Skip to content

README.md

Latest commit

 

History

History
30 lines (24 loc) · 1.16 KB

README.md

File metadata and controls

30 lines (24 loc) · 1.16 KB

HRRRParser

A VirtualiZarr parser for NOAA HRRR GRIB files.

Status

Experimental, proof-of-concept.

Usage

See VirtualiZarr for documentation on how parsers work to generate virtual representations of datasets.

  • Currently the HRRRParser only supports hourly files. Sub-hourly support should be added soon.

  • Using the generated VirtualiZarr ManifstStore or virtual dataset requires use of the included HRRRGribberishCodec for decoding.

    This codec can be registered with zarr-python via

    from zarr.registry import register_codec
from hrrrparser.codecs import CODEC_ID, HRRRGribberishCodec
register_codec(CODEC_ID, HRRRGribberishCodec)

  • The resulting virtual Zarr stores contain a "reference time" and "step" dimension (time and step respectively) that can be utilized for different “Forecast Model Run Collections” (FMRC) approaches. See https://xarray-indexes.readthedocs.io/earth/forecast.html for use with valid time approaches.

  • See /examples/icechunk_region.ipynb and /examples/manifeststore.ipynb for more detailed usage and file concatenation options.