Add lunar DEM file to remote server

[willschlitzer]

Per a discussion at the forum, I'm hoping to add a lunar topographic relief grid to the GMT remote server.

[Esteban82]

We need to define the moon cpt to finish this #182.

I think a grayscale is the default option. But, its a bit boring to me.

But, what do you think about this version with blue? I prefer this. I think that its nicer than the gray. Also, it is very easy to change to the grayscale if a user want it.

In the below link its says they used this "cpt":

Elevation_m Red Green Blue
-9000.0 56 2 123
-8000.0 0 0 89
-7000.0 0 94 187
-5000.0 60 157 255
0.0 215 255 255
2000.0 255 255 255
4000.0 192 192 192
6000.0 128 128 128
7000.0 83 83 83
8000.0 51 51 51
10000.0 52 31 31

Link: https://astrogeology.usgs.gov/search/map/Moon/LMMP/LOLA-derived/Lunar_LRO_LOLA_ClrShade_Global_64ppd_BlueSteel

In this link of the USGS there are posters of both versions (gray and with blue): https://pubs.usgs.gov/sim/3316/

[PaulWessel]

Nice, i like it! Ok with @joa-quim ?

[joa-quim]

Yes, looks nice. And we can say the blue is due to transparency and what is seen is the Moon background.

[Esteban82]

Great!!

Anyone else? @willschlitzer ?

[anbj]

Will this be added to https://docs.generic-mapping-tools.org/dev/datasets/remote-data.html#currently-available-remote-data-sets?

[PaulWessel]

@Esteban82, make sure you add that CPT (but after formatting it to our standards -see the others in cache), then add it to cache.

[PaulWessel]

Will this be added to https://docs.generic-mapping-tools.org/dev/datasets/remote-data.html#currently-available-remote-data-sets?

Yes, with Mars, etc etc

[anbj]

Will this be added to https://docs.generic-mapping-tools.org/dev/datasets/remote-data.html#currently-available-remote-data-sets?

Yes, with Mars, etc etc

rejoice: engage.

[Esteban82]

@Esteban82, make sure you add that CPT (but after formatting it to our standards -see the others in cache), then add it to cache.

Yes!! I am on it.

I would need to know the extreme values or the original grid to make the cpt. In the figure below its says (-9130/10780). Which is similar to Paul's proposal (-9200/10800).

[Esteban82]

There is an issue with the colorbar

The values published above go from -9000 to 10000 but the grid

Here I compare both colorbar:

Mine with original range values (from -9000 to 10000):

Extended up from -9200 to 11000 (I repeat the colors from the lowest and highest up to those values, see CPT below).

Here is the original legend from USGS (Link)

Moon CPT

Here is the moon CPT in case you want to see it.

# Default CPT for @moon_relief ...
# See ...
# Original done by ...
# Converted to GMT format by Federico Esteban, March 10 2023.
#
# COLOR_MODEL = RGB
-9200   56/2/123    -9000   56/2/123
-9000   56/2/123    -8000   0/0/89
-8000   0/0/89      -7000   0/94/187
-7000   0/94/187    -5000   60/157/255
-5000   60/157/255  0       215/255/255
0       215/255/255 2000    255/255/255
2000    255/255/255 4000    192/192/192
4000    192/192/192 6000    128/128/128
6000    128/128/128 7000    83/83/83
7000    83/83/83    8000    51/51/51
8000    51/51/51    10000   52/31/31
10000   52/31/31    11000   52/31/31
B   56/2/123
F   52/31/31
N   white

[PaulWessel]

I ould not repeat the last slice. maybe use makecpt to resample their original to the chosen limits?

[Esteban82]

I ould not repeat the last slice.

I repeat them in order to get those values in the colorbar. For the map is the same because I use the extrem color for F and B.

[Esteban82]

Both CPT: Repeating slice (above) and resampling the value (Paul's idea, below).

For me, they look the same, so I would go for Paul's idea (CPT below)

# COLOR_MODEL = RGB
-9200   56/2/123    -8000   0/0/89
-8000   0/0/89      -7000   0/94/187
-7000   0/94/187    -5000   60/157/255
-5000   60/157/255  0       215/255/255
0       215/255/255 2000    255/255/255
2000    255/255/255 4000    192/192/192
4000    192/192/192 6000    128/128/128
6000    128/128/128 7000    83/83/83
7000    83/83/83    8000    51/51/51
8000    51/51/51    11000   52/31/31
B   56/2/123
F   52/31/31
N   white

[PaulWessel]

Yep, cannot tell but good to not add in a constant slice after all the increasing ones.

[joa-quim]

There is one thing wrong with this CPT. We should have a kind of hinge at zero. I mean, the blues should stop exactly at 0 and the gray scale from 0 up. Even if this makes it different from the NASA colormap.

[PaulWessel]

Hm, what does z = 0 mean for the moon? Does it really require a hinge? But it is close so perhaps not crazy.

[joa-quim]

I didn't mean to put a true hinge in the cpt but just to let it be defined like that. There is nothing preventing us to use this cpt on Earth where it likely will also make nice figures.

[PaulWessel]

@Esteban82, do you want to try that and put up a moon map with white at zero?

[Esteban82]

Here are the maps.

CPT with white at 0 used for map below.

