BedMachine Antarctica

Since 2014, we have been working on a self-consistent dataset of the Antarctic Ice Sheet based on the conservation of mass. A first version will be soon released to the community.

Note: keep in mind that not all of the bed is mapped using mass conservation (MC). If there is a fast flowing region that is currently not mapped with MC, let me know and I will try to add it in the next release. If you have ice thickness data that is not included, I would also be more than happy to add it to the mapping to further improve the topography.


The data are in one single file in NetCDF format (3.0 Gb) and all heights are in meters above mean sea level (the geoid used is provided in the NetCDF file). All the data use the same 450 m-resolution grid although the “true” resolution of the bedrock may vary depending on the method used to map the bed. This dataset uses data from 1993 to 2016 and has a nominal date of 2012 (same as REMA).

Antarctic mask
The ice/land masks are from ADD rock outcrop, and the floating ice is derived from InSAR grounding lines (pers. comm.). 0 = ocean, 1 = ice-free land, 2 = grounded ice, 3 = floating ice, 4 = lake Vostok
Surface elevation
The surface dem is from REMA.
Method used to calculate ice thickness: 0 = none, 1 = REMA/IBCSO, 2 = Mass conservation, 3 = interpolation, 4 = hydrostatic equilibrium, 5 = streamline diffusion, 6 = gravity, 10+ = bathymetry data
Ice thickness
The ice thickness is inferred using mass conservation along the peryphery of the ice sheet and ordinary kriging in the interior.
Bed topography
The bed elevation is calculated by subtracting the ice thickness from the surface elevation data.
Error map
Ice thickness and bed topography error.

For the hydrostatic equilibrium calculation, we used a density of ice ρice=917 kg/m3, and an ocean water density of ρocean=1023 kg/m3.

As any model output, there are errors in these maps (there is an estimate included in the dataset). Feedback is more than welcome.


  • TBD


The projection is Polar Stereographic South (71ºS, 0ºE), which corresponds to ESPG 3031

Reading with MATLAB

MATLAB now has an extensive library for NetCDF files.
filename = '';
x = ncread(filename,'x');
y = ncread(filename,'y');
bed = ncread(filename,'bed')'; %Do not forget to transpose (MATLAB is column oriented)
%Display bed elevation
imagesc(x,y,bed); axis xy equal; caxis([-1000 1800]);

Converting heights to WGS84

All heights are referenced to mean sea level (using the geoid EIGEN-6C4). To convert the heights to heights referenced to the WGS84 ellipsoid, simply add the geoid height:

\(z_{ellipsoid}=z_{geoid} + geoid\)


The ice thickness and bed topography are model outputs and are not free of error (especially in regions where ice thickness measurements are sparse). This dataset is a work in progress and we encourage users to send us feedback so that we keep improving it.

Acknowledgements and References

This project is performed at the University of California Irvine under a contract with the National Aeronautics and Space Administration (Sea Level Rise Program #NNX14AN03G and MEaSURES-3) and the National Science Foundation (Thwaites #1739031).

The ice thickness data are from:

The bathymetry data are from:

The surface velocity data used by MC are from:

The surface mass balance is from RACMO 2.3 1 km:

Ice thickening rate is from SPOT dem differencing:

The surface topography map is from:

  • REMA

The mask is derived from: