PDA

View Full Version : Elevation, Weather Charts for Entire Fictional Planet



Podcreature
09-28-2013, 07:58 AM
So I’m making a map of Pern, as from the novel series “The Dragonriders of Pern” by the late Anne McCaffrey. *Hat-tip to the first woman to receive the Hugo and Nebula awards for her science fiction.*

Okay, so I’m a big fan and I’m not a map-maker, but I need a map to suit my purposes. We'll just see how it turns out in the end! I’ll share what I come up with as I go, and list some of my main goals, and proposed methods for how I might do things. Whatever anyone wants to throw out there, advice or criticism, is more than welcome. If there’s a better way to do something, I’m likely to be ignorant of it, lol.

Realistic stuff:
Firstly, my goals and needs are to create a detailed height map and also weather charts for an entire planet.
That being said, if anyone knows of some good climate maps, or guides, I'm new and need to learn as much as I can... :)
Canonical references for the map I'm making can be found here: The Pern Museum/Maps and Charts/Maps from the Books (http://www.agriphoto.nl/pma/Maps%20and%20Charts/Maps%20from%20the%20Books/index.html) and this section has some exciting maps from the book "The Atlas of Pern" which I think many cartographers would have fun using as reference to make an updated versinon. http://www.agriphoto.nl/pma/Maps%20and%20Charts/The%20Atlas%20of%20Pern/index.html I know I'll be needing these.
I plan to use Wilbur. I want the mountains to appear to be the realistic scale, not stylized and over-sized. I'm new to this software, and I don't really know what to expect. (Wilbur does work fine with windows 7??)

Style stuff:
I want to create a more simple map for demographics, roads, labels etc. Marrying the realistic terrains I plan to generate with a stylized map will be tricky for me, as I want the theme to be unified and cohesive, but when it comes to presentation and styles, I don’t know much. So suggestions for fonts, effects, presentation will be super helpful as I make updates.

Tools: Photoshop CS6, Wilbur, and later, possibly other freeware as I learn about them.
Right now the map project is 23760 x 11880 pixels. I need to learn more about file size... whether this is right for a detailed map of a whole world, I don't know. I definitely want lines and texts to show up crisp and clear even when scaled out. Still newish to Photoshop so here's hoping I make the best use of it!
Well that’s enough for now, feel free to take a look at the base map I’ve created from the canon ones in the books using the same longitude and Latitude lines.

Update 9/29/13: Better template image for building my heightmap:
58033

This is the old map I made, using Gimp only:

58009

Thanks for reading. :)

Podcreature
10-02-2013, 12:48 AM
Ok, first off, I have to say that I'm so happy I found Wilbur. Knowing that what I'm trying to do is even possible has been like Christmas every day since I started working on this project again!

So there's a long way to go, but I thought I'd share what I've learned so far.

I had two methods for creating my map. I could have hand drawn a heightmap, and brought it into Wilbur and gotten pretty good results. I should never have looked at NASA's heightmap of the earth. It was super detailed, and I could zoom right in and look at all the flowing, sweeping land forms, and I really didn't think I could get such diversity in Wilbur.

58108

So, I went on impulse and made my heightmap from bits of Earth, overlapped them, warped, skewed, flipped, rotated them until it looked different enough.

58109

58113
(just a test, still exciting)

But it wasn't really Pern, it was Earth and that was bugging me.

So I blurred it and got something like this...

58110

Then in Wilbur generated something like this,

58111

Not bad, but I decided to keep the detail and made this from a non-blurred piece,

58112

I'm excited about this stage because it's really become it's own planet. Any parts that looked hodge-podge are fully merged. Still playing with different techniques to get the size of mountains and rivers I want though. Lots of trial and error.

I plan to overlap these, and use some pieces from each to make the final version. I'm going to do the oceans separately and overlay it. To make the mountains and sea level areas more dynamic, I may use curves to select only certain ranges in Photoshop to darken them or lighten the grayscale heightmaps. Then I can bring it all into Wilbur one more time and raise lower a little more. Hold on to your butts!

Viking
10-02-2013, 01:59 AM
I like where this is going :) Please keep working on this and keeping us updated!

arsheesh
10-03-2013, 12:37 AM
This looks fantastic! You've done a splendid job combining bits and pieces of height maps of earth to form your own height map. I'm impressed!

