technique of Thick Source Alpha Counting (TSAC) has the advantage of being simple, inexpensive and reliable. The sample to background ratio is high and by employing the so-called pairs technique, U- and Th-contents can be discriminated (Aitken 1985). Disadvantages are long counting times that together with background and statistically viable pairs measurements can take up to a week, followed by arithmetic work load. TSAC is based on the principle that alpha particles emitted by a representative dry and powdered sediment sample placed in close contact with a zinc sulphide scintillation screen cause scintillations of light. Amplified by a photomultiplier and converted to sizeable electrical pulses, the alpha activity of the sample is counted. Zinc sulphide is largely insensitive to beta and gamma radiation, resulting only in minor signals that are rejected by a discriminator unit (Aitken 1985). The combined U- and Th-alpha activity can be discriminated by the pairs technique making use of the fact that ~3 % of counts related to the Th decay chain occur as near pairs (within 0.14 ms of each other) as the result
62
of the very short half life of alpha emitter polonium-216. Pairs are separately counted and with the help of an electronic coincidence circuit used to infer separate U- and Th-contents. An inexpensive laboratory method to measure the K-content is by X-Ray Fluorescence spectrometry (XRF). It is based on characteristic "secondary" (or fluorescent) X-rays produced by primary X-rays striking a thick and very homogenous sample (Hutton & Prescott 1992). The XRF machine used at the Mawson Laboratories of the University of Adelaide determines potassium-40 contents with a precision of >97 %. Detailed protocols for both TSAC and XRF analyses are given in appendices 5.2.1.3 and 5.2.1.4. The final conversion results based on Adamiec and Aitken (1998) are listed in appendices 5.2.1.5 and 5.2.1.6. Effect of water content variations Water attenuates ionising radiation more strongly than constituent mineral grains of the sedimentary unit and introduces the largest uncertainty in the calculation of accurate DR estimates (Aitken 1998). While present-day water contents can readily be measured in the laboratory and adequately taken into account in the final DR calculation, they do not necessarily represent the average water content within the sediment over its time of burial (Li et al. 2008). A simple test, using constrained K-, U-, and Th-values and a specific De estimate, was conducted with the AGE99 software (Grn 1994). The result emphasises the sensitivity of the DR and thereby OSL estimate to such variation (see Fig. 2.1A). Accordingly, an increase in the water content from zero to 10 % results in an increase of the apparent OSL-age by 6.9 %.
63
Fig. 2.1A) The isolated effect of water content variations on the apparent OSL-age for sample WL07FP6 (SA).
Equivalent-Dose Estimation
The intensity of the luminescence signal increases with the radiation dose in a nonlinear way differing for every sub-sample of quartz grains. In order to translate the measured natural luminescence signal into an accurate laboratory equivalent-dose (De), the aliquot-specific response to a sequence of administered radiation doses is established. Complications arise from sensitivity changes and partial transfer of deeply trapped charges within the mineral lattice over the sequential process of irradiation and resetting. Single-Aliquot Regenerative-Dose Protocol A recent analytical breakthrough in OSL dating was achieved by development of a method that accounts for both intrinsic variations in the luminescence intensity between grains and the sensitivity changes that occur throughout the dating procedure (Murray & Wintle 2000). The fundamental idea of the single-aliquot regenerative-dose protocol (SAR) is to compare the initially measured natural OSL signal with equivalent sensitivity-corrected signals from laboratoryadministered irradiation-dose responses performed on the same aliquot. Doses are incrementally increased beyond the value of the natural signal resulting in a dose-response growth curve. The De-value for a single aliquot is obtained by projecting the natural sensitivity-corrected OSL signal onto the growth curve and identifying the point of interception. Changes in luminescence that result through thermally resetting the luminescence signal between measurements, i.e. eliminating all
64
residual trapped charges, are monitored by interposing fixed test doses. The measured response is used to normalise the sequence (Murray & Olley 2002). The SAR protocol is employed for all the present samples and the data are processed with the scientific software Analyst 3.22b (Duller 2005). One fundamental advantage of the SAR protocol over previous methods is that it allows for internal checks to exclude deficient aliquots based on their recycling ratio, test dose errors, background difference and curve-fit (Wintle & Murray 2006). Only a small number of quartz grains result in acceptable growth curves that can be used to infer reliable De-values. Tests checking the integrity of aliquots performed on the presented data are based on the rejection criteria recommended by Jacobs et al. (2006). The most important are the sensitivity correction-recycling ratio test performed by repeating the first measurement at the end of the sequence of progressive and corrected sensitivity changes, here set not to exceed >10 %, and the signal intensity test set to demand a minimum intrinsic brightness of valid aliquots of more than three times the background. Aliquot Selection and De Calculation Variations between De-values of multiple aliquots for any given sample result from insufficient exposure to daylight, variable exposure to the flux of ionising radiation within the sedimentary unit, and contamination with more recently bleached material. Only in exceptional geomorphic circumstances are all grains sufficiently exposed to reset their pre-existing luminescence signals. Particularly in fluvial and colluvial environments, a potentially large fraction retain prior De-values. Including such poorly-bleached aliquots in the final De calculation for the sample will result in a potentially significant OSL-age overestimation (Olley et al. 1998). It follows that it is crucial to establish the variability in distribution of De estimates and, by making assumptions on the cause of the observed heterogeneity, determine the one representative of the depositional event to be dated. Variability of De-values from a sample is best addressed by reducing the aliquot size, thereby minimising the effect of averaging. With the development of luminescence readers designed for rapid single grain measurements (Duller et al. 1999), this can now be achieved on the basis of single grains. However, since only a small number of quartz grains produce a valid signal (Duller 2004), similar results are achieved by reducing the aliquot size to the smallest effective number of grains. This is demonstrated by OSL sample WL07-FP 6, for which the De was independently established by single grains and small aliquots of ~20 grains (see Fig. 2.1B). It is now common practice to run at least 12 replicate measurements on small aliquots to assess the error on the overall De estimate (Duller 2004).
65
Fig. 2.1B) Comparison of the variability in De-values between single grains and small aliquots (~20 grains) for sample WL07-FP 6 in the form of probability density plots using Analyst 3.22b. Note that calibrated overall De estimates are near identical when converted to Gy (56.44 6.19 Gy vs. 57.38 3.51 Gy). Sample De estimates based on averaging De-values across all reliable aliquots can result in OSL-ages with little relevance to the actual depositional event. Tails of large De-values in frequency distributions reflect partially-bleached aliquots (see Fig. 2.1B). Depending on the sedimentological context, a fraction of exceptionally small De-values can further reflect post-depositional processes, i.e. bleached grains subsequently introduced by pedogenic activity such as infiltration, desiccation cracking and bioturbation. It is therefore essential to establish discrete populations of De-values and limit the De calculation to fully-bleached aliquots representative of the depositional event to be dated. With prior knowledge of the sedimentological context of the sample, misleading De-values can be manually excluded by graphically- and statistically-aided classification estimates based on probability density and radial plots. Probability density plots replace the De-value of each aliquot with a Gaussian density function with a standard deviation equal to the error of the De estimate. All De-values are summed point-wise to produce a continuous frequency distribution curve. De-values with respective error bars can be added as interval plots, so that the number and spread of components forming modes can be assessed (see Figs. 2.1B & C). While probability density plots are visually intuitive and produce an overall error for selected De-values, they are problematic in that they confound De variation with error estimation, or as (Galbraith 1998) points out, ... obscure[ing] good information in the data by inappropriately weighting in with poor information. More importantly, weighted histograms offer no sound statistical basis to the interpreter to decide upon which group a given aliquot belongs to. Radial plots on the other hand display variability in De-values for a sample by plotting aliquots as points with their precision against the abscissa and their De-value against the ordinate. The latter is
66
scaled in terms of standard deviations from a given central value or radial line (see Fig. 2.1C). More precise De-values plot towards the right and statistically concordant ages fall within a band of two standard deviations about the radial line (Galbraith 1990). Different radial lines correspond to different central De-values. Populations, where more than 5 % of the constituent aliquots fall outside the two standard deviations band, are termed over dispersed. Ideally, the interpretive classification estimate results in a single discrete population of De-value with a Gaussian distribution. Multiple populations, representing a more complex geomorphic history and the occasional younger and older outliers stand out clearly (see Fig. 2.1C).
Fig. 2.1C) Comparison between the probability density plot comprising all reliable aliquots and the corresponding radial plot for sample WL07-G 1. De-values fall either within one of two discrete nearGaussian populations around a younger (red) and an older (green) radial line, or are rejected as outliers. The interim and final results of the De-value selection and corresponding age estimates are presented as probability density and radial plots in appendix 5.2.1.7. Equivalent-Dose Sample Preparation All samples are prepared for De estimation following the protocol of the University of Adelaide luminescence laboratory (see appendix 5.2.1.8). Physical and chemical pre-treatments are conducted with the aim of isolating a homogenous fraction of clean quartz grains for small aliquot (~20 grains) and single grain measurements. The partially-aggregated nature of the sample material requires thorough disintegration by ultrasonic bath and alkali washes (NaOH). Carbonates and organic matter are removed in 20 % HCl and 30 % H2O2, respectively. Treatment with 40 % HF dissolves feldspars and etches the outer rind of the quartz grains affected by alpha radiation. Heavy liquid separation employing Lithium Polytungstate with a specific gravity 2.67 g/cm3 and magnetic separation are used to remove heavier minerals and quartz grains with ferrous coatings or
67
inclusions. Owing to the very fine nature of the sediments, the chemical treatment was conducted on two separate size fractions; the preferred 180-212 m aliquot, and the finer 125-180 m aliquot as a fallback option. In order to remove potential broken quartz grains and only partially dissolved feldspars, both size fractions are re-sieved using 150 m and 90 m sieves, respectively. Final sample quantities are generally less than 0.5 g for the larger (>150 m) fraction, and in two instances the more abundant finer fraction (>90 m) fraction had to be used (appendix 5.2.1). A few samples did not yield sufficient grains altogether and had to be re-sampled or substituted by alternative samples.
OSL data discussion

According to single grain studies performed on samples from the Wilkawillina Silts, on average only one in ten grains of quartz records a valid photon count. This value varies from 4 to 16 % among samples. Partial bleaching is the norm and addressed by minimising the aliquot size to an effective number of ~20 grains where no single grain analyses could be performed. Independent results for single grain and small aliquot analyses are obtained for methodological comparison for sample WL07-FP 6. While the spread in De-values is most pronounced and therefore best recorded by single grains, the weighted overall De estimates turn out statistically identical (see Fig. 2.1B). Wherever feasible, OSL-ages are paired with or bracketed by calibrated AMS-ages. It is important to keep in mind that thereby dated events, i.e. concealment of quartz grains from daylight and the last uptake of carbon from the atmosphere by embedded organic material, are not necessarily synchronous (see section 2.2). A twofold approach was applied to obtain OSL-ages with a meaningful measure of uncertainty. Firstly, separate DR estimates are calculated for all independent laboratory and field measurements. These are corrected for a range of potential water contents over the burial time of the sedimentary unit, since this variable constitutes the largest source of uncertainty on part of the DR estimation. Measured present-day water contents range from ~2.5 to ~10 % across all samples and, where resampled, are found to vary ~5 % between seasons. Hence, these values are used to calculate a maximum and a minimum DR estimate for each sample. In addition, a most likely estimate is determined by averaging the DR estimate for the somewhat arbitrary measured present-day water content with the median DR estimate for a balancing 5 % water content (see Fig. 2.1D and appendix 5.2.1).
68
Fig. 2.1D) Flinders Silts optically-stimulated luminescence (OSL) data, colour-coded by stratigraphic section. Age estimates referring to the last depositional event are depicted as circles, flanked by rectangles indicating the impact of changes in water content for 2.5 % and 10 %. Significant inherited De-populations are plotted as diamonds and significant post-depositional Depopulations as triangles. The linear regression for AMS-dated shells is indicated for comparison with Fig. 2.2D). Secondly, De estimates are calculated by the following three-fold approach. Sets of all reliable aliquots based on the rejection criteria suggested by Jacobs et al. (2006) are compiled from best curve-fits using Analyst 3.22b (Duller 2005). These are independently analysed by the two statistical OSL-age models most commonly applied to partially-bleached sediments; the Minimum Age Model (Olley et al. 1998) and Finite Mixture Model (Galbraith 1988; Galbraith & Green 1990). The appropriate application of statistical age models in determining the population of De-values most likely to represent the last depositional event depends on the sedimentological context and can vary throughout the aggradational sequence. The Minimum Age Model (MAM) can be problematic in the context of post-depositional disturbances (Jacobs et al. 2006). The Finite Mixture Model (FMM) generates multiple populations of aliquots with similar apparent ages, requiring an interpretation within the stratigraphic context.
69
In an alternative procedure, the OSL data are first analysed by an iterative manual approach determining discrete De populations to which the Common Age Model (Galbraith et al. 1999) is applied. Accordingly, sets of all reliable aliquots are displayed as probability density plots and radial plots as automatically generated in Analyst 3.22b, providing a general idea of the spread of aliquots and number of modes. By applying the Common Age Model (CAM), i.e. the weighted mean across all reliable De-values, an initial estimate is calculated and recorded (see appendices 5.2.1 and 5.2.1.7). Consequently, discrete populations are discriminated. While cumulative weighted histograms are not an appropriate means by which to establish discrete populations per se (Galbraith 1998) they can assist in laying out the radial plots. Apparent modes of aliquots are entered as central values for radial lines in Radial Plot 1.3 (Olley & Reed 2003) to see if they encompass the inferred populations and, if so, establish which of the aliquots fall within two standard deviations. Subsequently, outliers are excluded from the probability density plots and the central age of the subsets is determined. The results are entered as central values in the radial plot. The process is repeated if new aliquots are encompassed by the shift of the two standard deviations wide band around the radial line. Finally, aliquots with large errors and little effect on the overall De estimate are excluded from the final calculations in order to reduce the OSL-age uncertainty. This step is followed by last adjustments of radial lines according to the final weighted means of the Devalue populations. The aliquots within the radial plots are colour-coded for documentation purpose; those within two standard deviations and part of the final De estimation in red, those falling outside but still part of the final De estimation in blue, and all aliquots rejected on account of being significant outliers or contributing only noise are depicted in grey. Some samples were found to encompass two significant discrete populations of De-values; a younger displayed in red and an older in green (see Fig. 2.1C). Averaging across such populations would result in an overall De estimate that would translate into an erroneous OSL-age, i.e. an age estimate with no significance to any single bleaching process. Where two populations are encountered, the younger is interpreted to represent the last transport/depositional event or significant post-depositional disturbance. Consequently, the older is interpreted to represent a previous bleaching event unaffected, or only partially affected by later transport process(es) and therefore possibly relating to the age of the sediment load prior to its re-deposition. Both overall De estimates are recorded (see 5.2.1 and 5.2.1.7). The manual approach is further complemented by the Finite Mixture Model (Galbraith 1988; Galbraith & Green 1990) applied to obtain De estimations for all OSL samples. The protocol is discussed together with the results in the manuscript of section 3.1. Full documentation is given in appendices 5.2.1 and 5.2.1.9.
70
References
Adamiec, G., Aitken, M.J. 1998. Dose-rate conversion factors: an update. Ancient TL 16(2), 37-50. Aitken, M.J. 1985. Thermoluminescence Dating. Oxford, Academic Press. Aitken, M.J. 1998. An Introduction to Optical Dating. Oxford, University Press. Bourke, M.C., Child, A., Stokes, S. 2003. Optical age estimates for hyper-arid fluvial deposits at Homeb, Namibia. Quaternary Science Reviews 22, 1099-1103. Duller, G.A.T. 2004. Luminescence dating of Quaternary sediments: recent advances. Journal of Quaternary Science 19(2), 183-192. Duller, G.A.T. 2005. Analyst 3.22b. Aberystwyth, United Kingdom, Institute of Geography and Earth Sciences, University of Wales. Duller, G.A.T., Botter-Jensen, L., Murray, A.S., Truscott, A.J. 1999. Single grain laser luminescence (SGLL) measurements using a novel automated reader. Nuclear Instruments and Methods in Physics Research 155(4), 506-514. Galbraith, R.F. 1988. Graphical display of estimates having differing standard errors. Technometrics 30, 271-281. Galbraith, R.F. 1990. The radial plot: graphical assessment of spread in ages. Nuclear Tracks and Radiation Measurements 17(3), 207-214. Galbraith, R.F. 1998. The trouble with "probability density" plots of fission track ages. Radiation Measurements 29(2), 125-131. Galbraith, R.F., Green, P.F. 1990. Estimating the component ages in a finite mixture. Nuclear Tracks and Radiation Measurements 17(3), 197-206. Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., Olley, J.M. 1999. Optical dating and multiple grains of quartz from Jinmium rock shelter, northern Australia: part 1, experimental design and statistical models. Archaeometry 41(2), 339-364. Grn, R. 1994. Age99, version 09/11/1994, Risoe Laboratories. Huntley, D.J., Godfrey-Smith, D.I., Thewalt, M.L.W. 1985. Optical dating of sediments. Nature 313, 105-107. Hutton, J.T., Prescott, J.R. 1992. Field and laboratory measurements of low-level thorium, uranium and potassium. Nuclear Tracks and Radiation Measurements 20(2), 367-370. Jacobs, Z., Duller, G.A.T., Wintle, A.G. 2006. Interpretation of single grain De distributions and calculation of De. Radiation Measurements 41, 264-277. Li, B., Li, S., Wintle, A.G. 2008. Overcoming environmental dose rate changes in luminescence dating of waterlain deposits. Geochronometria 30, 33-40.
71
Murray, A.S., Olley, J.M. 2002. Precision and accuracy in the optically stimulated luminescence dating of sedimentary quartz: a status review. Geochronometria 21, 1-16. Murray, A.S., Wintle, A.G. 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32, 57-73. Olley, J.M., Caitcheon, G., Murray, A.S. 1998. The distribution of apparent dose as determined by optically stimulated luminescence in small aliquots of fluvial quartz: implications for dating young sediments. Quaternary Geochronology 17, 1033-1040. Olley, J.M., Reed, M. 2003. Radial Plot 1.3, CSIRO Land and Water. Prescott, J.R., Clay, R.W. 2000. Cosmic ray dose rates for luminescence and ESR dating: measured with a scintillation counter. Ancient TL 18(1), 11-14. Prescott, J.R., Hutton, J.T. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements 23, 497-500. Prescott, J.R., Hutton, J.T. 1995. Environmental dose rates and radioactive disequilibrium from some Australian luminescence dating sites. Quaternary Science Reviews 14, 439-448. Prescott, J.R., Hutton, J.T. 1987. Low level measurements of potassium, thorium and uranium by means of scintillometry in the field. Proceedings, Fifth Australian Conference on Nuclear Techniques of Analysis, Lucas Heights, AINSE, Sydney, 76-78. Prescott, J.R., Robertson, G.B. 1997. Sediment dating by luminescence: a review. Radiation Measurements 27(5/6), 893-922. Walker, M. 2005. Quaternary dating methods: an introduction. John Wiley & Sons. Wintle, A.G., Murray, A.S. 2006. A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols. Radiation Measurements 41, 369-391.
72
73
2.2
Accelerator Mass Spectrometry radiocarbon dating
A fool sees not the same tree that a wise man sees William Blake Radiocarbon dating is the oldest radiometric dating method with its fundamental concept formulated in the aftermath of WWII (Libby 1946). It is based on the radioactive carbon isotope carbon-14 (14C) which is formed in the upper atmosphere through neutron bombardment of nitrogen by cosmic rays. Carbon-14 rapidly oxidises to
14
CO2 and disseminates throughout the
atmosphere, hydrosphere and biosphere, while slowly decaying to nitrogen-14 (14N) and beta radiation (formula [1] in Libby 1952). The radioactive decay is compensated by a steady production, resulting in a near-equilibrium throughout the global carbon cycle. Plants fix an equivalent fraction of 14CO2 during photosynthesis and are consumed by other organisms. Hence, 14C can be found in all living matter in a near-equal concentration, or as Libby puts it Since plants live off the carbon dioxide, all plants will be radioactive; since the animals on earth live off the plants, all animals will be radioactive. Thus we conclude that all living things will be rendered radioactive by the cosmic radiation. (Libby 1952). Death results in isolation from the global carbon cycle, followed by an exponential depletion of radiocarbon which has a half life of ~5568 a (Arnold & Libby 1951). Radiocarbon dating measures the residual activity or concentration of 14C and compares it to that of living organisms, thereby making it possible to infer the time of isolation, i.e. the radiocarbon age (see formula [5] in Libby 1952). Conventional radiocarbon dating determines the
14
C-activity by measuring the count rate of beta
radiation. Since the late seventies, an analytical breakthrough in trace element analysis made it possible to directly measure 14C-concentrations, more specifically to differentiate it from 14N with a similar charge-to-mass ratio (Bennett et al. 1977; Muller 1977). This is achieved by accelerating the carbon particles (in form of negatively ionised graphite) to high energies and measuring isotopedependent ionisation rates. The main advantage of atom counting by mass spectrometry over decay counting is an improved efficiency in the order of three magnitudes. In terms of sample quantity requirements, this translates to less than a tenth of a milligram of carbon instead of the few grams required for conventional radiocarbon dating, which in turn allows for a more targeted selection of samples and rigorous pre-treatment. The precision of AMS radiocarbon dating is better than 1 % and reflects the combined errors and uncertainties from counting statistics, standards, measurements and the natural background (Beukens 1992). It is expressed either as one (68.27 %) or two (95.45 %)
14
74 C and thereby
standard deviations from the best AMS-age estimate. The detection limit of
radiocarbon dating is generally given as eight half lives; i.e. ~46 ka BP (Walker 2005). Ages are presented in radiocarbon years before present (BP = 1950). Natural short- and long-term variations in global
14
C-concentration result in discrepancies with calendar years and require
calibration with an independent numerical dating method. Further, plants preferably take up the lighter carbon isotopes 12C and 13C resulting in variation in the equilibrium distribution of the 14C/12Cratio. This isotopic fractionation is largely dependent on the photosynthetic pathway of the plant species and results in variable depletion of radiocarbon. The correction is based on the principle that fractionation between 14C and 12C is approximately double that of the stable isotopes ratio 13C/12C which can easily be measured. Comparison of the isotopic composition with a standard allows for normalisation and is conducted by the radiocarbon facility as a standard correction (Gupta & Polach 1985).
Sample-specific sources of error

In radiocarbon dating, contamination refers to any alteration of the 14C/12C-ratio after death of the organism other than radioactive decay. It is often difficult to detect, gauge and correct and there is a great number of mechanical, chemical and microbiological post-depositional processes with such potential. Old samples with low residual 14C-activity are particularly susceptible to contamination by modern material. The laboratory approach to this problem is sample material- and context-specific, with the general aim to reduce the sample to the one carbon component most likely referring to the event to be dated (Hedges 1992). In any stratigraphic context the implicit assumption of concurrence of the radiocarbon-dated material and depositional event is made. However, it has long been demonstrated for fluvial deposits that this does not necessarily hold true (Blong & Gillespie 1978). Confidence in the reliability of a
14
C-based chronostratigraphy can be significantly increased by pairing radiocarbon
dates based on different sample materials and with independent numerical dating methods such as luminescence dating. Charcoal and charred plant macrofossils: Discrete charcoal pieces (see Fig. 2.2A) are generally considered the preferred sample material in terms of laboratory treatment and radiocarbon dating due to its high inert carbon content and relative low contamination risk (Hedges 1992). However, owing to its high surface area charcoal also has a strong adsorption potential of environmental contaminants and often comprises a wide spectrum of impurities. The standard pre-treatment procedure consists of an iterative acid-alkali-
75
acid (AAA) treatment; hot hydrochloric acid is used to eliminate secondary carbonates and for final neutralisation after warm alkali washes (NaOH) to flush out humic adsorbents. The very longevity of charcoal allows it potentially to endure multiple cycles of sediment storage, erosion and redeposition, presenting a considerable interpretative challenge in terms of the significance of the obtained age in a stratigraphic context (Blong & Gillespie 1978).
Fig. 2.2A) Large discrete piece of charcoal (sample OZJ904 from BRA07-SD) In addition to discrete charcoal pieces, veneers of charred organic detritus were sampled from laminated stratigraphic sections (see Fig. 2.2B). The potential spectrum of burned and degraded plant macrofossils is wide and hence problematic. However, the organic veneers readily disintegrate in water. While this creates procedural difficulties, it also presents interpretative advantages over discrete charcoal pieces in that such fragile components are unlikely to survive reworking.
76
Fig. 2.2B) Dark veneers of organic detritus partly covered by tufa precipitation as seen from the top and side (indicated). Carbonate shells: Another type of material sampled for radiocarbon dating consists of small carbonate shells from aquatic gastropods (see Fig. 2.2C). Carbonate shells are generally perceived as problematic since the gastropods feed on sub-aquatic plants not necessarily in equilibrium with the atmospheric radiocarbon concentration. In the vicinity of groundwater discharge containing dissolved carbonates from limestone formations the aquatic fauna is likely to ingest dead carbon. Finally, shells are prone to diagenetic re-crystallisation with soil carbonates (Walker 2005). However, contamination of carbonate shells very much depends on the geological and depositional environment and preservation. Contamination with pedogenic calcite can easily be differentiated from the aragonite shells by X-ray diffraction and eliminated by the standard chemical pre-treatment procedure removing everything but the core of the shell by acid hydrolysis (Gupta & Polach 1985). As with charcoal pieces, shells have the potential to survive erosion and re-deposition, especially when transported in suspension or by soil creep. However, the degree of fragmentation can be more readily assessed and broken shell fragments excluded from the bulk sample during physical pretreatment.
77
1 mm
Fig. 2.2C) Carbonate gastropod shell sample [Charopidae sp. (left) and Austropyrgus sp. (right)] from the basal unit of BRA-SD prior to chemical pre-treatment.
Calibration
It has early been recognised that the global 14C-concentration fluctuated over the past ~46 ka and at times diverged sharply from the measured historic concentrations (de Vries 1959). Long-term variations are caused by modulations of the cosmic ray flux linked to variations in the earths magnetic field and solar activity affecting the 14C-production rate, and by changes of the reservoir capacity of oceans, rivers and lakes reflecting climate-related impacts on vertical mixing (Hughen et al. 2004). Radiocarbon concentrations are elevated throughout most of the effective time range of radiocarbon dating, particularly during the LGM and its lead-up. This results in considerable underestimation in terms of calendar years. The non-linear offset therefore requires calibration with an independent numerical dating method such as annual tree-rings (Ferguson et al. 1966) now extended to ~12.4 ka, and varved marine sediments extended to ~14.7 ka (Reimer et al. 2004). Beyond,
14 230
C- and
Th-dated fossil corals provide a marine-based calibration of less precision

14
(Fairbanks et al. 2005). A high-resolution marine-derived
C-dated sedimentological and
geochemical record from Cariaco Basin, initially correlated to Dansgaard-Oeschger events recorded in the Greenland GISP2 ice core (Hughen et al. 2004) and now tied to the high-resolution 230Thdated Hulu Cave speleothem record, provides the latest improvement in terms of reliability and precision (Hughen et al. 2006). The updated Cariaco results are in good agreement with independent
78
fossil coral records (e.g. Fairbanks et al. 2005) and employed in the CalPal-2007Hulu-calibration data set (Weninger & Jris 2008) as part of the CalPal-2007 calibration and palaeoclimate research software package (Weninger et al. 2008), used in this study to calibrate all 14C-ages. Despite recent advances, uncertainties in calibrating the older half of the
14
C time scale remain (Hoffman et al.
2008). As a result, the error margin is here presented as two standard deviations and calibrated dates are expressed as a range between minimum and maximum ages before 1950 (cal BP). Abrupt large shifts in 14C-concentration beyond 28 ka cal BP result in prolonged age plateaus, reflected in large error ranges (see Fig. 2.2D).
AMS data discussion

The chronostratigraphic description, interpretation and discussion of the Flinders Silts are based on 93 AMS samples from 14 stratigraphic type sections (section 3.1). In addition to the 30 OSL samples, 43 radiocarbon dates were obtained within the scope of this PhD study; 27 by Single Stage Accelerator Mass Spectrometry (SSAMS) at the Australian National University (ANU) in collaboration with the Cooperative Research Centre for Landscape Environments and Mineral Exploration (CRC LEME), 11 in collaboration with the AMS facility at the Australian Nuclear Science and Technology Organisation (ANSTO) at Lucas Heights, and five commissioned to the Waikato radiocarbon facility in New Zealand. The remainder is re-calibrated from published reports (23 in Cock et al. 1999 and Williams et al. 2001) or presented in co-authorship with Martin A. J. Williams (14) and Peter Glasby (13). The individual AMS dates are best discussed for each stratigraphic type section respectively. However, some overall observations concerning the reliability of different sample material in terms of chronostratigraphic significance are best made in the form of a synopsis. Below, all AMS-dates from the Flinders Silts are plotted as a function of depth below the top of respective stratigraphic sections and sample material (see Fig. 2.2D). Despite this rough comparative approach, some observations made in most individual stratigraphic sections are well illustrated for the whole set. Firstly, the dates obtained from carbonate shells and veneers of organic detritus correlate reasonably well along an approximately linear trend. This suggests similar deposition rates and concurrent termination of the fine-grained aggradational sequences across a range of geomorphologic settings and different catchments. In contrast, charcoal-derived ages are scattered in relation to depth. Dates both older than 30 ka cal BP and younger than 18 ka cal BP are derived from sample depths ranging from 25 to 1,300 cm below top. Throughout the LGM, they largely deviate from the approximate general chronostratigraphic trend of AMS-ages based on shells and organic veneers. This observation is supported by individual pieces of charcoal sampled from organic
79
veneers that returned much older ages presenting outliers in the respective chronostratigraphic sequences. In conclusion, radiocarbon ages derived from charcoal in the studied depositional environment must be treated strictly as post quem dates in any stratigraphical context. While they represent a problematic proxy for dating depositional events, charcoal pieces are to some degree of palaeo-environmental significance for their cumulative occurrence is likely to reflect the presence of woody vegetation and wildfires within the catchment.
Fig. 2.2D) Flinders Silts calibrated radiocarbon data colour-coded by sample material: carbonate shells (blue triangles); charcoal pieces (green circles); veneers of plant detritus (red diamonds). The linear regression for the shell samples is indicated by the dotted line.
80
References
Arnold, J.R., Libby, W.F. 1951. Radiocarbon dates. Science 113, 111-120. Bennett, C.L., Beukens, R.P., Clover, M.R., Gove, H.E., Liebert, R.B., Litherland, A.E., Purser, K.H., Sondheim, W.E. 1977. Radiocarbon dating using electrostatic accelerators: negative ions provide the key. Science 198, 508-510. Beukens, R.P. 1992. Radiocarbon Accelerator Mass Spectrometry: background, precision and accuracy. In: Tayler, R.E., Long, A., Kra, R.S. (eds.), Radiocarbon after four decades: an interdisciplinary perspective. Springer, pp. 230-239. Blong, R.J., Gillespie, R., 1978. Fluvially transported charcoal gives erroneous deposits. Nature 271, 739-741.
14
C ages for recent
Cock, B.J., Williams, M.A.J., Adamson, D.A. 1999. Pleistocene Lake Brachina: a preliminary stratigraphy and chronology of lacustrine sediments from the central Flinders Ranges, South Australia. Australian Journal of Earth Sciences 46, 61-69. Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Kaplan, A., Guilderson, T.P., Fairbanks, T.W., Bloom, A., Grootes, P.M., Nadeau, M.-J. 2005. Radiocarbon calibration curve spanning 0 to 50,000 years BP based on paired 230Th/234U/238U and 14C dates on pristine corals. Quaternary Science Reviews 24, 1781-1796. Ferguson, C.W., Huber, B., Suess, H.E. 1966. Determination of the age of Swiss lake dwellings as an example of dendrochronologically-calibrated radiocarbon dating. Zeitschrift fr Naturforschung Teil A 21(34), 1173-1177. Gupta, S.H., Polach, H.A. 1985. Radiocarbon dating practices at ANU. Canberra, Australian National University Press. Hedges, R.E.M. 1992. Sample treatment strategies in radiocarbon dating. In: Tayler, R.E., Long, A., Kra, R.S. (eds.), Radiocarbon after four decades: an interdisciplinary perspective. Springer, pp. 165-183. Hoffman, D.L., Beck, J.W., Richards, D.A., Smart, P.L., Mattey, D.P., Paterson, B.A., Hawkesworth, C.J. 2008. Atmospheric radiocarbon variation between 44 and 28 ka based on U-series dated speleothem. Geophysical Research Abstracts 10. Hughen, K., Southon, J., Lehman, S., Bertrand, C., Turnbull, J. 2006. Marine-derived 14C calibration and activity record for the past 50,000 years updated from the Cariaco Basin. Quaternary Science Reviews 25, 3216-3227. Hughen, K., Lehman, S., Southon, J., Overpeck, J., Marchal, O., Herring, C., Turnbull, J. 2004. activity and global carbon cycle changes over the past 50,000 years. Science 303, 202-207.
14
Libby, W.F. 1946. Atmospheric helium three and radiocarbon from cosmic radiation. Physical Review 69, 671-672. Libby, W.F. 1952. Radiocarbon dating, 1st ed. Chicago, U.S.A., The University of Chicago Press.
81
Muller, R.A. 1977. Radioisotope dating with a cyclotron. Science 196, 489-494. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., Buck, C.E., Burr, G.S., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hogg, A.G., Hughen, K.A., Kromer, B., McCormac, F.G., Manning, S.W., Ramsey, C.B., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F. W., van der Plicht, J., Weyhenmeyer, C.E. 2004. IntCal04 atmospheric radiocarbon age calibration, 26-0 cal kyr BP. Radiocarbon 46, 1029-1058. de Vries, H.L. 1959. Variation in concentration of radiocarbon with time and location on Earth. In: Abelson, P.H. (ed.), Researches in Geochemistry. New York, John Wiley & Sons, pp. 180. Walker, M., 2005. Quaternary dating methods: an introduction, John Wiley & Sons. Weninger, B., Jris, O. 2008. A 14C age calibration curve for the last 60 ka: the Greenland-Hulu U/Th timescale and its impact on understanding the Middle to Upper Paleolithic transition in Western Eurasia. Journal of Human Evolution 55(5), 772-781. Weninger, B., Jris, O., Danzeglocke, U. 2008. CalPal-2007. Cologne Radiocarbon Calibration & Palaeoclimate Research Package. Available at: http://www.calpal.de/ [Accessed July 28, 2008]. Williams, M.A.J., Prescott, J.R., Chappell, J., Adamson, D., Cock, B., Walker, K., Gell, P. 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International 83-85, 129-144.
Lo ess an d Flo o d s: (Place holder)
82
83
2.3 Particle-Size Analysis

Great things are done by a series of small things brought together Vincent van Gogh The largely structureless and homogenous nature of the fine-grained valley-fill deposits limits the potential of field texturing and sedimentological observations for inferring former depositional environments. Apart from rare faint current cross-bedding structures observed in sandy intercalated beds, and more common imbricated gravel exposures of clast-supported cut-and-fill structures, the fine-grained sediments themselves offer little clue to the interpreter. Field texturing invariably returns silty to sandy loam, sometimes becoming heavier with kneading. This texture change is widely reported in south-eastern Australia from soils associated with loess mantles (e.g. Beattie 1970; McIntyre 1976). Butler introduced the term subplasticity to describe the phenomenon which reflects the presence of clay aggregates resisting mechanical dispersion (Butler 1955; Butler 1976). In order to address the lack of macroscopic sedimentary structures and the deficiencies of field texturing, high-resolution laboratory particle-size analysis (PSA) was performed. The two most important objectives hereby addressed are the inferred aeolian provenance of the material, and the former modes of transport and deposition. The layered to laminated stratigraphic section BRA-AR (Fig. 2.3A) at the confluence of the Aroona and Brachina Creeks was selected for a detailed study on vertical particle size-trends, in order to establish the depositional processes behind observed stacked laminations similar to those discussed for stratigraphic section BRA-SD (section 3.2). In order to decide whether the sediment samples should be sized in their naturally-occurring partially-aggregated condition or in a fully-dispersed state, it proved imperative to first resolve if the aggregates are a pre- or a post-depositional phenomenon. In other words, it needed to be established if the sediments were transported as sand-sized aggregates (Schieber et al. 2007), in which case they can be assumed to record the size- and density-dependent sorting processes, or if the aggregates are the product of in-situ pedogenic processes (Rust and Nanson 1989), in which case they would obscure the sought-after information. The difficulties faced in sizing partially-aggregated sediments are best illustrated by a quote from Dare-Edwards: ... there is no correct particle size distribution for subplastic samples, only a pattern that corresponds to the (dispersion) technique used (1982; 181). However, a large number of previous studies addressed disaggregation of subplastic sediments, providing dispersion protocols (e.g. McIntyre 1976; Norrish and Tiller 1976; Chittleborough 1982).
84
Fig. 2.3A) Layered to laminated stratigraphic section BRA-AR situated few meters upstream of the present confluence of the Aroona Creek tributary with the Brachina Creek trunk channel. The studied interval of laminations is indicated. The narrow size-range and subtle differences between the individual samples called for a particlesizing technique particularly adapted to silt-sized sediments; providing exceptional high resolution, and the largest possible number of discriminated size classes (Goossens 2008). The employed Multisizer 3 COULTER COUNTER, based on the electrical sensing zone principle (Beckman Coulter 2002), proved ideal for the scope of this study (McTainsh et al. 1997). In collaboration with Grant McTainsh, the latest instrument and an array of aperture tubes specifically selected for the study of dust-derived sediments (560, 280, 140 and 50 m) were used over repeated laboratory visits to the Australian School of Environmental Studies at Griffith University, Brisbane. Sample preparation and sizing protocols were adapted from McTainsh et al. (1997), and further developed as described in detail in appendix 5.2.3.1. The particle-size distributions (PSD) were analysed using a parametric statistical approach inspired by Leys et al. (2005). Reported shortcomings, such as restricted data resolution and limited control on curve-fitting algorithms, were addressed by using the open source Mixdist library (Macdonald & Du 2004) within the R environment for statistical computing (R Foundation 2008). Finally, an original protocol allowing the comparison of partially-aggregated with fully-dispersed samples was developed (appendix 5.2.3.2). The methodological advances of the
85
sediment-sizing study were deemed important enough to be presented as a separate paper. The following manuscript provides a review of the historical development of statistical approaches towards analysing PSDs, the new protocol allowing quantification of size distributions and relative abundances of aggregates within samples, and the results for the studied laminations of stratigraphic section BRA-AR (Fig. 2.3B). Detailed protocols are presented in appendices 5.2.3, 5.2.3.1 and 5.2.3.2.
Fig. 2.3B) Laminations of the studied sub-section of BRA-AR, intercalated with orange-mottled gravel sheets and white clasts of local spring tufa.
86
Quantifying particle aggregation in fluvial sediments

Lo ess an d Flo o d s: Ch ap t er 2.3 (Met h o d s) (similar version in revision with Sedimentology) David Haberlah 1, 2, Grant H. McTainsh 3 2) Cooperative Research Centre for Landscape Environments and Mineral Exploration 3) Griffith School of Environment, Griffith University, Brisbane, QLD 4111, Australia
87
1) School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia
Abstract
Sediments largely occur as non-normal size distributions composed of discrete, partially-aggregated particle populations. These populations reflect provenance, dispersal pathways and their depositional environments. The experimental laboratory studies by Schieber and co-authors (Science 318, 2007) describing mud flocculation in turbulent marine systems, prompted us to investigate the potential of aggregates to record size-sensitive depositional processes in a terrestrial fluvial system. Here we use decompositional statistical analysis of sediment-size distributions in their natural condition of particle-aggregate mixtures, which has not been attempted before. We outline and field test a practical and freely-available parametric approach which allows a sediment to be viewed in both its conventional particulate form and as a naturally-occurring mixture of transport-stable aggregates and particles. We also demonstrate that it is only by quantifying aggregation that the provenance and depositional processes forming sediments with complex fluvial and aeolian origins can be understood.
Keywords: aggregation, particle-size analysis, grain-size analysis, component population analysis,
decomposition, mud aggregates.
88
Introduction
The earths surface is covered by sediments and soils composed of mixtures of elementary particles and aggregates. It is well established that parent rocks and weathering processes determine the range of particle sizes and that the size-selective processes of entrainment, transport and deposition by water and wind control the particle-size distribution of the resultant sediments (Udden 1914). The study of particle-size distributions therefore has the potential to unlock the provenance and geomorphic history of sediments. While a large number of particle-size studies have significantly improved our understanding of the formation of deposits from around the world, they have not realised their full potential because of their narrow focus upon elementary particles at the expense of aggregates. Recent flume experimental studies have provided a paradigm shift in our understanding of how mudstones are formed, by demonstrating that mud aggregates flocculate in turbulent marine conditions (Schieber et al. 2007). This observation obliges earth scientists to fundamentally review their long-held assumption that mudstones reflect low energy, usually deep offshore marine environments. These flume studies however stop short of providing a method to measure aggregate-size. In the present study, we come to a similar conclusion about the importance of aggregates in turbulent fluvial terrestrial environments, and demonstrate a method that describes and quantitatively resolves their presence in natural sediments. Based on a field study of flood deposits in the Flinders Ranges, South Australia, we present an innovative approach to sediment-size analysis (as distinct from particle-size analysis) and show that measuring aggregate populations within sediment-size distributions provides new perspectives on the formation of fluvial and aeolian deposits.
Review of earlier studies

There are two main reasons why sediment-size studies have focused upon particle size at the expense of aggregate-size, and these relate to laboratory sizing methods and the statistical analysis of size data. Because of practical constraints on traditional laboratory particle-size analysis, the vast majority of studies have been performed on sediment samples in a dispersed particulate condition. Traditional particle-size analysis involves multiple steps and different methods; such as dry and wetsieving of the sand-sized fraction (>62.5 m), and the use of a pipette or sedimentation balance for the finer silt- (62.5-3.9 m) and clay- (<3.9 m) sized fractions. The exposure of partially-aggregated sediments to such diverse analytical procedures results in uncontrolled breakup of mud aggregates
89
at different steps in the analysis; with unpredictable impacts on resultant size distributions. Aggregates pose an additional problem in analyses where particle size is estimated from settling velocity based on Stokes Law, by not complying with the basic requirement for an overall uniform particle density. The specific density of quartz, which is commonly assumed for elementary particles, significantly over-estimates that of mud aggregates. Conventional sample pre-treatment protocols therefore involve physical and chemical dispersion of sediments into elementary particles that will remain unaltered during the sizing process. With the introduction of sophisticated sediment-sizing instruments based upon the principles of electrical sensing, light and laser diffraction, fine sediment samples can now be sized using a single method (Goossens 2008). These instruments significantly reduce uncontrolled disaggregation during size analysis and circumvent the need for the uniform density assumption. As a result, sediments can be analysed in a variety of sample conditions ranging from minimally-dispersed/ partially-aggregated to fully-dispersed/ particulate conditions, simply by changing the sample pre-treatment (McTainsh et al. 1997). The second reason why sediment-size studies have focused upon particle size at the expense of aggregate-size, relates to past limitations in our capacity to statistically analyse complex particle-size distributions (PSDs). To explain this, a brief overview of the history of statistical particle-size data analysis is presented. Udden (Udden 1898) and Wentworth (Wentworth 1922) were among the first to conclude that PSDs are best presented on log-normal scale, emphasising changes in ratios between size classes rather than their absolute differences. It was early recognized that well-sorted sediments approximate log-normal size distributions, taking up the classic Gaussian bell-shape in a log-normal frequency histogram, or a straight line in a cumulative plot (Richardson 1903) (Otto 1939). This paradigm shift was however soon tempered by the realisation that most sediments depart from this ideal, giving rise to asymmetric or polymodal PSDs in frequency histograms, and multiple straight segments forming zig-zag lines in cumulative plots. During the forties and fifties, such deviations were recognised to reflect mixtures of discrete sedimentary populations that can be approached in three fundamental ways. 1) (Inman 1952) and (Folk & Ward 1957) statistically characterised departures from log-normality of the total-sample PSD by developing a range of graphical measures of mean size, sorting, skewness and kurtosis statistics. 2) (Bagnold 1937) and others (Bagnold & Barndorff-Nielsen 1980)) looked for alternative distribution functions that match particular PSDs best. However, in the early sixties McCammon (McCammon 1962) demonstrated that statistical descriptions of multi-component PSDs are potentially misleading, pointing towards an earlier conceptual approach of 3) resolving PSDs into discrete component populations (CPs) (Bagnold 1941) (Doeglas 1946) (Crocker 1946). In these pre-computer days, researchers faced considerable practical data analysis obstacles. All early graphical, i.e. (Sindowski 1957) (Harris 1958) (Tanner 1958)
90
(Walger 1962) (Fuller 1962) (Spencer 1963) (Visher 1969), and numerical (Clark 1976) decompositional PSD studies were limited to dissecting cumulative curves at break points into multiple truncated Gaussian components, despite Bagnolds early statement that frequency histograms are clearly preferable for the more practical reverse process of finding the regular components of a [sediment] sample (Bagnold 1941):119). In the early seventies, a significant advance was made by Folk in Australia (Folk 1971), by graphically decomposing PSDs in the more realistic scenario of a series of overlapping log-normal curves corresponding to discrete CPs. This approach was followed by a decade of qualitative inverse modelling using the Du Pont 310 Curve Resolver, an analogue computer capable of graphically resolving and quantifying overlapping lognormal CPs in polymodal frequency histograms (Oser 1972) (van Andel 1973) (Dauphin 1980). Possibly due to the requirement of expensive dedicated hardware, a more widespread application was not realised despite the genetically meaningful results of these pioneering studies. With the spread of desktop computers, more rigorous parametric approaches based on analytical inverse curve-fitting models became available, calculating best-fitting sets of continuous unimodal probability distributions (Sheridan et al. 1987). At present, there has been a small number of pioneering decompositional analytical studies that have demonstrated the potential of the method, by accurately decomposing known synthetic complex particle mixtures (Leys et al. 2005), and showing how understanding of the provenance of a wide range of sediments can be improved (Sun et al. 2002). However, these published statistical software solutions either remain commercially unavailable (private VBA script in case of (Sun et al. 2002), or depend on outdated operating systems (DOS in case of Leys et al., 2005) and are restricted by limitations on size-class data. In the present study, we use the public domain Mixdist library (Macdonald & Du 2004) which is part of the R environment for statistical computing (R Development Core Team 2008) distributed online via CRAN. Mixdist was originally developed for ecological population studies (Macdonald & Green 1988) and can be adapted to other needs under the GNU General Public License. Based on the method of maximum likelihood, Mixdist fits finite mixture distribution models to grouped and conditional data using a Newton-type and the Expectation-Maximization algorithms.
Site and Samples

In the present study, we sampled the vertical face of a layered, fine-grained and partially-aggregated alluvial terrace remnant at the confluence of the Brachina and Aroona Creeks in the semi-arid Flinders Ranges of South Australia (S 31.329/ E 138.586). Mouths of tributary streams are often back-flooded by trunk streams, resulting in a layered succession of sediments, referred to as
91
slackwater deposits (Kochel & Baker 1982). A series of flood events can produce stacked finingupward flood couplets which reflect the differential size-, and density-sensitive settling of sediments from a suspension. Typically, fine-grained bands within such couplets reflect multiple fluxes of sediment, separated by dark veneers of organic flotsam plus sediment (Kochel & Baker 1988). As such, slackwater deposits provide an excellent opportunity for detailed analysis of vertical sizetrends. Within a continuous succession of three, complex but largely undisturbed flood couplets, six light bands topped by five dark bands were sampled (see appendix 5.2.3). The dark bands display multiple veneers of plant detritus, interpreted as organic flotsam which has settled out of suspension last, and the associated sediments are assumed to be finer than those in the underlying light bands. The largely undisturbed laminated nature of the veneers indicates minimal postdepositional disturbance.
Methodology
The Coulter Multisizer 3 used to analyse these samples emulates a turbulent fluvial environment during the sizing-process (Beckman Coulter 2002). Samples are immersed in a weak electrolyte (1% NaCl), required for the electrical sensing zone principle to operate, and stirred in a baffled beaker to maintain a uniform suspension. This inevitably breaks up loosely-bound aggregates, but samples quickly reach a stable state in which further dispersion is resisted. This we describe as its transportstable condition, because such a sample would survive transport in non-laminar water flow. Subsequently, we fully disperse the sample into its particulate condition. Physical pre-treatment consists of an ultrasonic bath (Branson 2200 sonifier, 472Hz/ 60W for 30 minutes). Chemical pretreatment consists of removing carbonate cementation in hydrochloric acid (1% HCl). Calgon (a standard dispersant) is not added since this resulted in flocculation of the sample suspension. Reproducibility of laboratory results is confirmed here for both pre-treatment protocols by timetransgressive test runs and from inspection of samples by binocular microscope (Fig. 1A and D). As the Coulter Multisizer 3 requires only very small quantities of sample, the continuous vertical succession of centimetre-thin bands within the slackwater deposit could be analysed. The size distributions of each band are based on more than a million individual particle-size counts and volume measurements, numerically merged by the Beckman Coulter Multisizer software (v.3.51) into 256 discrete size classes (see appendix 5.2.3). Decompositional sediment-size analysis (referred to here as Component Population Analysis - CPA) is performed on both minimally- and fully-dispersed size distributions of the fluvial sediments. This parametric analysis is initiated in unconstrained modus, specifying three hypothetical evenly-spaced Gaussian end-member populations as starting points of the computation. There is no compelling
92
reason to limit CPA to Gaussian distributions. In fact, it has long been demonstrated for the more readily-sizable gravel fraction in fluvial systems that component populations (CPs) are best described by the stretched-exponential Weibull (Rosin) distribution (Krumbein & Tisdel 1940) (Ibbeken 1983) (Kondolf & Adhikari 2000), a claim more recently made for disaggregated fine sediments from a variety of environments (Sun et al. 2002). The process is hence repeated for Weibull distributions, and statistics are compiled for both sample expressions. We observe that Weibull distributions result in slightly superior size correlation among similar sediment populations (see supporting information). We then progress from this unconstrained black-box approach (Weltje & Prins 2007) by using the advantage that each sample is described by two size distributions; reflecting its fully-dispersed, particulate condition and minimally-dispersed, mixed aggregate-particle condition. This computation is initiated assuming a larger than expected number of evenly distributed hypothetical CPs in the fully-dispersed size distribution. Unlikely small and near-identical CPs that produce warning messages from the software are rejected. The means (modes) of the remaining CPs that describe the fully-dispersed size distribution best are then used to partially constrain the calculation of finer CPs in the minimally-dispersed size distribution, leaving additional CPs unconstrained. The rationale behind this approach is that the same elementary particle populations, or composites of them if aggregated, must be present in both distributions. Therefore, CPs comprising entirely of aggregates, in a minimally-dispersed size distribution simply register with a near-zero relative proportion. This approach has the additional important advantage that as the same CPs are quantified in both sample expressions, aggregation per se can be quantified.
93
Fig. 1) Comparison of minimally- and fully-dispersed expressions of sample BRA07-AR 328 from the lowermost light band; (A) minimally-dispersed sample in microscopic view, (B) as decomposed sediment-size distribution, (C) as a pie-chart of relative proportions of its component populations (ca - coarse aggregates, fa - fine aggregates, cp - coarse particles, mp - moderate particles, fp - fine particles). Likewise, the depiction of the fully-dispersed sample expression (D, E, F).
Results
The postulated upward-fining successions between light and overlying dark bands were not detectable from conventional fully-dispersed particle-size analyses and total-sample statistics. The mean, median and modal size of the fully-dispersed light and dark bands are, within statistical error ranges, identical (Table 1A). In contrast, vertical size-trends from light to dark bands are readily apparent from Component Population Analysis. The initial unconstrained CPA assuming three CPs shows a clear upward-fining trend (Table 1B). A more detailed and genetically more meaningful picture of fining- upward trends emerges from the actual sediment populations in the partiallyconstrained CPA runs (Table 1C). The bands analysed in this manner consist of four to five discrete CPs. The results for a single exemplary sample from the lowermost light band are shown in Figure 1. The minimally-dispersed size distribution has two aggregate CPs; coarse aggregates (ca; ~165 m)
94
and fine aggregates (fa; ~89 m), (Fig. 1A and B), which disappear when fully dispersed (Fig. 1D and E). The three fine CPs are largely composed of particles; coarse particles (cp; ~51 m), moderate particles (mp; ~14 m) and fine particles (fp; ~3 m). However, the ~51 m CP comprises a mixture of similar-sized particles and aggregates reflected in the considerable overlap of the populations (cp) and (fa), (Fig. 1B) and as evident from microscopic observations (Fig. 1A and D).
Table 1) Summary statistics: (A) Average mean, median and mode of conventional fully-dispersed (FD) particle-size distributions for the samples of light and dark bands, respectively. The corresponding results of both (B) initial unconstrained (unconst.) and (C) final partially-constrained (p. constr.) Component Populations Analyses are presented as averaged modes of discrete component populations. The relative proportions in both minimally-dispersed (MD) and fullydispersed sample conditions are given. Proportional decreases and increases that result from the breakdown of aggregates are quantified for all corresponding component populations. The CP characteristics of all samples are shown in Figure 2, by plotting the individual modes of all CPs against their relative proportions in each sample. The CPs are grouped into aggregate (circles) and particulate (squares) populations; with component populations of light bands (shades of orange) differentiated from dark bands (shades of blue). The breakup of the aggregate CPs is evident from their disappearance following dispersion with an equivalent proportional increase of particulate CPs in the fully-dispersed samples (triangles, with colours corresponding to light and dark bands).
95
Fig. 2) Cluster plot of all component population modes (on logarithmic scale) versus their corresponding relative proportions. Light and dark bands are colour-coded for both minimallydispersed (MD) and fully-dispersed (FD) sample expressions. Medians (MDN) across component populations of related size classes are indicated. The percentage of aggregates within a sample can be calculated by taking the relative proportions of each CP containing aggregates in the minimally-dispersed size distribution (Fig. 1B) and subtracting their fully-dispersed equivalent CP (Fig. 1E). In the example, the fully-aggregated ~165 m (ca) and ~89 m (fa) CPs contribute ~28.5% to the total size distribution, and aggregates within the ~51 m CP make up an additional ~17.4%. Table 1C shows that for all light band samples, the coarsest two CPs are entirely composed of aggregates and make up ~56% of the total size distribution. Overall, an average of only ~3% of the equivalent ~51 m (cp) CP is aggregated. In contrast, the dark bands display only one aggregate (a) population making up 65% of the total size distribution. With an average mode of ~60 m, it is slightly smaller than the finer aggregated population of the light bands. Its dispersion contributes to all three elementary particle populations which, within statistical error ranges, are identical to those of the underlying light bands. It should be noted that the finest clay-sized CP is close to the detection limit of the sizing technique used here, as indicated by the abrupt drop in the plotted curve (Fig. 1B and E). The actual mode and relative proportion of this CP will therefore differ slightly from the obtained results. Across all samples, the particles making up
96
the aggregates are largely silt-sized particles (~74-~84%), with clay-sized particles contributing the remainder (Table 1; Fig. 2).
Discussion
The results of this Component Population Analysis can be used to infer the depositional environment and provenance of the samples. While the total sample statistics from a conventional particle-size analysis (i.e. fully-dispersed) fail to register vertical size-trends between light and dark bands, the CPA performed on minimally- and fully-dispersed size distributions detects clear upward-fining trends. The detailed process interpretation is that with each successive flood flux, aggregates, as the coarsest constituents, settle out of suspension first. This is reflected in the two sand-sized aggregate CPs in the light bands, while the overlying dark bands feature only one finer silt-sized aggregate CP (a). The fluvial deposition process is therefore best reflected by the aggregates in both bands. In contrast, the particulate CPs do not reflect any vertical size trends between light and dark bands and have very similar size modes. This relative uniformity of the particle size within both the light and dark bands (Fig. 2) is in contrast to the local catchment lithology and supports the earlier inference of Williams and Nitschke (Williams & Nitschke 2005) that the fine-grained alluvial valley-fill deposits of the Flinders Ranges have an allochthonous aeolian origin (loess). The modal sizes of these particulate CPs compare well with those of fluvially-reworked loess in China (Sun et al. 2002), carried out fully-dispersed. Perhaps most significantly, the present field study supports the evidence from the laboratory experiments of Schieber and colleagues (Schieber et al. 2007) that transport-stable mud aggregates are formed by flocculation in turbulent waters, and applies it to terrestrial fluvial environments.
Conclusions
The sediment sizing protocol outlined here represents a breakthrough in the size analysis of alluvial sediments. It is now possible to establish the provenance and geomorphic processes operating in the formation of partially-aggregated sediments in a range of environments; from late Pleistocene finegrained valley-fills (K. Heine & J. T. Heine 2002) in particular, to the mudstones in the geological record in general (Macquaker & Bohacs 2007)(Wright & Marriot 2007). The method also resolves numerical instabilities inherent in the parametric decomposition of single unimodal size distributions (Weltje & Prins 2007), by correlating size distributions of different dispersion states of each sample. The very nature of aggregation can now be examined with a freely-available quantitative tool, capable of resolving discrete aggregate and elementary particle populations. This new methodology
97
has the potential to open the door to a wide range of studies, striving to relate sediment size signatures to sources and formative processes within the landscape.
Acknowledgements
We thank CRC LEME (Cooperative Research Centre for Landscape Environments and Mineral Exploration), the Australian Research Council (ARC grant #DP0559577) and the ARC Environmental Futures Network (grant #58103136) for generous financial support. Thanks also to the Flinders Ranges National Park authorities for research permits and friendly accommodation, Craig Strong for introduction and assistance in the laboratory, and Peter D. M. MacDonald for inspiring correspondence concerning the Mixdist software package.
98
References
van Andel, T.H. (1973) Texture and dispersal of sediments in the Panama Basin. J. Geol., 81, 434-457. Bagnold, R.A. (1937) The size-grading of sand by wind. Proceedings of the Royal Society of London, A 163, 252-264. Bagnold, R.A. (1941) The physics of blown sand and desert dunes, London, Methuen & Co. Ltd, pp. 265. Bagnold, R.A. and Barndorff-Nielsen, O. (1980) The pattern of natural size distributions. Sedimentology, 27, 199-207. Beattie, J.A. (1970). Peculiar features of soil development in parna deposits in the eastern Riverina, N.S.W. Aust. J. Soil Res., 8, 145-156. Beckman Coulter (2002) Multisizer 3 Operator's Manual PN8321681. Fullerton, California, Beckman Coulter, Inc. Butler, B.E. (1955). A system for the description of soil structure and consistence in the field. J. Aust. Inst. of Agricultural Science, 1, 231-252. Butler, B.E. (1976). Subplasticity in Australian Soils. Introduction. Aust. J. Soil Res., 14, 225-226. Clark, M.W. (1976) Some methods for statistical analysis of multimodal distributions and their application to grain-size data. Math. Geol., 8(3), 267-282. Crocker, R.L. (1946) The Simpson Desert Expedition, 1939 Scientific Reports: No. 8 - The soils and vegetation of the Simpson Desert and its borders. Trans. Roy. Soc. S. Aust., 70(2), 235-264. Dare-Edwards, A.J. (1982). Clay pellets of clay dunes: types, mineralogy, origin and effect of pedogenesis. In: Quaternary dust mantles of China, New Zealand and Australia (Ed.) R.J. Wasson, Canberra: Australian National University, pp. 179-189. Dauphin, J.P. (1980) Size distribution of chemically extracted quartz used to characterize finegrained sediments. J. Sed. Petrol., 50(1), 205-214. Doeglas, D.J. (1946) Interpretation of the results of mechanical analyses. J. Sed. Petrol., 16(1), 19-40. Folk, R.L. (1971) Longitudinal dunes of the northwestern edge of the Simpson Desert, Northern Territory, Australia. 1. Geomorphology and grain size relationships. Sedimentology, 16, 5-54. Folk, R.L. and Ward, W.C. (1957) Brazos River bar: a study in the significance of grain size parameters. J. Sed. Petrol., 27(1), 3-26. Fuller, A.O. (1962) Systematic fractionation of sand in the shallow marine and beach environment off the South African Coast. J. Sed. Petrol., 32(3), 602-606. Goossens, D. (2008) Techniques to measure grain-size distributions of loamy sediments: a comparative study of ten instruments for wet analysis. Sedimentology, 55, 65-96.
99
Harris, S.A. (1958) Probability curves and the recognition of adjustment to depositional environment. J. Sed. Petrol., 28(2), 151-163. Heine, K. and Heine, J.T. (2002) A paleohydrologic reinterpretation of the Homeb Silts, Kuiseb River, central Namib Desert (Namibia) and paleoclimatic implications. Catena, 48, 107-130. Ibbeken, H. (1983) Jointed source rock and fluvial gravels controlled by Rosin's law: a grain-size study in Calabria, South Italy. J. Sed. Res., 53(4), 1213-1231. Inman, D.L. (1952) Measures of describing the size distribution of sediments. J. Sed. Petrol., 22(3), 125-145. Kochel, R.C. and Baker, V.R. (1982) Paleoflood analysis. Science, 215, 353-361. Kochel, R.C. and Baker, V.R. (1988) Paleoflood analysis using slackwater deposits. In: Flood Geomorphology (Eds. V.R. Baker and R.C. Kochel), New York: John Wiley & Sons, pp. 357-376. Kondolf, G.M. and Adhikari, A. (2000) Weibull vs. lognormal distributions for fluvial gravels. J. Sed. Petrol., 70(3), 456-460. Krumbein, W.C. and Tisdel, F.W. (1940) Size distributions of source rocks of sediments. Am. J. Sci., 238, 296-305. Leys, J., McTainsh, G.H., Koen, T., Mooney, B. and Strong, C. (2005) Testing a statistical curve-fitting procedure for quantifying sediment populations within multi-modal particle-size distributions. Earth Surf. Proc. Land., 30, 579-590. Macdonald, P.D.M. and Green, P.E.J. (1988) User's guide to program MIX: an interactive program for fitting mixtures of distributions, Ontario, Canada: Ichthus Data Systems. Macdonald, P.D.M. and Du, J. (2004) Mixdist: Mixture Distribution Models. R package version 0.5-2. http://cran.r-project.org/web/packages/mixdist/index.html [Accessed September 29, 2008]. Macquaker, J.H.S. and Bohacs, K.M. (2007) On the accumulation of mud. Science, 318, 1734-1735. McCammon, R.B. (1962) Moment measures and the shape of size frequency distributions. J. Geol., 70, 89-92. McIntyre, D.S. (1976). Subplasticity in Australian soils. 1 description, occurrence, and some properties. Aust. J. Soil Res., 14, 227-236. McTainsh, G.H., Lynch, A.W. and Hales, R. (1997) Particle-size analysis of aeolian dusts, soils and sediments in very small quantities using a Coulter Multisizer. Earth Surf. Proc. Land., 22, 12071216. Oser, R.K. (1972) Sedimentary components of northwest Pacific pelagic sediments. J. Sed. Petrol., 42(2), 461-467. Otto, G.H. (1939) A modified logarithmic probability graph for the interpretation of mechanical analyses of sediments. J. Sed. Petrol., 9(2), 62-76.
100
R Development Core Team (2008) R: a language and environment for statistical computing. In: R Foundation for Statistical Computing, Vienna, Austria, http://www.r-project.org/. Richardson, H. (1903) Sea sand. A lecture before the Yorkshire Philosophical Society, Dec. 1902. Annual Report of the Yorkshire Philosophical Society, 1902, 43-58. Rust, B.R. and Nanson, G.C. (1989). Bedload transport of mud as pedogenic aggregates in modern and ancient rivers. Sedimentology, 36, 291-306. Schieber, J., Southard, J. and Thaisen, K. 2007. Accretion of mudstone beds from migrating floccule ripples. Science, 318, 1760-1763. Sheridan, M.F., Wohletz, K.H. and Dehn, J. (1987) Discrimination of grain-size subpopulations in pyroclastic deposits. Geology, 15(4), 367-370. Sindowski, K.H. (1957) Die synoptische Methode des Kornkurven-Vergleichs zur Ausdeutung fossiler Sedimentionsrume. Geol. Jb., 73, 235-275. Spencer, D.W. (1963) The interpretation of grain size distribution curves of clastic sediments. J. Sed. Petrol., 33(1), 180-190. Sun, D., Rea, D.K., Vandenberghe, J., Jiang, F., An, Z. and Su, R. (2002) Grain-size distribution function of polymodal sediments in hydraulic and aeolian environments, and numerical partitioning of the sedimentary components. Sed. Geol., 152, 263-277. Tanner, W.F. (1958) The zig-zag nature of Type I and Type IV curves. J. Sed. Petrol., 28(3), 372-375. Udden, J.A. (1898) The mechanical composition of wind deposits. Augustina Library Publ., 1, 1-69. Udden, J.A. (1914) Mechanical composition of clastic sediments. Geol. Soc. Am. Bull., 25, 655-744. Visher, G.S. (1969) Grain size distributions and depositional processes. J. Sed. Petrol., 39(3), 10741106. Walger, E. (1962) Die Korngrssenverteilung von Einzellagen sandiger Sedimente und ihre genetische Bedeutung. Geol. Rundsch., 51(2), 494-507. Weltje, G.J. and Prins, M.A. (2007) Genetically meaningful decomposition of grain-size distributions. Sed. Geol., 202, 409-424. Wentworth, C.K. (1922) The shape of beach pebbles. US Geol. Surv. Prof. Pap., 131(C), 75-83. Williams, M.A.J. and Nitschke, N. (2005) Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. S. Aust. Geogr. J., 104, 25-36. Wright, P. and Marriot, S.B. (2007) The dangers of taking mud for granted: lessons from Lower Old Red Sandstone dryland river systems of South Wales. Sed. Geol., 195, 91-100.
3. Results and Discussion

