Scratching the Surface: A Marine Sediment Provenance Record From the Continental Slope of Central Wilkes Land, East Antarctica

The geology of Wilkes Land, East Antarctica, is masked by kilometers of ice and remains largely unexplored. Defining the sediment provenance adjacent to this hidden region is important for distinguishing the proximal subglacial basement terranes and refining the dynamic regional glaciological history. This study presents a detrital sediment provenance record spanning c. 23.5 ka from the continental slope of central Wilkes Land. Sediment provenance was characterized using U‐Pb geochronology and trace element geochemistry from detrital zircon, titanite and apatite, and Pb isotopic signatures from detrital feldspar. These data were compared with new feldspar Pb‐isotopic signatures and existing U‐Pb zircon data sets from rare nearby coastal outcrop. A principally igneous source was revealed with dominant age populations between c. 1,360‐1,100 Ma and c. 1,620‐1,490 Ma, characteristic of rocks of the proximal Wilkes and Banzare provinces, respectively. Minor detritus was additionally sourced from the proximal Nuyina Province (c. 1,450‐1,390 Ma). Temporal variation in the climate and ice sheet configuration are likely responsible for subtle downcore changes observed in detrital sediment provenance. High sedimentation rates during the glacial period suggest reworking of continental shelf sediments and downslope transport in debris flows during ice sheet advance. Glacial meltwater fluxes fed largely by the Totten Glacier were responsible for supplying detritus during deglaciation. During interglacials, detritus was derived from a broad coastal region and delivered to the slope via multiple glacial outlets. These results present the first substantial offshore evidence to support recent interpretations that the subglacial crust of central Wilkes Land has a dominantly Mesoproterozoic history.

. Additionally, deciphering the bedrock composition can help identify sediment erosion and transport pathways and constrain subglacial conditions (e.g., geothermal heat flow, topography, and basal erosion) deemed to have impacted Antarctic Ice Sheet evolution (Golledge et al., 2013;Jamieson et al., 2010;Lowry et al., 2020). In this context, the Sabrina Subglacial Basin in Wilkes Land (Aitken et al., 2016b; Figure 1a) is of particular interest, forming part of an extensive glacial drainage system that is predominantly situated below sea-level (Fretwell et al., 2013;Young et al., 2011). This region drains into the ocean via the rapidly retreating Totten Glacier (Rignot et al., 2019;Figure 1b) and has the capacity to raise global sea-level by 3.5 m (Greenbaum et al., 2015).
Multiple sedimentary features of the continental slope of the Sabrina Coast provide evidence for sediment delivery via meltwater fluxes supplied by major glacier systems throughout the Late Cenozoic (c. 34 Ma to present; Donda et al., 2020). The Totten Glacier is considered to have played a central role in controlling these sediment-laden meltwater fluxes that have historically shaped the submarine landscape. The slope bathymetry in this region is characterized by two large contourite ridges to the west and a dendritic network of gullies that develop into submarine canyons to the east Post et al., 2020;Figure 1a). These sedimentary features have been observed and interpreted in other regions around East Antarctica (e.g., De Santis et al., 2007;Donda et al., 2020;O'Brien et al., 2016) and West Antarctica (Amblas & Canals, 2016;Dowdeswell et al., 2004Dowdeswell et al., , 2006Gales et al., 2013) and provide records of sediment erosion during the evolution of the ice sheet.
Studies of marine sediment provenance around Antarctica have been used to investigate the subglacial terrain in the absence of onshore geological data and to complement nonunique geophysical signatures (Licht & Hemming, 2017;Pierce et al., 2014;Roy et al., 2007;Williams et al., 2010). In this study, we aim to test existing interpretations of subglacial bedrock composition and erosion pathways in central Wilkes Land. We do this by characterizing the isotopic and geochemical signature of detrital grains in a sediment core recovered from the eastern flank of a large contourite ridge situated on the continental slope of the Sabrina Coast in central Wilkes Land (Figure 1a).

Background: A Brief Glaciological and Geological History of Wilkes Land
Our interpretation of the composition, provenance, and transport of continentally derived detritus from Antarctica is informed by current understanding of the temporal changes to the Antarctic Ice Sheet and the regional geological history. Here, we provide a brief synopsis of the East Antarctic Ice Sheet configuration from its maximum extent during the last glacial period to present, followed by a summary of the geological evolution of central Wilkes Land, East Antarctica.

