Closed doblerone closed 5 months ago
Thanks @doblerone for pointing this out. I think you are right, there should a depth dimension and coordinate variable (like in the CMIP6 cmor tables that were adapted here). This information is still a little incomplete in the data request table for CORDEX and might got lost. Let me move this issue there since this is not CV related.
fixed using:
import pandas as pd
df = pd.read_csv("CORDEX-CMIP6/data-request.csv")
def insert_sdepth(dims):
split = dims.split()
if len(split) == 3: # contains time
split.insert(-1, "sdepth")
else:
split.append("sdepth")
return " ".join(split)
out_names = ["tsl", "mrsfl", "mrsol"]
sdepths = df[df.out_name.isin(out_names)].dimensions.map(insert_sdepth)
df.loc[sdepths.index, "dimensions"] = sdepths
df.to_csv("CORDEX-CMIP6/data-request.csv", index=False)
@doblerone just to let you know, i kept the meta data of the sdepth
coordinate entry like it is in CMIP6:
meaning that this coordinate also will require boundaries ("must_have_bounds": "yes
). I think it makes sense since it clearly defines also the thickness of the soil layers.
The Tier1 request includes three variables that need to be stored at each soil layer (3D variable): tsl, mrsfl and mrsol.
Should the vertical soil coordinate be defined in the archive specifications? And added here: https://github.com/WCRP-CORDEX/cordex-cmip6-cmor-tables/blob/main/Tables/CORDEX_coordinate.json Could be (just a guess)
And then add that dimension here https://github.com/WCRP-CORDEX/cordex-cmip6-cmor-tables/blob/main/Tables/CORDEX_6hr.json for the three variables i.e. "dimensions": "longitude latitude depth time1",