Winter seal-based observations reveal glacial meltwater surfacing in the southeastern Amundsen Sea

The rapid thinning of Antarctic ice shelves, particularly in West Antarctica (e.g., Pine Island Ice Shelf (PIIS) and Thwaites Ice Shelf), is primarily due to basal melting caused by the intrusion of warm Circumpolar Deep Water (mCDW). This warm water enters the ice-shelf cavity, melts the ice from below, and creates a meltwater-rich water mass. The behavior of this meltwater is critical because it can impact sea ice formation, influence air-sea heat exchange, and potentially alter regional and circumpolar circulation. While meltwater has been observed in front of PIIS, its spatial distribution, especially during winter, remains poorly understood. Summer observations are often confounded by solar warming, making it difficult to isolate the meltwater signal. This study addresses this knowledge gap by utilizing novel wintertime observations to determine the distribution of meltwater exiting from beneath PIIS.

Previous research has highlighted the significant role of warm Circumpolar Deep Water (mCDW) in basal melting beneath Antarctic ice shelves, particularly the PIIS. Studies using summer hydrographic data have identified meltwater plumes, but the interpretation is often complicated by solar warming effects in the upper ocean. The impact of meltwater on regional and global climate is multifaceted. It can affect the formation of Antarctic Bottom Water, influence sea ice extent through polynya formation, and contribute to nutrient cycling. Existing studies have focused primarily on the deep meltwater layer, with limited understanding of the near-surface meltwater distribution, particularly during the winter months when surface processes are minimized.

This study leverages a unique dataset of 625 full-depth profiles of salinity and temperature collected by sensors attached to three seals during the austral winter of 2014. These seal-based measurements provide unprecedented spatial coverage in the challenging winter conditions of the Amundsen Sea. The data are complemented by existing summer data collected via ship-based Conductivity-Temperature-Depth (CTD) sensors and a Vertical-Microstructure Profiler (VMP), along with noble gas measurements. The meltwater content is calculated using a composite-tracer method employing conservative temperature and absolute salinity. The analysis focuses on four sections across Pine Island Bay (PIB), revealing the spatial distribution of meltwater in both winter and summer. Summer meltwater content is calculated from both hydrographic and noble gas data, with the latter considered more reliable in the near-surface layer due to the lack of solar warming influence. Uncertainties associated with seal-tag data accuracy and the choice of end-points for water mass calculations are addressed through Monte Carlo simulations and sensitivity analysis of end-point selection. The impact of winter surface cooling on meltwater estimates is also quantified.

The study reveals a stark contrast in meltwater distribution between winter and summer. In summer, the meltwater is relatively uniformly distributed below the 27.55 isopycnal (approximately 450 m), while in winter, a highly variable pattern emerges. Winter observations show two distinct meltwater-rich layers: a near-surface layer (upper 250 m) and a deeper layer around 450 m. These layers are connected by scattered meltwater-rich columns extending through the intervening meltwater-poor Winter Water (WW) layer. The near-surface meltwater layer in winter is unambiguously identified based on temperature data alone, eliminating ambiguity caused by solar heating. The meltwater's buoyancy in winter appears to facilitate its rise to the near-surface without significant lateral mixing. This is in contrast to summer, where the stratified upper ocean leads to significant mixing and a relatively homogenous meltwater distribution. The PIB gyre's influence on meltwater distribution is evident in both seasons, but the patterns differ. Analysis of multiple sections reveals that meltwater extends further away from the PIIS along sections aligned with the PIB gyre circulation compared to sections running counter to the gyre. The spatial extent of the near-surface meltwater layer is potentially underestimated due to surface cooling and brine rejection during sea ice formation. The presence of two meltwater-rich layers in winter suggests different mechanisms are controlling their formation and that the pycnocline at the mCDW-WW interface may trap some meltwater, forming the deeper layer.

The findings demonstrate the critical role of upper-ocean stratification in shaping meltwater distribution. The homogeneous winter WW layer allows for the upward transport of meltwater through columns with minimal mixing, while the stratified summer layer results in extensive mixing and a relatively uniform meltwater distribution. This contrasts with the assumptions in many Earth system models where meltwater is often represented as uniform layers at specific depths. The near-surface meltwater in winter is crucial for maintaining polynyas by providing a heat source that prevents sea ice formation. This suggests that the actual heat and freshwater fluxes from meltwater could be significantly underestimated in studies that do not consider this winter phenomenon. The enhanced near-surface meltwater concentration in winter may contribute to increased biological productivity in the spring, as it maintains nutrients such as iron in the euphotic zone. The observed processes highlight the limitations of current modeling approaches, as uniform layer representations fail to capture the seasonality and spatial heterogeneity of meltwater distribution, which are vital for accurate simulations of ocean circulation, ice-shelf interactions, and surface heat exchanges.

This study presents compelling evidence of a highly variable, yet significant, near-surface meltwater distribution in the Amundsen Sea during winter. The use of seal-mounted sensors reveals meltwater layers and columns not previously observed. The buoyant nature of the meltwater in winter, combined with weak mixing in the relatively homogeneous upper ocean, facilitates the upward transport of substantial quantities of heat and freshwater, impacting sea ice formation and influencing air-sea fluxes. The findings underscore the necessity of incorporating this nuanced spatial and temporal meltwater distribution into Earth system models to improve the accuracy of projections related to Antarctic ice sheet stability and its effects on global climate. Further research is needed to investigate the interannual variability of this phenomenon and to better understand the detailed processes controlling the formation of the meltwater columns.

This study is based on one year of winter data, which might not be sufficient to capture the full range of interannual variability driven by processes like ENSO and the SAM. More extensive wintertime datasets are required to improve robustness. The calculated meltwater content is sensitive to the choice of end-points used in the composite tracer method, although the spatial distribution patterns remain consistent across different end-point selections. Furthermore, the surface processes of cooling and brine rejection likely affect the near-surface meltwater estimates, potentially underestimating the actual meltwater contribution.