Paired Pb isotopes and Rb-Sr geochronology of individual ice-rafted feldspar grains from seafloor surface samples from the Pacific margin of West Antarctica

This dataset comprises measurements of lead (Pb) isotopic composition, rubidium-strontium (Rb-Sr) geochronology, and elemental geochemistry carried out on individual ice-rafted potassium feldspar grains separated from 25 seafloor surface sediments from the continental shelf along the Pacific margin of West Antarctica. Seafloor surface sediment samples were collected from eight expeditions: Operation Deep Freeze 1962 and 1983; NBP00-01 (2000) and NBP20-02 (2020) with RVIB Nathaniel B. Palmer; and JR104 (2004), JR141 (2006), JR179 (2008), and JR284 (2013) with RRS James Clark Ross. Rb-Sr dates and element data were determined on a subset of the same grains analysed for Pb isotopes, resulting in a coupled dataset. Clustering was applied to Pb isotopic data, identifying five distinct provenance sources. The data provide an indication of the variation in the subglacial geology of West Antarctica, useful for tectonic reconstruction and sedimentary provenance work. NERC PhD studentship (NE/S007210/1); the THOR project, a component of the International Thwaites Glacier Collaboration (ITGC) with support from National Science Foundation (NSF: Grant 1738942) and NERC (grant nos. NE/S006664/1, and NE/S006672/1).

Seafloor surface sediment samples were collected by the British Antarctic Survey (BAS), US National Science Foundation (NSF) funded expeditions and the Thwaites Offshore Research (THOR) project using Box-, Gravity-, Vibro-, Piston-, Kasten-, or Multi-corers, as detailed in the sample metadata table. Cores collected by NSF-funded expeditions (including the THOR cores) were curated and stored at the Marine Geology Repository hosted at Florida State University, and later at Oregon State University. Those collected by BAS were curated and stored in the UK either at BAS or at the British Ocean Sediment Core Research Facility. The sediment samples were washed and dry-sieved to separate the 250-500 µm fraction, extended to 150-600 µm in sand-poor samples. Feldspar grains were then picked under a binocular microscope, mounted in resin, ground to expose their centres, and polished. Scanning Electron Microscopy with Energy Dispersive Spectroscopy (SEM-EDS) was used to screen the grains and K-feldspar phases were targeted. Pb isotope analysis used laser-ablation multi-collector inductively-coupled plasma mass spectrometry (LA-MC-ICP-MS). Sample-standard bracketing with NIST 610 was used for primary calibration, correcting for system mass bias and drift. Rb-Sr geochronological analysis used laser-ablation reaction-cell triple-quadrupole inductively-coupled plasma tandem mass spectrometry (LA-ICP-MS/MS), with SF6 + He (10:90) and H2 reaction gasses (Hogmalm et al., 2017). Bracketing analyses of NIST 610 were used for initial isotope ratio normalization and drift correction. Matrix-matched 87Rb/86Sr calibration used a sample of Dartmoor Granite (Darbyshire & Shepherd, 1985) from Haytor West Quarry, Dartmoor (Devon, UK). Detrital Rb-Sr ages were calculated in IsoplotR (Vermeesch, 2018) using two-point isochrons anchored to a fixed initial Sr ratio of 0.7068 ± 0.0018, calculated as the weighted mean of 67 87Sr/86Sr measurements in plagioclase grains from the same samples and with uncertainties fully propagated. Elemental data were measured alongside the Rb-Sr isotopes, and are calibrated to NIST 610 using Si as an internal standard (assuming 60 wt% for unknown feldspars). Pb isotopic data were grouped using a Gaussian Mixture Model (GMM) to determine clusters in 206Pb/207Pb-208Pb/207Pb space. Five distinct groups were detected (Groups A-E). Group D was further subdivided into geographical subgroups: D1 (in eastern Marie Byrd Land and the Amundsen Sea Embayment), D2 (in the Bellingshausen Sea), and D3 (in western Marie Byrd Land and the Ross Sea). The core of Group A (those grains with 206Pb/207Pb less than or equal to 1.171 and 208Pb/207Pb less than or equal to 2.43) was designated Group A1, and the remainder Group A2. A log-likelihood threshold containing 99.9% of the GMM was defined, and any data points with a log-likelihood value below the threshold were left unclassified (Group X). Pb isotopic data with 2SE uncertainties >0.003 (206Pb/207Pb) or >0.004 (208Pb/207Pb) were not included in the clustering, but are given in the dataset (3.4% of data).

Quality of Pb isotopic measurements was assessed by comparison of standard reference materials NIST 612 and USGS BCR-2G to known values. For all 204Pb-normalized ratios, NIST 612 precision was ~1.2% and accuracy was <0.25%. BCR 2G precision was ~2.7% and accuracy was ~0.01-0.02% Quality of Rb-Sr geochronology was assessed by dating a sample of the Mountsorrel Granodiorite (Leicestershire, UK) and comparing to the published age reported by Cribb (1975). Accuracy across both analytical sessions was -7% and -3%. Quality of elemental data was assessed using USGS BCR-2G: precision was ~10% and accuracy of all elements was less than or equal to 10%, except Pb (11%) and K (17%).

Elemental Scientific Lasers NWR 193 nm UV Excimer laser Teledyne Iridia 193 nm UV Excimer laser Thermo Scientific Neptune Plus MC-ICP-MS Agilent 8900 ICP-MS/MS Iolite version 4 IsoplotR (version 6.5.1; R version 4.4.1)