Coral Larval Connectivity

[Prejie]

Question

Question

Hey, so I'm trying to model coral larvae as particles that drift. Check out the video I embedded! I think I got pretty far!

So really, I'm looking for suggestions to move forward.

Right now, what I've done is to rasterize the shape file of reefs that I have. Then give properties to each particle such as its source and sink reefs and its age and whether it is dead or alive.

Then I have some kernels for settlement and aging. Then I execute them.

But it seems nothing is changing. No settlement occurs, and it proceed as if I have no kernels involved.

I think I might be missing something.

Anyway, what I do want to happen really is that every time step, when a particle enters a polygon (from the shape file) it has a percent chance of settling there (by settling, I mean it gets "stuck" there), otherwise it keeps flowing until it encounters a new polygon again. Then at the end I could record where the particle came from and where it landed. That way, I'd have a connectivity matrix.

Any suggestions or materials or literature I could look into? I want to try some approach that has worked before.

Thanks so much!

barc_animation.mp4

[erikvansebille]

Very cool animation, @Prejie. But if you want help on why your custom kernels are not executed, you will need to provide us with some code (ideally a Minimal Reproducible Example) that lets us have a quick look what may be going on

Note, by the way, that in many cases you could also compute settlement based on the zarr output data in the analysis phase, after you compute the particle trajectories. That may give you a bit more flexibility?

[Prejie]

Thanks Erik! I got settlement going now, but I think my rastered maps are too small to be landed by particle, so I'm trying to remap that out.

Anyway, I got curious about your second statement. Can you please show me examples of how I can compute settlement probabilities after having the zarr output? Thanks so much!