Methodology:

In total 5 different datasets were acquired which are used and will be presented in Schlegel et al (submitted). The 5 datasets are described as follows:

Main grid/DEL500 acquired over a 3 x 17 km area (Schlegel et al., 2022), with 500 m line spacing (parallel to ice flow).

Upstream grid, composed of DEL20UP and DEL50UP which were acquired in 2017/18 and 2016/17 respectively. DEL20UP covers a 1 x 3 km area with sampling compatible with 3D migration (2 m along-track, 20 m across-track). DEL50UP uses a 50 m across-track spacing, extending the DEL20UP coverage 2 km further downstream. The acquisition procedure and parameters are otherwise same as DEL500. Processing of the densely spaced grids includes regularisation of the acquisition grid by interpolation, bandpass filtering, correction for geometric spreading and 3D Kirchhoff Time migration (horizontal resolution of 24 m in both along- and cross-track directions).

Downstream grid/DEL20DW: this grid covers a 1 x 3 km area, with the same acquisition parameters and resulting resolution as DEL20UP.

UpB09 and TOL05 are 4.6 km and 3.6 km long 2D migrated seismic reflection profiles (Smith & Murray, 2009), respectively. They were acquired during the 2009/10 and 2004/05 Antarctic summer seasons.

These three datasets were combined into three different sub-datasets which were then further analysed in Schlegel et al. (submitted):

1. High-resolution topography from 3D migrated radar data:

2. Radar reflectivity from 2D migrated radar data:

3. Seismic acoustic impedance:

3D topography: Data in DEL20UP, Del50UP and DEL20DW are compliant with 3D migration. For the application of the migration data were interpolated into a regular grid with regular line and trace spacing. Following this, data were frequency filtered and 3D migrated using a Kirchhoff Time migration approach (migration velocity of 0.168 m ns-1) in SeisSpace/ProMAX. The bed reflection was picked using a semi-automated 3D seeded picking algorithm in Petrel software (Landmark). The two-way traveltime of the bed reflection was then interpolated to a surface with 65 x 20 m cell size.

Radar reflectivity: DEL500, DEL20UP, Del50UP and DEL20DW were frequency filtered and 2D migrated (using a finite difference approach and migration velocity of 0.168 m ns-1), followed by the picking of the bed reflection using ReflexW software (Sandmeier Scientific Software). Reflectivity values were calculated as described in Schlegel et al. (2022) (using a 280 ns window).

Seismic acoustic impedance: UpB09 and TOL05 are both seismic reflection profiles. Acquisition and procession are as described in Smith et al., 2007.

Data collection:

Instrumentation:

Surface radar data were acquired using the British Antarctic Survey (BAS) DELORES radar system (King et al., 2007). Acoustic impedance values are calculated from the bed reflection acquired using seismic reflection profiles as described in Smith et al. (2007).

The centre frequency of the DELORES data presented here is 3 MHz, the wavelength in ice (wave speed in ice=0.168 m ns-1) is 56 m. Lines were acquired perpendicular to the ice flow, with a 500, 50 and 20 m spacing between them. The along track sampling is around 2 m. Topography data has been 3D migrated, the theoretical horizontal resolution of the data after 3D migration is 12 m.

The centre frequency of the seismic data is 150 Hz, the wavelength in ice (wave speed in ice=3841 m s-1) is 25 m. Seismic lines presented here were acquired along ice flow. 2D migration of the data results in a theoretical horizontal resolution of 6.4 m.

Software:

ReflexW (Sandmeier Scientific Software)

MATLAB for the calculation of the received power

SeisSpace/ProMAX (LGCHalliburton Software)

Petrel (Schlumberger Software)

Data quality:

The 3D migrated radar data shows more details of the bed topography compared to the 2D migrated data. Even though most parts of the bed show a clear reflection, the steep flanks and parts of the steep stoss end of the Bump lack a clear return in the 3D migrated data. Similarly, the bed reflection along the stoss end of the Bump is not identifiable in the seismic line.

Missing values in the data are presented as #NV.