The role of ocean and atmospheric dynamics in the marine-based collapse of the last Eurasian Ice Sheet

Rising global sea levels pose a significant threat, prompting extensive research into ice sheet stability and dynamics. Understanding the complex interplay between oceanic and atmospheric circulation and ice sheets is crucial for predicting future sea-level rise. While the retreat of terrestrial parts of Northern Hemisphere ice sheets, such as the Eurasian and Laurentide ice sheets, is relatively well-documented, the marine sectors remain less understood. This study addresses this gap by focusing on the last deglaciation of the Eurasian Ice Sheet (EIS) and its disintegration into three separate ice masses: the British-Irish Ice Sheet (BIIS), the Fennoscandian Ice Sheet (FIS), and the Kara-Barents Sea-Svalbard Ice Sheet (KBSIS). Previous reconstructions have relied heavily on radiocarbon dating and extrapolation from land areas. This research utilizes new data from the BritIce-Chrono and GLANAM programs, along with high-resolution seismic data, to provide a refined reconstruction of EIS disintegration from 20 to 14 ka, integrated with existing reconstructions for earlier and later periods. This improved understanding of past ice sheet behavior will offer valuable insights for predicting future sea-level changes and the responses of contemporary ice sheets to environmental forcing.

Existing literature on the deglaciation of the Eurasian Ice Sheet (EIS) has primarily focused on terrestrial sectors, leaving a knowledge gap regarding the marine parts. Studies such as Hughes et al. (2016) have provided chronological databases and time-slice reconstructions (DATED-1) for the entire EIS, including the marine sectors; however, these often rely on extrapolations from adjacent land areas and sparse data from the marine environment. Other studies have focused on specific regions (e.g., Stroeven et al., 2016 for Fennoscandia), providing regional insights but lacking a comprehensive, integrated view of the entire marine-based EIS. Recent research, including the BritIce-Chrono and GLANAM programs, has generated new high-resolution data on the northwestern European shelf areas, filling some of this knowledge gap. These projects have provided crucial chronological and geomorphological information concerning ice sheet extent and ice stream dynamics. Therefore, this study leverages this newer, more detailed information to build a more accurate and nuanced understanding of the disintegration of the marine-based EIS.

This study combined several approaches to reconstruct the deglaciation of the Eurasian Ice Sheet's marine sectors. First, it compiled existing and new data on glacial landforms, including ice marginal features (terminal moraines, grounding zone wedges) and glacial lineations from the North Sea, Mid-Norwegian margin, Svalbard continental shelves, and the Barents Sea. These data were mapped using ArcMAP and QGIS software. Geochronological data, primarily radiocarbon dates from marine carbonates (mollusks and foraminifera), were compiled, recalibrated using the Marine20 curve and OxCal program. The researchers acknowledged the uncertainties inherent in radiocarbon dating of marine carbonates, particularly regarding the marine reservoir effect (ΔR). A conservative approach was adopted, setting ΔR to zero, and carefully considering the stratigraphic context and potential for mixing of samples to provide the most reliable age constraints for the time slices. A steady-state ice sheet model (ICESHEET v1.0) was used to simulate the evolution of the EIS ice sheet from 25 to 10 ka, utilizing perfectly plastic ice rheology. Model input included bed topography (GEBCO2020 grid), isostatically adjusted topography (accounting for glacial isostatic adjustment using SELEN v4), and basal shear stress values parameterised based on topography and sediment cover. Sensitivity experiments were conducted by perturbing basal shear stress values by ±20%. Summer sea surface temperature (SSST) data from two deep-sea cores (LINK17 and JM373) were used to assess the influence of ocean conditions on ice sheet retreat, using planktonic foraminifera assemblages and C² program. The resulting ice volume changes were used to estimate contributions to eustatic sea level rise, with careful consideration of the contributions of both terrestrial and marine-based ice sectors. The model outputs include estimates of the total ice volume changes, partitioning of ice volume changes between different ice sheet sectors and length of the marine ice margins exposed to the ocean.