Evolution of the East Antarctic Ice Sheet from the Late Pleistocene to Present
The expansion and retreat of the ice sheet over glacial to interglacial timescales has left an imprint on the sedimentary basin infill in central Wilkes Land (Aitken et al., 2016b). Region A of the Sabrina Sedimentary Basin (Figure 1a), is characterized by a thin subglacial sedimentary rock veneer (<1 km in thickness), in some places exposing crystalline basement rock at the base of the ice sheet (Aitken et al., 2016b). The scarcity of sedimentary rock between the base of the ice sheet and basement geology is interpreted as indicating repeated erosion during cyclic waxing and waning of the ice sheet in this region (Aitken et al., 2016b). Ancient shelf sedimentary megasequences provide additional evidence of repeated ice sheet advance and retreat from the beginning of the Oligocene to the mid-Miocene (Gulick et al., 2017).
Sedimentary bedforms on the continental shelf and slope of central Wilkes Land also preserve a dynamic history of sediment erosion, transport, and deposition. Throughout the Late Cenozoic (phases 2-4 of Donda et al., 2007), the East Antarctic Ice Sheet was unstable and sediment-laden meltwater fluxes sourced from major outlet glaciers (Figure 1b) had an important influence on the morphology of the seabed on the continental shelf, slope, and rise Leitchenkov, 2015). In the west of the study region, large sediment contourites formed on the continental slope and are maintained by debris flows and fallout from sediment-rich meltwater plumes O'Brien et al., 2020). Donda et al. (2020) suggested that the proximal Totten Glacier likely supplies a significant component of detritus to the slope in sediment-laden meltwater fluxes, maintaining these submarine landforms. Gullies developed in the east of TOOZE ET AL.
2 of 23  Morlighem et al. (2020). High-resolution bathymetry of the continental slope was acquired from RV Investigator (voyage IN2017_V01; Armand et al., 2018) and of the continental shelf from the RV Nathaniel B. Palmer (voyage NBP1402; Fernandez et al., 2018). Simplified outline of the Sabrina Sedimentary Basin (bold dark blue line) is from Maritati et al. (2019) after Aitken et al. (2016b). Region A, an area of thin sedimentary rock cover (<1-km thick) and basement rock exposure is interpreted by Aitken et al. (2016b). Gray lines mark the boundaries of ice drainage basins (Mouginot et al., 2017 the study region, where turbidite flows and meltwater fed by major glacial systems scoured into the seabed O'Brien et al., 2020;Post et al., 2020).
Previous studies on the continental shelf of the Sabrina Coast provide evidence for maximum expansion of the ice sheet to within 2-5 km of the shelf break during past glaciations (Fernandez et al., 2018;Post et al., 2020), but uncertainty remains around the exact position of the maximum grounding line extent. During the last glacial period, as the ice sheet expanded close to the edge of the present-day continental shelf, the advancing ice remobilized sediments on the continental shelf and redeposited them on the continental slope . As the ice sheet retreated, rising sea-level and sea-surface temperatures in the Southern Ocean promoted iceberg calving and meltwater production (Weber et al., 2014). The supply of detritus to the continental slope during deglaciation was controlled by erosion at the base of the grounded ice sheet, and the release of englacial debris in meltwater plumes O'Brien et al., 2020).

The Subglacial Geology of Central Wilkes Land
Throughout successive periods of supercontinent assembly (e.g., Nuna, Rodinia, and Gondwana), much of the East Antarctic and southern Australian lithosphere accreted and consolidated forming a contiguous Australo-Antarctic component (e.g., Aitken et al., 2016a;Boger, 2011;Fitzsimons, 2003). This ancient tectonic relationship has cemented strong geological and structural affinities between the two now conjugate margins (Williams et al., 2019). Hampering our understanding, the basement rock geology is concealed under sedimentary basins and regolith in southern Australia and thick ice in East Antarctica. Existing interpretations in Wilkes Land rely on analysis of rare coastal rock outcrop (e.g., Maritati et al., 2019;Morrissey et al., 2017;Mulder et al., 2019;Tucker et al., 2017) combined with borehole analyses from southern Australia (e.g., Kirkland et al., 2017;Wingate et al., 2015a) and airborne geophysical data (Aitken et al., 2014). From west to east, three geologically distinct Mesoproterozoic provinces have been identified using these methods: the Albany-Fraser Orogen (c. 1,300-1,140 Ma), Madura Province (c. 1,475-1,390 Ma), and Coompana Province (c. 1,610-1,490 Ma); their Antarctic equivalents defined as the Wilkes, Nuyina, and Banzare provinces, respectively (e.g., Aitken et al., 2014;Fitzsimons, 2003;Maritati et al., 2019). The geological relationships between southern Australia and East Antarctica are summarized below ( Figure 2) to provide context for our provenance analysis. A more comprehensive regional geological history can be found in the references herein.
The Albany-Fraser Orogen ( Figure 2) is characterized by reworked Archean and Proterozoic crust . Exposed metasedimentary rocks of the Windmill Islands ( Figure 1) within the Wilkes Province in East Antarctica compare to those of the Mount Ragged Formation (Clark et al., 2000) in the Albany-Fraser Orogen, on the basis of shared detrital U-Pb age signatures (Fitzsimons, 2003;Morrissey et al., 2017). Granite and charnockite intrusions in the Windmill Islands in East Antarctica have crystallization ages between c. 1,325-1,170 Ma, coeval with the Recherche (c. 1,300-1,290 Ma, Nelson et al., 1995) and Esperance (c. 1,200-1,140 Ma, Nelson et al., 1995) supersuites of the Nornalup Zone, the easternmost zone of the Albany-Fraser Orogen (Morrissey et al., 2017).

Sampling Locations and Descriptions
We report results from a marine sediment core recovered from the continental slope adjacent to Sabrina Coast of central Wilkes Land, East Antarctica, and nearby basement rocks ( Figure 1 and Table S1). Core IN2017_V01-A042-KC14 (hereafter referred to as KC14) was recovered in 2017 employing a 15-cm 2 Kasten corer from aboard the RV Investigator (Armand et al., 2018). KC14 was targeted for this study due to its proximity to Totten Glacier and its continuous sequence of sediment deposition. The core is 3.38-m long and was recovered from the eastern flank of a large contourite ridge 2,100 m below sea level. The core sedimentology consists predominantly of glacigenic clay and silty clay overlain by biosiliceous silty clay in the Holocene (Armand et al., 2018). To complement the offshore samples, the nearest onshore basement rocks from Chick Island (Maritati et al., 2019) and Balaena Islets (Cameron et al., 1958) were sampled for feldspar Pb isotope analysis.

