British–Irish Ice Sheet and polar front history of the Goban Spur, offshore southwest Ireland over the last 250 000 years

Deep Sea Drilling Program (DSDP) Site 548 was cored in 1984 at a water depth of 1256 m on the Goban Spur, offshore southwest Ireland. Coring retrieved a ~100‐m‐thick Pleistocene contourite sequence. This study uses planktonic foraminiferal assemblage and benthic foraminiferal oxygen isotope analyses to establish an age model for the upper 40 m of this core. This site's multidisciplinary analyses of planktonic foraminiferal assemblages, lithic grains, facies and calcium carbonate concentration reveal a 250 000‐year record of the North Atlantic polar front variability and British–Irish Ice Sheet (BIIS) history. The sequence is characterized by alternations of ice rafted debris (IRD) laden pelagic mud facies with calcium carbonate‐rich silty sand contourite facies that track glacial/interglacial cycles. The polar front migrated southward across the area several times during glacial maxima and stadial periods, while warmer Mediterranean Outflow Water (MOW) flowed northward across the region during interglacial and interstadial periods depositing contourites. Lithic analyses reveal a complex history of IRD deposition associated with iceberg calving from the Laurentide Ice Sheet and northwest European ice sheets, mainly the BIIS. Comparison between the Goban Spur (DSDP Site 548) and the Celtic Margin (MD03‐2692) and central North Atlantic Integrated Ocean Drilling Program (IODP) Site U1308 suggests differences between the ‘non‐Laurentide Ice Sheet’ Heinrich Events (HE) 6 and 3 at the Goban Spur, with IRD from the BIIS being prominent during HE 6 and IRD from other European ice sheets north of the BIIS likely being more dominant during HE 3. The nature of lithics in IRD‐rich horizons during Terminations 3, 3A, 2 and 1 suggests significant iceberg calving episodes preceding BIIS retreat during the onset of interstadial intervals.

