Exploration of an ice-cliff grounding zone in Antarctica reveals frozen-on meltwater and high productivity

The study addresses how poorly known grounding zones of Antarctic tidewater ice cliffs influence ice-sheet dynamics and biogeochemical processes. Despite advances in remote sensing and modelling, key small-scale processes at the ice-bed-ocean interface (e.g., subglacial hydrology, basal sliding, freeze-on) remain uncertain. Access is usually hindered by thick ice and extensive ice shelves. The authors aimed to directly observe a rare accessible tidewater grounding zone in Coats Land (Weddell Sea) to (1) characterize basal ice facies and grounding geometry, (2) assess calving front dynamics, seafloor disturbance, and benthic communities, and (3) evaluate implications for productivity and carbon sequestration.

Previous work improved knowledge of Antarctic cryosphere trends and modelling but highlighted uncertainties at ice-sheet boundaries. Drilling and borehole/ROV studies beneath large ice shelves and glaciers have documented laminated basal ice facies, subglacial water, and biologically connected seafloors. Transparent basal ice facies have been observed in slow- and fast-moving ice streams, often as multilayer sequences with debris-rich layers. Tidewater termini provide potential direct access but are rarely explored due to calving risks. Biological studies show high Southern Ocean productivity in polynyas fueled by iron from glacial sources, strong bentho-pelagic coupling, and the role of iceberg scour in shaping benthic communities and carbon burial. Existing grounding-zone benthic observations are scarce (e.g., Mackay Glacier, Ross Sea).

- Site: Coats Land tidewater ice cliff, Weddell Sea, approached within ~50 m during RV Polarstern cruise PS111; ROV dive at the calving front and a ~230 m transect offshore. - ROV and sensors: Ocean Modules V8Sii with 4k camera (25 fps; 16 MP stills), reference lasers, LED lighting; Seabird 19plusV2 CTD mounted on ROV for temperature, salinity, oxygen; GAPS USBL for positioning. - Grounding-zone observations: ROV surveyed the ice face and seafloor to 148 m depth; photogrammetric reconstruction of a 4 × 1 × 1 m section using Agisoft Metashape from 57 frames; lasers and altimeter used for scale; gap height and basal ice thickness measured on the 3D model. - Freeboard and geometry: Freeboard estimated from calibrated ROV deck-height reference (5.6 ± 0.2 m above sea level) and image geometry; two-dimensional scaling from laser spacing and altimeter; image width-height functions derived; ice density parameterized as a function of thickness after Orheim. - Hydrography: CTD profiles collected near the ice front and in open water (~250 m away) to assess temperature-salinity structure, melt signatures (TEOS-10), oxygen, and chlorophyll a (fluorescence). Analysis of compensation and critical depths for phytoplankton; comparison of salinity and temperature profiles for freshwater discharge signatures. - Seafloor characterization: ROV transect normal to the front (230 m) documented scour marks, clast shapes, benthic cover and taxa; clast roundness classified from images (n=1026) using Powers’ scale; Parasound sub-bottom profiling interpreted for sediment structure and till. - Ice dynamics and modelling: Satellite velocity fields (Rignot, Mouginot) used to compute reverse flow lines and travel times over final 18 km; velocity increase towards terminus quantified; calving front extrapolated from 2015 to 2018 using rbf and linear interpolation of velocity vectors; crevasse spacing assessed from satellite imagery. - Finite Element Analysis (FEA): Ice-sheet flexure and terminus interaction with seabed modeled with Calculix via FreeCAD; geometry from IBCSO bed and ice thickness; loads from ice weight minus buoyancy using seawater density 1028 kg m−3 and ice density-thickness relation; mesh with Netgen tetrahedra; linear elastic properties (E=9 GPa, ν=0.3); analytical beam validation performed. - Basal friction and freeze-on model: Friction factor derived from slope, velocity, and ice flux modifying Coulomb friction law; along-flow profiles of friction, velocity, and potential basal ice accretion computed; conservative accretion rate set to 1 mm yr−1 to estimate cumulative freeze-on initiation point and thickness (assuming availability of subglacial meltwater; temperature below pressure-melting point from ice thickness-dependent depression of melting point). - Algal cover quantification: ROV video frames (n=1158) processed with FFmpeg and Scilab/IPCV; RGB normalization for lighting variation; HSV threshold (hue<132) to classify algal pixels; relative cover vs depth computed. - Tides: Tide Model Driver (TMD 2.5) used to estimate local tidal range; comparison to FEA-derived flexure and observed gap to evaluate tidal pumping potential.