Lo ess an d Flo o d s: to Adolph von Morlot 10th Oct. 1844)
Ch ap t er 3 (Resu lt s an d Discu ssio n )
101
Pray observe I do not pretend to say your theories are not right, but a substratum of facts ought surely to be first given.[ ...] Again I am sure the publication of your Loess views in their present state would injure your reputation: it is a most curious and difficult subject. Charles Darwin (Letter 780
The outcomes of the litho- and chronostratigraphic mapping and geophysical and geochemical laboratory analyses are presented and discussed in three separate papers. The results of the sediment-sizing study from the layered to laminated stratigraphic section at the confluence of the Aroona and Brachina Creeks are presented in the previous section 2.3, because of the overall methodological focus of this paper. The chronostratigraphy of BRA-AR and 12 other sections from all three catchments are discussed in detail in the book chapter of section 3.1. Finally, a high-resolution multi-proxy study performed on the layered to laminated section BRA-SD is presented as a manuscript in section 3.2. Here, the results of detailed lithostratigraphic mapping, parametric sediment-size analysis, quantitative mineral spectroscopy, induced magnetic susceptibility and carbon isotopic geochemistry are discussed in terms of potential palaeo-environmental scenarios.
Ch ap t er 3 (Resu lt s an d Discu ssio n )
102
Of droughts and flooding rains*: an alluvial loess record from central South Australia spanning the last glacial cycle
Lo ess an d Flo o d s: Ch ap t er 3.1 (Result s an d Discu ssio n ) Society, Special Publications, London, in press) David Haberlah
1, 2
103
(similar version published in: Bishop, P., Pillans, B. (eds) 2009, Australian Landscapes. Geological
, Peter Glasby 3, Martin A.J. Williams 3, Steven M. Hill
1, 2
, Frances Williams 4,
Edward J. Rhodes 5, Victor Gostin 1, Anthony OFlaherty 6, Geraldine E. Jacobsen 7

1
Geology & Geophysics, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia (david.haberlah@adelaide.edu.au) Cooperative Research Centre for Landscape Environments and Mineral Exploration
Geographical & Environmental Studies, School of Social Sciences, University of Adelaide, Adelaide, SA 5005, Australia Archaeometry Laboratory, School of Chemistry & Physics, University of Adelaide, Adelaide, SA 5005, Australia
Department of Environmental & Geographical Sciences, Manchester Metropolitan University, Chester Street, Manchester, M1 5GD, UK
6
TAFE South Australia, Adelaide, Australia Institute for Environmental Research, ANSTO, PMB 1, Menai, NSW 2234, Australia
Ch ap t er 3.1 (Result s an d Discu ssio n )
104
Summary
Deposits of proximal dust-derived alluvium (alluvial loess) within the catchments of the now semiarid Flinders Ranges in South Australia record regionally synchronous intervals of fluvial entrainment, aggradation and down-cutting spanning the last glacial cycle. Today, these floodplain remnants are deeply entrenched and laterally eroded by ephemeral traction load streams. The north-south aligned ranges are strategically situated within the present-day transitional zone receiving both topographically-enhanced winter rainfall from the southwest and convectional downpours from summer monsoonal incursions from the north. We develop a regional chronostratigraphy of depositional and erosional events emphasising the Last Glacial Maximum (LGM). Based on 124 ages (94 AMS radiocarbon and 30 OSL) from the most significant terrace remnants on both sides of the Ranges, we conclude that the last glacial cycle including the LGM was characterised by major environmental changes. Two pronounced periods of pedogenesis between ~36-30 ka were followed by widespread erosion and reworking. A short-lived interval of climatic stability before ~24 ka was followed by conditions in which large amounts of proximal dust (loess) were deposited across the catchments. These loess mantles were rapidly redistributed and episodically transported downstream by floods. The termination of this regime ~18-16 ka was marked by rapid incision.
Ch ap t er 3.1 (Result s an d Discu ssio n ) I love a sunburnt country, A land of sweeping plains, Of ragged mountain ranges, * Of droughts and flooding rains. I love her far horizons, I love her jewel-sea, Her beauty and her terror The wide brown land for me! Dorothea Mackellar (1907)
105
Thirty years have elapsed since Bowler (1978) demonstrated the importance of Quaternary depositional history in clarifying some of the climatic and tectonic influences responsible for fashioning the Riverine Plain of south-eastern Australia, observing that while denudation chronology may indeed be dead in some parts of the world; its closest relative, depositional chronology, is alive and well in others (1978, p. 72). Over thirty years earlier Crocker (1946) had argued that the presence of calcium carbonate at depth in the red-brown earth soils of southeastern Australia reflected the past influx of calcareous loess derived from continental shelves and beach deposits exposed during times of lower glacial sea-levels, a conclusion endorsed by Sprigg (1979) and confirmed by recent strontium isotopic analysis of calcretes in semi-arid South Australia (Dart et al. 2007). Our aim in this chapter is to integrate these two themes of depositional chronology and loess influx into an analysis of the late Quaternary depositional history of the semiarid Flinders Ranges of South Australia. Studies of desert margin systems in and outside Australia are highly informative of past climatic influences (Williams et al. 1991; Talbot et al. 1994) since they not only reflect the interplay between alternating and contrasting morphogenetic systems, but the depositional evidence in question has often been well preserved as a result of low rates of erosion linked to the prevailing aridity. Our understanding of landscape evolution is only as good as our understanding of the timing, rate and duration of depositional and erosional events. This is particularly the case for landscape features for which no modern analogue exists such as late Pleistocene loess-derived valley-fills and floodplains. This is illustrated by the longstanding controversy over the nature of the Namib Silts (Srivastava et al. 2006) which have been re-interpreted over the past decades as lacustrine (Goudie 1972), low-energy alluvial (Vogel 1982; Eitel et al. 2005) and high-energy flood deposits (Ward 1987; Smith et al. 1993; Heine & Heine 2002; Leopold et al. 2006). Similar late Pleistocene fine-grained valley-fills from the Sinai Peninsula (Issar & Eckstein 1969; Issar & Bruins 1983; Rgner et al. 2004) and those, largely unnoticed but equally spectacular, from the Flinders Ranges, South Australia, have
106
been re-interpreted in a similar manner (Callen & Reid 1994; Cock et al. 1999; Preiss 1999; Williams et al. 2001; see Haberlah 2006). The most recent studies from each of these sites independently suggest an aeolian source (Eitel et al. 2001; Rgner et al. 2004; Williams & Nitschke 2005) and converge on a depositional model of flood deposition with back-flooded areas where slackwater deposits are preserved (Heine & Heine 2002; Haberlah et al. 2007). Our present study focuses on the fine-grained aggradational sequence from central South Australia, hereafter referred to as the Flinders Silts. In a collaborative effort, we have established a regional chronostratigraphy based on 14 strategically selected stratigraphic sections from major catchments across the Flinders Ranges (Fig. 1). The Flinders Ranges is the longest and highest mountain range in South Australia, extending as a series of mainly north-south trending strike-ridges for over 400 km from the Spencer Gulf inlet of the Indian Ocean deep into the arid heart of the Australian continent. The Ranges consist of a series of tilted, uplifted and dissected weakly metamorphosed Precambrian and Palaeozoic sedimentary rocks (Preiss 1987), rising abruptly from the surrounding plains reaching 1170 m at St Mary Peak (Ngarri Mudlanha). To the west, north and east, the Flinders Ranges are flanked by large structural basins occupied by the playas (salt flats) Lake Torrens, Lake Callabonna/Lake Blanche and Lake Frome.
107
Fig. 1) Overview of field area. Location of stratigraphic sections (red dots); HK refers to sections within the Hookina catchment, BRA to sections within the Brachina catchment, WL to sections within the Wilkawillina catchment, and CAS-1 is situated within the Parachilna catchment. The inset figure indicates the general location of the study area in Australia in terms of relief (SRTM data) and seasonality (adapted from Gentilli 1986). The 124 numerical ages, most of which have not been presented before (Table 1), are discussed and interpreted in terms of regional periods of aeolian and fluvial aggradation, and of reworking and down-cutting events. Insights are thus provided that potentially resolve questions associated with the less-well dated Namib and Sinai Silts. The catchment-based chronostratigraphic record for the Flinders Silts further assists in interpreting continental scale records from adjacent arid central
108
Australia (Lake Eyre) and SE Australia (Willandra Lakes) that register intricate responses to both regional and distal climatic events and which are currently being re-assessed by independent and more accurate dating methods (e.g. Nanson et al. 2008; Kemp & Spooner 2007). The catchments studied here occur across a region trending from the SW to NE, roughly parallel to contemporary (Schwerdtfeger & Curran 1996) and palaeo-wind directions, as inferred from dune patterns in the region (Jennings 1968). The regional synchronicity of recorded events is tested and catchmentspecific neotectonic, as well as sedimentological, threshold overprints of climatic signals are ruled out for the recorded period (see also Quigley et al. 2007). While the timing of deposition of the Flinders Silts covers most of the last glacial cycle, the main focus of the dating program is on the exceptionally well-preserved continuous sequence covering the Last Glacial Maximum (LGM) (212 ka: Mix et al. 2001); from an early lead-up right throughout the period from ~24-18 ka into the early Deglacial. The aim of this paper is to describe the present chronostratigraphic record and discuss the timing, nature and scope of events resulting in the formation of regolith in this part of the world. It provides a framework for ongoing and future palaeo-environmental studies drawing on the rich embedded biological, geophysical and geochemical records.
Field area
Despite early work on Australian loess (Crocker 1946; see Haberlah 2007, 2008), the full significance of the contribution of silt-sized dust on landscape evolution in Australia remains poorly understood (Hesse & McTainsh 2003). Recently described dust records offshore from South Australia (Gingele & De Deckker 2005) and downwind from subtropical eastern Australia (Petherick et al. 2008 a, b, 2009) that span the LGM and beyond, highlight sources of significant amounts of distal fine-grained dust from central and SE Australia. The Flinders Ranges are strategically located in the midst of this late Pleistocene dust bowl. The more than 400 km long sequence of mountain ridges forms a longitudinal topographic barrier with a general elevation above 300 m ASL and a chain of summits culminating at 1170 m. Hence, the Ranges efficiently harvest both low-lying clouds and proximal dust entrained from adjacent plains (Goossens 1988). As such, the Flinders Silts are a proximal equivalent to distal dust records that reach as far as Antarctica (Revel-Rolland et al. 2006). It follows that a different scale of sedimentation rates is recorded, with >10 m thick sequences of loess washed into the valleys and plains and choking the narrow gorges. At specific intervals over the last glacial cycle, silt- and fine sand-sized dust fallout that mantled the slopes was entrained by runoff and episodically transported downstream. Today, the slopes are largely stripped of dust mantles and ephemeral traction load channels deeply entrench the silts. The aeolian provenance of the Flinders Silts is evident from a number of field and analytical observations, particularly their widespread
109
occurrence regardless of the underlying bedrock lithology (see Williams & Nitschke 2005; Williams et al. 2001, 2006). The fine-grained deposits make a significant contribution to local closed catchments within narrow quartzite synclines (e.g. Wilpena Pound) and they interfinger with angular and probably frost-shattered quartzite clasts at the base of the westernmost range front. Pockets of loess are preserved on all major ridges and return geochemical signatures inconsistent with in situ weathering (Williams & Nitschke 2005). Finally, aggradation rates far exceed weathering rates within the respective catchments (Williams et al. 2001).
Age proxies and models

The chronostratigraphy of the Flinders Silts is based on two independent dating methods: radiocarbon dating and optically stimulated luminescence dating (OSL). The 94 radiocarbon age estimates are based on different types of material, including organic materials (charcoal [39], veneers of charred plant detritus [13], and organic clay [1]) and calcium carbonates (aquatic gastropod carbonate shells [39], tufa [1] and emu egg shell [1]). The 30 OSL-age estimates are largely based on small aliquots, i.e. ~15 grains (19), and single grains of quartz (5), 180-212 m in size or, where insufficient, 125-180 m, and are analysed by various age models (Table 1). A brief discussion of their significance and reliability as proxies to infer the depositional age of host sediments will facilitate the interpretation of the stratigraphic sections.
Flinders Silts radiocarbon dates

Radiocarbon samples were pre-treated, the carbon extracted and converted to graphite using standard methods (Hua et al. 2001). The graphite was dated by Accelerator Mass Spectrometry (AMS), by measuring the amount of residual
14 14
C in the material and calculating the age in
radiocarbon years (yrs BP), (Tuniz et al. 1998). In most cases the precision, expressed as one standard deviation from the best C-age estimate is better than 1 % and reflects the combined
14
errors and uncertainties from counting statistics, standards, measurements and the natural background. All ages fall within the detection limit of C, generally given as eight half lives (i.e.
~46,000 radiocarbon years), (Walker 2005). Atmospheric radiocarbon concentrations are elevated throughout the time range covered, particularly during the LGM and its lead-up. This results in considerable underestimation in terms of calendar years. Recently, calibration using independent numerical dating methods beyond the limits of annual tree-rings (Ferguson et al. 1966) and varved marine sediments which extend to only to ~14.7 ka (Reimer et al. 2004), now allow the calibration of
14
C-dates from the LGM and beyond (Fairbanks et al. 2005). All 14C-ages in this study are calibrated
by the currently most reliable and precise high-resolution marine-derived 14C-dated sedimentological

230
110 Th-dated Hulu Cave
and geochemical record from the Cariaco Basin, tied to the high-resolution
speleothem record (Hughen et al. 2006), using the integrated CalPal-2007Hulu-calibration data set (Weninger & Jris 2008) as part of the CalPal-2007 calibration and palaeoclimate research software package (Weninger et al. 2008). Despite all recent advances, uncertainties in calibrating the older half of the 14C time scale remain (Hoffman et al. 2008). As a consequence, the error margin of the calibrated 14C-age estimates is presented as two standard deviations, and is expressed as a range between minimum and maximum ages before 1950 (cal BP), (see Table 1). Abrupt and large shifts in
14
C concentration beyond 28 ka cal BP result in prolonged age plateaus, reflected in large error
ranges. When all 14C-age estimates from the Flinders Silts are plotted as a function of depth below the top of respective stratigraphic sections and sample material, two important conclusions arise. Firstly, the dates obtained from carbonate shells and veneers of organic detritus correlate reasonably well along an approximately linear trend (Pearson coefficient r = 0.79, excluding 2 Holocene ages), suggesting similar deposition rates and concurrent termination of the fine-grained aggradational sequences across a range of geomorphological settings and different catchments (Fig. 2). Secondly and in contrast, the charcoal-derived age estimates are scattered with regard to depth (r = 0.32, excluding 2 Holocene ages). Here, dates both older than 30 ka cal BP and younger than 18 ka cal BP are derived from similar sample depths ranging from ~25-1,300 cm below top. Throughout the LGM, many dates obtained from charcoal depart from the stratigraphic age-depth trend based on 14C-ages from shells and organic veneers, as well as on luminescence ages.
111
Fig. 2) Flinders Silts accelerator mass spectrometry (AMS) radiocarbon data, colour-coded by sample material. The trend line (linear regression) for AMS-dated veneers of organic detritus and carbonate shells (excluding two Holocene samples) is indicated. This observation is consistent with individual pieces of charcoal sampled from organic veneers that, according to their low 13-values, represent woody elements of vegetation and return much older ages, presenting outliers in the respective chronostratigraphic sequences (e.g. bold green diamonds in Fig. 2 & Table 1). It is likely that this reflects reworking of charcoal caused by erosion and subsequent re-deposition (Blong & Gillespie 1978). In conclusion, 14C-ages derived from charcoal in a depositional environment of alluvial loess need to be treated with caution and interpreted as maximum ages. While charcoal fragments represent a poor proxy for dating depositional events, collectively they do reflect intervals of widely-occurring woody vegetation burnt by wild fires within the catchment. By plotting the frequency distribution of all 14C-ages based on charcoal irrespective of their sample position versus those based on shell, an interesting palaeo-environmental picture emerges. According to the timeline, ages based on shell closely succeed those based on charcoal (Fig. 3). From this it can be inferred that intervals marked by droughts, dying wood and wildfires (charcoal) are followed by floods (shells) that entrain and redistribute the age proxies along with readily erodible sediments.
112
Fig. 3) Frequency distribution of calibrated 14C-ages based on charcoal pieces (brown) and shell (blue) on a linear scale from 40-10 ka cal BP.
Flinders Silts luminescence dates

Optically-stimulated luminescence dating takes advantage of the characteristic of common mineral grains, such as quartz sand, to accumulate stored energy in their crystal lattices which is released on exposure to light. Consequently, sediments buried in a depositional event gradually acquire an increasing energy that can be freed and measured in the laboratory by stimulation. Light-sensitive electrons can be released by optical stimulation (Huntley et al. 1985). The energy absorbed by the sample mineral depends on its exposure to radioactivity from the decay processes of naturallyoccurring radioactive elements within the sediments and on cosmic irradiation, which together make up the environmental dose rate (Dr). Except for the oldest Hookina Floodplain samples, the Dr was remarkably similar across all samples discussed, probably a reflection of their common aeolian source (Table 2). At the altitudes of the study area, cosmic ray contribution is largely a function of overburden thickness (Prescott & Hutton 1994) and is trivial relative to the radioactive contribution of the fine-grained sediments. The largest uncertainty in the calculation of accurate Dr is water content. Water attenuates the ionising radiation more than constituent mineral grains of the sedimentary unit (Aitken 1998). While present-day water contents can readily be measured in the laboratory and adequately taken into account in the final Dr calculation, they do not necessarily represent the average water content within the sediments over its time of burial (Li et al. 2008). Measured present-day water contents range from ~2.5-10 % across all recently collected samples and, where re-sampled, were found to vary up to ~5 % between seasons. Hence, these values were used to calculate a maximum and a minimum Dr-estimate for each sample, and to determine a most likely estimate by averaging the Dr-estimate for the measured present-day water content with that obtained with 5 % water content (Fig. 4 & Table 2). Quartz grains for this study were sampled, extracted and prepared using standard techniques (see for example Huntley et al. 1993). The intensity of the luminescence signal increases with the radiation dose in a nonlinear way, differing for every sub-sample of quartz grains. In order to translate the measured natural luminescence signal into an accurate laboratory equivalent-dose (De), the aliquot-specific response to a sequence of laboratory-administered radiation doses is
113
established by the single-aliquot regenerative-dose protocol (SAR), comparing initially measured natural OSL-signals with equivalent sensitivity-corrected signals from irradiation-dose responses performed on the same aliquot (Murray & Wintle 2000). The SAR protocol allows for internal checks to exclude deficient aliquots based on their recycling ratio, test dose errors, background difference and curve-fit (Wintle & Murray 2006). For the Flinders Silts samples, we used the rejection criteria recommended by Jacobs et al. (2006). Significant variations between De-values of multiple aliquots of a sample either reflect insufficient exposure to daylight during the transport process or postdepositional introduction of more recently bleached sediments by infiltration, desiccation cracking and bioturbation. Only in exceptional geomorphological circumstances are all grains sufficiently exposed to reset their pre-existing luminescence signals. Particularly in fluvial and colluvial environments, a potentially large fraction retain prior De-values. Averaging such poorly-bleached sample De-values will result in a potentially significant OSL-age overestimation (Olley et al. 1998). It follows that it is crucial to establish the variability in distribution of De-values statistically. A prerequisite for this is small aliquot sizes, minimising the effect of averaging. With the development of luminescence readers designed for rapid single grain measurements (Duller et al. 1999), this can now be achieved on the basis of single grains. However, only a small number of quartz grains produce a valid signal (Duller 2004), so that similar results are obtained by reducing the aliquot size to the smallest effective number of grains or small aliquots. This is demonstrated by OSL-sample WL07-FP 6, for which De-values were independently established from single grains and small aliquots of ~15 grains (Table 1 & 2). Consequently, the application of appropriate statistical age models to infer the depositional population is crucial to the interpretation of the data. The variability in De-values is best displayed in radial plots by plotting aliquots as points with their precision against the x-axis (increasing towards the right) and their De-value against the y-axis (Olley & Reed 2003). The latter is scaled in terms of standard deviations (STs) from a given central value or radial line (Table 2, final column). More precise De-values plot towards the right and statistically concordant ages fall within a band of two STs about the radial line (Galbraith 1990). Different radial lines correspond to different central Devalues and OSL-ages. Populations where more than 5 % of the constituent aliquots fall outside the two STs band are termed over-dispersed. In the present study, the number of populations and the degree of over-dispersion is statistically determined by the best data fit (i.e. the lowest Bayesian information criterion value (BIC) while observing the maximum log-likelihood (Lmax) to avoid overfitting), using the Finite Mixture Model (FMM), (Galbraith & Green 1990; Galbraith 2005). The resulting discrete populations are presented both numerically and as radial plots (Table 2). This approach allows the evaluation of any De-populations within a stratigraphic context as either
114
depositional, inherited or post-depositional. In this study, only discrete De-populations exceeding 25 % are discussed in terms of ages, the remainder is accounted for as outliers. Some caution needs to be exercised in interpreting inherited populations as proxies indicative of residual ages of the reworked sediments, for any previous Dr environments are likely to have differed from the present. However, the environmental dose rates for the Flinders Silts samples are very similar and cosmic ray contributions (and variations) minor. According to results of the FMM analysis, two samples are best described by a single population (K1887, K1888), for which the Central Dose Model (CDM) was employed (Galbraith et al. 1999; Galbraith 2005), (Table 2). In the most recent chronological studies of the Namib Silts, the Minimum Age Model (Olley et al. 1998; Galbraith 2005) is used (Bourke et al. 2003; Srivastava et al. 2006), revising and shifting a previously established LGM chronostratigraphy based on
14
C-ages (Vogel 1982) to the Holocene. To allow for a comparison,
results returned by this age model are reported as well (Table 1). However, large unsystematic discrepancies to otherwise orderly and paired internal chronostratigraphies reveal the inadequacy of the Minimum Age Model in the context of alluvial loess sedimentation.
Fig. 4) Flinders Silts optically-stimulated luminescence (OSL) age-depth plot, colour-coded by stratigraphic section. Age estimates referring to the last depositional event are depicted as circles flanked by rectangles indicating the impact of changes in water content for 2.5 % and 10 %. Significant "inherited" De-populations are plotted as diamonds and significant "post-depositional" Depopulations as triangles. The trend line (linear regression) for AMS-dated veneers of organic detritus and carbonate shells (excluding two Holocene samples) is indicated for comparison with Fig. 3.
115
Chronostratigraphy of the Flinders Silts

The description of stratigraphic sections follows a transect from the western piedmont plain across into the Brachina catchment to the Wilkawillina catchment on the eastern side, which is approximately in line with prevailing contemporary (Schwerdtfeger & Curran 1996) and palaeo-wind directions dominated by the winter westerlies (Jennings 1968; Shulmeister et al. 2004), (Fig. 1). While the overall thickness of the reworked loess deposits decreases in a downwind direction, most evident in a comparison between the western and eastern piedmont plains, the actual depth of the silts in the stratigraphic sections is a complex function of the size of the upstream catchment, the local topography and the degree of incision.
116
Hookina Silts
Lake Torrens
Hookina Cr
Fig. 5) Hookina Silts stratigraphic sections as seen from the south at an altitude of ~10.16 km on Cnes/Spot imagery (23/10/2004) projected on the SRTM DEM (elevation exaggeration factor 3) in Google Earth 4.3, with Lake Torrens playa to the west and the location of the Brachina Silts stratigraphic sections in the background. Locations of all sections within the Hookina catchment are indicated by red pointers with letters corresponding to section names (see also Fig. 1). Hookina Creek is deeply entrenched into a broad, gently rolling floodplain. A sequence of stacked calcareous palaeosols is exposed along vertical banks of unconsolidated fine-grained alluvium with near-horizontal bedding planes and localised cross-stratified sandy bands. Earlier attempts to date these regional Bca-horizons, which can be traced laterally for hundreds of metres, faced the constraints of radiocarbon dating of pedogenic carbonate (Williams 1973), (for discussion of the limitations and assumptions involved see Williams & Polach 1971; Callen et al. 1983; Fontes & Gasse 1989), and sample quantities required for conventional radiocarbon dating (Williams 1982). Three stratigraphic sections along the thalweg of the trunk channel are presented and OSL-dated (Fig. 5), establishing the onset and termination of Hookina Silts aggradation and the age of the parent
117
material of the major palaeosols. HK07-D is the most upstream stratigraphic section from the western piedmont plain within the Hookina catchment and is less than two kilometres west of the range front. Its surface is concordant with the highest and oldest floodplain and exhibits the uppermost well-developed Bca-horizon, which is also developed within both downstream sections. Approximately six kilometres downstream, stratigraphic section HK07-L functions as a spatial link between HK07-D and HK07-M, another kilometre downstream and separated by a partially loessdraped quartzite ridge.
Fig. 6) Hookina Silts stratigraphic sections; main lithostratigraphic units and chronostratigraphy.
118
The 8.5 m high vertical cliff face of HK07-D hosts a sequence of at least four Bca-horizons (Fig. 6) that extend laterally for ~500 m. Each horizon is topped by ~30 cm thick continuous red (e.g. 5YR 5/6 dry 4/6 moist) sheets consisting of sand-sized mud aggregates with lenses of rounded and platy pebbles of variable lithology, hereafter referred to as chromatic bands to include similar colours. The uppermost band mantles the present surface and exhibits a platy structure with rootlets and insect burrows. The buried chromatic bands are more compacted, of blocky to prismatic structure and in places have slickensides and organic streaks. All, including the chromatic surface drape, are associated with discontinuous gravel. A central unit between ~100-375 cm is sandier, with poorly preserved cross-stratification and intercalated silt drapes. A thin gravel sheet at ~230 cm depth forms a local disconformity, in places scouring up to ~30 cm into the underlying sediments. The imbricated, clast-supported, mostly platy, gravel indicates flow directions from the NE to NNE. The three OSL-ages from the basal Bca-horizon and the two chromatic bands topping the well-developed central Bca-horizons indicate rapid aggradation of the host material between 84.6 7.0 ka (K1887) and 82.5 13.5 ka (K1889), although the large error ranges potentially extend that interval to ~20 ka. The second sample from the base (K1888) is close to saturation and the few valid De-aliquots are likely to reflect only an inherited age of 119.7 12.8 ka. The base of the parent material of the uppermost Bca is dated to 19.3 1.6 ka (K1890), with a large inherited population suggesting a source of reworked material last deposited 38.9 2.7 ka. Derived from the foot of the ruins of the old Hookina Hotel, abandoned half a century ago, HK07-L displays a distinct twin chromatic band associated with and resting on a prominent composite Bcahorizon (Photo 1). This band, at a depth of ~275-325 cm (Fig. 6) peters out a few hundred metres upstream but can be traced laterally for more than one kilometre downstream, where it terminates in stratigraphic section HK07-M. The lower band is dated to 36.1 3.0 ka (K1891) and the upper, separated by a light carbonate-indurated silt drape, to 30.2 1.7 ka (K1892). Abundant artefacts from the former hotel are incorporated in the uppermost chromatic surface drape. The freshly collapsed 11 m high vertical stratigraphic section HK07-M presents the most complete record from the Hookina Floodplain. Here, the Hookina Silts rest on fluvial sands and gravel, and continue without obvious erosional breaks towards the top of the section, level with the highest floodplain surface (Fig. 6). The lowermost ~4 m are layered, comprising multiple sheets of rolled detrital calcareous nodules (transported nodular calcrete) from reworked Bca-horizons, which alternate with seven distinct ~20 cm thick undulating red bands consisting of well-sorted fine sand and silt. In rare instances, these bands are associated with narrow gravel-filled chutes and overlie in situ calcareous rhizocretions. Sheets of nodular calcrete presently occur on and downstream of
119
erosional surfaces truncating the uppermost calcareous Bca-horizon as the result of rare overbank flood events. No modern analogue for the red bands was observed, but colour and modal sizes match the fine component of dune fields that extend from Lake Torrens towards the range front (Fig. 1). The fluvial sands below the silts are dated to the onset of the last glacial at 108.2 11.1 ka (AdGL-08004) and the uppermost red band was deposited at 85.6 17.9 ka (AdGL-08005), with a dominant inherited component aged 125.1 ka 9.6 ka. The lowermost sample (AdGL-08004) was taken less than a metre below a large cut-and-fill structure from which large mottled root casts extend and its minor post-depositional age 23.0 3.0 ka is likely to reflect rapid early LGM-infilling with reworked fine-grained material. Deposition of the topmost Bca-horizon is dated to 18.0 1.6 ka (AdGL-08006) and contains a significant post-depositional population of 8.3 0.9 ka, attributed to bioturbation and pedogenesis, possibly linked to the formation of the Bca-horizon.
120
Brachina Silts
Rawnsley Quartzite Aroona Cr
Nuccaleena Dolomite
Brachina Cr
Etina Cr ABC Quartzite
Fig. 7) Brachina Silts stratigraphic sections as seen from the south at an altitude of ~3.4 km on Cnes/Spot imagery (23/10/2004) projected on the SRTM DEM (elevation exaggeration factor 3) in Google Earth 4.3, with Aroona Creek joining in from the north between the steep ABC Range Quartzite and Rawnsley Quartzite ranges. Locations of all sections within the Brachina catchment are indicated by red pointers with letters corresponding to section names (see also Fig. 1). The fine-grained valley-fill formations within the Brachina Creek catchment form an intra-montane floodplain remnant upstream of the narrow Brachina Gorge that cuts through a sequence of resistant ridges (Fig. 7). They were initially mapped and described as lacustrine deposits (Callen & Reid 1994; Cock et al. 1999). A theodolite survey and dating program in 2000 established that they form a gently sloping surface parallel to older rock-cut terraces and the present thalweg, extending into the gorge. This led to their interpretation as alluvial wetland deposits (Williams et al. 2001). Additional dating, DGPS-surveys, detailed sedimentological descriptions and geophysical and geochemical analyses of six stratigraphic sections, has led to the present re-interpretation of the Brachina Silts as a succession of two intra-montane floodplains. Areas upstream of bedrock
121
constrictions, tributary mouths and protected embayments back-flooded during major precipitation events, resulting in laminated slackwater deposits. Three stratigraphic sections are presented from the Brachina Floodplain upstream of the ABC Range: 1) BRA-SA situated upstream where the floodplain is confined by a rise of the Nuccaleena Dolomite Formation; 2) BRA-SG half-way along Etina Creek, which according to imbrication patterns of cut-and-fill structures embedded within the floodplain may have persisted as a main line of flow throughout the floodplain aggradation; and, 3) BRA07-SD which is within a slight depression ~300 m to the west of the trunk channel at the foot of the steep eastern valley slope in an area back-flooded by the constriction posed by the ABC Range. BRA07-SD is entrenched by the present channel of Brachina Creek. Three stratigraphic sections follow downstream from remnants within the Brachina Gorge. BRA-LW is within the deeply entrenched trunk channel near the eastern mouth of the gorge. Downstream of the constriction formed by the quartzite ABC Range, Brachina Creek meets the wide Aroona valley, which is eroded into shales of the Bunyeroo Formation. Stratigraphic section BRA07-AR is located where Aroona Creek joins Brachina Creek. Finally, just upstream of the steeply dipping Rawnsley Quartzite, Brachina Creek makes a sharp turn into an embayment eroded by three minor tributaries. Within this back-flooded area, the ~18 m high terrace remnant BRA07-G is preserved.
Fig. 8) Brachina Silts stratigraphic sections; main lithostratigraphic units and chronostratigraphy. For legend see Fig. 6. BRA-SA is a ~6 m sequence of Brachina Silts and gravel (Fig. 8). Aggradation rates correlate well with the two other sections from the Brachina Floodplain. Two features set it apart, however: the more
122
oxidised (red) colour of the silts (7.5YR 5/8 dry - 4/6 moist); and, three to four generations of intercalated widespread cut-and-fill structures of clast-supported gravels. The lowermost is below fine-grained sediments dated to 24.8 0.1 ka cal BP (Beta-84141), followed by two generations of chutes cutting into material aged between 22.2 0.2 ka cal BP (OZC709) and 23.0 0.2 ka cal BP (OZC704). The uppermost gravel spreads out as a continuous sheet on the silts. The sequence is topped by another metre of fine-grained sediments with a Bca-horizon developed below a chromatic surface drape (5YR 6/6 dry - 4/6 moist). The uppermost two metres have been dated in detail in the downstream section BRA-SG and show continuous aggradation until 18.1 0.2 ka cal BP (Wk-6548). BRA-SG is marked by multiple thin layers of detrital calcareous casts of sedges and tufa clasts. Here, the Brachina Silts are part of an overall greyish unit that increases in thickness towards the range front, and indicates waterlogged conditions (Cock et al. 1999). However, as the undisturbed sequence of
14
C-dates based on
gastropod shells indicates, bioturbation must be assumed minimal and the growth of the now fossilised vegetation was restricted. BRA07-SD is further downstream and across the floodplain in a slight depression at the base of a bedrock slope (Photo 2). This stratigraphic section is distinctly layered to laminated in its upper five metres and massive and more chromatic in its lower part (Fig. 8). The sharp boundary between these units is marked by a ~10-20 cm thick, brown (7.5YR 5/2 dry - 4/2 moist) chromatic band previously interpreted as a palaeosol (Williams et al. 2001). The microfossil record of BRA07-SD indicates a major palaeo-environmental change at and above the palaeosol, from a large diversity of unbroken freshwater species of ostracods and gastropods to a lower diversity reflecting high energy (fragmented shells) more saline and intermittent conditions (Glasby et al. 2007). Despite the close spacing of
14
C-dates, it proved impossible to quantify the hiatus between the two
aggradational units with certainty. Age estimates based on the lowermost veneer of organic detritus indicate an onset of laminated aggradation at 23.8 0.3 ka cal BP (SSAMS ANU-4205). Older ages further up the sequence are based on charcoal pieces. The top of the lower unit (i.e. the material of the palaeosol) is dated to 25.7 0.5 ka cal BP (OZJ909) based on shell, followed by large charcoal pieces returning ages between 32.2 0.3 ka cal BP (SSAMS ANU 2036) and 33.6 0.4 ka cal BP (OZJ904). Associated shells are dated to 32.0 0.2 ka cal BP (OZJ907) and 32.6 0.6 ka cal BP (OZJ908), suggesting rapid aggradation. This inference is consistent with the OSL-age of 32.8 2.8 ka (AdGL-96003) based on large aliquots. However, a number of indicators bring into question the depositional significance of these age proxies; two further OSL-samples below return unequivocally ages of 24.9 ka 1.4 ka (AdGL-96006) and 24.9 ka 2.5 ka (AdGL-96002), which in more water
123
saturated conditions could prove a slight underestimate (Williams et al. 2001). AdGL-96003 is sampled from a layer marked by transported nodular calcrete most likely derived from former Bcahorizons stripped from the slope or upstream remnants of a previous generation of floodplain. This suggests that the older ages are inherited and the sediment reworked. Finally, two paired 14C-ages from shell and small charcoal pieces collected at the present base return ages of 28.6 0 0.4 ka cal BP (SSAMS ANU-2039) and 28.1 0.1 cal BP (SSAMS ANU-1811), respectively. The layered upper sequence of BRA07-SD aggraded episodically throughout the LGM. At intervals, discrete light bands consisting of silt and fine-sand subdivide darker laminated units. These light bands are continuous for over 100 m and can be traced towards the trunk channel to the SW, increasing in thickness and finally associated with gravel and cross-stratification. They are topped by undulating couplets of grey fine-grained sediments, organic veneers and streaks of tufa. The lower 3 m of the layered upper unit aggraded prior to ~22 ka (Fig.8). Towards 19.3 0.2 ka cal BP (SSAMS ANU-4109), a final compact succession of light bands was deposited. In contrast to underlying sediments, it is affected by pedogenesis and bioturbation extending from the uppermost Bcahorizon. The layered and laminated sequence terminates at 18.3 0.2 ka cal BP (SSAMS ANU-4107) with lenses of gravel and is partially mantled by a chromatic surface drape. Situated near the mouth of Brachina Gorge within the trunk channel is BRA-LW, a section of Brachina Silts entrenched to more than 13.5 m (Fig. 8). Four
14
C-ages based on charcoal pieces
collected from within a metre return ages ranging from 30.2 0.3 ka cal BP (Wk-6562) to 32.3 0.2 ka cal BP (Beta-96171), similar to those in BRA07-SD but 6 m lower within the sequence. However, samples closer to the base dated to 19.7 0.2 ka cal BP (Wk-6564) and 16.9 0.1 ka cal BP (Wk6561) again call into question the extent that these reflect the depositional event or the age of reworked sediments. The Brachina Silts spread across the second intra-montane floodplain where the Brachina Creek debouches from the narrow gap through the ABC Range Quartzite Formation and is met by Aroona Creek. BRA07-AR is situated immediately upstream of the junction of the Brachina trunk channel with the Aroona Creek tributary (Fig. 7). The stratigraphic section consists of a succession of weathered gravel bands intercalated by laminated silt couplets (Fig. 8). The gravel is derived from local shales of the Bunyeroo Formation and interspersed with clasts of tufa from benches that crop out upstream. The fine-grained sediments reflect multiple fluxes of suspension load with vertical fining-upward trends from sand to silt draped by veneers of organic flotsam. At the base, near level with the present creek bed, large pebbles and cobbles of variable lithology were excavated and identified as traction load of the Brachina Creek and dated to 20.2 0.2 ka cal BP (OZK517).
124
According to five ages based on shell and charcoal, the overlying fine sequence aggraded rapidly between 19.5 0.2 ka cal BP (OZK002) and 17.7 0.1 ka cal BP (SSAMS ANU-1812). Repeated age inversions indicate that the material is from reworked sediments. The ~40 m wide section marked by the mottled gravel bands appears to be a cut-and-fill formed within a floodplain that may be older. Less than three kilometres downstream, Brachina Creek is once more confined to a narrow gorge cutting through the Rawnsley Quartzite. Approaching the range front, it makes a sharp turn and floods into an embayment eroded by three minor creeks coming out of the Wonoka Limestone and Siltstone (Fig. 7). A complex terrace remnant of Brachina Silts up to ~18 m high (Williams et al. 2001) is dated close to its downstream termination. Two OSL-ages from the base of BRA07-G put the onset of the fine-grained aggradation to 32.4 1.7 ka (AdGL-08001) and 30.7 2.7 ka (AdGL-08002). Both samples are only partially bleached with a significant inherited De-population dating to 38.9 4.4 ka (AdGL-08002, see Table 1), possibly indicating similarly aged reworked sediments as discussed for sections further upstream. Shells bracketed by the OSL-samples return ages of 34.6 0.4 ka cal BP (ANU BG27) and 28.9 4.4 ka cal BP (ANU BG42). The upper part of the sequence consists of silts intercalating with alluvial fan matrix-supported gravels from a tributary valley. Around 16.1 1.3 ka (AdGL-08003), the gravels spread out across the section.
125
Wilkawillina Silts
ABC Quartzite
Wonoka Formation 3rd Plain Cr
2nd Plain Cr
Moodlatanna Cr Bonney Mt Billy Cr Sandstone
Fig. 9) Wilkawillina Silts stratigraphic sections as seen from the south at an altitude of ~2.66 km projected on the SRTM DEM (elevation exaggeration factor 3) in Google Earth 4.3, with the Second Plain Creek and Mt Billy Creek coming in from the west upstream the Wilkawillina Gorge. Locations of all sections within the Wilkawillina catchment are indicated by red pointers with letters corresponding to section names (see also Fig. 1). The central Flinders Ranges form a large eroded anticline structurally controlled by the Oraparinna Diapir (Callen & Reid 1994) (Fig. 1). As a result, a similar sequence of wide valleys and steep ridges to those described from the Brachina Creek catchment occurs on the eastern side of the Ranges. Here, in mirror image to the Brachina and Etina Creeks, the Second Plain and Moodlatanna Creeks converge in a fine-grained floodplain and are joined by Mt Billy Creek upstream of Wilkawillina Gorge (Fig. 9). Four stratigraphic sections from the Wilkawillina Floodplain are described and dated at similar geomorphological sites with the aim of testing to what extent aggradational and erosional events were synchronous across different catchments. From the distal position within the floodplain, WL08-UFP is from within a bedrock embayment in the immediate vicinity of the present confluence with the Second Plain Creek, and situated less than 200 m upstream of a bedrock constriction cutting
126
through a resistant ridge of the ABC Range Quartzite (Callen & Reid 1994). The depositional setting is in many ways comparable to that of BRA07-SD, and both sections display a number of stratigraphic similarities. Stratigraphic section WL07-FP is a further 1,500 m downstream, and is situated at the downstream end of an interfluve flowing from Mt Billy Creek and joining the trunk channel ahead of a less pronounced ridge of ABC Range Quartzite. The composite section is marked by multiple generations of gravel sheets and lies in a comparable landscape setting (interfluve) to BRA-SA and BRA-SG. Further downstream, the Wilkawillina Creek cuts through the Bonney Sandstone but fails to breach the massive Rawnsley Quartzite. At the gorge entrance, the Third Plain Creek, occupying a valley incised within the Wonoka Formation, joins the trunk channel. Here, in a mirror image to BRA07-AR, a slackwater sequence with bands of mottled gravel is situated at section WL07-S. In contrast, this section occurs opposite the largely eroded confluence in a back-flooded embayment (Fig. 9). Finally, as an analogue to BRA07-G, the remnant of WL07-G lies in a protected embayment within Wilkawillina Gorge.
127
Fig. 10) Wilkawillina Silts stratigraphic sections; main lithostratigraphic units and chronostratigraphy. With more than 12 m of vertical exposure and well-preserved laminations in the lower sequence, WL08-UFP has the potential to complement and extend the high-resolution record of BRA07-SD (Photo 3). Perhaps the most apparent commonality between the two is the presence of distinct brown (10YR 5/3 dry - 4/4 moist) chromatic bands or horizons, displaying transitional lower but clearcut upper boundaries associated with marked colour changes between underlying and overlying units. WL08-UFP exhibits three such palaeosols varying in thickness and maturity, the best-defined at ~750-765 cm (Fig. 10). Shells collected above this chromatic band marking the onset of a more chromatic unit returned an age of 30.0 0.2 ka cal BP (Wk23524), and shells below from the Bcahorizon of 22.9 0 0.2 ka cal BP (Wk23467). A discrete pocket of intact shells from the lower of the closely spaced chromatic bands at ~855-890 cm returned an age of 35.3 0.4 ka cal BP (Wk23525). Disregarding the shells from the Bca-horizon, the ages of the closely spaced chromatic bands
128
compare well with OSL-ages obtained from the twin chromatic band from the Hookina Creek Floodplain. Both chromatic bands are associated with angular clasts and pebbles derived from local outcrops, and increase in thickness and angularity towards the bedrock slope. The fine-grained sediments underlying the chromatic bands are cemented by carbonate, in places forming large in situ calcareous rhizocretions, but less so below the uppermost at ~440-460 cm. This chromatic band marks the boundary between the lower more chromatic unit and an overlying grey unit, and, in terms of colour association and depth below the top, resembles the palaeosol of BRA07-SD (Williams et al. 2001). The minimum age for the end of the fine-grained aggradational regime on the distal floodplain comes from the age of 15.6 0.1 ka cal BP (Wk23465) from the matrix-supported gravel sheets overlying the Wilkawillina Silts. A chromatic surface drape (5YR 5/6 dry - 4/6 moist) with embedded Aboriginal artefacts overlies the sequence. Large intact snail shells sampled from it return an early Holocene age of ~8.2 0.1 ka cal BP (Wk23464). The ~15 m thick composite section of WL07-FP occurs at the downstream end of the largest interfluve of the Wilkawillina Floodplain. Its base rests on gravels and cobbles embedded in a matrix of fluvial sands (Fig. 10). The lowermost fine-grained sediments contain multiple hardpans of carbonate and return an OSL-age of 46.7 4.6 ka (AdGL-08007), followed by 45.5 3.9 ka (AdGL07006) with a dominant residual signal of 72.1 5.9 ka from the overlying lowermost Bca-horizon. Above, a sequence of pristine light-coloured silts is marked by sheets of gravel that terminate a few metres downstream. Further upstream, multiple generations of clast-supported gravel-filled chutes testify to persistent flow directions from the NE across the floodplain, and to continuous aggradation of the silts. Over two intervals, the sheets of gravel spread out. The termination of the lower gravel sheet is associated with charcoal dated to 41.7 0.4 ka cal BP (OZK010). Subsequently, fine-grained sediments aggraded between 38.9 2.9 ka (AdGL-07007) and 36.7 0.9 ka cal BP (OZK013), (excluding an unlikely young shell age). Interestingly, the OSL-age here comprises a small, but wellbleached, population of grains indicating an age of 22.7 1.4 ka (Table 2 & 1). The most likely explanation is that at this time, when no more gravel was transported across the interfluve, the trunk channel incised to this depth. The second generation of gravel sheets is covered by >5 m of exclusively fine-grained sediments preserved in the form of insular terrace remnants. These consist of grey silts, layered in appearance and displaying continuous light bands of silt and fine sand. Two pieces of charcoal collected from the base of this unit returned ages of 29.3 0.4 ka cal BP (SSAMS ANU-2033) and 27.5 0.4 ka cal BP (SSAMS ANU-2032). Above, a close sequence of horizontally laminated silt-and fine-sand bands, hosting many intact gastropods, was dated to 22.4 0.1 ka cal BP (SSAMS ANU-1813). A pronounced intercalated red light band returned an OSL-age of 24.9 1.4 ka (AdGL-07008). The termination in a similar red light silt-and fine-sand band from which deep
129
desiccation cracks extend was dated by two independent OSL-samples; one based on single grains (CLW-95/4) and the other on small aliquots (AdGL-07009). The determined ages of 17.1 1.6 ka and 17.4 1.7 ka are identical within error ranges. The same applies to an inherited early LGM age (however of significant proportion only in AdGL-07009), suggesting reworking of upstream sediments. Section WL07-S is banked against the steep ridge of the Bonney Sandstone, at the downstream termination of the Wilkawillina Floodplain where the channel is constricted to less than ~30 m. It is situated opposite the largely eroded confluence of the trunk channel with its final major tributary within an embayment filled by many metres of grey silts. From the level of the present channel, closely spaced couplets of mottled gravel and laminated fine-grained sediments alternate (Fig. 10), comparable in appearance to those of BRA07-AR (Fig. 8). The age of a basal veneer of organic detritus is 29.4 0.3 ka calBP (SSAMS ANU-2229), followed by 25.4 0.2 ka calBP (SSAMS ANU2227). A piece of charcoal extracted from the uppermost preserved veneer returned an age of 33.9 0.4 ka calBP (SSAMS ANU-2001), likely inherited. Within a protected embayment around the first bend within Wilkawillina Gorge lies the layered terrace remnant WL07-G (Photo 4). The sequence was excavated down to bedrock that in places is capped by thick carbonate (tufa) benches incorporating metre-sized boulders. The lower ~4 m of the section consists of ~25 % of thin bands of reworked fine-grained sediments. These alternate with matrix-supported bands of gravel, in places incorporating large slabs of limestone that slid down from the adjacent slope into the lower overall more chromatic unit (Fig. 10). Imbrication patterns indicate flow directions to the SW, similar to the present day system. The single-grain OSL analysis for the base of the Silt aggradation (CLW-95/1) returned two dominant De-populations (Table 2 & 1): the first aged at 43.4 2.6 ka; and the second at 24.0 2.0 ka. The lower unit terminates in a brown chromatic band (10YR 5/2 dry - 4/2 moist) upon which rests a lighter greyish and distinctly layered unit. This upper unit contains small lenses of fine clast-supported gravel, in places displaying herringbone cross-stratification indicating flow in and out of the embayment. The base of this unit was dated on two separate pieces of charcoal to 23.7 0.2 ka calBP (OZK019, OZK020), an age confirmed by a single-grain OSL-sample of 24.0 1.2 ka (CLW-95/2). The OSL-sample was taken from a more chromatic (2.5Y 6/4 dry - 4/4 moist) unit marked by rolled detrital calcareous nodules likely derived from redistributed slope mantles from the embayment, an inference consistent with a significant inherited age of 32.6 1.8 ka. The more massive fine-grained sediments give way to a unit of layered silts topped by thick veneers of organic detritus, the base of which is single-grain OSL-
130
dated to 18.3 1.4 ka (CLW-95/3), followed by the topmost discrete light band deposited at 17.1 1.4 ka (CLW-95/5).
Discussion of the alluvial record from the Flinders Ranges

