Roshar Map Reimagined

[Otto Didact]

A little while ago I noticed that the map of the Rosharan supercontinent had latitude lines called out along the sides, so I decided to digitize it and make a "modern-style" map of Roshar. The map I ended up making is reprojected into an orthographic projection (meaning the map seems like we're looking at the planet from space). I really had a good time making it and wanted to share with the community. Enjoy!

PS - I checked with the Inestimable Peter Ahlstrom and he said my placement of the landmass on the latitude/longitude lines was "very close to spot-on."
 

[StrikerEZ]

That's actually really cool! Upvote for you!

Have we always known that Roshar has been that south? Makes sense with the Frostlands, I guess, but I can't remember if there's a WoB about it somewhere.

[Otto Didact]

As far as I'm aware, it's pretty common information that the Rosharan continent is almost entirely in the Southern hemisphere of the planet.

[StrikerEZ]

4 minutes ago, Otto Didact said:

As far as I'm aware, it's pretty common information that the Rosharan continent is almost entirely in the Southern hemisphere of the planet.

Well, whoops. I figured it was. 

 

[Otto Didact]

No worries. There's always new tidbits that I find out, only to realize that most people already knew it way before. I love seeing just how far south it really is!

[Jofwu]

@Otto Didact, do you have or can you make an equirectangular projection of the map?

Edit: You inspired me to make a globe, using this cool site I found: https://www.maptoglobe.com/r1hnSN7VZ

It's pinched at the poles. Didn't unfold it properly. WoP is that the Roshar maps are azumithal equidistant, I believe. (did you ask him that?) And the website needs equirectangular. I didn't really have the time to find a way to stretch it out properly. Shallow Cripts should be slightly further west than the eastern edge of the mountains (as you've shown it) but on my globe it's a bit further west than the Shattered Plains. And the overall longitudinal scale is off in the first place. My Roshar wraps around about 150 deg while yours wraps closer to 180 it looks like.

But it was close enough to make me happy.

If I have time I want to properly convert the map somehow (if someone with Matlab wants to help out...), get the scale right, and make a color version.

[Otto Didact]

13 hours ago, jofwu said:

@Otto Didact, do you have or can you make an equirectangular projection of the map?

I'd like to make an equirectangular world map in the style of classroom maps of Earth if I have some time. I also have vague plans to make an interactive globe using Javascript + D3 at some point. I've attached a very early, badly symbolized version of an equirectangular projection. It's not very good stylistically, but it is projected correctly.

13 hours ago, jofwu said:

WoP is that the Roshar maps are azumithal equidistant, I believe. (did you ask him that?)

Yeah, that was part of what I asked when I was getting the positioning and scale correct. You can see below the positioning that I showed Peter on Twitter when I got his approval. Also, I have shapefiles that I'm happy to share with anyone who wants to play around with them.

[soulcastJam]

So, do we know that this is the only continent? Is there, perhaps, another continent waiting to be discovered? 

[Jofwu]

WoB is that there could be smaller islands and that sort of thing. But not another continent.

[Airhawk360]

This looks awesome!

[Jofwu]

Updated my globe: https://www.maptoglobe.com/ryovymN4-

The projection "fix" is still a rough hack. Don't have the tools or time to do better. But it's closer than what I made before, and in color! (One interesting observation is that the azumith lines on the color map don't appear to be correct.)

@Otto Didact, I think you have it stretching too far longitudinally. In my search for a method to fix adjust the projection I came across this post: https://www.cartographersguild.com/showthread.php?t=17310. Note the username of the poster.

So unless things have changed (the 60 degree number looks accurate still, so I doubt it) the continent is 120 degrees across. Yours looks more like 150 degrees. (excluding Aimia) Someone on Discord also had a good point: polar caps?

@PeterAhlstrom, any chance you'd care to comment on the 120 degree number or the possibility of ice caps?

[Luck Spren]

This is amazing @Otto Didact I think I'll be using this map from now on!

[Otto Didact]

23 minutes ago, jofwu said:

@Otto Didact, I think you have it stretching too far longitudinally. In my search for a method to fix adjust the projection I came across this post: https://www.cartographersguild.com/showthread.php?t=17310. Note the username of the poster.

So unless things have changed (the 60 degree number looks accurate still, so I doubt it) the continent is 120 degrees across. Yours looks more like 150 degrees. (excluding Aimia)

 

Hmm, interesting. I mean, all I did to make the alignment was to resize the existing map so it matched up with the latitudes, presuming that the longitudes would line up as well. In order for the continent to fit within 60 deg. of latitude and 120 deg. of longitude, I would have had to squish the original map, changing the aspect ratio. Weird.

23 minutes ago, jofwu said:

Someone on Discord also had a good point: polar caps?

I think someone mentions somewhere in the books that if you go south far enough, the sea freezes, so I'm sure there are polar ice caps. I didn't feel like speculating as to their shape for this map, though.

