Holocene relative sea-level change and deglaciation on Alexander Island, Antarctic Peninsula, from elevated lake deltas

Geomorphology 112 (2009) 122–134

Contents lists available at ScienceDirect

Geomorphology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / g e o m o r p h

Holocene relative sea-level change and deglaciation on Alexander Island, Antarctic Peninsula, from elevated lake deltas S.J. Roberts a,⁎, D.A. Hodgson a, M.J. Bentley b,a, D.C.W. Sanderson c, G. Milne d, J.A. Smith a, E. Verleyen e, A. Balbo a a

British Antarctic Survey (BAS), Natural Environmental Research Council (NERC) High Cross, Madingley Road, Cambridge, CB3 0ET, UK Department of Geography, University of Durham, South Road, Durham, DH1 3LE, UK Scottish Universities Environmental Research Centre (SUERC), Rankine Avenue, East Kilbride, UK d Department of Earth Sciences, University of Ottawa, Ottawa ON, K1N 6N5, Canada e Department of Biology, Section of Protistology and Aquatic Ecology, University of Ghent, Krijgslaan 281 S8, 9000 Gent, Belgium b c

a r t i c l e

i n f o

Article history: Received 6 June 2008 Received in revised form 20 May 2009 Accepted 22 May 2009 Available online 30 May 2009 Keywords: Antarctica Hypsithermal Luminescence dating Subglacial lakes

a b s t r a c t Field data constraining the rate and spatial pattern of deglaciation and relative sea-level (RSL) change on the Antarctic Peninsula (AP) are relatively sparse, but are needed to improve regional ice sheet and RSL change models, and contribute to better model predictions of future sea-level rise. We investigated the geomorphology, sedimentology and quartz-fraction single aliquot regeneration optically-stimulated luminescence (SAR-OSL) geochronology of elevated deltas around two epishelf lakes, Ablation Lake (AL) and Moutonnée Lake (ML), Alexander Island, Antarctic Peninsula. AL and ML are dammed by George VI Ice Shelf, and maintain a direct hydraulic connection to the sea; hence, their water levels are controlled by changes in RSL. Our aim was to provide new terrestrial constraints on RSL and deglaciation for the southern AP by comparing the formation processes, age and altitude of the AL and ML deltas with: (1) existing RSL curves for the AP; (2) isostatically-coupled AP ice sheet models, and (3) existing AP deglaciation history and SAR-OSL ages from elevated deltas around the nearby inland Hodgson Lake (HL). Although there was insufficient quartz in the ML samples for SAR-OSL dating, the 4.6 ± 0.4 ka SAR-OSL age of the elevated delta at AL represents the last time active deltas were forming higher than present day lake level, and implies: (1) a fall in RSL of up to 14.4 m since the mid-Holocene in this part of Alexander Island, which is consistent with existing field-based RSL chronologies for the AP; (2) relatively smaller ice masses than suggested by some (but not all) isostatically-coupled ice sheet models since the mid-Holocene, and (3) significant mid-Holocene thinning of the AP ice sheets, which is consistent with regional sediment core data and cosmogenic exposure ages, and the 4.4 ± 0.7 ka SAR-OSL age of the lowermost HL delta. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Sea-level rise is currently one of the biggest uncertainties in predictions of future global climate change. The Fourth Assessment Report of the Intergovernmental Panel on Climate Change (Bindoff et al., 2007; IPCC, 2007) suggested that decaying ice sheets may contribute more to future sea-level rise than previously thought, but concluded that current ‘limited understanding of the relevant processes’ (Bindoff et al., 2007 p. 409) makes reliable future predictions challenging (Arthern and Hindmarsh, 2006). The recent retreat and disintegration of many Antarctic Peninsula (AP) ice shelves (Vaughan, 1993; Rott et al., 1996; Vaughan and Doake, 1996; Vaughan et al., 2003; Cook et al., 2005) has highlighted how the accelerated discharge of glaciers from the Antarctic continent can potentially increase rates of global sea-level rise (De Angelis and ⁎ Corresponding author. Tel.: +44 1223 221 635; fax: +44 1223 221 259. E-mail address: [email protected] (S.J. Roberts). 0169-555X/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2009.05.011

Skvarca, 2003; Rignot et al., 2004; Scambos et al., 2004; Rignot et al., 2005; Arthern and Hindmarsh, 2006; Vaughan, 2006; Pritchard and Vaughan, 2007; Hock et al., 2009). In addition, the regions around large ice sheets are important areas for studying past sea-level change because they often show the greatest changes in relative sea-level (RSL), especially during glacial to interglacial transitions (Huybrechts, 2002; Bassett et al., 2005). Palaeo-records and isostatically-coupled ice sheet models of postLast Glacial Maximum (LGM) deglaciation and RSL change have provided a better understanding of the potential causes and consequences of recent ice mass loss on the AP, and additional insights into how sea level might change with further deglaciation of Antarctica (Payne et al., 1989; James and Ivins, 1998; Bentley, 1999; Ingólfsson and Hjort, 1999; Ivins et al., 2000; Bassett et al., 2005; Bentley et al., 2005a; Bentley et al., 2005b; Ivins and James, 2005; Bassett et al., 2007; Smith et al., 2007a). Earlier ice sheet modelling experiments suggest deglaciation of the AP to its present day limits occurred between 14 and 6.5 ka

Fig. 1. Location map of (a) Antarctic Peninsula (AP); (b) Ablation Lake (AL), Moutonnée Lake (ML), Hodgson Lake (HL) on Alexander Island and other place names mentioned in text; (c) summary geomorphological map superimposed on aerial photo Ablation Point area covering Ablation Lake and Moutonnée Lake; and (d) summary geomorphological map of the Citadel Bastion area superimposed on an aerial photo showing the location of Hodgson Lake.

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134 123

124

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

(Payne et al., 1989). Other models showed most uplift occurs within the first few thousand years following deglaciation (Lambeck, 1993), and that post-LGM crustal rebound was a relatively slow process compared to the more rapid rise in sea-level associated with large volumes of meltwater discharged into the Southern Ocean. These findings are broadly consistent with field-based reconstructions of RSL curves and some recent model-based RSL reconstructions for the AP, which show an initial rapid rise in RSL during the LGMHolocene transition, followed by a gradual fall in RSL during the Holocene, and a levelling out at, or near, current sea level sometime during the mid- to late Holocene (Bentley et al., 2005a; Bassett et al., 2007). 1.1. Rationale Field constraints on Holocene rates and spatial patterns of deglaciation and RSL change on the AP remain relatively sparse, however, and new data are needed to further improve the predictive ability of isostatically-coupled ice sheet models. Therefore, this paper aims, firstly, to establish the formation process, provenance and geochronology of elevated deltas in the catchments of Ablation Lake (AL), Moutonnée Lake (ML), and Hodgson Lake (HL) on Alexander Island, AP; and secondly, to determine whether these can be used to provide chronological constraints on Holocene deglaciation and RSL change for the southern AP. Elevated deltas, lake terraces, and palaeo-shorelines are common features of lake basins in Antarctica, providing evidence of past changes in shoreline elevation and semi-permanent (non-seasonal) changes in hydrological equilibrium (Hodgson et al., 2004), as well as important constraints on climate change (Hall et al., 2001, 2002), the timing and scale of local and regional deglaciation, and RSL (Hall and Denton, 1999; Hall et al., 2004; Bentley et al., 2005a). The formation of elevated deltas at AL and ML was probably related to changes in RSL. AL and ML are epishelf lakes formed on the east coast of Alexander Island, at the Ablation Point Massif where George VI Ice Shelf (GVI-IS) dams two marine embayments (Fig. 1). Epishelf lakes are tidal, stratified water bodies with a lower marine layer that extends under the ice-shelf to the ocean and an upper fresh meltwater layer whose maximum thickness is determined by the draught of the ice shelf (Heywood, 1977; Smith et al., 2007b). Current lake levels in AL and ML are controlled by changes in sea level via their sub-ice-shelf hydraulic connection to George VI Sound (Smith et al., 2006). Therefore, elevated delta deposits in the AL and ML catchments likely represent a former sea-level high-stand, which was abandoned when RSL fell. If accurately dated, these deltas can provide, firstly, the most southerly field constraints on RSL change on the AP (cf. Bentley et al., 2005a); and, secondly, key field data to constrain AP deglaciation history, and isostatically-coupled ice sheet models. In contrast, HL is a recently discovered former sub-glacial lake, land-locked and dammed by the Saturn Glacier, with no hydraulic sub-ice-shelf connection to the sea (Hodgson et al., in press-a,b). The formation of elevated deltas at HL is probably controlled by meltwater availability and/or variations in the thickness of the Saturn Glacier ice dam; hence, the post-LGM deglaciation history and formation of elevated deltas at HL provides a field constraint on the style and rate of deglaciation for south-eastern Alexander Island independent of RSL changes. Dating elevated deltas at this site constrains when the lake ice surface was higher than present because the deltas were probably abandoned when the ice over the lake retreated and the lake level fell, possibly discharging water via its palaeo-spillway (Fig. 2c) (Hodgson et al., in press-a,b). In the absence of organic matter for radiocarbon dating, we used optically stimulated luminescence (OSL) to establish the geochronology of the elevated deltas in this study. OSL is a property of certain ionic minerals such as quartz and feldspar, which can be used to date the time elapsed since the sediment was last exposed to light (known as