Rapid late Cenozoic climatic variations markedly affected global oceanography and ice-sheet behaviour in the North Atlantic Ocean and are well recorded in marine sedimentary strata (Ruddiman et al. 1989) and ice-cores (Grootes et al. 1993).In particular, the area around Greenland and northeastern Canada hosts the North Atlantic polar front (Fig. 1), a physical oceanographic feature separating two surface water-masses: cold, low-salinity modified Polar Water and warmer high-salinity Atlantic Water (Eynaud et al. 2007;Holliday et al. 2008).In the past, the polar front migrated southwards as far as the Iberian Margin (Eynaud et al. 2009; Fig. 1).The dynamics of the polar front in the northeast Atlantic Ocean during periods of rapid climatic variability have been the subject of several studies (Eynaud et al. 2009;Bashirova et al. 2014;Marchal et al. 2016).A few studies have documented the polar front history on the Celtic Margin (Scourse et al. 2000;Mojtahid et al. 2005).However, these studies do not document its history beyond the last glacial cycle.
One result of the climate oscillations in the North Atlantic is the waxing, waning and eventual demise of the British-Irish Ice Sheet (BIIS).The BIIS is a historical ice sheet that periodically covered parts of Britain and Ireland since the Early Pleistocene (∼2.5 Ma; Peters et al. 2015;Scourse et al. 2021;Clark et al. 2022;Gibbard et al. 2022; Fig. 2).The Quaternary evolution and estimates of the maximum extent of the BIIS have been the subject of several studies (Knutz et al. 2007;Scourse et al. 2009Scourse et al. , 2021;;Thierens et al. 2012;Peters et al. 2015Peters et al. , 2016;;Clark et al. 2022).Despite this, the history of the BIIS is not as well known prior to the last glacial cycle (Gibbard et al. 2022).After the Middle Pleistocene, at least three stable ice-sheet phases during different glacial periods covered Britain and Ireland, the Anglian Ice Sheet (MIS 12) (Toucanne et al. 2009), Wolstonian Ice Sheet (MIS 10-6) and the Devensian Ice Sheet (MIS 5d-MIS 2) (Lee et al. 2011) (Fig. 2, Table 1).
The Wolstonian BIIS (Fig. 2) spanned multiple glacial cycles from MIS 10-6 (Gibbard et al. 2022), and during its maximum, it expanded eastwards in the North Sea and combined with Drenthe glaciation in the Netherlands (Gibbard et al. 2009(Gibbard et al. , 2022;;Moreau et al. 2012).The detailed history of the Wolstonian BIIS inception, demise and the history of its older phases are not well known, including records from MIS 7d, an anomalous stadial event (Toucanne et al. 2009;Hughes et al. 2020) within the interglacial MIS 7. The last BIIS phase, known as the Devensian (MIS 5d-MIS 2; Fig. 2), has been studied extensively (Eynaud et al. 2007;Clark et al. 2012;Peters et al. 2016;Bradwell et al. 2021;Scourse et al. 2021).Its evolution between 30 000-15 000 years ago is chronicled in detail by the BRITICE-CHRONO project (Clark et al. 2022), but its pre-LGM history is not as well known due to the sparseness of associated terrestrial glacial deposits (Gibbard et al. 2022).Thus, looking at proximal marine sedimentary records is necessary to fill in the gaps of the BIIS chronicle, particularly those older than the last glaciation cycle.
One result of the interaction between abrupt climatic oscillations and ice-sheet bodies in the North Atlantic is the periodic deposition of coarse-grained lithics from calving ice sheets known as ice rafted debris (IRD), such as those deposited within an area known as Ruddiman's IRD Belt (Ruddiman et al. 1989) during Heinrich Events (HE) (Heinrich 1988).Mud-rich marine sediments  An IRD record from MD01-2461 (Porcupine Sea Bight; Fig. 1) indicates, despite its proximity to the BIIS, multiple IRD sources other than the BIIS, including the LIS, Icelandic Ice Sheet and the Fennoscandian Ice Sheet (Peck et al. 2006).Haapaniemi et al. (2010) found minor yet distinct (<4% total IRD observed) peaks of LISsourced IRD in sediments coeval to HE 2 and 1, within predominantly BIIS and Irish Sea Ice Stream sourced IRD from OMEX-2K (down-slope, east Goban Spur) (Figs 1, 2).DSDP 548, another core from the Goban Spur (Fig. 1), is south of MD01-2461 and east of OMEX-2K; therefore, it is further from the LIS source, suggesting less IRD from the LIS.In addition, DSDP 548 is very close (∼100 km east) to the maximum extent of the BIIS (Clark et al. 2022;Fig. 2) and directly downstream from the Irish Sea Ice Stream (Scourse et al. 2021;Fig. 1), making it an ideal archive for documenting IRD deposition from the BIIS with minimal influence of LIS sourced IRD.
DSDP Site 548's mud and marl sequences were suggested to be related to glacial-interglacial cycles but lackedagedataforMarineIsotopeStagesassignment (De Graciansky et al. 1985).Seismic profile interpretation at Site 548 identified Mediterranean Outflow Water (MOW) associated contourite sequences (Delivet et al. 2016).The strengthening of the MOWduring interglacial periods (Kaboth et al. 2017) is associated with coarsegrained contourite sedimentation along its flow path (Van Rooij et al. 2010).Thus, the alternation of IRDladen sediments with coarser contouritic sequences differentiates glacial-interglacial sediments of DSDP Core548,warrantingamoredetailedagemodeltoextract a detailed palaeoclimatic and palaeoceanographic history, including iceberg calving from the BIIS.This study thus has several aims.Firstly, to establish an age model for the upper 40 m of DSDP Site 548 based on benthic foraminiferal stable isotope correlation.Secondly, to identify the facies and their association with glacialinterglacial cycles.Thirdly, to examine the distribution of IRD lithics and their connection to Heinrich Events and BIIS iceberg calving periods.Lastly, to interpret the history of the BIIS and the polar front in the Goban Spur region for the past 250 000 years.

Setting
The Goban Spur (Williamson et al. 2011; Fig. 3A) is a topographic high seabed feature on the Celtic Margin 200 km offshore southwest Ireland, between 900 and 1600 m water depth (Colin et al. 1992).To the north, a gentle northwardly dipping slope connects it with the Porcupine Sea Bight, while two meandering submarine canyons erode its terraced western slope (Delivet et al. 2016).The eastern North Atlantic Ocean's continental margin between 26°N and 56°N is a 'glacially influenced margin' (Weaver et al. 2000) where underwater channels and canyons dominate the down-slope sedimentation  Shotton (1983) processes, particularly during glacial periods.The Goban Spur lacks submarine channel systems (Fig. 3B) and thus represents a low-energy environment protected from active down-slope processes and sedimentation (Loubere 1987), making it ideal for recording IRD deposition during the glacial period.
DSDP Site 548 was cored during the DSDP Leg 80 in 1984 at 12°09.84 0 Wat a water depth of 1256 m (Fig. 3A).The downhole wireline logging was conducted at Site 548A (48°54.93 0N, 12°09.87 0W), 30 m west of the cored section.This study focuses on the top 40 m of DSDP Core 548, consisting of alternations of Pleistocene to Holocene marly ooze and mud (Chennaux et al. 1985;De Graciansky et al. 1985).Preliminary biostratigraphy indicates a Pleistocene age for Unit 1 overlain by 28 cm of Holocene unlithified brown mud (De Graciansky et al. 1985).

Facies analyses
The facies of the 45 m DSDP Site 548 were logged and sampled for microfossils, lithic grains (IRD), and calcium carbonate content (%CaCO 3 ).The %CaCO 3 was measured on 89 samples using a modified volumetric technique of Huelsemann (1966) (cf.Wallace et al. 2002).