The onset of the aggradation of proximal dust-derived alluvium in the Flinders Ranges is recorded from the western piedmont plain from stratigraphic sections less than 50 km downwind of a major potential source of the material: the Lake Torrens playa. At this stage, the low resolution of OSLdates only allows us to state that between the end of the last Interglacial optimum and ~83 ka a ~4 m thick sequence of fine-grained sediments was widely deposited over the Hookina Creek Floodplain. Towards the end of this interval, the upstream section (HK07-D) records multiple pronounced Bca-horizons associated with chromatic bands. At the same time, the downstream section (HK07-M) records fluvial activity and reworking of dunes and calcareous palaeosols. As for the nature of the chromatic bands, also described as palaeosols (Williams 1973; Williams 1982; Williams et al. 2001), and in the case of the Namib Silts interpreted as cambic horizons (Brunotte et al. 2009), the chromatic surface drapes that mantle most of the present floodplain surfaces may provide a modern analogue. A number of characteristics clearly distinguish both from the underlying silts: their red colour characteristic of oxidation; association with small pebbles; and, their degree of aggregation (subplasticity). Typically, but not in all cases, they overlie Bca-horizons, which on first impression suggests that they represent the equivalent A-horizon. However, observations on the formation of the modern analogue indicate a different origin. Local land owners describe how after heavy rainfall the surface levels itself out and mounds and trenches are reduced to a common surface. As for the source of the material, after major flood events such as the most recent one in February 2007, large parts of the trunk channel were draped by silts entrained from the slumping banks. With desiccation, these may have been rapidly entrained by gusts of wind and recycled in the form of aggregated dust over the floodplain. Support for this hypothesis comes from ruins of buildings, such as the abandoned Hookina Hotel. Wherever the roof is missing, the interior is covered by a similar thin chromatic sediment drape. The associated gravel is surface lag left by overbank floods and often concentrated linearly by runoff in the waning stage. Hence, such chromatic bands or recycled silts reflect intervals characteristic of both surface stability and fluvial erosion. When buried, these principally aeolian sheets make excellent marker horizons. Within the Ranges, the onset of the aggradation of the Flinders Silts is dated to ~47 ka (WL07-FP). However, inherited ages indicate that the fine-grained sediments are possibly reworked material last deposited ~70 ka. Between ~47 ka and ~37 ka, close to 7-8 m of light-coloured fine-grained sediments intercalated with gravel sheets were deposited within the Wilkawillina catchment.
131
The next youngest succession of chromatic bands is closely-spaced and dated to ~36 ka and ~30 ka, respectively (HK07-L and WL08-UFP). Other stratigraphic sections that span this interval record fluvial activity reflected in cross-stratified sands (HK07-D) and multiple extensive gravel sheets and chutes (WL07-FP). Where these aeolian accessions mantled bedrock slopes, in many cases they subsequently became incorporated as slopewash in the valley-fills. The chromatic lower unit of BRA07-SD with sheets of carbonate is likely to reflect this interval. At section WL07-FP, this took place between ~29 ka and ~25 ka. The 6 m thick channel fill of stratigraphic section CAS06-1 in the upper Parachilna Creek catchment (Fig. 1) records a similar rapid depositional regime, the onset of which is dated to 29.4 2.1 ka (AdGL07001, Table 1). Intercalated and often matrix-supported sheets of gravel within the gorges, recorded in WL07-S and BRA07-G, point to fluvial redistribution of sediment in the lead-up to the LGM. The brown chromatic band (palaeosol) in BRA07-SD, the ~440-460 cm chromatic band in WL08-UFP (undated), and the twin bands of WL07-FP suggest an interval of surface stability prior to ~24 ka. This was followed regionally by rapid aggradation of finegrained layered and laminated unweathered sediments. According to the ~4 m succession of laminated slackwater deposits in BRA07-SD, this occurred during the course of at least 12 highmagnitude and numerous smaller floods, most pronounced between ~24 ka and ~21 ka and continuing until ~18 ka. To what extent these choked the gorges remains to be established by higher resolution dating of WL07-G. Over the Hookina Floodplain, ~2 m aggraded during the LGM proper (HK07-L), and large channels were rapidly filled with silts (HK07-M). According to results from BRA07-AR and BRA-LW, subsequent incision within the gorges is likely to have reached near present level at ~20 ka. Headward erosion during the peak of the LGM is also reflected further upstream by gravel chutes and sheets (BRA-SA) incising into and spreading across the floodplain, possibly indicating sediment starvation as its cause. This interval gave way to rapid aggradation of the trunk channel raising the base level for tributaries recorded in slackwater deposition from both sides of the Ranges (BRA07-AR, BRA-LW and WL07-G). All floodplain sections document near synchronous final aggradation of silts between ~19 ka and ~17 ka. Within the gorges and distal parts of the floodplain, aggradation of a fast decreasing fine component and increasing gravel influx persisted until ~16 ka (BRA07-G, WL07-G, WL08-UFP). After that, the termination of this fine-grained aggradational regime by incision was final.
Conclusions
The alluvial record from the Flinders Ranges presents a largely continuous palaeo-environmental record from semi-arid southern Australia for the period between the last interglacial and the Deglacial. Based on 124 ages (94 AMS and 30 OSL) from different catchments, regional synchronicity
132
of periods of aggradation, surface stability and incision can be inferred, reflecting major changes of the palaeo-environment. The current chronological resolution is highest between the lead-up to the Last Glacial Maximum and the early Deglacial. The LGM presents itself as an interval characterised by a highly variable climate, which is in accordance with other recent studies from Australia (Petherick et al. 2008 a, b, 2009) and New Zealand (Newnham et al. 2007). Periods of relative surface stability reflected in two closely-spaced intervals of pedogenic activity between ~36-30 ka were followed by widespread erosion and reworking. Another shorter-lived interval of stability prior to ~24 ka gave way to a climate in which large amounts of proximal dust were deposited across the catchments. These loess mantles were redistributed episodically and transported downstream by at least a dozen high-magnitude and numerous smaller floods. The peak of the LGM at ~20 ka was marked locally by rapid incision within the gorges and gravel alluviation on the floodplains. Fine-grained aggradation resumed (or continued) at a slower rate after this. The termination of this aggradational regime ~1816 ka was final and culminated in episodic incision. The Late Pleistocene floodplain remnants are today deeply entrenched and in process of being laterally eroded by ephemeral traction load streams. Future work analysing independent palaeo-environmental proxies collected from the stratigraphic sections will test and refine this model of late Quaternary landscape evolution. A major unanswered question concerns the extent to which these intervals of aggradation, surface stability and erosion reflect changes in the precipitation regime, the vegetation cover and the sediment supply. At this stage we can only state with confidence that the influx of large quantities of loess must be considered as an important variable. Our approach to dating a variety of buried materials enables us to critically assess their utility as proxies to infer the timing of deposition. Radiocarbon dates on charcoal proved to be problematic proxies in an environment prone to sediment redistribution, given the potential of charcoal to survive multiple episodes of reworking. Overall, the dated charcoal within the catchment reflects times of locally abundant woody vegetation burnt by wildfires. The most reliable approximation of depositional ages came from OSL dating. Partial bleaching during fluvial transport is common, but given the homogenous nature of the Flinders Silts this is not a disadvantage. By applying the Finite Mixture Model (Galbraith 2005), multiple ages can be obtained from many samples relating to inherited, depositional and post-depositional bleaching histories of the sampled quartz grains. The inferred scenario of LGM floods redistributing loess mantles, choking narrow gorges and causing widespread back-flooding that resulted in laminated slackwater deposits and ephemeral wetlands is new for Australia. Abnormal wet features within the generally more arid LGM landscape have so far
133
been answered with the minevaporal model proposed by Galloway (1965), by suggesting that a reduction of evaporation in a colder climate can locally outweigh precipitation losses in the overall water balance. However, there is an increasing realisation that droughts and floods go together (Lancaster 2002), and that semi-arid landscapes are to a great extent shaped by such events during climatic deterioration towards drier and cooler conditions (Zielhofer & Faust 2008). The rapid aggradation of the fine-grained valley-fills in the Flinders Ranges prompts us to ask whether shortlived flood events might not explain the widespread beach ridges of the playa lakes in the flanking basins. Our chronostratigraphic approach could also provide a useful working model for similar finegrained valley-fills outside Australia such as the Namib Silts and the Sinai Silts.
Global Context
This contribution is one of an increasing number from both hemispheres demonstrating that the LGM, as defined by Mix et al. (2001), was far more complex than earlier workers had realised (e.g. Farrera et al. 1999), as were the times immediately before and after it. A further set of issues that has attracted considerable recent attention is the question of whether global climatic changes were driven primarily by changes in North Atlantic ocean circulation and whether or not such changes were synchronous from pole to equator across both Hemispheres. Results from an ice core from Dronning Maud Land, Antarctica (EDML core), at comparable resolution to previously published Greenland ice core records, showed a one-to-one coupling between all Antarctic warm events and Greenland Dansgaard-Oeschger cold events, consistent with the bi-polar sea-saw, perhaps reflecting a reduction in meridional overturning in the Atlantic (EPICA Community Members 2006). A recent compilation of
10
Be exposure-dates for the onset of major retreat in mid-latitude LGM mountain
glaciers gives a mean age of 17.30.5 ka for the Southern Hemisphere and 17.40.5 ka for the Northern Hemisphere, showing that the retreat was synchronous in both hemispheres (Schaefer et al. 2006). The onset of glacier retreat therefore coincided with the onset of postglacial warming revealed in the Antarctic high-resolution EPICA Dome C ice core record but with a cooling trend in the Greenland GISP 2 ice core, where warming did not begin until the onset of the Blling/Allerd (B/A) interstadial event at 14.7 ka. What is now needed to advance palaeoclimatic research in Australia is a series of well-dated proxy records along a set of transects from temperate south to tropical north, with the capacity to provide seasonal temperature and precipitation signals.
Lo ess an d Flo o d s: Table 1) Flinders Silts Age Sheet
Ch ap t er 3.1 (Resu lt s an d Discu ssio n )
134
(Hookina Silts)
135
(Brachina Silts)
136
(Wilkawillina Silts)
137
(Parachilna Silts and other locations)
Table 1) Flinders Silts Age Sheet: all age estimates are presented within stratigraphic order of the sections. Sample material is indicated; Q - quartz (SA small aliquots, SG - single grains, LA -large aliquots), C - charcoal, S - carbonate shells of gastropods, OV - veneers of organic detritus, OC - organic clay, T tufa, E - emu egg shell. All 14C-ages are calibrated using the CalPal-2007Hulu-calibration data set (Weninger & Jris 2008). OSL-ages are calculated using the Finite Mixture Model (FMM) of (Galbraith 2005), interpreting resultant Depopulations in terms of "depositional", "inherited" and "post-depositional" age estimates (if relative proportion >25 %). Where De-values fall into a single population, the Central Dose Model (CDM) is employed (Galbraith 2005). For comparison, results of the Minimum Age Model (MAM) are given (Galbraith 2005).
Sample Codes Irradiation data (& designation) (TSAC and XRS) K1890 (HK07-D5) small aliquots K1889 (HK07-D4) small aliquots K1888 (HK07-D2) small aliquots K1887 (HK07-D1) small aliquots K1892 (HK07-L4) small aliquots K1891 (HK07-L2) small aliquots AdGL-08006 (HK07-M5) small aliquots U (ppm) 1.67 0.40

Environmental Dr FMM/CDM De-pop. OD BIC Radial Plots
138
(H2O %
Gy/ka) (Gy/ relative prop. %)

99.82 5.84 49.60 3.55 63 % 37 % 0.17 4.18 0.12 11.85
2.5 % 2.51 0.14 5.1 % 2.44 0.14 10 % 2.32 0.13 AV % 2.57 0.10 2.5 % 2.64 0.18 3.7 % 2.61 0.17 10 % 2.44 0.16 AV % 2.59 0.17 2.5 % 3.24 0.21 3.4 % 3.27 0.21 10 % 3.02 0.20 AV % 3.21 0.21 2.5 % 3.38 0.14 5.9 % 3.51 0.15 10 % 3.23 0.14 AV % 3.40 0.15 2.5 % 3.49 0.11 7.9 % 3.29 0.11
Th (ppm) 6.95 1.31 K (%) 1.50 0.05
Cosmic Dr 192 19.2 (Gy/a) U (ppm) 0.95 0.49 Th (ppm) 11.19 1.65 K (%) 1.56 0.05
213.43 31.84 76 % 274.03 82.05 24 % -
384.35 32.86 100 % 0 -
287.44 20.29 100 % 0 100.97 4.46 38.86 4.27 108.86 8.16 92 % 8% 90 %
Cosmic Dr 100 10.0 (Gy/a) U (ppm) 2.39 0.27 Th (ppm) 9.08 0.90 K (%)
0.09 1.88
2.204 0.066 10 % 3.22 0.11 AV % 3.34 0.11 2.5 % 3.06 0.16 6.0 % 2.95 0.15 10 % 2.83 0.14 AV % 3.01 0.10 2.5 % 3.55 0.24 7.4 % 3.37 0.23 10 % 3.28 0.22 AV % 3.42 0.23
0.20 11.12
258.20 68.22 10 % 61.61 3.59 28.18 2.27 54 % 33 % 0.15 58.22
180.34 27.57 9 % 9.45 2.04 5%
Cosmic Dr 201 20.1 (Gy/a)
AdGL-08005 (HK07-M3) small aliquots AdGL-08004 (HK07-M1) small aliquots AdGL-08003 (BRA07-G4) small aliquots AdGL-08002 (BRA07-G2) small aliquots AdGL-08001 (BRA07-G 1) small aliquots U (ppm) 2.55 0.30

1.9 % 2.38 0.11 2.5 % 2.36 0.11 10 % 2.18 0.10 AV % 2.34 0.10 2.5 % 2.95 0.19 5.3 % 2.86 0.19 10 % 2.72 0.18 AV % 2.86 0.19 2.3 % 3.33 0.11 2.5 % 3.33 0.11 10 % 3.07 0.10 AV % 3.28 0.11 2.5 % 3.16 0.11 6.6 % 3.02 0.10 10 % 2.91 0.10 AV % 3.05 0.10 2.5 % 3.26 0.13 3.8 % 3.21 0.13 10 % 3.00 0.12 AV % 3.19 0.13 2.5 % 3.36 0.13 9.4 % 3.13 0.12 10 % 3.11 0.12 AV % 3.26 0.09 2.5 % 3.36 0.13 9.7 % 3.12 0.12 10 % 3.11 0.12 AV % 3.26 0.09 2.5 % 3.49 0.09 8.4 % 3.27 0.09 10 % 3.22 0.09 AV % 3.33 0.09 292.64 18.35 90 % 200.15 40.85 10 % 0.20 36.91 0.15 3.23
139
Th (ppm) 6.27 0.98 K (%) 1.32 0.04
309.90 24.76 57 % 65.78 7.43 25 %
162.45 28.33 18 % 52.89 3.85 87.79 14.49 93.55 7.69 74 % 26 % 47 % 0.05 8.64 0.15 12.26
118.45 12.86 43 % 249.64 37.12 10 % 103.07 3.21 68 % 0.05 20.26
160.15 12.78 18 % 68.94 5.56 55.63 5.03 24.12 3.62 76.12 37.35 56.57 5.12 80.25 7.95 83.01 3.93 14 % 83 % 9% 8% 63 % 37 % 91 % 0.10 2.19 0.1 11.85 0.22 82.98
Cosmic Dr 52 5.2 (Gy/a) U (ppm) 2.33 0.33
CLW-95/4
(WL07-FP6-SG) Th (ppm) 8.68 1.09 K (%) single grains AdGL-07009 2.06 0.06
Cosmic Dr 202 20.2/ (Gy/a) 190 19.0 U (ppm) 2.33 0.33
(WL07-FP6-SA) Th (ppm) 8.68 1.09 K (%) small aliquots AdGL-07008 (WL07-FP5) small aliquots 2.06 0.06
Cosmic Dr 202 20.2/ (Gy/a) 190 19.0 U (ppm) 2.59 0.25 Th (ppm) 10.45 0.45 K (%) 2.10 0.05
167.61 30.19 9 % -
Cosmic Dr 189 18.9/ (Gy/a) 160 16.0
AdGL-07007 (WL07-FP3) small aliquots AdGL-07006 (WL07-FP1) small aliquots AdGL-08007 (WL07-FP0) small aliquots CLW-95/5 (WL07-G6) single grains CLW-95/3 (WL07-G5) single grains CLW-95/2 (WL07-G3) single grains CLW-95/1 (WL07-G1) single grains AdGL-07002 (CAS06-1 3) small aliquots U (ppm) 2.21 0.32

2.5 % 3.44 0.13 2.5 % 3.44 0.13 10 % 3.17 0.12 AV % 3.24 0.09 2.5 % 3.04 0.13 3.0 % 3.02 0.13 10 % 2.80 0.12 AV % 3.02 0.05 0.1 % 3.17 0.13 2.5 % 3.07 0.13 10 % 2.83 0.12 AV % 3.07 0.13 2.5 % 3.69 0.14 3.9 % 3.63 0.14 10 % 3.40 0.13 AV % 3.61 0.14 2.5 % 3.47 0.13 4.8 % 3.38 0.12 10 % 3.20 0.12 AV % 3.38 0.12 2.5 % 3.35 0.13 3.0 % 3.33 0.13 10 % 3.09 0.12 AV % 3.30 0.13 2.5 % 3.04 0.16 9.8 % 2.80 0.15 10 % 2.80 0.15 AV % 2.88 0.16 126.05 8.62 73.71 3.93 69 % 31 % 0 3.87 0.05 6.00
140
Th (ppm) 10.42 1.08 K (%) 2.19 0.07
217.69 17.42 77 % 137.53 11.63 23 % -
143.61 12.88 59 % 202.11 21.61 41 % 59.48 4.41 67 %
0.10 7.38
0.15 24.42
132.83 15.75 33 % 99.21 6.51 61.93 4.06 79.03 2.31 107.59 4.21 45.60 3.62 124.91 3.36 69.00 4.38 46.07 4.66 61 % 39 % 64 % 27 % 8% 59 % 32 % 9% 54 % 39 % 0.02 2.40 0 24.98 0.03 15.28 0.15 76.54
2.5 % 3.360 0.137 76.43 1.84 3.3 % 3.330 0.136 99.98 4.38
10 % 3.105 0.126 173.63 28.70 7 % AV % 3.300 0.135 -
AdGL-07001 (CAS06-1 1) small aliquots U (ppm) 1.99 0.34

2.5 % 2.588 0.121 97.08 4.97 3.3 % 2.574 0.120 74.78 4.10 10 % 2.388 0.111 AV % 2.546 0.119 51 % 49 % 0.06 -1.25
141
Th (ppm) 9.33 1.12 K (%) 1.44 0.04
Tab. 2) OSL-data: Environmental Dose Rate data (Dr) for variable water contents and corresponding Equivalent Dose data (De); Discrete De-populations are calculated using the Finite Mixture Model (FMM) fmix.s (Galbraith 2005), and numerically listed by their relative proportion in the sample. The number of De-populations and the over-dispersion factor (OD) is largely determined by the lowest Bayesian information criterion value (BIC). Where De -values fall into one discrete population, the Central Dose Model (CDM) cdose.s (Galbraith 2005) is employed. All aliquots/grains are presented as radial plots, i.e. as points with precision values plotted against the x-axis (increasing towards the right), and De-values plotted against the y-axis (bands indicating 2 STs from a given central value or radial line), (Olley & Reed 2003). The De-population interpreted to relate to the last deposition of the sediment is emphasised in black. De-populations consisting of <25 % are indicated in grey and treated as outliers and not converted into ages.
142
twin chromatic band
Photo Plate 1) Hookina Floodplain looking towards the Range front and stratigraphic section HK07-L.
palaeosol
Photo Plate 2) The layered and laminated slackwater stratigraphic section BRA07-SD.
143
440-460 cm band 750-765 cm band
855-890 cm band
Photo Plate 3) Stratigraphic section WL08-UFP from the distal Wilkawillina Floodplain.
Photo Plate 3) Stratigraphic sectionWL07-G within the first embayment of Wilkawillina Gorge.
144
Acknowledgements
We thank CRC LEME (Cooperative Research Centre for Landscape Environments and Mineral Exploration), the Australian Research Council (ARC grant #DP0559577) and the Australian Institute of Nuclear Science and Engineering (AINSE Grants 96/192R, 99/001 and 07/160) for generous financial support. Thanks also to the Flinders Ranges National Park authorities for research permits and friendly accommodation, station owners Richard Spears and Rex for letting us work on their land, and to Jutta von dem Bussche, Tom Brookman and Lyndon Parham for assisting in the field. In particular, we are indebted to Tim Pietsch for volunteering to run all single grain samples at the luminescence laboratory of CSIRO Land and Water in Canberra, John Prescott for kindly providing the facilities, encouragement and supervision at the Adelaide luminescence laboratory, and Stewart Fallon from the Australian National University Radiocarbon Dating Laboratory for preparing most of the labour-intensive radiocarbon samples based on organic veneers.
References
Aitken, M.J. 1998. An Introduction to Optical Dating, Oxford: University Press. Blong, R.J. & Gillespie, R. 1978. Fluvially transported charcoal gives erroneous 14C ages for recent deposits. Nature, 271(23), 739-741. Bourke, M.C., Child, A. & Stokes, S. 2003. Optical age estimates for hyper-arid fluvial deposits at Homeb, Namibia. Quaternary Science Reviews, 22, 1099-1103. Bowler, J.M. 1978. Quaternary climate and tectonics in the evolution of the Riverine Plain, southeastern Australia. In: Davies, J.L. & Williams, M.A.J. (eds) Landform Evolution in Australasia. Australian National University Press, Canberra, 70-112. Brunotte, E., Maurer, B., Fischer, P., Lomax, J. & Sander, H. 2009. A sequence uvial and aeolian of deposits (desert loess) and palaeosoils covering the last 60 ka in the Opuwo basin Kaokoland/Kunene Region, Namibia) based on luminescence dating. Quaternary International, 196(1-2), 71-85. Callen, R.A. & Reid, P.W. 1994. Geology of the Flinders Ranges National Park. South Australian Geological Survey: Special Map 1:75,000. Callen, R.A., Wasson, R.J. & Gillespie, R. 1983. Reliability of radiocarbon dating of pedogenic carbonate in the Australian arid zone. Sedimentary Geology, 35, 1-14. Cock, B.J., Williams, M.A.J. & Adamson, D.A. 1999. Pleistocene Lake Brachina: a preliminary stratigraphy and chronology of lacustrine sediments from the central Flinders Ranges, South Australia. Australian Journal of Earth Sciences, 46, 61-69.
145
Crocker, R.L. 1946. Post-miocene climatic and geologic history and its significance in relation to the genesis of the major soil types of South Australia. In: Council for Scientific and Industrial Research Bulletin, 5-65. Dart, R.C., Barovich, K.M., Chittleborough, D.J. & Hill, S.M. 2007. Calcium in regolith carbonates of central and southern Australia: Its source and implications for the global carbon cycle. Palaeogeography, Palaeoclimatology, Palaeoecology, 249, 322334. Duller, G.A.T. 2004. Luminescence dating of Quaternary sediments: recent advances. Journal of Quaternary Science, 19(2), 183-192. Duller, G.A.T., Botter-Jensen, L., Murray, A.S. & Truscott, A.J. 1999. Single grain laser luminescence (SGLL) measurements using a novel automated reader. Nuclear Instruments and Methods in Physics Research, 155(4), 506-514. Eitel, B., Blmel, W.D., Hser, K. & Mauz, B. 2001. Dust and loessic alluvial deposits in Northwestern Namibia (Damaraland, Kaokoveld): sedimentology and palaeoclimatic evidence based on luminescence data. Quaternary International, 76, 57-65. Eitel, B., Kadereit, A., Blmel, W.D., Hser, K. & Kromer, B. 2005. The Amspoort Silts, northern Namib desert (Namibia): formation, age and palaeoclimatic evidence of river-end deposits. Geomorphology, 64(3-4), 299-314. EPICA Community members, 2006. One-to-one coupling of glacial climate variability in Greenland and Antarctica. Nature 444, 195-198. Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Kaplan, A., Guilderson, T.P., Fairbanks, T.W., Bloom, A., Grootes, P.M. & Nadeau, M.-J. 2005. Radiocarbon calibration curve spanning 0 to 50,000 years BP based on paired 230Th/234U/238U and 14C dates on pristine corals. Quaternary Science Reviews, 24, 1781-1796. Ferguson, C.W., Huber, B. & Suess, H.E. 1966. Determination of the age of Swiss lake dwellings as an example of dendrochronologically-calibrated radiocarbon dating. Zeitschrift fr Naturforschung Teil A, 21(34), 1173-1177. Farrera, I., Harrison, S.P., Prentice, I.C., Ramstein, G., Guiot, J., Bartlein, P.J., Bonnefille, R., Bush, M., Cramer, W., von Grafenstein, U., Holmgren, K., Hooghiemstra, H., Hope, G., Jolly, D., Lauritzen, S.E., Ono, Y., Pinot, S., Stute, M. & Yu, G. 1999. Tropical climates of the Last Glacial Maximum: a new synthesis of terrestrial palaeoclimatic data. I. Vegetation, lake levels and geochemistry. Climate Dynamics, 15, 823-856. Fontes, J.-C. & Gasse, F. 1989. On the ages of humid Holocene and Late Pleistocene phases in North Africa - remarks on "Late Quaternary climatic reconstruction for the Maghreb (North Africa)" by P. Rognon. Palaeogeography, Palaeoclimatology, Palaeoecology, 70, 393-398. Galbraith, R.F. 2005. Statistics for Fission Track Analysis, Boca Raton: Chapman & Hall/CRC. Galbraith, R.F. 1990. The radial plot: graphical assessment of spread in ages. Nuclear Tracks and Radiation Measurements, 17(3), 207-214.
146
Galbraith, R.F. & Green, P.F. 1990. Estimating the component ages in a finite mixture. Nuclear Tracks and Radiation Measurements, 17(3), 197-206. Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H. & Olley, J.M. 1999. Optical dating and multiple grains of quartz from Jinmium rock shelter, northern Australia: part 1, experimental design and statistical models. Archaeometry, 41(2), 339-364. Galloway, R.W. 1965. Late Quaternary climates in Australia. Journal of Geology, 73(4), 603618. Gentilli, J. 1986. Climate. In: D.N. Jeans ed, Australia - A Geography. The Natural Environment. Sydney University Press, Sydney, 1448. Gingele, F.X. & De Deckker, P. 2005. Late Quaternary fluctuations of palaeoproductivity in the Murray Canyons area, South Australian continental margin. Palaeogeography, Palaeoclimatology, Palaeoecology, 220(3-4), 361-373. Glasby, P., Williams, M.A.J., McKirdy, D., Symonds, R. & Chivas, A.R. 2007. Late Pleistocene environments in the Flinders Ranges, Australia: preliminary evidence from microfossils and stable isotopes. Quaternary Australasia, 24(2), 19-28. Goossens, D. 1988. The effect of surface curvature on the deposition of loess: a physical model. Catena 15(2), 179-194. Goudie, A.S. 1972. Climate, weathering, crust formation, dunes and fluvial features of the Central Namib Desert, near Gobabeb, South West Africa. Madoqua II, 1, 1531. Haberlah, D. 2006. Depositional models of late Pleistocene fine-grained valley-fill formations in the Flinders Ranges, SA. In: R. W. Fitzpatrick & P. Shand (eds) Regolith 2006 - Consolidation and dispersion of ideas. Western Australia: Cooperative Research Centre for Landscape Environments and Mineral Exploration, 122-126. Haberlah, D. 2007. A call for Australian loess. Area, 39(2), 224-229. Haberlah, D., Williams, M.A.J., Hill, S.M., Halverson, G. & Glasby, P. 2007. A terminal Last Glacial Maximum (LGM) loess-derived palaeoflood record from South Australia? Quaternary International, 167-168, 150. Haberlah, D. 2008. Response to Smalley's discussion of 'A call for Australian loess'. Area, 40(1), 135136. Heine, K. & Heine, J.T. 2002. A paleohydrologic reinterpretation of the Homeb Silts, Kuiseb River, central Namib Desert (Namibia) and paleoclimatic implications. Catena, 48(1-2), 107-130. Hesse, P.P. & McTainsh G.H. 2003. Australian dust deposits: modern processes and the Quaternary record. Quaternary Science Reviews, 22, 2007-2035. Hoffman, D.L., Beck, J.W., Richards, D.A., Smart, P.L., Mattey, D.P., Paterson, B.A. & Hawkesworth, C.J. 2008. Atmospheric radiocarbon variation between 44 and 28 ka based on U-series dated speleothem. Geophysical Research Abstracts, 10.
147
Hua, Q., Jacobsen, G.E., Zoppi, U., Lawson, E.M., Williams, A.A., Smith, A.M. & McGann, M.J. 2001. Progress in radiocarbon target preparation at the ANTARES AMS centre. Radiocarbon, 43(2A), 275282. Hughen, K., Southon, J., Lehman, S., Bertrand, C. & Turnbull, J. 2006. Marine-derived 14C calibration and activity record for the past 50,000 years updated from the Cariaco Basin. Quaternary Science Reviews, 25, 3216-3227. Huntley, D.J., Godfrey-Smith, D.I. & Thewalt, M.L.W. 1985. Optical dating of sediments. Nature, 313, 105-107. Huntley, D.J., Hutton, J.T. & Prescott, J.R. 1993. The stranded beach-dune sequence of south-east Australia: a test of thermoluminescence dating, 0-800 ka. Quaternary Science Reviews, 12, 120. Issar, A.S. & Bruins, H.J. 1983. Special climatological conditions in the deserts of Sinai and the Negev during the latest Pleistocene. Palaeogeography, Palaeoclimatology, Palaeoecology, 43(1/2), 63-72. Issar, A.S. & Eckstein, Y. 1969. The lacustrine beds of Wadi Feiran, Sinai: their origin and significance. Israel Journal of Earth Sciences, 18(1), 29-32. Jacobs, Z., Duller, G.A.T. & Wintle, A.G. 2006. Interpretation of single grain De distributions and calculation of De. Radiation Measurements, 41, 264-277. Jennings, J.N. 1968. A revised map of the desert dunes of Australia. Australian Geographer, 10(5), 408-409. Kemp, J. & Spooner, N.A. 2007. Evidence for regionally wet conditions before the LGM in southeast Australia: OSL ages from a large palaeochannel in the Lachlan Valley. Journal of Quaternary Science, 22(5), 423-427. Lancaster, N. 2002. How dry was dry?Late Pleistocene palaeoclimates in the Namib Desert. Quaternary Science Reviews, 21(7), 769-782. Leopold, M., Vlkel, J. & Heine, K. 2006. A ground penetrating radar survey of late Holocene fluvial sediments in northwest Namibian river valleys: characterisation and comparison. Journal of the Geological Society, 163, 923-936. Li, B., Li, S. & Wintle, A.G. 2008. Overcoming environmental dose rate changes in luminescence dating of waterlain deposits. Geochronometria, 30, 33-40. Mix, A.C., Bard, E. & Schneider, R. 2001. Environmental processes of the ice age: land, oceans, glaciers (EPILOG). Quaternary Science Reviews, 20, 627-657. Mackellar, D. 1971. The poems of Dorothea Mackellar. Adelaide, Australia: Rigby. Murray, A.S. & Wintle, A.G. 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements, 32, 57-73.
148
Nanson, G.C., Price, D.M., Jones, B.G., Maroulis, J.C., Coleman, M., Bowman, H., Cohen, T.J., Pietsch, T.J. & Larsen J.R. 2008. Alluvial evidence for major climate and flow regime changes during the middle and late Quaternary in eastern central Australia. Geomorphology, 101, 109-129. Newnham, R.M., Lowe, D.J., Giles, T. & Alloway, B.V. 2007. Vegetation and climate of Auckland, New Zealand, since ca. 32 000 cal. yr ago: support for an extended LGM. Journal of Quaternary Science, 22(5), 517-534. Olley, J.M., Caitcheon, G. & Murray, A.S. 1998. The distribution of apparent dose as determined by optically stimulated luminescence in small aliquots of fluvial quartz: implications for dating young sediments. Quaternary Geochronology, 17, 1033-1040. Olley, J.M. & Reed, M. 2003. Radial Plot 1.3, CSIRO Land & Water. Petherick, L.M., McGowan, H.A., Moss, P.T. & Kamber, B.S. 2008 a. Late Quaternary aridity and dust transport pathways in eastern Australia. Quaternary Australasia, 25(1), 2-11. Petherick, L.M., McGowan, H.A. & Moss, P. 2008 b. Climate variability during the Last Glacial Maximum in eastern Australia: evidence of two stadials? Journal of Quaternary Science, 23, 787-802. Petherick, L.M., McGowan, H.A. & Kamber, B.S. 2009. Reconstructing transport pathways for late Quaternary dust from eastern Australia using the composition of trace elements of long traveled dusts. Geomorphology, 105(1-2), 67-79. Preiss, W.V. 1987. The Adelaide Geosyncline - late Proterozoic stratigraphy, sedimentation, palaeontology and tectonics. Geological Survey of South Australia, Bulletin 53, 438. Preiss, W.V. 1999. Parachilna, South Australia. 1: 250 000 Geological Series Explanatory Notes. Prescott, J.R. & Hutton, J.T. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long-term time variations. Radiation Measurements, 23, 497-500. Quigley, M.C., Sandiford, M. & Cupper, M.L. 2007. Distinguishing tectonic from climatic controls on range-front sedimentation. Basin Research, 19(4), 491-505. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C.J.H., Blackwell, P.G., et al.2004. IntCal04 atmospheric radiocarbon age calibration, 26-0 cal kyr BP. Radiocarbon, 46, 1029-1058. Revel-Rolland, M., De Deckker, P., Delmonte, B., Hesse, P.P., Magee, J.W., Basile-Doelsch, I., Grousset, F. & Bosch, D. 2006. Eastern Australia: a possible source of dust in East Antarctica interglacial ice. Earth and Planetary Science Letters, 249(1-2), 1-13. Rgner, K., Knabe, K., Roscher, B., Smykatz-Kloss, W. & Zller, L. 2004. Alluvial loess in the Central Sinai: occurrence, origin, and palaeoclimatological consideration. In: Paleoecology of Quaternary Drylands. Lecture Notes in Earth Sciences, 102. 79-99.
149
Schaefer, J.M., Denton, G.H., Barrell, D.J.A., Ivy-Ochs, S., Kubik, P.W., Andersen, B.G., Phillips, F.M., Lowell, T.V., & Schlchter, C. 2006. Near-synchronous interhemispheric termination of the Last Glacial Maximum in mid-latitudes. Science, 312, 1510-1513. Schwerdtfeger, P. & Curran, E. 1996. Climate of the Flinders Ranges. In: M. Davies, C. R. Twidale, & M. J. Tyler (eds) Natural history of the Flinders Ranges. Royal Society of South Australia Occasional Publication. Adelaide, Australia: Royal Society of South Australia, 63-75. Shulmeister, J., Goodwin, I., Renwick, J., Harle, K., Armand, L., McGlone, M.S., Cook, E., Dodson, J., Hesse, P.P., Mayewski, P. & Curran, M. 2004. The Southern Hemisphere westerlies in the Australasian sector over the last glacial cycle: a synthesis. Quaternary International, 118-119, 23-53. Smith, R.M.H., Mason, T.R. & Ward, J.D. 1993. Flash-flood sediments and ichnofacies of the Late Pleistocene Homeb Silts, Kuiseb River, Namibia. Sedimentary Geology, 85, 579-599. Sprigg, R.C. 1979. Stranded and submerged sea-beach systems of southeast South Australia and the aeolian desert cycle. Sedimentary Geology, 22, 53-96. Srivastava, P., Brook, G.A., Marais, E., Morthekai, P. & Singhvi, A.K. 2006. Depositional environment and OSL chronology of the Homeb silt deposits, Kuiseb River, Namibia. Quaternary Research, 65(3), 478-491. Talbot, M.R., Holm, K. & Williams, M.A.J. (1994). Sedimentation in low gradient desert margin systems: a comparison of the late Triassic of north-west Somerset (England) and the late Quaternary of east-central Australia. Geological Society of America Special Paper 289, 97-117. Tuniz, C., Bird, J.R., Herzog, G.F. & Fink, D. 1998. Accelerator Mass Spectrometry: Ultrasensitive analysis for global science 1st ed., CRC Press, Boca Raton, Florida. Vogel, J.C. 1982. The age of the Kuiseb River silt terrace at Homeb. Palaeoecology of Africa, 15, 201209. Walker, M. 2005. Quaternary dating methods: an introduction, John Wiley & Sons. Ward, J.D. 1987. The Cenozoic succession in the Kuiseb Valley, central Namib Desert. Geological Survey of South West Africa/ Namibia Memoir, 9, 1-124. Weninger, B. & Jris, O. 2008. A 14C age calibration curve for the last 60 ka: the Greenland-Hulu U/Th timescale and its impact on understanding the Middle to Upper Paleolithic transition in Western Eurasia. Journal of Human Evolution, 55(5), 772-781. Weninger, B., Jris, O. & Danzeglocke, U. 2008. CalPal-2007. Cologne Radiocarbon Calibration & Palaeoclimate Research Package. Available at: http://www.calpal.de/ [Accessed July 28, 2008]. Williams, D.L.G. 1982. Late Pleistocene vertebrates and palaeo-environments of the Flinders and Mount Lofty Ranges. Unpublished PhD thesis. Flinders University of South Australia.
150
Williams, G.E. 1973. Late Quaternary piedmont sedimentation, soil formation and paleoclimates in arid South Australia. Zeitschrift fr Geomorphologie N.F., 17, 102-125. Williams, G.E. & Polach, H.A. 1971. Radiocarbon dating of arid-zone calcareous paleosols. Geological Society of America Bulletin, 82, 3069-3086. Williams, M.A.J. & Nitschke, N. 2005. Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. South Australian Geographical Journal, 104, 25-36. Williams, M.A.J., De Deckker, P., Adamson, D.A. & Talbot, M.R. 1991. Episodic fluviatile, lacustrine and aeolian sedimentation in a late Quaternary desert margin system, central western New South Wales. In: Williams, M.A.J., De Deckker, P. & Kershaw, A.P. (eds) The Cainozoic in Australia: A re-appraisal of the evidence. Geological Society of Australia Special Publication 18, 258-287. Williams, M.A.J., Prescott, J.R., Chappell, J., Adamson, D., Cock, B., Walker, K. & Gell, P. 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International, 83-85, 129-144. Williams, M.A.J., Nitschke, N. & Chor, C. 2006. Complex geomorphic response to late Pleistocene climatic changes in the arid Flinders Ranges of South Australia. Gomorphologie: relief, processus, environnement, 4, 249-258. Wintle, A.G. & Murray, A.S. 2006. A review of quartz optically stimulated luminescence characteristics and their relevance in single-aliquot regeneration dating protocols. Radiation Measurements, 41, 369-391. Zielhofer, C. & Faust, D. 2008. Mid- and Late Holocene fluvial chronology of Tunisia. Quaternary Science Reviews, 27(5-6), 580-588.
Loess and floods: high-resolution multiproxy data of Last Glacial Maximum (LGM) slackwater deposition in the Flinders Ranges, semi-arid South Australia
Lo ess an d Flo o d s: Ch ap t er 3.2 (Result s an d Discu ssio n ) (similar version in submission with Quaternary Science Reviews) David Haberlah
1, 2
151
, Martin A.J. Williams 3, Galen Halverson 1, Steven M. Hill
1, 2
, Tomas Hrstka
4, 5
Alan R. Butcher 6, Grant H. McTainsh 7, Peter Glasby 3

1
Geology & Geophysics, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia (david.haberlah@adelaide.edu.au, +61 (0)8 8303-8022) Cooperative Research Centre for Landscape Environments and Mineral Exploration
Geographical & Environmental Studies, School of Social Sciences, University of Adelaide, Adelaide, SA 5005, Australia SGS Minerals Services, QLD 4064, Australia
Institute of Geology, Czech Academy of Science, Rozvojova 269, 165 02 Prague 6-Lysolaje, Czech Republic
6
FEI Australia, Brisbane, QLD 4064, Australia Atmospheric Environment Research Centre, Griffith University, Brisbane, QLD 4111, Australia
152
Abstract
Terrace remnants of late Pleistocene fine-grained valley-fill formations (Silts) deeply entrenched by ephemeral traction load streams in arid areas remain a puzzle. They have been attributed to a variety of origins ranging from lacustrine to alluvial floodplains. We here report a centimetre-scale multi-proxy study of a 7 m section of similar Silts in the semi-arid Flinders Ranges of South Australia, which span the lead-up to and peak of the Last Glacial Maximum. The results of detailed lithostratigraphic mapping, high-resolution parametric particle-size analysis, quantitative spectral mineralogy, magnetic susceptibility, carbon stable isotope geochemistry, and a chronostratigraphy based on 27 AMS radiocarbon and 6 luminescence ages are discussed in terms of sediment provenance, depositional environment, weathering and local hydrology with the aim of reconstructing the regional hydroclimatic history. The data are consistent with inferences of a fluctuating aeolian-fluvial interplay dominating the extended LGM environment with a greater impact on the landscape than all geomorphic processes since then. Accordingly, proximal dust accessions (loess mantles) were eroded and entrained by numerous small and at least a dozen largescale flood events and trapped in an intra-montane floodplain extending into Brachina Gorge. Upstream of this narrow constriction, recurrent backflooding is discussed resulting in a thick sequence of slackwater couplets. Aggradation and degradation of the valley-fills appear to be largely controlled by sediment supply from the valley slopes, possibly replenished by aeolian dust accessions from upwind deflated terminal playa lakes and dunefields. In conclusion, this study demonstrates how dust storms and flooding rains can account for pluvial features previously explained by the opposing effects of reduced precipitation and evaporation in the colder more arid glacial landscape of southern Australia.
153
Introduction
The contrast between gently sloping, deeply entrenched fine-grained valley-fills and rugged chains of weathering-resistant ridges that inspired the painter Sir Hans Heysen (1877-1968) in the Flinders Ranges of South Australia have prompted more recent scientific interest. Cock et al. (1999) first described a laminated sequence of silts within Brachina Gorge in the central Flinders Ranges as spanning the Last Glacial Maximum (LGM) (Fig. 1; Plate 1). The LGM (24-18 ka: EPILOG LGM chronozone Level 2; Mix et al., 2001) is elsewhere in south-eastern Australia marked by widespread erosion, deflation of lake beds and salt pans (playas), and actively migrating dunefields (e.g. Fitzsimmons et al., 2007). Yet, here is a continuous more than 7 m thick aggradational sequence of well-defined centimetre-scale laminations, extending horizontally for hundreds of metres and, judging by its fine-grained texture and redoximorphic colours, suggestive of lacustrine deposition (Cock et al., 1999). A year later, a theodolite survey established that the surface of the Brachina Silts is inclined westwards with a mean gradient of 1:87, parallel to the present-day thalweg and older rock-cut benches (Williams et al., 2001). The spatially restricted, layered to laminated facies is embedded in a fluvial massive tabular facies that extends into Brachina Gorge where it is preserved as terraces up to 18 m high. Williams et al. (2001) discuss the enigma of a late Pleistocene wetland and concluded that an aggraded surface with this slope was incompatible with sedimentation in a shallow lake (op. cit., p.133). They concluded that the clays and fine sands accumulated in a fluvial wetland, essentially unchannelled in the fine-textured reaches but entered at its margins by channels carrying coarser sediment from tributary streams and fans (op. cit., p.133). Three hypotheses were proposed addressing provenance, depositional nature and the demise of the valley-fills. 1) The fine-grained material, in stark contrast to todays gravel- and sand-dominated stream bed, entered the catchment as proximal dust. 2) The aeolian accessions were eroded by gentle winter rains, trapped by swamp vegetation and aggraded in the form of a mainly channel-free fluvial wetland. 3) Terminal incision, heralded by an influx of coarse alluvium, was triggered by storm-driven floods likened to the monsoonal incursions that at present inundate floodplains in the region at decadal to centennial frequency (McCarthy et al., 2006). The first hypothesis was supported by: a) aggradation rates far in excess of long-term catchment erosion rates as inferred from in situ cosmogenic
10
Be-measurements (Williams et al., 2001); and, b) by a comparative
geochemical study of loess patches from the slopes, valley-fill remnants and the bedrock geology (Williams and Nitschke, 2005). The second hypothesis complies with the minevaporal theory of Galloway (1965), often invoked in Australia to explain pluvial anomalies such as high lake beach ridges dated to the glacial interval otherwise characterised by peak aridity. It was since confirmed as
154
a potential catchment scenario by modelling the impact of lower glacial temperatures and the retreat of the dominant riparian tree river red gum (Eucalyptus camaldulensis) on the local hydrology by: a) lowering the cloud base, thereby increasing orographically enhanced drizzling winter rainfall; and, b) by significantly reducing evapo-transpiration and raising local water tables (Williams et al., 2006). However, this low-energy depositional scenario contrasts with most recent interpretations of similar LGM deposits of loess-derived alluvium in Namibia (Eitel et al., 2001) and the Sinai Peninsula (Rgner et al., 2004), suggested to reflect more frequent and intense run-off events (Lancaster, 2002; Heine and Heine, 2002; Srivastava et al., 2005). Furthermore, the very stratigraphic type section that prompted the initial scientific interest (Cock et al., 1999) remained at odds with a wetland scenario, in which swamp vegetation and other forms of bioturbation would have destroyed the laminations. This study addresses the layered to laminated aggradational sequence by an integrated lithostratigraphic, chronostratigraphic, geochemical and geophysical approach, aiming to establish the provenance of the material, its depositional, weathering and hydrological history, changes in the vegetation cover of the catchment and the regional climate throughout the LGM, which according to several recent reviews was far more complex than previously assumed (Gasse et al., 2008; Williams et al., 2009).
155
Fig. 1) Study area: location of the layered to laminated stratigraphic section BRA-SD within the Brachina catchment as outlined in yellow. Major creeks and water bodies draining the Flinders Ranges to the west towards terminal playa Lake Torrens, and to the east towards terminal playa Lake Frome are projected in blue. The inset figure places the study area on the Australian continent (SRTM DEM) and in context of present-day seasonality (adapted from Gentilli, 1986).
Stratigraphic type section BRA-SD