- Grounding-zone geometry and basal ice: At 148 m depth, a 0.3–0.5 m gap exists between ice and seafloor. A ~0.4 m thick, transparent, bubble- and debris-free basal ice facies is frozen to the underside of meteoric ice, with sharp transition and square-edged base, indicating freeze-on origin and sustained subglacial water pressure allowing ice-bed separation. - Freeze-on history and origin: Model results suggest basal accretion initiated ~6.5 km upstream near the subglacial shoreline in the marine portion of the ice sheet, producing up to 400 mm of clear basal ice over ~400 years at 1 mm yr−1. Maximum ice thickness along flow (~510 m) implies ~−0.34 °C depression of melting point; low deformation stresses considered. - Calving front dynamics: Ice takes >3500 years to traverse final 18 km; velocity increases toward the terminus, thinning the last 6–7 km to ~half its former thickness. Estimated meteoric ice density at the cliff is ~901 kg m−3 for a 170 m thickness and freeboard 20.8 ± 0.7 m. Calving bands likely follow ~150 m crevasse spacing; with advance rate ~29.8 m yr−1, calving occurs on average every ~5 years. FEA predicts maximum spring tidal flexure ~2 mm, far less than the observed 0.3–0.5 m gap; tidal pumping at the terminus is unlikely. - Hydrography and melt: Upper 10–40 m T–S characteristics align with expected melt mixing lines; temperatures near cliff are cooler than offshore while salinities are similar or higher, indicating melt without substantial freshwater discharge at the grounding line. No runnels or suspended matter observed along 90 m of grounding line. - Seafloor and benthos: Seafloor slope ~2.5% matches sub-ice slope; heavily scoured with subparallel NW–W oriented scour marks; compacted till with coarse sediments, half-buried subrounded clasts (basal shear/crushing origin). Sub-bottom profiling shows strong surface reflector, low penetration, no internal reflectors. Sessile fauna scarce near front (<5% cover): Molgula sp., Homaxinella balfourensis, scattered polychaete tubes and bryozoans—indicative of early post-scour stages. Further offshore, stage 1–2 communities appear with gorgonians (Thouarella spp.), Stylocordyla chupachups, epizoic holothurians; denser fish present. - Primary producers on ice cliff and in water column: Abundant algal tufts line scalloped melt ridges of meteoric ice, especially in upper marine portion; inferred to be diatom-dominated blooms seeded from the water column and fueled by iron from melting meteoric ice. Chlorophyll a indicates an active summer bloom with >1 mg m−3 peak at ~20 m (thermo-/halocline); algal tuft lower limit at ~90 m marks compensation depth; phytoplankton remain abundant to the seabed (critical depth >148 m). Juvenile Pagothenia borchgrevinki shelter in melt cusps. - Carbon export implications: High primary production at the calving front combined with frequent scour (low benthic consumer biomass and turnover) likely enhances organic matter export and burial. Given that grounded tidewater ice cliffs rim ~38% of the Antarctic coastline, grounding zones may represent significant, underappreciated carbon sequestration hotspots.

Direct observations reveal a persistent gap and a clear, debris-free basal ice layer consistent with centuries-long freeze-on fed by subglacial meltwater in the marine portion of the ice sheet. This finding constrains subglacial hydrology near grounding zones and supports models where groundwater head can replace denser seawater to maintain pressurized basal water and freeze-on. The lack of tidal flexure relative to the gap size indicates that tidal pumping is negligible at this stable, buttressed tidewater front, refining expectations of grounding-line mobility. The seafloor scouring pattern and sparse pioneer benthos demonstrate frequent advance–retreat cycles of the calving front, resetting benthic succession at intervals of ~5–10 years. Concurrently, iron-rich melt fuels algal growth on the ice face and a strong water-column bloom, with production extending to depth, enhancing vertical flux of phytodetritus. The combination of high organic supply and low consumer biomass due to recurrent disturbance likely increases carbon burial efficiency in grounding zones, linking glaciological processes to the regional carbon pump and potentially to global climate feedbacks.

This study provides rare in situ observations of a tidewater grounding zone, documenting (1) a 0.4 m transparent basal ice layer formed by freeze-on over ~400 years, (2) a stable, non–tidally pumped grounding geometry with a measurable basal gap, (3) intense seafloor scouring that suppresses benthic biomass, and (4) unexpectedly rich algal growth on the ice cliff coincident with a water-column bloom, likely fueled by iron from meteoric ice melt. Together, these results illuminate how grounding-zone processes modulate ice dynamics, ecosystem structure, and carbon sequestration. Future work should extend observations across multiple Antarctic tidewater sectors, directly quantify iron fluxes and primary productivity on ice faces, measure organic carbon fluxes and burial rates, and integrate high-resolution subglacial hydrology with ice–ocean models to generalize grounding-zone behavior and feedbacks.

- Spatial and temporal scope: Observations are from a single site and limited transects (≈90 m of grounding line inspected; ~230 m seafloor transect) during one cruise period, which may limit generalization across Antarctica and seasons. - Inference of subglacial discharge: No direct measurement of subglacial freshwater outflow was detected; the freeze-on history and water source are inferred from ice facies, hydrography, and modelling assumptions. - Modelling assumptions: Basal accretion thickness assumes a constant conservative rate (1 mm yr−1) and availability of subglacial meltwater; friction law simplifications, linear elastic properties (FEA), parameterized ice density–thickness relation, and bathymetric/ice geometry inputs introduce uncertainties. - Remote sensing extrapolation: Ice-front evolution was extrapolated from 2015 to 2018 using velocity fields and excluded boundary data; unmodeled calving events may affect local comparisons. - Biogeochemical proxies: Algal identification and iron supply are inferred (no direct iron or taxonomic analyses reported); chlorophyll a from CTD fluorescence is a proxy and may carry calibration uncertainties.