Foraminiferal analysis
Eighty-nine samples from DSDP Site 548 were dried, weighed, reacted with hydrogen peroxide, wet sieved using a 63-μm sieve, dried and split using a micro splitter for quantitative analysis.Approximately 150 to 200 planktonic foraminifera were counted in the >150 μm fraction, a size fraction typically used in palaeoceanographic studies (Kucera 2007).
We used Kucera's (2007) classification to identify planktonic foraminiferal assemblages based on their SST distribution and van Kreveld's (1996) classification to identify North Atlantic Current assemblages, which contains taxa from the sub-polar and transitional assemblages (Fig. 5A, Table 2).
Planktonic foraminiferal diversity is estimated using the criteria of Spellerberg & Fedor (2003), where the term 'diversity' refers to the number of species and individuals in each sample and not 'species richness', which commonly refers to the number of species observed in a population.Following the recommendation of Jost (2006), we use the effective number of species (ENS), calculated from the Shannon-Wiener index (Shannon 1948), as the measure of diversity.
We calculated diversity initially using the Shannon-Wiener index (Shannon 1948): where H: Shannon-Wiener index, n: total number of species in a sample, and pi: proportion of individuals belonging to the i-th species.Followed by conversion to ENS (Jost 2006): or: where ENS: effective number of species, H: Shannon-Wiener index, n: total number of species in a sample, and pi: proportion of individuals belonging to the i-th species.

Lithic grain/ice rafted debris (IRD) analyses
A lithic grain census was carried out on an aliquot of the microfossil residue.Lithic concentration is expressed as the number of IRD per gram of dried sediment.The lithic grains were counted in the >250 and >150 μm size fractions.The >250 μm fraction was counted to determine the stratigraphical distribution of IRD, as the grains in this size fraction are not likely to be transported to DSDP Site 548 by means other than calving icebergs due to their size and weight.Splits of microfossil residue with a minimum of 200 counts of lithic grains were also counted in the >150 μm size fraction.The >150 μm size fraction is typically counted for IRD lithic analyses in the northwest Atlantic (Andrews 2000;Peck et al. 2007;Scourse et al. 2009).The 150 μm lithics were classified using the Peck et al. (2007) classification (Table 3) and can be divided into clear quartz, stained quartz, 'Heinrichlayer'-related, BIIS-related, and volcanic grains.Unidentified lithic grains with multiple possible origins were classified as 'other'.

Benthic foraminifera stable isotope analysis
Carbon and oxygen isotope (δ 13 C and δ 18 O) measurements on the benthic foraminifera taxa Cibicides spp., Cibicidoides spp., Cassidulina laevigata and Bullimina spp.were carried out at the Institute of Geology and Palaeontology at the University of M ünster (Germany) using a GasBench connected to a ThermoScientific Delta V Plus mass spectrometer via a ConFlo III interface.Individual measurements were carried out on sonicated foraminiferal tests, with sample weights between 50 and 180 μg per analysis.Results are reported in the standard delta notation as per mil difference relative to V-PDB (Vienna Pee Dee Belemnite).Accuracy was checked against two in-house standards, as well as against NBS-19.Reproducibility, as determined through replicate measurements of standards that were not used for the correction scheme, was better than AE0.16‰ for δ 18 O and AE0.06‰ for δ 13 C (2σ).

Age model
The new age model for the top 40 m of DSDP Site 548 is principally based on the correlation of multispecies benthic oxygen isotope ratio, including Cibicides spp., Additional datums forour age model were obtained by correlating %NPS trends to MD03-2692 (Mojtahid et al. 2005) in the northern part of the Bay of Biscay, 320 km southeast of the Goban Spur (Figs 1, S4).The tie points used to constrain the age are in Table 4.The age model for the studied interval and its linear sedimentation rate are then derived by linearly extrapolating the tie points (Fig. S5).Comparison to the data sets from ODP 983 (Barker et al. 2019), MD01-2461 (Peck et al. 2007) and MD03-2692 (Mojtahid et al. 2005) (Fig. S6) shows that the age model for DSDP 548 conforms well with regional trends.

Facies and physical property variability
Four alternating facies are present in the top 45 m at DSDP Site 548, organic-rich mud, calcareous mud, calcareous silt, and calcareous silty sand (Figs 7A,8).Bedding contact between these lithofacies is gradational or sharp and often bioturbated (Figs 7B, 9).Bioturbation features include ellipsoids and vertical and horizontal burrows infilled with adjacent facies (Fig. 7B).Occasional mud intervals are weakly laminated with large (up to 2 cm in diameter) lithic clasts (Figs 7A, 9).The % CaCO 3 positively correlates with sediment grain size, ranging from a minimum of 5% in the mud to 65% in the silt/sand facies (Fig. 8).%CaCO 3 in the mud reaches a maximum of 30%, whereas, in the calcareous silt, it varies from 30 to 50% and 40 to 50% in the silty sand.The organic-rich mud facies (Fig. 7A) typically yield abundant %NPS, IRD and elevated NGR values with low  diversity planktonic foraminiferal assemblages (Fig. 9).In contrast, calcareous silty sand facies (Fig. 7A) are characterized by minor %NPS and IRD, low NGR values and high planktonic foraminiferal diversity and abundance and are associated with warmer interglacial intervals (Fig. 8).