The main body of the Brachina Silts spreads for ~2 500 m from east to west across the middle reaches of the Brachina valley, increasing in thickness and converging upon the steep weatheringresistant range front of the ABC Quartzite (Cock et al., 1999) (Fig. 1 & 2). The layered to laminated
156
facies is spatially restricted to the Brachina-Etina confluence upstream of the narrow Brachina Gorge that cuts through the quartzite ridge. The stratigraphic type section BRA-SD describes the distal reaches of a ~235 m long semi-circular vertical cliff face more than 7 m thick consisting of alternating light and dark layers and laminations banked against the steep vegetated valley slope (Photo 1). The sediments are undercut and eroded by the present course of Brachina Creek except where unconformably resting upon jagged bedrock of Neoproterozoic shales that underlie the Silts (Photo 2A, Fig. 2) (Preiss, 1987).
Lithostratigraphy
The stratigraphic section includes five lithostratigraphic units (Photo 1; appendix 5.3A). From the base up these consist of: 1) Basal unit (I) 700-506 cm (below top): Overall red-brown sequence of decimetre-thick alternating lighter and darker bands with transitional boundaries terminating in a well-defined dark brown/grey (5YR 6/1
dry
/ 7.5YR 5/1
moist)
band of blocky to prismatic structure and
slickensides at 506 cm, previously interpreted as a palaeosol (Cock et al., 1999; Williams et al., 2001). The thickness of the transitional bands decreases towards the top of the unit. In contrast to the exclusively fine-grained units above, multiple sheets of rolled detrital calcareous nodules (transported pedogenic carbonate sourced from reworked Bca-horizons) (Photo 2B), large (>5 cm) pieces of charcoal (Photo 2C), and a single line of well-rounded cobbles and tufa clasts (Photo 2D) are incorporated. In rare instances, current cross-bedding is faintly preserved. At the base, vertical in situ calcareous rhizomorphs are exposed (Photo 2E). A few large pseudogleyic root casts extend from the overlying sediments into the basal unit (Photo 2F). 2) Transitional unit (II) 506-470 cm: Onset of laminated aggradation bracketed by a light yellowish brown yellow band at 500 cm and a light reddish brown red band at 470 cm (Photo 3A). The sequence differs from the laminated unit (III) above in its overall more oxidised appearance reflecting that of the basal unit (I) (Photo 1), and in that the yellow and red bands thicken towards the upstream valley slope. 3) Laminated unit (III) 470-90 cm: Alternating light and dark laminations including ten ~10 cm thick yellow bands that stretch across the width of the section and extend towards the present-day Brachina-Etina confluence, increasing in thicknesses to >20 cm. Where the laminations are disturbed, the sharp boundaries grade into thicker pale and dark bands. 4) Pedogenic unit (IV) 90-0: The laminations become increasingly disturbed by pedogenesis and bioturbation. Precipitation of soft pedogenic carbonate nodules (Bca-horizon), root casts, animal burrows and desiccation cracks obliterate veneers of organic detritus above ~60 cm. At this
157
level, the sudden appearance of mostly locally-derived gravel supersedes the exclusively finegrained depositional regime. 5) Surface drape (V) up to 10 cm thick: Yellowish red platy fine-grained mantle that covers the top of the section and most of the alluvial plain. In places, it drapes steep erosional gully banks (Photo 3B), indicating post-incisional deposition yet to be dated. The central sequence of laminations (II&III) includes three lithofacies (Photo 3A): a) Yellow bands (red band): Light yellowish brown (2.5Y 6/4 dry/ 5/4 moist) and light reddish brown (5YR 6/3 dry/ 7.5YR 6/3 moist) friable granular silt and very-fine sand mantling underlying sediments as discrete sheets. A dozen are close to 10 cm thick and continuous (Photo 1). Despite their overall excellent preservation, sedimentary structures, such as cross-stratification, were not identified. In some parts, small pieces of charcoal are incorporated. Together with 22 discrete, only centimetre-thin, and 20 more disturbed often mottled yellow bands, this lithofacies makes up two thirds of the laminated sequence (appendix 5.3A). Upper boundaries are typically blurred, with vertical millimetre-thick channels extending from the overlying lithofacies, introducing dark organic-rich material. b) Organic veneers: Similar fine-grained sediments hosting stacked undulating veneers of black plant detritus, dominated by elongated phytoliths with bulliform and quadrilateral cells characteristic of grasses (Williams et al., 2001). The 33 best preserved veneers mostly overlie yellow bands and range in thickness between <1-4 cm. Together with 25 dark bands of disturbed veneers of up to 5 cm thickness, organic veneers contribute to a quarter of the laminated sequence (II&III) (appendix 5.3A). The veneers are best preserved where associated with tufa. c) Tufa: White discontinuous streaks of carbonate precipitation, loosely cemented and typically associated with organic detritus. The tufa closely resembles carbonate precipitation observed west of the Ranges in shallow evaporating pools in the aftermath of the January 2007 flood (Photo 2G). Tufa bands of up to 5 cm thickness provide ~8 % of the laminated sequence (II&III) (appendix 5.3A), but only two bands are continuous enough to be included in the type section drawing (Photo 1). The three lithofacies largely follow a cyclic stacking pattern: yellow bands are topped by organic veneers that alternate and interfinger with streaks and thin sheets of tufa (Photo 3A). Transitions between the different lithofacies are usually sharp and, except for the 12 continuous yellow bands, undulating. No erosional contacts were observed, but upper boundaries of the yellow bands are often blurred by bioturbation.
158
Imbrication Patterns
Discrete gravel facies were restricted to one narrow line of well-rounded cobbles (Photo 1; 2D) within the stratigraphic section. Elsewhere across the deeply dissected alluvial plain however, numerous clast-supported gravel-fills are exposed as chutes and sheets. The main imbrication direction was mapped for all significant gravel exposures with the aim to reconstruct dominant flow patterns throughout the aggradation of the Silts (appendix 5.3B). The gravel occurrences are grouped into three colour-coded categories as a function of their depth below surface (i.e. their relative age): 1) onset of aggradation (green); 2) main aggradational interval (yellow); and, 3) termination (red) (Fig. 2). The lowermost gravel exposures largely consist of thick narrow channelfills, approximating present-day flow directions of the entrenched Brachina and Etina Creeks and in places fan out across the bedrock surfaces. With advanced Silt aggradation, there is a marked deviation from the present-day channel network. Accordingly, the Brachina tributary discharge flowed across the alluvial plain into Etina upstream of the modern confluence (Fig. 2). In contrast, the main flow axis of the Etina Creek remained much the same. The locality of the laminated stratigraphic section was partially shielded by raised bedrock outcrops and occupied a peripheral position. According to a sequence of shallow superposed gravel beds in the proximity of the modern confluence (Fig. 2) that display reversed directions of inclination (herringbone pattern) (Photo 3C), the site was repeatedly inundated by water flowing out of and into the main axis of flow. Towards the termination of the fine-grained depositional regime, the Etina/Brachina overflowed the complex bedrock topography of todays confluence. Further upstream, the termination is characterised by gravel sheets extending across the alluvial plain, mantled only by the surface drape. The long -axes of the inclined gravel clasts are orientated more often parallel than perpendicular to the palaeoflow direction, possibly indicating a hyperconcentrated high-energy flood depositional environment (Rust, 1972; Hartley et al., 2005), however further quantitative analysis of the gravel lithofacies is required.
159
Fig. 2) Drainage pattern: evolution of flow directions throughout the aggradation of the Brachina Silts as inferred from main imbrication directions measured from inset gravel exposures along the Brachina and Etina Creek beds. The fine-grained valley-fills are colour-coded in shaded red and the stratigraphic section BRA-SD is indicated in white. The arrows point towards palaeo-flow directions of mapped gravel chutes and sheets. Their relative depth below the highest floodplain surface is presented by three colours and corresponds to the main lithostratigraphic units of BRA-SD (see appendix 5.3B): green refers to the onset of fine-grained aggradation with gravel exposures resting on or closely above the bedrock (= basal unit I); yellow refers to gravel exposures cut into or aggrading throughout the main aggradational phase (= laminated units II&III); and, red refers to gravel sheets that spread out across the fine-grained floodplain, usually mantled only by the surface drape (= pedogenic unit IV).
Sediment-size analysis
Texturing of the fine-grained valley-fills presents a challenge in that the material invariably consists of silt loam, sometimes becoming heavier (more clayey) with continued working by hand. This soil property, termed subplasticity, is widely reported from loess-derived sediments throughout southeastern Australia (e.g. Butler, 1955; McIntyre, 1976). In order to obtain a detailed and genetically more meaningful picture of the sediment composition, a two-step approach was performed involving: 1) high-resolution 3-D sediment-sizing by the electrical sensing zone method employing
160
the Multisizer 3 COULTER COUNTER (Beckman Coulter, 2002); and, 2) conventional and parametric analyses of the statistical particle-size distributions. Eleven samples from key yellow bands and other significant strata were first described by standard statistical derivatives. Then, discrete principle particle populations that make up the size distribution were quantified for both partially aggregated (minimally-dispersed) and particulate (fully-dispersed) sample conditions. The granulometric results are interpreted in terms of sediment provenance and depositional mode. Protocol The consolidated sample material was immersed in ISOTON II (1 g/100 ml), left to slake, and stirred in a baffled beaker, maintaining a uniform suspension throughout sub-sampling by pipette. This treatment breaks up loosely-bound aggregates, but further dispersion is soon resisted. The sample state is described as its minimally-dispersed (MD) condition (Leys et al., 2005) and assumed to survive non-laminar water flow (Schieber et al., 2007; Haberlah and McTainsh, subm.). Consequently, the sediment suspension was dispersed by means of ultrasonic bath (Branson 2 200 sonifier, 472 Hz/ 60 W for 30 min) and hydrochloric acid (1 % HCl), removing carbonate cementation, tufa and shells. Sample suspensions were diluted to concentrations of 5-10 % and passed through an array of calibrated orifice tubes (560, 280, 140 & 50 m) covering the nominal particle-sizing range between 1-336 m with generous overlap. A minimum sizing threshold of 3-4 m was employed to eliminate the effect of electronic interference caused by iron-rich particles that resulted in some abnormal electrical pulses towards the lower size limit. Hence, clay particle contributions (<3.9 m) are truncated. The size fraction >200 m was wet-sieved for ~300 g of sample material. Results are expressed in weight percentage (appendix 5.3C), described, and photographed under binocular microscope. The silt- to sand-sized particle-size distributions (PSDs) are based on >1 000 000 analogue pulses that were converted and merged into 256 discrete size classes of sub-micron resolution. Standard statistical derivatives such as percentiles, mean, mode, skewness and kurtosis (Inman, 1952; Folk and Ward, 1957) were described by the Beckman Coulter Multisizer software (Beckman Coulter, 2002) (Fig. 3; appendix 5.3C). Consequently, the PSDs were converted into log10listings and imported into the Mixdist library (Macdonald and Du, 2004) as part of the R environment for statistical computing (R Team, 2008). The Particle Size Distributions (PSDs) were resolved into 2-4 discrete stretched-exponential (Weibull) distributions interpreted as the principal particle populations that make up the sediment (Folk, 1971; Sun et al., 2002; Leys et al., 2005; Haberlah and McTainsh, subm.). The iterative computation, employing a Newton-type algorithm and the expectation-maximisation algorithm, was initiated by specifying 5 evenly-spaced unconstrained end-members, and progressed by eliminating insignificant (<1 %) and redundant populations (Haberlah and McTainsh, subm.) (appendix 5.3C).
Lo ess an d Flo o d s: Results
161
All fully-dispersed (FD) samples are predominantly composed of silt (3.9-62.5 m) (Fig. 3). Most yellow bands further display a significant very-fine sand component (62.5-125 m) but, in contrast to samples from the basal unit (I) and the uppermost section (IV&V), little to no coarse material. The coarser 275-285 yellow band is an exception by comprising ~8 % fine sand. The PSDs of the yellow bands are narrow and peaked, with coarse silt to very-fine sand (20-125 m) invariably making up >75 % of the sediment volume. All sample modes fall within a narrow range of 50-63 m, except for the 470-479 red band. Here, a 71 m mode indicates the presence of coarser material, but the median particle size of 60 m still lies within the silt size range. Negative skewness, here expressed by modes larger than corresponding median values (Fig. 3), is evident in all but the near-symmetrical distribution of the lowermost sample (650-660) and reflects the presence of additional fine sediment. In the top sample (000-010), this fine component is particularly pronounced, resulting in a bimodal distribution. Aggregation is assessed by juxtaposing corresponding FD and minimallydispersed (MD) particle-size distributions and their statistical derivatives (Fig. 3). For most samples, differences between the two are negligible. An increase of >5 % in the silt fractions is only recorded for the coarser 275-285 yellow band and the two uppermost samples (IV&V). This increase corresponds with a decrease in the >62.5 m fractions, indicating the presence of sand-sized mud aggregates. A comparison between ratios of mode and median values for both sample conditions show that PSDs are more symmetrical in MD condition. A genetically meaningful quantitative description of the sediment composition and nature of aggregation is attempted by parametrically resolving the distributions into their principle particle populations. Accordingly, the primary particle population consists of coarse silt with modes ranging from 40-64 m contributing >75 % to all yellow bands (Fig. 3). The 348-355 yellow band, in which this population attains only 72 %, wedges out in the section face and exhibits a larger fine silt particle population. This poorly-sorted secondary population is present in all samples with an average mean and mode of 16 m. In samples from the palaeosol, transitional unit (II), and termination of the sequence (VI&V), the fine silt particle population is more abundant than in the yellow bands of the laminated unit (III). The two lowermost and uppermost samples and the coarser 275-285 yellow band further comprise a fine sand tertiary particle population (Fig. 3). With the exception of the basal unit (I), this coarse particle population is resolved in corresponding FD sample conditions, indicating that it is made up of mud aggregates. With the exception of the coarser 275-285 yellow band, >99 % of the material of the yellow bands is made up of particles <200 m. This is in contrast to the basal (I) and uppermost sediments (IV&V) that respectively consist of 10 and 3 of sand -sized well-rounded quartz grains (Photo 3D),
gastropod shells (Photo 3E), charcoal (Photo 2C) and tufa fragments, and lithic clasts (appendix
162
5.3C). Between 506-380 cm, a succession of thin, less-continuous laminations, in places incorporating streaks of tufa and topped by veneers of organic detritus (Photo 3A), was sized in MD condition (Haberlah et al., 2007). Results indicate a succession of upward-fining couplets, as expressed by a decrease in the coarse silt primary particle population accompanied by an increase in the fine silt secondary particle population. In the representative example (Fig. 3), the volume contribution of the 43 m primary particle population decreases with subsequent deposition of material associated with in situ tufa formation and plant detritus, to the point where the 11 m secondary particle population makes up the bulk of the material. Data interpretation The texture of the sediments remains remarkably similar throughout the aggradational sequence, principally consisting of coarse silt and very-fine sand. The narrow size range and the steep drop-off in coarse material are attributed to the aeolian provenance of the material. As proximal dust, the material was carried in suspension across the steep range front into the catchment, with the upper size limit controlled by maximum wind velocities. The high-resolution particle-size distributions, and their decomposition into principle particle populations, reveal subtle but important variations in sorting and the presence of transport-stable mud aggregates. The sedimentary fraction that straddles the Wentworth boundary between silt and sand (62.5 m) dominates all samples but is particularly expressed in the narrow and peaked (leptokurtic) particle-size distributions of the yellow bands. Their well-sorted composition is best explained by additional sorting of the former loess mantles during fluvial entrainment, transport and deposition. The primary and secondary particle populations were collectively transported in suspension by turbulent flow. With the reduction in flow capacities as a result of backflooding from the narrow gorge entrance, the coarse silt particle population settled out of the water column more rapidly than the fine silt population. This interpretation is consistent with an increase in the fine silt population in overlying material associated with plant flotsam, and in the finer 348-355 yellow band that pinches out in the section face and hence reflects the distal reach of the associated inundation event. In contrast, the lowermost (I) and uppermost (IV&V) samples are less sorted and incorporate lithic sands. They are more likely to have been deposited as fluvial bedload or colluvium from adjacent loess-mantled slopes. A comparison between corresponding MD and FD particle-size distributions suggests that these samples, and the 244-250 tufa-bearing yellow band, also comprise transport-stable mud aggregates primarily made up of fine silt. All minimally-dispersed sample expressions exhibit the better defined modes as expressed by a shift from platykurtic to leptokurtic, or in the case of sample (000-010), from bimodal to unimodal particle-size distributions. This observation is interpreted to reflect pre-depositional aggregation, with mud aggregates possibly formed during fluvial transport
163
(Schieber et al., 2007), and corroborates a study conducted downstream concluding that loessderived alluvium is best characterised by minimally-dispersed sediments (Haberlah and McTainsh, subm.).
(see opposite page) Fig. 3) Multisizer 3 COULTER COUNTER samples: particle-size distributions (PSDs), key statistical derivatives and parametrically-resolved principle particle populations. Mode (MD) and median (d50) particle sizes are listed for both samples states, with black numbers referring to minimally-dispersed and white numbers to fully-dispersed samples. The PSDs are presented as particle sizes (in m) on a logarithmic scale versus the normalised volume, with size boundaries of clay/silt (3.9 m), silt/veryfine sand (62.5 m) and very-fine sand/fine sand (125 m) indicated by dotted lines. The sedimentary fractions are quantified in terms of fine (3.9-20 m) and coarse silt (20-62.5 m), very-fine sand (62.5-125 m) and fine sand (125-250 m), and depicted as pie charts with blue colours referring to minimally-dispersed, and red colours to fully-dispersed sample modes (see legend). Parametricallyresolved, partially-overlapping principle particle populations are plotted in blue for minimallydispersed, and red for fully-dispersed sample states. The population modes are indicated as black, white and grey triangles (see legend). The dominant principle particle population is listed as a mode/mean value (in m) in the PSDs, with corresponding relative percentage to the right. Minimally-dispersed PSDs of a succession of laminations forming a representative slackwater couplet are superimposed, with partially-constrained modes highlighting the depositional size trends in the coarse silt and fine silt particle populations settling out over the course of flood fluxes (see legend).
164
165
Mineralogy
Mineral spectroscopy was performed employing the QEMSCAN technique which combines electron beam and backscattered electron (BSE) systems with scanning electron microscopy (SEM), X-ray mapping and liquid nitrogen-free energy dispersive spectrometers (EDS) (Gottlieb et al., 2000). Primary rock-forming minerals, secondary clay minerals, evaporites and iron oxides were quantitatively assessed in 13 undisturbed consolidated hand specimens, including all previously sized section face samples, the underlying bedrock and loose dune sand from the western piedmont plain. The polished sample surfaces were mapped by dedicated software (iDiscover), providing additional spatial context, size and texture attributes. The mineral maps are interpreted in terms of their degree of particle sorting and the presence and nature of aggregation, bioturbation, and in situ formation of clay minerals, carbonates and iron oxides. Size dependencies of detrital primary mineral particle abundances are discussed in terms of sediment provenance, and as a potential proxy for palaeowind velocities (Kandler et al., 2009). Protocol The QEMSCAN samples were ground and polished following an adapted vacuum epoxy treatment optimised to fixate larger, loosely-bound particles. Sample surfaces were scanned by QEMSCAN 310-series instrument fitted with SIRIUS 10/SUTW detectors, employing electron beam stepping intervals of 15 m over a total area of 100 mm2. A site-specific regolith mineral Species Identification Profile (SIP) library was compiled. The ten most abundant minerals, as identified by bulk mineralogical analysis, are presented as colour-coded mineral maps. For all major constituents (>10 %), average mineral particle sizes were stereologically calculated from intercept lengths and boundary transitions (Sutherland, 2007). Results All specimens include an identical suite of dominant minerals except for a fluctuating carbonate content (Fig. 4). Quartz and feldspars account for more than half the sediment volume in all but sample (650-660) from the basal unit (I), the lowermost (492-500) and uppermost (093-100) yellow bands, in which clay minerals constitute the majority. The quartz fraction appears to control the overall composition more than any other constituent and has an inverse relationship with carbonates, but for the carbonate-free palaeosol and surface drape. The latter two stand out by a near-identical composition and corresponding average mineral particle sizes. The quartz fraction invariably constitutes the coarsest major component. Throughout laminated aggradation (II&III), mean quartz particle sizes gradually increase, peak and decrease, approximating the trend of their overall abundance in the samples (Fig. 4). The same size-trend is mirrored by K-feldspars, here
166
however independent of their relative volume contribution. In contrast, the clay minerals do not exhibit vertical size-trends and fall within a narrow size range. Minerals of the bedrock sample remain below corresponding average mineral particle sizes in the overburden. Single mean particlesize values are problematic in samples with polymodal particle-size distributions, as evident in the dune sand. Its quartz component was digitally resolved into a mineral-size distribution that highlights the abundance of material in the silt to fine-sand size range (Fig. 5). The generation of mineral maps from polished surfaces of the samples provides additional size, shape and contextual information (Fig. 6). The high degree of sorting in the yellow (and red) bands throughout the interval of laminated aggradation (II&III) is in contrast to samples from the basal unit (I) and the uppermost sequence (IV&V) that incorporate sand-sized particles of quartz, mud aggregates, and carbonate and lithic clasts. All section samples comprise some quartz and K-feldspar particles larger in size than present in the homogenous fine-grained bedrock specimen, however abundant in the dune sample. The yellow bands differ in their degree of post-depositional bioturbation and pedogenesis. Irregular vertical channels of varying thickness lined by secondary clay minerals characterise the lower and uppermost yellow bands. The 470-479 red band and the 348-355 yellow band exhibit detrital root pseudomorphs cemented by iron oxide precipitation (Photo 3F). Carbonates occur either as homogenously dispersed clasts slightly larger in size than corresponding quartz particles (650-660, 470-479, 000-010), or as in situ precipitates (348-355, 093-100). The latter, wet-sieved and examined under binocular microscope, morphologically resembles cemented paludal tufa (Photo 3G) (Pedley, 2009).
167
Fig. 4) QEMSCAN samples: bulk mineralogy and mineral particle sizes. Minerals are colour-coded (see legend) and presented as relative volume percentages in the sample. Stochastic mean mineral particle sizes are listed for all major constituents (>10 %) in the corresponding bar segments.
(see opposite page) Fig. 5) QEMSCAN dune sample: virtually-resolved particle-size distribution of the quartz component, highlighting its abundance in the coarse silt- (20-62.5 m) and very-fine sand (62.5-125 m) -sized fraction.
168
Lo ess an d Flo o d s: Data interpretation
169
The lowermost (I&II) and uppermost (IV&V) units are more weathered than the laminated sequence. The question of pre-burial weathering versus in situ post-depositional alteration can be approached by comparing the clay spectral features, and the nature and abundance of carbonates susceptible to dissolution and reprecipitation. The samples from the basal unit (650-660) and the uppermost yellow band (093-100) that exhibit the highest clay contents are both associated with high carbonate concentrations (Fig. 4). However, carbonates in the lower units (I&II) occur as detrital largely undissolved particles as opposed to the in situ precipitates in the upper section (III&IV) (Fig. 6). Clay mineral formation within the laminations (II&III) is spatially restricted to traces of bioactivity and therefore took place in situ. Consequently, clay minerals in the lower sequence are interpreted to be primarily inherited from reworked material mixed with carbonate clasts, while throughout laminated aggradation a sizable contribution was formed in situ dissolving all carbonate particles in the affected zones. Carbonate and gypsum precipitation around tubular structures (weathered plant material?) is clearly evident in the 244-250 tufa-bearing yellow band. The clay mineral contents of the palaeosol and surface drape (Fig. 4) are among the lowest and largely consigned to mud aggregates (Fig. 6). Their wide range of mineral particle sizes suggests reworking of various source materials. The large primary mineral particles stand in contrast to the fine-grained homogenous bedrock specimen, but compare well with the finer fraction of the dune sample. The correlation between quartz size and abundance within the laminated sequence (II&III) possibly reflects palaeowind speeds in a scenario in which quartz sand, as the dominant dune mineral constituent, would have been winnowed from upwind dunefields.
Fig. 6) QEMSCAN field scans: mineral maps of the undisturbed polished sample specimens. The 10 most abundant minerals are colour-coded against the white background (see legend). A graphic scale is presented in form of maximum silt-, very-fine sand-, and sand-sized particle diameters (Wentworth
170
classification). Key features discussed in the text are pointed out by annotated yellow markers: a) sand-sized well-rounded quartz and K-feldspars; b) mud-aggregates composed of fine sand, silt and clay particles; c) lithic sand and gravel; d) unaltered detrital clasts of carbonate; e) trace fossils/channels filled and lined by secondary clay minerals; f) in situ carbonate/gypsum/iron oxide precipitation around tubular voids; and, g) iron oxide-cemented plant pseudomorphs.
Magnetostratigraphy
Sedimentary deposits such as loess/palaeosol sequences (Heller and Liu, 1984) and loess-derived slackwater aggradation (Huang et al., 2007) record vertical trends in measured variations in magnetic susceptibility that reflect climate and sediment unit-specific information (Maher, 1998). Magnetic susceptibility () is defined as the magnetisation (M) of the sediment material by an induced magnetic field (H), both measured in amperes per metre; = M/H. The degree of magnetisation induced by the applied magnetic field is related to the mineralogy and particle-size distribution of constituent iron oxides. There is little if any correlation between total iron content and magnetic susceptibility (Fine et al., 1995) because the strong ferrimagnets magnetite (Fe3O4) and, by subsequent oxidation, maghemite (Fe2O3) that dominate measured soil magnetic remanence and susceptibility occur only in trace amounts (Maher, 1998). Loess may also have a weak susceptibility component from hematite and goethite (Zhou et al., 1990). Significant magnetic enhancement may result from pedogenic formation of ultra-fine magnetite or maghemite, biogenic contributions (bacterial Fe3O4), or burning of the (vegetated) surface producing trace amounts of ultrafine ferromagnetic minerals. Pedogenic formation of ferrimagnets is favoured in well-drained soils while prolonged waterlogging results in iron reduction and magnetic depletion by favouring the formation (conversion) of weakly magnetic Fe3+ sulphides or by dissolution (Maher, 1998). The magnetostratigraphy of the section can therefore be interpreted as a proxy for the degree of pedogenesis experienced, the hydrological history of the site, or in terms of clastic magnetic mineralogy and/or the concentration of magnetic minerals. Protocol The Bartington Instruments MS2 Magnetic Susceptibility System was employed in conjunction with the MS2E sensor with a probe area of 3.8 mm x 10.5 mm designed to perform high-resolution volume susceptibility measurements on the surface of cores (Bartington, 2007). The sensor was moved incrementally in 1 cm intervals along the full length of the air-dried surface of overlapping section cuttings (obtained from the cleaned section face by P.G. using an angle grinder), and calibrated with a standard between each reading. The process was repeated three times across different transects offset from the central line, and mean and median values were plotted (Fig. 7).
Lo ess an d Flo o d s: Results
171
The depth plot of measured magnetic susceptibility values shows considerable variation between individual readings, but also indicates major trends. The sequence can be subdivided into three segments, largely corresponding with the stratigraphic units I, II&III, and IV/V, based on susceptibility values above and below 10-4 SI units (Fig. 7). The basal unit (I) is marked by elevated values (mean = 12.3 x 10-5 SI) decreasing towards the palaeosol. One narrow band of values that falls below 10 x 10-5 SI correlates laterally with the sheet of tufa clasts and small pebbles (Photo 1). A significant change is recorded with the onset of laminated aggradation (II&III), with susceptibility values fluctuating around a mean of 5.8 x 10-5 SI. Depleted values coincide with bands characterised by in situ tufa formation, most pronounced in the transitional unit (II) and the 244-250 tufa-bearing yellow band. Above the last discernible organic veneer, susceptibility values rise (mean = 14.4 x 10-5 SI) throughout the pedogenic unit (IV), with a dramatic increase (mean = 81.3 x 10-5 SI) towards the surface drape (V). The ratio of 2.6 between the elevated units (I&IV) and the laminated sequence (II&III) is identical to the ratio between 178 palaeosol samples and 227 pristine loess samples calculated for the Chinese loess deposits (Heller and Liu, 1984). While caution must be exercised interpreting individual readings which can be influenced by the uneven surface of the continuous section cutting and small scale mottles, the three independent readings correlate strongly with a Pearson coefficient of r=0.9. Data interpretation The magnetostratigraphy correlates closely with the lithostratigraphy. The magnetic susceptibility data can be interpreted in terms of: 1) elevated values against a background indicative of pedogenic activity in well-drained soils; and, 2) depleted values indicative of site-specific waterlogging. Accordingly, the basal (I) and the upper units (I&IV) are both characterised by soil formation processes, with one important distinction: while the uppermost Bca-horizon records a sharp increase towards the top of the section, the lower unit records a gradual decrease towards the top of the palaeosol. This can be explained by assuming that the lower sequence (I) consists of reworked sediments. During the erosion of former floodplain remnants and/or valley slope mantles, soil horizons with the highest susceptibility values are likely to become redeposited first. Alternatively, the susceptibility decrease could reflect the exposure of the palaeosol to stagnating water. In the overburden, a change in sediment supply relating to the influx of pristine loess is recorded. Pronounced negative excursions in otherwise relatively uniform background susceptibility values throughout laminated aggradation (II&III) are interpreted to reflect temporal waterlogging. Bands of depleted values are often associated with paludal tufa, interpreted to have formed in ponding conditions that by modern analogy must have lasted for weeks. The exceptionally high values
172
towards the surface drape (V) can be explained by the assumption that the material was recycled between creek beds and floodplain, reflecting multiple iron reduction/oxidation cycles (Haberlah et al., in press).
Fig. 7) Magnetic susceptibility log: mean (black) and median values (red) of three repeated magnetic susceptibility readings at 1 cm increments plotted against depth in the stratigraphic type section. Towards the surface drape (V), the values increase markedly, cumulating in 98.1 SI as indicated by an arrow. Boundaries between the main lithostratigraphic units are projected in blue.
173
Carbon isotopic geochemistry

Trends in carbon isotope ratios of detrital plant matter preserved in sediment sequences record changes within the palaeo-floral composition of the catchment. Aside from a few succulents and epiphytes, terrestrial plants fix atmospheric CO2 using two distinct photosynthetic pathways expressed by characteristic isotopic fractionation and ratio of depletion of 13C (Farquhar et al., 1989). Each group, the so-called C3- and C4-plants, can be distinguished by a mutually exclusive range of 13C-values (Marshall et al., 2007). In a recent study, the isotopic composition of plants of the Lake Eyre region flanking the Flinders Ranges to the north was measured (Johnson et al., 2005). The C3plants comprise all trees, shrubs, dominant chenopods and forbs, and the C4-plants all the grasses, herbs and a few sedges. The average 13C-values for these C3-plants (-26.8 1.8 ) and C4-plants (13.8 0.8 ) are here used as endmembers in a regional isotope mass balance model. The relative percentage of the C3-plant composition can be calculated as follows:
with Corg sample referring to the measured 13C-values of the analysed veneers of organic detritus and pieces of charcoal expressed in relative to the Peedee Belemnite Standard (PDB). The isotopic composition of sampled bulk organic matter is interpreted to reflect the relative C3/C4-plant abundance in the Brachina catchment at the time of entrainment and deposition in the aggradational sequence. Preferential incorporation and/or preservation of C3- versus C4-plants in organic debris, with the potential to systematically offset the equation, is addressed by analysing a representative (~200 g) sample of flotsam collected on-site in the aftermath of the once-in-ahundred-years flood that occurred in January 2007. The model does not take into account any postdepositional modification of the isotopic signatures. However, significant alteration by biogenic activity and pedogenesis is unlikely given the excellent preservation of the fragile organic veneers. Contributions of aquatic plants such as phytoplankton and macrophytes cannot be ruled out, but the reported range of 13C-values, particularly throughout laminated aggradation, is too small and isotopically enriched for such a scenario (Leng et al., 2006; Finlay and Kendall, 2007). The palaeofloral composition of C3- and C4-plants is the product of global, regional and local meteorological and climatic controls. The C4 photosynthetic pathway is favoured by low atmospheric pCO2, high water stress and high temperature (Ehleringer et al., 1997; Sage and Coleman, 2001). Trees and other higher C3-flora are less efficient in water use and do not recover as easily from wildfires (Tipple and Pagani, 2007). The sensitivity of C3- and C4-plants towards temperature changes and water stress is also a function of the seasonal rainfall regime (Liu et al., 2005; Finlay and Kendall, 2007). For a given
174
pCO2, cooler, wetter conditions favour C3-plants, while warmer, more arid conditions favour C4plants. Protocol Several grams of bulk organic matter, targeting every discrete veneer and piece of charred vegetation, were collected from the continuous section cutting and across the full section face. The samples were pulverised, then acidified several times in 1 m HCl to remove carbonates, rinsed twice in deionised water, dried and weighed into tin capsules. All samples were run by continuous flow on a Fisons Optima stable isotope-ratio mass spectrometer coupled to a Fisons elemental analyser. Sample data were corrected for internal fractionation during each run based on the measured values of regularly spaced standards whose isotopic compositions bracketed those of the samples. Multiple house standards were measured during each analytical run to ensure consistent reproducibility, which was 0.6 (1; n=13) over the course of the analyses. Duplicates were measured on more than 10 % of the samples as an additional assurance of sample reproducibility. Results The samples from the section cuttings (grey triangles) comprise the 32 best-developed and most continuous organic veneers from the laminated sequence (II&III) (Fig. 8). An additional 58 samples extend and complement this record by including more disturbed and partly discontinuous veneers, detrital charcoal from the basal unit (I) and lower yellow bands (II), and modern flotsam (black triangles). One third of those samples (19, white triangles) were also sent to The Australian National University (ANU) for radiocarbon dating on the Single Stage Accelerator Mass Spectrometer (SSAMS), which reports the precision of the 13C/12C ratio within 2 (Fifield et al., 2007). The pretreatment for radiocarbon dating involved alkaline leaching of humic acids with the aim to concentrate the insoluble charred organic fraction (Olsson, 1986). The 13C-values associated with the isolated plant fractions exceed the range of values reported for the bulk organic matter (Fig. 8), and were excluded from the following calculations. With an average 13C-value of -23.6 , the material sampled from the basal unit (I) reflects a predominant C3-plant composition (>75 %). The onset of laminated aggradation (II) coincides with a marked shift in the palaeo-floral composition towards C4-grasses at the expense of C3-plants, declining from -26.4 (93 %) to -20.2 (<50 %) just above the 470-479 red band. Subsequently, the strongly fluctuating 13C-values become, on average, more 13C-depleted towards the uppermost laminations, reaching -22.3 (65 %) at 169 cm. This overall trend marking the laminations can be subdivided into a lower part, characterised by an average 13C-value of -22.0 corresponding to a 63 % C3-plant contribution reaching a plateau at ~320 cm, with values decreasing from -22.1 -
175
24.2 , and an upper part, with an average of -22.7 corresponding to a 68 % C3-plant abundance terminating in a plateau at 169 cm, with values increasing from -20.9 -23.4 . The uppermost sequence (IV), hosting only a few poorly-preserved organic veneers that terminate by 51 cm depth, reflects an increase in C3-plants with an average 13C-value of -23.1 translating to a C3-plant contribution of 72 %. Finally, the 13C-value of -26.8 obtained from present-day flotsam suggests a 100 % C3-plant composition. Data interpretation While small-scale fluctuations in 13C-values may reflect short-term variations in the palaeo-floral composition, the larger scale plateaus at the onset of laminated aggradation at 506 cm, and just below 300 cm, are best explained within a chronological context. Calibrated radiocarbon ages of concentrated C3-plant fragments (charcoal) predate those of concentrated C4-plant fragments, invariably returning pre-24 ka cal BP ages for the lower half of the laminations (Fig. 8). Hence, these isotopically depleted vegetation elements predate the laminated depositional regime, representing sequestered residual and possibly reworked charcoal. Reworking is also envisaged for C3-plant material dominated samples from the basal unit (I) that are older than the lowermost radiocarbon age, i.e. within the range of ~34-29 ka cal BP. If so, the Brachina catchment is characterised by a relatively C4-dominated floral composition throughout the lead-up and peak of the LGM, with a fluctuating but overall continuous increase in C3-plant contribution to the total sedimentary organic carbon. C3-vegetation possibly prevailed during peak glacial conditions ~320 cm, but still remained ~20 % below the isotopic composition of modern flotsam. Considering the sizeable quantity of C3grasses growing within the catchment today, the 100 % C3-plant equivalent 13C-value from the collected modern organic debris suggests that transport and depositional processes and/or organic decay systematically discriminate against C4-plants in flotsam, possibly reflecting the higher preservation potential of lignin and cellulose which is more significant in C3-plants. The relative predominance of C4-plants throughout the LGM despite prevailing lower temperatures cannot be explained by the lower pCO2 alone, but rather indicates seasonally and inter-annually more arid conditions and possibly more frequent wildfires.
176
Fig. 8) Carbon isotopic log: 13C-values of bulk organic matter plotted against depth in the stratigraphic type section. Samples collected from the section face are presented in black, samples from the continuous section cuttings in grey, and dated sub-samples in white listing their corresponding 14C-ages in ka cal BP. The lower abscissa presents their relative percentage of C3-plant contribution according to regional isotope mass balance model based on data from Johnson et al. (2005). Plateaus in 13C-values discussed in the text are projected in blue.
Chronostratigraphy
In order to establish the timing and rates of deposition of the main stratigraphic units, 33 age estimates were obtained: 27 by accelerator mass spectrometry (AMS) radiocarbon dating, and 6 by optically stimulated luminescence (OSL) dating. Within the radiocarbon dating program, three types of organic materials were sampled from the section: discrete charcoal pieces, bulk plant detritus from organic veneers, and intact freshwater gastropod carbonate shells. The fundamental principle behind radiocarbon dating is that the radioactive isotope 14C decays at a constant rate from an initial concentration in near-equilibrium with a known atmospheric concentration starting with the death of the organism (i.e. cessation of photosynthesis or the intake of biomass) (Bowman, 1990). Hence,
14
C-dating of charcoal establishes the age of cell growth, 14C-dating of organic veneers averages the
time of death of plant litter (and charcoal) entrained by the flood event, and 14C-dating of carbonate shells reflect the age of carbonate precipitation. None of these are necessarily coeval with the depositional event, as demonstrated in a study on contemporaneous charcoal in a fluvial
177
environment by Blong and Gillespie (1978). The relative resilience and longevity of charcoal make it prone to endure multiple cycles of erosion and redeposition. The same principle applies to shells and plant litter, albeit perhaps to a lesser degree considering their relative fragility. Therefore, 14C-ages must be strictly treated as terminus post quem in any chronostratigraphic context. The OSL-ages date the timing when the sampled subset of quartz grains were last exposed to light prior to burial and consequent irradiation (Aitken, 1998). The timing of exposure can vary from grain to grain, particularly if the sediment only experienced short-distance transport with high suspension load concentrations (Olley et al., 1998). Post-depositional contamination by infiltration of younger grains by seepage, pedogenesis or bioturbation further complicates the analysis. With recent technical advances in luminescence readers, dating protocols and age models, these can now be accounted for on a grain by grain basis (Duller, 2004). However, a decade ago when the OSL-ages for BRA-SD were analysed, the luminescence signal could only be averaged across large aliquots of grains mounted on 1 cm discs (see Williams et al., 2001 for details). Protocol Radiocarbon samples were pre-treated, the carbon extracted and converted to graphite using standard methods (Hua et al., 2001). The samples were either dated on the ANTARES AMS facility at ANSTO (Fink et al., 2004) (code: OZJ), or the recently installed SSAMS at the ANU (Fifield et al., 2007) (code: SSAMS). In addition, two previously published AMS-ages (Cock et al., 1999) (code: Beta) are included in the discussion and, like the other samples, calibrated using the integrated CalPal2007Hulu-calibration data set (Weninger and Jris, 2008), as part of the CalPal-2007 calibration and palaeoclimate research software package (Weninger et al., 2008). Results The independent age proxies allow quantification of the material-dependent residual times, best demonstrated by plotting age, sample material and method versus depth in the section (Fig. 9). In comparison to the calibrated AMS-ages, the OSL-ages appear to be younger by 2-3 ka (Table 1). AMS-ages based on both charcoal and shells, and OSL-ages describe a marked age inversion in the basal unit (V); from ~27-28 ka cal BP (assuming a systematic offset in the OSL data) via ~32-36 ka cal BP back to ~26 ka cal BP (OZJ909). Layered to laminated aggradation (II&III) sets in ~24 ka cal BP with the lowermost veneer of organic detritus dated to 23.83 0.25 and 23.71 0.20 ka cal BP (SSAMS ANU 4117, 4205). Large discrepancies between individual AMS-ages are apparent during the onset of the laminations. All discrete charcoal pieces (SSAMS ANU 2030, 2035) and the gastropod shell (OZJ905) sampled from yellow bands return ages a few thousand years older than those from corresponding organic veneers, suggesting reworking and long residence times of the material
178
within the catchment. Fragments of residual charcoal are also incorporated in some of the lower organic veneers. By using the reported 13C data and the regional isotope mass balance model (Fig. 8), a division was made into predominantly woody charred C3-plants, likely to survive multiple cycles of erosion and deposition, and the more fragile C4-plant-rich detritus (>-24 ppm PDB, <80 %), sensitive to such reworking (Fig. 9). The latter samples display a coherent age-depth trend throughout the laminated profile, with a linear regression suggesting a depositional rate of ~83 cm/ka (r=0.95). In contrast to the onset of layered to laminated aggradation, C3-plant based 14C-ages associated with the termination (IV) appear contemporary with the depositional events. Accordingly, a marked reduction in depositional rates to ~23 cm/ka (r=1) set in as early as ~22 ka cal BP (SSAMS ANU 4206) with layered to laminated aggradation terminating at 18.28 0.23 ka cal BP (SSAMS ANU 4107).
(see opposite page)
179
Table 1) BRA-SD radiocarbon ages: Accelerator mass spectrometry (AMS) radiocarbon data listed by sample material: C4-plants refer to bulk organic matter with 13C values >-24 , C3-plants to 13C values <-24 , and CaCO3 shells relate to unbroken carbonate shells of freshwater gastropods. All samples are correlated in depth to the stratigraphic type section (see appendix 5.3A). The 14C-ages are calibrated using the integrated CalPal-2007Hulu-calibration data set (Weninger and Jris, 2008), as part of the CalPal-2007 calibration and palaeoclimate research software package (Weninger et al., 2008).
Lab code Sample material below TOP (in cm) SSAMS ANU 4207 SSAMS ANU 4206 SSAMS ANU 4110 SSAMS ANU 4111 SSAMS ANU 4114 SSAMS ANU 4209 SSAMS ANU 4117 SSAMS ANU 4205 SSAMS ANU 4107 SSAMS ANU 4109 SSAMS ANU 4112 SSAMS ANU 4113 SSAMS ANU 4116 SSAMS ANU 2030 SSAMS ANU 2035 SSAMS ANU 2036 SSAMS ANU 2037 Beta-96679 OZJ904 SSAMS ANU 2039 OZJ905 OZJ909 Beta-96166 OZJ908 OZJ907 OZJ906 SSAMS ANU 1811 C4-plants C4-plants C4-plants C4-plants C4-plants C4-plants C4-plants C4-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants C3-plants CaCO3 shell CaCO3 shell CaCO3 shell CaCO3 shell CaCO3 shell CaCO3 shell CaCO3 shell 116 198 213 244 341 372 505 505 51 94 264 338 437 474 495 516 516 565 632 695 353 508 565 639 667 684 695 17.42 18.41 18.52 18.61 18.88 19.67 19.82 19.91 15.16 16.17 19.36 20.13 21.12 21.89 24.11 27.63 27.74 29.80 29.16 23.60 20.46 21.51 28.12 27.99 27.32 27.20 23.27 0.12 0.10 0.12 0.13 0.14 0.12 0.18 0.18 0.10 0.12 0.14 0.16 0.15 0.16 0.21 0.28 0.29 0.18 0.38 0.30 0.14 0.31 0.16 0.71 0.28 0.50 0.09 20.92 22.11 22.23 22.41 22.68 23.54 23.71 23.83 18.28 19.34 23.23 24.08 25.16 26.26 28.99 32.21 32.31 33.56 33.57 28.56 24.45 25.66 32.57 32.64 31.97 31.9 28.08 0.17 0.25 0.26 0.18 0.15 0.12 0.20 0.25 0.23 0.20 0.15 0.24 0.24 0.31 0.39 0.26 0.29 0.30 0.41 0.42 0.19 0.46 0.26 0.63 0.20 0.39 0.08 -22 -20 -19 -14 -22 -18 -15 -21 -26 -26 -27 -25 -29 -28 -26 -24 -29 -25 -27 Age (in ka BP) Error (1 SD) Age (ka cal BP) Error (1 SD) 13C (in )
Lo ess an d Flo o d s: Data interpretation
180
By increasing the resolution and variety of dated samples, a more detailed but also more complex chronostratigraphic picture emerges than previously assumed (Williams et al., 2001). Irrespective of sample type and dating method, the basal unit (I) displays a marked age inversion best explained by incorporation of reworked older material. The coincident poorly-bleached OSL sample AdGL-96003, and the unbroken condition of the large pieces of charcoal and shells suggest short-distance transport, possibly from adjacent valley slope mantles or nearby remnants of a former floodplain. While a tentative hiatus lasting for ~2 ka between deposition of the palaeosol sediments and the onset of the laminations can be inferred from shell sample OZJ909 obtained from the uppermost palaeosol, further dating by means of single-grain OSL-age estimates is required. Laminated aggradation setting in ~24 ka was rapid with close to four metres of deposition prior to peak LGM conditions ~21 ka. Significant amounts of residual charcoal and the occasional shell were initially sequestered with the coarsest suspension load (yellow bands). Organic veneers topping a closely spaced succession of yellow bands (SSAMS ANU 4111, 4110, 4106) indicate episodic aggradation, possibly of centennial frequency. The second half of the LGM experienced a marked reduction in deposition rates, and an increase in what appears to be contemporary elements of C3-plants. Finegrained aggradation terminated towards the end of the LGM ~18 ka. The surface drape remains to be dated.
Fig. 9) BRA-SD age-depth plot: accelerator mass spectrometry (AMS) radiocarbon data (see Table 1) are plotted as calibrated ages with error bars and colour-coded by sample material. Based on the regional isotope mass balance model generated from data published in Johnson et al. (2005), the
181
bulk organic matter is divided into predominantly woody C3-plants (13C <-24 ), including all discrete charcoal pieces depicted as brown squares, and organic detritus composed largely of C4plants (13C >-24 ) depicted as red diamonds. Dated carbonate shells from unbroken freshwater gastropods are depicted as blue triangles. In addition, OSL-ages based on three independent assessments of the environmental dose rate (see Williams et al., 2001) are plotted. Two interpreted depositional rates (linear regressions) are indicated in blue and listed. AMS-ages discussed as most likely approximations of the depositional events are connected by a black line, dotted in the lower part to express increasing uncertainty.
Discussion
The layered to laminated stratigraphic section consists principally of coarse and fine silt-sized quartz, feldspars and carbonate particles, differing in their degree of sorting and weathering. A significant segment of the primary minerals are larger than those in the fine-grained homogenous shales that underlie the sequence and most of the upstream valley-fills. The full size range of the sediments can be found in dune sands extending from playa Lake Torrens towards the range front. The apparent size dependency of the quartz abundance in the laminated sequence is consistent with the inferred aeolian provenance of the material, with quartz content and its mean particle sizes increasing simultaneously as more and larger material is entrained and transported from the quartz-dominated dunefields by strong westerly winds. Winnowing of active dunefields is suggested as a source for well-studied glacial loess (Sun et al., 2002) and desert loess (Crouvi et al., 2008) occurrences, and provides a plausible provenance scenario for the similar Namib Silts (Eitel et al., 2001) and Sinai Silts (Rgner et al., 2004). The presence of rolled detrital pedogenic carbonates, clasts of tufa, lithic sand and gravel, and traces of current cross-bedding indicate that the basal unit (I) aggraded as fluvial bedload. The alternating, upward-thinning lighter and darker bands are interpreted as a sequence of flood couplets decreasing in thickness towards the palaeosol. The darker bands, including the palaeosol, inherited some of their colour from plant detritus, in places still preserved as discontinuous veneers. The decreasing thickness of the couplets possibly reflects an increasingly limited sediment supply and/or an increase in accommodation space as the bedrock topography of narrow rock-cut channels became filled in. The results of the mineral spectroscopy show that the basal unit (I) is rich in secondary clay minerals indicative of chemical weathering, an interpretation consistent with the magnetic susceptibility record. The hematite- and goethite-masked (red-brown) colour of the sediments and the presence of undissolved detrital carbonate particles and shells suggest that the material consists of reworked
182
well-aerated former floodplain remnants and/or valley slope mantles. This scenario is consistent with paired AMS- and OSL-age inversions, indicating inherited ages for material dated ~36-32 ka, with fluvial redistribution setting in ~28 ka cal BP or earlier. The 506-516 cm palaeosol, as the uppermost and best-defined darker band, marks a hiatus followed by a pronounced shift in the depositional environment. In terms of bulk mineralogy, overall particlesize distribution, poor sorting and the presence of mud aggregates and large well-rounded sand grains, the palaeosol presents a buried analogue of the present surface drape. However, there are discrepancies between the two: a) the dark brown/grey colour of the palaeosol versus the yellowish red colour of the surface drape, b) the blocky to prismatic structure with slickensides versus a granular to platy structure, and c) induced magnetic susceptibility values that decrease towards the top of the palaeosol but sharply increase towards the top of the section. These differences can perhaps be explained by the comparably short-lived nature of the hiatus, possibly limited to ~2 ka or less, and by the hydrological environment of subsequent deposits. Large pseudogleyic root casts that extend from above testify to the palaeosols inundation. While the onset of laminated aggradation marks a sudden change in the depositional mode, redoximorphic pale and lepidocrocite-dominated (orange) colours that characterise the >4 m of laminations dominate above the transitional unit (II). Inherited AMS-dated charcoal and shells are consistent with a scenario of a gradual decrease in reworked material and sequestered organic matter. The sediments of the laminated unit (III) above 470 cm appear to have experienced little prior weathering, with secondary mineral formations restricted to traces of roots and burial activity, and magnetic susceptibility values common for pristine loess accessions. The laminated aggradation (II&III) is characterised by an absence of coarse material, cross-bedding and erosional contacts, all indicating a low-energy depositional environment. The stacked fining-upward pattern of the laminations is interpreted to reflect the density-sensitive settling velocities of suspended sediments. The yellow bands settled out of the water column first, effectively blanketing the flooded surface as flow capacities were reduced by backflooding from the gorge entrance. Their remarkably narrow absolute size-range is interpreted to reflect: a) an upper particle-size range constrained by wind velocities; and, b) additional sorting controlled by Stokes' law. Subsequently, floating plant litter settled out with the bulk of the fine silt-sized population. Tufa formed in situ during desiccation. Under the present climate, similar tufa precipitates over days and weeks in pools of water lining the creek bed in the aftermath of a major flood event. The thickness of the slackwater couplets relates to the suspension load, which is a function of the magnitude of the flood event and distance from the main flow. The more discontinuous laminations record either discrete smaller-scale inundations
183
or multiple sediment fluxes over the course of a single large-magnitude flood event (Kochel and Baker, 1988). The slackwater scenario is consistent with the increase in thickness of the yellow bands and the overall sequence downstream towards the present-day confluence with the Etina Creek. According to the gravel imbrication pattern, this channel persisted as the main axis of flow throughout the LGM, episodically inundating the peripheral depression that is now occupied by the stratigraphic section. The change in deposition from flood couplets to slackwater couplets corresponds with: a) a substantial increase in accommodation space as the pre-existing rock-cut channels were filled and the flow of the Brachina tributary spread out across the width of the valley; b) a major influx of proximal dust; and, c) a decrease in perennial C3-vegetation within the catchment. Water stagnation was short-lived and penetrated only newly deposited sediments, as manifested in the unaltered state of fine carbonate particles and hematite coatings of the 470-479 red band. Bioturbation likewise was limited in duration and penetration. In contrast, the termination of the fine-grained aggradational regime is characterised by pronounced pedogenesis and bioturbation coeval with an increase in C3-plants. The pedogenic unit (IV) records a fourfold decrease in sedimentation rates and a sudden presence of gravels. This is consistent with fine sediment starvation, i.e. the final stripping of loess-mantled valley slopes and consequent headward erosion, incision and realignment of the Brachina Creek along the northern flank of the valley to its present (and possibly pre-aggradational) course. Fluvial aggradation is superseded by the aeolian/colluvial deposition of a thin surface drape, possibly sourced from desiccating creek beds during dust storms. Palaeoclimatic and palaeo-environmental controls promoted the widespread aggradation of fines, as opposed to their erosion by ephemeral traction load streams under the current climate. Since the 1960s, pluvial anomalies in the overall more arid glacial landscape of south-eastern Australian are discussed as expressions of complex local hydrological responses to lower temperatures (Galloway, 1965). This minevaporal theory was invoked to explain the Brachina Silts in terms of a low-energy perennial wetland with swamp vegetation trapping dust washed in by gentle winter rainfall (Williams et al., 2001; 2006). However, reduced evapo-transpiration alone does not produce run-off capable of inundating and backflooding wide reaches of the valleys. The recently established coeval aggradation of extensive floodplains on the piedmont plain (Haberlah et al., in press) further emphasises a simple fact: flood deposits involve flooding rains. The LGM sequence of slackwater couplets indicates that the frequency and magnitude of exceptional rainfalls was similar or even of larger magnitude from today, raising the question about the controls that determine their erosive capacity. Over an interval of relative tectonic stability
184
(Quigley et al., 2007), aggradation and degradation of the fluvial system are determined by discharge and erosional/transport capacity, i.e. the rate of surface run-off and sediment supply. Surface run-off is inversely related to the vegetation cover (e.g. Lancaster, 2002). The dated isotopic record of bulk organic matter from the catchment suggests that the C3-vegetation that presently stabilises the hillslopes was burnt ~32 ka, and only recovered towards the Deglacial. The interval spanning the lead-up to and peak of the LGM was dominated by C4-grasses, possibly reflecting the higher tolerance of the C4-photosynthetic pathway to droughts and flooding (Baruch, 1994; Heckathorn et al., 1999). The transition from a temperate, more humid to a cold and arid climate characterised by the collapse of the perennial protective vegetation cover would have promoted valley slope erosion and, by increasing the sediment supply, an aggradational fluvial regime (Harvey and Wells, 1994; Zielhofer and Faust, 2008). The same scenario was suggested to explain palaeofloods larger in magnitude and further in reach than today depositing the lower Tsondab Silts in Namibia at ~28-26 ka and ~22-20 ka (Lancaster, 2002). In contrast, the recovery of perennial vegetation towards the end of the LGM would have favoured stream incision by reducing the sediment supply, thereby increasing the erosional capacity of the run-off (Harvey and Wells, 1994; Lancaster, 2002; Hessel, 2006). This is consistent with pollen and charcoal particle concentrations from terminal playa Lake Frome recording a marked increase in C3-vegetation from ~20 ka over a period of low wildfire frequency (Singh and Luly, 1991; Luly and Jacobson, 2000). The deposition/erosion rates can be expected to vary along the thalweg, depending on the local topography that controls flow velocities over the passage of the flood (van Maren et al., 2009). Slackwater deposits with the potential to record discrete flood events are restricted to backflooded reaches upstream of gorges, flooded tributary mouths and protected embayments (Kochel and Baker, 1988), and were described from all of these settings from multiple catchments in the Flinders Ranges (Haberlah et al., in press). Along the main flow, lateral floodplain accretion and, with increasing transmission losses, flood-outs will dominate aggradation, possibly explaining the range of depositional interpretations by different authors that studied different sections of the Namib Silts (for reviews see Lancaster, 2002; Srivastava et al., 2006; Haberlah, 2007; Heine and Vlkel, 2009). In the case of the Flinders Silts, the fluvial response to vegetation-controlled sediment supply may have been significantly amplified by fluxes of proximal dust. Accordingly, between ~29-18 ka the slope mantles were intermittently stripped of unconsolidated loess accessions and washed into the valley as hyperconcentrated sheet floods with little erosional capacity. As long as the fluvial regime remained charged with an excess load of loess it aggraded. Consequently, aggradation/degradation of the fine-grained regime is ultimately controlled by the dust supply and, by inference, the activity of downwind dunefields. Empirical studies and conceptual models suggest that dune activation is
185
principally controlled by thresholds in: 1) sediment supply; 2) the transport capacity of wind; and, 3) stabilisation by vegetation (e.g. Lancaster, 1994; Hugenholtz and Wolfe, 2005). These variables are strongly affected by the seasonal and inter-annual precipitation regime, but remain largely indifferent to rare exceptional rainfall events. The floor of Lake Torrens was lowered more than 2 m by deflation during the last glacial, as inferred from its former surface level preserved in form of a prominent travertine structure aged ~27.4 0.5 ka cal BP (Schmid, 1990). A dune remnant rising out of the deepest part of Lake Frome was dated to ~23.3 0.5 ka cal BP and presents a minimum age for the last regional deflation peak (Callen, 1984). Downwind, the source-bordering Strzelecki dunefield was reactivated throughout the LGM (Fitzsimmons et al., 2007). In conclusion, fine-grained valley-fill aggradation is restricted to geological intervals when regional water levels were low enough to expose the terminal playa lake floors, seasonal/inter-annual water stress (droughts) was severe enough to significantly reduce the plant cover, and strong westerly winds prevailed at least seasonally. When the threshold of any of those is not met, the dust source is switched off, which soon translates into the re-establishment of the erosional capacity of the fluvial system and valleyfill/floodplain incision. The inferred early lead-up to the LGM, with the fluvial system responding to barren erosive slopes by ~28 ka or earlier, compares well with high-resolution palynological records from westerliesdominated New Zealand. According to the tephra marker bed-controlled terrestrial Auckland pollen record (Newnham et al., 2007a) within the same mean zonal pressure belt of westerly airflows as the central Flinders Ranges (Shulmeister et al., 2004; Figure 1), dense forest cover was reduced to its lowest levels between ~29-25.5 ka and 24-20 ka, separated only by a short-lived mid-LGM warming. A major transition from cool climate grass- and shrub land to a podocarp forest is recorded for ~21-19 ka (Newnham et al., 2007a). Independent palynological records (Newnham et al., 2007a; b) and moraine mapping of glacial advances (Suggate and Almond, 2005) from both the North and South Islands support the wider, possibly southern hemispheric relevance of these palaeo-environmental changes which correspond to peak global ice-levels maintained throughout ~30-19 ka (Lambeck et al., 2002).
Conclusions
The lithostratigraphic, chronostratigraphic, geochemical and geophysical data highlight different aspects of the depositional, diagenetic and hydrological history of the layered to laminated sedimentary sequence embedded within the fine-grained valley-fills upstream of the Brachina Gorge in the Flinders Ranges. Collectively, they provide new insights about the palaeoclimate and palaeo-
186
environment of the region throughout the lead-up and peak of the LGM. The results of highresolution particle-size analyses and mineral spectroscopy are consistent with earlier inferences (Williams et al., 2001; Williams and Nitschke, 2005; Haberlah et al., in press) that the bulk of the finegrained material consists of wind-blown dust. Dunefields that extend from the downwind margin of Lake Torrens towards the Ranges are identified as a potential source, holding significant quantities of dust and silt from the deflated playa lake floor and the coalescing alluvial fans. In accordance with the chronostratigraphy, large quantities of proximal dust were deposited into the Brachina catchment in the lead-up to ~32 ka and ~21 ka, two ages correlating closely with the first-order dust peaks in the high-resolution continental distal dust record, interpreted as cold dry events or stadials (Petherick et al., 2008). The ~32 ka dust peak is preserved within the lower sequence, which, according to lithofacies observations, bulk mineralogy and induced magnetic susceptibility, experienced advanced weathering in a comparably well-aerated temperate environment. A separate regional chronostratigraphic study on the Flinders Silts concluded that the interval between ~36-30 ka is marked by calcareous pedogenesis coinciding with fluvial incision. For the interval between ~29-25 ka widespread erosion and reworking were inferred (Haberlah et al., in press). The sediments in the stratigraphic section are interpreted to be reworked and flushed into the pre-existing rock-cut channel from ~28 ka (or earlier). An interval of relative surface stability, lasting for up to ~2 ka prior to ~24 ka, marks the transition from poorly-sorted flood deposition to cyclic upward-fining slackwater deposition, with rapid aggradation rates (~80 cm/ka) throughout the first half of the LGM. Elsewhere within the floodplain, this interval is characterised by a realignment of the drainage pattern, accommodating the ~21 ka dust influx filling in the rugged topography to a point where sediment-charged floods migrated laterally across the width of the valley. A dozen large-scale and numerous smaller flood events are recorded in the form of extensive sheets of suspension fall-out. Their redoximorphic colours reflect a cool, temporarily stagnating weathering environment. The excellent state of preservation of the laminations and the hematite-masked band with undissolved carbonate particles suggests that inundation events and bioturbation were short-lived and limited to the uppermost centimetres of the aggrading sediments. The second half of the LGM is marked by a decrease in sedimentation rates (~20 cm/ka), terminating with an influx of gravel at ~18 ka. Bioturbation and pedogenesis have blurred the laminations of the uppermost metre into more transitional light and dark bands. Termination of fine-grained deposition at ~18 ka is synchronous for all valley-fills across multiple catchments in the Flinders Ranges (Haberlah et al., in press). The fine-grained valley-fill formations in the Flinders Ranges are here interpreted as the fluvial response to glacial aridity-induced dust deposition and devegetation. Consequently, the Brachina slackwater deposit presents a proximal equivalent to regional marine and continental terrestrial
187
distal dust records spanning the LGM (Revel-Rolland et al., 2006; Petherick et al., 2008; 2009). Heightened glacial aridity and dust storms are here shown to have coexisted with large-magnitude floods. As an alternative to linking the termination of the fine-grained aggradational regime to an inferred assumption of higher intensity rainfalls (Williams et al., 2001), fine sediment starvation over the course of Deglacial revegetation of dunes and valley slopes, and rising regional groundwater levels covering the terminal playa lake floors is suggested. More information on the source and season of the precipitation regime controlling vegetation growth and generating floods must be obtained by future research. However, the present study of one continuous sedimentary sequence has confirmed the complex environmental response to climatic variability that characterised the LGM and the times immediately before and after it, as recently outlined by other authors for Africa and Australasia (Gasse et al., 2008; Williams et al., 2009).
(see opposite page) Photo Plate 1) View of the vertical outcrop of stratigraphic section BRA-SD (S 31.33730/E 138.60655, TOP 338.1 m a.s.l.) looking upstream towards the vegetated valley slope. The five lithostratigraphic units (basal unit I, transitional unit II, laminated unit III, pedogenic unit IV, and surface drape V) are indicated and juxtaposed with the stratigraphic type section. On the section drawing, 12 lithofacies are differentiated (see legend).
188
189
(see opposite page) Photo Plate 2) Section details: A) downstream end of stratigraphic section BRA-SD unconformably resting on the Neoproterozoic Brachina Shale Formation; B) detrital pedogenic carbonate nodules, reworked from former Bca-horizons and deposited as discontinuous sheets throughout the basal unit (I); C) large (8x2 cm) piece of charred wood dated to ~32.5 ka cal BP (OZJ904); D) narrow line of wellrounded cobbles and large clasts of spring tufa within the basal unit (I) below the palaeosol; E) vertical in situ calcareous rhizomorphs excavated by recent floods at the base of the stratigraphic section; F) pseudogleyic root cast extending from the transitional unit (II) into the basal unit (I). Stagnating surface water infiltrated the root channels leaching the soluble Fe2+-ions which in contact with trapped soil air precipitated as orange lepidocrocite-dominated zones. The void left by microbial decomposition is filled by material similar in colour to the 470-479 red band; G) tufa precipitation in an evaporating pool of water after a 100-year flood in January 2007 observed in Hookina Creek to the SW. Similar tufa crust protected underlying veneers of organic flotsam from erosion by subsequent flood fluxes, with new material infiltrating into the loose structure.
190
191
(see opposite page) Photo Plate 3) Lithostratigraphic details: A) detail of the central aggradational sequence, indicating the boundaries between the lower lithostratigraphic units (I-II) above the palaeosol and the laminated units (II-III) above the 470-479 red band. The lithofacies that comprise the laminations are deposited as a cyclic sequence of yellow bands, veneers of organic detritus and partially discontinuous sheets of in situ tufa precipitates; B) insular remnant of the red surface drape, here sloping down the bank of Etina Creek, dissected by steep gullies of possibly post-European age; C) gravel exposure close to the present confluence of the Brachina and Etina Creeks with a general direction towards BRA-SD. Reversed directions of inclinations (herringbone pattern) suggest flow directions in and out; D) wet-sieved residua (>200 m) from the palaeosol (506-516) consisting of well-rounded sand-sized quartz particles; E) wet-sieved residua (>200 m) from the basal unit (650660) comprising two intact undissolved carbonate shells of the snails Charopidae sp. (left) and Austropyrgus sp. (Hydrobiidae) (right), today associated with damp vegetation and springs in the catchments of the Flinders Ranges (Glasby et al., 2007); F) wet-sieved residua (>200 m) from the 470-479 red band comprising abundant detrital iron oxide-cemented pseudomorphs; G) wet-sieved residua (>200 m) from the 373-380 yellow band comprising in situ tubular tufa cementation.
192
193
Acknowledgements
We thank CRC LEME (Cooperative Research Centre for Landscape Environments and Mineral Exploration), the Australian Research Council (ARC Environmental Futures Network grant DP0559577) and the Australian Institute of Nuclear Science and Engineering (AINSE Grants 96/192R, 99/001 and 07/160) for generous financial are consistent with in dating the stratigraphic section. The International Association of Sedimentologists (1st IAS 2008 postgraduate grant scheme) and Intellection Pty Ltd. are thanked for covering the costs pioneering the application of QEMSCAN to questions of Quaternary stratigraphy and climate change, the Flinders Ranges National Park authorities for research permits and friendly accommodation, Jayesh Pillarisetty, Frances Williams, Rosi Glasby and Jutta von dem Bussche, for assisting in the field. In particular, we are indebted to Amy Suto from the University of Adelaide for helping to prepare the carbon stable isotope samples and Stewart Fallon from the Australian National University Radiocarbon Dating Laboratory for his work on the labour-intensive radiocarbon samples based on organic veneers, Craig C. Strong from Griffith University for introduction and assistance in the Particle Sizing Laboratory, and Rob Fitzpatrick from the University of Adelaide for kindly providing instrument time on his Bartington Magnetic Susceptibility System.
194
References
Aitken, M.J., 1998. An Introduction to Optical Dating. University Press, Oxford. Bartington Instruments, 2007. MS2 Magnetic Susceptibility System. Bartington Instruments Ltd, Oxford. Baruch, Z., 1994. Responses to drought and flooding in tropical forage grasses. Plant and Soil. 164(1), 87-96. Beckman Coulter, 2002. Multisizer 3 Operator's Manual PN8321681 Rev. B. Beckman Coulter, Inc, Fullerton, California. Blong, R.J., Gillespie, R., 1978. Fluvially transported charcoal gives erroneous deposits. Nature. 271, 739-741.
14
C ages for recent
Bowman, S., 1990. Interpreting the Past: Radiocarbon Dating. University of California Press, Berkeley. Butler, B.E., 1955. A system for the description of soil structure and consistence in the field. Journal of the Australian Institute of Agricultural Science. 1, 231-252. Callen, R.A., 1984. The islands of Lake Frome. Geological Survey of South Australia. Quarterly Geological Notes. 88, 2-8. Cock, B.J., Williams, M.A.J., Adamson, D.A., 1999. Pleistocene Lake Brachina: a preliminary stratigraphy and chronology of lacustrine sediments from the central Flinders Ranges, South Australia. Australian Journal of Earth Sciences. 46, 61-69. Crouvi, O., Amit, R., Enzel, Y., Porat, N., Sandler, A., 2008. Sand dunes as a major proximal dust source for late Pleistocene loess in the Negev Desert, Israel. Quaternary Research. 70 (2), 275282. Duller, G.A.T., 2004. Luminescence dating of Quaternary sediments: recent advances. Journal of Quaternary Science. 19 (2), 183-192. Ehleringer, J.R., Cerling, T.E., Helliker, B.R., 1997. C4 photosynthesis, atmospheric CO2, and climate. Oecologia. 112 (3), 285-299. Eitel, B., Blmel, W.D., Hser, K., Mauz, B., 2001. Dust and loessic alluvial deposits in Northwestern Namibia (Damaraland, Kaokoveld): sedimentology and palaeoclimatic evidence based on luminescence data. Quaternary International. 76, 57-65. Farquhar, G.D., Ehleringer, J.R., Hubick, K.T., 1989. Carbon isotope discrimination and photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology. 40, 503-537. Folk, R.L., 1971. Longitudinal dunes of the northwestern edge of the Simpson Desert, Northern Territory, Australia. 1. Geomorphology and grain size relationships. Sedimentology. 16, 5-54.
195
Folk, R.L., Ward, W.C., 1957. Brazos River bar: a study in the significance of grain size parameters. Journal of Sedimentary Petrology. 27 (1), 3-26. Fifield, K., Chappell, J., Fallon, S.J., 2007. The new radiocarbon AMS system at the ANU. Quaternary International. 167-168, 116. Fine, P., Verosub, K.L., Singer, M.J., 1995. Pedogenic and lithogenic contributions to the magnetic susceptibility record of the Chinese loess/palaeosol sequence. Geophysical Journal International. 122 (1), 97-107. Fink, D., Hotchkis, M., Hua, Q., Jacobsen, G., Smith, A.M., Zoppi, U., Child, D., Mifsud, C., van der Gaast, H., Williams, A., Williams, M., 2004. The ANTARES AMS facility at ANSTO. Nuclear Instruments and Methods in Physics Research Section B. 223-224, 109-115. Finlay, J.C., Kendall, C., 2007. Stable isotope tracing of temporal and spatial variability in organic matter sources to freshwater ecosystems. In: Michener, R., Lajtha, K. (Eds.), Stable isotopes in Ecology and Environmental Science, 2nd Edition. Wiley-Blackwell, pp. 283-333. Fitzsimmons, K.E., Bowler, J.M., Rhodes, E.J., Magee, J.M., 2007. Relationships between desert dunes during the late Quaternary in the Lake Frome region, Strzelecki Desert, Australia. Journal of Quaternary Science. 22 (5), 549-558. Galloway, R.W., 1965. Late Quaternary climates in Australia. Journal of Geology. 73 (4), 603618. Gasse, F., Chali, F., Vincens, A., Williams, M.A.J., Williamson, D., 2008. Climatic patterns in equatorial and southern Africa from 30,000 to 10,000 years ago reconstructed from terrestrial and near-shore proxy data. Quaternary Science Reviews. 27, 2316-2340. Gentilli, J., 1986. Climate. In: Jeans, D.N. (Ed.), Australia - A Geography. The Natural Environment. Sydney University Press, Sydney, pp. 1448. Glasby, P., Williams, M.A.J., McKirdy, D., Symonds, R., Chivas, A.R., 2007. Late Pleistocene environments in the Flinders Ranges, Australia: preliminary evidence from microfossils and stable isotopes. Quaternary Australasia. 24 (2), 19-28. Gottlieb, P., Wilkie, G., Sutherland, D., Ho-Tun, E., Suthers, S., Perera, K., Jenkins, B., Spencer, S., Butcher, A., Rayner, J., 2000. Using quantitative electron microscopy for process mineralogy applications. JOM Journal of the Minerals, Metals and Materials Society. 52 (4), 24-25. Haberlah, D., 2007: Depositional models of late Pleistocene fine-grained valley-fill deposits in the Flinders Ranges, SA. In: Fitzpatrick, R.W., Shand, P. (Eds) 2007, Regolith 2006Consolidation and dispersion of ideas. Cooperative Research Centre for Landscape Environments and Mineral Exploration, Perth, pp. 122-126 Haberlah, D., McTainsh, G.H., submitted. Quantifying aggregation in sediments. Haberlah, D., Williams, M.A.J., Hill, S.M., Halverson, G., Glasby, P., 2007. A terminal Last Glacial Maximum (LGM) loess-derived palaeoflood record from South Australia? Quaternary International. 167-168 Supplement, 150.
196
Haberlah, D., Glasby, P., Williams, M.A.J., Hill, S.M., Williams, F., Rhodes, E.J., Gostin, V., OFlanery, A., Jacobsen, G.E., in press. 'Of droughts and flooding rains'*: an alluvial loess record from central South Australia spanning the last glacial cycle. In: Bishop, P., Pillans, B. (Eds.), Australian Landscapes. Geological Society Special Publications, London. Hartley, A.J., Mather, A.E., Jolley, E., Turner, P., 2005. Climatic controls on alluvial-fan activity, Coastal Cordillera, northern Chile. Geological Society London Special Publications. 251 (1), 95116. Harvey, A.M., Wells, S.G., 1994. Late Pleistocene and Holocene changes in hillslope sediment supply to alluvial fan systems: Zzyzx, CA. In: Millington, A.C., Pye, K. (Eds.), Environmental Change in Drylands. Wiley, New York, pp. 6684. Heckathorn, S.A., McNaughton, S.J., Coleman, J.S., 1999. C4 plants and herbivory. In: Sage, R.F., Monson, K. (Eds.), C4 Plant Biology. Academic Press Ltd, London, pp. 285312. Heine, K., Heine, J.T., 2002. A paleohydrologic reinterpretation of the Homeb Silts, Kuiseb River, central Namib Desert (Namibia) and paleoclimatic implications. Catena. 48, 107-130. Heine, K., Vlkel, J. 2009. Desert flash flood series Slackwater deposits and floodouts in Namibia: their significance for palaeoclimatic reconstructions. Zentralblatt fr Geologie und Palontologie, Teil 1. 3-4, 287-308. Heller, F., Liu, T., 1984. Magnetism of Chinese loess deposits. Geophysical Journal International. 77 (1), 125-141. Hessel, R., 2006. Consequences of hyperconcentrated flow for process-based soil erosion modelling on the Chinese Loess Plateau. Earth Surface Processes and Landforms. 31 (9), 1100-1114. Hua, Q., Jacobsen, G.E., Zoppi, U., Lawson, E.M., Williams, A.A., Smith, A.M., McGann, M.J., 2001. Progress in radiocarbon target preparation at the ANTARES AMS centre. Radiocarbon. 43 (2A), 275282. Huang, C.C., Pang, J., Zha, X., Su, H., Jia, Y., Zhu, Y., 2007. Impact of monsoonal climatic change on Holocene overbank flooding along Sushui River, middle reach of the Yellow River, China. Quaternary Science Reviews. 26 (17-18), 2247-2264. Hugenholtz, C.H., Wolfe, S.A., 2005. Biogeomorphic model of dunefield activation and stabilization on the northern Great Plains. Geomorphology. 70, 5370. Inman, D.L., 1952. Measures of describing the size distribution of sediments. Journal of Sedimentary Petrology. 22 (3), 125-145. Johnson, B.J., Miller, G.H., Magee, J.W., Gagan, M.K., Fogel, M.L., Quay, P.D., 2005. Carbon isotope evidence for an abrupt reduction in grasses coincident with European settlement of Lake Eyre, South Australia. The Holocene. 15 (6), 888-896. Kandler, K., Schtz, L., Deutscher, C., Ebert, M., Hofmann, H., Jckel, S., Jaenicke, R., Knippertz, P., Lieke, K., Massling, A., Petzold, A., Schladitz, A., Weinzierl, B., Wiedensohler, A., Zorn, S.,
197
Weinbruch, S., 2009. Size distribution, mass concentration, chemical and mineralogical composition and derived optical parameters of the boundary layer aerosol at Tinfou, Morocco, during SAMUM 2006. Tellus. 61B, 32-50. Kochel, R.C., Baker, V.R., 1988. Paleoflood analysis using slackwater deposits. In: Baker, V. R., Kochel, R.C. (Eds.), Flood Geomorphology. Wiley, New York, pp. 357-376. Lambeck, K., Yokoyama, Y., Purcell, T., 2002. Into and out of the Last Glacial Maximum: sea-level change during Oxygen Isotope Stages 3 and 2. Quaternary Science Reviews. 21, 343360. Lancaster, N., 1994. Controls on aeolian activity: some new perspectives from the Kelso Dunes, Mojave Desert, California. Journal of Arid Environments. 27, 113-125. Lancaster, N., 2002. How dry was dry? Late Pleistocene palaeoclimates in the Namib Desert. Quaternary Science Reviews. 21 (7), 769-782. Leng, M.J., Lamb, A.L., Heaton, T.H.E., Marshall, J.D., Wolfe, B.B., Jones, M.D., Holmes, J.A., Arrowsmith, C., 2005. Isotopes in lake sediments. In: Leng, M.J. (Ed.), Isotopes in Palaeoenvironmental Research. Springer, Dordrecht, The Netherlands, pp. 148-176. Leys, J., McTainsh, G., Koen, T., Mooney, B., Strong, C., 2005. Testing a statistical curve-fitting procedure for quantifying sediment populations within multi-modal particle-size distributions. Earth Surface Processes and Landforms. 30, 579-590. Liu, W., Huang, Y., An, Z., Clemens, S.C., Li, L., Prell, W.L., Ning, Y., 2005. Summer monsoon intensity controls C4/C3 plant abundance during the last 35 ka in the Chinese Loess Plateau: Carbon isotope evidence from bulk organic matter and individual leaf waxes. Palaeogeography, Palaeoclimatology, Palaeoecology. 220 (3-4), 243-254. Luly, J.G., Jacobson, G., 2000. Two new AMS dates from Lake Frome, arid South Australia. Quaternary Australasia. 18, 29-33. Macdonald, P.D.M., Du, J., 2004. Mixdist: Mixture Distribution Models. R package version 0.5-1. Available at: http://www.math.mcmaster.ca/peter/mix/mix.html [Accessed 30/05/2008]. Maher, B.A., 1998. Magnetic properties of modern soils and Quaternary loessic paleosols: paleoclimatic implications. Palaeogeography, Palaeoclimatology, Palaeoecology. 137 (1-2), 2554. van Maren, D.S., Winterwerp, J.C., Wu, B.S., Zhou, J.J., 2009. Modelling hyperconcentrated flow in the Yellow River. Earth Surface Processes and Landforms. 34 (4), 596-612. Marshall, J.D., Brooks, J.R., Lajtha, K., 2007. Sources of variation in the stable isotope composition of plants. In: Michener, R., Lajtha, K. (Eds.), Stable isotopes in Ecology and Environmental Science. Wiley-Blackwell, Malden, MA, pp. 22-60. McCarthy, D., Rogers, T., Casperson, K. (Eds.), 2006. Floods in South Australia, 1836-2005. Bureau of Meterology South Australian Regional Office, Kent Town, South Australia.
198
McIntyre, D.S., 1976. Subplasticity in Australian soils. 1 description, occurrence, and some properties. Australian Journal of Soil Research. 14, 227-236. Mix, A.C., Bard, E., Schneider, R., 2001. Environmental processes of the ice age: land, oceans, glaciers (EPILOG). Quaternary Science Reviews. 20, 627-657. Newnham, R.M., Lowe, D.J., Giles, T., Alloway, B.V., 2007a. Vegetation and climate of Auckland, New Zealand, since ca. 32 000 cal. yr ago: are consistent with for an extended LGM. Journal of Quaternary Science. 22 (5), 517-534. Newnham, R.M., Vandergoes, M.J., Hendy, C.H., Lowe, D.J., Preusser, F., 2007b. A terrestrial palynological record for the last two glacial cycles from southwestern New Zealand. Quaternary Science Reviews. 26, 517535. Olley, J.M., Caitcheon, G., Murray, A.S., 1998. The distribution of apparent dose as determined by optically stimulated luminescence in small aliquots of fluvial quartz: implications for dating young sediments. Quaternary Geochronology. 17, 1033-1040. Olsson, I., 1986. Radiometric methods. In: Berglund, B.E., Ralska-Jasiewiczowa, M. (Eds.), Handbook of Holocene Palaeoecology and Palaeohydrology. Wiley, Chichester, pp. 273-312. Pedley, M., 2009. Tufas and travertines of the Mediterranean region: a testing ground for freshwater carbonate concepts and developments. Sedimentology. 56 (1), 221-246. Petherick, L.M., McGowan, H.A., Moss, P., 2008. Climate variability during the Last Glacial Maximum in eastern Australia: evidence of two stadials? Journal of Quaternary Science. 23, 787-802. Petherick, L.M., McGowan, H.A., Kamber, B.S., 2009. Reconstructing transport pathways for late Quaternary dust from eastern Australia using the composition of trace elements of long traveled dusts. Geomorphology. 105, 67-79. Quigley, M.C., Sandiford, M., Cupper, M.L., 2007. Distinguishing tectonic from climatic controls on range-front sedimentation. Basin Research. 19 (4), 491-505. Preiss, W.V. (Ed.), 1987. The Adelaide Geosyncline: Late Proterozoic stratigraphy, sedimentation, palaeontology and tectonics. Dept. of Mines and Energy, Geol. of Survey South Australia, Adelaide. R Development Core Team, 2008. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria, Available at: http://www.r-project.org/ [Accessed May 30, 2008]. Revel-Rolland, M., De Deckker, P., Delmonte, B., Hesse, P.P., Magee, J.W., Basile-Doelsch, I., Grousset, F., Bosch, D., 2006. Eastern Australia: a possible source of dust in East Antarctica interglacial ice. Earth and Planetary Science Letters. 249 (1-2), 1-13. Rgner, K., Knabe, K., Roscher, B., Smykatz-Kloss, W., Zller, L., 2004. Alluvial loess in the Central Sinai: occurrence, origin, and palaeoclimatological consideration. Paleoecology of Quaternary Drylands. Lecture Notes in Earth Sciences. 102, 79-99.
199
Rust, B.R., 1972., Pebble orientation in fluvial sediments. Journal of Sedimentary Petrology. 42 (2), 384-388. Sage, R.F., Coleman, J.R., 2001. Effects of low atmospheric CO2 on plants: more than a thing of the past. Trends in Plant Science. 6 (1), 18-24. Schieber, J., Southard, J., Thaisen, K., 2007. Accretion of mudstone beds from migrating floccule ripples. Science. 318, 1760-1763. Schmid, R.M., 1990. Absolute dating of sedimentation on Lake Torrens with spring deposits, South Australia. Hydrobiologia. 197, 305-308. Shulmeister, J., Goodwin, I., Renwick, J., Harle, K., Armand, L., McGlone, M.S., Cook, E., Dodson, J., Hesse, P.P., Mayewski, P., Curran, M., 2004. The Southern Hemisphere westerlies in the Australasian sector over the last glacial cycle: a synthesis. Quaternary International, 118-119, 23-53. Singh, G., Luly, J.G., 1991. Changes in vegetation and seasonal climate since the last full glacial at Lake Frome, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology. 84, 75-86. Srivastava, P., Brook, G.A., Marais, E., 2005. Depositional environment and luminescence chronology of the Hoarusib River Clay Castles sediments, northern Namib Desert, Namibia. Catena. 59 (2), 187-204. Srivastava, P., Brook, G.A., Marais, E., Morthekai, P., Singhvi, A.K., 2006. Depositional environment and OSL chronology of the Homeb silt deposits, Kuiseb River, Namibia. Quaternary Research. 65 (3), 478-491. Suggate, R.P., Almond, P.C., 2005. The Last Glacial Maximum (LGM) in western South Island, New Zealand: implications for the global LGM and MIS 2. Quaternary Science Reviews. 24, 1923 1940. Sun, J. 2002. Provenance of loess material and formation of loess deposits on the Chinese Loess Plateau. Earth and Planetary Science Letters. 203, 845-859. Sun, D., Bloemendal, J., Rea, D.K., Vandenberghe, J., Jiang, F., An, Z., Su, R., 2002. Grain-size distribution function of polymodal sediments in hydraulic and aeolian environments, and numerical partitioning of the sedimentary components. Sedimentary Geology. 152, 263-277. Sutherland, D., 2007. Estimation of mineral grain size using automated mineralogy. Minerals Engineering. 20 (5), 452-460. Tipple, B.J., Pagani, M., 2007. The early origins of terrestrial C4 photosynthesis. Annual Review of Earth and Planetary Sciences. 35, 435-461. Weninger, B., Jris, O., 2008. A 14C age calibration curve for the last 60 ka: the Greenland-Hulu U/Th timescale and its impact on understanding the Middle to Upper Paleolithic transition in Western Eurasia. Journal of Human Evolution. 55 (5), 772-781.
200
Weninger, B., Jris, O., Danzeglocke, U., 2008. CalPal-2007. Cologne Radiocarbon Calibration & Palaeoclimate Research Package. Available at: http://www.calpal.de [Accessed 28/04/2009]. Williams, M.A.J., Nitschke, N., 2005. Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. South Australian Geographical Journal. 104, 25-36. Williams, M.A.J., Prescott, J.R., Chappell, J., Adamson, D., Cock, B., Walker, K., Gell, P., 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International. 83-85, 129-144. Williams, M.A.J., Nitschke, N., Chor, C., 2006. Complex geomorphic response to late Pleistocene climatic changes in the arid Flinders Ranges of South Australia. Gomorphologie: relief, processus, environnement, 4, 249-258. Williams, M., Cook, E., van der Kaars, S., Barrows, T., Shulmeister, J., Kershaw, P., 2009. Glacial and deglacial climatic patterns in Australia and surrounding regions from 35 000 to 10 000 years ago reconstructed from terrestrial and near-shore proxy data. Quaternary Science Reviews. doi:10.1016/j.quascirev.2009.04.020 Zhou, L.P., Oldfield, F., Wintle, A.G., Robinson, S.G., Wang, J.T., 1990. Partly pedogenic origin of magnetic variations in Chinese loess. Nature. 346, 737-739. Zielhofer, C., Faust, D., 2008. Mid- and Late Holocene fluvial chronology of Tunisia. Quaternary Science Reviews. 27 (5-6), 580-588.
4. Conclusions and Outlook

