GRIB

GRIB is the WMO’s format for binary gridded data consisting of GRIB messages, which in the earthkit terminology are called fields. The earthkit-data GRIB interface is based on ecCodes and can handle both GRIB edition 1 and edition 2.

Fieldlists

We can read/retrieve GRIB data with from_source(). The resulting object will be a GribFieldList representing a list of GribFields, which we can iterate through:

>>> import earthkit.data
>>> ds = earthkit.data.from_source("file", "docs/examples/test.grib")

>>> len(ds)
2

>>> for f in ds:
...     print(f)
...
GribField(2t,1000,20200513,1200,0,0)
GribField(msl,1000,20200513,1200,0,0)

The following table gives us an overview of the GRIB FieldList API:

Highlights of the GRIB FieldList API

Methods/Operators	API
Conversion to scientific Python objects	to_xarray()
Concatenation
Iteration
Selection with [...]
Selection with .sel()	sel()
Ordering with .order_by()	order_by()
Accessing data values	to_numpy()
Accessing metadata	metadata()
Inspecting contents	ls(), head() and tail()

Fields

A GribField represent a single GRIB field. It primarily offers methods to:

• extract field values, such as to_numpy()

• extract field metadata, such as metadata()

Examples:

• Using GRIB data

• GRIB: inspecting contents

• GRIB: selection using metadata

• GRIB: missing values