Lithic grain distribution
The down-core lithic variability between the >150 and >250 μm fractions is similar (Fig. 10), and lithic grain abundance shows a strong positive correlation to %NPS (Fig. 11), suggesting a significant increase of both size fractions during the colder periods.Lithic grain assemblage in the >150 μm fractions is dominated by clear quartz, which makes up 65 to 75% of the total clasts (Fig. 10).The next most common grain type is stained quartz grains, with colours ranging from yellow to reddish-orange and occasionally light pink due to iron oxide staining.Other common lithic grains include mica schist, vesicular basalt and obsidian, rhyolitic tuff, granite and granodiorite, grey shale, chalk, and dark grey to light cream-coloured carbonates.BIIS lithics dominate IRD maxima, except at ∼14.6 ka, where volcanic clasts are common.The Heinrich lithics are a minor component of some IRD-rich horizons, such as at 60 and ∼33 ka, corresponding to HE 6 and 3 (Fig. 10).Heinrich lithics are rare, making up less than ∼2% of all lithics, suggesting minor LIS-sourced IRD at Site 548.The absolute abundance of the LIS-sourced IRD at the Goban Spur is less than at OMEX-2K (Haapaniemi et al. 2010) or MD01-2461 (Peck et al. 2007).The Wolstonian BIIS glacial shows two IRD maxima near the start and end (∼265 and 252 ka during MIS 8 and ∼155 and 132 ka during MIS 6; Fig. 10).In addition, an IRD maximum is at ∼221 ka during the MIS 7d stadial.IRD maxima at ∼252, 221 and 132 ka correlate to Terminations 3, 3A and 2. During the Devensian phase, several IRD peaks are equivalent to HE 3, 6 and 1 and the final phase of the last BIIS at ∼14.6 ka (Fig. 10).However, the sampling density in this interval is insufficient to determine the nature of other Heinrich Events (HE 5, 4 and 2).

Interpretation and discussion
The glacial-interglacial sedimentation at the Goban Spur Presently, the most active bottom current at Site 548 is the MOW (Fig. 4A, B).MOW intensity varied over the last ∼130 000 years (Kaboth et al. 2017), with stronger MOW regimes during the interglacial periods.The alternations of calcareous silty sand contourites (Fig. 9) with warmer water planktonic foraminifera (Fig. 12) during interglacial periods and glacial mud pelagites with abundant cold water species (Fig. 8) at Site 548 suggests that this glacial/interglacial modulation of relative MOW intensity extended back to at least 250 000 years ago.

Lithics and IRD at the Goban Spur
The strong positive correlation between lithic grain abundance of both size fractions and %NPS (Fig. 11) suggests that the lithic clasts in the uppermost 40 m of DSDP Core 548 are associated with glacial periods.The lack of submarine channels and canyons affecting the down-slope glacial sedimentation at the Goban Spur (Fig. 3B) suggests that these glacial lithic clasts are related to iceberg calving.The relative rarity of Hudson Bay-related lithics and the dominance of the BIIS lithics (Fig. 10) in IRD-rich mud horizons (Fig. 8) suggest that they were mainly sourced from the BIIS.The two Heinrich Event (HE 6 and 1) equivalent horizons sampled in the uppermost section are characterized by abundant IRD, >80%NPS, and low planktonic foraminifera productivity and species diversity.In comparison, the HE 3 equivalent horizon is marked by a slight increase in IRD and below polar level %NPS (Fig. 13).
At IODP Site U1308, IRD-rich horizons are typically characterized by high Si/Sr ratios, and Heinrich Event IRDs from Hudson Bay have high Ca/Sr ratios due to the abundance of detrital carbonate (Hodell et al. 2008).Thus, high Ca/Sr and Si/Sr ratios at IODP Site U1308 indicate increased IRD deposition from the LIS.Hodell et al. (2008) used these ratios to identify Heinrich Events in the mid-latitude Central Atlantic.Thus, by comparing the Si/Sr and Ca/Sr ratios at Site U1308 and the %NPS and IRD concentration at Site 548 and MD03-2692 (Fig. 13), it is possible to relate IRD events at Site 548 to LIS discharge, BIIS discharge, or other northwest European ice-sheet discharge.This relationship is essential during stadial periods as the westward flowing North Atlantic Current was not present north of 45°N, and the oceanography of the region was dominated by polar water (van Kreveld 1996; Fig. 5B); this allowed icebergs from northwest European ice sheets to travel westwards to the central Atlantic Ocean.We characterize IRD maxima with abundant %NPS horizons at Site 548 based on their association with MIS termination events (Tables 5, 6).We suggest IRD maxima not associated with MIS terminations could represent cold periods when ice shelves were forming around the BIIS (Batchelor et al. 2019), and their subsequent calving may relate to factors such as glacial-isostatic adjustment related to sea level change (Bradley et al. 2023) or megatides (Scourse et al. 2018).In comparison, IRD maxima at terminations represent the intense iceberg-calving episodes associated with the waning of the BIIS with the onset of interglacial climate.