Lo ess an d Flo o d s: (Scio me nescire) Socrates
Ch ap t er 4 (Co n clu sio n s an d Ou t lo o k)
201
Within the scope of this thesis, the fine-grained valley-fill deposits of the Flinders Ranges were first brought into a global context by comparing them with similar aeolian-fluvial sequences or so-called Silts in Namibia, Sinai and the Sahara (Haberlah 2007a), and by relating them to the concept of loess-derived alluvium (Haberlah 2007b; 2008). Key questions concerning their provenance, depositional environment and age were identified and addressed by a regional-scale study involving radiocarbon and luminescence dating, parametric sediment-size analysis, and a number of conventional and emerging geophysical and geochemical techniques. The main finding from the chronostratigraphic study based on 13 sections from three major catchments in the central Flinders Ranges is that the Flinders Silts aggraded throughout the last glacial cycle during intervals of rapid aggradation alternating with intervals of relative surface stability and erosion (Haberlah et al. 2009a). The final, most prominent and best preserved interval of aggradation spans the early lead-up and peak of the Last Glacial Maximum (LGM). Its nearsynchronous onset and termination across most sections suggests that the regional palaeoclimate exercised more control than catchment/site-specific morphologic thresholds and possible neotectonic activity. Additional insights were obtained by comparing and juxtaposing different age proxies and dating methods. Accordingly, radiocarbon ages based on charcoal may not relate to the depositional event, due to its resilience towards erosion and re-deposition within the studied sedimentary environment. The presence of charcoal further reflects the abundance of woody vegetation within the catchment which varied throughout the last glacial cycle, and appears to be largely absent for the intervals of aggradation. Intact carbonate shells of aquatic gastropods appear to be more reliable depositional age proxies, as well as fine organic flotsam where preserved as undisturbed veneers. While optically stimulated luminescence dating is often less precise, with error ranges generally a magnitude larger than those reported for the radiocarbon ages, overall the luminescence age estimates seem to be more accurately reflecting the timing of the last depositional event. However, partial bleaching is common in the studied transport and depositional environment, necessitating careful interpretations of the De distributions by appropriate protocols and statistical age models.
202
The main finding from the sediment-sizing study is that laminated deposits situated in former protected embayments and tributary reaches of some confluences display cyclic sequences of upward-fining trends which are likely to reflect episodic backflooding (Haberlah and McTainsh 2009). Accordingly, these conspicuous sedimentary sequences comprise slackwater deposits, recording flood events charged with suspension load. It proved important to consider aggregation by an original approach involving comparative parametric statistics, because aggregates comprised the coarsest transported sediment fraction reflecting the former fluvial depositional environment best. In contrast, the fully-dispersed sediments were extraordinarily well-sorted and homogenous, with upper size ranges supporting the previously inferred aeolian provenance of the material (Williams and Nitschke 2005). Additional geophysical and geochemical analyses performed on the layered to laminated stratigraphic section BRA-SD provided important detail on the local vegetation, hydrological and weathering history, and improved the reconstruction of the former depositional environment (Haberlah et al. 2009b). Accordingly, at ~24 ka recurring bedload deposition of reworked weathered loess mantles and/or former fine-grained terrace remnants was superseded by suspension fall-out during episodic flood fluxes charged with unweathered proximal dust. Peak aggradation rates coincide with peak continental aridity and aeolian activity, recorded by the isotopic plant composition in the catchment, and elsewhere by the deflation of adjacent terminal playa lakes (e.g. Bowler et al. 1986; Schmid 1990), the reactivation of source-bordering dune fields (e.g. Fitzsimmons et al. 2007), and in continental (e.g. Petherick et al. 2008), marine (Gingele and De Deckker 2005) and Antarctic (e.g. Revel-Rolland et al. 2006) distal dust records. Deposition was episodic, involving numerous smaller and a dozen or more large flood events preserved in form of stacked upwardfining sediment layers and veneers. The mineral spectroscopy results corroborate lithostratigraphic observations in that both indicate ponding surface water, however limited in penetration and duration not to cause the dissolution of underlying fine detrital carbonates and the alteration of hematite-masked sediment sheets. In conclusion, this study extends the record of fine-grained valley-fill deposits in the Flinders Ranges to the last glacial cycle and across multiple catchments. It presents a depositional model that explains the aggradation of a layered to laminated facies embedded in tabular fine-grained floodplain remnants which extend into the gorges throughout intervals of increased aridity, as opposed to their subsequent erosion under the present climate. The depositional model resolves inconsistencies in the previously suggested wetland scenario (Williams et al. 2001); mainly that bioturbation appears to have been limited in most reaches and in particular in the well-preserved
203
laminated facies. However, gleyic colours, precipitation of groundwater calcrete and considerable mottling in reaches directly upstream of narrow gorge entrances not studied here do not preclude a scenario in which low-frequency high-magnitude floods coupled with low evapo-transpiration could result in spatially restricted prolonged ponding. The thesis further draws attention to the complex interaction of aeolian-fluvial processes resulting in widespread deposition of loess-derived alluvium which have not been studied to this extent in the Australian context before. The inferred processes played an important role in shaping the landscape over the last glacial cycle. The Flinders Silts present continuous terrestrial archives throughout intervals of peak aridity, elsewhere marked by deflation and erosion. The regional study on the Flinders Silts can help resolving some ongoing debates on the age, source and nature of similar fine-grained valley-fill remnants in other arid mountainous regions (see Haberlah 2007a). The present results are consistent with recent inferences of the allochthonous aeolian origin of the silt terraces in the drainage lines of the Great Escarpment of Namibia, and in some of the creeks and tributaries cutting through the largely Precambrian basement of the southern Sinai Peninsula of Egypt. Dust is proposed to contribute significantly to the Namib Silts based on geochemical similarities between the allochthonous micritic calcites in the fine-grained valley-fills and the playa lakes and karstic calcrete surfaces of the western Kalahari (Eitel et al. 2001). For the Sinai Silts, Rgner et al. (2004) confirmed Nirs earlier hypothesis (1970) of an aeolian origin by Miocene foraminifera finds, and suggest the Gulf of Suez to the west as a dust source throughout glacial intervals of low sea levels exposing unconsolidated Miocene globigerina marls. While the latest papers from Namibia and Sinai (e.g. Srivastava et al. 2006; Heine and Vlkel 2009; Rgner et al. 2004) agree with the pioneering studies that described the silt terraces as fluvial sedimentary sequences deposited by flood events (Klaer 1962; Ollier 1977), uncertainty remains on how exactly the sediments were deposited, and how to reconcile rapid aggradation rates with peak aridity. This resulted in the contested re-evaluation of existing depositional models and chronostratigraphies (Bourke et al. 2003; Srivastava et al. 2006). The present regional chronology from semi-arid South Australia and its discussion of age proxies and age models can contribute to this debate. The forwarded palaeo-environmental model could help to resolve some of the prima facie incompatible observations, by discussing the complex impact of climate and vegetation changes on both sediment supply and the stability of valley slope mantles in a comparable environment. This study sets the stage for future research linking the erosion of the Flinders Silts with terminal LGM/early Deglacial rapid lacustrine deposition recorded for the adjacent terminal playas Lake Frome and Lake Torrens involving comparative chronostratigraphic, granulometric and mineralogical
204
work (appendix 5.4.2 and 5.4.3). The latest preliminary sediment-size study is consistent with the inferred aeolian transect from west to east (appendix 5.4.4) but requires further work. The tempospatial variability of flood deposition and fluvial gradients needs to be established in more detail and for the catchments extending to the increasingly arid north and humid south to infer the nature and extent of the palaeo-circulation patterns. Finally, aggradation and erosion of the Flinders Silts throughout the first half of the last glacial cycle remain sketchy, and need to be tied in with the stacked palaeosol record in the piedmont plains (Williams 1973; Callen et al. 1983) by additional single grain luminescence dating and geochemical analyses.
References
Bourke, M.C., Child, A., Stokes, S. 2003. Optical age estimates for hyper-arid fluvial deposits at Homeb, Namibia. Quaternary Science Reviews 22, 1099-1103. Bowler, J.M., Huang, Q., Chen, K., Head, J., Yuan, B. 1986. Radiocarbon dating of playa-lake hydrologic changes: Examples from northwestern China and central Australia.
Palaeogeography, Palaeoclimatology, Palaeoecology 54, 241-260. Callen, R.A., Wasson, R.J., Gillespie, R. 1983. Reliability of radiocarbon dating of pedogenic carbonate in the Australian arid zone. Sedimentary Geology 35, 1-14. Eitel, B., Blmel, W.D., Hser, K., Mauz, B. 2001. Dust and loessic alluvial deposits in northwestern Namibia (Damaraland, Kaokoveld): sedimentology and palaeoclimatic evidence based on luminescence data. Quaternary International 76/77, 57-65. Fitzsimmons, K.E., Bowler, J.M., Rhodes, E.J., Magee, J.M. 2007. Relationships between desert dunes during the late Quaternary in the Lake Frome region, Strzelecki Desert, Australia. Journal of Quaternary Science 22 (5), 549-558. Gingele, F.X., De Deckker, P. 2005. Late Quaternary fluctuations of palaeoproductivity in the Murray Canyons area, South Australian continental margin. Palaeogeography, Palaeoclimatology, Palaeoecology 220 (3-4), 361-373. Haberlah, D. 2007a: Depositional models of late Pleistocene fine-grained valley-fill deposits in the Flinders Ranges, SA. In: Fitzpatrick R.W., Shand P. (eds) 2007, Regolith 2006Consolidation and dispersion of ideas. Cooperative Research Centre for Landscape Environments and Mineral Exploration, Perth, 122-126. Haberlah, D. 2007b: A call for Australian loess. Area 39.2, 224-229.
205
Haberlah, D. 2008. Response to Smalleys discussion of A call for Australian loess. Area 40.1, 135136. Haberlah, D., McTainsh, G.H. 2009. Quantifying particle aggregation in sediments. Sedimentology (in revision). Haberlah, D., Glasby, P., Williams, M.A.J., Hill, S.M., Williams, F., Rhodes, E.J., Gostin, V., OFlanery, A., Jacobsen, G.E. 2009a. Of droughts and flooding rains*: an alluvial loess record from central South Australia spanning the last glacial cycle. In: Bishop, P., Pillans, B. (eds) 2009, Australian Landscapes. Geological Society, London, Special Publications, (in press). Haberlah, D., Williams, M.A.J., Halverson, G., Hrstka, T., Butcher, A.R., McTainsh, G.H., Hill, S.M., Glasby, P. 2009b. Loess and floods: high-resolution multi-proxy data of Last Glacial Maximum (LGM) slackwater deposition in the Flinders Ranges, semi-arid South Australia. Quaternary Science Reviews (in revision). Heine, K., Vlkel, J. 2009. Desert flash flood series Slackwater deposits and floodouts in Namibia: their significance for palaeoclimatic reconstructions. Zentralblatt fr Geologie und Palontologie (1) 3-4, 287-308. Klaer, W. 1962. Untersuchungen zur klimagenetischen Geomorphologie in den Hochgebirgen Vorderasiens. Heidelberger geographische Arbeiten 11, Heidelberg, Keysersche
Verlagsbuchhandlung. Nir, D. 1970. Les lacs Quaternaires dans la region de Feiran (Sinai Central). Revue de Gographie Physique et de Gologie Dynamiques (2) 12 (4), 335-346. Ollier, C.D. 1977. Outline geological and geomorphological history of the Central Namib Desert. Madoqua 10(3), 207-212. Petherick, L.M., McGowan, H.A., Moss, P. 2008. Climate variability during the Last Glacial Maximum in eastern Australia: evidence of two stadials? Journal of Quaternary Science 23, 787-802. Revel-Rolland, M., De Deckker, P., Delmonte, B., Hesse, P.P., Magee, J.W., Basile-Doelsch, I., Grousset, F., Bosch, D. 2006. Eastern Australia: a possible source of dust in East Antarctica interglacial ice. Earth and Planetary Science Letters 249 (1-2), 1-13. Rgner, K., Knabe, K., Roscher, B., Smykatz-Kloss, W., Zller, L. 2004. Alluvial loess in the Central Sinai: Occurrence, origin, and palaeoclimatological consideration. Lecture Notes in Earth Sciences 102, 79-100.
206
Schmid, R.M. 1990. Absolute dating of sedimentation on Lake Torrens with spring deposits, South Australia. Hydrobiologia 197, 305-308. Srivastava, P., Brook, G.A., Marais, E., Morthekai, P., Singhvi, A.K. 2006. Depositional environment and OSL chronology of the Homeb silt deposits, Kuiseb River, Namibia. Quaternary Research, 65(3), 478-491. Williams, G.E. 1973. Late Quaternary piedmont sedimentation, soil formation and paleoclimates in arid South Australia. Zeitschrift fr Geomorphologie N.F. 17, 102-125. Williams, M.A.J., Prescott, J.R., Chappell, J., Adamson, D., Cock, B., Walker, K., Gell, P. 2001. The enigma of a late Pleistocene wetland in the Flinders Ranges, South Australia. Quaternary International 83-85, 129-144. Williams, M.A.J., Nitschke, N. 2005. Influence of wind-blown dust on landscape evolution in the Flinders Ranges, South Australia. South Australian Geographical Journal 104, 25-36. (Place holder)
Appendices
Ap p en d ices (5)
207
Instructions
Ap p en d ices (5.1 In st r u ct io n s)
208
The numbering of the appendices following the prefix 5 reflects the chapter and subchapter structure of the main thesis. For example, section (5.2.1) and subsections therein present additional material referenced in the Methods chapter Optically Stimulated Luminescence dating (2.1).
Methods
Ap p en d ices (5.2 Met h o d s)
209
Ap p en d ices (5.2 Met h o d s)
210
Appendix 5.2.1 SampleID fraction quantity method (inm) (ing) 180212 180212 180212 180212 180212 180212 180212 180212 0.003 XRS+TSAC 0.231 XRS+TSAC 0.265 XRS+TSAC 0.42 gammacorr 0.42 XRS+TSAC 0.54 gammacorr 0.54 XRS+TSAC 0.04 XRS+TSAC ? XRS+TSAC ? XRS+TSAC ? XRS+TSAC ? XRS+TSAC ? XRS+TSAC ? XRS+TSAC K
DR, De and OSL Age calculation Table K U U Th Th dose error error error error (in%) (3%) (inppm) (inppm) (inGy/ka) 2.184 0.066 2.562 0.394 10.131 1.308 2.041 0.061 1.562 0.047 1.360 0.041 1.496 0.045 1.709 0.043 1.847 0.055 2.204 0.066 1.733 0.052 1.324 0.040 2.016 0.060 1.905 1.765 1.978 0.057 0.053 0.059 2.060 0.590 0.948 0.491 2.098 0.218 1.669 0.395 2.035 0.223 2.280 0.420 2.393 0.273 2.110 0.569 2.549 0.300 2.013 0.725 2.656 3.175 3.225 0.368 0.285 0.282 10.015 1.964 11.193 1.650 9.738 0.396 6.947 1.313 10.719 0.405 8.360 1.393 9.084 0.904 11.220 1.896 6.265 0.978 14.755 2.426 11.910 1.232 10.560 0.939 8.695 0.927 H2O (in%) 5.88 3.37 3.68 5.05 5.05 6.01 6.01 7.91 5.33 1.94 7.40 3.78 6.63 2.34
211 Depth cosmicray (incm) (Gy/ka) (min/ max) (max/ min) 765 0.100 490 410 85 85 235 240 175 1025 640 40 1620 1330 375 325 325 265 0.128 0.138 0.192 0.192 0.164 0.164 0.175 0.081 0.112 0.201 0.052 0.064 0.145
HK07D1 HK07D2 HK07D4 HK07D5
1.052 0.023 1.212 0.027
HK07L2
HK07L4
HK07M1 180212 M1(bothpop.) HK07M3 180212 M1(bothpop.) HK07M5 180212 M5(bothpop.) BRA07G1 180212 G1(bothpop.) BRA07G2 180212 BRA07G4 180212
Appendix 5.2.1 H2O (inGy/ka) (mean) 3.380 3.240 2.607 2.689 2.440 3.054 2.947 3.289 2.858 2.378 3.371 3.207 3.019 3.332 error
DR, De and OSL Age calculation Table
212 FMMerr. FMM% (MDN)wght.MEAN FMM Pop.1 Pop.1 Pop.1 DR (=CAM) (inGy/ka) 20.29 3.397 0.145 287.44 3.211 2.589 2.565 2.565 3.014 3.014 3.341 2.864 2.864 2.339 2.339 3.415 3.415 3.185 3.185 3.047 3.284 0.208 0.173 0.096 0.096 0.102 0.102 0.110 0.186 0.186 0.104 0.104 0.234 0.234 0.130 0.130 0.103 0.108 384.35 213.43 99.82 99.82 108.86 108.86 100.97 309.90 292.64 61.61 103.07 93.55 52.89 32.86 31.84 5.84 5.84 8.16 8.16 4.46 24.76 18.35 3.59 3.21 7.69 3.85 0.76 0.63 0.63 0.90 0.90 0.92 0.57 0.90 0.54 0.68 0.47 0.74
DR error DR error 10%H2O error DR error 2.5%H2O 5%H2O (inGy/ka) (inGy/ka) (mean) (mean) (mean) (mean) (mean) (mean) (mean) /2) 0.144 3.510 0.150 3.413 0.146 3.234 0.138 3.397 0.145 0.210 0.174 0.052 0.140 0.054 0.149 0.108 0.186 0.106 0.231 0.131 0.102 0.110 3.271 2.641 2.733 2.507 3.125 3.061 3.487 2.951 2.363 3.553 3.255 3.162 3.326 0.212 0.177 0.053 0.144 0.055 0.155 0.114 0.192 0.105 0.244 0.132 0.107 0.109 3.182 2.570 2.689 2.441 3.074 2.979 3.392 2.869 2.299 3.458 3.163 3.074 3.235 0.206 0.172 0.052 0.140 0.054 0.151 0.111 0.186 0.102 0.237 0.129 0.104 0.106 3.017 2.439 2.609 2.320 2.980 2.827 3.219 2.719 2.181 3.282 2.996 2.911 3.068 0.195 0.163 0.050 0.133 0.052 0.143 0.105 0.176 0.097 0.224 0.122 0.098 0.101 3.211 2.589 2.689 2.441 3.064 2.963 0.208 0.173 0.052 0.14 0.054 0.15
3.341 0.1095 2.864 2.339 3.415 3.185 3.047 3.284 0.186 0.104 0.234 0.13 0.103 0.108
Appendix 5.2.1 FMM Pop.2 274.03 49.60 49.60 258.20 258.20 38.86 65.78 200.15 28.18 160.15 118.45 87.79 FMMerr. Pop.2 82.05 3.55 3.55 68.22 68.22 4.27 7.43 40.85 2.27 12.78 12.86 14.49 FMM% Pop.2 FMM Pop.3 0.24 0.37 0.37 0.10 0.10 0.08 0.25 0.10 0.33 0.18 0.43 0.26 180.34 68.94 249.64 162.45 27.57 5.56 37.12 FMMerr. Pop.3 28.33
DR, De and OSL Age calculation Table FMM% Pop.3 0.18 0.09 0.14 0.10 9.45 2.04 0.05 FMM Pop.4 FMMerr. Pop.4 FMM% Pop.4 FMM BIC# 4.18 11.85 11.85 11.12 11.12 1.88 36.91 3.23 58.22 20.26 8.64 12.26 sigmab Minimum Min.Age AgeDe (lower) 0.17 0.12 0.12 0.20 0.20 0.09 0.20 0.15 0.15 0.05 0.05 0.15 265.06 348.27 178.06 44.56 226.72 299.83 155.78 38.82
213 Min.Age (upper) 290.31 381.28 201.35 50.11
86.50
69.76
92.23
50.71 57.09 228.71 15.04 75.98 83.62 43.25
41.80 48.31 207.65 9.03 67.66 73.49 38.30
59.84 67.10 247.25 18.11 83.89 94.02 47.90
Appendix 5.2.1 MAM min. max. FMM Pop.1 (=CAM) 84.63 119.70 82.45 38.92
DR, De and OSL Age calculation Table FMMerr. MIN Pop.1 FMM (inka) 6.98 81.89 12.84 13.48 2.70 sd MAX FMM (inka) 6.76 88.88 sd FMM Pop.2 105.87 19.34 FMMerr. Pop.2 32.48 1.56 FMM Pop.3 FMMerr. Pop.3 FMM Pop.4
214 FMMerr. Pop.4
78.04 108.46 68.79 17.37
66.75 93.38 60.18 15.14
85.47 118.74 77.79 19.54
7.33
80.81 13.22 18.15 19.78 34.83 35.56 28.96 1.35 1.82 2.68 3.22 1.59
87.51 14.31 19.01 21.38 36.53 38.51 31.37 1.41 1.96 2.81 3.48 1.72
28.70
23.15
30.61
36.12
2.97
85.68
22.82
15.18 19.94 97.80 4.40 23.86 27.45 13.17
12.51 16.87 88.80 2.64 21.24 24.12 11.66
17.91 23.43 105.73 5.30 26.34 30.86 14.59
30.23 108.23 125.14 18.04 32.36 30.71 16.11
1.66 11.14 9.62 1.62 1.66 2.73 1.29
11.63 22.97 85.59 8.25 50.28 38.88 26.74
1.33 2.99 17.88 0.87 4.51 4.42 4.50
56.73 52.82 21.65 81.94
10.56 8.85 1.96 12.50
2.8
0.6
105.02 10.82 84.70 17.69 17.34 31.67 29.59 15.90 1.56 1.62 2.63 1.27
113.98 11.72 91.77 19.17 18.77 34.40 32.14 17.24 1.68 1.76 2.85 1.38
Appendix 5.2.1 totalaliquots reliant aliquot# 12 (5);2,6,7,9,11 12 (5);17,18,19,20,23 12 (9);25,26,27,28,31,32,33,35,36 12 (9);37,38,39,40,41,42,45,46,47
DR, De and OSL Age calculation Table selected aliquot# (4);6,7,9,11 (2);18,19 (6);25,26,27,32,33,36 (3);38,41,45
215
12 (8);1,3,7,9,11,13,21,23 12 (8);1,3,7,9,11,13,21,23 12 (12);112 26 26 26 26 26 26 26 26 13 (12);1,2,3,4,5,7,9,16,17,18,22,23 (12);1,2,3,4,5,7,9,16,17,18,22,23 (17);1,2,3,4,5,6,10,11,12,13,14,17,18,19,21,22,26 (17);1,2,3,4,5,6,10,11,12,13,14,17,18,19,21,22,26 (13);3,4,5,9,10,11,12,13,18,19,21,23,24,26 (13);3,4,5,9,10,11,12,13,18,19,21,23,24,26 5 (15);3,4,5,7,8,9,12,13,15,16,19,21,22,23,26 (10);1,2,3,4,5,8,9,11,12,13
(5);1,3,11,13,21 (5);1,3,11,13,21 (11);25,27,29,31,33,35,37,39,43,45,47 (8);1,2,4,5,9,16,17,18 (12);1,2,3,4,5,7,9,16,17,18,22,23 (3);1,12,14 (15);1,4,5,6,10,11,12,13,14,17,18,19,21,22,26 (8);4,5,11,12,18,19,21,26 (13);3,4,5,9,10,11,12,13,18,19,21,24,26 (5);3,13,15,16,21 (12);3,4,7,9,13,15,16,19,21,22,23,26 (5);4,5,8,9,12 (8),1,3,4,10,12,16,18,19
21 (13),1,2,3,4,6,10,12,13,14,16,18,19,20
Appendix 5.2.1 SampleID fraction quantity method (inm) (ing) 180212 ? XRS+TSAC 0.249 gammacorr gammaspec 0.249 gammacorr 0.249 XRS+TSAC 0.506 gammacorr 0.506 gammacorr 0.506 XRS+TSAC 0.228 gammacorr 0.074 0.074 0.134 0.134 gammacorr XRS+TSAC gammacorr XRS+TSAC K
DR, De and OSL Age calculation Table K U U Th Th dose error error error 2.542 0.347 7.495 1.143 2.022 0.061 1.653 1.600 1.681 1.779 1.940 1.959 2.187 2.099 1.930 2.064 1.930 2.064 0.041 0.040 0.042 0.053 0.045 0.048 0.066 0.051 0.046 0.062 0.046 0.062 2.765 2.677 2.509 2.187 2.169 2.412 2.205 2.590 2.335 2.331 2.335 2.331 0.216 0.209 0.225 0.335 0.224 0.245 0.324 0.253 0.232 0.329 0.232 0.329 8.750 8.474 10.304 10.208 0.366 0.354 0.391 1.115 1.199 1.159 1.210 1.332 1.338 1.369 1.266 1.266 error 0.026 0.025 0.026 0.029 0.029 0.030 0.028 0.028 H2O 0.05 2.99 2.99 2.99 2.99 2.50 2.50 2.50 8.38 9.69 9.69 9.37 9.37 9.82 2.98 4.78 3.86 3.31 3.01
216 Depth cosmicray (incm) (Gy/ka) 765 1465 0.100 0.058 620 620 620 620 155 155 155 205 35 35 35 35 715 310 160 110 22 1320 1200 1320 1320 855 855 855 255 85 85 85 85 0.114 0.114 0.114 0.114 0.069 0.069 0.069 0.069
WL07FP0 FP0(bothpop.) WL07FP1 180212 180212 (2ndreading) 180212 180212 WL07FP3 (graveltopped) 125180 125180 125180 125180
11.022 0.390 10.663 0.442 10.419 1.079 10.449 0.451 10.310 8.683 10.310 8.683 0.412 1.091 0.412 1.091
0.179 0.093 0.179 0.093 0.179 0.093 0.189 0.161 0.202 0.202 0.202 0.202 0.104 0.153 0.178 0.187 0.205 0.115 0.192 0.192 0.192 0.192
WL07FP5 WL07FP6(SA) WL07FP6(SG)
125180 125180 125180 125180 180212 180212 180212 180212 125180 125180
WL07G1 G1(bothpop.) WL07G3 WL07G5 G5(bothpop.) WL07G6 CAS063 CAS063 CAS061
0.205 XRS+TSAC 0.143 XRS+TSAC 0.158 XRS+TSAC 0.262 XRS+TSAC ? XRS+TSAC ? XRS+TSAC
1.926 0.058 2.183 0.065 2.149 0.064 2.354 0.070 1.918 1.441 0.05 0.04
1.546 0.446 1.878 0.331 2.795 0.318 2.049 0.344 1.542 1.994 0.372 0.335
10.349 1.494 9.556 1.103 8.113 1.047 10.450 1.147 14.215 9.328 1.25 1.115
609
Appendix 5.2.1 H2O (inGy/ka) 3.166 3.042 3.077 3.019 3.238 3.288 3.439 3.274 3.268 3.116 3.275 3.126 2.804 3.333 3.382 3.510 3.330 2.574 error
DR, De and OSL Age calculation Table
217 FMMerr. FMM% (MDN)wght.MEAN FMM Pop.1 Pop.1 Pop.1 DR 143.61 12.88 0.59 0.128 3.074 0.128 3.074 17.42 0.77 217.69 3.021 0.052 3.021 3.021 3.243 3.243 3.243 3.334 3.259 3.259 3.259 3.259 2.878 2.878 3.297 3.378 3.378 3.489 3.300 2.546 0.052 0.052 0.086 0.086 0.086 0.088 0.092 0.092 0.092 0.092 0.156 0.156 0.129 0.123 0.123 0.129 0.135 0.119 217.69 217.69 126.05 126.05 126.05 83.01 56.57 56.57 55.63 55.63 124.91 79.03 99.21 59.48 76.43 85.81 97.08 17.42 17.42 8.62 8.62 8.62 3.93 5.12 5.12 5.03 5.03 3.36 2.31 6.51 4.41 1.84 3.58 4.97 0.77 0.77 0.69 0.69 0.69 0.91 0.63 0.63 0.83 0.83 0.59 0.64 0.62 0.67 0.54 0.93 0.51
DR error DR error 10%H2O error DR error 2.5%H2O 5%H2O (inGy/ka) (inGy/ka) 0.132 3.067 0.127 2.982 0.124 2.826 0.117 3.074 0.1278 0.052 0.053 0.126 3.053 3.087 3.036 3.238 3.288 3.439 3.488 3.425 3.362 3.425 3.362 3.036 3.351 3.468 3.563 3.360 2.588 0.052 0.053 0.127 0.082 0.088 0.129 0.092 0.061 0.131 0.061 0.131 0.164 0.131 0.126 0.131 0.137 0.121 3.001 3.035 2.952 3.150 3.198 3.344 3.393 3.368 3.272 3.368 3.272 2.952 3.260 3.374 3.467 3.270 2.518 0.051 0.052 0.123 0.079 0.085 0.126 0.089 0.060 0.127 0.060 0.127 0.160 0.127 0.123 0.128 0.133 0.118 2.906 2.938 2.798 2.987 3.033 3.171 3.220 3.262 3.106 3.262 3.106 2.799 3.092 3.202 3.290 3.105 2.388 0.049 0.050 0.117 0.075 0.058 0.119 0.085 0.058 0.121 0.058 0.121 0.151 0.121 0.116 0.121 0.126 0.111 3.021 3.056 2.985 0.051 0.052 0.125
0.082 0.088 0.129 0.086 0.058 0.121 0.058 0.121 0.152 0.130 0.123 0.129 0.136 0.120
3.194 0.0803 3.243 0.0863 3.392 0.127 3.334 0.0875 3.318 3.194 3.322 3.199 2.878 0.059 0.124 0.059 0.124 0.156
3.297 0.1285 3.378 0.123
3.489 0.1285 3.300 0.1345 2.546 0.119
Appendix 5.2.1 FMMerr. FMM% FMM FMM Pop.2 Pop.2 Pop.2 Pop.3 202.11 21.61 0.41 137.53 137.53 137.53 73.71 73.71 73.71 167.61 80.25 80.25 24.12 24.12 69.00 107.59 61.93 132.83 99.98 168.68 74.78 11.63 11.63 11.63 3.93 3.93 3.93 30.19 7.95 7.95 3.62 3.62 4.38 4.21 4.06 15.75 4.38 42.41 4.10 0.23 0.23 0.23 0.31 0.31 0.31 0.09 0.37 0.37 0.09 0.09 0.32 0.27 0.39 0.33 0.39 0.07 0.49 173.63 FMMerr. Pop.3 76.12 76.12 46.07 45.60 28.70 37.35 37.35 4.66 3.62
DR, De and OSL Age calculation Table FMM% Pop.3 FMM Pop.4 0.08 0.08 0.09 0.08 0.07 FMMerr. Pop.4 FMM% Pop.4 FMM BIC#
218 sigmab Minimum Min.Age Min.Age AgeDe (lower) (upper) 7.38 0.10 127.43 113.60 140.25 3.87 3.87 3.87 6.00 6.00 6.00 2.19 11.04 11.04 82.98 82.98 24.98 15.28 76.54 24.42 2.40 1.54 1.25 0.00 0.00 0.00 0.05 0.05 0.05 0.10 0.10 0.10 0.22 0.22 0.00 0.03 0.15 0.15 0.02 0.13 0.06 75.53 65.36 84.65 145.09 119.38 168.52
69.38 47.58 31.97
61.24 41.14 29.61
76.90 53.55 34.30
54.89 60.63 56.26 45.46 71.56 71.56 71.12
47.94 56.03 53.24 39.33 65.43 65.43 64.11
61.93 65.03 59.26 49.11 77.12 77.12 77.26
Appendix 5.2.1 MAM 41.46 48.03 min. 36.96 39.52 max. 45.63 55.78 FMM Pop.1 46.72 72.06
DR, De and OSL Age calculation Table FMMerr. MIN Pop.1 FMM 4.62 46.82 5.90 45.05 44.56 45.31 sd MAX FMM 4.63 50.82 47.33 46.82 49.16 42.20 41.56 39.75 25.78 17.34 18.21 17.05 17.91 24.65 25.56 19.34 18.08 24.61 27.64 31.31 sd 5.02 4.08 4.04 4.64 3.07 2.95 3.10 1.39 1.60 1.79 1.57 1.76 2.05 1.25 1.45 1.50 1.16 1.61 2.25 22.73 1.35 FMM Pop.2 65.75 45.52 FMMerr. Pop.2 7.54 3.93 FMM Pop.3 FMMerr. Pop.3 FMM Pop.4
219 FMMerr. Pop.4
3.89 3.84 4.27 2.84 2.81 2.86 1.29 1.52 1.66 1.50 1.63 1.89 1.15 1.34 1.38 1.08 1.49 2.08
23.29
20.15
26.10
38.87
2.85
38.93 38.33 36.65 23.80 16.52 16.83 16.24 16.55 22.73 23.58 17.86 16.69 22.75 25.54 28.89
20.81 14.60 9.81
18.37 12.63 9.09
23.07 16.43 10.53
24.90 17.36 17.07
1.35 1.65 1.62
50.28 24.63 7.40
9.15 2.54 1.13
23.36
11.48
19.07 18.39 16.65 13.03 21.68 21.68 27.93
16.66 17.00 15.76 11.27 19.83 19.83 25.18
21.52 19.73 17.54 14.08 23.37 23.37 30.35
43.40 23.97 29.37 17.05 23.16 26.00 38.13
2.63 1.17 2.20 1.41 1.10 1.52 2.64
23.97 32.64 18.33 38.08 30.30 51.12 29.37
2.00 1.80 1.38 4.73 1.81 13.02 2.12
16.01 13.83 52.62
1.84 1.22 8.96
Appendix 5.2.1
DR, De and OSL Age calculation Table selected aliquot# (6);2,4,5,10,14,20 (10);2,4,5,10,11,14,17,20,23,26 (3);8,9,10 (3);8,9,10 (3);8,9,10 (6);3,8,9,10,14,15 (6);3,8,9,10,14,15 (6);3,8,9,10,14,15 (9);1,2,3,4,6,7,10,11,12 (6);1,2,5,7,12,13 (6);1,2,5,7,12,13 28grains 28grains 6grains 15grains 20grains 21grains 36grains 7grains
220
totalaliquots reliant aliquot# 26 (13);2,3,4,5,9,10,11,14,17,18,20,23,26 26 (13);2,3,4,5,9,10,11,14,17,18,20,23,26 11 (5);1,8,9,10,11 11 (5);1,8,9,10,11 11 (5);1,8,9,10,11 16 (10);1,2,3,6,8,9,10,13,14,15, 16 (10);1,2,3,6,8,9,10,13,14,15, 16 (10);1,2,3,6,8,9,10,13,14,15, 13 (11);1,2,3,4,6,7,8,10,11,12,13 13 13 5 5 (12);1,2,3,4,5,7,8,9,10,11,12,13 (12);1,2,3,4,5,7,8,9,10,11,12,13 70grains 70grains
500 18grains 18grains 500 35grains 500 78grains 78grains 300 15grains 14 (14);114 14 (12);110,12,13
Appendix 5.2.1.1 SampleID

HK07-D1 HK07-D2 HK07-D4 HK07-D5 HK07-L2 HK07-L4 HK07-M1 HK07-M3 HK07-M5
Cosmic Ray DR contribution Table Density

1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
221 C/denom Bx Bx (C=6072) (B=5.5x104)

0.101 0.129 0.139 0.192 0.164 0.175 0.081 0.112 0.201 0.0063 0.0040 0.0034 0.0007 0.0020 0.0014 0.0085 0.0053 0.0003 0.0134 0.0110 0.0030 0.0063 0.0121 0.0043 0.0101 0.0013 0.0071 0.0021 0.0007 0.0059 0.0026 0.0013 0.0009 0.0002 0.0050 -0.0063 -0.0040 -0.0034 -0.0007
Depth inm
7.65 4.9 4.1 0.85 2.4 1.75 10.25 6.4 0.4
Depthx x+d (x+d) (x) (d=11.69

11.5 7.4 6.2 1.3 3.6 2.6 15.4 9.6 0.6 24.3 20.0 5.5 11.5 22.0 7.8 18.3 2.3 12.8 3.8 1.3 10.7 4.7 2.4 1.7 0.3 9.1 23.1 19.0 17.8 12.9 15.2 14.2 27.0 21.2 12.2 35.9 31.6 17.1 23.1 33.6 19.4 29.9 13.9 24.4 15.4 12.9 22.3 16.3 14.0 13.3 11.9 20.7 195.0 140.1 125.5 73.2 96.7 86.5 253.5 169.1 66.8 409.8 329.9 117.6 195.0 366.2 145.7 301.4 83.5 214.6 99.1 73.2 184.5 108.2 84.2 76.8 64.4 163.0
(x+d)+a x+H (a=75) (H=212)