Cheers,
-Arsheesh

Podcreature
10-05-2013, 09:19 AM
Thanks so much for the encouragement guys! It's helped, particularly when I forgot to save a grayscale verison of the map, and closed Wilbur(I need both the gray and the elevation .png, matching.) I had to re-do the region just when I was feeling proud of it. They "joy" of Wilbur, I'm learning, is that if you're new, there's so many numerical settings to memorize, particularly with "incise flow." If you're not the technical sort, ahem, you'll spend days trying to remember what you did to repeat an effect... Headaches aside, I've established steps that I'll be able to repeat with all the other regions. So hopefully I'll be able to update again soon with some big progress. :D

It's not a big update, but proof I'm making it past the humps, and because pictures are fun...

Here's the region before Wilbur:
58196

Here's the map I was "proud of." (The ocean has a weird growth - it doesn't matter because I'm masking out the ocean anyway.)
58195

And here's what I was able to recreate. It's not the same but I accepted it because it looks more realistic to me. Maybe not as dramatic in the same way, but there's more... diversity? Thoughts and opinions welcomed. ;p
58197

The grayscale version looks a little wacky, but I'm sure it will look much better when colored.
58198

Finally off to repeat this with all the other regions.

Thank you Arsheesh, for your super helpful Eriond tutorial, it's what got me going in the right direction. And Viking, your map of Skenth was one of the first I found when I stumbled on the forum. It's amazing! Glad there's competent mappers here to supervise my novice attempt, lol.

Update:

58199
Some un-Wilberized map bits look better when brought back in.

Nolan75
10-05-2013, 10:12 PM
Just arrived on this forum and this looks great, hope I can do stuff like this one day xD

Thorf
10-09-2013, 05:22 AM
Very interesting... I am working on a similar kind of project, mapping out a whole world using Photoshop and Wilbur, like you. And I too began with the continental outlines/sea mask, and decided to build a height map in Photoshop for further processing in Wilbur. I considered using real world data, as you have so effectively demonstrated, but ended up going with my own model based on rendered clouds in Photoshop, with lots of masks and level adjustments.

In any case, that part is different, but the erosion part is very much the same. I'd love to hear what settings you're using for erosion. I have experimented with various different "recipes", using both precipiton erosion and incise flow, noting the settings each time. But so far I've only found one set that gives me the kind of results I'm looking for. The results of your process look rather nice.

Regarding the resolution you've chosen, I'm assuming you went with those figures to be on a similar scale to the next generation Blue Marble imagery? I made a similar decision to go with a base 21,600 x 10,800 world map. It's big, but not so big as to be unwieldy. However, I found that the erosion produced by Wilbur does not scale properly - the valleys carved by incise flow end up being far too big, and not detailed enough. I later read on another thread by Wilbur programmer Joe Slayton that the in-built "resolution" of Wilbur's erosion is 1 - 50 metres per pixel. Blue Marble is roughly 2 km per pixel. That's a huge difference in scale, but it's simply not feasible for a world map to be 50 metres per pixel. A roughly earth-sized world would end up being in the realm of 864,000 x 432,000 pixels - ten times the size of the already unwieldy full resolution of NASA's latest Blue Marble!

I have decided to compromise, and go with a full resolution of 86,400 x 43,200 - handily maintaining compatibility with full resolution Blue Marble - but split into continental parts to make it feasible to work on even at that rather massive resolution. I've found that Wilbur works best with pieces of 9,999 x 9,999 or less, so I have split my continents up into four to six parts each, overlapping as you are doing.

I'll be following your project with interest. I'm working on some posts about my own project to post soon, which I hope you might find useful too.

waldronate
10-09-2013, 06:45 AM
It's not so much that Wilbur has a preference for a particular resolution. It's more that the kind of model in use generates results that I think appear truly plausible only on roughly that scale. Note that Wilbur is missing huge kinds of effects that would appear on the micro scale such as meanders or other effects that have to do with density of flow and with mass wasting (critical physical processes for real-world erosion). Wilbur's precipiton thingy is the simplest algorithm that I could get to generate something that looked reasonably plausible to me. Incise flow is just as bad in terms of having a physical meaning...

Podcreature
02-12-2014, 08:32 PM
Sorry it's been so long without an update. I've been working on the masking regions, as well as generating heightmap and elevation maps for each region, which is just a long grind.

I'm having to just guess the southern territories, I think. They don't have to be well-formed because they're essentially uninhabited. Eeach major region was named in "official" maps, but borders weren't drawn, and I assume they are probably somewhat contested, because in the books they were never really colonized, so I've allowed a lot of the terrain to speak for the borders, stopping at barrier ranges and land I can only assume no one wants. I don't really know much about how territory borders are drawn or decided, aside from rivers, mountains, and long./lat. lines. I guess that's all that's typically used though?

I'm also grafting on the detailed heightmaps. I'm leaving unimportant areas mainly unedited.

