You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There are issues when regridding certain datasets of the data collections
C3S-CMIP6
C3S-CMIP5
C3S-CORDEX
These encompass:
Some datasets include unmasked missing values in the latitude and longitude coordinate arrays
Some datasets include erroneous / CF uncompliant metadata (eg. standard_name longitude for the grid_longitudeand longitude variables - cf_xarray uses units, long_name and standard_name to identify lat/lon coordinate variables)
There is an issue with the detection of degenerated cells in the clisops.core.regrid.Grid class:
Not all cells collapsing to lines are detected
ESMF additionally considers "smashed" cells as degenerate, which are quadrilateral cells where the opposite bounds are close to each other
There is an issue with adjusting the longitude frame for datasets with NaNs in the longitude array
Some datasets have invalid bounds (eg. vertices with dimensionality of 2 instead of 4)
What I did
This is just a log of the problems.
Generally, I tried:
import xarray as xr
import clisops.core as clore
ds=xr.open_dataset("some_dataset")
g=clore.Grid(ds=ds)
c3s-cmip5/output1/BCC/bcc-csm1-1/rcp85/mon/ocean/Omon/r1i1p1/zos/v20120705
-> failed because: rlat and lat identified as latitude
cmip5/output1/BCC/bcc-csm1-1/decadal1991/mon/ocean/Omon/r3i1p1/v20121026
-> failed because: rlat and lat identified as latitude
cmip5/output1/LASG-CESS/FGOALS-g2/rcp85/mon/ocean/Omon/r1i1p1/v2/thetao
-> failed because of time.calendar and time.units ("noleap" "days since 0001-01")
for this, the time.units have to be corrected in the metadata (xarraythrows a ValueError upon opening the file), so out of scope
CMIP6/HighResMIP/BCC/BCC-CSM2-HR/hist-1950/r1i1p1f1/Omon/tos/gn/v20200922
-> failed because lat and latitude are identified as latitude
CMIP6/C4MIP/BCC/BCC-CSM2-MR/1pctCO2-rad/r1i1p1f1/Omon/pbo/gn/v20190514
-> failed because lat and latitude are identified as latitude
CMIP6/AerChemMIP/BCC/BCC-ESM1/ssp370/r1i1p1f1/Omon/pbo/gn/v20190624
-> failed because lat and latitude are identified as latitude
CMIP6/HighResMIP/NCAR/CESM1-CAM5-SE-HR/control-1950/r1i1p1f1/Omon/tos/gn/v20200724
-> failed because extent ranges from [-1, 359.9] (legal is [-180,360] but max - min may not exceed 360 degree plus grace
will have to see if (x)ESMF accepts this with an extent >360 degrees
CMIP6/CMIP/NCAR/CESM2-FV2/1pctCO2/r1i1p1f1/Omon/intpn2/gn/v20200310
-> failed because of unmasked missing values in the longitude and latitude arrays (no _FillValue / missing_value attribute)
CMIP6/OMIP/NOAA-GFDL/GFDL-OM4p5B/omip1/r1i1p1f1/Omon/volcello/gn/v20180701
-> failed because extent -299.9143 to 1e+20 (unmasked missing values - no _FillValue / missing_value attribute)
c3s-cordex/output/ARC-44/BCCR/ECMWF-ERAINT/evaluation/r1i1p1/BCCR-WRF331/v1/day/tas/v20200915
-> failed because vertices have only 2 vertices instead of 4: lat_bnds(rlat, rlon, bnds) with bnds=2
The text was updated successfully, but these errors were encountered:
Description
There are issues when regridding certain datasets of the data collections
These encompass:
longitude
for thegrid_longitude
andlongitude
variables -cf_xarray
uses units, long_name and standard_name to identify lat/lon coordinate variables)clisops.core.regrid.Grid
class:ESMF
additionally considers "smashed" cells as degenerate, which are quadrilateral cells where the opposite bounds are close to each otherWhat I did
This is just a log of the problems.
Generally, I tried:
-> failed because: rlat and lat identified as latitude
-> failed because: rlat and lat identified as latitude
-> failed because of time.calendar and time.units ("noleap" "days since 0001-01")
time.units
have to be corrected in the metadata (xarray
throws aValueError
upon opening the file), so out of scope-> failed because lat and latitude are identified as latitude
-> failed because lat and latitude are identified as latitude
-> failed because lat and latitude are identified as latitude
-> failed because extent ranges from [-1, 359.9] (legal is [-180,360] but max - min may not exceed 360 degree plus grace
(x)ESMF
accepts this with an extent>360 degrees
-> failed because of unmasked missing values in the longitude and latitude arrays (no _FillValue / missing_value attribute)
-> failed because extent -299.9143 to 1e+20 (unmasked missing values - no _FillValue / missing_value attribute)
-> failed because vertices have only 2 vertices instead of 4: lat_bnds(rlat, rlon, bnds) with bnds=2
The text was updated successfully, but these errors were encountered: