Closed SadeghTabas-NOAA closed 4 months ago
@SadeghTabas-NOAA Does vertical interpolation work for this purpose?
@SadeghTabas-NOAA Does vertical interpolation work for this purpose?
we have these pressure levels in the master file, we just did not export them to the grib2 files, we will talk about this on Monday.
@SadeghTabas-NOAA @LinlinCui-NOAA There are two gdas grib files:
gdas.t00z.pgrb2.0p25.f000 gdas.t00z.pgrb2b.0p25.f000
The model fields on those pressure levels listed in the issue are in gdas.t00z.pgrb2b.0p25.f000. Please let me know if there is any missing field. Thanks
@SadeghTabas-NOAA @LinlinCui-NOAA There are two gdas grib files:
gdas.t00z.pgrb2.0p25.f000 gdas.t00z.pgrb2b.0p25.f000
The model fields on those pressure levels listed in the issue are in gdas.t00z.pgrb2b.0p25.f000. Please let me know if there is any missing field. Thanks
Thank you so much @junwang-noaa for this. we will work on it.
@junwang-noaa I dont see gdas.t00z.pgrb2b.0p25.f000
on s3 or NOMADS. Could you please double check? Thanks
The GDAS grib2 data does not account for a few levels (e.g., 125, 175, 225, 775, 825, 875) that are needed for the current graphcast pre-trained model. As we have these levels in the master file from post processing in Gaussian grid, the
gdas_utility.py
for 37 PL, needs to be updated in a way that, it gets the data for those levels in the Gaussian grid and interpolates them to latlon grid to be used for graphcast 37 PL inputs (wgrib2
is a potential tool for this task).