Planning on using these maps to make 3D terrain which I'll touch up in Unity later too.

This is not the intended appearance, just how I'm picking out the borders based on topology. It will later be colored, with a bump-map on it. I'll make a new map for each region. Again, I need to make a global heightmap, global elevation map, global climate map, global colored map, and demographic colored maps for each of the regions and the whole map... I'll probably have to make closer up heightmaps and elevation maps for each region too, for threadfall charts, but we'll see. That would be more work to throw on an already huge project. :P

EDIT: @ Thorf: Oh, thanks for that info by the way... the reason I chose the resolution I did was simply to keep detail from the original heightmap. Since I'm going to later use these to generate 3D landscapes, I might need as high of resolution as possible. Since I'm not sure how much detail I will need when I get to that point, I would rather work with a big file now than find out later that it's not big enough. I hear you on the unwieldy part. Fortunately my computer can handle chunks of these maps and it's easy enough to overlay them in PS. True that Wilbur doesn't behave dynamically with different file sizes. I need as tiny non-stylized rivers as possible, and I just assumed bigger file, smaller details. I'd definitely love to see what you are doing in Wilbur.

EDIT (2/13/2014): @Waldronate

Note that Wilbur is missing huge kinds of effects that would appear on the micro scale such as meanders or other effects that have to do with density of flow and with mass wasting (critical physical processes for real-world erosion). Wilbur's precipiton thingy is the simplest algorithm that I could get to generate something that looked reasonably plausible to me. Incise flow is just as bad in terms of having a physical meaning...

That's some cool knowledge! I really should study this sort of thing more. Just makes me a little glad I'm not entirely relying on Wilbur at least. I'm trying to let the earth images speak for realistic texture, but then let Wilbur make it it's own land. Makes me worry that the terrain pieces I picked didn't match the climate of the places I was grafting them onto for Pern. Just went with what looked nice. :/

Podcreature
07-27-2014, 10:56 AM
Hey guys, sorry it's been so long with no update. I've finished (enough) using Wilbur to merge terrain, and I've gotten pretty comfortable with it, though there's still some thick river flow zones, it doesn't look terrible with the glacial look. It won't really matter when it's bumped and colored.

Only major stuff left to do is touch-up the borders of the Northern territories, fix up rivers, edit the coastline to check that it's canon since it deformed, and I need to go over the islands with Wilbur. Oh, and I also need to hand-draw major rivers and roads on their own layer.

After that is other political details.


I'd like to render significant areas from this terrain using programs like World Creator, Terragen 3, World Machine and Terrasculptor.. I don't have any experience with those though. I need to settle on one that can give me a large map, with color, edit terrain and something to render it... I'd rather not pay an arm and a leg, and I also really care that it's going to be able to be brought easily enough into Unity Pro (which I am saving for.) I'm kindof thinking I could cut out a rendering program, just buy a terrain editor, and use Outerra, but I don't know anything about how possible or easy it is to get a custom world into that either, I suspect it would be way too technical, but if anyone has advice there, I invite it.

In the meantime, back to the world maps...

6602566026

Naima
07-27-2014, 11:13 AM
looks very nice .
For render you can use Terragen 3 , for Further on regional detailing , you can use World machine to build up further detailed erosion . World machine can create stunning erosion flows but on a more local scale , using it on large worlds ends up with not realistic results .
for small scales I mean something like an island or the like where you can clearly see the river basins etc .

Podcreature
07-29-2014, 02:20 PM
Awesome, thanks for good advice! I will definitely need more localized detail, as I convert the global topology to something someone actually runs around in and on. I'll keep my eye on World Machine, or out for something with that characteristic. Looks like you can get World Machine stuff into Unity: https://www.world-machine.com/learn.php?page=workflow&workflow=wfunity, so that's good news. I already liked WM for the price tag too, lol but without being able to compare these all on my computer, it's really helpful to hear endorsements from people who tried them or know their finer points.

Podcreature
09-11-2014, 12:58 AM
So I am mostly finished with the elevation map. The sea is still a mockup, I haven't even tried to think critically about the continental shift yet... It's a fan-map so I was more worried about putting mountainas and valleys where the canon maps say they should be. So the final step will be attempting to make scientific sense of it.

Any suggestions or critiques about what I have so far will be very much appreciated.

Need to use the band-aid to fix some unnaturally straight edges along the river basins. I did decide one thing about those steep eroded trenches, which I'd failed to soften in Wilbur. At first they bugged me, but, they may not be so unrealistic for this particular map, considering the alien menace "thred" began to fall between thousands and millions of years ago, and the forests were pretty much wiped out by the time colonists arrived. I read about Iceland: a mostly barren landscape that was also once covered in forest. It's experienced severe erosion that's had a great impact on it's terrain. The erosion has actually made large areas of land essentially uninhabitable. On a planetwide scale, I can't imagine the impact all that displaced sediment would have had on ocean ecosystems. I suppose whatever survived would have adapted.