the ‘resetting’ or ‘bleaching’ event) (Godfrey-Smith et al., 1988; Aitken, 1998). The total radiation dose received by sediments (palaeodose) since the last ‘bleaching’ event can be estimated in the laboratory by forming a dose-dependent growth curve from luminescence signal measurements to produce an equivalent dose (De) value, measured in SI unit grays (Gy), and represents the absorbed energy of radiation, where 1 Gy = 1 J kg− 1 = 1 m2 s− 2. If the environmental dose rate (DT) (measured in mGy a− 1) during burial can be evaluated reliably from field and laboratory measurements, a luminescence age can be calculated as De /DT (Aitken, 1998). Sediments whose OSL clock has been completely reset by exposure to infra-red, visible and/or ultra violet electromagnetic radiation are said to be ‘fully-bleached’. Sediments that are incompletely reset are said to be ‘partially bleached’ and provide unreliable depositional ages (Spencer and Owen, 2004). 1.2. Post-LGM palaeoenvironmental change on Alexander Island Deglaciation from LGM limits, which began as far back as c. 18 ka, accelerated during the early Holocene when ice rapidly retreated from AP continental shelf areas in and around Marguerite Bay (Ingólfsson et al., 1998; Ingólfsson and Hjort, 2002; Ingólfsson et al., 2003; Bentley et al., 2006, 2009 and references therein) (Fig. 1). Bentley et al. (2009) proposed that the general pattern of Holocene climate change on the AP, and, in particular, the two most prominent warm periods of the early and mid-Holocene, were probably driven by a combination of atmospheric forcing (mainly by latitudinal shifts in Southern Westerly wind circulation position and patterns) and oceanographic forcing (in the form of variable Circum-polar Deep Water position and flow rates; e.g., Smith et al., 2007b). The mid-Holocene period, in particular, was characterised by a period of cross-Peninsula warming c. 5–2 ka (Hodgson et al., 2004). However, most reconstructions of Holocene AP palaeoenvironmental change are based on records located north of Marguerite Bay, which has different physiographic, climatic, glaciological, and geological setting (Fig. 1). Less is known about mid-Holocene to present palaeoenvironmental change on Alexander Island, due, mainly, to a lack of dateable organic material in moraine and core sediments, and concerns about radiocarbon dating of bulk sediments, which might be influenced by ‘old’ carbon (Bentley et al., 2005b; Smith et al., 2007a; Roberts et al., 2008). Alexander Island is a large (c. 400 × 80 km), mostly glaciated island situated to the west of the Antarctic Peninsula between 68.7° and 72.6° S (Fig. 1). It is predominantly ice covered with some exposed nunataks and a few notable ice-free areas, such as the Ablation Point Massif. Numerous glaciers flow off Alexander Island; west into the Bach and Wilkins Ice Shelves and Bellingshausen Sea, and east into GVI-IS. GVI-IS is fed both by outlet glaciers from Alexander Island and the ice cap on Palmer Land, an extension of the main West Antarctic Ice Sheet. Cosmogenic surface exposure ages from moraines and exposed bedrock on Alexander Island (Bentley et al., 2006; Hodgson et al., in press-b) suggest significant ice free areas had appeared by the start of the Holocene. Multi-proxy analysis and radiocarbon dating of lake sediments from ML and shells within lake catchment moraines from the Ablation Point area (Fig. 1c) show that GVI-IS retreated to at least the Ablation Point Massif area only once in the last c. 11.5 ka. This occurred c. 9600– 7500 yr BP, during the early Holocene hypsithermal, after a phase of atmospheric warmth and the intrusion of warm Circum-polar Deep Water onto the continental shelf (Clapperton and Sugden, 1982, 1983; Bentley et al., 2005b; Smith et al., 2006, 2007a,b; Roberts et al., 2008). Roberts et al. (2008) described the sedimentology of sediment cores from AL, and found that the numerous coarse-grained (N8 mm) clasts found in the main lake core were predominantly of Alexander Island origin. These sediment cores remain undated, however, due to the coarse clastic nature of the sediments and a lack of macrofossils and organic compounds suitable for radiocarbon dating. Bulk surface

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

125

Fig. 2. Oblique aerial views of (a) Ablation Lake; (b) Moutonnée Lake (both looking approximately west, from a north–south flight-line along George VI Sound, towards Alexander Island); (c) Hodgson Lake (looking approximately north–east). For scale, Citadel Bastion is 465 m a.s.l. The lake is approximately 2 m a.s.l. and its dimensions are 1.87 km (NE–SW) by 1.6 km (NW–SE); (d) Ablation Lake: Sampling hole for AB1 elevated delta; (e) Moutonnée Lake: Sample points for ML1B elevated delta; (f) Hodgson Lake: CB6.5 elevated delta OSL sample positions and their relationship to the relict lake ice/sediment interface and its relationship to the elevated delta surface at 7.6 m above present lake level (a.p.l.l.).

sediment ages are N10 ka, suggesting a substantial ‘old’ radiocarbon effect exists at this site (Roberts et al., 2008). At HL, multi-proxy analysis of the lake sediments and hydrological system, and cosmogenic surface exposure ages from the surrounding ice-free ridges, suggest HL was once a sub-glacial lake buried beneath the overriding Last Glacial Maximum (LGM) ice sheet, emerging to become a perennially ice covered lake system sometime during the late Holocene (Hodgson et al., in press-a,b).

2. Site descriptions 2.1. Ablation Lake (AL) (70° 48.974' S, 68° 27.143 W) Ablation Lake is located south of Ablation Point at the entrance to Ablation Valley (Figs. 1c and 2a). The geology of the area was mapped by Horne (1968) and the geomorphology by Sugden and Clapperton (1981) and Clapperton and Sugden (1983).

126

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

At AL, GVI-IS is only partially grounded across the entrance of the embayment, creating a continuous hydraulic connection to the seawater in George VI Sound and an in-phase tidal regime between George VI Sound and the lake (Sugden and Clapperton, 1981; Smith et al., 2006). Due to the absence of a continuous bedrock barrier, GVIIS penetrates into the northern part of the lake basin forming an iceshelf tongue, which deposits sediments into the lake via an ice conveyor, and into ice push deposits formed by the tidal pressure ridges on the lake shore (Sugden and Clapperton, 1981; Hall et al., 2006; Smith et al., 2006). These deposits can be coarse gravel to boulder in size. Some are of igneous origin, transported englacially from the west coast of Palmer Land through GVI-IS, or reworked from well-developed ice-shelf moraines along the coast of Alexander Island (Clapperton and Sugden, 1983; Smith et al., 2007b; Roberts et al., 2008). Two catchment glaciers currently terminate close to the lake. The elevated delta at the western shore of AL forms a gently dipping 200 m-long by 60 m-wide terrace, c. 14.4 m above present lake level (a.p.l.l.), bisected by the main lake inflow stream at the western rim of the lake. 2.2. Moutonnée Lake (ML) (70° 52.123' S, 68° 19.635' W) Moutonnée Lake is located south of Ablation Point in Moutonnée Valley (Figs. 1c and 2b). The geology and geomorphology of the ML catchment were first described by Horne (1968), Elliott (1974) and Clapperton and Sugden (1983). The ML catchment bedrock forms part of the Fossil Bluff Group, and, although a small active glacier exists near the head of the valley, larger glaciers have remained relatively stable c. 2 km up-valley for at least 40 years (Fig. 1c) (Horne, 1968; Elliott, 1974; Clapperton and Sugden, 1983; Butterworth, 1991; Browne, 1996). The catchment is composed largely of bedrock to the south, but reworked glacial sediments have been found to the west, in Moutonnée Valley, and north of the main lake (Clapperton and Sugden, 1983). At ML, GVI-IS is grounded on a partially submerged bedrock sill damming the embayment. However, the lake retains a limited hydraulic connection to the seawater in George VI Sound as revealed by the offset tidal regime that occurs between the lake and George VI Sound (Pearson and Rose, 1983). Several elevated deltas exist in the ML catchment. The largest two form wide, gently sloping terraces c. 6 m a.p.l.l. at the western shore of ML: one (MLA) is 400 m long by 40 m wide (and has previously been used as an airstrip), the other (MLB) is 150 m long by 50 m. 2.3. Hodgson Lake (HL) (72° 00.549′ S, 068° 27.708′ W) Hodgson Lake (HL) is located on the southeastern coast of Alexander Island, further inland than AL and ML. HL was first mapped in 2001–2003, and the lake sediments cored in 2002–3 (Hodgson et al., in press-a,b). It is enclosed on three sides by Citadel Bastion and Corner Cliffs and dammed along its northern edge by the Saturn Glacier (Fig. 1d). The cirque occupied by HL is one of two north-facing cirques abutting the southern margin of the Saturn Glacier adjacent to George VI Sound. The western cirque at nearby Hall Cliff is filled with locally derived ice that abuts the Saturn Glacier. In contrast, the cirque at HL includes a number of ice-free slopes, and most of the ridges at the rim of the cirque are ice-free (Fig. 2c). There are two small remnant catchment glaciers (unofficial names: Corner Cliffs Glacier and Citadel Bastion Glacier, Figs. 1d and 2c) discharging into the southwest corner of the lake. A spillway for lake outflow exists in the eastern corner, where the lake abuts Corner Cliffs and the Saturn Glacier (Fig. 1c). The thickness of the Saturn Glacier at this point appears to be critical in controlling the level of water in Hodgson Lake: if it were to thin, the lake could