Radiocarbon Dating
The bulk acid-insoluble organic carbon fraction of six samples from KC14 was radiocarbon-dated. Prior to dating, all samples (weighing ∼5 g) were dried in an oven at 60°C, ground and submerged in 2M HCl for several hours on a hotblock at 80°C for carbonate removal. Each sample was subsequently rinsed with Milli-Q water to remove any residual acid and dried at 60°C. Radiocarbon analysis by Accelerator Mass Spectrometry was performed by DirectAMS. The raw age data were corrected for the Marine Reservoir Effect and calibrated using the Marine13 calibration curve in MatCal 2.41 (Lougheed & Obrochta, 2016). A uniform marine reservoir age of 917.5 ± 30 year was applied as this is the median of the range of reservoir ages currently defined for the Southern Ocean (Paterne et al., 2019). Median calibrated ages are reported with a 95% confidence limit.

Quantification of Iceberg-Rafted Debris Flux
Iceberg-rafted debris (IBRD) flux was calculated at 10 cm intervals along the core following the protocol of Krissek (1995) and Patterson et al. (2014): The coarse sand fraction (CSF; 250 μm-2 mm) was isolated by wet-sieving a known weight of each sample. Authigenic grains and microfossils were removed from the samples under a microscope. The remaining material was weighed on a microbalance and calculated as a fraction of the bulk sample. Cuvettes with a precise volume of 5.49 cm 3 were used to calculate the dry bulk density (DBD). The linear sedimentation rate (LSR) was calculated at intervals between the calibrated bulk organic carbon ages.

Biogenic Silica Concentrations
Biogenic silica concentrations were measured at 10-cm intervals along KC14 following a protocol based on DeMaster (1981) and Mortlock and Froelich (1989). Samples (30 ± 3 mg) were leached in 30-mL 1M NaCaO 3 solution and maintained at 85°C on a hotblock. After 1 h, a 1-mL aliquot was centrifuged at 10,000 rpm for 45 s to separate the sediment from the leached biogenic silica. A 200-μL aliquot of the solution was TOOZE ET AL. 6 of 23 10.1029/2020GC009156 then diluted in 9 mL of Milli-Q water and the dissolved silica was analyzed immediately using a Quickchem 8500 Series 2 FIA (Method 31-114-27-1-A, Diamond, 2002). This process was repeated for a total of 5 h. The biogenic silica in all samples was fully leached after 3 h. The percent standard deviation of two replicate analyses of in-house sediment standards, over a 2-day period, was 15% and 7%, respectively, for an Antarctic margin (Collaborative East Antarctic Marine Census) standard (15.2% biogenic silica) and a Chile margin sediment standard (6.6% biogenic silica). The biogenic silica flux was calculated in the same way as the IBRD flux, replacing the CSF with the biogenic silica mass fraction from each sample.

Sample Preparation
The fine (<63 μm), intermediate (63-250 μm), and coarse (>250 μm-2 mm) fractions were acquired from 2-cm-thick sediment samples (weighing ∼20 g) at 10-cm intervals along KC14 by wet-sieving. For subsequent analysis and interpretation, all the sampled 10-cm intervals were used from the intermediate size fraction, supplemented by four samples at ∼1-m intervals from each of the coarse and fine fractions. In the fine fraction, detrital grains were separated from clay minerals and subsequently mounted in resin. In the intermediate fraction, the heavy and light mineral fractions were manually separated by suspending the grains in Milli-Q water and rotating them around a watchglass. Nonmagnetic and weakly magnetic grains were then separated from the dried heavy mineral fraction using a Fe-B-Nd hand magnet and mounted in epoxy resin. For the coarse fraction, the CSF samples used to calculate the IBRD flux were mounted in epoxy resin. The one-inch-round resin discs were then sanded and polished to expose the interior of the grains. Basement rock samples were similarly prepared with rock chips from each sample mounted in resin before being dried and polished.

Sample Imaging
Sample mounts were imaged and analyzed on a FEI MLA650 environmental scanning electron microscope at the Central Science Laboratory, University of Tasmania (UTAS). Using Mineral Liberation Analysis (MLA) software, all mineral types in each mount were identified and color-coded using a predefined mineral library. In this way, detrital zircon, apatite, titanite, K-feldspar, and plagioclase feldspar from KC14, and K-feldspar and plagioclase feldspar from the four basement samples, were located in preparation for Laser Ablation-Inductively Coupled Plasma Mass Spectrometry (LA-ICPMS). Where possible, high-resolution backscatter and cathodoluminescence (CL) images of zircons from the intermediate (63-250 μm) fraction were acquired, capturing the external morphology, internal structure and zoning of the grains, and identifying optimal sites for LA-ICPMS analysis. Elongation ratios (length:width) for imaged zircons were quantified using the MLA software. The detrital titanite, apatite, and feldspar grain morphologies were examined under a microscope.