I'm worried the font in the gradient scale is too small.

67245

Podcreature
09-13-2014, 10:12 AM
Here's the ocean currents map. Pyer was a big help, sending me scans from his Atlas of Pern (which I do not own.) ;) Thanks, Pyer!

I didn't want to go against canon, so only made one semi-significant edit. The southern stream, on the East portion of the map, instead of going round and round the south, goes North, and enters the subsidence zone in the sub-arctic seas. This is because, the entire reason the North and South currents go in opposing directions is because it's all one big current going in a loop. Yet the southern current really didn't tie into the north in any of the original maps. Here's a great example of how this works on Earth: www.e-education.psu.edu/earth1…

Feel free to critique and let me know if I'm on the right track or not. :)

67305

Does the map title look like it's in a good place? I centered it before but it just seemed distracting that way. Maybe I need to spend more time on a different design for it. I really was in too much of a rush to get on to the other projects.

Azelor
09-13-2014, 12:49 PM
I think you should have 2 westward currents at the equator. Then the water turns poleward when reaching the westernmost part of the ocean.

like this:

http://upload.wikimedia.org/wikipedia/commons/6/67/Ocean_currents_1943_%28borderless%293.png

Podcreature
09-13-2014, 03:12 PM
I agree, Azelor. I kept thinking that over and over whenever I looked at the chart in the books. I will see what I can do, I didn't actually make up these currents, I am trying to stick to canon, but I will think more on this. Oh, I probably should have mentioned earlier somewhere that (edited) I think Pern rotates in the opposite direction of Earth, but I'm not sure. There's some statements I have to find again. That might affect current, if I'm not just copying an already made map.

ADDED: Maybe I should make a canon map and a non-canon one. Might be neat to have both on my group's resource site for people to think about.

Podcreature
09-16-2014, 01:29 PM
Question: Anyone got tips or links to how to make a sinusoidal or equal-area map from a square map? I'm thinking along the lines of the two halves design. I need this one for threadfall charts to really get a sense of mileage covered by the strips of fall, on a spherical object.

Falconius
09-16-2014, 01:36 PM
Like G. Projector (http://www.giss.nasa.gov/tools/gprojector/)?

Azelor
09-16-2014, 01:42 PM
the sinusoidal map projection looks awful. And if you want to split the map in two, I'm not sure that equal area projections are the best for that purpose. It depend what you have in mind.

Podcreature
09-17-2014, 02:29 AM
Thanks Falconius! That's exactly what I'm looking for! And I'm not worried about looks for this batch. They're purely for function. I need to be able to move a pattern of bands over the planet that are of equal length and be able to see pretty close to the actual surface area they'd cover. If that makes sense. It doesn't have to be *perfect* but more accurate is better. It cannot, however, be too interrupted (like this (http://www.progonos.com/furuti/MapProj/Normal/ProjInt/Img/mp_Sinusoidal-s75-i9.png)) or it will make moving the thread pattern around on the surface too tedious to calculate. I need to make a -lot- of these charts.

So I was thinking something like this: http://www.progonos.com/furuti/MapProj/Normal/ProjInt/Img/mp_QuaAuthalic-s75-I180-s180-R-z20.png .. but, is this not equal-area? Should I aim for at least three lobes?

I could find a happy medium with something like this: http://www.progonos.com/furuti/MapProj/Normal/ProjInt/Img/mp_Sinusoidal-s75-i3.png

I could just try to project all this onto a globe in ZBrush and have a REAL sphere to work with. But projecting these maps onto a ZB sphere with the Spotlight seems like it would be quite tedious.

waldronate
09-17-2014, 04:00 AM
The first interrupted map that you refer to isn't a Sinusoidal one; it's Quartic Authalic (which is indeed equal-area).The general idea of making an interrupted projection is to reduce distortion in areas of interest while putting breaks into areas that are less interesting. The second image that you reference just sort of arbitrarily breaks things at convenient locations. The breaks usually are put in more useful places, such as emphasizing land or ocean. The problem with the map you presented is that it doesn't have any natural break points that would be good for doing large splits (see north and south pole presumed reprojections, assuming that your original map was Equirectangular).
6743567434

A generic split shows some promise, but the split lines would need a little refinement.
67436