drain to a lower level. The lake ice surface is only 2.3–3.9 m above the WGS84 GPS datum. As the maximum measured depth of the lake is 89.9 m this means that the bottom of the lake is c. 86 m below the WGS84 GPS datum (Hodgson et al., in press-a,b). Along the south shore of HL are two prominent elevated deltas, the surfaces of which are 8.4 and 7.6 m a.p.l.l., and a prominent horizontal break in slope approximately 300 m in length and 13.4 m a.p.l.l. Active deltas are currently forming where small birds-foot deltas extend over the ice at the margin of the lake, and refreezing of meltwater and sediment associated with several ephemeral meltwater streams is creating sedimentary lobes in and around the lake (Hodgson et al., in press-b). Since the present delta is forming as lobes of sediment on top of the present lake ice, the age of the elevated deltas would be expected to increase with height. 3. Methods 3.1. Modern control samples To test where (and if) sediment was most likely to be exposed to light in the ML hydrological system, we took ‘light-tight’ luminescence dating samples from four modern depositional environments. Samples were taken using copper tubes, sealed at one end, from a shallow depositional inflow stream, a shallow meltwater pond depositing fine sediment, and from the centre of a glaciofluvial gravel deposit. In addition, a sediment trap was left for one year, suspended at a water depth of 47 m in the centre of the ML main basin (Smith et al., 2006). 3.2. Elevated delta samples The lake catchments, the altitudes of the elevated deltas, and the delta sampling sites were surveyed using an EDM and a Magellan GPS roving unit and, at HL, a high-precision Trimble GPS. Delta profiles were logged in the field and all samples taken from holes excavated for luminescence dating. Samples for luminescence dating were extracted from elevated deltas in the AB, ML and HL catchments by hammering four copper tubes (22 mm diameter × 20 cm length), sealed at one end, into excavated delta sections that had been cut back at least 0.5 m (Fig. 2d–f). Tubes were removed, sealed and transported frozen in the dark. Bulk sediment was excavated from the same position as the tubes for spectrometric analysis. Tubes were opened under darkroom conditions and 1–2 cm of sediment from the end of the sample, which could have been exposed to light during sampling, was removed from both ends. The single-aliquot regenerative-dose (SAR) OSL method (Murray and Wintle, 2000, 2003) was used to reconstruct equivalent dose (De) values from the 90–150 µm or 150–250 µm quartz fraction for modern and elevated delta samples. Infra-red stimulated luminescence (IRSL) was measured at the end of the SAR-OSL cycle to determine whether OSL came from an uncontaminated quartz source (Duller, 2003). Environmental dose rates (DT) were calculated from bulk geochemical composition data obtained by XRF and ICP-MS. Sample preparation procedures, and sampling errors, are described in Tables 1 and 2, and in the online supplementary data. Bulk sedimentological data were obtained following standard procedures (Fig. 3). 4. Results 4.1. Modern control samples Of the modern catchment samples collected, sufficient quartz for SAR-OSL analysis was extracted from the ML8 Pond sample. Dose rates for ML8 Pond are similar to those from the modern stream and fluvial gravel samples (Table 1), but significantly lower than elevated delta

ML1B(I)

1476⁎

5 ± 0.2 6 ± 0.3 9 ± 0.4

6.0a 6.0a 0.155 ± 0.015

0.155 ± 0.015

0.155 ± 0.015

0.192 ± 0.019

0.136 ± 0.014

0.142 ± 0.014

0.18 ± 0.018 0.18 ± 0.018 0.18 ± 0.018 0.024 ± 0.002

Cosmicf dose rate (mGya− 1 ± 1σ)

–

0.642 ± 0.004

0.686 ± 0.005

– – – –

0.905 ± 0.006

– – – –

ISGS gamma dose rate (mGya− 1 ± 1σ)

2.250 ± 0.053

1.934 ± 0.068

1.698 ± 0.074

2.548 ± 0.056

2.237 ± 0.094

2.102 ± 0.068

1.603 ± 0.118 – 1.950 ± 0.073 0.873 ± 0.035

Dry TSBCe dose rate (mGya− 1 ± 1σ)

(1) (2) (1) (2) (1) (2)

(1) (2) (1) (2)

1.93 1.951 2.11 1.917 2.32 1.778

2.29 2.385 2.71 2.274

(1) 2.54 (2) 2.202

1.85 1.78 2.29 0.40

K O/K (%)

2

38.1 – 65.0 – 71.5 –

61.62 – 97.16 –

42.0 –

41.9 56.9 68.8 13.3

Rb (ppm)

1.55 1.899 1.47 1.933 1.65 1.625

1.79 2.215 2.08 2.401

1.62 2.352

1.10 1.63 1.70 0.30

U (ppm)

Sample geochemistryc

5.28 6.541 6.05 6.333 6.96 5.828

7.80 9.394 9.71 9.474

5.77 7.792

4.10 5.94 6.31 0.92

Th (ppm)

0.817 ± 0.020 1.025 ± 0.014 0.882 ± 0.022 1.010 ± 0.014 0.988 ± 0.024 0.915 ± 0.012

1.038 ± 0.024 1.312 ± 0.016 1.247 ± 0.028 1.311 ± 0.016

0.972 ± 0.026 1.201 ± 0.016

0.694 ± 0.019 0.828 ± 0.018 0.956 ± 0.023 0.158 ± 0.004

Dry gammad dose rate (mGya− 1 ± 1σ)

1.636 ± 0.063 2.084 ± 0.031 1.772 ± 0.070 2.055 ± 0.031 1.962 ± 0.077 1.880 ± 0.030

1.982 ± 0.075 2.572 ± 0.035 2.363 ± 0.090 2.510 ± 0.035

2.057 ± 0.083 2.395 ± 0.035

1.488 ± 0.061 1.575 ± 0.059 1.931 ± 0.076 0.333 ± 0.013

Dry betad dose rate (mGya− 1 ± 1σ)

2.903 ± 0.035

2.822 ± 0.036

2.805 ± 0.036

3.374 ± 0.059

3.210 ± 0.065

3.407 ± 0.039

1.340 ± 0.044 1.416 ± 0.037 1.629 ± 0.044 0.536 ± 0.023

Mean totalg dose rate (mGya− 1 ± SE)

b

Depth (cm) below sediment surface; m a.p.l.l. = metres above present lake level. In-situ percentage water content with 10% error; attenuation factors in Aitken (1998) applied to dry dose rate values. c K2O data was obtained by standard XRF methods and U, Th, Rb by standard ICP-MS methods (described further in Roberts et al. (2008)). (1) = XRF and ICP-MS measurements, potassium as K2O, (2) = Laboratory-based gamma spectrometry measurements, potassium as K. d Dose rates based on chemical composition and calculated using conversion factors in Adamiec and Aitken (1998). Sediment core gamma dose-rate calculations assume that sediment in a 30 cm sphere surrounding the core is homogenous. Gamma spectrometry dose rate calculations assume Rb at 50 ppm. e TSBC = Thick Source Beta Counting, 20 g was used for delta samples (Sanderson, 1998). f The cosmic dose rate was calculated from 0.21exp[−0.07(dr) + 0.0005(dr)2] mGy a− 1 (Prescott and Hutton, 1994) where d is the current depth of burial and r is the density of the attenuating medium (r = 0.998 g cm− 3 for water and 2.6 g −3 for sediment). For core sediments, d = depth of water in the lake (89.8 m) + sediment depth and assumes no changes in water/sediment depth throughout burial. cm g Total dose-rate from means of measured beta, gamma and cosmic ray components incorporating external attenuation factors for U, Th and K taken from (Medahl, 1979) and based on a median grain-size of 100 µm; † = No material of sand fraction or greater (i.e., N 63 µm) was present in the ML sediment trap, but we were able to extract sufficient fine-grained sediment for ICP-MS analysis, and calculate a dose rate of 0.536 ± 0.023 mGya− 1. The low dose rate and anomalous bulk sediment geochemistry of this sample are due to encrustation with sea salt; TSBC n = 2 except SUTL1464 n = 4; ⁎Palaeodose reconstruction undertaken on asterisked samples. Other samples are included for completeness, but yielded insufficient quartz in the coarse (90–250 µm) fraction for SAR-OSL analysis. Methods are summarised in supplementary data.

a

Elevated delta

MLA2(I)

1472⁎

Elevated delta

Elevated delta

Moutonnée Lake 1468⁎ ML1A(I) 6.0a

16 ± 1

Elevated delta

8.4a

CB8.4

1464⁎

0.65 ± 0.03

8±1

14.4a

100 ± 5 100 ± 5 100 ± 5 100 ± 5

Waterb content (%)

6.5a

Elevated delta

0 0 0 47

Depth (m) Height (m a.p.l.l.)a

Hodgson Lake (Citadel Bastion) 1460⁎ CB6.5 Elevated delta (relict ice/sediment interface)

Ablation Lake 1456⁎ AB1(A)

control samples (all Moutonnée Lake catchment) ML8 Stream ML8 Pond ML7 Fluvial gravel ML47(II) Sediment trap†

Modern 1451 1452⁎ 1453 1455

Sample environment

Sample

SUTL no.

Table 1 Summary dose-rate data for elevated delta samples.

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134 127

128

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

Table 2 Summary equivalent dose (De) data (see online supplementary data for method summaries and full data). Elevated delta (Lab. ID)

No. of discs

Cluster group definition

Weighted mean values ± 2σ Recycling

TD7/1

De

SAR-OSL age

Min–Max group range

Cluster centroid

Ratio (RR)

ratio

(Gy)

(103 yr)

(103 yr)

(103 yr)

3.1–12

–

3.1–4.3

3.7

4.5–5.8

5.2

6.8–7.3

7.0

12 ± 4

12

AB1 (SUTL 1452) All data All RR passes Group A: all Group A: RR pass Group B: all Group B: RR passes Group C: all Group C: RR pass Outlier D

15 5 4 1 6 3 4 1 1

1.17 ± 0.12 1.02 ± 0.17 1.26 ± 0.26 1.12 ± 0.27 1.10 ± 0.16 1.00 ± 0.19 1.25 ± 0.27 1.01 ± 0.23 1.33 ± 0.62

1.23 ± 0.61 1.24 ± 0.57 1.27 ± 0.97 1.16 1.09 ± 0.41 1.11 ± 0.38 1.35 ± 0.45 1.69 1.06

16.8 ± 3.9 16.0 ± 5.3 11.7 ± 8.8 13.6 ± 10.8 16.3 ± 4.9 15.8 ± 5.6 24.3 ± 10.5 24.8 ± 13.7 41.4 ± 17.4

4.9 ± 0.4 4.6 ± 0.4 3.5 ± 1.9 4.0 ± 2.5 4.6 ± 0.4 4.6 ± 0.4 7.2 ± 1.2 7.3 ± 2.2 12 ± 4

CB6.5 SUTL 1460) All data All RR passes Group A: all Group A: RR passes Outlier B (RR pass) Outlier C