270.0 215.1 200.5 148.2 171.7 161.5 328.5 244.1 141.8 484.8 404.9 192.6 270.0 441.2 220.7 376.4 158.5 289.6 174.1 148.2 259.5 183.2 159.2 151.8 139.4 238.0 223.5 219.4 218.2 213.3 215.6 214.6 227.4 221.6 212.6 236.3 232.0 217.5 223.5 234.0 219.8 230.3 214.3 224.8 215.8 213.3 222.7 216.7 214.4 213.7 212.3 221.1
denom (GxH)
60342.6 47178.3 43739.7 31602.5 37021.9 34666.7 74697.6 54101.2 30156.8
exp(Bx) Cosmicrays (inGy/ka(JxM))

0.994 0.996 0.997 0.999 0.100 0.128 0.138 0.192 0.164 0.175 0.081 0.112 0.201 0.052 0.064 0.145 0.100 0.058 0.125 0.069 0.179 0.093 0.161 0.192 0.104 0.153 0.178 0.187 0.205 0.115
-0.0020 0.998 -0.0014 0.999 -0.0085 0.992 -0.0053 0.995 -0.0003 1.000 -0.0134 0.987 -0.0110 0.989 -0.0030 0.997 -0.0063 -0.0121 -0.0043 -0.0101 -0.0013 -0.0071 -0.0021 -0.0007 -0.0059 -0.0026 -0.0013 -0.0009 0.994 0.988 0.996 0.990 0.999 0.993 0.998 0.999 0.994 0.997 0.999 0.999
BRA07-G1 16.2 BRA07-G2 13.3 BRA07-G4 3.65 WL07-FP0 WL07-FP0 WL07-FP1 WL07-FP1 WL07-FP3 WL07-FP3 WL07-FP5 WL07-FP6 WL07-G1 WL07-G3 WL07-G5 WL07-G6 7.65 14.65 5.2 12.2 1.55 8.55 2.55 0.85 7.15 3.1 1.6 1.1
114555.3 0.053 93905.4 0.065 41883.3 0.145 60342.6 103227.5 48513.4 86682.4 33966.0 65101.8 37582.3 31602.5 57794.6 39690.6 34140.1 32430.6 29594.7 52619.8 0.101 0.059 0.125 0.070 0.179 0.093 0.162 0.192 0.105 0.153 0.178 0.187 0.205 0.115
CAS06-G3 0.22 CAS06-G1 6.09
-0.0002 1.000 -0.0050 0.995
Loess and Floods: (Placeholder)
Appendices (5.2.1.2 Methods)
222
Loess and Floods:
223
DoseRatecalculationwithAGE99softwareProtocol
InordertorunAge99(Grn1994)onWindowsVista,aDOSenvironmentneedstobeemulated usingthefreeMSDOSemulatorDOSBox(http://www.dosbox.com/).DOSbox.exeisstartedanda virtualdirectoryismountedbyentering: Z:\>MOUNTCC:\age99 Next,theage.exeisstartedbychangingthedirectorytothevirtualdriveandexecutingtheprogram file(assumingthatAge.exeislocatedonC:\age99\Age99.exe): Z:\>C: C:\>Age99.exe A new data file is created (F1), named (F2), and quartz entered as sample material (F2)+(F1). The corrected gamma spectrometry or XRS and TSAC values for U, Th and K are entered into the softwareasdemonstratedbythescreenshotbelow:
1) 2) 3) DosereferstotheequivalentdoseDe(onlyenteredifknown) U,Th,andKContentareleftblankbecausecombinedinDCF TherelativeAlphaEfficiencyamountsto0.050.02forquartzandhastoberesetwithevery iterationoftheprogram(F1followedbyF2) 4) 5) 6) Diameterreferstofullypreparedquartzgrains(forrange,inourexample125180m) LayerRemovedquantifiestheeffectofHCletchingongrains(inourexample9m) Density(quartz=2.65forg/cm3)
Loess and Floods: 7)
224
U, Th, and KSediment refer to the H2Ocontent corrected/ dry values of the gamma spectrometer/laboratorymeasurements
8)
Water/Dry Soil refers to the H2Ocontent of the sediment samples (either measured or assumed)
9) 10)
FOR/IRR.only?tobeansweredwithY(es)forthegammaspectrometerfollowedby Ext.GammareferstotheH2Ocontentcorrectedinsitumeasuredtotalexternalgammarate (convertGy/kaintoGy/abymultiplyingwith100)
11)
CosmicDRreferstothecosmicrayestimateapplyingastandarderrorof10%(convertGy/ka intoGy/abymultiplicationwith100)
The results of DR calculations for varying methodologies and H2Ocontents are presented in appendix5.2.1.
Loess and Floods:
225
Thick Source Alpha Counting (TSAC) Protocol

0B
Preparation for Thick Source -counting (TSAC) consists of measuring the background reading of a clean cell with a zinc sulphide scintillation screen (dull ZnS side up) for long enough to fall below 0.3 counts/ksec (excluding the first 16 ksec from the count). 1) About 50 g of the dry sample are reduced to fine powder. The sample is placed in a clean tungsten carbide vibrating ring mill for 30 sec. The mill is cleaned out by crushed quartz and ethyl alcohol. 2) When the background reading is low enough (usually overnight), the cell containing the ZnSscreen is removed from the counting chamber in subdued light and transferred to the laboratory to be loaded with the sample. 3) Here, after removing the cover disc, the ZnS-screen is covered by ~1 mm of the finely powdered sample, spread and compacted by a spatula. 4) The cell lid is replaced, labelled and dated (on the sticker), noting the ZnS-screen sheet number. 5) The cell is carefully transferred in horizontal position to the counting room and loaded into the chamber. 6) The -counter (DAYBREAK) is reset to zero and the printer switched on, noting sample ID, batch number, date and time at the head of the paper. 7) Counting continues at least until the total number of counts (excluding the first 16 ksec) exceeds 1000. 8) Finally, the loaded sample and the ZnS-screen are disposed off and the cell is cleaned out by 50 % HCl in a stainless steel box, wiping each item separately with tongs and Kimwipe taking care that the aluminium spacing ring is not exposed to the acid more than a few seconds. 9) 10) Each item is rinsed thoroughly under running water and finally cleaned with methanol. The clean and dry cell is transferred back to the laboratory where it is loaded with a new ZnSscreen (dull side up), covered by the spacing ring, retaining ring and the cardboard circle. 11) The screen batch number is recorded on a sticker on the lid and the cell is transferred to the counting room where it is loaded into the counting chamber for background counting. The calculation of U and Th concentrations follows manually on a designated sheet: 1) First, the background count is calculated by dividing the total counts (excluding the first 16 ksec) by time (in ksec).
226
Then, the total counts and total pairs are calculated, excluding the first 16 ksec and the intervals of anomalous counts. These follow from threshold values for both counts and pairs per interval by applying the Chauvenet's criterion to identify any outliers (Taylor, J R 1997. An Introduction to Error Analysis. 2nd edition. Sausolito, California. University Science Books, 166-8). The Chauvenet's Rejection Value (CRV) is calculated by the following equation:
Intervals where either counts or pairs fall below or exceed the threshold values are removed from the final equation. 3) 4) The background is subtracted from the counts less rejects/ ksecs. Pairs less rejects are corrected for accidental pairs which for the Daybreak counting chamber is calculated by the following equation:
5)
Errors are calculated in proportion to those of the total counts less rejects and the total pairs less rejects.
6)
The final pairs and counts and their respective values are entered into the conversion table (appendix 5.2.1.5), taking account of the ZnS-batch efficiency.
7)
The concentrations for Th and U (in ppm) with respective errors are used for the environmental dose rate calculations using AGE99 software (Grn 1994) following the protocol in appendix 5.2.1.2.
Loess and Floods:
227
XRFsamplepreparationProtocol
TwosetsofvialsarewashedforeachXRFsample(onerimmedwithoutlidandtheotherwithlid)by runningwarmwateranddetergentusingabrushforthevialsandsoftclothforthecaps.Bothsets arerinsedundercoldrunningtapwaterandconsequentlythreetimesunderfilteredwater.Clean vialsareplacedon(clean)glassdishesanddriedintheglasswaredryingovenat110.Plasticcaps aredriedonpapertowelsplacedonthetopoftheoven. 1) After~30min,therimmedvialsareremovedfromtheoven,labelledandfilledwith~2g(1 spatula)ofpowderedsamplematerial. 2) 3) Thesamplesaredriedinthesampledryingovenat110foratleast~120min. The dried samples and dry vials with lids are transferred to the balance room after cooling downtoroomtemperature. 4) 1gofsamplematerialisweightedout(tothe4thdigitontheSartoriusscale)andtransferred intotheclean(labelled)vialwiththehelpofacleanvibratingspatula.Spilledmaterialmustbe brushedoffthescaleandthefinalexactmeasurementsarecopied. 5) Consequently, after zeroing the balance 4 g of Xray flux (to the 4th digit) are added to the sample material, again taking note of the exact measurements. The flux consists of lithium salts (64.7 % lithium metaborate, lithium tetraborate 35.3 %) and must be kept in the desiccator. 6) Lids are placed on the sample vials and the sampleflux blend is shaken by a "super mixer" untildistributeduniformly. 7) Thesamplevialsaretransferredtotheanalyticallaboratoryforfusion.
Preparations for the fusion of XRF discs consist of lighting the pilot flame on the fusion machine, turningontheextractionfansandplacingabeakerwith50%HClintothesinknexttoabeakerwith runningcoldwater. 1) OnesampleatatimeistransferredintoacleanPt/Aucruciblebygentletappingoutthefull samplecontentoverblackpaper(toshowanyspill). 2) Thecruciblewiththesampleisplacedintothemetalsupportringoverthemainburnerofthe fusion machine by titanium tongs and rotated at an angle of ~45 to ensure complete and homogeneousmelting,andtobringairbubblestothesurface. 3) 4) Thefusionisoccasionallyagitatedfor45minutes. A matchstick headsized amount of Ammonium Iodide is added to the fusion by a small spatulatolowertheviscosity.
228
After another minute, a clean casting mould is placed on the right burner by pointed tongs andpreheatedfor~2minutes.
6)
Meanwhile, the fusion is swirled occasionally and eventually poured into the centre of the castingmould.
7)
Thecrucibleiscooledonthesilicatriangleandtherightburneristurnedoffandreplacedby theaircoolingjetsfor~2minutes.
8)
The sample disc is labelled on the upper (curved) surface and transferred into a press seal plasticbagwithouttouchingthelowersurface. The crucible is cleaned by transferring it to the water beaker with platinumtipped tongs, sprinklingNa2CO3alongtheinsidewalls,andheatingitabovea2cmhighbluepropaneoxygen flameuntilitmeltsanddissolvesallsampleresidue.Afterlettingitcooldownforaminuteona silicatriangle,itisdroppedbackintothewaterbeaker,transferredwiththeplasticcoatedtongs into50%HClacid.Once,theNa2CO3isfullydissolved,thecrucibleisrinsedunderrunningtab waterandfilteredwateranddriedwithKimwipe. FusedXRFdiscsmustbestoredinthedesiccatoruntilrunontheXRFmachine.Theresultsofall runsarepresentedinappendix5.2.1.6.
Appendix 5.2.1.5 SampleID HK07D1 HK07D2 HK07D4 HK07D5 HK07L2 HK07L4 HK07M1 HK07M3 HK07M5 WL07FP0 WL07FP1 WL07FP3 WL07FP5 WL07FP6 WL07G1 WL07G3 WL07G5 WL07G6 Pairs 0.173 0.171 0.191 0.118 0.142 0.155 0.195 0.109 0.256 0.130 0.174 0.178 0.151 0.148 0.176 0.166 0.141 0.178 error 0.022 0.033 0.028 0.022 0.024 0.015 0.033 0.017 0.042 0.020 0.019 0.018 0.024 0.019 0.025 0.019 0.018 0.020 Counts 7.906 7.135 6.025 5.295 6.762 7.226 7.775 6.332 9.119 6.838 7.398 7.513 7.277 6.970 6.534 6.740 7.465 7.300
U and Th contents TSAC conversion Table error 0.160 0.226 0.162 0.155 0.172 0.111 0.218 0.144 0.264 0.151 0.133 0.126 0.174 0.133 0.160 0.127 0.141 0.131 Batch efficiency 1.16 1.16 1.16 1.16 1.16 1.16 1.15 1.15 1.15 1.15 1.16 1.16 1.16 1.16 1.16 1.15 1.15 1.16 U (inppm) 2.562 2.060 0.948 1.669 2.280 2.393 2.110 2.549 2.013 2.542 2.187 2.205 2.493 2.331 1.546 1.878 2.795 2.049 error 0.394 0.590 0.491 0.395 0.420 0.273 0.569 0.300 0.725 0.347 0.335 0.324 0.418 0.329 0.446 0.331 0.318 0.344 Th (inppm) 10.131 10.015 11.193 6.947 8.360 9.084 11.220 6.265 14.755 7.495 10.208 10.419 8.859 8.683 10.349 9.556 8.113 10.450 error 1.308 1.964 1.650 1.313 1.393 0.904 1.896 0.978 2.426 1.143 1.115 1.079 1.385 1.091 1.494 1.103 1.047 1.147
229
Batcheff.Sq 1.346 1.346 1.346 1.346 1.346 1.346 1.323 1.323 1.323 1.323 1.346 1.346 1.346 1.346 1.346 1.323 1.323 1.346
Appendix 5.2.1.5 Prsxbesq 0.232 0.230 0.257 0.159 0.192 0.208 0.257 0.144 0.338 0.172 0.234 0.239 0.203 0.199 0.237 0.219 0.186 0.240 Ctsxbe 9.171 8.277 6.989 6.142 7.843 8.383 8.941 7.282 10.487 7.864 8.582 8.715 8.441 8.085 7.579 7.751 8.585 8.468 Th 10.131 10.015 11.193 6.947 8.360 9.084 11.220 6.265 14.755 7.495 10.208 10.419 8.859 8.683 10.349 9.556 8.113 10.450 Thx0.483 4.893 4.837 5.406 3.355 4.038 4.388 5.419 3.026 7.127 3.620 4.931 5.033 4.279 4.194 4.999 4.616 3.918 5.047
U and Th contents TSAC conversion Table subtract 4.278 3.439 1.583 2.787 3.806 3.995 3.522 4.256 3.360 4.244 3.651 3.682 4.162 3.891 2.581 3.136 4.667 3.421 U 2.562 2.060 0.948 1.669 2.280 2.393 2.110 2.549 2.013 2.542 2.187 2.205 2.493 2.331 1.546 1.878 2.795 2.049 Cterrxbe 0.185 0.262 0.188 0.179 0.199 0.129 0.250 0.165 0.303 0.174 0.154 0.146 0.202 0.154 0.186 0.146 0.162 0.152 Therr 1.308 1.964 1.650 1.313 1.393 0.904 1.896 0.978 2.426 1.143 1.115 1.079 1.385 1.091 1.494 1.103 1.047 1.147 0.632 0.949 0.797 0.634 0.673 0.437 0.916 0.473 1.172 0.552 0.538 0.521 0.669 0.527 0.721 0.533 0.506 0.554 0.658 0.984 0.819 0.659 0.702 0.455 0.950 0.501 1.210 0.579 0.560 0.541 0.699 0.549 0.745 0.553 0.531 0.574 0.394347744 0.589545634 0.490527122 0.394692277 0.420395352 0.272677961 0.568803113 0.299910959 0.724906069 0.346615219 0.335479792 0.324355149 0.418426874 0.328958312 0.446253502 0.33103226 0.31810412 0.344044047
230
Appendix 5.2.1.6 SampleID HK07D1 HK07D2 HK07D4 HK07D5 HK07L2 HK07L4 HK07M1 HK07M3 HK07M5 WL07FP0 WL07FP1 WL07FP3 WL07FP4 WL07FP6 WL07G1 WL07G1(2nd) WL07G3 WL07G3(2nd) WL07G5 WL07G5(2nd) WL07G6 WL07G6(2nd)
XRF results Table
231
SiO2 Al2O3 Fe2O3T MnO MgO CaO Na2O K2O K error TiO2 P2O5 SO3 LOI Total reproducability (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) (in%) 55.47 13.67 6.06 0.09 2.67 6.36 1.60 2.632 2.184 0.066 0.84 0.13 0.35 0.00 89.86 58.70 12.80 5.63 0.08 2.27 6.39 1.09 2.459 2.041 0.061 0.72 0.14 0.33 0.00 90.60 60.11 10.47 4.50 0.06 1.94 7.86 1.26 1.882 1.562 0.047 0.57 0.09 0.12 0.00 88.86 50.88 10.95 4.72 0.06 2.71 11.69 1.42 1.803 1.496 0.045 0.63 0.10 0.12 0.00 85.07 0.000 58.28 9.96 4.26 0.06 2.43 9.35 1.04 2.226 1.847 0.055 0.60 0.14 0.18 0.00 88.54 57.33 13.52 5.82 0.08 2.86 5.69 1.11 2.656 2.204 0.066 0.76 0.15 0.52 0.00 90.48 81.14 81.39 63.48 66.23 54.81 60.07 50.78 57.41 49.86 48.71 60.43 59.96 60.12 60.36 66.57 67.54 7.27 5.85 11.57 9.71 8.94 10.31 14.58 11.11 11.93 11.66 11.59 11.57 12.08 12.28 13.26 12.80 3.43 2.54 4.96 5.38 4.64 5.36 6.72 5.33 6.28 6.14 4.99 4.92 5.46 5.53 4.35 4.06 0.02 0.03 0.07 0.06 0.06 0.08 0.26 0.09 0.11 0.10 0.06 0.06 0.04 0.03 0.04 0.03 1.25 1.04 2.24 2.18 2.58 2.85 3.78 4.06 3.13 3.12 2.83 2.83 2.81 2.84 2.46 2.37 1.08 2.79 4.90 5.17 11.47 7.19 5.02 5.94 10.17 11.18 5.63 5.89 5.28 4.97 1.85 1.91 0.55 0.53 1.12 1.05 1.33 1.33 1.56 1.55 1.36 1.27 1.40 1.36 1.32 1.30 1.30 1.32 2.088 1.595 2.429 2.436 2.144 2.635 2.639 2.487 2.321 2.289 2.631 2.613 2.589 2.612 2.837 2.812 1.733 1.324 2.016 2.022 1.779 2.187 2.190 2.064 1.926 1.900 2.183 2.168 2.149 2.168 2.354 2.334 0.052 0.040 0.060 0.061 0.053 0.066 0.066 0.062 0.058 0.057 0.066 0.065 0.064 0.065 0.071 0.070 0.53 0.38 0.75 0.78 0.67 0.79 0.83 0.73 0.71 0.69 0.81 0.80 0.79 0.81 0.88 0.88 0.05 0.06 0.12 0.09 0.09 0.12 0.12 0.14 0.12 0.13 0.11 0.11 0.10 0.10 0.07 0.08 0.06 0.05 0.10 0.09 0.30 0.13 4.71 0.14 0.42 0.37 1.30 1.61 0.13 0.12 0.06 0.06 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 97.46 96.23 91.73 93.17 87.03 90.84 91.01 89.00 86.39 85.66 91.77 91.72 90.71 90.94 93.67 93.83
1.398 0.689 0.881 0.889
232
Loess and Floods:
233
ManualDevalueselectionusingAnalyst3.22andRadialPlot1.3Documentation
SampleID Weightedhistograms (allfunctionalgrains/ smallaliquots) HK07D1 Weightedhistograms (selectedpopulation(s)of grains/smallaliquots) Radialplotsautomatic (allfunctionalgrains/ smallaliquots) Radialplots (selectedpopulation(s)out ofallfunctionalaliquots) De/Age Forall (inGy) 276.28 42.49 81.34 12.98 373.54 49.56 116.33 17.18 allfunctionalaliquots HK07D4 allselectedaliquots allfunctionalaliquots finalpopulation 214.73 49.31 82.95 19.84 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 194.90 21.84 75.30 9.83 De/Age selected (inGy) 271.94 42.31 80.06 12.92
allfunctionalaliquots HK07D2
allselectedaliquots
allfunctionalaliquots
finalpopulation
344.79 25.47 107.38 10.55
Loess and Floods:
234
HK07D5
59.61 22.09 23.24 8.66 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 98.85 25.30 32.80 8.47 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 55.18 28.53 16.52 8.56 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 106.09 81.23 37.05 28.47 allfunctionalaliquots allselectedpopulations allfunctionalaliquots finalpopulations
47.14 1.81 18.38 0.99
HK07L2
89.94 7.81 29.85 2.78
HK07L4
94.75 9.99 28.36 3.13
HK07M1
274.49 41.25& 68.61 9.41 95.86 15.69&
Loess and Floods:
235
23.96 3.64
depositionalpopulation
allselectedpopulations
incisionalpopulation HK07M3 248.17 57.74 106.12 25.14 allfunctionalaliquots allselectedpopulations allfunctionalaliquots finalpopulations 189.43 8.80& 296.20 25.29 81.01 5.21& 126.66 12.19
Loess and Floods:
236
dunepopulation HK07M5 109.66 87.96 32.11 25.86 allfunctionalaliquots allselectedpopulations allfunctionalaliquots finalpopulations 64.41 7.13& 17.97 3.92 18.86 2.46& 5.26 1.52
pedogenicpopulation
Loess and Floods:
237
BRA07G1
95.22 23.69 29.90 7.54 allfunctionalaliquots allselectedpopulations allfunctionalaliquots finalpopulations
78.87 14.47& 111.62 10.87 24.76 4.66& 35.04 3.70
inheritedpopulation BRA07G2 100.13 22.71 32.87 7.54 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 92.79 8.88 30.46 3.09
Loess and Floods:
238
BRA07G4
52.64 11.05 16.03 3.41 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 154.37 36.07 50.22 11.92 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation
53.13 4.91 16.18 1.59
CAS071 CAS072 WL07FP0
136.03 19.41& 203.79 26.94 44.25 6.58& 66.30 9.19
inheritedpopulation
Loess and Floods:
239
WL07FP1
158.48 42.59 52.46 14.13 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 76.76 18.33 23.67 5.69 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 83.32 11.48 24.99 3.50 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 61.13 11.97 18.76 3.71 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation
146.44 25.65 48.47 8.53
WL07FP3
126.08 7.05 38.88 2.61
WL07FP5
82.72 8.63 24.81 2.67
WL07FP6 (SA)
57.38 3.51 17.61 1.19
Loess and Floods:
240
WL07FP6 (SG)
42.64 17.93 13.09 5.51 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 85.92 36.20 29.85 12.68 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulations
56.44 6.19 17.32 1.96
WL07G1
70.10 8.10& 119.97 13.63 24.36 3.11& 41.69 5.25
allselectedaliquots
inheritedpopulation
Loess and Floods:
241
WL07G3
75.97 21.85 23.04 6.69 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finalpopulation 65.51 19.34 19.39 5.70 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finaltwopopulations
76.92 7.84 23.33 2.55
WL07G5
60.63 5.48& 84.95 6.85 17.95 1.75& 25.15 2.23
allselectedaliquots
inheritedpopulation
Loess and Floods:
242
WL07G6
57.79 22.75 16.02 6.33 allfunctionalaliquots allselectedaliquots allfunctionalaliquots finaltwopopulations
allselectedaliquots
57.44 9.09& 135.30 26.81 15.92 2.59& 40.05 8.07
inheritedpopulation ForfinalDecalculations,laboratoryspecificsourcecalibrationfactors(0.09369forsamplesfromHK07DandHK07L;0.1414forsamplesfromHK07Mand WL07FP0;0.01430fortheremainingsamplesfromWL07FPandallsamplesfromBRA07G;0.105forsamplesfromWL07G)needtobeapplied(see appendix5.2.1).
Loess and Floods:
243
EquivalentDosesamplepreparationProtocol
The sample is extracted from the black plastic capped steel cylinders (12.5 cm x 6 cm) under controlledredlightconditionsinthedarklaboratory.Theoutwardfacing23cmandinwardfacing1 cmareremovedforimmediatemoisturecontentmeasurementsandsubsequentDRanalysis. 1) 2) Carbonatesaredissolvedin20%HCl(whilestirringandheatingthesamplefor15min). The fine suspension load is flushed by slowly decanting the sample (for a minimum of 5 repetitionsuntilthewaterisclear). 3) Clay aggregates are disintegrated by the addition of ~6 NaOH pellets and ultrasonic bath treatment(whilestirringthesample). 4) The newlyreleasedsuspensionloadis flushed (byaminimumof3repetitionsuntilwater is clear). 5) Organiccomplexesaredisintegratedbytheadditionof30%H2O2(~1cmabovethesample whilestirringandheatinguntilvapourdevelops,followedby~15minrest). 6) The newlyreleasedsuspensionloadis flushed (byaminimumof5repetitionsuntilwater is clear). 7) The sample is rinsed with filtered water, Methanol and Acetone and dried overnight (at ~60C). 8) The sample is sieved into two grain size fractions: 125180 m, and 180212 m (treated separatelyfromhereon). 9) The outer rind of the grains affected by alpha radiation is etched and feldspar minerals are dissolved in 50 % HF (using a plastic beaker, ~1cm above sample, for 40 min while stirring every5min). 10) Thesampleistransferredbacktoaglassbeakerwithfilteredwaterandrepeatedlyrinsedwith 20%HCl(thesecondtimewhilestirringandheatinguntilvapourdevelops,followedby~15 minrest). 11) ThesampleisrinsedwithROwater(3times),MethanolandAcetoneanddriedfor34hrs(at ~60C). 12) HeavymineralsareseparatedbysuspensioninLithiumPolytungstate(withaspecificgravity 2.67,allowing34hrsforsettling). 13) ThesampleisrinsedwithROwater(3times),MethanolandAcetoneanddriedfor30min(at ~60C). 14) Mineral grains with ferrous contaminations that survived the heavy liquid separation are excludedbymagneticseparation.
244
Thefullytreatedsampleisresievedusing90mand150msievesforthe125180mand 180212mfractionsrespectively,removingbrokenquartzandfeldspargrainsthatsurvived theHFtreatment.
16)
The final sample is weighted, packed and labeled and transferred in a dark box to dating laboratorywhereitismountedonstainlesssteeldiscseitherassmallaliquots(~20grains)or singlegrains(100grains)fortheSARprocedure.
Loess and Floods:
245
Finite Mixture Model OSL data Table

Environmental Dose Rates (DR) for variable water contents and corresponding Equivalent Dose estimates (ED):
Sample Code K1890 (HK07-D5) Irradiation data U (ppm) Th (ppm) K (%) 1.669 0.395 6.947 1.313 1.496 0.045 Environmental DR FMM/ CDM ED pop. 63 % 37 % 0.17 4.18 OD BIC Radial Plot
2.5 % 2.507 0.144 99.82 5.84 5.1 % 2.440 0.140 49.60 3.55 10 % 2.320 0.133 AV % 2.565 0.096 -
0.12 11.85
Cosmic DR 192 19.2 (Gy/a) K1889 (HK07-D4) U (ppm) Th (ppm) K (%) 0.948 0.491
2.5 % 2.641 0.177 213.43 31.84 76 %
11.193 1.650 3.7 % 2.607 0.174 274.03 82.05 24 % 1.562 0.047 10 % 2.439 0.163 AV % 2.589 0.173 -
2.5 % 3.240 0.210 384.35 32.86 100 % 0 -
10.015 1.964 3.4 % 3.271 0.212 2.041 0.061 10 % 3.017 0.195 AV % 3.211 0.208 -
2.5 % 3.380 0.144 287.44 20.29 100 % 0 92 % 8% 90 %
10.131 1.308 5.9 % 3.510 0.150 2.184 0.066 10 % 3.234 0.138 AV % 3.397 0.145 2.5 % 3.487 0.114 100.97 4.46 7.9 % 3.289 0.108 38.86 4.27 10 % 3.219 0.105 AV % 3.341 0.110 2.5 % 3.061 0.155 108.86 8.16
Cosmic DR 100 10.0 (Gy/a) K1892 (HK07-L4) U (ppm) Th (ppm) K (%) 2.393 0.273 9.084 0.904 2.204 0.066
0.09 1.88
Cosmic DR 175 17.5 (Gy/a) K1891 (HK07-L2) U (ppm) Th (ppm) K (%) 2.280 0.420 8.360 1.393 1.847 0.055
0.20 11.12
6.0 % 2.947 0.149 258.20 68.22 10 % 10 % 2.827 0.143 AV % 3.014 0.102 2.5 % 3.553 0.244 61.61 3.59 54 % 33 % 0.15 58.22
Cosmic DR 164 16.4 (Gy/a) AdGL-08006 (HK07-M5) U (ppm) Th (ppm) K (%) 2.013 0.725
14.755 2.426 7.4 % 3.371 0.231 28.18 2.27 2.016 0.060
10 % 3.282 0.224 180.34 27.57 9 % AV % 3.415 0.234 9.45 2.04 5%
Cosmic DR 201 20.1 (Gy/a)
Loess and Floods:

AdGL-08005 (HK07-M3) U (ppm) Th (ppm) K (%) 2.549 0.300 6.265 0.978 1.324 0.040

1.9 % 2.378 0.106 292.64 18.35 90 % 2.5 % 2.363 0.105 200.15 40.85 10 % 10 % 2.181 0.097 AV % 2.339 0.104 0.20 36.91 0.15 3.23
246
Cosmic DR 112 11.2 (Gy/a) AdGL-08004 (HK07-M1) U (ppm) Th (ppm) K (%) 2.110 0.569
2.5 % 2.951 0.192 309.90 24.76 57 % 25 %
11.220 1.896 5.3 % 2.858 0.186 65.78 7.43 1.733 0.052
10 % 2.719 0.176 162.45 28.33 18 % AV % 2.864 0.186 2.3 % 3.332 0.110 52.89 3.85 2.5 % 3.326 0.109 87.79 14.49 10 % 3.068 0.101 AV % 3.284 0.108 2.5 % 3.162 0.107 93.55 7.69 74 % 26 % 47 % 0.05 8.64 0.15 12.26
Cosmic DR 81 8.1 (Gy/a) AdGL08004 (BRA07-G4) U (ppm) Th (ppm) K (%) 3.225 0.282 8.695 0.927 1.978 0.059
Cosmic DR 145 14.45 (Gy/a) AdGL08002 (BRA07-G2) U (ppm) Th (ppm) K (%) 3.175 0.285
10.560 0.939 6.6 % 3.019 0.102 118.45 12.86 43 % 1.765 0.053 10 % 2.911 0.098 249.64 37.12 10 % AV % 3.047 0.103 2.5 % 3.255 0.132 103.07 3.21 68 % 0.05 20.26
Cosmic DR 64 6.4 (Gy/a) AdGL08001 (BRA07-G 1) U (ppm) Th (ppm) K (%) 2.656 0.368
11.910 1.232 3.8 % 3.207 0.131 160.15 12.78 18 % 1.905 0.057 10 % 2.996 0.122 68.94 5.56 AV % 3.185 0.130 14 % -
Cosmic DR 52 5.2 (Gy/a) CLW-95/4 U (ppm) 2.331 0.329 8.683 1.091 2.064 0.062
2.5 % 3.362 0.131 55.63 5.03 9.4 % 3.126 0.121 24.12 3.62 10 % 3.106 0.121 76.12 37.35 AV % 3.259 0.092 2.5 % 3.362 0.131 56.57 5.12 9.7 % 3.116 0.121 80.25 7.95 10 % 3.106 0.121 AV % 3.259 0.092 2.5 % 3.488 0.092 83.01 3.93
83 % 9% 8% 63 % 37 % 91 %
0.22 82.98
(WL07-FP6-SG) Th (ppm) K (%)
Cosmic DR 202 20.2/ (Gy/a) 190 19.0 AdGL-08009 U (ppm) 2.331 0.329 8.683 1.091 2.064 0.062
0.1
11.85
(WL07-FP6-SA) Th (ppm) K (%)
Cosmic DR 202 20.2/ (Gy/a) 190 19.0 AdGL-07008 (WL07-FP5) U (ppm) Th (ppm) K (%) 2.590 0.253
0.10 2.19
10.449 0.451 8.4 % 3.274 0.086 167.61 30.19 9 % 2.099 0.051 10 % 3.220 0.085 AV % 3.334 0.088 -
Cosmic DR 189 18.9/ (Gy/a) 160 16.0
Loess and Floods:

AdGL-07007 (WL07-FP3) U (ppm) Th (ppm) K (%) 2.205 0.324

2.5 % 3.439 0.129 126.05 8.62 69 % 31 % 0 3.87 0.05 6.00
247
10.419 1.079 2.5 % 3.439 0.129 73.71 3.93 2.187 0.066 10 % 3.171 0.119 AV % 3.243 0.086 -
Cosmic DR 179 17.9/ (Gy/a) 90 9.0 AdGL-07006 (WL07-FP1) U (ppm) Th (ppm) K (%) 2.187 0.335
2.5 % 3.036 0.127 217.69 17.42 77 %
10.208 1.115 3.0 % 3.019 0.126 137.53 11.63 23 % 1.779 0.053 10 % 2.798 0.117 AV % 3.021 0.052 0.10 7.38
Cosmic DR 114 11.4/ (Gy/a) 70 0.7 AdGL-08007 (WL07-FP0) U (ppm) Th (ppm) K (%) 2.542 0.347 7.495 1.143 2.022 0.061
0.1 % 3.166 0.132 143.61 12.88 59 % 2.5 % 3.067 0.127 202.11 21.61 41 % 10 % 2.826 0.117 AV % 3.074 0.128 2.5 % 3.685 0.139 59.48 4.41 67 %
Cosmic DR 100 10.0/ (Gy/a) 60 0.6 CLW-95/5 (WL07-G6) U (ppm) Th (ppm) K (%) 2.049 0.344
0.15 24.42
10.450 1.147 3.9 % 3.630 0.137 132.83 15.75 33 % 2.354 0.070 10 % 3.402 0.128 AV % 3.608 0.136 2.5 % 3.468 0.126 99.21 6.51 4.8 % 3.382 0.123 61.93 4.06 10 % 3.202 0.116 AV % 3.378 0.123 2.5 % 3.351 0.131 79.03 2.31 3.0 % 3.333 0.130 107.59 4.21 10 % 3.092 0.121 45.60 3.62 AV % 3.297 0.129 2.5 % 3.036 0.164 124.91 3.36 61 % 39 % 64 % 27 % 8% 59 % 32 % 9% 54 % 39 % 0.02 2.40 0 24.98 0.03 15.28 0.15 76.54
Cosmic DR 187 18.7 (Gy/a) CLW-95/3 (WL07-G5) U (ppm) Th (ppm) K (%) 2.795 0.318 8.113 1.047 2.149 0.064
Cosmic DR 178 17.8 (Gy/a) CLW-95/2 (WL07-G3) U (ppm) Th (ppm) K (%) 1.878 0.331 9.556 1.103 2.183 0.065
Cosmic DR 153 15.3 (Gy/a) CLW-95/1 (WL07-G1) U (ppm) Th (ppm) K (%) 1.546 0.446
10.349 1.494 9.8 % 2.804 0.152 69.00 4.38 1.926 0.058 10 % 2.799 0.151 46.07 4.66 AV % 2.878 0.156 2.5 % 3.360 0.137 76.43 1.84 3.3 % 3.330 0.136 99.98 4.38
Cosmic DR 104 10.4 (Gy/a) AdGl07002 (CAS06-1 3) U (ppm) Th (ppm) K (%) 1.542 0.372 14.215 1.25 1.918 0.05
10 % 3.105 0.126 173.63 28.70 7 % AV % 3.300 0.135 -
Loess and Floods:

AdGl07001 (CAS06-1 1) U (ppm) Th (ppm) K (%) 1.994 0.335 9.328 1.115 1.441 0.04

2.5 % 2.588 0.121 97.08 4.97 3.3 % 2.574 0.120 74.78 4.10 10 % 2.388 0.111 AV % 2.546 0.119 51 % 49 % 0.06 -1.25
248
Discrete De-populations are calculated using the Finite Mixture Model (FMM) fmix.s (Galbraith 2005), and numerically listed by their relative proportion in the sample. The number of De-populations and the over-dispersion factor (OD) is largely determined by the lowest Bayesian information criterion value (BIC). Where De -values fall into one discrete population, the Central Dose Model (CDM) cdose.s (Galbraith 2005) is employed. All aliquots/grains are presented as radial plots, i.e. as points with precision values plotted against the x-axis (increasing towards the right), and De-values plotted against the y-axis (bands indicating 2 STs from a given central value or radial line), (Olley & Reed 2003). The De-population interpreted to relate to the last deposition of the sediment is emphasised in black. De-populations consisting of <25 % are indicated in grey and treated as outliers and not converted into ages.
Appendix 5.2.2
Section name Sample code elevation of Sample TOP (in m) material Depth from TOP (in cm) elev. 193 85 410 490 765 elev. 162 265 325 elev. 154 40 637 1075 14C-Age Estimates (14C a BP err) 14C-Age Estimates (a calBP, 1 STD) OSL-Age Estimates (FMM depositional population/ CDM)
Flinders Silts Age Sheet

14C-Age Estimates (a calBP, 2 STD) OSL-Age Estimates (MAM with 1 STD range) OSL-Age Estimates (FMM inherited populations) OSL-Age Estimates (FMM postdepositional populations) d13C (in ppm) H2 O content (in %) Comments
249
HK07-D K1890 K1889 K1888 K1887 a1139869 HK07-L K1892 K1891 HK07-M AdGL-08006 AdGL-08005 AdGL-08004 BRA-SA OZE022 OZC706 OZC705 OZC704 Beta-84140 Beta-84141 BRA-SG Wk-6548 Wk-6554 Wk-6555 Wk-6552 Wk-6550 Wk-6549 Wk-6553 Wk-6556 BRA07-SD SSAMS ANU-4107 SSAMS ANU-4109 SSAMS ANU-4207 AdGL-96005 SSAMS ANU-4206 SSAMS ANU-4110 SSAMS ANU-4111 SSAMS ANU-4112 AdGL-96007 SSAMS ANU-4113 SSAMS ANU-4114 OZJ905 SSAMS ANU-4209 SSAMS ANU-4116 AdGL-96004
(S 31.77733, E 138.31807) Q SA Q SA Q SA Q SA (S 31.80446, E 138.26218) Q SA Q SA (S 31.79532, E 138.25908) Q SA Q SA Q SA ~(S 31.337, E 138.619) S 18,500 180 C 17,300 200 C 18,150 350 S 19,100 180 C 20,320 90 C 20,840 90 (S 31.3388, E 138.611) S 14,827 87 S 15,891 85 S 16,173 89 S 16,172 93 S 16,365 94 S 16,960 100 S 16,928 94 S 17,148 96 (S 31.337339, E 138.60660) OV 15,160 100 C 16,170 120 OV 17,420 120 Q LA OV 18,410 100 OV 18,520 120 OV 18,610 130 OV 19,360 140 Q LA OV 20,130 160 C 18,880 140 S 20,460 140 OV 19,670 120 OV/ C 21,120 150 Q LA
19,340 1,560 82,450 13,480 84,630 6,980
17,370 (15,140 - 19,540) 68,790 (60,180 - 77,790) 108,460 (93,380 - 118,740) 78,040 (66,750 - 85,470)
38,920 2,700 105,870 32,480 119,700 12,840 -
5.1 7.2 4.2 5.5
Hookina Section D below uppermost Bca-horizon chromatic band topping 2 pronounced Bca-horizons chromatic band between 2 pronounced Bca-horizons base of lowermost Bca-horizon Hookina Section L upper band of chromatic "twin marker horizon" lower band of chromatic "twin marker horizon" Hookina Section M uppermost Bca-horizon "orange Sf band" (reworked dune?) onset of Silts, below major cut & fill Brachina Section A (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) Brachina Section G (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) (Williams et al. 2001) Slippery Dip topmost undisturbed organic veneer topmost undisturbed light band base of central light band (Williams et al. 2001) based on TSAC/ XRS DR top of light "5er marker bands" base of light "5er marker bands" top of continuous "tufa band" top of light "4er marker bands" (Williams et al. 2001) based on TSAC/ XRS DR topping central light band of "3er marker bands" central light band of "3er marker bands" from monolith tufa capping lower light band of "3er marker bands" top of "pink band" (Williams et al. 2001) based on TSAC/ XRS DR
30,230 1,660 36,120 2,970
15,180 (12,510 - 17,910) 28,700 (23,150 - 30,610)
85,680 22,820
11,630 1,330 6.5 5.5
18,040 1,620 85,590 17,880 108,230 11,140
4,400 (2,640 - 5,300) 97,800 (88,800 - 105,730) 19,940 (16,870 - 23,430)
52,820 8,850 125,140 9,620 -
8,250 0,870 6.2 3.5 22,970 2,990 5.2 BRA-SA
405 440 440 440 550 600 elev. 344 20 55 72 78 93 117 150 165 elev. 340 51 94 116 ~160 198 213 244 264 ~325 338 341 ~353 372 437 ~475
22,170 310 20,790 240 21,860 450 22,980 210 24,290 180 24,830 80
21,550 - 22,790 20,310 - 21,270 20,960 - 22,760 22,560 - 23,400 23,930 - 24,650 24,670 - 24,990
-4.74 -25 -25 0 -
18,100 240 18,970 110 19,340 160 19,340 160 19,620 160 20,390 110 20,360 100 20,600 120
17,620 - 18,580 18,750 - 19,190 19,020 - 19,660 19,020 - 19,660 19,300 - 19,940 20,170 - 20,610 20,160 - 20,560 20,360 - 20,840
-6.6 0.2 -8.1 0.2 -7.6 0.2 -7.4 0.2 -8.3 0.2 -9.1 0.2 -6.8 0.2 -7.7 0.2
18,280 230 19,340 200 20,920 170 18,000 1,200 22,110 250 22,230 260 22,410 180 23,230 150 22,400 2,300 24,080 240 22,680 150 24,450 190 23,540 120 25,160 240 21,600 1,200
17,820 - 18,740 18,940 - 19,740 20,580 - 21,260 21,610 - 22,610 21,710 - 22,750 22,050 - 22,770 22,930 - 23,530 24,560 - 23,600 22,380 - 22,980 24,070 - 24,830 23,300 - 23,780 24,680 - 25,640 -
-26 3 -26 3 -22 3 5.3 -20 3 -19 3 -14 3 -27 3 23.7 -25 3 -22 3 -7.4 -18 3 -29 3 9.5
Appendix 5.2.2
Section name Sample code elevation of TOP (in m) Depth from TOP (in cm) 474 495 505 505 508 516 516 ~565 ~565 ~600 ~635 632 639 667 684 695 695 ~830 ~830 elev. 341 ~120 ~145 ~195 255 285 ~455 elev. 333 1150 1230 1265 1265 1275 1300 elev. 316 15 52 elev. 300 155 253 338 380 380 508 Sample 14C-Age material Estimates (14C a BP err) C 21,890 160 C 24,110 210 OV 19,820 180 OV 19,910 180 S 21,510 310 C 27,630 280 C 27,740 290 S 28,120 160 C 29.,80 180 Q LA Q LA C 29,160 380 S 27,990 710 S 27,320 280 S 27,200 500 C 23,600 300 S 23,270 90 Q LA coarse Q LA ~(S 31.33935, E 138.60737) S 16,150 80 S 16,200 60 S 17,070 70 S 17,450 350 S 17,650 140 S 20,650 80 ~(S 31.33593, E 138.60231) C 26,430 350 C 25,330 310 C 27,100 260 C 27,710 220 C 13,715 98 C 16,380 120 ~(S 31.3334, E 138.594) T 11,650 110 OC 15,780 140 (S 31.32897, E 138.58563) S 15,550 130 S 16,280 130 S 15,750 120 S 14,470 60 C 15,210 220 C 16,820 180 14C-Age Estimates (a calBP, 1 STD) OSL-Age Estimates (FMM depositional population/ CDM) 26,260 310 28,990 390 23,710 200 23,830 250 25,660 460 32,210 260 32,310 290 32,570 260 33,560 300 32,800 2,800 24,900 1,400 33,570 410 32,640 630 31,970 200 31,900 390 28,560 420 28,080 80 24,000 1,600 24,900 2,500

14C-Age Estimates (a calBP, 2 STD) OSL-Age Estimates (MAM with 1 STD range) 25,640 - 26,880 28,210 - 29,770 23,310 - 24,110 23,330 - 24,330 24,740 - 26,580 31,690 - 32,730 31,730 - 32,890 32,050 - 33,090 32,960 - 34,160 32,750 - 34,390 31,380 - 33,900 31,570 - 32,370 31,120 - 32,680 27,720 - 29,400 27,920 - 28,240 OSL-Age Estimates (FMM inherited populations) OSL-Age Estimates (FMM postdepositional populations) d13C (in ppm) H2 O content (in %) -28 2 -26 3 -15 3 -21 3 -24 2 -29 2 -6 -25.8 18.3 17.4 -24.5 -7.1 -27 2 -23 2 19.4 19.4 Comments
250
SSAMS ANU-2030 SSAMS ANU-2035 SSAMS ANU-4117 SSAMS ANU-4205 OZJ909 SSAMS ANU-2036 SSAMS ANU-2037 Beta-96166 Beta-96679 AdGL-96003 AdGL-96006 OZJ904 OZJ908 OZJ907 OZJ906 SSAMS ANU-2039 SSAMS ANU-1811 AdGL-96008 AdGL-96002 Southern End (Section D) Beta-96169 Beta-96168 Beta-96167 OZC710 OZC709 Beta-96170 BRA-LW Wk-6558 Wk-6562 OZC707 Beta-96171 Wk-6561 Wk-6564 BRA-ABC OZE086 Wk-7295 BRA07-AR OZK516 OZK002 OZK003 SSAMS ANU-1812 SSAMS ANU-2038 OZK517
"pink band" base of lowermost light band lowermost organic veneer lowermost organic veneer top of "palaeosol" base of "palaeosol" independent run of SSAMS ANU-2036 (Cock et al. 1999) - Section C (Cock et al. 1999) - Section C (Williams et al. 2001) based on TSAC/ XRS DR (Williams et al. 2001) based on TSAC/ XRS DR large piece of charcoal (8 x 2 cm) Lower Unit lighter band Lower Unit darker band Lower Unit darker band with large calc. rhizocretes present base of Lower Unit present base of Lower Unit independent run of AdGL-96002 on coarser fraction (Williams et al. 2001) based on TSAC/ XRS DR
19,300 140 19,380 120 20,500 90 20,930 390 21,160 170 24,680 80
19,020 - 19,580 19,140 - 19,620 20,320 - 20,680 20,150 - 21,710 20,820 - 21,500 24,520 - 24,840
-8.2 -9.1 -7.9 0 0 -9.6
(Cock et al. 1999), ~250 m downstream of BRA07-SD (Cock et al. 1999), ~250 m downstream of BRA07-SD (Cock et al. 1999), ~250 m downstream of BRA07-SD (Cock et al. 1999), ~250 m downstream of BRA07-SD Lubra Waterhole disturbed Silts towards base of 1350 cm section disturbed Silts towards base of 1350 cm section disturbed Silts towards base of 1350 cm section disturbed Silts towards base of 1350 cm section disturbed Silts towards base of 1350 cm section disturbed Silts towards base of 1350 cm section
31,210 380 30,240 290 31,820 160 32,250 230 16,940 70 19,650 190
30,450 - 31,970 29,660 - 30,820 31,500 - 32,140 32,710 - 31,790 16,800 - 17,080 19,270 - 20,030
-26.0 0.2 -23.1 0.2 -25 -24.9 -24.4 0.2 -23.0 0.2
13,540 130 18,910 140
13,280 - 13,800 18,630 - 19,190
18,710 70 19,500 210 18,860 110 17,720 60 18,310 270 20,210 220
18,570 - 18,850 19,080 - 19,920 18,640 - 19,080 17,600 - 17,840 17,770 - 18,850 19,770 - 20,650
Brachina ABC Narrows 0 topmost tufa cap -23.8 0.2 30 cm black organic clay band below tufa (Williams et al. 2001) Brachina AR -9.3 uppermost gravel-Silts couplet -7.6 gravel-Silts couplet -8.5 gravel-Silts couplet -17 1 lowermost gravel-Silts couplet -23 2 lowermost gravel-Silts couplet -24.9 clast-supported basal gravel (up to ~25 cm)
Appendix 5.2.2
Section name Sample code elevation of Sample TOP (in m) material Depth from TOP (in cm) elev. 270 375 1350 1400 1620 14C-Age Estimates (14C a BP err) 14C-Age Estimates (a calBP, 1 STD) OSL-Age Estimates (FMM depositional population/ CDM) 16,110 1,290 30,710 2,730 28,900 410 34,630 400 32,360 1,660

14C-Age Estimates (a calBP, 2 STD) OSL-Age Estimates (MAM with 1 STD range) OSL-Age Estimates (FMM inherited populations) OSL-Age Estimates (FMM postdepositional populations) d13C (in ppm) H2 O content (in %) 3.7 5.8 4.4 2 Comments
251
BRA07-G AdGL-08003 AdGL-08002 ANU BG42 ANU BG27 AdGL-08001
(S 31.34134, E 138.56968) Q SA Q SA S 24,000 240 S 30,400 480 Q SA
13,170 (11,660 - 14,590) 27,450 ( 24,120 - 30,860) 28,080 - 29,720 33,830 - 35,430 23,860 (21,240 - 26,340)
26,740 4,500 38,880 4,420
50,280 4,510
1,650 1,960
Brachina 18M Section towards base of final wedge of local colluvium below onset of multiple gravel sheets (Williams et al. 2001) (Williams et al. 2001) lowermost exposed band of Silts
WL08-UFP Wk23464 Wk23465 Wk23524 Wk23467 Wk23525 WL07-FP CLW-95/4 AdGL-07009 SSAMS ANU-1813 AdGL-07008 OZK014 SSAMS ANU-2033 SSAMS ANU-2032 OZK012 OZK013 OZK011 AdGL-07007 OZK010 AdGL-07006 AdGL-08007 WL07-S SSAMS ANU-2001 SSAMS ANU-2227 SSAMS ANU-2229 WL07-G CLW-95/5 CLW-95/3 CLW-95/2 OZK019 OZK020 CLW-95/1
elev. 399 20 75 733 780 853 elev. 382 85 85 240 255 268 500 520 745 750 760 855 880 1320 1467 elev. 376 85 235 275 elev. 367 107 156 310 360 370 715
(S 31.27082, E 138.85221) S 7,390 45 C 13,021 52 S 25,084 176 S 19,034 127 S 31,398 367 (S 31.26960, E 138.86783) Q SG Q SA S 18,500 60 Q SA S 18,610 180 C 24,440 200 C 22,800 170 C 32,760 430 C 32,210 400 S 29,300 300 Q SA C 36,630 580 Q SA Q SA (S 31.26887, E 138.88176) C/ OV 29,590 370 OV 21,410 120 OV 24,520 190 (S 31.27243, E 138.87921) Q SG Q SG Q SG C 19,800 160 C 19,760 210 Q SG
8,240 60 15,640 100 30,010 160 22,930 180 35,300 390
8,120 - 8,360 15,440 - 15,840 29,690 - 30,330 22,570 - 23,290 34,520 - 36,080
-25.5 0.2 -7.1 0.2
17,070 1,620 17,360 1,650 22,350 110 24,900 1,350 22,310 310 29,290 360 27,500 360 37,120 840 36,650 930 33,700 350 38,870 2,850 41,690 410 45,520 3,930 46,720 4,620
9,810 (9,090 - 10,530) 14,600 (12,630 - 16,430) 22,130 - 22,570 20,810 (18,370 - 23,070) 21,690 - 22,930 28,570 - 30,010 26,780 - 28,220 35,440 - 38,800 34,790 - 38,510 33,000 - 34,400 23,290 (20,150 - 26,100) 40,870 - 42,510 48,030 (39,520 - 55,780) 41,460 (36,960 - 45,630)
23,360 11,480 24,630 2,540 50,280 9,150
7,400 1.130 -
72,060 5,900 65,750 7,540
7.2 7.3 -6 1 6.7 -5.7 -26 2 -23 2 -24.7 -25.5 -2.2 22,730 1,350 3.8 -23.8 4.0 2.5
Wilkawillina Upper FP uppermost "red drape" towards base of Silts-capping gravel band within "chromatic band" within incipient Bca-horizon below "chromatic band" pocket of intact shells below 2nd Bca-horizon in lowermost "chromatic band" Wilkawillina Floodplain 2-3cm below "red drape" within incipient Bca-horizon 2-3cm below "red drape" within incipient Bca-horizon upper of red "twin Sf marker bands" between red "twin Sf marker bands" lower of red "twin Sf marker bands" onset of Silts above widespread gravel sheet onset of Silts above widespread gravel sheet massive Silts below multiple gravel bands massive Silts below multiple gravel bands massive Silts below multiple gravel bands massive Silts below gravel drape of lower section massive Silts below gravel drape of lower section lowermost Bca-horizon onset of Silts aggradation below calctrete pans Wilkawillina Slackwater topmost organic veneer above more "chromatic Silts" lowermost gravel-Silts couplet lowermost organic veneer Wilkawillina Gorge distinct light band sandwiched by organic veneers termination of massive disturbed Silts below gravel light Silts with organic veneers onset of light disurbed Silts onset of light disurbed Silts second brown Silts band from base incorporating platy gravel and kankar
33,890 360 25,440 230 29,410 300
33,170 - 34,610 24,980 - 25,900 28,810 - 30,010
-25 2 -15.6 -14.7
17,050 1,410 18,330 1,380 23,970 1,170 23,670 170 23,660 220 23,970 2,000
13,030 (11,270 - 14,080) 16,650 (15,760 - 17,540) 18,390 (17,000 - 19,730) 23,330 - 24,010 23,220 - 24,100 19,070 (16,660 - 21,520)
38,080 4,730 29,370 2,200 32,640 1,800
43,400 2,630
4.4 4.9 13,830 1,220 4.0 -25.8 -24.6 16,010 1,840 7.4
Appendix 5.2.2
Section name Sample code elevation of Sample TOP (in m) material Depth from TOP (in cm) elev. 332 22 75 225 265 300 382 477 609 14C-Age Estimates (14C a BP err)

14C-Age Estimates 14C-Age Estimates (a calBP, 1 STD) (a calBP, 2 STD) OSL-Age Estimates OSL-Age Estimates (FMM depositional (MAM with 1 STD range) population/ CDM) (S 31.15334, E 138.56692)/ (S 31.15365, E 138.57011) Q SA 23,160 1,100 21,680 (19,830 - 23,370) C 15,080 110 18,250 230 17,790 - 18,710 C 21,560 250 25,670 400 24,870 - 26,470 C 22,590 260 27,310 400 26,510 - 28,110 C 3,940 190 4,400 280 3,840 - 4,960 C 10,370 200 12,160 350 11,460 - 12,860 C 12,790 180 15,310 270 14,770 - 15,850 Q SA 29,370 2,120 27,930 (25,180 - 30,350) OSL-Age Estimates (FMM inherited populations) OSL-Age Estimates (FMM postdepositional populations) d13C (in ppm) H2 O content (in %) 4.2 -25 -25.4 -24 -25 -25 -25 4.0 Comments
252
CAS06-1 AdGL07002 OZJ893 OZJ895 OZJ897 OZJ892 OZJ894 OZJ896 AdGL07001 Other OZE026 OZE025 OZC713 OZJ902 OZE021 OZE023 OZE024
30,300 1,810
38,130 2,640
Cascades 1 TOP without Bca-horizon <100 g! run on ANTARES accelerator run on STAR accelerator run on STAR accelerator <100 g! run on ANTARES accelerator <100 g! run on ANTARES <100 g! run on ANTARES disturbed Silts at base of section Other between eroded burial and palaeosol between eroded burial and palaeosol dune site on Hawker/ Parachilna road (palaeosol) termination of Silts capped by tufa bench at CAS06-2 from 10 m terrace sampled 1998 Price Creek bank section, 5.8 m terrace Bunyeroo Creek, 10 m terrace, red-brown clay
70 70 TOP ~700 420 850
C S E S S C C
300 60 3,450 70 40,300 3,300 14,620 140 33,250 1,050 27,900 300 28,900 500
390 70 3,730 90 44,470 2,870 17,770 90 38,170 1,800 32,430 320 33,330 520
530 - 250 3,910 - 3,550 38,730 - 50,210 17,590 - 17,950 34,570 - 41,770 31,790 - 33,070 32,290 - 34,370
-9.84 -1.65 0 -8.2 -8.59 -26.05 -25.0
Tab.1)FlindersSiltsAgeSheetlistingallageestimateswithinstratigraphicorderintypesections.Samplematerialisindicated; Q quartz(SA small aliquots,SGsinglegrains,LAlargealiquots),Ccharcoal,Scarbonateshellsofaquaticgastropods, OVveneersoforganicdetritus,OCorganic clay,Ttufa,Eemueggshell.AllradiocarbonagesarecalibratedusingtheCalPal2007Hulucalibrationdataset(WeningerandJris2008)aspartof theCalPal2007calibrationandpalaeoclimateresearchsoftwarepackage(Weningeretal.2008).TheOSLageestimatesarecalculatedusingtheFinite MixtureModel(FMM)ofGalbraith(2005).Thediscrete"agepopulations"areinterpretedas"depositional","inherited"and"postdepositional". Wherevertheequivalentdoseratesarerestrictedtoasinglepopulation,theCentralDoseModel( CDM)isemployed(Galbraith2005).Forcomparison,
Appendix 5.2.3
A B C
Partially-constrained Component Population Analysis (CPA)