Polar front and BIIS discharge history at the Goban Spur
This section outlines 250 000 years of the polar front and BIIS iceberg discharge history at the Goban Spur.
We follow the BIIS chronology of Hughes et al. (2022) and Clark et al. (2022), the MIS chronology of Lisiecki & Raymo (2005), where evenly numbered MIS correspond to colder periods, and odd numbers indicate warmer periods along with the standardized marine isotope substages naming convention of Railsback et al. (2015).In addition, we also use stage names specific to the British Isles for each MIS (Table 1).Our results are then compared to MD03-2692 (Figs 1, 13) to determine the polar front history of the Celtic Margin.
Marine Isotope Stage 8 (Middle Wolstonian BIIS).-The polar front was south of the Goban Spur at ∼277 ka, from ∼265 to ∼263 ka, and at ∼243 ka, with IRD peaks at ∼277, ∼265 and ∼253 ka.These IRD maxima correlate to periods with high Si/Sr and low Ca/Sr ratios at IODP Site 1308 (Hodell et al. 2008;Fig. 13), suggesting increased IRD input to the North Atlantic from sources other than the LIS, possibly the northwest European ice sheet.The ∼265 ka IRD peak has an equivalent at MD03-2692 in the Bay of Biscay (Mojtahid et al. 2005), interpreted to represent the BIIS maximum during MIS 8, equivalent to the Fennoscandian Ice Sheet maximum extent (Olsen et al. 2013).Subsequently, the BIIS disintegrated rapidly during Termination 3 following the terminal ice rafting event at ∼252 ka.
The migration of the polar front south of the Goban Spur at ∼243 ka is atypical as it does not coincide with significant IRD deposition at the Goban Spur or MD03-2692 in the Bay of Biscay (Mojtahid et al. 2005;Fig. 13).This event was relatively short-lived as, ∼3000 years later, the %NPS fell from ∼94 to ∼9%, equivalent to a sea surface temperature (SST) increase of >5 °C (Govin   1985).API = American Petroleum Institute (a unit for natural gamma radiation).D. DSDP 548 %CaCO 3 .E. DSDP 548 planktonic foraminiferal abundance.F. DSDP 548 ice rafted debris (counts/gram dried sediments).G. DSDP 548 %NPS (this work).H. MD03-2692 %NPS.I. MD03-2692 ice rafted debris (counts/gram dried sediments) (Mojtahid et al. 2005).J. IODP Site U1308 Si/Sr ratio.K. IODP Site U1308 Ca/Sr ratio (Hodell et al. 2008).Yellow circles and blue shading indicate periods when the polar front was south of Site 548, associated with IRD maxima.Yellow squares indicate high IRD and %NPS horizons associated with MIS terminations.The termination events are in red.Grey shading indicates glacial MIS.Red bars around the LR04 curve indicate interstadial intervals, and dark blue bars indicate stadial intervals (Railsback et al. 2015).

