Plotting data from different formats - ECMWF internal earthkit training, March 2026

earthkit-plots leverages the power of earthkit-data to make plotting data from various formats a completely equivalent process.

In this example, we will access the same GRIB dataset above from a NetCDF file, an xarray dataset, and NumPy arrays. In each case, plotting the data should give exactly the same result.

GRIB¶

import earthkit.data as ekd

era5_2t_grib = ekd.from_source("sample", "era5-monthly-mean-2t-199312.grib")

NetCDF¶

The data used in this example comes from the ERA5 monthly averaged data on single levels from 1940 to present dataset in the Copernicus Climate Change Service Climate Data Store, where it is available in both GRIB and netCDF. For convenience, the same timestep as we accessed in GRIB above is available in netCDF as a sample.

era5_2t_nc = ekd.from_source("sample", "era5-monthly-mean-2t-199312.nc")

xarray¶

Using earthkit-data, we can easily convert our netCDF data to xarray with:

era5_2t_ds = era5_2t_nc.to_xarray()
era5_2t_ds

NumPy¶

Similarly, we are able to transform into NumPy arrays with .to_numpy().

Since a NumPy array of values has no geographical information, we will also need to extract the latitude and longitude values by converting to a FieldList object and using the geography component.

Finally, NumPy arrays have no metadata attached to them. If we want automatic styles, unit conversion and metadata formatting, we should build a dictionary of the minimal metadata we need - in this case the units, name and time information.

from datetime import datetime

lats, lons = era5_2t_nc.to_fieldlist().geography.latlons()
era5_2t_arr = era5_2t_nc.to_numpy().squeeze()

metadata = {
    "units": "K",
    "long_name": "2 metre temperature",
    "time": datetime(1993, 12, 1),
}

Putting it all together¶

Now, let’s plot data from all these different sources as subplots on a Figure. Notice that (with the slight exception of NumPy) we use exactly the same methods in exactly the same way for each source type.

import earthkit.plots as ekp

fig = ekp.Figure(domain="Europe", rows=2, columns=2)

for fmt, data in zip (["GRIB", "netCDF", "xarray"], [era5_2t_grib, era5_2t_nc, era5_2t_ds]):
    subplot = fig.add_map()
    subplot.plot(data, units="celsius")
    subplot.title(f"Source: {fmt}\n{{variable_name}} - {{time:%b %Y}}")

# NumPy is slightly different as we need to manually pass the coordinates and metadata
subplot = fig.add_map()
subplot.plot(era5_2t_arr, x=lons, y=lats, metadata=metadata, units="celsius")
subplot.title("Source: NumPy\n{variable_name} - {time:%b %Y}")

fig.coastlines()
fig.legend()
    
fig.show()

Conclusion¶

This example demonstrates that earthkit-plots promotes format-agnosticism when writing visualisation workflows. Creating maps follows the same approach from different sources like GRIB files, netCDF files, xarray datasets and NumPy arrays (which inevitably require a little more work!).