14 9 12 8 1 1

1.08 ± 0.07 1.04 ± 0.08 1.08 ± 0.08 1.04 ± 0.09 1.04 ± 0.12 1.07 ± 0.13

1.33 ± 0.70 1.18 ± 0.44 1.40 ± 0.65 1.25 ± 0.38 0.93 0.90

14.8 ± 1.7 16.3 ± 2.0 13.2 ± 1.7 14.1 ± 2.1 35.8 ± 4.1 97.0 ± 9.9

4.6 ± 0.5 5.1 ± 0.6 4.1 ± 0.5 4.4 ± 0.7 11 ± 1 30 ± 3

2.5–30

–

2.5–6.2

4.2

11 ± 1 30 ± 3

11 30

– – – –

– – – –

15 ± 3 30 ± 7 58 ± 8

11–19 26–37 58 ± 8

12 31 58

CB8.4 (SUTL 1464) Group A: all Group B: all Outlier C Group D: all

5 4 1 5

50.7 ± 8.8 100.2 ± 22.1 195.9 ± 26.5 OSLnatural N saturation

The most reliable ages are shown in bold (see Appendix A).

samples. Four discs of purified quartz were made from the ML8 Pond 90–150 µm or 150–250 µm fraction quartz extracts. Two (of four) discs produced insufficient OSL for analysis. The SAR-OSL ages of the two discs that retained a luminescence signal and passed internal validity tests are 7.3 ± 4.2 ka and 9.7 ± 3.4 ka (see Appendix A). 4.2. Elevated delta samples 4.2.1. Ablation Lake (AL) The elevated delta (AB1) sample at c. 14 m a.p.l.l. consists of predominantly well-sorted sandy gravel sand with sporadic cobbles

Fig. 3. Grain size profiles (log scale) for the b 2 mm fraction for representative samples from delta samples from Ablation Lake (AL), Moutonnée Lake (ML) and Hodgson Lake (HL). Samples for grain size analysis were split, and dry sieved through a 2 mm mesh. The b 2 mm fraction was treated with 10% HCl (to remove carbonates) and 30% H2O2 (to remove residual organic material) and analysed with a Coulter laser granulometer following standard procedures. The data shown illustrate the relative lack of 90–250 µm sediment in the AL and ML samples compared to the HL delta samples. ML1B (SUTL1476) was not analysed due to insufficient sample volume.

and small boulders. Overall, the grain size is unimodally distributed, with nearly 55% of the AB1 sample N2 mm (Fig. 3). The AB1 OSL sample was taken from the base of the most prominent elevated delta whose surface is c. 14.4 m a.p.l.l. In-situ gamma spectrometry measurements were made in the excavated OSL sample hole. We extracted 50–100 mg of purified 90–250 µm fraction quartz from AB1 for SAR-OSL dating, sufficient for a 16 small aliquot disc SAR-OSL run. Equivalent dose (De) values for AB1 were mainly in the 10–20 Gy range and, excluding statistical outliers, unimodally distributed (Fig. 4a). We separated individual ages into groups, defined by which discs passed or failed internal validity tests (recycling ratio—RR in Table 2) using Ward's method-Euclidean distance cluster analysis, weighted by the individual age errors. Five discs (out of 16) passed internal validity (recycling ratio) tests (Table 2). The weighted mean age of five discs that passed recycling ratio tests is 4.6 ± 0.4 ka, compared to a weighted mean age from all data of 4.9 ± 0.4 ka (Table 2). Three cluster groups (A–C) exist, with a valid age of 4.0 ± 2.5 ka, and valid weighted mean ages of 4.6 ± 0.4 ka and 7.3 ± 2.2 ka, respectively. One outlier (D) has an age of 12 ± 4 ka. 4.2.2. Moutonnée Lake (ML) Three samples were taken from the two most prominent raised deltas at the western end of ML, c. 6 m a.p.l.l. Two samples (MLA1 and MLA2) were taken 60 m apart. Sample MLA1 (S 70° 51.796, W 068° 21.099) comes from a part of the delta that is composed of sandy gravel deposits, crudely bedded with angular to sub-angular clasts. Sample MLA2 (S 70° 51.797, W 068° 21.166) is from a part of the same delta composed of coarse sand and gravel. The third sample (ML1B; S 70° 51.893, 68° 21.538) was extracted from MLB, which was composed predominantly of coarse sand and gravel but with a crudely bedded, silty matrix and some cobbles on its upper surface. The MLA1 and MLA2 samples are composed, respectively, of c.73% and c.87% of grains N2 mm (Table 1), and are a well-sorted, unimodally distributed fine to medium sandy gravel (Fig. 3). Three OSL dating samples (ML1A, MLA2 and ML1B) were taken from the two elevated deltas (6 m a.p.l.l.) at the western end of Moutonnée Lake. In situ gamma spectrometry measurements made in excavated OSL sample holes for ML1A and MLA2, but not in ML1B due

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

129

the surface sediment has been disturbed, and, therefore, is not suitable for OSL sampling. The interface between the buried relict lake-ice and the overlying sediment is at an altitude of 6.5 m a.p.l.l. Samples were taken from the 7.6 m a.p.l.l. delta, at c. 6.5 m a.p.l.l. (CB6.5) (buried ice–sediment interface; henceforth referred to as the ‘lower HL delta’) and from the c. 8.4 m a.p.l.l. delta (CB8.4) (near upper delta surface; henceforth referred to as the ‘upper HL delta’). Both samples are composed of light brown medium to coarse sand; 1% of the CB6.5 sample and 27% of the CB8.4 sample are composed of material N2 mm (Fig. 3). The 90–250 µm fraction of the upper and lower delta OSL samples constituted c. 37% and 30%, respectively, of the bulk sediment (Fig. 3). Quartz was rare, but c.100 mg, sufficient for a 16-disk small aliquot SAR-OSL run, was extracted. De values for the lower HL delta were mostly in the 10–20 Gy range and, excluding two statistical outliers, unimodally distributed (Fig. 4b). One main cluster group (A) and two outliers (B, C) were identified (Table 2). Eight samples in Group A that passed validity tests had a weighted mean age of 4.4 ± 0.7 ka, compared to a weighted mean age for all data of 4.6 ± 0.5 ka (Table 2). Two outliers had ages of 11 ± 1 (valid) and 30 ± 3 (not valid) ka. De values and age distributions for the upper HL delta were larger and more dispersed than the lower delta (Fig. 4c; Table 2). Recycling ratio data were not produced for these samples due to an equipment malfunction; hence it was not possible to determine which discs passed internal validation tests. Nevertheless, it was possible to calculate mean age estimates from ten sample runs with equivalent dose errors b25% (Fig. 4b). Weighted Euclidean distance cluster analysis of the De data identified three main groups, two (A, B) with age ranges of 11–19 ka and 26–37 ka, respectively, and a ‘beyond OSL saturation’ group (D) of indeterminate age. One outlier (C) has an age of 58 ± 8 ka (Table 2). We discuss the validity and implications of all age data further in the next section. 5. Discussion 5.1. Assessment of OSL data

Fig. 4. Equivalent dose (De) Histogram data for elevated deltas from (a) AL delta and (b) HL CB6.5 delta and (c) HL CB8.4 delta illustrating the relatively-well constrained SAR-OSL De values obtained from the delta sample in AL and the lower 6.5 m relict ice/ delta sediment interface in HL compared to more widely-scattered De values obtained from the upper (8.4 m) delta sample in the HL catchment.

to sediment collapse. Dose rates were calculated (Table 1), but we were not able to extract sufficient quartz from the ML samples for SAROSL analysis. 4.2.3. Hodgson Lake (HL) Samples were taken from a 2 m section of sediment overlying relict lake-ice, exposed where a gully cuts into the lower delta ~100 m from the marginal moraine of the ‘Corner Cliffs Glacier’ (Fig. 2f). The section comprises debris-free transparent relict lake-ice overlain by a unit of crudely laminated fine sands, silts and inter-bedded gravel layers. Isotopic data (δ18O and δ2H) confirms that the buried ice is of lacustrine origin, with values more similar to contemporary lake ice than contemporary lake water, and dissimilar to regional values for compressed snow, glaciers and ice shelf ice (Hodgson et al., in press-b). The topmost 30 cm of the sediment is cryoturbated and is capped by a coarse gravel lag on the upper surface. The uppermost sediment is not glacially-deformed but, as is common in Antarctic environments,

The quartz fraction was initially chosen for dating analysis because coarser-grained quartz is thought to bleach more quickly than feldspar, and not expected to suffer from anomalous fading (Godfrey-Smith et al., 1988; Aitken, 1998; Wallinga, 2002; Duller, 2003). Unlike quartz, feldspar bleaching rates are not highly dependent on exposure to the ultra-violet component of daylight (Sanderson et al., 2007), which is potentially advantageous in glaciogenic settings and/ or high latitude areas with restricted daylight. Preliminary polymineral TL, OSL, IRSL screening of HL core sediments suggested feldspars deposited in HL were more poorly bleached than quartz (Hodgson et al., in press-a), but further feldspar-based luminescence analyses might be useful, especially for ML samples with insufficient quartz. During sampling we observed sediment in shallow streams being exposed to strong sunlight while forming currently active deltas, suggesting sufficient exposure to reset the quartz OSL signal exists in this environment, at least during the ‘austral’ summer. Quartz was largely absent in both modern and elevated delta samples, however, and the quartz extracted had a lower sensitivity (OSL per unit dose) compared to other Antarctic environments investigated (e.g., Roberts et al., 2006). Consequently, most aliquots used in this study failed standard SAR-protocol internal quality checks. Fifty-percent or more of the recycling and recuperation data for each sample were incomplete or outside commonly accepted error ranges, suggesting that quartz in the AL and HL catchments may not be sufficiently sensitive for use in further single-grain analysis. 5.1.1. Modern OSL data The accuracy of OSL ages in glaciated and polar regions can be compromised where deposition occurred beneath ice and/or during

