Implementing correct depth-to-sigma calculation

[erikvansebille]

This PR fixes the way that the conversion is done for CROCO fieldsets from particle.depth (in meters) to local sigma-coordinate. Where this was initially assumed to be a simple linear conversion, this turned out to be much more complex.

After some discussion on the myroms forum https://www.myroms.org/forum/viewtopic.php?p=25752#p25752, the best option is to calculate the local depth of each sigma layer and then linearly interpolate the particle depth to find the fractional sigma depth.

This fixes #1763

[VeckoTheGecko]

Looks good! Just some things that I think would be good to clarify before merging.

I'm not 100% happy with the API changes with from_croco. I feel that having the hc= parameter is a deviation from the rest of the from_* functions. Though I don't have a better suggestion on how we could implement it given our limitations 🤔

[erikvansebille]

I'm not 100% happy with the API changes with from_croco. I feel that having the hc= parameter is a deviation from the rest of the from_* functions. Though I don't have a better suggestion on how we could implement it given our limitations 🤔

Yes, I fully understand; I'm not 100% happy either.

This would actually be another reason to move to xarray for FieldSet creation in Parcels... The xarray DataSet can be expected to contain all these variables (hc, Cs_w, Zeta, H, etc) if the data comes from CROCO/Roms, so we can take them directly from there...

[VeckoTheGecko]

This would actually be another reason to move to xarray for FieldSet creation in Parcels...

I 100% agree. I'm currently scoping that out