Nikolai Lofving Hersfeldt
patreon
Teacup Ae Terrain Preview
🕑 Added 2024-05-08 08:47:37 +0000 UTCBit of a preview of the terrain work I've been doing. This is still wip; each continent has been run at 3 km/pixel resolution for 400,000 years in gospl with uniform precipitation and no uplift. I ran them in individual hammer projections as per my tutorial, but stitched them together here. I haven't touched the seafloor much, you can see the seams between the continent sections if you look closely. You can also see my attempt to make something vaguely resembling mid-ocean ridges, which I don't expect to hold up well on close inspection but gives the seafloor a bit of texture.
Now I'll take the terrain and run it through some climate modelling and then some ice sheet modelling with PISM (parallel ice sheet model), which will help inform some final runs for another 120k years, which will also include some tweaking to encourage lake formation and add in better coastal and river delta landforms and hopefully a slightly finer resolution. I'll make some tutorials out on the whole process once I've figured it out; I'm also working on my own code to work out proper lake levels and river discharge from the final data, as no existing software seems to quite meet our requirements there, and take a whack at some kind of "true color" mapping. Maybe once it's all done I'll have a few posts just exploring the different resulting terrains and my overall choices in making them.
Comments
Nikolai Lofving Hersfeldt
It's been a while since I did those first tests but I don't remember making much of a point to directly compare that.
Nikolai Lofving Hersfeldt
As with many glacier models, IGM seems focused on modelling glaciers at small scale over regions small enough to have negligible climate variation; I picked PISM specifically for its focus on whole ice sheets over continental scales; and it also models isostatic rebound. gospl managed lakebed sediment on its own, but that can tend to fill in whole lakes quite quickly unless you attempt to add in some kind of uplift and subsidence, which I will experiment with. The model I want to build is just for tracking water flow, essentially.
Will B
Looks awesome! Not sure how far along you've gotten with PISM but I recently found the Instructed Glacier Model (https://github.com/jouvetg/igm) which approximates 3D glacier evolution like a full-Stokes model but has a pretrained ML component which supposedly greatly speeds up computation. And it has a built-in flexure module via gflex, with might help simulate isostatic rebound. Just a thought! Also curious about how you're thinking of handling lakes. With your custom method will you be tracking sediment deposition onto lakebeds?
Seristal
I recall you saying something about the benefit to arid areas through the use of uniform precipitation over the EPS precip maps. Conversely, did you notice much of a difference in the terrain that did receive more regular rainfall when comparing them to uniform and EPS maps?