130

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

extended periods of darkness in the austral winter (Doran et al., 1999; Berger and Anderson, 2000; Berger and Doran, 2001; Gore et al., 2001). ‘Old’ ages from modern samples are a persistent problem for OSL dating studies of young (glacio)fluvial sediments (Wallinga, 2002), suggesting that some modern glacial environments do not provide sufficient light exposure to reset the OSL clock to zero. Meaningful conclusions from our modern samples are limited by the lack of sediment and quartz. The only sample with sufficient quartz, ML8 Pond, was taken from a pool of turbid silt-rich water up to 30 cm deep. If the OSL signal was fully reset by light exposure in a modern environment, we would expect these samples to return zero or near zero OSL ages. However, the turbidity of this pond probably produced differential bleaching of coarser material that had settled out on the bottom of the pond. Therefore, we consider the 7.3 ± 4.2 ka and the 9.7 ± 3.4 ka SAR-OSL ages from two discs (of four), which retained an OSL signal from ML8 Pond to be unreliable as the larger than expected De values most likely relate to incomplete bleaching. Furthermore, both discs failed recuperation tests, which suggests changes in luminescence sensitivity and/or thermal transfer occurred during the measurement process (Rhodes and Bailey, 1997; Rhodes, 2000). The dose rate of our modern pond (ML8) sample of 1.416 ± 0.037 mGya− 1 does not include any contribution made by the underlying bedrock or take account of potential changes in the rate or nature of sediment deposition over time. Consequently, this dose rate might not be representative of the actual environmental dose rate. Using the mean dose rate of ML delta sediments (2.843 ± 0.036 mGya− 1) provides a more accurate assessment of the likely environmental dose rate and reduces the ‘zero-age’ offset to c. 4.2 ± 2.4 ka and 5.5 ± 1.9 ka (Appendix A). This age estimate is closer to ‘zero-age’ offsets encountered in other luminescence studies in other parts of Antarctica. For example, Berger and Doran (2001) obtained ‘modern’ ages of between 600 and 3000 years from zero-age testing experiments in Lake Hoare in the McMurdo Dry Valleys of East Antarctica, by applying TL and IRSL dating methods to fine-grained sediments. The other two discs analysed produced insufficient OSL for analysis, consistent with recent bleaching in a modern depositional setting, and our original ‘zero-age’ hypothesis for these samples. 5.1.2. Elevated delta OSL data Apart from single-grain dating, the SAR-OSL method is the most reliable method currently available for dating glacial sediments (Klasen et al., 2007). Small aliquot samples used are thought to provide the most reliable multiple-grain equivalent dose results from complex depositional environments (Wallinga, 2002; Spencer et al., 2003; Rhodes, 2007). Nevertheless, ambiguity in the interpretation of OSL ages arises from partially bleached sediments since they carry residual signals from transport and depositional cycles earlier in the history of the minerals. Recent OSL dating studies on relict proglacial and glaciofluvial/deltaic sediments have encountered problems with incomplete bleaching of the luminescence signal prior to deposition, leading to anomalously old ages in some samples, and mixtures of differentially-bleached quartz grains (Rendell et al., 1994; Rhodes and Pownall, 1994; Berger and Anderson, 2000; Rhodes, 2000; Bitinas et al., 2001; Spencer and Owen, 2004; Raukas and Stankowski, 2005; Alexanderson and Murray, 2007; Boe et al., 2007; Gemmell et al., 2007; Klasen et al., 2007; Lukas et al., 2007; Thrasher et al., 2009). Fast component OSL quartz grains and grains whose OSL derives from other components could also exist within a deposit that has been evenly bleached, creating an ‘apparently’ inconsistently bleached sediment (Bailey, 2000; Murray and Wintle, 2003; Singarayer and Bailey, 2003). Gemmell et al. (2007) suggested that incompletely bleached samples can still provide an approximation of the maximum residual dose, and, where the dose rate can be accurately assessed, a good approximation of a maximum age for the deposit being dated.

Moreover, in SAR-OSL analysis, each aliquot produces a ‘date’; hence, the age distribution can be used to assess whether the sample contains a heterogeneous mixture differentially bleached sediment. Analysis of the quartz OSL decay shape can also give some insight into whether partial bleaching might have occurred in more homogeneous cases (Sanderson et al., 2001; Sommerville et al., 2003, 2007). The age distribution profile of the AL delta sample was relatively well-defined, compared to the HL upper delta age profile (see below), with a weighted mean age of 4.6 ± 0.4 ka for all data that passed internal validation tests (Table 2). However, cluster analysis of the AL delta profile highlights the existence of three possible age-related groups centred on c. 3.7 ka, 5.2 ka, and 7.0 ka (Table 2). This suggests that three stages of well-bleached deposition could have existed (a scenario which is partially backed-up by a disc at c. 7 ka that passed the internal validity tests), and/or incomplete resetting of quartz during the final phase of sedimentary deposition at c. 3–4 ka. The maximum possible age of the AL delta, from our limited dataset, is c. 12 ka; consistent with what is known about the deglaciation history of this part of Alexander Island (see Section 5.2). Where sufficient luminescence sensitivity exists in individual quartz grains, singlegrain OSL dating could be used to further characterise the age-profile of the AL delta. The lower HL delta was the most homogenously bleached sample examined in this study. Aliquots which passed validity tests have a weighted mean age of 4.4 ± 0.7 ka, which is similar to the 4.6 ± 0.5 ka weighted mean age obtained from all aliquots (Table 2). Two outlier ages (c. 11 ka and c. 30 ka) are consistent with age cluster groups from the upper HL delta, suggesting that the lower HL delta sediments could have been derived from the upper HL delta. However, the lower HL delta has a significantly higher C/N ratio of 34.8 compared to the upper HL delta bulk and b2 mm fractions, which have C/N ratios of 6.7 and 10.4, respectively (Hodgson et al., in press-a). This is probably due to the presence of calcite grains or a source of old carbon in the bulk sediment, suggesting different sediment sources for the upper and lower deltas, each with potentially different bleaching rates and/or bleaching environment. The wide scatter in the upper HL data probably reflects reworking of the delta deposit resulting in multiple bleaching histories. Therefore, we regard the youngest age of c. 11 ka as a maximum age estimate for the upper delta. Older ages from lower and upper HL deltas probably represent sediment that was incompletely-bleached during the final formation of the upper delta. 5.2. Elevated delta formation, RSL change and deglaciation on Alexander Island We propose that the AL (and ML) elevated deltas are the end product of glaciofluvial deposition from the catchment and reworking of sediment transported across the lake by lake-ice conveyors (Hall et al., 2006; Smith et al., 2006). After formation, the AL (and ML) elevated deltas were abandoned by changes in local RSL, resulting from LGM deglaciation and isostatic rebound of Alexander Island. The lake level high-stand at AL therefore provides an independent constraint on RSL change. It seems likely that a large percentage of RSL change and glacioisostatic rebound on Alexander Island will have been completed by the mid- to late Holocene since, globally, deglaciation-related sealevel rise was effectively complete by the mid-Holocene (Hall and Denton, 1999). Deglaciation and isostatic readjustment models for the AP (based largely on compilations of ‘palaeo’ data) show that grounded ice sheets extended to the continental shelf break at the LGM, and reached their current configuration in the early to midHolocene (Lambeck, 1993; Bentley, 1999; Ivins and James, 2005). Although models show that most uplift occurs within the first few thousand years following deglaciation (Lambeck, 1993), changes in the AL lake level were probably driven primarily by changes in

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

isostatic uplift because sea-level remained largely constant from the mid Holocene onwards (Bassett et al., 2007). The closest available RSL curve from the AP is from Marguerite Bay (Bentley et al., 2005a). The mid-shelf area of Marguerite Bay was deglaciated by c. 12,000 14C yr. BP (c. 11,700–12,100 cal. yr. B.P.) (Anderson et al., 2002; Ó Cofaigh et al., 2005). The inner part of Marguerite Bay area deglaciated by c. 9000 14C yr. BP, (c. 7800– 8400 cal. yr. B.P.) with relative sea level up to c. 41 m, or more, above present sea level (Bentley et al., 2005a). Alexander Island and the Marguerite Bay area have different postLGM deglaciation histories. Isostatic rebound may have been greater on islands in Marguerite Bay, which deglaciated earlier and more fully. Nevertheless, our AL elevated delta RSL age constraint is consistent with the field-based deglaciation and RSL history of Marguerite Bay. The 4.6 ± 0.4 ka constraint on 14.4 m asl of RSL is bracketed by the two most reliable existing constraints on RSL change in the region (Bentley et al., 2005b): the first occurrence of completely freshwater conditions after the transition from marine to terrestrial sediments in cores extracted from uplifted isolation basins on Pourquoi-Pas Island (19.4 m asl at 6420 ± 50 14C yr. BP, c. 5320–5480 cal. yr. B.P.) (Bentley et al., 2005b) and Horseshoe Island (3.5 m asl at 1320 ± 70 14C yr. BP, c. 600–880 cal. yr. B.P.) (Wasell and Hakansson, 1992). Different isostatically-coupled ice sheets models predict different amounts of uplift for the AP (e.g., James and Ivins, 1998; Ivins et al., 2000; Ivins and James, 2005). Predictions using the glacial isostatic adjustment models described in Bassett et al. (2007) and the ICE-5GVM2 model (90 km-thick lithosphere) (Peltier, 2004) define a range of uplift magnitudes for the late Holocene that are compatible with the observed value of 14.4 m of uplift since c. 4.6 ka. Field constraints on RSL change from elevated deltas on Alexander Island are important, therefore, because they can reduce the number of possible model scenarios, and improve the predictive ability of future isostaticallycoupled ice sheet and sea-level change models. The mid-Holocene formation of elevated deltas at AL is also consistent with the deglaciation history of the Antarctic Peninsula ice sheet since the LGM. Cosmogenic exposure ages and sediment core data suggest: Alexander Island was overridden with ice, over 470 m thick, at the LGM (c. 21 ka) (Bentley et al., 2006); deglaciation of some upper ridges above c. 400 m occurred by c. 13.5 ka BP (Hodgson et al., in press-a); and the present ice sheet configuration existed sometime after c. 7.2 ka (Bentley et al., 2006; Smith et al., 2006, 2007a,b). This timescale of events is in broad agreement with some ice sheet models (e.g., Payne et al., 1989; Huybrechts, 2002), as well as the proposed deglaciation history of HL (Hodgson et al., in press-a,b). During post-LGM deglaciation, there was a progressive drop in lake level at HL as the Saturn Glacier thinned (Hodgson et al., in press-a,b). To account for the formation of upper and lower deltas, it is likely that both the Saturn Glacier and the HL catchment glaciers remained relatively thick between the exposure of the col until at least 13.5 ka (Hodgson et al., in press-a), during the early Holocene retreat of GVI-IS (Bentley et al., 2005a). The SAR-OSL age of the upper HL delta suggests deltas were forming on the elevated lake ice surface from c. 11 ka. This is consistent with cosmogenic exposure ages from the HL area, which suggests that the ridge on the western shore of HL emerged sometime after c. 13 ka. The well-defined SAR-OSL age of the lower HL delta places it in the mid-Holocene ‘hypsithermal’, a ‘warm period’ widely documented in climate records from the northern Antarctic Peninsula and sub-Antarctic islands (see Hodgson et al. (2004) and Bentley et al. (2009) and for summaries). The age of the lower HL delta implies HL reached its present configuration as a c. 90 m-deep perennially ice covered lake sometime after 4.4 ± 0.7 ka, when the lower delta became isolated from the current hydrological system, and after significant post-LGM and Holocene thinning of AP ice sheets had occurred.

