cfgrib is failing to recognize data variable

Question

Data source: http://edc.occ-data.org/nexrad/mosaic/#data-updates

• Using 'MRMS_PrecipRate_00.00_20010101-120000.grib2' which is the first file in the 2001 dataset

Problem: The 'PrecipRate' attribute I know to be in the grib file is unrecognized. The data are there but I'd like to preserve the metadata in the xarray dataset.

import xarray as xr

ds = xr.open_dataset('MRMS_PrecipRate_00.00_20010101-120000.grib2', engine = 'cfgrib')

print(ds)

<xarray.Dataset>
Dimensions:     (latitude: 3500, longitude: 7000)
Coordinates:
    time        datetime64[ns] 2001-01-01T12:00:00
    step        timedelta64[ns] 00:00:00
    surface     float64 0.0
  * latitude    (latitude) float64 54.99 54.98 54.98 54.97 ... 20.03 20.02 20.01
  * longitude   (longitude) float64 230.0 230.0 230.0 ... 300.0 300.0 300.0
    valid_time  datetime64[ns] 2001-01-01T12:00:00
Data variables:
    unknown     (latitude, longitude) float32 ...
Attributes:
    GRIB_edition:            2
    GRIB_centre:             161
    GRIB_centreDescription:  161
    GRIB_subCentre:          0
    Conventions:             CF-1.7
    institution:             161
    history:                 2021-05-04T13:11 GRIB to CDM+CF via cfgrib-0.9.9...

Answer 1

This is a common failure working with grib data. Extract grib data requires grib tables. In these tables the parameter identifier are explained and mapped to their names.

So what you have to do, is get in contact with the publisher/provider of the grib files and ask for the corresponding grib tables. I guess pygrib uses public ncep grib tables.

Additionally I recommend to use xarray in combination with cfgrib. It based on the powerful ECCODES Library by ECMWF. Here you have to set an additional ECCODES_DEFINTION_PATH to where the grib tables stored. Maybe eccodes contains the right grib tables by default. Otherwise get in contact with the grib file provider.