LA-ICPMS Analysis
LA-ICPMS measurements were made using an ASI Resolution S-155 System with a large format cell and a Coherent CompexPro 110 Excimer laser coupled with an Agilent 7900 ICPMS at the Discipline of Earth Sciences/Centre of Ore Deposit and Earth Sciences (CODES), UTAS. Specific instrumental parameters and standard reference materials used are listed in Table S2. Detrital zircon, apatite, and titanite were targeted from all size fractions for U-Pb analysis. Detrital plagioclase and K-feldspars were selected from the CSF only for Pb-Pb analysis.
The experimental setup for U-Pb and trace element analysis of zircon, apatite, and titanite follows Thompson et al. (2018). A glass primary standard, NIST610, was used to calculate a correction factor for the 207Pb/206Pb ratio using values proposed by Baker et al. (2004). Primary standard 91500 (Horstwood et al., 2016) was used for the calibration of all U-Pb ratios in zircon, with Plešovice (Sláma et al., 2008) and Temora (Black et al., 2003) used as secondary reference materials and treated as unknowns. During the analysis of apatite, OD306  was used as a primary standard and Otter Lake (Barfod et al., 2005), Emerald Lake (Coulson et al., 2002), 401 apatite , McClure Mountain (Schoene & Bowring, 2006), and Durango (McDowell et al., 2005) as secondary reference materials. An in-house primary standard, 19686-titanite, was used for titanite analysis with Mt. Dromedary (Miller et al., 1985)  Initial U-Pb-trace element data reduction follows Thompson et al. (2018) and Halpin et al. (2014) using Microsoft Excel macro spreadsheets created at the UTAS. Following these data reduction steps, we filtered the zircon data set to exclude: (1) processed signal intervals spanning less than 50% of the overall analysis time, (2) high 204 Pb values exceeding 100 cps, (3) concordance outside the range 80%-110%, and (4) grains displaying high Light Rare Earth Element (LREE) and P-contents (n = 15) symptomatic of mixed ablation of apatite or monazite microinclusions. A total of 139 analyses (44% of the data set) were eliminated using these criteria and were excluded from interpretation. All rare earth element (REE) data were chondrite-normalized using values proposed by Sun and McDonough (1989). A common-Pb correction procedure was applied for the detrital apatite and titanite U-Pb-trace element data sets, which were discordant due to common Pb. We followed the five-stage iterative 207Pb-correction method of Chew et al. (2011), with estimates of initial Pb-isotopic compositions from the Stacey and Kramers (1975) crustal Pb evolution model. Some analyses recorded significant differences in the calculated 207 Pb-corrected 206 Pb/ 238 U ages after the fifth iteration. To account for this, the uncertainty for common Pb on all analyses was increased using the absolute difference in the common Pb predicted by the fourth and fifth iterations. Subsequent to this common Pb correction procedure, we further filtered the apatite and titanite data set to exclude processed signal intervals spanning less than 50% of the total analysis time, which eliminated 22% and 10% of the apatite and titanite data sets, respectively. Probability-density plots were produced using DensityPlotter version 8.4 (Vermeesch, 2018 Figure S1) were constructed using IsoplotR (Vermeesch, 2018).
A comprehensive description of the method used for Pb-Pb analysis of detrital feldspar using the UTAS LA-ICPMS system is given by Mulder et al. (2019). Lines, rather than spots, were used for feldspar ablation to maintain the signal intensity and smooth, unaltered areas of individual grains were targeted for analysis. Data reduction also followed Mulder et al. (2019); only one analysis with a signal less than 50% of the total analysis time was excluded from interpretation. No correction for the in-growth of radiogenic Pb was applied due to the avoidance of U and Th inclusions. NIST612 was ablated for the calibration of all Pb-Pb ratios, using values proposed by Baker et al. (2004) and glass reference materials GSD-1g (Jochum et al., 2011) and BCR-2g (Jochum et al., 2016) were used as secondary standards and treated as unknowns. No feldspar reference standard was available during this study, however, the UTAS facility uses glass reference materials for Pb-isotopic compositions on a wide range of minerals without a measurable "matrix effect" on the quadrupole ICPMS system (e.g., monazite: Berry et al., 2016;pyrite: Hawke et al., 2015;zircon: Thompson et al., 2018).
Weighted average values for all LA-ICPMS standards were calculated using the Microsoft Excel add-in Isoplot v. 4.15 (Ludwig, 2012), all of which fall within 2σ error of the published reference values (Table S3).

Bulk Organic Carbon Ages, IBRD Flux, and Biogenic Silica Concentrations
The median calibrated calendar ages for KC14 capture a Quaternary record spanning c. 23,500 cal years BP. Based on the bulk radiocarbon ages and low biogenic silica concentrations (representing low productivity characteristic of glacial conditions), the base of the core to 170 cm represents the last glacial period (interval 4, Figure 3). At 170 cm, biogenic silica concentrations start to increase, indicating increased productivity, which we infer represents the deglaciation and a transition into the Holocene (interval 3, Figure 3). Estimated LSRs were highest during the glacial (377 cm/Kyr) and moderate in the Holocene (10-11 cm/Kyr). A sharp decline in the LSR to 8 cm/Kyr followed by an increase to 14 cm/Kyr marks the deglacial transition between the last glacial period and the Holocene (Figure 3). The CSF ranges between 0.0001 and 0.0147, with an average of 0.003 in the glacial and 0.001 in the Holocene (Figure 3). The biogenic silica flux averages 7.6 g/cm 2 /Kyr in the glacial and 2.0 g/cm 2 /Kyr in the Holocene and positively correlates with the IBRD flux. The IBRD flux ranges between 0.001 and 3.4 g/cm 2 /Kyr with an average of 0.4 g/cm 2 / Kyr during the glacial period and 0.004 g/cm 2 /Kyr during the Holocene (Figure 3).