D E F G H I J K
253
CPAscreenshot
SampleID
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
relativeprop. CPmodes (in%) (inm)

0.0 4.2 22.8 66.3 6.7 15.3 69.3 15.4 0.0 0.0 0.2 9.1 23.7 64.5 2.4 16.1 69.6 14.3 0.0 0.0 0.5 12.0 20.6 66.9 21.5 56.0 22.5 0.0 0.0 4.4 22.1 60.0 13.5 3 13 48 74 167 3 13 48 74 167 3 16 44 74 176 3 16 44 74 176 3 16 46 59 3 16 46 59 3 12 51 75 170
pi
0.0000 0.0424 0.2277 0.6627 0.0672 0.1530 0.6930 0.1540 0.0000 0.0000 0.0017 0.0913 0.2374 0.6452 0.0244 0.1610 0.6961 0.1429 0.0000 0.0000 0.0053 0.1196 0.2057 0.6694 0.2150 0.5600 0.2250 0.0000 0.0000 0.0442 0.2211 0.5995 0.1352
mu(log10)
0.4771 1.1292 1.6812 1.8715 2.2227 0.4771 1.1292 1.6812 1.8715 2.2227 0.4771 1.2133 1.6435 1.8680 2.2463 0.4771 1.2133 1.6435 1.8680 2.2463 0.4771 1.2172 1.6628 1.7743 0.4771 1.2172 1.6628 1.7743 0.4771 1.0736 1.7076 1.8765 2.2292
sigma
0.2142 0.3050 0.2436 0.2367 0.1345 0.2000 0.3150 0.1350 0.0000 0.0000 0.0997 0.2941 0.1912 0.2080 0.0830 0.3550 0.3300 0.1400 0.0000 0.0000 0.1095 0.3144 0.1750 0.2146 0.2950 0.3425 0.1671 0.0000 0.1925 0.2640 0.2837 0.2180 0.1546
fixpi
FALSE FALSE FALSE FALSE FALSE TRUE FALSE TRUE FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE FALSE
fixmu
TRUE TRUE TRUE FALSE TRUE TRUE FALSE TRUE TRUE TRUE TRUE FALSE FALSE TRUE FALSE TRUE TRUE TRUE TRUE FALSE TRUE TRUE TRUE TRUE TRUE TRUE FALSE TRUE
fixsigma Mixdistparameters
FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE mix(AR300_MD,mixparam(
BRA07300MD
BRA07300FD
mix(AR300_FD,mixparam(
BRA07303MD
mix(AR303_MD,mixparam(
BRA07303FD
BRA07304MD
BRA07304FD
BRA07305MD
Appendix 5.2.3
A 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 B C

D E F G H I J K
254
BRA07305FD
BRA07306MD
6.5 57.7 35.7 0.0 0.0 0.5 10.2 10.2 79.1 12.3 52.0 35.7 0.0 0.0 5.1 21.1 55.8 18.0 7.3 60.5 32.2 0.0 0.0 2.1 20.2 42.3 30.1 5.4 16.0 41.8 42.2 0.0 0.0
3 12 51 75 170 4 16 47 60 4 16 47 60 3 12 51 69 141 3 12 51 69 141 4 18 60 61 194 4 18 60 61 194
0.0655 0.5774 0.3572 0.0000 0.0000 0.0046 0.1025 0.1022 0.7907 0.1230 0.5200 0.3570 0.0000 0.0000 0.0511 0.2111 0.5575 0.1803 0.0726 0.6050 0.3224 0.0000 0.0000 0.0205 0.2020 0.4226 0.3011 0.0539 0.1600 0.4182 0.4218 0.0000 0.0000
0.4771 1.0736 1.7076 1.8765 2.2292 0.6021 1.2153 1.6677 1.7790 0.6021 1.2153 1.6677 1.7790 0.4771 1.0717 1.7034 1.8405 2.1487 0.4771 1.0717 1.7034 1.8405 2.1487 0.6021 1.2660 1.7776 1.7825 2.2877 0.6021 1.2669 1.7776 1.7825 2.2877
0.1500 0.3150 0.2050 0.0000 0.0000 0.1107 0.2646 0.1559 0.2644 0.2800 0.2900 0.1850 0.0000 0.1880 0.2816 0.3099 0.2322 0.1793 0.1450 0.3242 0.1950 0.0000 0.0000 0.1769 0.3310 0.2638 0.2627 0.0936 0.2800 0.3500 0.1600 0.0000 0.0000
FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE FALSE TRUE FALSE FALSE
TRUE FALSE TRUE TRUE TRUE TRUE FALSE TRUE FALSE TRUE TRUE TRUE TRUE FALSE TRUE TRUE FALSE TRUE TRUE TRUE TRUE FALSE TRUE TRUE TRUE FALSE
TRUE TRUE TRUE FALSE FALSE FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE TRUE FALSE TRUE FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE
mix(AR305_FD,mixparam( mix(AR306_MD,mixparam(
BRA07306FD
BRA07307MD
BRA07307FD
BRA07314MD
BRA07314FD
Appendix 5.2.3
A 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 B C

D E F G H I J K
255
BRA07317MD
0.9 21.1 6.3 71.7 11.5 67.9 20.6 0.0 0.8 7.1 42.3 26.8 23.1 7.8 69.5 22.7 0.0 0.0 0.7 16.0 40.7 42.6 6.7 77.4 15.9 0.0 0.5 10.8 60.3 25.9 2.6
3 17 47 54 3 17 47 54 4 13 46 95 158 4 13 46 95 158 3 17 48 81 3 17 48 81 3 14 51 89 165
0.0094 0.2105 0.0629 0.7171 0.1150 0.6786 0.2064 0.0000 0.0080 0.0710 0.4228 0.2675 0.2307 0.0780 0.6948 0.2272 0.0000 0.0000 0.0067 0.1601 0.4071 0.4262 0.0670 0.7736 0.1594 0.0000 0.0048 0.1078 0.6026 0.2588 0.0259
0.4771 1.2428 1.6758 1.7358 0.4771 1.2428 1.6758 1.7358 0.5769 1.1260 1.6625 1.9766 2.1987 0.5769 1.1260 1.6625 1.9766 2.1987 0.4771 1.2422 1.6836 1.9086 0.4771 1.2422 1.6836 1.9086 0.4771 1.1555 1.7076 1.9513 2.2185
0.1096 0.3439 0.1114 0.2492 0.2250 0.3650 0.1610 0.0000 0.1925 0.2415 0.2327 0.1643 0.1613 0.1900 0.2850 0.2000 0.0000 0.0000 0.1079 0.3402 0.2003 0.2304 0.1900 0.3850 0.1450 0.0000 0.1055 0.3071 0.2312 0.1773 0.0874
FALSE FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
TRUE TRUE TRUE FALSE TRUE TRUE FALSE TRUE TRUE TRUE FALSE TRUE FALSE TRUE TRUE TRUE TRUE TRUE FALSE TRUE FALSE TRUE TRUE TRUE TRUE FALSE FALSE
FALSE FALSE FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE FALSE FALSE
BRA07317FD
BRA07324MD
BRA07324FD
BRA07327MD
BRA07327FD
BRA07328MD
Appendix 5.2.3
A 107 BRA07328FD 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 B C

D E F G H I J K
256
5.0 52.1 42.9 0.0 0.0
3 14 51 89 165
0.0500 0.5210 0.4290 0.0000 0.0000
0.4771 1.1544 1.7076 1.9513 2.2185
0.1830 0.3100 0.2020 0.0000 0.0000
TRUE FALSE FALSE
TRUE FALSE TRUE
TRUE TRUE TRUE
CPAStatistics:
lightMDca lightMDfa lightMDcp lightMDmp lightMDfp darkMDa darkMDcp darkMDmp darkMDfp
CPAMEAN (Weibull)
166 77 51 14 3 64 47 17 3
CPAMEDIAN (Weibull) 166 75 51 13 3 60 47 17 3
CPASTDEV (Weibull)
16 12 4 2 0 10 1 1 0
CPAMEAN CPAMEAN relative%MD relative%FD 12 0 44 0 35 32 9 58 1 10 65 0 19 24 15 63 1 13
Changes MD=>FD(in%) 12 44 3 +49 +9 65 +5 +48 +12
Appendix 5.2.3
257
Cell: B1 Comment: relative proportions of component populations (Weibull distribution) Cell: C1 Comment: means (modes) of component populations (Weibull distribution) Cell: D1 Comment: relative proportions of component populations (Weibull distribution) Cell: E1 Comment: means (modes) of component populations (Weibull distribution, log10) Cell: F1 Comment: Weibull distributions of component populations Cell: G1 Comment: constraints on relative proportions (Weibull distribution) Cell: H1 Comment: constraints on means (modes, Weibull distribution) Cell: I1 Comment: constraints on sigma values (Weibull distribution) Cell: J1 Comment: partially-constrained input parameters (final command line) Cell: K1 Comment: curve-fits for MD and FD sample expressions (Weibull distribution, partially-constrained CPA) Cell: A113 Comment: summary statistics of partially-constrained Component Population Analysis Cell: B113 Comment: averaged mean (mode) across all corresponding component populations (Weibull distribution, in m) Cell: C113 Comment: averaged median across all corresponding component populations (Weibull distribution, in m) Cell: D113 Comment: averaged standard deviation across all corresponding component populations (Weibull distribution, in m) Cell: E113 Comment: average relative proportions of corresponding component populations of minimally-dispersed samples (Weibull distribution) Cell: F113 Comment: average relative proportions of corresponding component populations of fully-dispersed samples (Weibull distribution) Cell: G113
Appendix 5.2.3
258
Comment: relative decreases and increases within corresponding component populations from minimally- to fully-dispersed condition Cell: A114 Comment: coarse aggregated component population in minimally-dispersed "light bands" Cell: A115 Comment: fine aggregated component population in minimally-dispersed "light bands" Cell: A116 Comment: coarse particulate component population in "light bands" Cell: A117 Comment: moderate particulate component population in "light bands" Cell: A118 Comment: fine particulate component population in "light bands" Cell: A119 Comment: single aggregated component population in minimally-dispersed "dark bands" Cell: A120 Comment: coarse particulate component population in "dark bands" Cell: A121 Comment: moderate particulate component population in "dark bands" Cell: A122 Comment: fine particulate component population in "dark bands"
Appendix 5.2.3
A B C
Unconstrained Component Population Analysis (CPA)

D E F G H I J K
259
CPAscreenshot CPAscreenshot (Gaussian) (Weibull)
SampleID
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
pi sigma (Gaussian) mu(Gaussian) (Gaussian) conpi,conmu,consigma="NONE"

0.1606 0.6287 0.2107 0.12 0.7044 0.1756 0.1382 0.4992 0.3626 0.0952 0.6399 0.2649 0.0469 0.2924 0.6607 0.1436 0.5737 0.2827 0.20326 0.7221 0.07464 0.08853 0.62688 0.28459 0.03078 0.37788 0.59134 0.09914 0.59512 0.30574 0.2194 0.6423 0.1383 1.432 1.826 2.1 0.478 1.103 1.685 1.317 1.735 1.968 0.4994 1.1385 1.6408 0.8799 1.4998 1.8138 0.4854 1.1152 1.6671 1.474 1.916 2.254 0.5024 1.1413 1.7669 0.868 1.522 1.865 0.6168 1.2523 1.7177 1.438 1.883 2.187 0.3497 0.2115 0.1491 0.1122 0.3068 0.1431 0.3503 0.2085 0.1618 0.1227 0.2972 0.1627 0.274 0.2343 0.1623 0.122 0.3136 0.1541 0.3624 0.2097 0.1177 0.1354 0.3202 0.1606 0.2151 0.2629 0.1961 0.1388 0.2887 0.1535 0.3638 0.2061 0.1263
3CPmodes (Gaussian,m)
27 67 126 3 13 48 21 54 93 3 14 44 8 32 65 3 13 46 30 82 179 3 14 58 7 33 73 4 18 52 27 76 154
pi mu sigma (Weibull) (Weibull) (Weibull) conpi,conmu,consigma="NONE"

0.05635 0.87594 0.06772 0.08625 0.68195 0.23179 0.07439 0.90177 0.02384 0.06044 0.5996 0.33996 0.09704 0.20281 0.70015 0.08605 0.53357 0.38038 0.04943 0.91766 0.03291 0.05351 0.58242 0.36407 0.3193 0.5521 0.1286 0.06057 0.42893 0.5105 0.1047 0.7947 0.1005 1.24 1.827 2.222 0.4475 1.0482 1.6269 1.19 1.797 2.23 0.4571 1.0765 1.5824 1.146 1.735 1.735 0.434 1.004 1.6 1.102 1.873 2.369 0.4506 1.0613 1.7051 1.519 1.73 2.057 0.5692 1.0994 1.6172 1.268 1.85 2.233 0.3482 0.251 0.1334 0.09693 0.30952 0.19497 0.34072 0.23975 0.09063 0.1001 0.309 0.2122 0.3652 0.1762 0.2341 0.09112 0.32191 0.20762 0.28528 0.26056 0.08346 0.1083 0.3167 0.2135 0.3793 0.2074 0.1327 0.1146 0.2735 0.2253 0.3477 0.2443 0.1361
3CPmodes (Weibull,m)
17 67 167 3 11 42 15 63 170 3 12 38 14 54 54 3 10 40 13 75 234 3 12 51 33 54 114 4 13 41 19 71 171
BRA07300MD
BRA07300FD
BRA07303MD
BRA07303FD
BRA07304MD
BRA07304FD
BRA07305MD
BRA07305FD
BRA07306MD
BRA07306FD
BRA07307MD
Appendix 5.2.3
A 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 B C

D E F G H I J K
260
3 11 48 14 58 196 4 10 51 13 48 105 3 11 40 19 66 158 3 13 45 13 57 163 2 11 40 16 60 139 3 12 46
BRA07307FD
BRA07314MD
BRA07314FD
BRA07317MD
BRA07317FD
BRA07324MD
BRA07324FD
BRA07327MD
BRA07327FD
BRA07328MD
BRA07328FD
0.1013 0.6193 0.2794 0.31158 0.66168 0.02673 0.09628 0.42405 0.47967 0.03812 0.33888 0.62299 0.1198 0.5763 0.3039 0.2131 0.5383 0.2486 0.09904 0.74382 0.15713 0.021 0.2538 0.7252 0.07352 0.60687 0.31961 0.1197 0.405 0.4753 0.06664 0.56719 0.36617
0.5045 1.1213 1.7487 1.298 1.833 2.321 0.6039 1.1833 1.775 0.6642 1.3926 1.7961 0.5219 1.1805 1.6877 1.382 1.836 2.176 0.5728 1.1827 1.7204 0.619 1.396 1.84 0.4628 1.1359 1.6863 1.144 1.644 1.927 0.5276 1.2123 1.7491
0.1383 0.3164 0.1663 0.3976 0.2396 0.0489 0.1318 0.3054 0.1583 0.1915 0.28 0.1917 0.161 0.3068 0.1586 0.3827 0.2303 0.1443 0.1684 0.2899 0.1534 0.1789 0.3074 0.2134 0.1312 0.3225 0.1831 0.3506 0.2316 0.1724 0.1565 0.3117 0.1791
3 13 56 20 68 209 4 15 60 5 25 63 3 15 49 24 69 150 4 15 53 4 25 69 3 14 49 14 44 85 3 16 56
0.06307 0.57303 0.3639 0.1746 0.777 0.0484 0.05889 0.3278 0.61331 0.16083 0.77021 0.06896 0.04968 0.51254 0.43777 0.1164 0.7044 0.1792 0.05154 0.73342 0.21505 0.118 0.8392 0.0427 0.03276 0.53019 0.43705 0.1293 0.8239 0.0468 0.03029 0.46597 0.50374
0.4576 1.0366 1.6803 1.149 1.76 2.292 0.5539 1.0095 1.7086 1.116 1.685 2.023 0.4188 1.0406 1.6033 1.272 1.819 2.2 0.5033 1.1242 1.6573 1.128 1.756 2.212 0.3774 1.0375 1.6022 1.204 1.775 2.144 0.434 1.096 1.667
0.114 0.3117 0.2221 0.3871 0.2772 0.0905 0.1046 0.264 0.2128 0.3656 0.2383 0.1504 0.1047 0.329 0.2227 0.3917 0.2758 0.1629 0.1396 0.3002 0.2087 0.3723 0.2449 0.1017 0.08102 0.33472 0.24541 0.3731 0.2421 0.1215 0.1095 0.3162 0.2422
Appendix 5.2.3
A B C

D E F G H I J K
261
3CPsRange (Gaussian)
85209 4482 1430 4860 1316 34 6373 2533 48 4952 1318 34
unconstrained 70 CPAStatistics:
71 72 73 74 75 76 77 78 79 80 81 82
lightMDc lightMDm lightMDf lightFDc lightFDm lightFDf darkMDc darkMDm darkMDf darkFDc darkFDm darkFDf
3CPMEAN 3CPMEDIAN 3CP%MEAN (Gaussian) (Gaussian) (Gaussian) 151 152 123 68 68 70 24 26 24 55 56 56 14 15 14 3 3 3 68 67 67 29 28 29 6 6 6 49 49 49 15 14 15 3 3 3
3CPSTDEV (Gaussian)
39 12 5 4 1 0 4 4 2 2 2 0
3CPMEAN 3CPMEDIAN 3CP%MEAN (Weibull) (Weibull) (Weibull) 178 169 169 66 67 66 16 17 16 47 47 48 12 11 12 3 3 3 109 110 71 53 54 53 18 14 22 40 40 40 11 11 11 3 3 3
3CPSTDEV (Weibull)
30 6 2 3 1 0 39 3 8 1 1 1
3CPsRange (Weibull)
139234 5875 1319 4251 1013 34 54163 5748 1333 4041 1013 24
Appendix 5.2.3
262
Cell: B1 Comment: relative proportions of 3 unconstrained component populations (Gaussian distribution) Cell: C1 Comment: means (modes) of 3 unconstrained component populations (Gaussian distribution, log10) Cell: D1 Comment: sigma values of 3 unconstrained component populations (Gaussian distribution) Cell: E1 Comment: means (modes) of 3 component populations (Gaussian distribution, in m) Cell: F1 Comment: relative proportions of 3 unconstrained component populations (Weibull distribution) Cell: G1 Comment: means (modes) of 3 unconstrained component populations (Weibull distribution, log10) Cell: H1 Comment: sigma values of 3 unconstrained component populations (Weibull distribution) Cell: I1 Comment: means (modes) of 3 component populations (Weibull distribution, in m) Cell: J1 Comment: curve-fits for MD and FD sample expressions (Gaussian distribution, unconstrained CPA) Cell: K1 Comment: curve-fits for MD and FD sample expressions (Weibull distribution, unconstrained CPA) Cell: I16 Comment: single population? Cell: I17 Comment: single population? Cell: A70 Comment: summary statistics of unconstrained Component Population Analysis Cell: B70 Comment: averaged mean (mode) across 3 corresponding unconstrained component populations (Gaussian distribution, in m) Cell: C70 Comment: averaged median across 3 corresponding unconstrained component populations (Gaussian distribution, in m) Cell: D70 Comment: averaged weighted mean (mode) across 3 corresponding unconstrained component populations (Gaussian distribution, in m) Cell: E70
Appendix 5.2.3
263
Comment: averaged standard deviation across 3 corresponding unconstrained component populations (Gaussian distribution, in m) Cell: F70 Comment: averaged mean (mode) across 3 corresponding unconstrained component populations (Weibull distribution, in m) Cell: G70 Comment: averaged median across 3 corresponding unconstrained component populations (Weibull distribution, in m) Cell: H70 Comment: averaged weighted mean (mode) across 3 corresponding unconstrained component populations (Weibull distribution, in m) Cell: I70 Comment: averaged standard deviation across 3 corresponding unconstrained component populations (Weibull distribution, in m) Cell: J70 Comment: range of modes within 3 corresponding unconstrained component populations (Gaussian distribution, in m) Cell: K70 Comment: range of modes within 3 corresponding unconstrained component populations (Weibull distribution, in m) Cell: A71 Comment: coarse component population in minimally-dispersed "light bands" Cell: A72 Comment: moderate component population in minimally-dispersed "light bands" Cell: A73 Comment: fine component population in minimally-dispersed "light bands" Cell: A74 Comment: coarse component population in fully-dispersed "light bands" Cell: A75 Comment: moderate component population in fully-dispersed "light bands" Cell: A76 Comment: fine component population in fully-dispersed "light bands" Cell: A77 Comment: coarse component population in minimally-dispersed "dark bands" Cell: I77 Comment: high value reflecting BRA07-AR 304 (MD) lacking coarse population? Cell: A78 Comment: moderate component population in minimally-dispersed "dark bands" Cell: A79 Comment: fine component population in minimally-dispersed "dark bands"
Appendix 5.2.3
264
Cell: A80 Comment: coarse component population in fully-dispersed "dark bands" Cell: A81 Comment: moderate component population in fully-dispersed "dark bands" Cell: A82 Comment: fine component population in fully-dispersed "dark bands"
Appendix 5.2.3
A B C
Partially-constrained Conventional Particle-Size Analysis (PSA)

D E F
265
SampleID
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
Mean (inm)
83.01
Median (inm)
73.07
Mode (inm)
76.05
PSAscreenshot
BRA07300MD
BRA07300FD
21.05
13.98
17.86
BRA07303MD
71.82
65.01
74
BRA07303FD
24.14
17.86
37.95
BRA07304MD
58.27
55.11
61.65
BRA07304FD
24.41
17
44.84
BRA07305MD
90.84
79.37
83.52
BRA07305FD
29.95
19.25
58.98
BRA07306MD
64.47
56.86
62.79
BRA07306FD
31.13
24.4
47.67
BRA07307MD
87.12
76.24
100.3
BRA07307FD
28.31
17.87
46.5
Appendix 5.2.3
A 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 B C

D E F
266
BRA07314MD
68.41
55.82
67.91
BRA07314FD
39.78
34.63
68.7
BRA07317MD
54.34
49
58.45
BRA07317FD
27.84
20.81
51.06
BRA07324MD
91.04
76.78
109.9
BRA07324FD
23.7
16.61
14.91
BRA07327MD
65.92
58.28
64.34
BRA07327FD
28.52
20.44
54.34
BRA07328MD
67.5
60.78
68.03
BRA07328FD
35.11
26.52
51.54
PSAstatistics:
lightMD lightFD darkMD darkFD
PSDMEAN 81 30 61 28
PSDMEDIAN 70 21 55 21
PSDMODE 84 43 62 49
Appendix 5.2.3
267
Cell: B1 Comment: means for minimally- and fully-dispersed samples returned by the Beckman Coulter Multisizer software (v.3.51) Cell: C1 Comment: medians for minimally- and fully-dispersed samples returned by the Beckman Coulter Multisizer software (v.3.51) Cell: D1 Comment: modes for minimally- and fully-dispersed samples returned by the Beckman Coulter Multisizer software (v.3.51) Cell: E1 Comment: Coulter Multisizer 3 frequency particle size-distributions (PSD) of minimally- and fully-dispersed samples (log normal) Cell: A69 Comment: summary statistics of standard Particle Size Analysis (PSA) performed by the Beckman Coulter Multisizer software (v.3.51) Cell: B69 Comment: average mean across corresponding samples Cell: C69 Comment: average median across corresponding samples Cell: D69 Comment: average mode across corresponding samples Cell: A70 Comment: minimally-dispersed "light bands" Cell: A71 Comment: fully-dispersed "light bands" Cell: A72 Comment: minimally-dispersed "dark bands" Cell: A73 Comment: fully-dispersed "dark bands"
Appendix 5.2.3
A B C
Sediment Sample Descriptions

D
268
SampleID
1 2 3 4 5 6 7 8
Depth(incm)
CPAshorthand Detaileddescription
massivelayerofclastsupportedorangemottledpebbles(Bunyerooshales),pinchingoutinupstreamsectionface,increaseinsubangular tufaclaststowardsbase,erosivecontacttounderlying Siltbandmarkedbytufafragments,lightlydisturbedbymaterialincorporatedfromoverlyinggravellayer discontinuouslensesofsmalltufafragments,partiallyinterfingeringwithveneersoforganicdetritus Siltbandmarkedbydistinctcomplexofmultipleveneersoforganicdetritus,3mmthickwhereprotectedbyoverlyingtufa lightbandofgreyfinegrainedSilts,largelyundisturbed Siltbandmarkedbythin(<1mm)butdistinctcomplexofmultipleveneersoforganicdetritus lightbandofgreyfinegrainedSilts,partiallymottled,somecharcoalfragmentsandgastropodshellstowardsbase thinlayerofclastsupportedorangemottledpebbles(Bunyerooshales),increaseinsubangulartufaclaststowardstopandbase,erosive contacttounderlying lightbandofgreyfinegrainedSilts,partiallydisturbedbymaterialincorporatedfromoverlyinggravellayer,fewremnantsoforganic veneers,gastropodshellfragments Siltbandmarkedbythin(<1mm)butdistinctcomplexofmultipleveneersoforganicdetritus,thelowercomplexmantlingthenear horizontalgravelunit thinlayerofclastsupportedorangemottledpebbles(Bunyerooshales),increaseinsubangulartufaclasts(<4cm)towardsbase,clearcut contactwithoverlyingbuterosivecontacttounderlyingunit(partlyincorporatingSilts) lightbandoffinegrainedSiltswithtufafragments,partiallydisturbedwithmaterialincorporatedfromoverlyinggravel,wherethegravel layerpinchesoutitslateralequivalentismarkedby3veneersoforganicdetritus,tufaandgastropodshellfragments undulatingSiltbandmarkedbycomplexofmultipleveneersoforganicdetritus,thinbutdistinct(<1mm)whereundisturbed lightbandofgreyfinegrainedSilts,inplacestracesofupto3veneersoforganicdetritus massivelayerofclastsupportedorangemottledpebbles(Bunyerooshales),occasionalroundedtosubroundedquartzpebbles,large(~6cm x4cm)clastsoftufaconcentratingtowardsbase,imbricationindicatingpaleoflowfrompresentdayBrachinatrunkchannelintoAroona tributary(50),fewcharcoalpieces,upperpartofunitinplacesmatrixsupportedbygreySilts,basaltransitioniserosionalandinplaces mixedwithunderlyingbrownSilts
BRA07280 BRA07300 BRA07303 BRA07304 BRA07305 BRA07306 BRA07307
280300 300303 303304 304305 305306 306307 307310 310314 314317 317318 318324 324327 327328 328330
gravelband lightband tufaband darkband lightband darkband lightband gravelband lightband darkband gravelband lightband darkband lightband
9 BRA07310 10 BRA07314 11 BRA07317 12 BRA07318 13 BRA07324 14 BRA07327 15 BRA07328
16 BRA07330
330340
gravelband
Appendix 5.2.3
Cell: B1 Comment: CPA sub-section specific depth (from TOP)
Sediment Sample Descriptions
269
Cell: A13 Comment: 2 separate samples collected and sized to test consistency of PSA/ CPA results
Appendix 5.2.3
A B C D
Stratigraphic Section Description

E
270
Depth MunsellSoil AMSages Shorthand 1 (cmfromTOP) Colour(moist) (inacalBP) TOP060 3 060090

2 4 5 6 7 8 9 10
Detaileddescription
reddishsandyunit,inplacesupto100cmthickwithsmallcalcareousnodules(weaklydevelopedBcahorizon),furtherdownstream erosionalgravelfilledchutes lensesofunweatheredmatrixsupportedgravel,incisingintounderlyingunit,lithologyindicativeofAroonacatchment disturbedtransitionalzone,largelyconsistingofSilts,channelsandburrowswith"pink"infill(fromabove),gypsumprecipitationon channelwalls greySilts,hostingtopmostdiscernibleremnantsofveneersoforganicdetritusinsection,intactgastropods,fewtufafragments layerofclastsupportedorangemottledpebbles(Bunyerooshales),greatlyvaryinginthickness(channellenses) greySilts,thinundulatingbandsandwichedbetweengravelunits layerofclastsupportedorangemottledpebbles(Bunyerooshales),discontinuous,pinchingoutinupstreamsection greySilts,thinundulatingbandsandwichedbetweengravelunits massivelayerofclastsupportedorangemottledpebbles(Bunyerooshales),inplacestoppedbytufafragments,onelargely unweatheredsubangularcoarsegravel(~20cm,greyshale)placedontop
7.5YR5/44/4
"reddrape" freshgravel transitional Silts gravel 18.820227 Silts gravel Silts gravel
090130 130140 140155 155160 160170 170175 175200
7.5YR6/35/4 2.5YR7/25/2 2.5Y6/85/6 2.5YR7/25/2 2.5Y6/85/6 2.5YR7/25/2 2.5Y6/85/6
11 12 13 14 15 16 17 18
200220 220230 230240 240250 250255 255275

2.5Y6/85/6 2.5Y6/85/6
2.5Y6/85/6
275370 370380 19 380400 20 400415

21 22 23
7.5YR4/2 7.5YR4/2 10YR5/6
415425 425440 440455
10YR5/6
greySiltshostingtwoprominentbandsofclastsupportedorangemottledpebbles(Bunyerooshales),pinchingoutindownstream section,gravellaterallyreplacedbySiltswithtufafragmentspartiallyprotectingunderlyingorganicveneersfromerosion,topmostwell Silts/gravel definedbandofmultipleveneersoforganicdetritusinsection layerofclastsupportedorangemottledpebbles(Bunyerooshales),continuousandnearlevel,increaseinsubangulartufaclasts gravel/tufa towardsbase(upto8cm) Silts/tufa greySilts,lowerhalfhostingstreaksoftufa(upto5cmthick,partiallyformedinsituandtoppedbyveneersoforganicdetritus) layerofclastsupportedorangemottledshalepebbles,nearlevel,pinchingoutinupstreamsection,increaseinsubangulartufaclasts gravel towardsbase thingreySiltswithmarkedbandoftufa(upto5cmthick),toppingunderlyinggravelunit,tufapartlyformedinsitubutincorporating 19.465337 Silts/tufa clasts gravel,mud massivelayerofclastsupportedorangemottledpebbles(Bunyerooshales),tufaclasts,fewpurple(5YR6/24/2))mudpebbles(~5cm) pebbles towardsbase stackedunitofsix~5cmthickundulatingbandsofclastsupportedpebbles(Bunyerooshales)withtufaclasts,eachtoppedbygreySilts 18.991238 6couplets comprisingupto3complexesofmultipleveneersoforganicdetritus 18.357303 brownSilts brownSilts,lowermostdistinctmassiveSiltunit,chromatic,disturbed,remnantsofveneersoforganicdetritus gravel mottledmatrixsupportedfine(<2cm)gravelindarkclayeysiltymatrix,charcoalpieces mud darkclayeysiltyunit,disturbed,occasionalmottledgravel layerofmottledmatrixsupportedsmall(<2cm)gravelindarkclayeysiltymatrix,charcoalpieces,graduallyupwardfiningsequence gravel withorangemottledsandandfinegravel gravel, layerofclastsupportedorangemottledpebbles(Bunyerooshales),~40%freshlargersubangulartosubroundedclastsfromvarious calcrete upstreamprovenances(greyABCQuartzite,redBonneySandstone),restingon~1cmcalcretepan finegravel orangemottledfinegravel(<1cm),~30%pocketsofpurple(5YR4/2)Silts multipleunitsofupwardfiningsuccessionsofclastsupportedgravel,largebasalclasts(upto~25cm)sourcedfromvariableupstream 20.039333 coarsegravel lithology(Brachinatrunkchannel),pocketsofpurpleSilts(5YR4/2,mudpebbles)andthinSiltbands,large(~1.5cm)charcoalpieces
24
455520
Appendix 5.2.3
Stratigraphic Section Description
271
Cell: C7 Comment: OZK516, AMS sample from intact gastropods in topmost Silt layer underlying orange-mottled shales, tufa fragments (@150-160) Cell: C15 Comment: OZK002, AMS sample from broken gastropods in Silt layer overlying topmost orange-mottled shales of CPA sub-section (@250-255/ [275-280 in CPA section]) Cell: C17 Comment: OZK003, AMS sample from broken gastropods in Silt layer underlying basal orange-mottled shales of CPA sub-section (@~335-340/ [340-345 in PSA section]) Cell: C18 Comment: SSAMS1812 & SSAMS 2038, 2 AMS samples from fragile pieces of charcoal from veneer of organic detritus and 2 small unbroken gastropods, respectively Cell: D22 Comment: shift in imbrication: lower large allochthonous gravel indicate paleoflow direction towards 170 (S), mottled gravel units above paleoflow directions between 230 (SW) and 50 (NE) Cell: C24 Comment: OZK517, AMS sample from large (~1.5 cm) piece of charcoal at base of pit (@ 507-508) in purple silt lens level with large (~25 cm) gravel clasts
Ap p en d ices (5.2.3.1 Met h o d s)
272
273
Particle-Sizing Protocol
The Multisizer 3 COULTER COUNTER instrument here employed to size the samples is based on the Coulter electrical impedance method coupled with a Digital Pulse Processing technology. It provides 3-D particle sizing and counting statistics (Beckman Coulter 2002). The sizing range is limited by the aperture diameter of the selected orifice tubes separating the two electrodes. The size-range restrictions prevent blocking of the orifice and ensure normalised sizing sensitivities. The maximum nominal particle diameters for all orifice tubes selected for the present study are listed below (Table 1). All samples were wet-sieved into the listed size fractions and sized by the full range of orifice tubes. Table 1) aperture (in m) 50 140 280 560 min. nominal range (in m) 1.0 2.8 5.6 11.2 max. nominal range (in m) 30.0 84.0 168.0 336.0 calibration beads 5-10 (10) 14-28 (20) 28-56 (43) 56-112 (90) sieve size (in m) < 38 < 90 < 180 < 355
Nominal orifice tube data for the Multisizer 3 COULTER COUNTER The electrolyte used for wet-sieving and diluting the sample suspensions is triple-filtered (0.22 m Millipore nitrocellulose membranes) 1% NaCl (ISOTON II). Particle-suspension concentrations were maintained at 5-10%. The results from the array of orifice tubes were merged into sample particlesize distributions using the Beckman Coulter Multisizer 3 software (version 3.51).
Calibration
Prior to sizing the samples, all orifice tubes were calibrated with COULTER latex beads in ISOTON II (Table 1). Nine calibration runs were carried out for each orifice tube. The median was calculated and entered in the Change Aperture Tube Wizard.
Sample preparation
Two separate sample pre-treatment protocols were applied to all samples; the first sizing transportstable aggregates among particles to describe the minimally-dispersed (MD) sample expression, the second sizing only elementary particles to describe the fully-dispersed (FD) sample expression.
274
Pre-treatment and sizing of minimally-dispersed samples Minimally-dispersed (MD) samples were treated with the aim to size partially-aggregated suspension load which would behave transport-stable within turbulent fluvial flow. 1) A representative subsample of the sediments is reduced to ~2-3 g avoiding direct touch or cutting. 2) An equivalent of 100 ml of electrolyte per gram of sample material is added in a clean baffled beaker and left to soak for 60 seconds. 3) 4) The slaked sample is stirred into a uniform suspension by a magnetic flee for 60 secs. One subsample per orifice tube are collected by pipette ~1/3 above and to the side of the magnetic stirring flee. Sample quantities vary between ~3-10 ml, depending on the orifice diameter and targeted total counts. 5) The suspension is gently passed through respective sieves (Table 1) by a squeeze bottle filled with electrolyte. Rare particles retained by the largest sieve (355 m) are collected for observation under binocular microscope and recorded. 6) Sample suspensions are diluted with electrolyte until they reach a concentration of 5-10% and sized in a clean Multisizer beaker by the selected array of orifice tubes, starting with the 560 m aperture. The stirrer of the Multisizer 3 COULTER COUNTER is adjusted to maintain the sample in suspension without introducing excess turbulence and air bubbles; i.e. 40=560 m tube, 35=280 m tube, 30=140 m tube, 15-20=50 m tube. 7) All subsamples are sized to their maximum count (or a minimum of 300,000 individual coincidence-corrected counts). Due to the tendency of the 50 m aperture to block, necessitating a reset of the count and purge of the system, fewer counts may be recorded for the smallest orifice tube. The electric pulses are saved as RAW data and converted to 300 discrete size bins per tube. Subsequently, the spread in bins is manually limited to the sampleand tube-specific effective size range. The data is plotted as particle diameters (in m) against the x-axis (log scale) and volume (in %) against the y-axis (linear scale). 8) The individual orifice tube measurements for each sample are merged using the Multi-tube Overlap function. Software-generated defaults based on best curve-fits are documented and, where necessary, adjusted by user-adjusted reruns (Fig. 1)
275
Fig. 1) Software-generated default (in blue) and user-adjusted multi-tube merges (in red) across individual orifice tube data (in grey) for sample BRA07-AR 300 (FD). Pre-treatment and sizing of fully-dispersed samples Fully-dispersed (FD) samples are treated with the aim to size disaggregated suspension load consisting entirely of elementary particles. 1) 2) 3 ml of 32% HCl are added per 100 ml of sample suspension. The beaker with the acidified sample is immersed into an ultrasonic bath (Branson 2200 sonifier, 472Hz/ 60W) for 30 min, while stirring the suspension every 10 minutes. 3) The sample suspension is analysed under binocular microscope and if found thoroughly dispersed subsampled, sized and analysed according to the same protocol employed for the MD sample expressions.
Cliffing effect
Some samples recorded an abnormal high count towards the fine end of the sizing range, thereby skewing the overall particle-size distribution. This effect proved particularly problematic for the smallest diameter orifice tube (50 m). Comparison with ICP-MS data suggested that elevated iron contents might be responsible by interfering with the electrical sensing method. As a result, the data required correction prior to the multi-tube merge and calculation of statistics. The lowermost affected size bins were truncated by manually adjusting the lower sizing threshold to the onset of the cliffing effect, and reconverting the remaining continuous pulse data to 300 discrete size bins. Sufficient overlap was provided by employing the large number of orifice tubes. The correction worked fine with all samples but limited the lowermost effective size range to ~3 m.
276
277
Component Population Analysis Protocol

In order to establish the principle particle size populations that make up the sediment samples, the particle-size distributions (PSDs) are statistically resolved into discrete component populations (CPs) of aggregates and elementary particles. The present parametric approach is based on the open source software package Mixdist version 0.52 (Macdonald & Du 2004), compiled to run in the R software environment for statistical computing (R Foundation for Statistical Computing 2009) and freely distributed online via CRAN (The Comprehensive R Archive Network). Alternatively, it can be downloaded here: http://cran.r-project.org/web/packages/mixdist/index.html (10/2009). Mixdist fits finite mixture distribution models to grouped and conditional data using a Newton-type and Expectation-Maximisation algorithms based on the method of maximum likelihood. Detailed documentation of the software package and syntax is presented here:
http://rss.acs.unt.edu/Rdoc/library/mixdist/html/00Index.html (10/2009).
Data conversion from Multisizer 3 COULTER COUNTER to Mixdist

Mixdist can only read a specific tab-delimited text file format, requiring prior data conversion: 1) Merged Multisizer 3 data sets (both minimally- and fully-dispersed expressions) are exported as tab-delimited size lists (run file -> export data -> tab-delimited size listing), opened in OpenOffice Calc (or Microsoft Office Excel) and reduced to two columns; 256 size fractions reflecting lower size-bin diameters (in m) and corresponding volume percentages. The base-10 logarithm is calculated for all size fractions in a separate column to the left of the volume using the formula [=log10(cell)]. 2) All 256 size fractions (as expressed as base-10 logarithm values) and their corresponding volume percentages are limited to 10 decimal places and copied into a tab-delimited text file via the clipboard function. This text file is saved in the data directory of the Mixdist package (e.g. C:\Program Files\R\R-2.7.2\library\mixdist\data\sample_name.txt).
Unconstrained parametric analysis using Mixdist

R is opened and the Mixdist package installed either from the local zip-file or via the selected CRAN mirror. 1) The Mixdist package is loaded by entering the command > library(mixdist) 2) The sample data file is imported > data(sample_name) and converted into a Mixdist object by entering
278
> sample_name <- as.mixdata(sample_name) 3) The PSD of the converted data is plotted > plot(sample_name) and compared for consistency with the original Multisizer 3 PSD. 4) The PSD is now decomposed into multiple discrete populations by entering the following syntax > new_sample_name <mix(sample_name,mixparam(mu=c(...),sigma=c(...),pi=c(...)),"sample_distribution",mixconstr(c onpi="NONE", conmu="NONE", consigma="NONE", fixpi=c(NULL), fixmu=c(NULL), fixsigma=c(NULL))) Mixdist requires a set of initial assumptions to start the computation; i.e. expected number, means and vectors of standard deviations (sigma values) of CPs, and the type of probability distribution function. This step has potentially large implications for the final outcome of the analysis and must be considered by an experimental design. It is here suggested to start with an unlikely high number of populations set equal distances apart; e.g. for a unimodal distribution ranging from 2350 m five populations are suggested with (converted base-10 log) mu [means of component distribution] =c(0.3,0.75,1.25,1.75,2.5). Vectors of standard deviations are kept minimal so that potential minor (narrow) components are registered; e.g. sigma=c(0.2). The relative proportions of all populations are expressed as a near equal contributions, letting the software determine main and minor components; e.g. pi [mixing proportions of components] =c(0.2). Finally, the type of probability function is specified; e.g. Gaussian (norm) or Weibull power distribution (weibull). All parameters are initial suggestions only and must be left unconstrained by defining the arguments conpi, conmu and consigma as ="NONE" and corresponding fixpi, fixmu and fixsigma as =c(NULL). 5) The computed populations are plotted by entering > plot(new_sample_name) and activating the graphics window in R. 6) Subsequently, the parameters for all computed CPs are listed numerically >list(new_sample_name) 7) Improbable small and near-duplicate population (resulting in error messages) are excluded from consequent runs by adjusting the input parameters. The protocol is repeated until the minimum number of modes resulting in a good description of the PSD is established. The process may be duplicated for alternative probability distribution types (e.g.
279
Weibull or Gaussian). Repeatability and reliability of the final results are tested by adjusting the returned values for mean, sigma and proportions to different values. On the matter of choosing a distribution type, our experimental observations suggest that partiallyaggregated sediments are best described by the asymmetric Weibull power distribution, perhaps by allowing for tails of break-up products that must be expected from the sample pre-treatment and sizing protocol (see section 5.2.3.1). On the other hand, populations of elementary particles might better be described by Gaussian distributions. The distribution type with superior curve-fit can be visually determined. Mixdist also analyses the variance table and returns goodness-of-fit chi-square (Chisq) statistics which can be accessed by entering: > anova(new_sample_name) However, strictly statistically this value is not valid for data in which bin frequencies represent volume percentages. Neither are the returned values for standard errors associated with each population (accessed via > summary(new_sample_name)). Chi-square and standard errors are based on multi-nomial distributions that apply only when bin frequencies are expressed as number of individuals. 8) The numerical results are transferred to a spreadsheet. The modes of populations are converted back into m by the function [=POWER(10;(cell)], and relative proportions are presented as percentages. 9) It is good practise to copy the syntax line for documentation and save a screenshot of the plot via the GUI.
Partially-constrained parametric analysis using Mixdist

Corresponding CPs of minimally- and fully-dispersed sample expressions can be compared by partially constraining individual parameters; i.e. modes, relative proportions and/or sigma values of specific populations. Such data manipulation can alter the results and must be addressed by a documented experimental design. For the present study, parameters are partially constrained to account for the adverse effect of truncated populations at the fine end of the PSD and to quantify aggregation within sediment samples. Partial constraints were imposed by the following protocol: 1) Both partially-aggregated/minimally-dispersed (MD) and fully-dispersed (FD) expressions of the sample are analysed in unconstrained mode as detailed above and results are recorded. 2) Subsequently, the modes returned for the FD sample expression are copied and used as constraints to resolve corresponding MD particle-size distributions.
280
The rationale behind this approach is that the same elementary particle populations, or composites of them if aggregated, must be present in both distributions. Therefore, individual CPs comprising entirely of aggregates, in the MD size distributions simply register with a near-zero relative proportion in corresponding FD size distributions. This approach has the advantage that identical CPs are quantified in both sample expressions, thereby allowing aggregation per se to be quantified. 3) Data not accounted for by constrained elementary particle modes reflect additional aggregated CPs absent in the FD sample expression. These must be quantified unconstrained, employing the same protocol as described above. Here is an example of the syntax for a MD sample with three constrained fine modes and two additional unconstrained aggregate populations. Only the means of the finer modes (a, b and c) are fixed. Their relative proportions and sigma values remain unconstrained for they are likely to change. Note also that conmu is expressed as MFX with corresponding fixmu=c(TRUE,TRUE,TRUE for the three constrained modes a,b and c, and FALSE,FALSE for the two unconstrained coarse modes d and e). > new_sample_name_FD <mix(sample_name_MD,mixparam(mu=c(a,b,c,d,e),sigma=c(0.2),pi=c(0.2)),"weibull",mixconstr( conpi="NONE", conmu="MFX", consigma="NONE", fixpi=(NULL), fixmu=c(TRUE,TRUE,TRUE,FALSE,FALSE,FALSE), fixsigma=c(NULL))) 4) Results are copied to a spreadsheet and converted. Corresponding minimally- and fullydispersed sample expressions are compared by establishing the differences in their relative sample contributions, i.e. increases in elementary particle populations and disappearance of aggregated populations. The workflow can be improved by using arrow keys to replicate (and change) previous command lines, and by the use of single letter file names that are overwritten in subsequent runs.
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) surfacedrape surface 000010 (5YR6/6dry drape 5YR4/6 moist) 000010 transitionalsilts surface 000010 drape 010035 calcichorizon transitional 010050 silts
BRA-SD Tabular Section Description

Finegrainedunits description Graveldescription Pedogenic description 13Csamples description 14Csample references OSLages
(Williamset al.2001)
281
TopSection TopSection
veryloosesoftsilts(sodic?) brownmottledsiltmatrixwithminorsand component,platystructure platy,largelyangularand fewsubroundedlocally derivedshales(purpleand grey)oftheBrachina Formation,mostclastsare CaCO3encrustedand occurassheetssub paralleltothesurface, someconcentrationsin minorchannels distinctundulatingsheet ofpebbles calcareouscutansand towardsthebasesoft verticalcalcareous nodules
TopSection
035040 calcichorizon TopSection 040056 calcichorizon
transitional 010050 silts
brownsiltmatrix,pronouncedorangemottles
TopSection
TopSection
transitional 010050 silts 056057 disturbedorganic disturbed 050078 veneers silts
brownsiltmatrix uppermostdiscernableveneersoforganicdetritus, disturbed
pronouncedonsetof vertical(insitu) calcareousnodulesat ~38cm calcifiedrootcastsand nodules charcoalfragments SSAMSANU fromdisturbed 4107(051) organicveneers calcifiedrootcasts
TopSection TopSection TopSection TopSection TopSection TopSection
TopSection
disturbedmassive disturbed silts&organic silts veneers 066067 organicveneers disturbed silts 067068 disturbedyellow disturbed band silts 068077 disturbedmassive disturbed silts&organic silts veneers 077078 disturbedorganic organic veneers veneers 078088 disturbedmassive organic silts&organic veneers veneers 088095 continuousyellow yellowband band
05766
050078
massiveorangemottledsiltswithtracesofveneersof organicdetritus uppermostpartiallyintactpronouncedveneerof organicdetritus uppermostdiscreteyellowband,thinanddisturbed overalldarkerdisturbedsilts
050078 050078 050078
fewcalcifiedrootcasts disturbedorganic veneers fewcalcifiedrootcasts fewcalcifiedrootcasts
078093 078093
partiallypronouncedbutlargelydisturbedveneersof organicdetritus lighterdisturbedsilts
fewcalcifiedrootcasts fewcalcifiedrootcasts
093100
uppermostdistinctcontinuousyellowbandof massivesilts,blurredupperboundary(bioturbation)
lowermostoccurence ofpedogenic carbonates disturbedorganic veneers
SSAMSANU 4109(094)
TopSection
095098 organicveneers
organic veneers
100105
distinctmassiveunitofmultipleveneersoforganic detritus,partiallydisturbed
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 098100 yellowband organic 100105 veneers 100104 organicveneers organic 100105 veneers 104111 yellowband yellowband 105117

282
TopSection TopSection TopSection TopSection TopSection TopSection TopSection TopSection TopSection TopSection
thinyellowbandofmassivesilts darkbandofdisturbedveneersoforganicdetritusin massivesilts,morepronouncedtowardstop distinctcontinuousyellowbandofmassivesilts, blurredupperboundary(bioturbation) partiallydisturbedundulatingdarkbandofveneersof organicdetritus undulatingpartiallydisturbedyellowbandofmassive silts darkbandofveneersoforganicdetritus,inplaces wellpreserved overalldarkerdisturbedandorangemottledsiltsof varyingthickness,tracesofveneersoforganic detritus disturbedundulatingdarkerbandofveneersof organicdetritus morepronouncedthinpartlyorangemottledmassive band disturbedundulatingdarkerbandofveneersof organicdetritus,veneerstowardsbaseandtopmore pronounced distinctyellowbandwithblurredboundaries
111113 disturbedorganic yellowband 105117 veneers 113115 yellowband organic 117125 veneers 115117 organicveneers organic 117125 veneers 117137 disturbedmassive disturbed 125135 silts&organic yellowband veneers 137138 disturbedorganic organic 135145 veneers veneers 138141 disturbedyellow organic 135145 band veneers 141143 disturbedorganic organic 135145 veneers veneers 143145 yellowband disturbed silts disturbed silts disturbed silts disturbed silts disturbed silts disturbed silts 145160
wellpreserved organicveneers
SSAMSANU 4207(116)
disturbedorganic veneers,mottled, gypsum 17.700 1.100 (145)
TopSection TopSection TopSection TopSection TopSection 145150 disturbedorganic veneers 150155 disturbedyellow band 155157 disturbedorganic veneers 157164 disturbedmassive silts&organic veneers 164168 disturbedorganic veneers 168172 yellowband
145160 145160 145160 160187
disturbedundulatingdarkerbandofveneersof organicdetritus poorlypronouncedorangemottledmassiveband poorlypronounceddarkerbandofveneersoforganic detritus orangemottledmassivesiltswithtracesofveneersof organicdetritus darkbandofdisturbedveneersoforganicdetritus, likelytwodiscretepopulations distinctyellowbandofmassivesilts,orangemottled towardsbase darkbandofveneersoforganicdetritus,partially disturbed undulatingpartiallydiscontinuousyellowbandof massivesilts wedgeofdisturbedveneersoforganicdetritus undulatingpartiallydiscontinuousyellowbandof massivesilts disturbedorganic veneers,mottled, gypsum
160187
TopSection TopSection TopSection TopSection TopSection TopSection
disturbed silts 172174 disturbedorganic disturbed veneers silts 174178 yellowband disturbed silts 178179 disturbedorganic disturbed veneers silts 179182 yellowband disturbed silts
160187 160187 160187 160187 160187
littledisturbed organicveneers
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 182183 disturbedorganic disturbed 160187 veneers silts 183199 disturbedmassive disturbed 187195 silts&organic silts veneers 199201 organicveneers 201203 yellowband 203204 organicveneers 204208 yellowband 208209 organicveneers 209210 yellowband TopSection TopSection TopSection TopSection 215220 disturbedyellow band 220225 disturbedorganic veneers 225229 disturbedyellow band 229231 disturbedorganic veneers 231233 disturbedyellow band 233234 organicveneers 210211 organicveneers 211212 yellowband 212215 organicveneers 195200 organic veneers yellowband 200208 yellowband 200208 yellowband 200208 organic veneers organic veneers organic veneers disturbed silts disturbed silts disturbed silts disturbed silts disturbed silts disturbed silts disturbed silts 208210 208210

283
TopSection
TopSection
TopSection TopSection TopSection TopSection TopSection
darkbandofveneersoforganicdetritus,partially disturbedanddiscontinuous orangemottleddisturbedunit,verticaltubes extendingfromoverlyingbandoforganicveneers (bioturbation),gypsumefflorescence,tracesof veneersoforganicdetritus darkbandofveneersoforganicdetritus wellpreservedthincoupletsofyellowbandsof massivesilts darkbandofveneersoforganicdetritus wellpreservedthincoupletsofyellowbandsof massivesilts darkbandofveneersoforganicdetritus wellpreservedthincoupletsofyellowbandsof massivesilts darkbandofveneersoforganicdetritus wellpreservedthincoupletsofyellowbandsof massivesilts darkbandoftwoveneersoforganicdetritus, sandwichingorangemottledmassiveband orangemottledmassivesilts,verticaltubesextending fromoverlyingbandoforganicveneers (bioturbation),gypsumefflorescence darkbandof34veneersoforganicdetritusin massive,partiallyorangemottledmatrix orangemottledmassivesilts,gypsumefflorescence darkbandof2veneersoforganicdetritusinmassive, partiallyorangemottledmatrix orangemottledmassivesilts,gypsumlinedvertical tubesextendingfromoverlyingbandoforganic veneers(bioturbation) discontinuousdarkbandofveneersoforganic detritus thinbandoforangemottledmassivesilts darkbandofveneersoforganicdetritus orangemottledmassivesilts,indistincttracesof veneersoforganicdetritus
discontinuous organicveneers
multipleorganic veneers
SSAMSANU 4206(198)
lessdistinctorganic veneers 20.100 2.100 (210)
208210 210244 210244
multipleorganic veneers(baseof 4er)
SSAMSANU 4110(213)
210244
TopSection TopSection TopSection TopSection TopSection TopSection TopSection TopSection TopSection
210244 210244 210244 210244
multipleorganic veneers(224225)
disturbed silts 234235 disturbedyellow disturbed band silts disturbed 235236 organicveneers silts 236240 disturbedmassive disturbed silts&organic silts veneers
210244 210244 210244 210244
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 240241 organicveneers disturbed 210244 silts 241244 disturbedmassive disturbed 210244 silts&organic silts veneers 244245 organicveneers disturbed 210244 silts

284
TopSection TopSection
darkbandof2veneersoforganicdetritusinmassive, partiallyorangemottledmatrix orangemottledmassivesilts,indistincttracesof veneersoforganicdetritus darkbandofmultipleveneersoforganicdetritus overlyingtufamarkerlayer multipleorganic veneers(248249 UpperSection) SSAMSANU 4111(244)
TopSection
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 249251 tufaband tufaband 244250 251252 organicveneers 252254 yellowband 254256 organicveneers 256260 disturbedyellow band 260262 disturbedtufa band 262270 disturbedmassive silts&organic veneers 270271 organicveneers tufaband 244250

285
Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section
distinctcontinuousbandoftufainfinegrainedmatrix darkbandofmultipleveneersoforganicdetritus underlyingtufamarkerlayer distinctcontinuousyellowbandofmassivesilts darkbandofmultiplecompactveneersoforganic detritus massiveorangemottledsiltswithstreaksoftufa discontinuousbandoftufa orangemottledmassivesiltswithstreaksoftufaand patchesofveneersoforganicdetritus darkbandof2veneersoforganicdetritus lessdistinctorganic SSAMSANU veneerbelow 4112(264) mottledzone multipleorganic veneers
yellowband 250255 yellowband 250255 yellowband 250255 laminated unit laminated unit laminated unit 255270 255270
255270
271274 yellowband 274275 organicveneers 275278 yellowband 278281 organicveneers 281291 yellowband
laminated unit laminated unit laminated unit organic veneers yellowband
255270 255270 255270 270275 275285
massivesiltswithtracesofveneersoforganicdetritus distinctdarkbandofmultipleveneersoforganic detritus massivesiltswithtracesofveneersoforganicdetritus darkbandofmultiplecompactveneersoforganic detritus distinctcontinuousyellowbandofmassivesilts,in placessandwichingupto2cmthickpatchesof veneersoforganicdetritus indistinctdarkbandofveneersoforganicdetritus massivesiltsdarkeningbecomingincreasingly disturbedandmixedwithoverlyingveneersof organicdetritus darkbandofmultipleveneersoforganicdetritus multipleorganic veneers
291292 disturbedorganic yellowband 275285 veneers 292298 disturbedyellow yellowband 275285 band 298299 organicveneers disturbed silts 285305
multipleorganic veneerstopping tufa(orange mottled)
299301 disturbedtufa band 301306 disturbedyellow band 306307 disturbedorganic veneers 307315 disturbedyellow band
disturbed silts disturbed silts disturbed silts disturbed silts
285305 285305
discontinuousbandoftufainorangemottledsilts orangemottledmassivesilts,verticaltubesextending fromoverlyingbandoforganicveneers (bioturbation),gypsumefflorescence darkbandofveneersoforganicdetritus,partially disturbedanddiscontinuous orangemottledmassivesilts,tracesoforganic veneers,gypsumefflorescenceandrootcasts
285305 285305
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 315317 organicveneers disturbed 285305 silts

