Well, that was a bit of a longer pause, than I had hoped. So much for my “best intentions of documenting my progress”…

But I’ve finally managed to implement a working prove-of-concept for the plate simulation to generate terrain, also it’s still a bit rough around the edges.

While there are still many problems that need more work^[1], the general approach appears to work, at least.

But one thing that I noticed now is that the current tooling to visualize the results is not enough to test and debug the algorithm. So, before I continue working on the algorithm itself, I’ve decided to refactor and improve the editor to better visualize the generated data and allow users to modify the intermediary results. But first, it’s probably a good idea to document how the current implementation of the plate simulation works^[2].

The first aspect we will have to take a closer look at, is how we can generate our starting point, i.e. the initial state before the simulation is run.

We’ve already discussed the first half of that in a previous post, where we’ve generated a simple triangle mesh of a sphere. So, I’ll first recap the most important part of that and then dive into how we can populate that mesh with the information we need for the actual simulation, like plate-IDs, plate-velocities and elevations.

Now that we have a way to generate some (albeit a bit dull) tectonic plates, we can use an iterative algorithm that simulates plate movements and interactions, to refine it into (hopefully) realistic looking landforms.

Because the simulation itself is relatively complex^[1], I’ve split it into three separate posts:

• Moving the plates and determining how they should interact with each other (this post)
• Updating the mesh after vertices have been moved, which includes the creation and destruction of crust material on plate boundaries, as well as detecting and handling collisions between plates
• Simulating large-scale plate interactions like rifting and suturing