Podcreature
09-17-2014, 05:43 AM
This is fantastic, thank you so much... The folks in this forum just blow me away. Thanks for pointing out that the first map is not Sinusoidal. Still learning terminology and I think what I'm really looking for is just an "interrupted" map then? :)

The maps you've showed me look -great-! Now the tricky part for my brain is, to make these "threadfall charts" I need to take a pattern of bands, (like a series of meteor showers that typically have the same configuration) and move them based on the rotation of the planet, as they move in a SE direction - X degrees East and X degrees South.. need to read my notes again to know the exact numbers. Each chart represents a "fall cycle" which is a span of 50 days, before the cycle restarts, probably with some variation in the arrangement or "fall pattern." So the pattern of falls is a long strip with these bands scattered around on it that I need to move around the planet, over and down before each "fall." Fall representing when the strips of meteors actually get close enough to the planet to hit the atmosphere and rain down on the planet below. Hope I'm describing things well. I'm an artist, not a mapper, or a scientist, but this project has been so fascinating. Here's a threadfall chart from the book, which doesn't actually make any practical sense: http://www.agriphoto.nl/pma/Maps%20and%20Charts/Maps%20from%20the%20Books/slides/PernMap005_DDcUS.jpg

With that being said, I suppose I could work with either type of map. But I have an easier time wrapping my head around the second map, being that it's.. horizontal? I don't know the words I need, haha. Still a bit broken up, however. Would putting it into 3 parts instead of 6 cause a lot more distortion? (Some distortion is ok.)

I can't express how thankful I am for any and all input, but do not feel obligated to go out of your way to answer my noobish questions. :) Thank you guys for everything.

I just downloaded G.Projector, and I've got updated Java, but still getting error message on start up, so I'll be trying to sort that out now. Can't wait to make some incredible (at least to me) maps like those you did there, Waldronate. And, I am quite sure that my map is Equirectangular. (I had to look that word up.) These are the maps I've been working from: Pern Map - Current Stage by AMCAlmaron on deviantART (http://www.deviantart.com/art/Pern-Map-Current-Stage-216164719) and http://i121.photobucket.com/albums/o226/Sylthezyne/mergemaplonglatlines_zpse7818756.png~original

Azelor
09-17-2014, 11:50 AM
No, It's not equirectangular because the spacing of the latitudes is not constant.

waldronate
09-17-2014, 01:41 PM
The threadfall map that you linked to has a graticule on in, which makes it easier to get a reasonable reprojection. The two attached images show that map reprojected to equirectangular and to orthographic projections (10 degree graticule, and it looks like I missed the scaling a little bit).
6743967438

Podcreature
09-17-2014, 02:37 PM
No, It's not equirectangular because the spacing of the latitudes is not constant.

Ohh, ok. I see now. Hmm, ok, so I think I follow. I need to convert my original map into being equrectangular if I want to make an accurate split that centers on the North and South poles. Or I could go with the "generic split" that was shown, which I somewhat favor, though the previous is really visually cool. I see what you meant before, Waldronate, that this layout (of continents) is inconvenient, because you can't make an interrupted map without breaking them up.

So.. will an interrupted map in three sections be equal area -enough-? And am I getting you right: G.Projector can convert my map to being equirectangular before it's wrapped around a globe? Is that what you're showing me there? Edit: would an equal area map in one big piece be a better way to go than an interrupted map?

Really enjoying looking at these projections. I'm itching to try this program out.

waldronate
09-17-2014, 05:50 PM
Be warned that nothing I've done here uses G.Projector; it's all software that I wrote long ago (ReprojectImage to get things into Equirectangular and Fractal Terrains to get things into other projections). I've never used G.Projector beyond observing that it uses Java and so can't be installed on my machine. I am of the opinion that Java makes even Linux or Flash look good on the security front and I haven't had it installed since I was forced to in order to pass a few classes back when I was getting an IT degree.

I'm not sure that I understand what you want to do with your map (or set of maps, really). Any equal-area projection is equal-area everywhere, interrupted or not. Interruptions on an equal-area projection are used for minimizing shape distortion, not area distortion. The projection that you pick must ultimately be one that you like and that has useful qualities. And by "useful qualities" I mean qualities that support the purpose for which you are making the map.

Podcreature
09-18-2014, 01:46 AM
Thanks. :) Sorry if I'm not very good at describing things. I think I have learned enough to make an informed decision. In this case, less interrupted is probably better. Two or even one solid part would probably be best.

Podcreature
10-03-2014, 05:54 PM
Quick update:

68125
68126

Looks pretty warped by the poles and a bit in general. Will need to blend the square map better. I think I'm going to try to figure out a way to create threadfall charts from 3D renders of an actual gobe.