 

EDIT: Also, someone pointed out to me that I misspelled Thaylenah on the original map, so here's an updated one (web resolution again—if anyone wants one in print resolution, feel free to ask).

Edited June 30, 2017 by Otto Didact
Added updated map.

[trimeta]

It's interesting to see Roshar in the context of a spherical world, but since it was very carefully designed to be a specific 2D slice of a 4D Julia set, any transformation which deforms it would make it inaccurate. For normal maps, the 2D representation is a distorted version of the 3D reality, but in this specific case, the 2D representation is canonical.

[Jofwu]

Eh... I don't think the Julia set thing is meant to be taken that far. It's not like it's a 100% direct representation of land vs. sea on the map.

[PeterAhlstrom]

I have an equirectangular map and a gores map, but they're not something we're prepared to share around.

[Jofwu]

@Otto Didact Really good new info here: https://www.reddit.com/r/Stormlight_Archive/comments/7by9pu/oathbringer_megathread/dpvedpn/

Newer map with better lat/long lines is here. Doesn't have the equator marked though, so you have to cross reference with other things to pin that down. Anyways, important point from Peter that they use 200 "degrees" longitude and 100 "degrees" latitude.

Shouldn't change your proportions of course, but I think that would make the continent a tiny bit smaller than you have it? If you assumed the longitude lines were measuring X degrees out of 360 while they were actually measuring X half-"degrees" out of 400, then I think you'd get something 10% larger than it should be. It would wrap around the planet a bit less and the southern regions would be pulled up a little bit.

[Otto Didact]

Ooh interesting tidbit. I kind of assumed that they'd have a system based around multiples of ten. I'll have to play around with the new grid and see what I come up with. I was going on the assumption that 90 degrees was the south pole, but in Rosharan degrees it's 100, so yeah, the continent is likely a bit smaller than what I came up with before. I'm a little confused by the canon grid, though, because in an azimuthal equidistant projection the equator should be a straight line rather than a curved one.

[Jofwu]

7 hours ago, Otto Didact said:

I'm a little confused by the canon grid, though, because in an azimuthal equidistant projection the equator should be a straight line rather than a curved one.

My interpretation here was that this map isn't azimuthal equidistant. I just assumed that any distortion relative to the other projections was subtle enough that I didn't notice. Would be interesting to overlay them and see if there are indeed some distortions I guess?

I'm curious what you come up with!

[Harakeke]

That's some pretty Awesome work!

On 6/30/2017 at 8:40 AM, Otto Didact said:

Also, I have shapefiles that I'm happy to share with anyone who wants to play around with them.

I'd love to mess around with the shapefiles if you get a chance. I've always wanted to take a crack at georectifying the various map images, but wasn't sure what projection to set the data frame in.

Any thoughts as to the blue and red markings on the Frostlands map? Do they match up with any interesting lines of lat/long?

[Jofwu]

@Harakeke I don't have the skills/knowledge/programs to properly adjust the Frostlands projection, but comparing them side by side it does look like some of the lines correspond to lat/long.

The three dark vertical lines definitely seem to be longitudes. As best I can tell, the points they pass through match what we see in the new map. And the numbers along the top and bottom seem to correlate. I count 5 "squares" horizontally between the longitude off the east coast of New Natanan and the longitude off the west coast of that unnamed island. I count another 5 "squares" horizontally from there to the very narrow inlet/river. This matches the numbers, putting those lines 10 units apart each. So each "square" on the big map is two "degrees", or else the numbers on the Frostlands map is counting half-"degrees".

The dark mostly-horizontal lines are roughly perpendicular, so it's no surprise those match expected latitudes. New Natanan's latitude does correspond with the northmost point of the Shattered Plains for example. Shallow Crypts is just barely north of the large inlet.

These dark perpendicular lines vaguely form squares, so if you assume latitudes mark off angles of the same magnitude, Shallow Crypts to New Natanan is maybe 17 ticks? On the new map I count 7.5 "squares" between them, so that doesn't quite match up. More (and different) error if you measure from Shallow Crypts to the southern tip of the Frostlands. (about 20 ticks? versus 5.5 "squares"). We don't have numbers on the Frostlands map latitude lines, so I guess we just have to assume they aren't spaced equally.

[Otto Didact]

I'd say the numbers on the Frostlands map are almost certainly Rosharan degrees of longitude, but I haven't fiddled with the new info yet so I can't be sure. I originally tried to get them to line up with Earth degrees and couldn't quite make it happen. It does also seem that the Frostlands map is in a different projection than the continent map—when I georeferenced the Frostlands map to line up with the Rosharan one, it deformed it a fair amount.

I'd be happy to send you the shapefiles I currently have, @Harakeke, but I am planning to fix them based on the new info we have from Oathbringer and Peter, so you may want to wait until that happens. Your call.