The study presents a refined reconstruction of the Eurasian Ice Sheet (EIS) deglaciation from 20 to 14 ka. The EIS, at its peak volume around 19 ka, held approximately 7.9 × 10⁶ km³ of ice, equivalent to 20 m of sea-level rise. Net ice loss increased steadily from this peak. The EIS's disintegration into separate ice sheets (BIIS, FIS, and KBSIS) was asynchronous, showing a latitudinal trend with more persistent marine-based ice in the north. The North Sea deglaciation occurred early (19-17 ka), likely driven by increased surface melting and changes in surface mass balance (SMB). The Barents Sea deglaciation was delayed, coinciding with subsurface water warming and a decrease in precipitation post-LGM. The collapse of the KBSIS, primarily marine-based, contributed to a significant increase in meltwater delivery to the Nordic Seas between 16 and 14 ka, potentially associated with Meltwater Pulse-1A. The ice sheet modeling indicates that peak contributions to eustatic sea-level rise occurred around 15 ka, with nearly equal contributions from the Fennoscandian and Barents Sea sectors. The model shows the increasing significance of marine sectors on deglaciation, highlighting how an increased proportion of the ice margin became marine-terminating (38% at 20 ka to 47% at 16-17 ka). The analysis of summer sea surface temperatures (SSST) reveals a complex interplay between atmospheric and oceanic conditions in driving the deglaciation, with the North Sea deglaciation seemingly more influenced by atmospheric conditions (surface melt) while the Barents Sea retreat was more influenced by warming subsurface waters. Analysis of the SSST data and ocean modeling results suggest that the meltwater pulse events from the EIS did not have an extensive or prolonged impact on the Atlantic Meridional Overturning Circulation (AMOC). This study highlights the significant role of major ice streams, such as the Norwegian Channel Ice Stream (NCIS) and the Bear Island Ice Stream (BeIS), in the deglaciation process. The NCIS is suggested to have been more susceptible to melt-driven feedbacks, whereas warming sub-surface waters seem to have played a crucial role in the retreat of the BeIS.

This study's findings advance our understanding of Eurasian Ice Sheet deglaciation by providing a more detailed and nuanced reconstruction of its marine-based sectors' collapse. The asynchronous nature of the deglaciation, with the North Sea retreating earlier than the Barents Sea, highlights the complexities of ice sheet response to climate forcing. The interplay of atmospheric (SMB) and oceanic (subsurface warming) controls underscores the need for integrated models that account for both processes. The model's representation of ice sheet dynamics, combined with the detailed chronological and geomorphological data, offers a framework for understanding and comparing different aspects of deglaciation. The relatively minor influence of the EIS meltwater pulses on AMOC strength suggests the importance of meltwater from other sources, such as Greenland and the Laurentide Ice Sheets, in shaping ocean circulation patterns during the deglaciation period. The results underscore the importance of detailed regional studies integrated with robust modeling for improving understanding of ice sheet response to environmental changes.

This study provides a comprehensive reconstruction of the last Eurasian Ice Sheet's marine-based deglaciation. It highlights the complex interplay between atmospheric, oceanic, and ice sheet dynamics controlling the timing and pattern of retreat. The findings emphasize the need for future research to incorporate these interacting factors into ice sheet models to improve predictions of future sea-level rise. Further research should focus on refining data resolution in under-sampled areas, particularly the eastern parts of the Barents and Kara Seas, and improving the integration of high-resolution ice sheet models with ocean circulation models to simulate the complex interactions between these systems.

The study acknowledges limitations related to data availability and uncertainties in radiocarbon dating of marine carbonates. Data density varies across the study area, with some areas having more limited data than others, potentially affecting the accuracy of the reconstructions in those regions. The use of a steady-state ice sheet model, while providing a first-order approximation, simplifies the complexities of ice sheet dynamics. Furthermore, the model's assumptions about basal shear stress and its parameterization may introduce some uncertainty into the results.