# COLOR_MODEL = RGB
-9200   56/2/123    -8000   0/0/89
-8000   0/0/89      -7000   0/94/187
-7000   0/94/187    -5000   60/157/255
-5000   60/157/255  -2000   215/255/255
-2000   215/255/255 0       255/255/255
0       255/255/255 4000    192/192/192
4000    192/192/192 6000    128/128/128
6000    128/128/128 7000    83/83/83
7000    83/83/83    8000    51/51/51
8000    51/51/51    11000   52/31/31

[Esteban82]

Yes, it is better with the white at 0. In this way, our map also is more similar to USGS maps posted yesterday.

Here I think that there is an issue. With white a 0, the maps look similar but the colorbars don't.

I suppose that:

1. there was an error in the USGS Maps (when they draw the legend and post the "CPT" values) or
2. the grids values are somehow different. The heights of the USGS moon is related to an sphercial moon (the USGS). Ours too or are related to the moon ellipsoid?

In any case, I think we should use white at zero for our moon cpt.

[PaulWessel]

From https://astrogeology.usgs.gov/search/details/Moon/LRO/LOLA/Lunar_LRO_LOLA_Global_LDEM_118m_Mar2014/cub it says spherical moon.

[Esteban82]

Yes, sphercial and both have the same radius (1737.4 km).

[Esteban82]

I read both description of the moon products and the only differences is the resolution of the final grid. So, we can discard option 1 (of my previous supposiotions)

[Esteban82]

I agree we should put white at 0. Now, my doubt is how to adapt the rest of the values? To start, I have to change the value of the 215/255/255 (ca. ligthcyian1 for GMT). In the previous map I use -2000. Also, I wonder if we should adapt the values for the gray colors?

Original NASA CPT with comments:

Elevation_m Red Green Blue
-9000.0 56 2 123
-8000.0 0 0 89
-7000.0 0 94 187
-5000.0 60 157 255
0.0 215 255 255           # Lightcyian1. At what value should it go??
2000.0 255 255 255     # White, must go to 0! 
4000.0 192 192 192     # Keep the same value or change it for a lower one (e.g. 2000)?
6000.0 128 128 128
7000.0 83 83 83
8000.0 51 51 51
10000.0 52 31 31

[Esteban82]

Here I made another maps from differentes CPTS

Without cyan:

With cyan at -1000

With cyan at -500

[PaulWessel]

I have no strong opinions on this so suggest you and @willschlitzer can decide what you prefer, then go with that.

[Esteban82]

I wonder if @joa-quim has something to say.

[Esteban82]

I will use the second CPT. Below a copy it.

Is the header ok? Line 2 is too long but is where the "cpt" was posted. I think to use this link instead (https://pubs.usgs.gov/sim/3316/) but it has no information of the colors.

# Default CPT for @moon_relief.
# Original table of colors post it in https://astrogeology.usgs.gov/search/map/Moon/LMMP/LOLA-derived/Lunar_LRO_LOLA_ClrShade_Global_64ppd_BlueSteelOriginal
# Colors modified and converted to GMT format by Federico Esteban, March 15 2023.
#
# COLOR_MODEL = RGB
-9200   56/2/123    -8000   0/0/89
-8000   0/0/89      -7000   0/94/187
-7000   0/94/187    -5000   60/157/255
-5000   60/157/255  -1000   215/255/255
-1000   215/255/255 0       255/255/255
0       255/255/255 4000    192/192/192
4000    192/192/192 6000    128/128/128
6000    128/128/128 7000    83/83/83
7000    83/83/83    8000    51/51/51
8000    51/51/51    11000   52/31/31
B   56/2/123
F   52/31/31
N   white

[joa-quim]

Sorry, had no time to look to this. My preferred would be the first (no cyan) but no strong opinion either.

No problem with that second long line. It's a comment line.

[Esteban82]

How can I test the moon_relief.cpt from the test chache server (/export/gmtserver/gmt/test/cache)? I upload the cpt file to the directory but it didin't work.

I supossed that I have to update the gmt_data_hash.txt. I try something but it didn't work.

The error that I got:

grdimage [ERROR]: Libcurl Error: HTTP response code said error
grdimage [ERROR]: Probably means @moon_relief does not exist on the remote server
grdimage [ERROR]: Unable to obtain remote file @moon_relief

[PaulWessel]

Not sure if cache testing works that way. Just add it to the regular cache area since we are going to do that anyway.

[Esteban82]

Now it is working.

[Esteban82]

Just add it to the regular cache area since we are going to do that anyway.

Just to know.

In order to add data in the cache, is it only necessary to add to the server? Or is there something else that need to be done (like updating the cache file)?

[PaulWessel]

Just add it to your working copy of the repo, just like you do when you add a file to the gmt repo, then make a PR to add the file, someone approves, and within 1 hour (I hink) the gmtserver will check and detect there have been changes, it will do a git pull and the new file will be in the cloud. So give it 1 hour or so after approval/merge for it to show.

[PaulWessel]

Hi @willschlitzer, could you please ssh into the gmtserver and run chmod -R g+w on the moon directory under the test server, i.e.

chmod -R g+w /export/gmtserver/gmt/test/server/moon

Without write-permission for the group the rest of us cannot do much there. We are inching closer to the releases.

[Esteban82]

It was already implemented.