I Thought Tectonics Would Be Easy But...

[Umbral Reaver]

I've been working on this world for a while and decided to redo it with more informed accuracy. So, to begin with, what's a realistic placement of mountains given this coastline? I have put in a few tentative plate boundaries. Any help would be appreciated!

Thanks.

[Chick]

If the red arcs are meant to be subduction zones, you have them backwards. A continent will never subduct under an ocean plate, because continental rock is so much lighter and more buoyant. You want to show the oceanic plate subducting under the continental (or another oceanic) plate.

[Umbral Reaver]

Er, yeah. That's what I meant. In any case, trying to fill in all the plates without running into a few with contradictory directions is difficult.

[Umbral Reaver]

A second attempt! I'm still sure there's a better arrangement for the plates than I've presented here, but I'm yet to find it. Here's the blank outline if anyone wants to have a scribble on it!. I appreciate assistance.

[Pixie]

Tectonics is the hardest part...

But it's doable. My first advice is to turn your mercator projection into a equiretangular projection and use map-to-globe to do the scribbling. It will also help to get the program g.plates and to learn about movement in a sphere (particularily, you need to know what an Euler pole is).

As for your map, since the most of the land is nearly fused together, I would consider the planet to be in a just-before-pangaea state, where all the continents are merging. Also, it will be easier if you are open to change the shape of some coasts.. consider that.

[Hai-Etlik]

There are three things to keep in mind about planning tectonics:

It should be done before you figure out the coastlines not retrofitted on to them. If you have coastlines you are better off just using rough heuristics like "mountains tend to follow coastlines".

It should be done on a globe. Maps will lead you astray here. Tectonics is entirely dependant on spherical geometry. If you think in terms of flat plates moving in straight lines on a flat map, your movements will be so wrong that you might as well have just placed things arbitrarily. In particular, it's important to understand that all plate movement is rotation and the precise direction of movement of any point on a plate varies depending on the point in a way that's not at all obvious on a map.

The shapes of plate boundaries depend on their relative movement, again in ways that tend to be hard to understand without a globe. This gets really complicated fast, and again, map projections will lead you astray.

[Umbral Reaver]

I used a fractal map generator to make the coast. Doing it manually is hard and takes time and I am super lazy. D:

[Umbral Reaver]

Okay, so I may be open to altering the coastlines. Or redoing the map entirely. Wilbur is not kind to lakes and inland seas. Or maybe I'm just doing it wrong.

Edit: A brief Wilburing of the land using Arsheesh's method:

The rougher-looking rivers are ones I drew in manually to try to connect lakes to the ocean. If I exclude the water mask from erosion, it produces strange, spiky formations around the water edges.

[acrosome]

There's a group working on a utility that generates random (well, seeded) tectonic plates and runs them through time to generate world maps. It's called WordEngine, and the group is Mindwerks. It looks exceptionally cool, and will obviously be far superior to fractal world generators. They're still in alpha and I can't get it to compile on my linux box. Grr. But it looks like it will be damned handy... some day.

There are various other tectonics-based utilities that precede WorldEngine, one is Lands and there are earlier ones. I can't get any to run, tho.

Regarding Wilbur:

Wilbur is awesome. However, it's creator has always honestly presented it's limitations. It's weathering functions do not work well on smaller-scale maps (that is, maps covering larger areas). They work best on highly zoomed-in local maps, not on whole worlds. Very zoomed-in maps. Like, 10m/pixel maps.

[Umbral Reaver]

I got WorldEngine working! The downside: It takes ages to generate a map (especially large resolution ones), is not scale nonvariant (i.e. the same seed at different sizes produces different maps) and there's no way to adjust the generation parameters (water/erosion amounts, etc.) that I can see.

It appears to run only on a single core, so I might be able to get several worlds to generate simultaneously.