[MASK]: BedMachine Greenland

[MathieuMorlighem]

Overview

• Product name and version: BedMachine Greenland v4
  • Paper: https://agupubs.onlinelibrary.wiley.com/doi/10.1002/2017GL074954
  • Data: https://nsidc.org/data/idbmg4/versions/5
• Purpose of your product: Bed/Bathymetry/ice thickness/surface topography of the Greenland Ice sheet
• Geographic region: Greenland

Upstream

• What upstream mask product did you use? GIMP mask v1 https://nsidc.org/data/nsidc-0714/versions/1

• Reason for using this mask. This mask was the most complete, accurate, highest resolution available at the time when BedMachine was released, and coincided temporally with other datasets.

• Effort required by you if upstream product changed (to a community standard). I had to make some minor corrections (e.g. remove supraglacial lakes)

• Type: raster

• Resolution if raster: 150 m

• Projection information: EPSG 3413

[mankoff]

Does BedMachine require a temporally aligned DEM for the mass conservation method? If we have a 2022 Greenland outline, do we need a 2022 DEM too? Which DEM is used in the latest BedMachine, and what is its temporal span? The 2017 BedMachine paper say,

We deduce the subglacial topography by subtracting our ice thickness map from a surface digital elevation model from the Greenland Ice Mapping Project (GIMP) (Howat et al., 2014), which has a nominal date of 2007

Is that still the case for v5?

I think generating a 2022 DEM is out-of-scope for this project, but I would guess someone is working on annual DEMs and perhaps we can support/motivate/leverage efforts between this group and any DEM creators.

[MathieuMorlighem]

Hi @mankoff ! Sorry for the delay. The way BedMachine works is that it computes the ice thickness for a given year and then we get bed topography by subtracting the calculated thickness from a surface dem. Ideally, all datasets should coincide temporally and I have tried to stick to ~2007 so that we would have some "hindcast" period to validate the models.

So, yes, I am still using GIMP and a nominal date of 2007, even in v5. That may change in the future though, I am contemplating switching to the Copernicus dem in the future.

[mankoff]

Hi @mmorligh,

I'm not familiar with the Copernicus DEM. If you mean https://spacedata.copernicus.eu/collections/copernicus-digital-elevation-model I note that it spans 2003 -- 2015. That's an even longer time span than the 2003 -- 2009 GIMP span, but may be a better product for other reasons.

https://doi.org/10.5194/essd-14-781-2022 is a DEM that spans 1 year - the only one I know of so far. I haven't used for anything (and wasn't involved in editing that ms), but did download it once and glance at it, and there may be some gaps at the edges of the ice sheet.

Off-topic request, but if anyone knows of a single-year DEM, or a project that is currently developing one, I would be glad to know of it.

PROMICE PRODEMs are annual, 2019 -- 2022 and continuing, but only cover the ice sheet edge, and are still in review: https://doi.org/10.5194/essd-2023-224

[Antoine-UGA]

Hi @mankoff I had in mind that Copernicus DEM was made with Tandem-X data from 2011 to 2015. Regards. Antoine

[JonathanCarrivick]

In my experience ALOS (AWD3D: https://www.eorc.jaxa.jp/ALOS/en/dataset/aw3d30/aw3d30_e.htm ) is excellent quality product, with really very few if any artefacts/errors and excellent coverage of mountain regions. Indeed it is a global dataset. It is a mosaic, generated from images collected using the Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM) aboard the Advanced Land Observing Satellite (ALOS) from 2006 to 2011.