New World Ocean Map on Web Site

[quadibloc]

Since there are problems with putting images in posts, or at least there were the last time I tried, I have to plug my web site.
In the discussion of August's Conformal Projection of the World on a Two-Cusped Epicycloid, I used G Projector to make a physical map of the world ocean following the aspect used by Athelstan Spilhaus.
However, I was dissatisfied with what I produced; while I used the best land and ocean shaded relief equirectangular base map that I could find, what I really wanted was one that showed the relief of the ocean floor a bit more dramatically...
And, needless to say, what I meant by "more dramatically" was "as much like that famous map by Heezen and Tharp as humanly possible".
Well, on Wikimedia Commons I found a more suitable equirectangular base map that did come much closer to what I wanted, and they claimed it was a public-domain product of the U.S. government.
So here is the page, and the location of the image to which I am referring should be obvious:
http://www.quadibloc.com/maps/mcf0702.htm
I've found a source for the base map I used; it's on NASA's web site instead of that of the USGS, which is the original source.

[Milo]

Huh, I've actually been using that same base map for a while. I managed to forget it didn't actually come standard with G.Projector. It's a nice map, although it has some visible artifacts near the cutoff meridian (use any cylindrical/etc. projection with nonzero central meridian and graticule turned off in order to highlight it), and despite the focus of bathymetry, some parts of the mid-ocean ridge system are rather subtle. Overall it's still better on land - stuff like the Tibetan Plateau comes out extremely crisply.

The choice of aspect is interesting. Using a transverse map centered almost but not exactly on the south pole for an ocean map, I don't think I've seen that one before, but I can see the appeal. It neatly splits the three major oceans (Pacific, Atlantic, Indian) into three recognizable "lobes", though it does mangle the South China Sea. I think the map could be improved by adjusting the mangled point north a little so it's actually in China (about 115°E 36°N looks good - you should understand why from looking at this).

Can't say I'm really fond of the August epicycloidal projection, but it's nice to see some atypical use cases, and trying to show the entire world ocean with minimal interruptions is an underappreciated challenge.

[quadibloc]

I don't think I've seen that one before, but I can see the appeal.

Well, it certainly isn't commonly used. Aside from Spilhaus' original use, it appeared at least once in Scientific American, and that's where I first saw it. I also later saw the original appearance reproduced in textbooks on cartography.
While August's conformal projection looks better for the world in standard aspect, for an unconventional map like this, the Eisenlohr would be a better choice. However, that projection is difficult to invert, since its forward version is the sum of two projections. Therefore, it's not available on G.Projector.
I've found the original online, so you can see it here:
https://www.geographyrealm.com/the-spilhaus-endless-ocean-map-projection/

[quadibloc]

I've made another update to my web site. It turns out that Justin Kunimune has made additional improvements to his program Map Designer, and by popular demand the Dixon conformal octahedral projection now produces output in the form of the Cahill butterfly directly! Yay!

[Atarimaster]

And, needless to say, what I meant by "more dramatically" was "as much like that famous map by Heezen and Tharp as humanly possible".

When we talked about the Heezen-Tharp map last year, I mentioned that I, too, would love to have a map like that. Recently I tried again to approximate that map’s dramatic look, but I kept failing. It then occurred to me that the reason for my failure was that for the sea floor, the Heezen-Tharp uses a representation that resembles a panoramic map rather than a shaded relief.
(I hope “panoramic map” is the right term here, I’m meaning something like this).
Since the software I used “only” created shaded relief maps, I gave up on trying.

However – yesterday I downloaded TopoToImage that is introduced in the thread next door, and I was able to create a map with a sea floor that, of course, still is no panoramic map, but looks more dramatic than anything I’ve created so far.

Here‘s a scaled-down version of the full map and an excerpt at original size (5400 × 2700).

(And yes, dropping shadows in the deep sea is unrealistic, but this image is not about realism.)

[quadibloc]

This is wonderful.
At first, I thought I would need to have Geocart in order to use it, but I see that it merely generates image files, which Geocart and other applications could use.
However, although they say that the source can be used in Python from Windows, they did not give instructions on how, from Windows, one addresses the dependencies of the package. There is no "pip" command in Windows.

[Atarimaster]

There is no "pip" command in Windows.

AFAICS you’ve got to install pip before you can install TopoToImage:
https://github.com/pypa/pip

[quadibloc]

Thank you very much. Now, thanks to you, I know I need to type

py -m pip install -r requirements.txt

and not just

pip install -r requirements.txt

...I think. The TopoToImage page on GitHub just didn't consider that dense Windows users like myself would be using it.

I tried, it almost worked. So now I'm working on installing and running gdal-installer. That didn't seem to work, but there are other options for installing GDAL. However, I'm beginning to suspect that the only real way to get this working is to install the entirety of OSGeo4W, which I am loath to do on my tiny laptop.

[quadibloc]

Recently I tried again to approximate that map’s dramatic look, but I kept failing. It then occurred to me that the reason for my failure was that for the sea floor, the Heezen-Tharp uses a representation that resembles a panoramic map rather than a shaded relief.
(I hope “panoramic map” is the right term here, I’m meaning something like this).
Since the software I used “only” created shaded relief maps, I gave up on trying.

I was a little puzzled by this.
However, I ran across a site which noted that Heinrich C. Berann, who painted the Heezen-Tharp map, was the "father of the modern panoramic map". However, what they were referring to was paintings of a landscape from an angled view rather than from a vertical one. I don't think the world ocean map is one of those.
However, in my recent searches, I've come across a mention of another kind of map which is kind of in between a conventional hillshade map and a panorama. It is called "plan oblique relief". The map is still laid out as a map, without perspective, but elevated regions are displaced in proportion to their elevation in addition to shaded relief being present, so that while you are looking at the land from above, you are seeing the mountains and valleys from the side.
EDIT: I didn't think that the Heezen-Tharp map was of this type, but I was just now looking at a video on how to make such maps in ArcGIS Pro, and the narrator specifically mentioned that the ocean floor maps (also by Berann) included with issues of National Geographic were of the "plan oblique" type.

[Milo]

However, in my recent searches, I've come across a mention of another kind of map which is kind of in between a conventional hillshade map and a panorama. It is called "plan oblique relief". The map is still laid out as a map, without perspective, but elevated regions are displaced in proportion to their elevation in addition to shaded relief being present, so that while you are looking at the land from above, you are seeing the mountains and valleys from the side.

That sounds a lot like how a lot of oldschool 2D video games look!

Meanwhile, the map Atarimaster posted above looks like if someone made a map with 3D relief as a physical scale model, and then took a photograph of that scale model.