286
darkbandofmultipleveneersoforganicdetritus
Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section Upper Section
lessdistinctorganic veneers,tiny charcoalfragments
317319 yellowband 319320
320323
323326
326330
disturbed silts organicveneers disturbed organic veneers yellowband disturbed organic veneers disturbedyellow disturbed band organic veneers disturbedorganic disturbed veneers organic veneers
285305 305315
undulatingyellowbandofmassivesilts,orange mottledtowardsbase darkbandofmultipleveneersoforganicdetritus
305315
massivesiltswithtracesofveneersoforganicdetritus andstreaksoftufa orangemottledmassivesilts,tracesoforganic veneers darkbandofveneersoforganicdetritusmixedwith massivesilts,topmostveneersisbestpreserved multipleorganic veneers,gypsum efflorescence(326 327)
305315
305315
330332 disturbedyellow band 332336 yellowband
disturbed 305315 organic veneers yellowband 315322
massivesiltsmixedwithveneersoforganicdetritus
distinctcontinuousyellowbandofmassivesilts
Appendix 5.3
Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) ExtendedPSA 344352 yellowband yellowband 315322 5Y7/25/2 Section 322335 352355 disturbedorganic disturbed ExtendedPSA veneers organic Section veneers 355359 disturbedtufa disturbed 322335 ExtendedPSA band organic Section veneers 322335 359363 disturbedorganic disturbed ExtendedPSA veneers organic Section veneers 363367 yellowband disturbed 322335 ExtendedPSA organic Section veneers 367368 organicveneers disturbed 322335 ExtendedPSA organic Section veneers ExtendedPSA 368370 yellowband yellowband 335343 Section ExtendedPSA 370371 organicveneers yellowband 335343 Section ExtendedPSA 371372 disturbedyellow yellowband 335343 Section band 372374 disturbedorganic yellowband 335343 ExtendedPSA veneers Section ExtendedPSA Section ExtendedPSA Section 374381 yellowband Subsection below TOP (incm)

287
distinctcontinuousyellowbandofmassivesilts darkbandofveneersoforganicdetritusinmassive, partiallyorangemottledmatrix bandofstreaksoftufaandveneersoforganic detritus darkbandofveneersoforganicdetritusinmassive, partiallyorangemottledmatrix,rootcastslinedwith gypsumefflorescence distinctalbeitdiscontinuousyellowbandofmassive silts,sandwichedbyveneersoforganicdetritus discontinuousdarkbandofmultiplecompactveneers oforganicdetritus thindiscontinuousyellowbandofmassivesilts discontinuousdarkbandofmultiplecompactveneers oforganicdetritus thinbandoforangemottledmassivesilts darkbandofveneersoforganicdetritusinmassive, partiallyorangemottledmatrixandstreaksoftufa multipleorganic veneers disturbedorganic veneers,mottled (361363) disturbedorganic veneers,mottled
multipleorganic veneers(topping continuousyellow band)
SSAMSANU 4113(338)
381383 organicveneers
ExtendedPSA 383387 yellowband Section 387401 disturbedorganic laminated unit veneers ExtendedPSA Section
yellowband 335343 2.5Y7/26/2 distinctcontinuousyellowbandofmassivesilts, morechromatic(pink)thenothermassivebandsin thesequence organic 343348 distinctdarkbandofmultipleveneersoforganic veneers detritus,partiallyinorangemottledmatrixand associatedwithstreaksoftufa yellowband 348355 undulatingyellowbandofmassivesilts 355373 massiveunitofmultipleveneersoforganicdetritus separatedbydiscontinuousbandsofmassivesilts
SSAMSANU 4114(341) multipleorganic veneers OZJ905(353) welldeveloped organicveneers, stronglyattachedto clay(391392)/ multipleorganic veneers(400401)
ExtendedPSA 401408 Section ExtendedPSA 408409 Section ExtendedPSA 409410 Section
disturbedyellow band organicveneers yellowband
laminated unit laminated unit laminated unit
355373 355373 355373
massivesilts,partiallyorangemottledandmixed withveneersoforganicdetritus discontinuousdarkbandofmultiplecompactveneers oforganicdetritus thindiscontinuousyellowbandofmassivesilts
Appendix 5.3
Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 410411 organicveneers laminated 355373 ExtendedPSA unit Section ExtendedPSA 411413 disturbedtufa Section band ExtendedPSA 413423 yellowband Section laminated 355373 unit yellowband 373380 Subsection below TOP (incm)

288
discontinuousdarkbandofmultiplecompactveneers oforganicdetritus
multipleorganic veneers(topping continuousyellow band)
SSAMSANU 4209(372)
weatheredbandofstreaksoftufaandveneersof organicdetritus distinctcontinuousyellowbandofmassivesilts
Appendix 5.3
Subsection below TOP (incm) 423426 426430 430431 431433 433438 Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) organicveneers laminated 380465 unit disturbedyellow laminated 380465 band unit disturbedorganic laminated 380465 veneers unit disturbedtufa laminated 380465 band unit disturbedyellow laminated 380465 band unit 380465 380465 380465 380465 380465 380465 380465

289
LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section
darkbandofmultipleveneersoforganicdetritus, cappedandprotectedbytufa orangemottledmassivesilts,tracesoforganic veneersanddispersedtufa darkbandofveneersoforganicdetritusmixedwith tufa,topmostveneersisbestpreserved discontinuousbandofstreaksoftufainorange mottledmatrix orangemottledmassivesilts,tracesoforganic veneersanddispersedtufa distinctbutdiscontinuousbandoftufainorange mottledsilts darkbandofmultipleveneersoforganicdetritus, streaksoftufa bandofmassivesilts,tracesofveneersoforganic detritusanddispersedtufa distinctbutdiscontinuousstreaksoftufa thindiscontinuousyellowbandofmassivesilts darkbandofveneersoforganicdetritus,disturbed orangemottledmassivesilts,tracesofveneersof organicdetritusandstreaksoftufa darkbandofmultipleveneersoforganicdetritus, streaksoftufaandpartiallymottled bandofmassivesilts,tracesofveneersoforganic detritus thinundulatingstreaksofdensetufa,partially associatedwithveneersoforganicdetritus bandofmassivesilts,tracesofveneersoforganic detritusanddispersedtufa darkbandofveneersoforganicdetritus,thinand discontinuous thinundulatingstreaksofdensetufa,partially associatedwithveneersoforganicdetritus bandofmassivesilts,tracesofveneersoforganic detritus darkbandofmultipleveneersoforganicdetritusin clayeymatrix distinctbutdiscontinuousstreaksoftufa thindiscontinuousyellowbandofmassivesilts sandwichedbyveneersoforganicdetritus
lensoforganic veneersatbaseof mottledunit
laminated unit 442443 organicveneers laminated unit 443447 yellowband laminated unit 447449 disturbedtufa laminated band unit 449452 yellowband laminated unit 452453 disturbedorganic laminated veneers unit 453458 disturbedmassive laminated silts&organic unit veneers 458460 organicveneers laminated unit 460461 yellowband laminated unit
438442 tufaband
380465
multipleorganic veneers,charcoal (457458)
SSAMSANU 4116(437) 21.600 1.000 (460)
380465
laminated unit laminated unit 465466 disturbedorganic laminated veneers unit laminated 466467 disturbedtufa band unit 467468 yellowband laminated unit 468469 organicveneers laminated unit 469470 disturbedtufa laminated band unit 470474 yellowband laminated unit
461462 disturbedtufa band 462465 yellowband
380465 380465 380465 380465 380465 380465 380465 380465
Appendix 5.3
Subsection below TOP (incm) Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 474475 organicveneers laminated 380465 unit 475479 disturbedorganic organic 465470 veneers veneers

290
LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section LowerPSA Section
darkbandofmultipleveneersoforganicdetritus multipledarkbandsofveneersoforganicdetritus, transitionalbasewithstreaksoftufa compositeof veneerstoppingred band(470479) SSAMSANU 2030(474) veneersatbaseof redband
479486 redband
redband
486488 organicveneers 488490 tufaband
laminated unit laminated unit laminated unit laminated unit laminated unit laminated unit laminated unit laminated unit laminated unit laminated unit yellowband
490492 disturbedyellow band 492493 tufaband 493494 disturbedyellow band 494495 organicveneers 495496 tufaband
470479 (5YR6/3dry distinctcontinuousredbandofmassivesilts 7.5YR6/3 moist) 479492 darkbandofmultipleveneersoforganicdetritus, streaksoftufa 479492 discontinuousbandoftufainfinegrainedmatrix, pocketsoflargertufaclasts,gradualtransition towardstop(upto485) 479492 undulatingbandoforangemottledmassivesilts, dispersedtufa,moreclayeytowardstop 479492 discontinuousbandoftufainfinegrainedmatrix 479492 479492 479492 undulatingbandoforangemottledmassivesilts, tracesofveneeroforganicdetritus darkbandofveneersoforganicdetritus discontinuousbandoftufainfinegrainedmatrix
charcoaltowards topofyellowband
496497 disturbedyellow band 497498 disturbedorganic veneers 498499 tufaband 499505 yellowband
479492 479492 479492
undulatingbandoforangemottledmassivesilts, dispersedtufa darkbandofveneersoforganicdetritus,thinand discontinuous discontinuousbandoftufainfinegrainedmatrix multipleorganic veneers(topping continuousyellow band) organicveneers toppingpalaeosol SSAMSANU 2035(495)
492500 2.5Y6/45/4 distinctcontinuousyellowbandofmassivesilts,in placestufafragmentfromoverlyingtufaband (bioturbation) 500506 thickbandofstreaksoftufa,veneersoforganic detritus,gradualtransitiontowardsbase
505510 tufaband
tufaband
SSAMSANU 4117&4205 (505)
Appendix 5.3
Subsection Lithostratigraphic Typesection Depthin MunsellSoil units units type Colors section (drymoist) (incm) 506516 palaeosol palaeosol 506516 (5YR6/1dry 7.5YR5/1 BaseSection moist) below TOP (incm)

291
distinctbrownhorizon/boundarybetweenunderlying chromaticbasalunit(I)andoverlyinglaminatedunits, wherewelldefinedprismatictoblockystructure
largecharcoalat baseofpalaeosol (514518)
OZJ909(508) & SSAMSANU 2036&2037 (516)
BaseSection BaseSection BaseSection BaseSection
516530 lighterband 530541 darkerband
lighterband 516530 indistinct darkerband
chromaticlighterbandofmassivesilts,gradualupper andlowertransition 530541 2.5Y6/25/2 indistinctdarkerband,wedgingoutdownstream
541548 lighterband 548570 darkerband 570575 tufaclasts
lighterband 541548 2.5Y7/35/3 chromaticlighterbandofmassivesilts,gradualupper andlowertransition darkerband 548570 2.5Y5/24/2 indistinctdarkerband,morepronouncedtowardstop lighterband 570603 2.5Y7/45/4 discontinuoussheetoftufaclastsatdownstream end,concentrationofshellsandcharcoal
BaseSection 575580 lighterband 580585 pebblesheet BaseSection 585603 lighterband BaseSection 603627 darkerband BaseSection
largecharcoal(542 543) smallcharcoal(563 564) largecharcoalat baseofdarkerband (572573),(573574)
BaseSection
lighterband 570604 2.5Y7/45/4 chromaticlighterbandofmassivesilts,gradualupper andlowertransition lighterband 570606 2.5Y7/45/4 discontinuoussheetof wellroundedsmall pebbles(mmupto occasional1cm),shells lighterband 570607 2.5Y7/45/4 chromaticlighterbandofmassivesilts,sheetsof rolleddetritalpedogeniccarbonates darkerband 603627 2.5Y6/25/3 darkerband,morepronouncedtowardstop
charcoalfragments Beta96679& atbaseofgravel 96116(565) sheet(584585) 36.100 2.700 (585) largecharcoal towardstopof 25.300 darkerband 1.400 (602603), (620) charcoal (617618) charcoalfragments OZJ904(632) (641644) & OZJ908(639) topfewwelldeveloped smallcharcoal OZJ907(667) large(fistsized) towardstopof calcareousrhizocretes darkerband (648649), charcoalfragments towardsbaseof darkerband (658661)
627647 lighterband BaseSection 647671 darkerband
lighterband 627647 2.5Y7/35/4 chromaticlighterbandofmassivesilts,gradualupper andlowertransition darkerband 647671 2.5Y6/25/3 darkerband,morepronouncedtowardstop
BaseSection
Appendix 5.3
Subsection below TOP (incm) 671

Lithostratigraphic Typesection Depthin MunsellSoil Finegrainedunits Graveldescription units units type Colors description section (drymoist) (incm) lighterband lighterband 671 2.5Y7/35/3 chromaticlighterbandofmassivesilts,gradualupper andlowertransition Pedogenic description 13Csamples description 14Csample references OSLages
292
welldevelopedlarge (fistsized)calcareous insiturhizomorphs
BaseSection
OZJ906(684) & SSAMS1811 (695) & 24.500 SSAMS2039 1.600 (695) (840)
Appendix 5.3
ID BG1 BG2 BG3 BG4 BG5 BG6 Imbrication Imbrication North East (in mag.) (true N in ) (in ) (in ) 250 31.3365 138.61281 243 31.3364 138.61443 31.3362 138.61447 31.3358 138.61534 31.3353 138.61604 31.334 138.61584 185 185 190 190 (180200) 10, upstream direction 290 135 178 178 183 183 3 Depth from TOP (in m) ?, but towards base 1 2 ?, but ~0.8 surface drape 3 2.5 lag, partially overlying bedrock outcrop Geometry
Gravel Imbrication Data

Lithology (remarks) channel-fill, 200 cm thick various sheet, 30 cm thick sheet, 30-40 cm thick sheet, 30-40 cm thick sheet, 40-50 cm thick various, largely < 5 cm various, largely < 5 cm various, largely < 5 cm various, largely < 5 cm, few local angular platy shales various, large, CaCO3-encrusted, mixed with large (50x30x20 cm) angular blocks of local shale in silty matrix various, large Location (remarks) main creek bank, lower gravel unit in side gully of Brachina Silts in side gully of Brachina Silts, fine gravel in grey unit towards TOP
293
in side gully of Brachina Silts, in red-brown unit in side gully of Brachina Silts, erosional basal contact in red-brown unit, sheet with small chutes in side gully from range front, mixed matrix-supported gravel unit is apparently deposited upstream into local gully from nearby older gravel terrace remnant?
BG7 BG8
31.3338 138.61569 31.3338 138.61569
283 128
2 m gravel-free Silts 3
30 cm upstream a gully sheet, 20-30 cm thick
BG9 (a)
31.3326
138.6188
230
223
5, 1 m above local bedrock
channel-fill, 40-50 cm
in side gully from range front, termination of allochthonous gravel band into sheltered embayment? matrix-supported mainly local angular in side gully of Brachina Silts platy shales, associated allochthonous gravel < 2 cm matrix-supported mainly local angular in main creek, small tributary (225) confluence, alluvial fan?, the large platy clasts (up to 10 cm), upstream end of the fan front is confined by a bedrock narrow (60 m) associated allochthonous gravel which is flanked by a massive 4-5 m thick and 30 m wide tufa indicating present trunk channel blockage over the aggradation of silts matrix-supported largely allochthonous gravel (shale?, but rounded), CaCO3-encrusted, matrixsupported large local angular clasts in lithic sands in main creek, small tributary (225) confluence, alluvial fan?, the upstream end of the fan front is confined by a bedrock narrow (60 m) which is flanked by a massive 4-5 m thick and 30 m wide tufa indicating present trunk channel blockage over the aggradation of silts in main creek, small tributary (225) confluence, alluvial fan?, the upstream end of the fan front is confined by a bedrock narrow (60 m) which is flanked by a massive 4-5 m thick and 30 m wide tufa indicating present trunk channel blockage over the aggradation of silts in main creek, small tributary (225) confluence, alluvial fan?, the upstream end of the fan front is confined by a bedrock narrow (60 m) which is flanked by a massive 4-5 m thick and 30 m wide tufa indicating present trunk channel blockage over the aggradation of silts
BG9 (b)
31.3326
138.6188
250
243
1.5
sheet, 30 cm thick
BG10 (a)
31.3333
138.6192
210
203
see BG9 (a)
150 cm thick
BG10 (b)
31.3333
138.6192
280
273
1.5
35 cm thick, 100 m wide
various, partially CaCO3-encrusted, rounded, well-rounded boulders towards base of band
BG11
31.3352 138.61862
230
BG12
31.336 138.61838
185
BG13 (a) BG13 (b) BG13 (c)
31.3365 138.61847 31.3365 138.61847 31.3365 138.61847
235 235 235
223 below surface drape, in upstream continuation overlying silts rapidly increase to up to 2 m 178 below surface drape, cutand-fill into red-brown silts 4 228 228 228 below surface drape
sheet, <200 c?, but only 40 m upstream thinning out to < 20 cm 40 cm thick, 10 m wide chute 200 cm thick, 4 m wide chute 100 cm thick, 3 m wide chute, similar loci sheet, discontinuous 20 cm thick
various, upward-fining with largest in main creek, widespread topmost gravel sheet above downstream clasts 10 cm, both allochthonous and confining bedrock outcrop, mantelled by surface drape local pebbles are well-rounded various in main creek, topmost gravel unit
various, coarse gravel ~ 10 cm various, finer gravel ~ 5 cm various
in main creek, lowermost gravel unit of three stacked generations of chutes in main creek, central gravel unit of three stacked generations of chutes in main creek, topmost gravel unit of three stacked generations of chutes
Appendix 5.3
Imbrication Imbrication North East (in mag.) (true N in ) (in ) (in ) BG14 31.3368 138.61858 190 183 ID BG15 (a) BG15 (b) BG16 BG17 31.3368 138.61876 31.3368 138.61876 31.3369 138.61928 31.3386 138.61227 230 170 195 265 223 163 188 258 Depth from TOP (in m) Geometry

Lithology (remarks) various, finer gravel ~ 5 cm various various various
294
Location (remarks) in main creek, central gravel unit of three stacked generations of chutes in main creek, central gravel unit of three stacked generations of chutes in main creek, topmost gravel unit of three stacked generations of chutes in main creek, central gravel unit of three stacked generations of chutes
200 cm thick, 4 m wide chute 150 cm thick, 4 m wide chute below surface drape 150 cm thick, 4 m wide chute 4 200 cm thick, 3 m wide chute 4 ?, but overlying outcrop 30 cm thick, 30 m wide
BG18
31.3412 138.60826
260
253
300 cm thick, 45 m wide channel
BG19 BG20
31.3417 138.60841 31.3402 138.60609
?? 255
248
gravel lag 4
widespread sheet small chute
up to 20 cm well-rounded partially in Etina Creek which here becomes a bedrock-confined channel (30 m, CaCO3-encrusted clasts among finer later 25 m width until and beyond downstream confluence) allochthonous gravel various, partially CaCO3-encrusted, in Etina Creek, 15 m downstream of bedrock-confined channel rounded, well-rounded boulders construction towards base and up to 100x40x30 cm large angular local shales in silt matrix topped by at least another meter of gravel various in Etina Creek at confluence with large local tributary platy angular to rounded pebbles of small chute just upstream Brachina of the confluence, in slackwater various lithology in lithic-sands matrix, deposit distinctly orange-mottled, at the base large angular clasts up to 10 cm various, well-rounded small resting on continuous bedrock outcrop separating Etina and Brachina Creeks resting on continuous bedrock outcrop separating Etina and Brachina Creeks continuation of gravel lag across the Etina from the tributary?
BG21
31.3401 138.60857
280-330
198
1.5
BG22
31.3403 138.60811
330
323
below surface drape
multiple discontinuous sheets, 10-20 cm thick, ~100 m wide additional uppermost sheet, up to 40 cm thick widespread sheet, m wide 120 cm thick, 25 m wide chute
various, well-rounded small
BG23 BG24
31.3407 138.60823 31.3379 138.63074
?? 285 278
gravel lag 1
various
BG25 (a) BG25 (b) BG26
31.3371 138.62952
265
258
31.3371 138.62952
220
213
2.2
120-150 cm thick, 60 m wide, multiple sheets of gravel chute among gravel sheets
various, mix of platy (~80%), largely 5-just 50 m downstream of Nuccaleena Formation (laminated dolomite 62010 cm 610 Ma) standing out as a prominent ridge confining the Brachina Silts limited to the Brachina Formation (siltstone & shale 600-610 Ma), perhaps the cut-and-fill) matrix-supported and open boxwork 20 m wide and 2 m thick tufa plug at downstream end with some reed fill mostly platy sub-rounded shales (at inclined to 190, in proximity to raised older gravel terrace remnants least 3 events) open boxwork 20 m wide and 2 m thick tufa plug at downstream end with some reed fill inclined to 190, in proximity to raised older gravel terrace remnants various, partially CaCO3-encrusted, topmost gravel sheet rounded, well-rounded boulders with ~30cm diameter, partially matrixsupported, pebbles ~3 cm up to 10 cm
31.3367 138.62749
220
213 0.3, below surface drape 30-40 cm thick, 25 m wide
Appendix 5.3
Imbrication Imbrication North East (in mag.) (true N in ) (in ) (in ) BG27 31.3371 138.62624 165-180 165 ID

Depth from TOP Geometry (in m) up to 200 cm thick, 70m top of gravel bench wide elevated above the Brachina Silts, bedrock terrace ~5 m from present level of incision Lithology (remarks) mix of rounded and angular large pebbles (~10 cm ) and boulders (some 40x20x20 cm)
295
Location (remarks) gravel unit resting unconformately on bedrock (Brachina shales), partially up to 1 m thick pockets of silt-coloured tufa matrix incorporating weathered (shattered) pebbles, pebble bench sourced from older gravel terrace remnants within the Brachina Silts. These older gravel terrace remnants consist of allochthonous, rounded, partially CaCO3-encrusted pebbles and boulders (up to ~1 m!), resting on bedrock outcrops (Brachina shales), and can attain heights of up to ~8 m above the Brachina Silts. They present a local source of CaCO3-encrusted rounded pebbles of various lithology and predate the fine-grained aggradation.
BG28
31.3385 138.62312
215
208
no TOP
~150 cm thick(?), ~100 m wide
Tufa dams occur upstream of narrow trunk channel confinements, are attached to the bedrock (Brachina shales), consist of at least 2 generation fluvial in nature, as demonstrated by imbricated near-horizontal gravel (angular local shales). Towards the top up the sequence, detrital tufa clasts and blocks are incorporated. various, partially CaCO3-encrusted, gully section from Etina Creek, gravel lag sourced from nearby older grave rounded gravel, some weathered and terrace remnant, few block were transported as tufa-cemented clasts shattered, mostly matrix-supported, few boulders up to 40x30x20 cm various, matrix-supported various, open boxwork topmost generation of gravel sheet in silts extending laterally above the tufa structure central chute, just downstream of bedrock outcrop with channel located between bedrock outcrop and tufa structure (incorporating local pebbles aligned a 10-15 degree slope lowermost generation of gravel sheet gravel sourced from nearby older gravel terrace remnant (100 m in 190, elevated ~2 m above Brachina Silts), strongly weathered and partially shattered ~1 m blocks (tillite?), tufa in side gully from Etina Creek, extensive gravel sheet partially resting on bedrock outcrop and covered by surface drape, sourced from nearby older gravel terrace remnant in side gully from Etina Creek, gravel unit resting on bedrock within grey unit The grey unit in locations of its furthest lateral extent rests on local bedroc (Brachina Shales) incorporating gravel in its discoloured matrix at its base, it is ~ 1.5 m thick followed by 0.5 m transitional brown and 2.5 m redbrown silts and likely represents a water-logged environment. It can be subdivided into a discreet unit with gravel alluviation at its fringes, followed by dark grey units with abundant gypsum efflorescence, larger root channels (partially lined with red-brown silts). The grey unit is clearly confined to a central unit aggrading over peak the LGM
BG29 (a) BG29 (b) BG29 (c) BG30
31.339 138.62244 31.339 138.62244
290-320 300
298 293
below surface drape 1
combined 250 cm thick, 70 m wide central chute up to 250 cm thick combined 250 cm thick, 70 m wide gravel lag, at least 100 m wide radius around elevated older gravel terrace remnant 200 cm thick, 20 m wide (outcroping?, but more extensive) gravel sheet 100 cm thick, 6 m wide chute
31.339 138.62244 31.3391 138.61951
290-320 ??
298
1.5 gravel lag
various, matrix-supported various, CaCO3-encrusted rounded pebbles and boulders, matrixsupported various, CaCO3-encrusted rounded pebbles (5-10 cm) and boulders (up to 30 cm), matrix-supported various, CaCO3-encrusted rounded pebbles (~5 cm, sorted) in grey unit
BG31
31.3406 138.61911
285
278
below surface drape
BG32
31.3406 138.61841
330
323
Appendix 5.3
Imbrication Imbrication North East (in mag.) (true N in ) (in ) (in ) BG33 31.3396 138.61501 350 (?) 343 ID Depth from TOP (in m) 4.5 Geometry

Lithology (remarks) 20-30 cm, all in all ~ 400 various, CaCO3-encrusted rounded m! pebbles (~5 cm, sorted) in grey unit 150 cm thick, 60 m wide
296
Location (remarks) extensive gravel sheet extending towards Etina Creek, sitting on yellow/orange-mottled silts, ~2.5 m of grey unit resting on gravel sheet above gravel chute topped by at least 1 m of silts
BG34
31.3356 138.62054
195-210
195
2 (?)
BG35
31.3371 138.60.689
90 or 270
83 0.25 below yellow band in 10 cm thick, 30 cm wide basal unit (I) of BRA-SD single line of up to 4 clasts
BG36 (a) BG36 (b) BG37
31.3391 138.60744
230-250
233
1.5
30 cm thick, ~5 m wide
31.3391 138.60744 31.3398 138.60623
270 270-280
263 268
0.5 1
10 cm thick, few m wide 100 cm thick, 30 m wide (?)
BG38
31.3401 138.60577
255
248 4 ?, but overlying outcrop 200 cm thick (5 couplets), 11 m wide
BG39
31.3395 138.60583
260-330, and 10
300 cm thick (multiple couplets), 30 m wide
even mix of large angular local platy shales (~40x20x10 cm) and ~10 cm rounded CaCO3-encrusted pebbles of various lithology (ranging from 5-20 cm), lower half finer, well-sorted and imbricated parallel rounded CaCO3-encrusted single narrow line of gravel coming out of BRA-SD in the uppermost basal clasts of various lithology and tufa in unit (I) silts of basal unit (I), partially weathered (from some local older gravel terrace remnant?) largely matrix-supported larger (5-10 lower restricted gravel sheet in grey unit cm) CaCO3-encrusted rounded grave of various lithology lithic sand-supported largely platy?, upper restricted gravel sheet just below uppermost well-developed calcic but rounded fine (1-2 cm) gravel horizon matrix-supported mainly local angular prominent gravel sheet platy shales (5-10 cm), ~20% allochthonous rounded gravel platy angular to rounded pebbles of side gully succession of at least 5 bands of orange-mottled gravels various lithology in lithic-sands matrix, separated by ~10 cm thick bands of light to purple silts, pebble units are distinctly orange-mottled, at the base associated with tufa clasts up to ~10 cm and often topped by large angular clasts up to 10 cm discontinuous sheets of tufa (in-situ), separating light Silt bands (Sf) are similar to massive yellow/ pink bands and partially exhibit faint crossbedding, not clearly a cut-and-fill structure, one line to BG20! platy angular to rounded pebbles of in side gully: succession of multiple bands of orange-mottled and fresh various lithology in lithic-sands matrix, gravel separated by ~10 cm light and darker mottled Silts, reverse flow distinctly orange-mottled, with imbrication (herringbone pattern) within grey silts, lateral continuation of unmottled units towards the top of the the orange-mottled bands in form of bog iron concretions in yellow silt unit matrix similar to those found in the basal (I) and uppermost (IV) units of BRA-SD downstream end of the section, while imbrication is ambiguous in the mottled units it is more clear in the fresh gravel (10), a laminated BRA SD equivalent section is close by (80 m) largely platy unsorted (1-10 cm) from Brachina Creek, channel-fill of 12 slackwater couplets resting on 2.5 matrix-/ lithic sand-supported gravel of m of grey silts which bank against bedrock (present exposure 3 m), gravel various lithology, partially orange occurs in bands of 10-50 cm thickness alternating with 5-10 cm thick mottled, some incorporating tufa massive light silts, gravel bands often topped by discontinuous sheets of clasts tufa
BG40
31.3401 138.60533
230-250, many reversals at 10
233
400 cm thick (12 couplets), 6 m wide
Appendix 5.3
297
Cell: E1 Comment: magnetic inclination corrected based on the latest International Geomagnetic Reference Field (IGRF) model (http://www.ngdc.noaa.gov/geomagmodels/struts/calcDeclination)
Appendix 5.3 A SampleID 1 (condition) 2 drape(FD) 3 drape(MD) 4 000010(FD) 5 000010(MD) 6 093100(FD) 7 093100(MD) 8 244250(FD) 9 244250(MD) 10 275285(FD) 11 275285(MD) 12 348355(FD) 13 348355(MD) 14 373380(FD) 15 373380(MD) 16 470479(FD) 17 470479(MD) 18 492500(FD) 19 492500(MD) 20 506516(FD) 21 506516(MD) 22 650660(FD) 23 650660(MD) B 1stpopulation (inm) 54 62 61 78 43 53 40 42 64 71 63 65 51 49 59 59 52 51 63 59 54 65 C proportion (in%) 66.6 72.8 42.8 75.3 78.6 54.3 78.8 65.1 78.0 80.9 72.2 66.8 82.3 94.5 89.3 93.4 61.7 68.3 65.3 81.1 56.1 82.2 D sigma 0.2 0.2 0.2 0.2 0.2 0.1 0.2 0.1 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.1 0.1 0.2 0.2 0.2 0.2
Sediment-Sizing Data . E F 2ndpopuplation proportion (inm) (in%) 14 20 15 25 11 30 14 18 16 24 23 32 12 12 13 15 30 27 13 12 85 14 29.0 21.2 50.4 21.0 18.6 41.1 19.1 29.2 16.9 11.6 24.4 29.1 15.2 5.6 9.0 4.8 31.4 27.5 22.8 14.1 34.5 10.6 G sigma 0.2 0.1 0.3 0.2 0.3 0.2 0.2 0.2 0.3 0.2 0.3 0.2 0.3 0.2 0.3 0.2 0.2 0.2 0.3 0.1 0.2 0.2 H 3rdpopulation (inm) 5 133 5 242 4 10 5 8 5 207 5 10 4 19 4 9 7 9 5 7 12 166 I proportion (in%) 3.9 11.2 6.8 3.7 2.7 4.6 2.0 5.8 3.1 4.5 3.4 4.1 2.5 0.0 1.7 1.7 6.9 4.2 11.0 4.7 9.4 7.3 J sigma 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.1 0.2 0.1 0.1 0.0 0.3 0.1
298
Appendix 5.3 A SampleID 1 (condition) 2 drape(FD) 3 drape(MD) 4 000010(FD) 5 000010(MD) 6 093100(FD) 7 093100(MD) 8 244250(FD) 9 244250(MD) 10 275285(FD) 11 275285(MD) 12 348355(FD) 13 348355(MD) 14 373380(FD) 15 373380(MD) 16 470479(FD) 17 470479(MD) 18 492500(FD) 19 492500(MD) 20 506516(FD) 21 506516(MD) 22 650660(FD) 23 650660(MD) K silt (in%) 73 52 77 44 82 81 89 92 56 46 59 60 68 68 53 52 79 80 61 61 51 48 L veryfine sand(in%) 23 37 18 39 15 19 10 8 36 40 38 38 29 32 45 46 19 20 30 31 39 38 M finesand (in%) 3 11 3 15 0 0 0 0 8 14 3 2 0 1 2 2 0 0 6 7 9 14 N rest (in%)
Sediment-Sizing Data . O >200m%, wetsieved 27.0 1 0 2 2 2 0 1 0 1 1 0 0 2 0 1 0 1 0 4 0 1 1 8.7 0.5 0.4 0.7 0.6 0.5 9.5 1.1 1.1 38.0 P mean (MS3) 46.8 69.5 39.4 82.7 40.1 45.1 37.6 35.4 63.6 77.9 57.0 58.0 48.5 52.8 60.1 61.8 45.1 46.1 52.4 58.7 68.8 75.6 Q median (MS3) 42.4 60.8 26.1 69.4 40.4 44.3 37.0 33.6 57.1 66.2 55.5 55.6 49.0 51.2 59.8 60.8 45.5 44.9 47.7 51.8 61.4 64.5 R mode (MS3) 55.2 62.3 50.8 78.7 53.1 51.3 50.2 43.6 62.3 71.6 63.2 62.2 59.0 59.6 71.2 69.3 54.1 54.4 62.4 60.1 61.8 64.2 S skewness (MS3,%) 1.4 1.1 1.6 1.5 0.2 0.6 0.3 0.7 1.5 1.3 0.6 0.6 0.2 0.5 0.2 0.4 0.2 0.5 1.2 1.4 1.4 1.3 T kurtosis (MS3,%) 3.1 1.2 2.7 2.5 0.4 0.8 0.4 0.4 3.4 1.9 0.6 0.4 0.3 0.4 0.1 0.3 0.1 0.5 2.3 2.8 2.8 1.8
299
Appendix 5.3 Cell: P1 Comment: smootheddata(groupsofseven) Cell: Q1 Comment: smootheddata(groupsofseven) Cell: R1 Comment: smootheddata(groupsofseven) Cell: S1 Comment: smootheddata(groupsofseven) Cell: T1 Comment: smootheddata(groupsofseven) Cell: H2 Comment: truncated Cell: H4 Comment: truncated Cell: H19 Comment: truncated Cell: H20 Comment: truncated Cell: H21 Comment: truncated
Sediment-Sizing Data .
300
Loess and Floods:
Appendices (5.4 Conference Abstracts)
301
Conferenceabstracts

Appendices (5.4 Conference Abstracts)
302
Ap p en d ices (5.4.1 Co n f er en ce Ab st r act )
303
A terminal Last Glacial Maximum (LGM) loess-derived palaeoflood record from South Australia?
(XVII INQUA Congress 2007, Cairns. Similar version published in Quaternary International (2007) 167168 Supplement, 15) David Haberlah, Martin A.J. Williams, Steven M. Hill, Galen Halverson, Peter Glasby Late Pleistocene fine-grained valley-fill formations have been reported in the uplands of present-day desert margins, including the Great Escarpment of Namibia, the Sinai Peninsula of Egypt, the Matmata Hills of Tunisia, and the Flinders Ranges of South Australia. The valley-fills or Silts mantled the underlying irregular bedrock topography and now form gently sloping terraces incised by ephemeral traction-load streams. Their homogeneous silty texture and massive to crude horizontal bedding, with occasional fine laminations that can be traced laterally for tens of metres, have prompted a number of conflicting depositional models (Haberlah 2006,
http://crcleme.org.au/Pubs/Monographs/regolith2006/Haberlah_D.pdf ). Reworked desert loess has been invoked as a source for these homogeneous Silts at all locations (Coud-Gaussen et al., 1984 SGB 37:1050; Eitel et al., 2001 QI 76/77:57; Rgner et al., 2004 LNES 102:79; Williams and Nitschke, 2005 SAGJ 104:25). This realisation has major implications for their palaeo-environmental interpretation. While most recent sedimentological studies suggest that the Silts represent loessderived alluvium (Rgner et al., 2004, Srivastava et al., 2006 QR 65:478), their ages are still contested and for example range at the Homeb Silts stratigraphic type section in Namibia from LGM (Vogel, 1982 Palaeoecol Afr 15:201; Eitel and Zller, 1996 MGG 137:245) to mid-Holocene (Bourke et al., 2003 QSR 22:1099). However, a chronostratigraphy based on 37 AMS
14
C-ages and 7 OSL dates,
exists for Silts in the Flinders Ranges (Williams et al., 2001 QI 83-85:129). New results from a continuous monolith 8m long obtained from the dated stratigraphic type locality of the Brachina Creek are presented. Hyperspectral mineralogical logging at 1cm intervals targeting alternating lightand dark-coloured bands are linked with results of Coulter Multisizer particle size and XRF/XRD analyses, suggesting slackwater deposition due to backflooding of a tributary mouth upstream of the loess-choked Brachina Gorge. Stable isotope and mollusc assemblage studies on carbonate shells are consistent with this model of palaeoflood deposition. Calibrated 14C-ages indicate three episodes of rapid aggradation at ~20.2 ka, ~19.4 ka and ~18.2 ka which, when compared with independent regional palaeo-environmental proxy data, suggest early deglacial orographically enhanced frontal precipitation events causing rapid erosion of LGM loess slope mantles. Other occurrences of Silts can now be compared with these fine-resolution sedimentological studies and palaeo-environmental
304
data embedded in a well-established chronostratigraphy. The loess-derived palaeoflood record also sheds new light on palaeoclimatic processes operating in southern Australia during one of the most complex intervals of the late Pleistocene.
305
The Flinders Silts: a last glacial alluvial loess record from South Australia
(7th International Conference on Geomorphology 2009, Melbourne) David Haberlah 1, 2, Martin A. J. Williams 3, Steven M. Hill 1, 2, Galen Halverson 1, Amy Suto 1, Peter Glasby 3, Alan R. Butcher 4, Tomas Hrstka 5, 6
1
Geology & Geophysics, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia (E-mail: david.haberlah@adelaide.edu.au) Cooperative Research Centre for Landscape Environments and Mineral Exploration (CRC LEME)
Geographical & Environmental Studies, School of Social Sciences, University of Adelaide, Adelaide, SA 5005, Australia FEI Australia, Brisbane, QLD 4064, Australia SGS Minerals Services, QLD 4064, Australia Institute of Geology, Czech Academy of Science, Rozvojova 269, 165 02 Prague 6-Lysolaje, Czech Republic
Remnants of late Pleistocene loess-derived valley-fills (Silts) are a common occurrence in mountainous catchments downwind of deserts, formerly exposed continental shelf areas and large playa lakes. The Namib Silts within the Great Escarpment of Namibia and the Sinai Silts of Egypt have been studied by Quaternary geomorphologists for decades. However, their depositional nature remains a matter of controversy, making it difficult to interpret their rich palaeo-environmental archives. Here we present the largely unnoticed but equally spectacular Flinders Silts from South Australia. Situated in the midst of the last glacial continental dust bowl, the longitudinal Flinders Ranges trapped large quantities of proximal dust. At present, the Silts are entrenched up to 18 meters by ephemeral traction load streams. Fifteen stratigraphic sections from four catchments were logged and dated by 94 AMS radiocarbon and 30 OSL ages. All sections are put in geomorphological context by stratigraphic mapping, and, within the Brachina catchment, by means of a differential GPS survey. Multi-proxy studies involving high-resolution parametric particle-size analyses, magnetic susceptibility, ICP-MS, carbon, oxygen, and nitrogen isotope geochemistry and spectral mineralogy employing HyChips and QEMSCAN technologies were performed. The results of the regional study indicate that over the last glacial cycle, particularly over its culmination in the Last Glacial Maximum (213 ka), loess blown into the catchments was episodically entrained and redeposited by floods, choking narrow gorges and backflooding into tributaries and embayments. Here, successive floods resulted in the deposition of layered to laminated slackwater deposits, presenting a near continuous terrestrial palaeo-environmental archive. The multi-proxy record
306
suggests that the last glacial represents an overall arid but complex interval, significantly altering the landscape by dust storms and flood events. The study of fine-grained valley-fills from south-eastern Australia can help to resolve the prolonged discussion over the nature of Silts in other parts of the world.
307
Dust fingerprinting in regolith: an integrated high-resolution parametric particle-size analysis quantitative spectral mineralogy approach
(7th International Conference on Geomorphology 2009, Melbourne) David Haberlah 1, 2, Steven M. Hill 1, 2, Craig Strong 3, Alan R. Butcher 4, Grant H. McTainsh 3, Tomas Hrstka 5, 6
1
Geology & Geophysics, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia (E-mail: david.haberlah@adelaide.edu.au) Cooperative Research Centre for Landscape Environments and Mineral Exploration (CRC LEME) Atmospheric Environment Research Centre, Griffith University, Brisbane, QLD 4111, Australia FEI Australia, Brisbane, QLD 4064, Australia SGS Minerals Services, QLD 4064, Australia Institute of Geology, Czech Academy of Science, Rozvojova 269, 165 02 Prague 6-Lysolaje, Czech Republic
Loess-derived alluvium is widely recognised in; mountainous catchments downwind of deserts, formerly exposed continental shelf areas and large playa lakes. A prominent example from South Australia is the late Pleistocene Flinders Silts; up to 18 m thick fine-grained valley-fill deposits choking narrow gorges and mantling the piedmont plains of the Flinders Ranges. In order to assess their impact on landscape evolution, deposition rates of dust need to be estimated for a given time period and area. This is best achieved by quantitatively differentiating the allochthonous aeolian component from in-situ weathering products within chronostratigraphic sequences along a downwind transect. Here, we present the results of an integrated approach employing highresolution three-dimensional particle-size analyses using a Multisizer 3 COULTER COUNTER and automated mineralogical QEMSCAN technology, a combination of features found in other analytic instruments such as Scanning Electron Microscopes (SEM) and Electron Probe Micro Analysers (EPMA). Parametric analysis of the particle-size data is performed on both fully-dispersed and minimally-dispersed sediment samples, the latter approximating transport-stable conditions in turbulent fluvial flow. Aggregation is further explored by mapping the sample mineralogy, thus visualising the spatial distribution of minerals in the form of discrete particles and compound aggregates. Disaggregation is digitally performed by the software package iDiscover. The results underline the importance of analysing soils and sediments in its naturally-occurring partiallyaggregated state, with certain forms of aggregation being indicative of fluvial transport. Apart from the palaeo-environmental implications, there are many other potential fields of application of
308
integrated quantitative dust fingerprinting in regolith, ranging from more efficient prospecting and mining to monitoring of human health and climate change.
(15th International Joint Seminar on the Regional Deposition Processes in the Atmosphere (RDPA) and Climate Change 2009, Taipei. Extended abstract) David Haberlah 1, Tomas Hrstka 2, Patricio Jaime 3, Alan R. Butcher 3, Grant H. McTainsh 4
Geology & Geophysics, School of Earth and Environmental Sciences, University of Adelaide, Adelaide, SA 5005, Australia (E-mail: david.haberlah@adelaide.edu.au)
2 1
Last glacial dust cycles in South Australia: employing advanced Automated Mineralogy and sediment-size analyses in the study of provenance, transport and depositional palaeo-environments
309
Institute of Geology, Czech Academy of Science, Rozvojova 269, 165 02 Prague 6-Lysolaje, Czech Republic FEI Australia, Brisbane, QLD 4064, Australia Atmospheric Environment Research Centre, Griffith University, Brisbane, QLD 4111, Australia
Abstract
More than 10 m thick mantles of loess-derived alluvium cover the plains and block the drainage lines of the semi-arid Flinders Ranges in South Australia. The fine-grained sequences aggraded episodically throughout the last glacial cycle and present an important palaeo-environmental archive of dust storms, fluvial activity and soil formation. In order to interpret this record in terms of climate change, the nature the homogenous sediments must be firmly established. Here we employ advanced quantitative mineralogy and parametric sediment-sizing to establish the sediment source, former transport pathways and depositional environments, and the subsequent impact of weathering and soil formation. The results indicate a largely aeolian provenance and downwindfining of the sediments, partially masked by fluvial reworking and deposition in floodplain and local slackwater environments. Automated Mineralogy and parametric sediment sizing are demonstrated to be powerful new tools in advancing our knowledge of the impact of late Quaternary dust cycles on the Australian landscape.
Keywords
Loess, Flinders Silts, Automated Mineralogy, parametric Particle-Size Analysis
310
Introduction
The last glacial cycle of the Flinders Ranges in semi-arid South Australia is characterised by multiple intervals of rapid aggradation of fine-grained valley-fill deposits (Flinders Silts), alternating with intervals of non-deposition, soil formation and erosion. The very homogenous aggradational sequences are preserved as up to 18 m thick terrace remnants, recording climate and environmental changes throughout the Last Glacial Maximum (LGM). Few other terrestrial deposits from the Southern Hemisphere provide a continuous record for this interval between 21 3 ka, marked by peak continental aridity, large-scale wind erosion of lake beds, and actively migrating dune fields. However, prior to this study the significance of the Flinders Silts as a palaeo-environmental archive was limited by uncertainties in the nature of the fine-grained material and its mode of deposition. Here, we combine an Automated Mineralogy technology (QEMSCAN) and a parametric particlesizing approach to infer provenance, transport and depositional environments of the fine-grained sediments.
Methods
Sediment samples of similar LGM age (~19 ka) established by optically-stimulated luminescence dating (Haberlah et al., in press) were collected from three prominent terrace remnants along an approximate downwind (SW-NE) transect in conjunction with underlying bedrock samples. All samples were analysed by the Automated Mineralogy QEMSCAN technique. QEMSCAN provides quantitative mineralogical data in spatial context by combining features found in other analytic instruments such as Scanning Electron Microscopes (SEM) and Electron Probe Micro Analysers (EPMA). Unmatched speed of analysis is achieved using four Energy Dispersive X-ray detectors. Purpose-built image analysis software allows the data on the mineralogy and textures of consolidated and dispersed sample specimens to be extracted on a particle-by-particle basis, sample for sample. Particles and aggregates are explored by micron-scale mapping of the mineralogy and texture of the surface of the sample, visualising the spatial distribution of minerals in form of discrete particles and compound aggregates (Fig.1). Disaggregation and the generation of 2-D size distributions are performed digitally by the iDiscover software (Fig.2). The data are used to discuss provenance, depositional environments and the nature of aggregation, weathering and diagenesis (see also Haberlah et al., accepted). 3-D Multisizer 3 COULTER COUNTER sediment-size data complement the study. Transport-stable mud aggregates that would survive turbulent fluvial flow are quantified in minimally-dispersed partially-aggregated sample condition and juxtaposed with corresponding fully-dispersed sample expressions (Haberlah and McTainsh, submitted), (Fig.3). Parametric resolution of the size distributions into individual Gaussian components interpreted as principal sediment populations has
311
recently been applied in the provenance study of Chinese loess (Sun et al., 2002). In a similar approach, we perform decompositional curve-fitting employing the freely-available open source statistical software Mixdist (Macdonald and Du, 2004) in the R environment of statistical computing (R Development Core Team, 2009). Here, three unconstrained Component Populations (CPs) are computed for a similarly aged subset of samples and compared (Figs. 3 and 4).
Results
The QEMSCAN mineral maps allow differentiation of the allochthonous aeolian component from insitu weathering products of the underlying bedrock by establishing respective mineral compositions and mineral-size ranges of detrital components (Fig.1). Winnowing of formerly active dune fields is suggested as a potential source of the aeolian material (Fig.2). Average quartz grain sizes are demonstrated to decrease along the downwind transect (Table 1). In a single stratigraphic context, average quartz grain sizes can potentially be used as a palaeowind proxy over time (Haberlah et al., accepted). Detailed 3-D size-information is provided by the parametrically-resolved Multisizer 3 data, suggesting a downwind-fining trend expressed by a gradual shift in the relative contribution of the coarse silt- and fine silt-sized component populations (Fig.3 and 4). Juxtaposition of corresponding minimally- versus fully-dispersed sample expressions suggests that aggregation predates fluvial deposition and, in some localities, occurred in backflooded slackwater environments (Haberlah and McTainsh, submitted). In contrast, post-depositional pedogenic aggregates in the present surface drape are larger in size and poorly sorted in terms of mineral composition (Fig.1).
Conclusion
The present study indicates that the fine-grained sediments are the product of windblown dust possibly generated from source-bordering dune fields downwind of deflated playa lake beds and intercepted by the longitudinal Flinders Ranges. Within the catchment of the Ranges, this proximal dust was deposited in form of loess mantles which were repeatedly eroded and reworked by highmagnitude low-frequency flood events. Hence, the Flinders Silts represent remnants of largely dustderived floodplains and slackwater deposits that extend from the inner ranges towards the piedmont plains. The onset of Deglacial climatic amelioration resulted in the spread of a perennial plant cover that stabilised dunes and valley slopes. This translated into dust starvation which resulted in the incision and ongoing entrenchment of the Flinders Silts. At present, ephemeral traction load streams erode and move the loess-derived alluvium towards the flanking terminal basins. The present study highlights the importance of analysing dust-derived soils and sediments in terms of individual mineral grain sizes and transport-stable aggregates. Apart from improving our palaeo-environmental understanding of past climate change, other potential applications for this
312
integrated quantitative dust fingerprinting approach, range from: more efficient regolith prospecting and mining to improved monitoring of human health in polluted urban environments.
References
Haberlah, D. and G. H. McTainsh, submitted: Quantifying aggregation in sediments. Sedimentology. Haberlah, D., Glasby, P., Williams, M. A. J., Hill, S. M., Williams, F., Rhodes, E. J., Gostin, V., OFlanery, A., Jacobsen, G. E., in press: Of droughts and flooding rains: an alluvial loess record from central South Australia spanning the last glacial cycle. In: Bishop, P. and B. Pillans (editors), Australian Landscapes. Geological Society Special Publications, London. Haberlah, D., Williams, M. A. J., Halverson, G., Hrstka, T., Butcher, A. R., McTainsh, G. H., Hill, S. M., Glasby, P., accepted: Loess and floods: high-resolution multi-proxy data of Last Glacial Maximum (LGM) slackwater deposition in the Flinders Ranges, semi-arid South Australia. Quaternary Science Reviews. Macdonald, P. D. M., Du, J., 2004: Mixdist. Mixture Distribution Models. R package version 0.5-1. http://www.math.mcmaster.ca/peter/mix/mix.html [last accessed 11/2009]. R Development Core Team, 2009: R. A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria. http://www.r-project.org/ [last accessed 11/2009]. Sun, J., 2002: Provenance of loess material and formation of loess deposits on the Chinese Loess Plateau. Earth and Planetary Science Letters, 203, 845-859.
313
Figure 1) QEMSCAN imagery of 3 LGM fluvio-aeolian regolith samples (top row), 3 underlying Palaeozoic bedrock specimens (middle row), and a palaeosol, surface drape and dune sand sample (bottom row). Samples were taken across an approximate downwind transect from the western piedmont plain Hookina catchment (H), the central ranges Brachina catchment (B) and the eastern ranges Wilkawillina catchment (W). Sample depths from the top of sections HK-M, BRA-SD and WLFP are indicated in centimeters and correlate in age to ~19 ka (Haberlah et al., in press). The QEMSCAN data is presented as colour-coded mineral maps showing textural size and shape attributes. A graphic scale is presented in form of maximum silt-, very-fine sand-, and sand-sized particle diameters (Wentworth classification).
314
Figure 2) Digitally-resolved 2-D particle-size mineral distribution of the quartz component of the dune sample, suggesting an abundance in material in the coarse silt- (20-62.5 m) and very-fine sand (62.5-125 m) -sized fraction.
315
Figure 3. 3-D sediment-size distributions of 3 LGM regolith samples. Two samples are partiallyaggregated with a well-defined unimodal distribution in minimally-dispersed condition (grey). Fullydispersed particle-size distributions are parametrically resolved into 3 unconstrained Component Populations; a primary coarse silt population (~51 m, in dark orange), a secondary fine silt population, (~16 m, in orange) and a minor truncated population of clays (in light orange).
316
Figure 4) Unconstrained Component Population modes (means) of the 3 LGM regolith samples plotted versus their relative contribution and colour-coded by catchment location. An apparent downwind decrease in the primary coarse silt population (red -> yellow -> green) is compensated by an increase in the two finer populations (see Table 1).
317
Table 1) Summary statistics of 3 LGM regolith samples. Unconstrained Component Population modes (means) based on 3-D Multisizer 3 COULTER COUNTER data are listed for corresponding minimally-dispersed and full-dispersed sample expressions and juxtaposed with 2-D average estimated grain sizes of the quartz component calculated from QEMSCAN mineral maps.
Sediment sample ID HK-M (050-060) MD (minimally-dispersed) (s.a.) (s.a.) HK-M (050-060) FD fully-dispersed (s.a.) (s.a.) BRA-SD (093-100) MD (minimally-dispersed) (s.a.) (s.a.) BRA-SD (093-100) FD (fully-dispersed) (s.a.) (s.a.) WL-FP (040-050) MD (minimally-dispersed) (s.a.) (s.a.) WL-FP (040-050) FD (fully-dispersed) (s.a.) (s.a.) Component Population modes (in m) 64 17 6 50 11 5 53 30 10 52 26 7 180 69 27 52 11 4 relative contributions (in %) 88.4 10.9 0.7 76.9 18.1 5 54.3 41.1 4.6 54.8 36.8 8.5 6.0 72.0 22.0 47.8 38.4 13.9 QEMSCAN quartz average grain size (in m) -
66
35
33

Loess and Floods: Late Pleistocene Fine-Grained Valley-Fill Deposits in The Flinders Ranges, South Australia (PHD Thesis)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Loess and Floods: Late Pleistocene Fine-Grained Valley-Fill Deposits in The Flinders Ranges, South Australia (PHD Thesis)

Uploaded by

Copyright:

Available Formats

Loess and floods: late Pleistocene fine-grained valley-fill deposits in the Flinders Ranges, South Australia

Thesis abstract Thesis declaration

1.1 A review of the last deglacial response in South Australia

2.1 Optically Stimulated Luminescence dating

5.2.1.5 U and Th contents TSAC conversion Table

................................................................. ............. ................................................................. ....................................... ..........................

A review of the last deglacial response in South Australia

Lake Frome physiography

Islands of Lake Frome

C-date, derived from inorganic carbonate partly

The lacustrine chronostratigraphy

Geochemical record from Lake Frome

The piedmont plain and palaeosol record

C-dates dates. Taken at face value, the

Fine-grained valley-fill formations of the Flinders Ranges

Fig. 3) 2-D dispersion of calibrated

C-dates for the Brachina Silts (modified diagram of

Interpretation of silty lacustrine sequence of Lake Frome

Th/234U dating. Accordingly, "greater effective precipitation levels for the

Be, (Fig. 1). Apart from low

Fine-grained valley-fill formations in the Finders Ranges

Depositional models and palaeo-environmental implications

A call for Australian loess

Introduction: the Australian liaison with loess

Loess a non-prescriptive definition

Discussion: linking loess-generating processes in periglacial and hot desert environments

Conclusion: loess not just another term but a concept

Response to Smalleys discussion of A call for Australian loess

Optically Stimulated Luminescence dating

Environmental dose-rate estimation

OSL data discussion

Accelerator Mass Spectrometry radiocarbon dating

CO2 and disseminates throughout the

C-activity by measuring the count rate of beta

C-concentration result in discrepancies with calendar years and require

Sample-specific sources of error

C-based chronostratigraphy can be significantly increased by pairing radiocarbon

Th-dated fossil corals provide a marine-based calibration of less precision

(Fairbanks et al. 2005). A high-resolution marine-derived

C-dated sedimentological and

C time scale remain (Hoffman et al.

AMS data discussion

C ages for recent

Lo ess an d Flo o d s: (Place holder)

2.3 Particle-Size Analysis

Lo ess an d Flo o d s: (Place holder)

Quantifying particle aggregation in fluvial sediments

Keywords: aggregation, particle-size analysis, grain-size analysis, component population analysis,

decomposition, mud aggregates.

Review of earlier studies

Site and Samples

3. Results and Discussion

Ch ap t er 3 (Resu lt s an d Discu ssio n )

Ch ap t er 3 (Resu lt s an d Discu ssio n )

, Peter Glasby 3, Martin A.J. Williams 3, Steven M. Hill

Edward J. Rhodes 5, Victor Gostin 1, Anthony OFlaherty 6, Geraldine E. Jacobsen 7

Ch ap t er 3.1 (Result s an d Discu ssio n )

Ch ap t er 3.1 (Result s an d Discu ssio n )

Ch ap t er 3.1 (Result s an d Discu ssio n )

Ch ap t er 3.1 (Result s an d Discu ssio n )

Ch ap t er 3.1 (Result s an d Discu ssio n )

Age proxies and models

Flinders Silts radiocarbon dates