Swirls and scoops: Ice base melt revealed by multibeam imagery of an Antarctic ice shelf

Abstract

Knowledge gaps about how the ocean melts Antarctica's ice shelves, borne from a lack of observations, lead to large uncertainties in sea level predictions. Using high-resolution maps of the underside of Dotson Ice Shelf, West Antarctica, we reveal the imprint that ice shelf basal melting leaves on the ice. Convection and intermittent warm water intrusions form widespread terraced features through slow melting in quiescent areas, while shear-driven turbulence rapidly melts smooth, eroded topographies in outflow areas, as well as enigmatic teardrop-shaped indentations that result from boundary-layer flow rotation. Full-thickness ice fractures, with bases modified by basal melting and convective processes, are observed throughout the area. This new wealth of processes, all active under a single ice shelf, must be considered to accurately predict future Antarctic ice shelf melt.

Fig. 1.. Dotson Ice Shelf.

(A and B) Reference Elevation Model of Antarctica mosaic (66) of DIS with the regions and features discussed in the text labeled: basal channel features (dashed lines), ~2000 Landsat-derived grounding line [thick black lines; (67)], and 1992–2020 Interferometric Synthetic Aperture Radar–derived grounding lines [pink lines; (68, 69)]. Thin black lines show the AUV mission paths, and red squares in (B) show the location of the zoomed-in areas in (C) to (F). Blue shades show bathymetry (70) [color bar in (A)]. [(C) to (F)] Zoomed-in areas showing the ice surface elevation at the survey areas, referenced to the WGS84 ellipsoid. The yellow star in (E) shows the location of the hot water drilling site. (G) Study area location.

Fig. 2.. Upward-looking multibeam sonar maps of the underside of DIS.

(A and B) One-meter multibeam grids showing ice base topography in the East region (E1 and C2). Color bars show the depth of the ice base. In (B), red crosses depict locations of channel-like features, and yellow star shows the location of the hot water drilling site. Profiles of the ice shelf base illustrate (C) channels and (D) terraces, with distance and depth in meters, which are also highlighted by slope derivations of the (E to G) ice base surface. Black bars in (A) and (B) are 1 km. Maps are projected in Universal Transverse Mercator zone 13S (WGS84 Datum).

Fig. 3.. Upward-looking multibeam sonar maps of the underside of the central part of DIS.

(A) One-meter multibeam grid from the central survey region (C1) showing full-thickness ice fractures, superposed on smaller terraces. The profile in (B) shows that the northernmost fractures are generally larger and have been widened and eroded at these flanks; themselves containing (B and C) smaller subterraces. Black bar in (A) is 1 km, and all maps projected in Universal Transverse Mercator zone 13S (WGS84 Datum).

Fig. 4.. Upward-looking multibeam sonar maps of the underside of the western part of DIS.

(A to C) One-meter multibeam maps from the Western region (W1 to W3) show a smooth and eroded ice topography, (D and E) enigmatic teardrops in clusters, and fractures. The 45° deviation from water flow of the teardrops is highlighted by the (D) ice base slope, and their morphology is shown in three-dimensional (3D) from a viewpoint above the (E) ice shelf base. Black bars in (A) to (C) are 1 km, and maps are projected in UTM zone 13S (WGS84 Datum).

Fig. 5.. Fracture age.

Approximate fracture age in the (A) C1 survey area and (B) E1 survey area. Dates show the approximate time period of fracture propagation through the ice that advected into the survey region, based on visible surface expression in Landsat imagery. Background is a Landsat-8 image from 15 February 2022.

Fig. 6.. Velocity, temperature, and meltwater content collected from the ship (Materials and Methods).

The viewpoint is looking north, and the distance on the x axis is along the ice front. Thin black lines show the depth at which the AUV moved, and cyan markers indicate the ice shelf base as measured by the AUV multibeam about 10 km into the cavity. (A) Eastward velocity component (m s⁻¹), with negative values indicating flow toward west. (B) Northward velocity component (m s⁻¹), with positive values indicating flow toward north. (C) Conservative temperature (°C). (D) Meltwater fraction (MWF).

Fig. 7.. Velocity and temperature of the water inside and outside DIS cavity.

(A) Map showing current speed in m s⁻¹ according to the right-hand color bar, with black arrows showing the main currents inferred from our data (see also Fig. 6). White semitransparent area indicates the ice shelf extent, and blue shading and dots (left hand color bar) indicates bathymetry from the present cruise (Materials and Methods) and from gravity inversion (62). Markers inside the ice shelf are AUV data from 20 to 80 m below the ice base (Materials and Methods), markers outside the ice shelf are current speed measurements made from the ship (61) at the level of the ice shelf base (Materials and Methods), and black markers show the location of the ship temperature and salinity profiles. The yellow star in the top right corner shows the location of the mooring (Materials and Methods). Black squares indicate the survey areas in (B) to (G). (B) MWF in survey area E1 and (C) the north-south component of the velocity (positive northward) according to the color bar. (D to G) Velocity arrows (black) together with temperature measured by the AUV (color bars). Note that the maps in (B) to (G) show distance (km) east and north instead of latitude and longitude and that the color scale in (D) is different from (E) to (G).

Fig. 8.. Water mass properties and basal topographic features inside the cavity.

(A) Conservative temperature as a function of depth outside the ice shelf front, color coded by longitude (color bar). (B) Conservative temperature as a function of depth inside the cavity, color coded by longitude (color bar), and with pale gray markers repeating the data in (A). Inset shows profiles of conservative temperature from the borehole site, collected by the AUV, and the borehole profile, colored coded by longitude, for the parameter space indicated by the black rectangle in (B). Red arrows in inset point to the larger vertical gradients in the stairstep structures. (C) Density as a function of depth for the data outside the cavity (black semitransparent) and for the AUV and borehole data inside the cavity, color coded by longitude (color bar). (D and E) Temperature-salinity plots of data (D) outside the cavity and (E) inside the cavity. Color indicates longitude (color bar). Black dashed line is the meltwater mixing line, and the yellow dashed line is the freezing point. In (E), black semitransparent markers repeat the data from (D).

Fig. 9.. Sketch showing the processes discussed in the paper.

Note that the vertical scale is exaggerated.