@seisman mentioned on GitHub that there had been an unfunded proposal to add other planetary datasets to the GMT remote datasets. I have no experience with this subject and do not know what work is required to add datasets, but I would be happy to contribute if anybody is working on it/has a roadmap of work to be done for it.
mmm, I suspect that the idea refers to add them as part of the remote data set.
So, probably you could start with this repository
If you get the gmtserver-admin repo then you can look in the recipes folder for earth_relief.recipe and make one called mars_relief.recipe, make lots of changes to metadata in the recipe, identify the best quality NASA/John Hopkins Mars DEM, find the URL etc (it was made by GMT so we could also ask Greg Newman about what to use) and then try to run the script/srv_downsampler_grid.sh with the recipe as argument to see how it might work. @maxrjones and @Esteban82 are both maintainers who have actually done this without me holding hands so they can help too. I had found the Messenger data for Mercury and the newest lunar data as well, so should be simple to do given this infrastructure. The remote-datasets repo is mostly for documentation of things later.
Thanks for the information! I’ll see if I can find a Martian DEM. Do you have a good URL for the lunar topography? I’ve only been able to find this NASA page, but I’m not seeing any NetCDF files, only JPEG2000 and FLOAT IMG.
No problem with jp2 files. That’s how they are stored in the Hawaiian server.
Not sure but probably this link
I am happy to assist and answer questions about the process needed but best to do so via issues on the gmtserver_admin repo I think. For the moon, the 256 samples per lunar degree means the LOLA DEM is ~14 arc seconds. Close to 15s which is the max res for many other (Earth) DEMs but resampling from 14 to 15 arc sec is annoying - we many need to play with this. But if we did 15 then the usual sequence of 30s, 01m, 02m etc can be used for the derivatives. We can also make that _14s to retain full resolution since high frequencies are always lost when downsampling which nobody likes. So we can make both and see if the difference is significant. If it is then we might do 14s, 30s, 01m, etc… and even allow _15s to be an alias for the 14s grid.