Detrital Zircon, Apatite, and Titanite Geochronology
Compiled U-Pb ages from all depth intervals of KC14 reveal dominant age populations between c. 1,360-1,100 Ma for detrital zircon and between c. 1,200-1,100 Ma for detrital titanite and apatite (Figure 5a). The detrital zircon age distribution exhibits strong multimodality; the dominant c. 1,360-1,100 Ma peaks representing 51% of the data set, with secondary peaks between c. 1,620-1,490 Ma (17%). A very small number of zircon analyses exhibit ages older than 1,620 Ma (8%) and between c. 700-500 Ma (5%). In contrast, the detrital titanite and apatite age distributions are largely unimodal, with most ages predominantly falling within the main c. 1,200-1,100 Ma population (66% and 86% for titanite and apatite respectively; Figure 5a). TOOZE ET AL. 11 of 23 10.1029/2020GC009156 Figure 5. Detrital age signatures in KC14. (a) Age spectra for detrital zircon, titanite, and apatite for the entire KC14 core presented as probability-density plots (black curve, gray fill). Kernel density curve is shown in red. (b) Probabilitydensity plots showing downcore age variability in detrital zircon, titanite, and apatite. Gray rectangles within the stylized core on left hand side show where sediment samples were taken from within each depth interval of KC14 for U-Pb analysis. Major age peaks and total number of analyses (n) are indicated. Age signatures of geological provinces in East Antarctica/southern Australia (as in Figure 2) are highlighted as colored bars for comparison. Although the Moodini Supersuite intrudes the Banzare/Coompana and Nuyina/Madura provinces, it is here grouped with the Wilkes Province/Albany-Fraser Orogen due to the overlap in age ranges. Pie charts show relative proportion of each mineral size fraction contributing to the age signature for each depth interval. Intervals 1-4 represent changes in biogenic silica concentration and are the same as those presented in Figure 3.
A minor portion of detrital titanite and apatite analyses are older than c. 1,300 Ma and a very small population between c. 750 and 500 Ma is present in the detrital apatite record (Figure 5a).
Although the main c. 1,360-1,100 Ma age population is prevalent throughout the detrital zircon record, subtle changes in the relative proportions of age populations are observed with depth. Detrital zircon ages between c. 1,620 and 1,490 Ma are more prominent at the top and bottom of the core (intervals 1 and 4, Figure 5b). Major peaks within this population occur at c. 1,620 Ma and c. 1,540 Ma within interval 1 and at c. 1,580 Ma and c. 1,495 Ma within interval 4. In intervals 2 and 3, where biogenic silica is most abundant, detrital zircon grains within the c. 1,620-1,490 Ma age range are significantly reduced, and a minor c. 1,450-1,390 Ma population can be seen (Figure 5b) that is otherwise not discernible in the whole core record (Figure 5a). This population is also present as a small peak in interval 1 (at c. 1,400 Ma) but is completely absent from interval 4. The youngest age population (c. 750-500 Ma) exists in all depth intervals for detrital zircon and most of the apatite record (intervals 2-4, Figure 5b), but only exists within interval 2 for titanite ( Figure 5b). These rare c. 750-500 Ma detrital zircon (n = 9), apatite (n = 10), and titanite (n = 1) grains were all obtained from the fine fraction. The detrital titanite and apatite age distributions are largely consistent throughout the transition from the glacial period through to the Holocene, with the addition of very minor peaks at c. 1,470 Ma and c. 1,450 Ma, respectively, in interval 1 of the core (Figure 5b).

Detrital Zircon, Apatite, and Titanite Trace Element Geochemistry
The detrital zircon grains in KC14 yield chondrite-normalized REE abundances between 10 −4 and 10 4 with steep heavy rare earth element (HREE) enrichment, a positive Ce anomaly and a negative Eu anomaly (Figure 6a), typical of magmatic zircon (e.g., Rubatto, 2017). A minor portion of detrital titanite grains reveal a positive Eu anomaly and HREE enrichment, associated with a metamorphic source rock type ( Figure S2b). The detrital apatite and zircon geochemical data reveal a dominantly granitoid source rock type using discrimination plots from Belousova et al. (2002aBelousova et al. ( , 2002b, respectively; Figures 6b,, and S2c). Most of the detrital titanite grains have high Fe concentrations ranging between 7,500 and 20,000 ppm, characteristic of igneous titanite (e.g., Olierook et al., 2019; Figure S2d).

Feldspar Pb-Isotopic Compositions
Feldspar analyzed from basement rocks from the Sabrina and Budd coasts (Figure 1) Figures 7a and 7b). Analyzed grains from both locations broadly overlap with Pb-isotopic signatures from other Australo-Antarctic domains (Mulder et al., 2019; Figure 7).
The Pb-isotopic compositions of feldspars in the offshore record are more diverse. Detrital K-feldspar and plagioclase from the coarse fraction in KC14 have 206 Pb/ 204 Pb of 15.9-17.5 and 207 Pb/ 204 Pb of 15.1-15.7 (Figure 7). These values predominantly span the mantle to average crust growth curves, with ∼20% of the detrital feldspar compositions from KC14 overlapping the average-to-upper crustal basement rock compositions analyzed from the Balaena Islets (and Windmill Islands, defining the broader Wilkes Province) and Chick Island (near the boundary between the Nuyina and Banzare provinces; Figure 7).

Discussion
The rapidly retreating, fast-flowing Totten Glacier (Rignot et al., 2011;Figure 1b) is the closest major outlet glacier to the KC14 core site and is an obvious candidate for detrital delivery from the Totten glacial drainage basin. Indeed, the Totten Glacier is interpreted to have discharged sediment-laden meltwater and plumes throughout the Late Cenozoic, from the early development of the East Antarctic Ice Sheet to the present day , and is suggested to have contributed to the contourite ridge where KC14 was recovered .
The most recent interpretation of the subglacial geology in central Wilkes Land from Maritati et al. (2019) is based on matching geophysical signatures across the margins of southern Australia and East Antarctica and U-Pb analyses of primary and inherited zircon from rare plutonic outcrop at Chick Island. This tectonic framework suggests that the Wilkes, Nuyina, and Banzare provinces underlie the Totten catchment area, with the Totten Glacier exploiting a suture between the Wilkes and Nuyina provinces in the coastal region of the catchment, and the Banzare Province situated in the hinterland.
Based on these recent interpretations, it is anticipated that detritus on the continental slope and specifically sampled at KC14 should have an age and composition that reflects the Totten catchment basement terranes. In this paradigm, given that erosion is largely focused along the ice sheet margins at the coast (Golledge  (Maritati et al., 2019). Typical REE compositions for magmatic and granulitic (i.e., high-temperature metamorphic) zircon from Rubatto (2017) are shown. (b, c) KC14 detrital zircon trace element data overlain on source rock discrimination diagrams from Belousova et al. (2002bBelousova et al. ( ). et al., 2013Jamieson et al., 2010), the source regions for detritus sampled at KC14 should be concentrated in the Wilkes and Nuyina provinces. We now explore this hypothesis in Section 5.1.