BOREAS
British-Irish Ice Sheet and polar front history of the Goban Spur, offshore SW Ireland North Atlantic during an 'older' Heinrich-like event late in MIS 8.
Marine Isotope Stage 7. -Compared to MIS 8, initial MIS 7 conditions at the Goban Spur were warmer.The relative absence of IRD suggests a restricted BIIS, which reached its minimum by ∼234 ka, along with the onset of interstadial conditions by MIS 7e (Fig. 13).However, the interstadial climate did not persist throughout MIS 7 as the polar front migrated south of the Goban Spur by ∼223 ka, associated with an IRD peak at ∼221 ka (Fig. 13), suggesting glacial conditions and an actively calving BIIS during the MIS 7d stadial.At MD03-2692, MIS 7d was marked by the influx of the cold water dinoflagellates such as Spiniferites elongatus and Spiniferites septentrionalis (Penaud et al. 2008) with %NPS maxima (Mojtahid et al. 2005), suggesting cooling in the Bay of Biscay.During this time, the pollen record from offshore Portugal (Roucoux et al. 2006) shows a significant reduction in tree density and steppe vegetation expansion, typical continental European interstadial dry and cold climates.However, the ∼221 ka IRD peak at Goban Spur is not present at MD03-2692 (Mojtahid et al. 2005), although an equivalent horizon occurs at IODP Site U1308, marked with high Si/Sr and low Ca/Sr ratio indicating increased IRD deposition from the northwest European ice sheet (Hodell et al. 2008) (Fig. 13).We suggest this variability represents significant calving events from the BIIS to the Goban Spur associated with an episode of rapid expansion during MIS 7d followed by its abrupt collapse by Termination 3A.
IRD and %NPS are rare in the Goban Spur and Bay of Biscay during the interstadial MIS 7c, suggesting warmer conditions at ∼212 ka.This period is associated with the influx of the interglacial climatic optimum indicator dinoflagellate taxon Spiniferites mirabilis at MD03-2692 (Penaud et al. 2008), while pollen records from northeastern Greece suggest a diverse forest flora during MIS 7c (Tzedakis et al. 2003).We suggest the Wolstonian BIIS was actively calving during the MIS 7d stadial with a terminal ice rafting episode associated with Termination 3A.In addition, this event is bounded by two episodes of significant reduction of IRD deposition at Site 548 during the interstadial climate of MIS 7e and MIS 7c (Fig. 13).
Marine Isotope Stage 6 (Late Wolstonian BIIS).-The polar front was at or south of Site 548 at ∼152 and ∼155 ka during MIS 6. IRD deposition at the Goban Spur reached its maximum by ∼155 ka associated with MIS 6 BIIS expansion (Fig. 13), contemporaneous with the Netherlands' Drenthe glacial advance (Gibbard et al. 2022).Toucanne et al. (2009) interpreted enhanced 'Fleuve Manche' fluvial activity east of the Bay of Biscay at ∼155 ka to be related to increased western European ice sheets' discharge intensity.High Si/Sr and low Ca/Sr ratios from equivalent horizons at IODP Site U1308 (Hodell et al. 2008;Fig. 13) suggest increased non-LIS IRD deposition to the region.Those observations suggest that during the coldest phase of MIS 6, the BIIS actively shed sediments into Britain's 'Fleuve Manche' tributary rivers while also discharging IRDbearing icebergs westwards into the central North Atlantic.
Subsequently, transitional species such as Globigerina bulloides dominated the planktonic foraminiferal assemblages at Site 548, associated with incursions of subtropical species such as Orbulina universa at ∼136 ka (Fig. 12).Mojtahid et al. (2005) also interpreted a warming trend at MD03-2692 from ∼151 to ∼135 ka.These authors suggested that climatic warming in the tropics related to a shift in the planet's eccentricity significantly warmed the North Atlantic ∼151 ka.Following this warm phase, cooling from ∼135 to ∼129 ka pushed the polar front  (Mojtahid et al. 2005).A significant IRD peak at ∼132 ka, equivalent to HE 11 (Govin et al. 2015;Mokeddem & McManus 2016), preceded the glacial conditions of ∼129 ka at Site 548.At IODP Site U1308, sediments dated at ∼132 ka have high Si/Sr yet low in Ca/Sr ratios (Hodell et al. 2008;Fig. 13), suggesting increased 'non-LIS' IRD discharge.The ∼132 ka IRD peak is interpreted as a sign of the final massive iceberg discharge event from the Wolstonian BIIS before its retreat during interglacial MIS 5e, followed by a significant reduction or even temporary cessation of IRD deposition from the BIIS.
Marine Isotope Stage 5e (Ipswichian Interglacial).-Following the demise of the Wolstonian BIIS by the end of MIS 6, the climate warmed rapidly and reached its optimum by ∼123 ka during MIS 5e.Shackleton et al. (2003) documented a marked increase in Euro-Siberian and Mediterranean vegetation and a sharp decline in steppe vegetation in the pollen record at MD95-2042 (Fig. 1; Tagus Abyssal Plain, west offshore of southern Portugal) during this time, suggesting a warming climate.MIS 5e warming began at ∼126 ka and culminated by ∼123 ka, associated with %NPS and IRD minima at Site 548 in the absence of a BIIS (Fig. 13).
