Growth, dynamics and deglaciation of the last British–Irish ice sheet: the deep-sea ice-rafted detritus record

The evolution and dynamics of the last British–Irish Ice Sheet (BIIS) have hitherto largely been reconstructed from onshore and shallow marine glacial geological and geomorphological data. This reconstruction has been problematic because these sequences and data are spatially and temporally incomplete and fragmentary. In order to enhance BIIS reconstruction, we present a compilation of new and previously published ice-rafted detritus (IRD) flux and concentration data from high-resolution sediment cores recovered from the NE Atlantic deep-sea continental slope adjacent to the last BIIS. These cores are situated adjacent to the full latitudinal extent of the last BIIS and cover Marine Isotope Stages (MIS) 2 and 3. Age models are based on radiocarbon dating and graphical tuning of abundances of the polar planktonic foraminifera Neogloboquadrina pachyderma sinistral (% Nps) to the Greenland GISP2 ice core record. Multiple IRD fingerprinting techniques indicate that, at the selected locations, most IRD are sourced from adjacent BIIS ice streams except in the centre of Heinrich (H) layers in which IRD shows a prominent Laurentide Ice Sheet provenance. IRD flux data are interpreted with reference to a conceptual model explaining the relations between flux, North Atlantic hydrography and ice dynamics. Both positive and rapid negative mass balance can cause increases, and prominent peaks, in IRD flux. First-order interpretation of the IRD record indicates the timing of the presence of the BIIS with an actively calving marine margin. The records show a coherent latitudinal, but partly phased, signal during MIS 3 and 2. Published data indicate that the last BIIS initiated during the MIS 5/4 cooling transition; renewed growth just before H5 (46 ka) was succeeded by very strong millennial-scale variability apparently corresponding with Dansgaard–Oeschger (DO) cycles closely coupled to millennial-scale climate variability in the North Atlantic region involving latitudinal migration of the North Atlantic Polar Front. This indicates that the previously defined “precursor events” are not uniquely associated with H events but are part of the millennial-scale variability. Major growth of the ice sheet occurred after 29 ka with the Barra Ice Stream attaining a shelf-edge position and generating turbiditic flows on the Barra–Donegal Fan at ∼27 ka. The ice sheet reached its maximum extent at H2 (24 ka), earlier than interpreted in previous studies. Rapid retreat, initially characterised by peak IRD flux, during Greenland Interstadial 2 (23 ka) was followed by readvance between 22 and 16 ka. Readvance during H1 was only characterised by BIIS ice streams draining central dome(s) of the ice sheet, and was followed by rapid deglaciation and ice exhaustion. The evidence for a calving margin and IRD supply from the BIIS during Greenland Stadial 1 (Younger Dryas event) is equivocal. The timing of the initiation, maximum extent, deglacial and readvance phases of the BIIS interpreted from the IRD flux record is strongly supported by recent independent data from both the Irish Sea and North Sea sectors of the ice sheet.