Was the Totten Glacier the Main Conduit for KC14 Detritus?
Previous work has identified that specific regions along the East Antarctic margin are characterized by distinct detrital zircon age signatures (Pierce et al, 2014), consistent with the known onshore geology (Veevers, 2012;Veevers & Saeed, 2011, and references therein Growth curves for the mantle (μ = 9), average crust (μ = 9.74), and upper crust (μ = 12) are plotted using the Pb isotope evolution model of Stacey and Kramers (1975 broadly corresponding to those allocated by Pierce et al. (2014; Figure 8). Rocks of the Prydz Bay sector have peak zircon age populations at c. 990-900 Ma and c. 650-500 Ma (Veevers, 2012;Veevers & Saeed, 2011, and references therein). The Eastern Sector is dominated by Cambrian granites (c. 600-400 Ma) that intruded during the Ross Orogeny, with secondary older (>1,610 Ma) age peaks representative of Archean-Paleoproterozoic rocks of Adélie Land (Veevers, 2012;Veevers & Saeed, 2011, and references therein). The Wilkes Sector has a signature dominated by rocks that are predominantly Late Mesoproterozoic in age (c. 1,200-1,000 Ma). The detritus from KC14 has an age signature that is clearly distinct from the Prydz Bay and Eastern sectors (inset in Figure 8).  (Orsi et al., 1995). Black arrows represent the westward-flowing Antarctic Coastal Current. Modeled basal velocity contours for an ice sheet slightly advanced relative to the present-day grounding line are from Aitken et al. (2016b) for the region between 90˚E and 180˚E.
KC14 matches well with other detrital sediment records from the Wilkes Sector further to the east (Pierce et al., 2014; Figure 8), but reveals additional prominent c. 1,360-1,220 Ma age peaks that are diagnostic of the Wilkes Province (Figures 5a and 8).
A derivation from the Totten Glacier catchment predicts mainly Wilkes and Nuyina-age signatures for KC14 detritus. The abundant c. 1,200-1,000 Ma zircon, apatite and titanite detritus in KC14 ( Figure 5) could potentially be derived from voluminous plutonic rocks in any or all three of the proximal Wilkes, Nuyina, and Banzare provinces (Figure 2). Therefore, this age group is not particularly useful for identifying unique There are a number of scenarios that could explain the scarcity of Nuyina-age detritus (which is a signature expected if the Totten Glacier was the main conduit) in addition to the abundance of Banzare-age detritus at the KC14 core site. One scenario is that some erosion was focused in the hinterland of the Totten catchment, encapsulating Banzare crust. However, much of this region is covered by kilometers of Sabrina Basin sedimentary strata (Aitken et al., 2016b) so we do not expect crystalline bedrock exposure. Furthermore, the high angularity and broken morphologies of most detrital grains from all size fractions in KC14 (Figure 4) supports a relatively local provenance and argues against recycling through preglacial sedimentary systems. This is further supported by the high abundance of detrital feldspar grains which are typically unable to withstand prolonged weathering and erosion (Johnson et al., 2018).
A second scenario could be envisaged whereby the Nuyina Province may be less extensive than interpreted by Maritati et al. (2019), such that basal erosion in the coastal region of the Totten catchment sources Banzare crust in addition to Wilkes. This interpretation could account for the strong Banzare-age signature and weak Nuyina-age signature in the detrital zircon record (Figure 5a). Testing of this hypothesis is difficult without higher-resolution geophysical data linking the conjugate margins of southern Australia and East Antarctica and/or geological sampling (e.g., dredging of basement rock ridges or sampling of proximal marine sediment cores).
A third scenario that could explain the scarcity of Nuyina-age detritus is an underrepresentation of mafic rock types in the detrital zircon record (Licht & Hemming, 2017;Shao et al., 2019). The Nuyina Province likely comprises juvenile mafic rocks similar to the Arubiddy Ophiolite Complex of the Madura Province ( Figure 2b). Although detailed analysis of zircon grains revealed well-developed oscillatory growth zoning ( Figure 4) and trace element signatures typical of a felsic igneous source (Figures 6 and S2), this could be a biased record related to the high abundance of zircon in felsic rocks compared to the typically small amounts of zircon contained in mafic rocks (Licht & Hemming, 2017;Shao et al., 2019). Another bias in favor of felsic rock compositions may also be introduced by the different weathering rates of mafic and felsic rock types. Mafic rocks have a lower resistance to weathering, leading to limited preservation of mafic rock components over time (Licht & Hemming, 2017). Derivation of KC14 detritus from at least some mafic-intermediate crust is supported by the presence of minor amphibole, pyroxene, and olivine grains. Evidence for mafic source components is also provided by the spread in detrital feldspar Pb-isotopic compositions (plagioclase in particular) to mantle-like values (Figure 7). This signature suggests the presence of unexposed juvenile rocks in the hinterland, likely within the Nuyina and Banzare provinces, related to reworking of oceanic crust during a series of accretionary tectonic events Maritati et al., 2019;Spaggiari et al., 2018).
Although the underrepresentation of mafic source rocks in the detrital zircon record potentially explains the paucity of Nuyina-age detritus in KC14, this scenario alone does not explain the abundance of Banzare-age detritus. A fourth scenario, and one that we suggest is most plausible, is that the Totten Glacier was one of a number of outlet glaciers that contributed detritus to KC14, as described in Section 5.2 below.