Marine Isotope Stage 5d-2 (Devensian BIIS).-Significant IRD deposition to the Goban Spur from the Devensian BIIS resumed at MIS 5d.Marked fluctuations in %NPS (from 90 to 7%) and IRD concentration (Fig. 13) characterize the instability of the Devensian BIIS initial growth phase.After a minimum during MIS 5e, %NPS and IRD increased until ∼111 ka.However, the %NPS was below the polar front threshold of 80%, which, based on the %NPS temperature estimations of Peck et al. (2008) and Govin et al. (2012), was equivalent to an average SST of ∼9 °C.In addition, the presence of transitional and subtropical planktonic foraminifers also suggests that the polar front was north of Site 548 (Fig. 12).Thus, the elevated IRD concentration at ∼111 ka in the Goban Spur likely originates from other northwestern European ice sheets north of the BIIS.One possible source for the IRD may be the disintegration of the Fennoscandian Ice Sheet (Fig. 1), which covered most of Norway and the highlands of Sweden during the Herning stadial at ∼110 ka (Mangerud et al. 2011;Wohlfarth 2013).Subsequently, %NPS minima and the influx of transitional and subtropical planktonic foraminifera suggest more temperate conditions at ∼109 ka (Fig. 12).Cooling followed, and by ∼99 ka, the polar front was at or south of DSDP Site 548.
The polar front was south of DSDP Site 548 at ∼85 and ∼60 ka.A significant IRD influx characterized the ∼60 ka polar front migration, equivalent to HE 6 (Heinrich 1988), one of the 'non-LIS' types of Heinrich Events (Grousset et al. 2000).At MD03-2692 (Bay of Biscay), HE 6 is characterized by high %NPS, increased magnetic intensity values and high IRD concentration with reduced foraminiferal productivity, while at IODP Site U1308, it is characterized by increased detrital carbonate, a magnetic intensity maximum, high Si/Sr and low Ca/Sr ratios (Fig. 13), indicating increased IRD input from the northwestern European ice sheets (Hodell et al. 2008).
Throughout MIS 3, %NPS never exceeded 80% at Site 548, indicating sub-polar SST just south of the polar front with the warmest period at ∼46 ka.IRD deposition was continuous throughout MIS 3; however, it was not abundant as in either MIS 2 or 4, suggesting a persistent smaller BIIS.The minor IRD peak at ∼33 ka may be equivalent to HE 3.However, the %NPS suggests that the polar front was north of Site 548, and the IRD concentration of this event (∼400 lithic grains g À1 ) is lower than the older HE 6 equivalent horizons (∼1800 lithic grains g À1 ) (Fig. 10).This difference suggests that the BIIS was a more prominent IRD contributor to the Atlantic during HE 6, while other northwest European ice sheets north of the BIIS (Fig. 1) may have been the primary HE 3 IRD source.This finding supports the suggestion that there were different IRD sources for HE 3 and HE 6 (Andrews 2000;Hemming 2004).Grousset et al. (1993) and Gwiazda et al. (1996) suggested that the Fennoscandian and Greenland ice sheets were the primary IRD sources during HE 3.
SST cooled after ∼18 ka during MIS 2, with the migration of the polar front to or south of DSDP Site 548 by ∼15.6 ka, equivalent to HE 1, associated with increased IRD concentration at Site 548 and MD03-2692 (Mojtahid et al. 2005), and increased (minor) HEtype lithics at Site 548.Similar IRD concentration maxima with a LIS signature are also reported at MD01-2461 (Peck et al. 2007) and OMEX-2K (Haapaniemi et al. 2010), suggesting synchronous iceberg calving from the BIIS and LIS during HE 1.
The increased IRD at ∼14.6 ka was associated with sub-polar (70%) %NPS (Fig. 13).This youngest peak IRD horizon may represent the last major iceberg calving from a BIIS that had already detached into discrete ice caps centred in the highlands of Ireland and Scotland by ∼15 ka (Clark et al. 2022) during the warming period following HE 1 (McCabe et al. 1998;Scourse et al. 2009).A similar trend is also interpreted from the laminated facies at MD03-2692 from ∼16 ka to Termination 1, which, according to Mojtahid et al. (2005), represents enhanced seasonal melting of the BIIS during the period.Three IRD horizons at Site 548 may be associated with episodes of BIIS calving at ∼60, ∼15.6 and ∼14.6 ka; associated with HE 6, 1 and the terminal iceberg calving activity from the already detached British and Irish ice sheets (Clark et al. 2022).
Marine Isotope Stage 1 (Holocene).-Dominant subpolar taxa characterize the MIS 1 record at Site 548, with minor %NPS and an influx of subtropical and transitional planktonic foraminiferal assemblages (Fig. 12), associated with minor IRD input at ∼12 ka (Fig. 13).The record suggests that the Devensian BIIS's final demise occurred during a Holocene warming background, accompanied by AMOC strengthening and more vigorous ENACW, MOW and NAC, which led to the present day climatic and oceanographic configuration (Figs 4A,5A).

