Several points

[CarlosGrohmann]

Hi, congrats on the initiative!

I'm not sure if I should make a PR or start a discussion...

There is a recent paper about the terminology of DEMs (I'm one of the authors) that might be interesting to add to the reference list:

Guth, P.L., Niekerk, A.V., Grohmann, C.H., Muller, J.-P., Hawker, L., Florinsky, I.V., Gesch, D., Reuter, H.I., Herrera-Cruz, V., Riazanoff, S., Lópéz-Vázquez, C., Carabajal, C.C., Albinet, C., Strobl, P., 2021 - Digital Elevation Models: Terminology and Definitions.
Remote Sensing, 13(18):3581.
https://doi.org/10.3390/rs13183581

From that paper, you'll see that we don't equate DEM with DTM. DEM is a generic term that can be used for any raster elevation dataset, while DTM should only be used when the data represents the terrain. All of the global or quasi-global DEMs (SRTM, Copernicus, ASTER, ALOS, NASADEM, TanDEM-X) are DSMs since they don't represent the topographic surface in vegetated or urban areas.

And here another paper comparing these global DEMs, this time for areas in Brazil:

Grohmann, C.H., 2018. Evaluation of TanDEM-X DEMs on selected Brazilian sites: comparison with SRTM, ASTER GDEM and ALOS AW3D30. Remote Sensing of Environment, 212C:121-133.
https://doi.org/10.1016/j.rse.2018.04.043

[DahnJ]

Hi @CarlosGrohmann, thank you for the comments! The paper looks very nice, I will take a proper look at it.

I agree with using DEM as a generic term and DTM as a model of the terrain (bare surface). The "definition" of a DEM in the repo that includes the word "terrain" should thus perhaps be rephrased

For SRTM etc. I initially thought of labelling them something like "DTM*", in that they do not specifically try to be a DSM, but end up being somewhere between a DTM and DSM due to the way they were collected/processed. For example, I thought that in forested areas, SRTM is somewhere between a DTM and a DSM. I am happy to be corrected on that though and simply call everything (in the global section) except FABDEM a DSM. What do you think would be the most accurate way of categorizing the DEMs?

I think the papers would make a great addition to the list.

I'm happy to receive suggestions on what changes to make or even a PR.

edit

After reading a bit more, I have made the following changes

• Replaced "terrain" with "surface" in the definition of a DEM
• Categorized everything except FABDEM as DSM
• Added the papers you have linked

Looking forward to your suggestions, it's great to have people help with this!

[DahnJ]

I have finished reading the terminology paper, thank you for sending it. It's a great overview of many of the considerations one should be aware of with DEMs.

I think having a section in the repo that outlines some of the important considerations for choosing a DEM could be really useful. Things such as

• Thinking about how hydrosphere, cryosphere, biosphere, and anthroposphere are represented in the DEM (do you think it would be possible to include Figure 1 in the repo?)
• The concept of a sensor-surface grid. Mentioning some consideration about e.g. how much an X- or C-band SAR penetrates forest canopy
• Point-based vs area-based sampling
• Possibility of pixel-level metadata
• Illustrations of when different DSM/DTM/DEMs might be useful

At this point, these are mainly nice-to-haves that I'd like to include at some point. Feel free to suggest anything that you think should really be corrected or included in the repo.