Thanks so much for taking the time to go through the setup and for all the great suggestions. I’m really happy to have you involved in this discussion. It’s super helpful and motivating.
About the parameters:
The very low prefactors (around 5.5e-51) I used were mainly to suppress diffusion creep in the continental and oceanic crust, since it didn’t seem very relevant based on what I saw in a few papers.
But I really like the values you suggested. I’ll definitely give them a try (for the mantle) and see how they affect the model.
As for the boundary velocity setup, I’m currently running the model with tangential velocities applied on all boundaries. That said, if you have an example of how to properly set up an outflow boundary, that would be really helpful! Also, I’ve removed the free surface section.
For the nonlinear solver, I’m using Iterated Advection and defect correction Stokes.
I’ll attach some snapshots of my current .prm file updated with some of your suggestions.
I haven’t yet run the model long enough to get really presentable results; things like plate movements and collision patterns aren’t showing clearly yet mostly because the runtime is quite slow.
If you have any advice on speeding things up, that would be awesome!
I may be able to help you with the world builder part of your question.
In your case there may be an easier approach than variable thicknesses to get rid of the overriding plate lithosphere below the subducting plate. You can just make the subducting plate thicker and add a layer to the subducting plate, which sets it to mantle composition and temperature.
If you want to use the variable thickness (which I guess is the cleaner solution), that should also work.
Since I see you use a value of 660e3 in the world builder input file, I just want to make sure that you know that the coordinates you input are always in map view. The depth is a separate parameter. See Adding your first tectonic feature — Manual GWB 1.0.0. I would also take the cross section in the center of the 3rd dimension and not at the edge to prevent rounding errors: "cross section":[[0,0e3],[4000e3,330e3]], (see 2D models — Manual GWB 1.0.0)
You probably want your max depth to look something like this:
What this does is it sets the boundary coordinates to 140 km depth. Next you define two coordinates at 1200km x (spanning the 3rd dimension direction) to be also 140km deep. Next you define at the end of the plate (where the slab is) that the thickness should be 35km (or 0 if you want it to taper to 0.
Thanks a lot for your suggestions on the world builder; they were really helpful!
I tried using the "max depth" structure you suggested, and even after adding the missing brackets, I couldn’t quite get it to work right. But your advice got me on the right track . I ended up reproducing the model from the GWB tutorial and tweaked it to fit my case.
Here’s what I used for "max depth":[
[100e3],
[140e3, [[1300e3, 0], [500e3, 0], [0, 1000e3]]]
]
I’ll attach a snapshot of the result.
Right now, I’m still trying to find ways to speed things up . After almost two days of running, I’ve only got the first output. If someone’s got any performance tips, I’d really appreciate them!