Conclusions
Using the presented age framework, we combined facies, planktonic foraminiferal assemblage and coarse lithic grain analyses to determine the 250 000-year history of the BIIS and North Atlantic polar front history for the Goban Spur and reached the following conclusions: • The alternation between pelagic mud and calcium carbonate-rich silty sand contouritic facies (Stow & Faugères 2008) within the top 40 m at Site 548 directly reflects the glacial-interglacial cycle for the past 250 000 years.

• %NPS record of Site 548 indicates multiple North
Atlantic polar front migrations south of the Goban Spur during glacial maxima and stadial periods, while the invigoration of the northward flowing MOW during interglacial periods generated contouritic sediments.Our interpretation of a stronger MOW during interglacial periods of the Late Pleistocene agrees with the findings of Kaboth et al. (2017).• Coarse lithic clasts census at Site 548, Ca/Sr and Si/Sr ratios from U1308 (Hodell et al. 2008) and IRD concentration at MD03-2692 (Mojtahid et al. 2005) reveal that the coarse lithic grains embedded in DSDP Core 548 are IRD from northwest European ice-sheet discharges, principally from the BIIS.• The peak IRD-rich horizons of Site 548 are grouped based on the %NPS value and their association with glacial MIS termination events.Those associated with the MIS termination events precede a significant reduction in IRD deposition to Site 548, suggesting episodes of major ice shelf collapse preceding a significant retreat of the BIIS.In comparison, those not associated with termination events reflect iceberg calving and IRD deposition related to possible ice shelf formation around the BIIS during the colder period.• The signatures of the two non-LIS ' Heinrich Events (HE), HE 6 and 3, are very different in Site 548.HE 6 (∼60 ka) has a polar level of %NPS and IRD concentration more than four times that of HE 3 (∼33 ka), and %NPS during HE 3 is below the polar level.Thus, BIIS discharge is a major IRD contributor to the Goban Spur during HE 6, while during HE 3, other northwestern European ice sheets north of the BIIS are more dominant.• Six distinct IRD peaks are associated with the middle to late Wolstonian BIIS, three of which correspond to the extreme glacial conditions at ∼267, ∼255 and ∼155 ka relating to the migration of the polar front south of Site 548.In contrast, the other three at ∼252, ∼221 and ∼132 ka correspond to Termination 3, 3A and 2, respectively.The IRD peak at ∼223 ka (MIS 7d) signifies increased BIIS activity during the extremely cold yet short-lived stadial period (Hughes et al. 2020), while the IRD peak at ∼132 ka signifies the last major iceberg calving from the Wolstonian BIIS.• Significant IRD deposition from the Devensian BIIS resumes from MIS 5d, with three significant peaks at 60, 15.5 and 14.6 ka.The two oldest peaks are coeval to HE 6 and 1, whilst the youngest peak reflects one of the last significant icebergs calving from the Devensian BIIS before its demise.

Fig. 3 .
Fig. 3. A. Sea floor profile of the Goban Spur and inset map indicating the location of the coring site during DSDP Leg 80, including Site 548 (this study), modified after Snyder & Waters (1985).m b.s.f.= metres below the sea floor.B. Submarine canyons and channel systems around the Goban Spur adapted from Kenyon et al. (1978), Weaver et al. (2000) and Verweirder et al. (2021).BIIS limit is from Clark et al. (2022).DSDP Site 548 location is adapted from De Graciansky et al. (1985).

Fig. 5 .
Fig. 5. A. Division of water-masses in the North Atlantic based on the planktonic foraminiferal assemblages, adapted from van Kreveld (1996); black arrows indicate the current direction.The present polar front location is adapted from Eynaud et al. (2009) and Barker et al. (2015).Red shading indicates subtropical water, yellow indicates transitional; green indicates North Atlantic Current water and light blue indicates sub-polar water.Polar water is the unshaded region north of the sub-polar water.B. Palaeoceanography of the North Atlantic during the Last Glacial Maximum, adapted from van Kreveld (1996), along with relevant core locations: DSDP Site 548 (De Graciansky et al. 1985), IODP Site U1308 (Hodell et al. 2008), MD03-2692 (Toucanne et al. 2009), OMEX-2K (Haapaniemi et al. 2010), and MD01-2461 (Peck et al. 2007; Scourse et al. 2009).Black arrows indicate the direction of surface currents.

Fig. 6 .
Fig. 6.The correlation between the relative abundance of Neogloboquadrina pachyderma (%NPS), %CaCO 3 , NGR (De Graciansky et al. 1985) at DSDP Site 548 and the LR04 stack of Lisiecki & Raymo (2005).Light blue shading indicates intervals with ≥80% NPS, interpreted to signify the presence of the polar front at or south of DSDP Site 548.Red dashed lines indicate the correlative horizons.m b.s.f.= metres below sea floor; API = American Petroleum Institute (a unit for natural gamma radiation).

Fig. 7 .
Fig. 7. A. Four types of facies in the upper 45 m at DSDP Site 548.B. Common bioturbation marks in the studied interval.The upper section of the core is on the left side of each image.

Fig. 10 .
Fig. 10.The abundance of lithics at Site 548 based on the classification of Peck et al. (2007).All the lithic grain groupings are from the >150 μm size fraction.

Table 1 .
Summary of Marine Isotope Stage names, general climatic conditions, and timing of the BIIS over the last 300 ka.
oxygen isotope values (Fig.6) interpreted to represent freshwater influx due to increased iceberg calving; hence those were not used to construct the age model.An attached Supporting Information file contains a detailed explanation of the age model (Data S1).

Table 5 .
List of peak IRD horizons at DSDP 548 with abundant NPS not associated with MIS terminations.Note that HE 2, 4 and 5 were not documented due to insufficient sampling resolution in the upper part of the core, whereas HE 3 is marked with lower than 80% NPS.

Table 6 .
List of peak IRD horizons at DSDP 548 with high %NPS associated with MIS terminations.