131

6. Conclusions (1) We investigated the geomorphology, sedimentology and geochronology of elevated deltas in the two epi-shelf lake catchments of Ablation Lake (AL) and Moutonnée Lake (ML) and compared their formation processes with elevated deltas in the ice-dammed inland Hodgson Lake (HL), providing new terrestrial constraints on relative sea level (RSL) and deglaciation for Alexander Island, Antarctic Peninsula. (2) Weighted mean SAR-OSL ages that passed internal validity tests from the elevated delta at AL (4.6 ± 0.4 ka) and lower HL delta (4.4 ± 0.7 ka) represent the last time active deltas were forming at higher than present day lake levels in the AL and HL catchments. (3) SAR-OSL data from the AL elevated delta place a new chronological and vertical constraint on regional RSL change, suggesting up to 14.4 m fall in RSL occurred in the last c. 4.6 ± 0.4 ka in this part of Alexander Island. This is broadly consistent with the field-based RSL and Holocene deglaciation history of Alexander Island and the Marguerite Bay region, as well as some glacial isostatic adjustment models (e.g., Peltier, 2004; Bassett et al., 2007), but suggests that some other models might have overestimated ice mass fluctuations and the rate of RSL change across the Antarctic Peninsula since c. 6 ka. Acknowledgements This study was funded by the BAS core program ‘Signals in Antarctica of past Global changES’ (SAGES)-10K, led by DH. Additional work was undertaken as part of the CACHE-PEP(B) program, led by DH/Eric Wolff, and as part of the NERC Antarctic Funding Initiative project (AFI2/69 and CGS00/06) led by DH, MB and David Sugden. Thanks to Pete Milner, Adam Hunt and the BAS Field Operations Staff for assistance in the field, and the Mapping and Geographic Information Centre (MAGIC) at the BAS (namely Adrian Fox, Peter Fretwell and Andreas Czifersky) for maps and additional field data from Citadel Bastion; Dr. Dodie James for assistance with XRF analysis and Valerie Olive for assistance with ICP-MS analysis. We are grateful to Eric Wolff, Alan Rodger and two reviewers for their constructive comments. Appendix A. Supplementary data Equivalent dose data for modern and elevated delta samples AB1, CB6.5 and CB8.4; TD 7/1 is the ratio of test doses 1 and 7; ⁎ based on the measured total dose rate value of 1.416 ± 0.037 mGya− 1 (Table 1); # based on the mean dose rate for ML1A, ML2A, and ML1B of 2.843 ± 0.036 (Table 1). Equivalent dose (De) method summary: To establish the ages of the deltas and modern samples, we undertook 90–150 or 150– 250 µm quartz fraction single aliquot regeneration-optically stimulated luminescence (SAR-OSL) analysis (Murray and Wintle, 2000, 2003) using modified standard procedures. These included recycling ratio and recuperation measurements to check that sensitivity changes induced by the measurement process are being adequately corrected. Samples were prepared following standard procedures (Sanderson et al., 2001), and analysed using an automated Risø TL/ OSL system (TL-DA-15) using blue LEDs with a wavelength of 470 nm for stimulation. After the natural signal was analysed (60 s; 60% power; 125 °C), a sequence of doses and preheats and read-outs were administered to reconstruct the regeneration line. A repeated dose was used to calculate recycling ratios that examine the effectiveness of sensitivity corrections. A normalisation test dose was added, preheated and read out after each dose point read-out. At the end of the run, a final dose was given, followed by an IRSL readout to check for possible feldspar contamination. Sequence

132

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

summary: Preheat 220–280 °C for 10 s, groups of 4 discs; readout 60% power, 60 s at 125 °C; test dose = 10 Gy; dose points 50, 100, 150, 200, 250, 0, 50, 10, 20, 30, 10 Gy + 10 Gy IRSL readout. The range of preheat temperatures were used to assess thermal transfer effects (Murray and Wintle, 2000; Rhodes, 2000; Murray and Wintle, 2003). Discs with a recycling ratio 0.90–1.10 (Spencer and Owen, 2004), and no evidence of feldspar contamination determined from the IRSL readout data (not shown), were classified as a ‘pass’ and used to calculate the ‘most reliable’ ages (shown in bold) for AB1 and CB6.5. Group structures and outliers were defined by cluster analysis using XLSTAT Version 2009.2.01 (http://www.xlstat.com/en/home/) using Wards' method, with Euclidean distance between individual age data points as a dissimilarity measure, weighted by individual age data point error values. Ward's method was considered most suitable because it uses analysis of within group variance to evaluate the distance between clusters. Environmental dose rate (DT) method summary: Values were based on gamma spectrometry (in-situ and laboratory-based) and laboratory-based Thick Source Beta Counting (TSBC) (Sanderson, 1998), coupled with X-ray fluorescence (XRF) and Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) measurements of elemental concentrations. For XRF, 2–5 g of bulk sediment was hand-crushed using an agate mortar and pestle. Ten major element run were analysed using fused-disc method in a Philips PW2404 wavelength-dispersive, sequential XRF spectrometer (Norrish and Hutton, 1969). Precision and accuracy are typically 5–10% and were assessed by measuring geostandard BIR1 routinely and quantified as the root mean square deviation (rmsd), expressed in wt.% of the calibration data about the regression line (Govindaraju, 1994; Fitton et al., 1998). ICP-MS analysis of a suite of trace and REE was undertaken on a VG Elemental PQ2 Plus mass spectrometer. Triplicate measurements were made on each sample, and bracketed by regular analyses of geostandards BCR, AGV-1, GA, BE-N, DR-N to assess precision and accuracy, which are typically 1–2%. In-situ gamma spectrometry was measured from deltas in AL and ML using a ORTEC gamma spectrometer with a 7.62 × 7.62 cm NaI sensor. Count time was 900 s. Laboratory-based gamma spectrometry and TSBC delta samples were crushed to a fine powder using a steel Tema Mill. TSBC samples were counted for six cycles of 300 s (Sanderson, 1988). Geochemical analyses were performed on a sub-sample of TSBC bulk material. Where applicable, effective total beta dose rate is calculated from water corrected means of TSBC values and chemically derived dry beta dose values. Water contents were measured as received. High resolution, low-level gamma spectrometry was undertaken on finely powdered bulk sediment with count times of 50,000 s or 100,000 s to determine the concentration of isotopes in the U- and Th-series decay chains and assess the degree of secular disequilibrium.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.geomorph.2009.05.011. References Adamiec, G., Aitken, M.J., 1998. Dose-rate conversion factors: new data. Ancient TL 16, 37–50. Aitken, M.J., 1998. An introduction to optical dating. The Dating of Quaternary Sediments by the Use of Photon-stimulated Luminescence. InOxford University Press, Oxford, p. 270. Alexanderson, H., Murray, A.S., 2007. Was southern Sweden ice free at 19–25 ka, or were the post LGM glacifluvial sediments incompletely bleached? Quaternary Geochronology 2, 229–236. Anderson, J.B., Shipp, S.S., Lowe, A.L., Wellner, J.S., Mosola, A.B., 2002. The Antarctic Ice Sheet during the Last Glacial Maximum and its subsequent retreat history: a review. Quaternary Science Reviews 21, 49–70.

