Landmass shrinkage

[TheNewParad1gm]

I've noticed that the amount of landmass always seems to considerably shrink over time.

At about 4Gy it's easily less than half what it was at the outset, and I reckon quite a bit more than that. I always end up with something like the attached

I'm sure I read in a response elsewhere that continental crust amount was meant to be preserved?

[redferret]

This doesn't appear to be an issue. The first setup of the planet is give it enough mass on the surface to work with. If the planet started out completely flat and bare it would take too long to get continential shields to form. If you really look at it your example shows a planet similar to Earth in terms of land coverage and ocean coverage. There appears to be a large continent submerged under water but this is still a continent and it's pretty cool you got that result. The North American continent was like this for millions of years.

[davidson16807]

We try to conserve continental crust as much as possible, but IIRC there are still leaks. Each plate is a grid with a slightly different orientation so the cells in the grid don't always line up against each other. Processes like erosion are stored using a single global raster since we need to transfer mass across plate boundaries, but that means we have to somehow distribute the mass amongst plates, and if one cell maps to several you might see gradual mass loss.

@redferret might be right, though, since it looks like that planet from the picture is really super eroded.

There's a way to test this. When the model starts up we store a variable that tracks how much mass we ought to expect per cell for the rest of the simulation. If you open up devtools while running the model you can see it by typing:

sim.focus().lithosphere.average_conserved_per_cell

Then you can compare it to the equivalent amount for the current point in time:

Crust.get_average_conserved_per_cell(sim.focus().lithosphere.total_crust)

If those number don't match then you have some sort of conservation error.

Anybody willing to give it a shot? I've been super intense with the C++ port lately and don't want to get side tracked.

[TheNewParad1gm]

I'd be happy to give it a shot. It really does feel to me that after a good while I have less continent. But, it could just be a trick of the mind!

[aobunau]

well i'm 2 years late to the party but you're absolutely correct and it is no trick of the mind you can see the result of 2.6 billion years of simulation here however, my starting heightmap is this image: you can see just how much land was lost between those 2 images and finally, i put those 2 lines into devtools and got this

i didnt really want to create a duplicate issue but at the same time i'm not really sure if this project is abandoned or not, i just hope somebody will fix this someday because it's a total dealbreaker

[ininja-ininja]

I think that if hotspots and oceanic plate rifting were added, we would see an increase in land and not a decrease. Thanks to oceanic plate rifting and hotspots we have on our Earth, the amount of landmass is changing, earth once had no continents and was just a water world but now has some large landmasses

[ininja-ininja]

This doesn't appear to be an issue. The first setup of the planet is give it enough mass on the surface to work with. If the planet started out completely flat and bare it would take too long to get continential shields to form. If you really look at it your example shows a planet similar to Earth in terms of land coverage and ocean coverage. There appears to be a large continent submerged under water but this is still a continent and it's pretty cool you got that result. The North American continent was like this for millions of years.

Well I do have to say though , if we had the features that i said in the previous comment, we'd be able to form continents from a bare, ocean world quickly, and that would realistically and easily mimic how it went for Earth.