parcels.FieldSet.from_croco() input files

[wavestorm]

Question

Question

Hi there

I'm in the process of setting up my environment to use parcels (3.1.1) for some experiments. On the parcels.FieldSet.from_croco() function, I have two questions:

1. I do not see an option to bring in the CROCO grid file, only the output file. The tutorial suggests that this function is expecting all of the grid file variables to be in the output file, but this is not generally how a typical CROCO output works. Grid parameters (lat, lon, hc, s-coordinates, etc.) are in a separate grid file to the data. Is there something that I am missing?

2. How does one load a sequence of CROCO files (such as a multi-year run split into 5-day output files)?

Thanks!

Supporting code/error messages

# Paste your code within this block

# Paste your error message within this block

[erikvansebille]

Thanks for considering to use Parcels, @wavestorm. And also thanks for raising this question. The support for CROCO is very new, and we don't use it much here in my own team, so I don't have much experience. The only output I saw was in a single file, but I believe you that there are other options

Parcels does support separate data and grid files for NEMO, I think that you could leverage that support. Following the 3D NEMO tutorial, you might be able to do something like

datafiles = sorted(glob("data*.nc"))  # replace with your list of filenames
gridfile = "grid.nc"  # replace with name of gridfile

filenames = {
    "U": {"lon": gridfile, "lat": gridfile, "depth": gridfile, "data": datafiles},
    "V": {"lon": gridfile, "lat": gridfile, "depth": gridfile, "data": datafiles},
    "W": {"lon": gridfile, "lat": gridfile, "depth": gridfile, "data": datafiles},
    "H": {"lon": gridfile, "lat": gridfile, "depth": gridfile, "data": datafiles},
    "Zeta": {"lon": gridfile, "lat": gridfile, "depth": gridfile, "data": datafiles},
    "Cs_w": {"lon": gridfile, "lat": gridfile, "depth": gridfile, "data": gridfile},
}

where you of course replace the correct filenames in the first two lines, and then use that filenames instead of file in your parcels.FieldSet.from_croco() call. Note that loading a sequence of files (your second question) is as simple as using wildcards like above, or providing a list of filenames.

Hope this helps; please let us know if this works!

[wavestorm]

Hi Erik

Fantastic, that's more than enough to get me going. I have the development version installed so I can experiment with getting my CROCO data into Parcels. I will report back on how it goes.