Arthern, R.J., Hindmarsh, R.C.A., 2006. Determining the contribution of Antarctica to sea level rise using data assimilation methods. Philosphical Transactions of the Royal Society A 364, 1841–1865. doi:10.1098/rsta.2006.1801. Bailey, R.M., 2000. The slow component of quartz optically stimulated luminescence. Radiation Measurements 32, 233–246. Bassett, S.E., Milne, G.A., Mitrovica, J.X., Clark, P.U., 2005. Ice Sheet and solid earth influences on far-field sea-level histories. Science 309, 925–928. doi:10.1126/science.1111575. Bassett, S.E., Milne, G.A., Bentley, M.J., Huybrechts, P., 2007. Modelling Antarctic sealevel data to explore the possibility of a dominant Antarctic contribution to meltwater pulse IA. Quaternary Science Reviews 26, 2113–2127. Bentley, M.J., 1999. Volume of Antarctic Ice at the Last Glacial Maximum, and its impact on global sea level change. Quaternary Science Reviews 18, 1569–1595. Bentley, M.J., Hodgson, D.A., Smith, J.A., Cox, N.J., 2005a. Relative sea level curves for the South Shetland Islands and Marguerite Bay, Antarctic Peninsula. Quaternary Science Reviews 24, 1203–1216. Bentley, M.J., Hodgson, D.A., Sugden, D.E., Roberts, S.J., Smith, J.A., Leng, M.J., Bryant, C., 2005b. Early Holocene retreat of the George VI Ice Shelf, Antarctic Peninsula. Geology 33, 173–176. Bentley, M.J., Fogwill, C.J., Kubik, P.W., Sugden, D.E., 2006. Geomorphological evidence and cosmogenic 10Be/26Al exposure ages for the Last Glacial Maximum and deglaciation of the Antarctic Peninsula Ice Sheet. Geological Society of America Bulletin 1149–1159. Bentley, M.J., Hodgson, D.A., Smith, J.A., Cofaigh, C.O., Domack, E.W., Larter, R.D., Roberts, S.J., Brachfeld, S., Leventer, A., Hjort, C., Hillenbrand, C.D., Evans, J., 2009. Mechanisms of Holocene palaeoenvironmental change in the Antarctic Peninsula region. The Holocene 19, 51–69. doi:10.1177/0959683608096603. Berger, G.W., Anderson, P.M., 2000. Extending the geochronometry of Arctic lake cores beyond the radiocarbon limit using thermoluminescence. Journal of Geophysical Research 105 (D12), 15,439–15,455. Berger, G.W., Doran, P.T., 2001. Luminescence-dating zeroing tests in Lake Hoare, Taylor Valley, Antarctica. Journal of Paleolimnology 25, 519–529. Bindoff, N.L., Willebrand, J., Artale, V., Cazenave, A., Gregory, J., Gulev, S., Hanawa, K., Quere, C.L., Levitus, S., Nojiri, Y., Shum, C.K., Talley, L.D., Unnikrishnan, A., 2007. Observations: oceanic climate change and sea level. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. Bitinas, A., Damusyte, A., Hutt, G., Jaek, I., Kabailiene, M., 2001. Application of the OSL dating for stratigraphic correlation of Late Weichselian and Holocene sediments in the Lithuanian Maritime Region. Quaternary Science Reviews 20, 767–772. Boe, A.-G., Murray, A., Dahl, S.O., 2007. Resetting of sediments mobilised by the LGM icesheet in southern Norway. Quaternary Geochronology 2, 222–228. Browne, J., 1996. Sandstone provenance and diagenesis of arc-related basins: James Ross Island and Alexander Island, Antarctica, Unpublished PhD Thesis. University of Exeter, Exeter, pp. 464. Butterworth, P.J., 1991. The role of eustasy in the development of a regional shallowing event in a tectonically active basin: Fossil Bluff Group (Jurassic–Cretaceous), Alexander Island, Antarctica. In: Macdonald, D.I.M. (Ed.), Sedimentation, Tectonics and Eustasy. Special Publication of the International Association of Sedimentologists. Blackwell Scientific Publications, Oxford, UK, pp. 307–329. Clapperton, C.M., Sugden, D.E., 1982. Late Quaternary glacial history of George VI Sound area, West Antarctica. Quaternary Research 18, 243–267. Clapperton, C.M., Sugden, D.E., 1983. Geomorphology of the Ablation Point massif, Alexander Island, Antarctica. Boreas 12, 125–135. Cook, A.J., Fox, A.J., Vaughan, D.G., Ferrigno, J.G., 2005. Retreating glacier fronts on the Antarctic Peninsula over the past half-century. Science 308, 541–544. De Angelis, H., Skvarca, P., 2003. Glacier surge after ice shelf collapse. Science 299, 1560–1562. Doran, P.T., Berger, G.W., Lyons, W.B., Wharton Jr., R.A., Davisson, M.L., Southon, J., Dibb, J.E., 1999. Dating Quaternary lacustrine sediments in the McMurdo Dry Valleys, Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology 147, 223–239. Duller, G.A.T., 2003. Distinguishing quartz and feldspar in single grain luminescence measurements. Radiation Measurements 37, 161–165. Elliott, M.H., 1974. Stratigraphy and sedimentary petrology of the Ablation Point area, Alexander Island. British Antarctic Survey Bulletin 39, 87–113. Fitton, J.G., Saunders, A.D., Larsen, L.M., Hardarson, B.S., Norry, M.J., 1998. Volcanic rocks from the southeast Greenland margin at 63 degrees N: composition, petrogenesis, and mantle sources. Proceedings of the Ocean Drilling Program, Scientific results, vol. 152. Gemmell, A.M.D., Murray, A.S., Connell, E.R., 2007. Devensian glacial events in Buchan (NE Scotland): a progress report on new OSL dates and their implications. Quaternary Geochronology 2, 237–242. Godfrey-Smith, D.I., Huntley, D.J., Chen, W.-H., 1988. Optical dating studies of quartz and feldspar sediment extracts. Quaternary Science Reviews 7, 373–380. Gore, D.B., Rhodes, E.J., Augustinus, P.C., Leishman, M.R., Colhoun, E.A., Rees-Jones, J., 2001. Bunger Hills, East Antarctica: ice free at the Last Glacial Maximum. Geology 29, 1103–1106. Govindaraju, K., 1994. Compilation of working values and sample description for 383 geostandards. Geostandards Newsletter 18, 1. Hall, B.L., Denton, G.H., 1999. New relative sea-level curves for the southern Scott Coast, Antarctica: evidence for Holocene deglaciation of the western Ross Sea. Journal of Quaternary Science 14, 641–650. Hall, B.L., Denton, G.H., Overturf, B., 2001. Glacial Lake Wright, a high-level antarctic lake during the LGM and early Holocene. Antarctic Science 13, 53–60.

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134 Hall, B.L., Denton, G.H., Overturf, B., Hendy, C.H., 2002. Glacial Lake Victoria, a high-level Antarctic lake inferred from lacustrine deposits in Victoria Valley. Journal of Quaternary Science 17, 697–706. Hall, B.L., Baroni, C., Denton, G.H., 2004. Holocene relative sea-level history of the Southern Victoria Land Coast, Antarctica. Global and Planetary Change 42, 241–263. Hall, B.L., Hendy, C.H., Denton, G.H., 2006. Lake-ice conveyor deposits: geomorphology, sedimentology, and importance in reconstructing the glacial history of the Dry Valleys. Geomorphology 75, 143–156. Heywood, R.B., 1977. A limnological survey of the Ablation Point area, Alexander Island, Antarctica. Philosophical Transactions of the Royal Society of London B279, 39–54. Hock, R., de Woul, M., Radic, V., Dyurgerov, M., 2009. Mountain glaciers and ice caps around Antarctica make a large sea level rise contribution. Geophysical Research Letters 36, L07501. doi:10.1029/2008GL037020. Hodgson, D.A., Doran, P.T., Roberts, D., McMinn, A., 2004. Paleolimnological studies from the Antarctic and sub-Antarctic islands. In: Pienitz, R., Douglas, M.S.V., Smol, J.P. (Eds.), Developments in Palaoenvironmental Research. Longterm Environmental Change in Arctic and Antarctic Lakes. Springer, Dordrecht, pp. 419–474. Hodgson, D.A., Roberts, S.J., Bentley, M.J., Smith, J.A., Johnson, J., Fox, A.J., Leng, M.J., Sanderson, D.C.W., Verleyen, E., Vyverman, W., Geissler, P., Carmichael, E.L., Cziferszky, A., Sugden, D.E., in press-a. Palaeolimnology of Citadel Lake: a former subglacial lake at 72° South? Quaternary Science Reviews. doi:10.1016/j. quascirev.2009.04.011. Hodgson, D.A., Roberts, S.J., Bentley, M.J., Smith, J.A., Johnson, J., Fox, A.J., Leng, M.J., Sanderson, D.C.W., Verleyen, E., Vyverman, W., Geissler, P., Carmichael, E.L., Cziferszky, A., Sugden, D.E., in press-b. Geomorphology and limnology of Citadel Lake: a former subglacial lake at 72° South? Quaternary Science Reviews. doi:10.1016/j.quascirev.2009.04.014. Horne, R.R., 1968. Petrology and provenance of the Cretaceous sediments of southeastern Alexander Island. British Antarctic Survey Bulletin 17, 73–82. Huybrechts, P., 2002. Sea-level changes at the LGM from ice-dynamic reconstructions of the Greenland and Antarctic ice sheets during the glacial cycles. Quaternary Science Reviews 21, 203–231. Ingólfsson, Ó., Hjort, C., 1999. The Antarctic contribution to Holocene global sea level rise. Polar Research 18, 323–330. Ingólfsson, Ó., Hjort, C., 2002. Glacial history of the Antarctic Peninsula since the Last Glacial Maximum—a synthesis. Polar Research 21, 227–234. Ingólfsson, Ó., Hjort, C., Berkman, P.A., Björck, S., Colhoun, E.A., Goodwin, I.D., Hall, B.L., Hirakawa, K., Melles, M., Möller, P., Prentice, M.L., 1998. Antarctic glacial history since the last glacial maximum: an overview of the record on land. Antarctic Science 10, 326–344. Ingólfsson, Ó., Hjort, C., Humlum, O., 2003. Glacial and climate history of the Antarctic Peninsula since the Last Glacial Maximum. Arctic, Antarctic, and Alpine Research 35, 175–186. IPCC, 2007. Summary for policymakers. In: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom. Ivins, E.R., James, T.S., 2005. Antarctic glacial isostatic adjustment: a new assessment. Antarctic Science 17, 541–553. Ivins, E.R., Raymond, C.A., James, T.S., 2000. The influence of 5000 year-old and younger glacial mass variability on present-day crustal rebound in the Antarctic Peninsula. Earth, Planets, and Space 52, 1023–1029. James, T.S., Ivins, E.R., 1998. Predictions of Antarctic crustal motions driven by presentday ice sheet evolution and by isostatic memory of the Last Glacial Maximum. Journal of Geophysical Research-Solid Earth 103, 4993–5017. Klasen, N., Fiebig, M., Preusser, F., Reitner, J.M., Radtke, U., 2007. Luminescence dating of proglacial sediments from the Eastern Alps. Quaternary International 164–65, 21–32. Lambeck, K., 1993. Glacial rebound and sea-level change—an example of a relationship between mantle and surface processes. Tectonophysics 223, 15–37. Lukas, S., Spencer, J.Q.G., Robinson, R.A.J., Benn, D.I., 2007. Problems associated with luminescence dating of Late Quaternary glacial sediments in the NW Scottish Highlands. Quaternary Geochronology 2, 243–248. Medahl, V., 1979. Thermoluminescence dating: beta attenuation in quartz grains. Archaeometry 21. Murray, A.S., Wintle, A.G., 2000. Luminescence dating of quartz using an improved single-aliquot regenerative-dose protocol. Radiation Measurements 32, 57–73. Murray, A.S., Wintle, A.G., 2003. The single aliquot regenerative dose protocol: potential for improvements in reliability. Radiation Measurements 37, 377–381. Norrish, K., Hutton, J.T., 1969. An accurate X-ray spectrographic method for the analysis of a wide range of geological samples. Geochemica et Cosmochimica Acta 33, 431–453. Ó Cofaigh, C., Dowdeswell, J.A., Allen, C.S., Hiemstra, J.F., Pudsey, C.J., Evans, J., Evans, D.J.A., 2005. Flow dynamics and till genesis associated with a marine-based Antarctic palaeo-ice stream. Quaternary Science Reviews 24, 709–740. Payne, A.J., Sugden, D.E., Clapperton, C.M., 1989. Modeling the growth and decay of the Antarctic Peninsula Ice Sheet. Quaternary Research 31, 119–134. Pearson, M.R., Rose, I.H., 1983. The dynamics of George VI ice shelf. British Antarctic Survey Bulletin 52, 205–220. Peltier, W.R., 2004. Global glacial isostacy and the surface of the Ice Age Earth: the ICE5G (VM2) Model and GRACE. Annual Reviews of Earth and Planetary Sciences 32, 111–149. doi:10.1146/annurev.earth.32.082503.144359.