A Case for a Distributed Wilkes Land Provenance for KC14
Two main age distribution patterns are observed in the detrital zircon record. In intervals 1 and 4 of KC14 (broadly equivalent to the Holocene and last glacial period, respectively), detrital zircon age peaks associated with the Banzare Province (c. 1,620-1,490 Ma) are prevalent alongside the c. 1,360-1,100 Ma Wilkes signature ( Figure 5b). In the middle biosiliceous portion of the core (intervals 2 and 3, Figure 5b), Banzare-age peaks are reduced and Nuyina-age peaks (c. 1,450-1,390 Ma) are more prominent. We suggest these subtle changes in the detrital zircon record over time support the supply of detritus from multiple source regions. Below, we relate downcore changes in sediment provenance to temporal changes in the climate and ice sheet configuration.

A Broad Source Region During Interglacials
The most recently deposited sediment is captured by KC14 interval 1 (Figure 5b), which should therefore yield an age signature that is most representative of the subglacial rock types being eroded at the base of the ice sheet in the present day. The Banzare-age detrital zircon signature in this interval suggests that some of this more recently deposited detritus was sourced from where the Banzare Province is exposed to the base of the ice sheet at or close to the coast. The interpreted extension of the Banzare Province beneath multiple subglacial catchments east of the Totten catchment ( Golledge et al. (2013). The modeled maximum extent of the ice sheet is indicated with a blue dashed line (Golledge et al., 2013). The position of a grounding zone wedge (GSW) identified by Post et al. (2020) is shown (white star), as new evidence indicating the former maximum extent of the ice sheet at that location.
butions from the Moscow University, Thompson, and Holmes glaciers (Figure 9). This idea is supported by recent interpretations of the continental slope morphology that suggest that the contourite ridge from which KC14 was recovered is fed not only by detritus from the Totten Glacier, but also by fine sediment plumes transported westward across the slope by the Antarctic Coastal Current . With an ice sheet slightly advanced relative to modern grounding line, the ice sheet model of Aitken et al. (2016b) predicts high basal erosion (using basal velocity as a proxy) across a broad coastal area of central Wilkes Land (Figure 9a). The dominance of Wilkes and Banzare-age peaks in interval 1 of our marine core record supports the predictions of this ice sheet model which shows high basal erosion focused in zones mostly underlain by the Wilkes and Banzare provinces (Figures 5b and 9a).

Debris Flows During the Last Glacial Period
Curiously, the detrital zircon signature during the glacial (interval 4) resembles that of the current interglacial (interval 1) described above (Figure 5b). We interpret the detrital age signal in the glacial section of KC14 (interval 4; Figures 3 and 5) to represent reworked sediment remobilized off the shelf and deposited downslope during glacial advance, a process recently inferred in this region Post et al., 2020) and in previous studies around the Antarctic margin (e.g., Diekmann et al., 2000;Dowdeswell et al., 2006Dowdeswell et al., , 2004Gales et al., 2013;Pudsey, 2000). This interpretation is supported by the high glacial sedimentation rate (377 cm/Kyr; interval 4, Figure 3), at a time when basal erosion by the ice sheet was minimum (Golledge et al., 2013;Figure 9b), suggesting that reworking of shelf sediments by debris flows was the most likely process contributing to the high sediment accumulation Post et al., 2020). The relatively small concentrations of biogenic silica measured in the glacial section of the core (interval 4%; 1.9%-7.9%) likely comprise reworked diatoms from earlier warm periods that were redeposited with the sediment from the continental shelf, explaining the high biogenic silica flux during this same glacial period (interval 4, Figure 3). Coarse debris (250 μm-2 mm) that matches the criteria of IBRD was likely mobilized by downslope sediment transport. This process explains the higher calculated "IBRD" flux during the glacial (Figure 3), when iceberg calving would have been infrequent and the quantity of IBRD is expected to be low Wilson et al., 2018). A mixture of this coarse debris and fallout of finer sediment from suspension in sediment-laden plumes likely contributed to the maintenance of the large contourite ridge where KC14 is located Post et al., 2020). The similarity in age distribution between intervals 1 and 4 suggests that the sediment reworked in debris flows as the ice sheet advanced was initially deposited during past warmer climate periods, when the position of the retreated ice sheet margin and the provenance of continentally derived detritus was similar to that of the present day.

A Proximal Provenance During Deglaciation
A prominent decline in the sedimentation rate from 377 cm/Kyr to 8 cm/Kyr at c. 23.2 cal years BP ( Figure  3) indicates the ceasing of abundant sediment supply in debris flows to the slope. We interpret this decrease in sedimentation as the point when the ice sheet reached its maximum extent, and the associated reworking of shelf sediments ceased. The onset of deglaciation (c. 16.9-14.2 cal years BP) is marked by a rise in biogenic silica concentrations representing a growth in primary productivity and warmer, ice-free conditions above the core site (intervals 2 and 3, Figure 3). During this time, the sedimentation rate increased to 14 cm/Kyr and IBRD became more abundant, reflecting a change in the mode of sediment transport to the continental slope. While sedimentation during the glacial period was dominated by the supply of reworked sediment in mass flows, deglacial sedimentation was likely controlled by the release of englacial IBRD in sediment-laden meltwater fluxes associated with climate warming and ice sheet retreat, as interpreted in other studies in East Antarctica (e.g., Bertram et al., 2018;Wilson et al., 2018).
The deglacial sediment provenance signature is represented by detrital ages from biosiliceous sediments in intervals 2 and 3 of the core (Figures 3 and 5b). In comparison to intervals 1 and 4, the more prominent Wilkes and Nuyina-age signatures, coupled with a significantly reduced Banzare-age signal in these deglacial intervals, suggests that detritus was sourced from the coastal region of the Totten catchment area. At the onset of deglaciation, with an ice sheet extended further offshore of its present-day posi-tion, the terminus of the Totten Glacier was situated in closer proximity to the KC14 core site and would likely be the principal outlet for detritus during this time. Recently, Prothro et al. (2018) interpreted that most englacial debris that melted out of the base of ice shelves in the Ross Sea region during retreat was deposited proximal to the grounding line. We interpret this to be the case in central Wilkes Land. With a narrower "cone of influence," the abundant detritus being supplied by the Totten Glacier to the slope would dampen the signal produced from detritus from more distal sources, explaining the relatively enhanced Nuyina-age signature and reduced Banzare-age signature in the detrital zircon record (intervals 2 and 3, Figure 5b).

Evidence for Minor Far-Field Provenance
Minor detrital c. 750-500 Ma ages are present throughout the detrital zircon record, that are not known to be characteristic of basement rocks of central Wilkes Land nor the conjugate regions in southern Australia. The compiled age spectra from the offshore (Pierce et al., 2014) and onshore (Veevers, 2012;Veevers & Saeed, 2011) records of the Eastern Sector show a dominant age population between c. 740-400 Ma (Figure 8). Westward transport of detritus over long distances from one sector to another by icebergs driven by the Antarctic Coastal Current is supported by evidence from warm climates and transitions of the Early Pliocene (Cook et al., 2014;Williams et al., 2010). Icebergs calved off from the Eastern Sector have been observed tracking westward along the shelf to the Wilkes Sector (Budge & Long, 2018). We propose westward transport of a minor component of detritus from the Eastern Sector via icebergs driven by the Antarctic Coastal Current to the KC14 core site. This hypothesis would account for the rare, but distinct c. 750-500 Ma detrital age population characteristic of the basement rock and intruded granites of George V Land (e.g., Lamarque et al., 2018). Provenance from the east could also account for minor older c. 2,200-1,700 Ma detrital zircon ages that correlate with Archean and Paleoproterozoic rocks of the Adélie Land coast (e.g., Ménot et al., 2007; Figure 8).

Conclusions
This investigation presents the first sediment provenance record from the continental slope of central Wilkes Land, defining the sedimentary environment, transport, and source of detritus to this part of the Wilkes Sector from c. 23.5 ka to present. Detailed multiproxy analyses allow the following conclusions to be drawn: 1. The main Mesoproterozoic (c. 1,360-1,000 Ma and c. 1,620-1,490 Ma) age populations in the KC14 detritus indicate sourcing from dominantly felsic granitoid rocks of the proximal Wilkes and Banzare provinces, with smaller contributions from the adjacent Nuyina Province (c. 1,450-1,390 Ma) in central Wilkes Land. The unexpectedly small contribution from Nuyina Province, and relative importance of material from the Banzare Province, is hypothesized to reflect sourcing of detritus from a broad source region, including the Totten Glacier catchment and subglacial catchments further east. 2. Debris flows instigated by the advancing ice sheet remobilized sediments from the continental shelf and redeposited them on the slope during the last glacial period. These sediments, deposited prior to glacial advance, have the same provenance signature as detritus deposited during the current interglacial. 3. The deglacial signature is dominated by detritus from the Wilkes and Nuyina provinces that underlie Totten Glacier, suggesting dominant sourcing of detritus from the coastal region of the Totten catchment area when the ice sheet margin was extended further offshore of its present-day position. This is likely due to the proximity of Totten Glacier to the core site and the dampening of detrital age signatures from more distal source regions during this time 4. Minor c. 750-500 Ma and 2,200-1,700 Ma detrital zircon in KC14 was likely derived from rocks in the Eastern Sector and transported to the core site via icebergs driven by the Antarctic Coastal Current This new sediment provenance record provides the first offshore evidence from the Sabrina Coast to support existing subglacial bedrock interpretations from available geophysical and rare outcrop geological data in central Wilkes Land. These findings can be used to support plate tectonic reconstructions and to help infer the properties of the Antarctic crust that are relevant to basal ice sheet conditions and associated ice sheet configuration. of Tasmania) and Nils Jansen (Institute for Marine and Antarctic Studies, University of Tasmania) for their expertise and technical assistance and A. Maritati for Pb-Pb analysis of the basement rock feldspars used in this study. We also thank the reviewers and editors for their constructive comments and suggestions on previous versions of this article. Sian Tooze was supported by a PhD scholarship from the Australian Research Council's Special Research Initiative for the Antarctic Gateway Partnership (Project ID SR140300001). Taryn L. Noble acknowledges support from the Science and Industry Endowment Fund, and Australian Antarctic Science Program Grant 4419. This research used samples (from Balaena Islets) provided by the Polar Rock Repository (PRR). The PRR is sponsored by the National Science Foundation Office of Polar Programs. All data to support our conclusions can be found in this article, in the supporting information and archived with Pangaea database.