133

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. Pritchard, H., Vaughan, D.G., 2007. Widespread acceleration of tidewater glaciers on the Antarctic Peninsula. Journal of Geophysical Research 112, F03S29. doi:10.1029/ 2006JF000597. Raukas, A., Stankowski, W., 2005. Influence of sedimentological composition on OSL dating of glaciofluvial deposits: examples from Estonia. Geological Quarterly 49, 463–470. Rendell, H.M., Webster, S.E., Sheffer, N.L., 1994. Underwater bleaching of signals from sediment grains—new experimental-data. Quaternary Science Reviews 13, 433–435. Rhodes, E.J., 2000. Observations of thermal transfer OSL signals in glacigenic quartz. Radiation Measurements 32, 595–602. Rhodes, E.J., 2007. Quartz single grain OSL sensitivity distributions: implications for multiple grain single aliquot dating. Geochronometria 26, 19–29. Rhodes, E.J., Pownall, L., 1994. Zeroing of the OSL signal in quartz from young glaciofluvial sediments. Radiation Measurements 23, 581–585. Rhodes, E.J., Bailey, R.M., 1997. The effect of thermal transfer on the zeroing of the luminescence of quartz from recent glaciofluvial sediments. Quaternary Science Reviews 16, 291–298. Rignot, E., Casassa, G., Gogineni, P., Krabill, W., Rivera, A., Thomas, R., 2004. Accelerated ice discharge from the Antarctic Peninsula following the collapse of Larsen B Ice Shelf. Geophysical Research Letters 31. Rignot, E., Casassa, G., Gogineni, S., Kanagaratnam, P., Krabill, W., Pritchard, H., Rivera, A., Thomas, R., Turner, J., Vaughan, D., 2005. Recent ice loss from the Fleming and other glaciers, Wordie Bay, West Antarctic Peninsula. Geophysical Research Letters 32. Roberts, S.J., Mulvaney, R., Hillenbrand, C.D., Triest, J., Hodgson, D.A., Sanderson, D.C.W., Sommerville, A.A., 2006. Preliminary analyses on quartz-rich sand from near the base of the Berkner Island Ice Core, Antarctica. Geophysical Research Abstracts X ISSN: 1029-7006; (Abstract no.: EGU06-A-01175). Roberts, S.J., Hodgson, D.A., Bentley, M.J., Smith, J.A., Millar, I., Olive, V., Sugden, D.E., 2008. The Holocene history of George VI Ice Shelf, Antarctic Peninsula from clastprovenance analysis of epishelf lake sediments. Palaeogeography Palaeoclimatology Palaeoecology 259, 258–283. Rott, H., Skvarca, P., Nagler, T., 1996. Rapid collapse of Northern Larsen Ice Shelf, Antarctica. Science 271, 788–792. Sanderson, D.C.W., 1988. Thick source beta counting (TSBC): a rapid method for measuring beta dose-rates. Nuclear Tracks and Radiation Measurements 14, 203–207. Sanderson, D.C.W., 1998. Thick source beta counting (TSBC): a rapid method for measuring beta dose-rates. Nuclear Tracks and Radiation Measurements 14, 203–207. Sanderson, D.C.W., Bishop, P., Houston, I., Boonsener, M., 2001. Luminescence characterisation of quartz-rich cover sands from NE Thailand. Quaternary Science Reviews 20, 893–900. Sanderson, D.C.W., Bishop, P., Stark, M., Alexander, S., Penny, D., 2007. Luminescence dating of canal sediments from Angkor borei, mekong delta, southern Cambodia. Quaternary Geochronology 2, 322–329. Scambos, T.A., Bohandler, J., Shuman, C.A., Skvarca, P., 2004. Glacier acceleration and thinning after ice shelf collapse in the Larsen B Embayment, Antarctica. Geophysical Research Letters 31, L18402. doi:10.1029/2004GL0260670. Singarayer, J.S., Bailey, R.M., 2003. Further investigations of the quartz optically stimulated luminescence components using linear modulation. Radiation Measurements 37, 451–458. Smith, J.A., Hodgson, D.A., Bentley, M.J., Verleyen, E., Leng, M.J., Roberts, S.J., 2006. Limnology of two Antarctic epishelf lakes and their potential to record periods of ice shelf loss. Journal of Paleolimnology 35, 373–394. Smith, J.A., Bentley, M.J., Hodgson, D.A., Cook, A.J., 2007a. George VI Ice Shelf: past history, present behaviour and potential mechanisms for future collapse. Antarctic Science 19, 131–142. Smith, J.A., Bentley, M.J., Hodgson, D.A., Roberts, S.J., Leng, M.J., Lloyd, J.M., Barrett, M.S., Bryant, C., Sugden, D.E., 2007b. Oceanic and atmospheric forcing of early Holocene ice shelf retreat, George VI Ice Shelf, Antarctica Peninsula. Quaternary Science Reviews 26, 500–516. Sommerville, A.A., Hansom, J.D., Sanderson, D.C.W., Housley, R.A., 2003. Optically stimulated luminescence dating of large storm events in Northern Scotland. Quaternary Science Reviews 22, 1085–1092. Sommerville, A.A., Hansom, J.D., Housley, R.A., Sanderson, D.C.W., 2007. Optically stimulated luminescence (OSL) dating of coastal aeolian sand accumulation in Sanday, Orkney Islands, Scotland. The Holocene 17, 627–637. doi:10.1177/ 0959683607078987. Spencer, J.Q., Owen, L.A., 2004. Optically stimulated luminescence dating of Late Quaternary glaciogenic sediments in the upper Hunza valley: validating the timing of glaciation and assessing dating methods. Quaternary Science Reviews 23, 175–191. Spencer, J.Q., Sanderson, D.C.W., Deckers, K., Sommerville, A.A., 2003. Assessing mixed dose distributions in young sediments identified using small aliquots and a simple two-step SAR procedure: the F-statistic as a diagnostic tool. Radiation Measurements 37, 425–431. Sugden, D.E., Clapperton, C.M., 1981. An ice-shelf moraine, George VI Sound, Antarctica. Annals of Glaciology 2, 135–143. Thrasher, I.M., Mauz, B., Chiverrell, R.C., Lang, A., Thomas, G.S.P., 2009. Testing an approach to OSL dating of Late Devensian glaciofluvial sediments of the British Isles. Journal of Quaternary Science. doi:10.1002/jqs.1253.

134

S.J. Roberts et al. / Geomorphology 112 (2009) 122–134

Vaughan, D.G., 1993. Implications of the break-up of Wordie Ice Shelf, Antarctica for sealevel. Antarctic Science 5, 403–408. Vaughan, D.G., 2006. Recent trends in melting conditions on the Antarctic Peninsula and their implications for ice-sheet mass balance and sea level. Arctic, Antarctic, and Alpine Research 38 (1), 147–152. Vaughan, D.G., Doake, C.S.M., 1996. Recent atmospheric warming and retreat of ice shelves on the Antarctic Peninsula. Nature 379, 328–331. Vaughan, D.G., Marshall, G., Connolley, W.M., Parkinson, C., Mulvaney, R., Hodgson, D.A., King, J.C., Pudsey, C.J., Turner, J., Wolff, E., 2003. Recent rapid regional climate warming on the Antarctic Peninsula. Climatic Change 60, 243–274.

Wallinga, J., 2002. Optically stimulated luminescence dating of fluvial deposits: a review. Boreas 31, 303–322. Wasell, A., Hakansson, H., 1992. Diatom stratigraphy in a lake on Horseshoe Island, Antarctica: a marine-brackish-freshwater transition with comments on the systematics and ecology of the most common diatoms. Diatom Research 7, 157–194.