Full documentation on NetCDF functionality in Metview is here.
Setup
Navigate into the 4_netcdf folder within Metview where you will find some data files and other icons.
Examine and visualise a geomatrix style netCDF file
Right-click on the file era5_2000_aug.nc and choose examine to see the structure of the file:
For more complete information, click on the NcDump tab.
You will need to tell Metview how to visualise this data, as there are multiple variables. Create a new NetCDF Visualiser icon, edit it and set the following parameters:
Parameter | Value |
---|---|
Netcdf Data | Drop the era5_2000_aug.nc icon into this box |
Netcdf Plot Type | Geomatrix |
Netcdf Latitude Variable | latitude |
Netcdf Longitude Variable | longitude |
Netcdf Value Variable | t2m |
Save the icon and visualise it. For fun, drop the supplied icons contour_t2m and mollweide into the plot window to obtain the following:
Examine and visualise a geopoints style netCDF file
Examine the file madis-maritime.nc. We will plot the temperature variable. As you can see, there are 1-dimensional variables for temperature, latitude and longitude. Create a copy of your previous NetCDF Visualiser icon and edit it as follows:
Parameter | Value |
---|---|
Netcdf Data | Right-click/remove the existing netCDF file from there, then drop madis-maritime.nc into this box |
Netcdf Plot Type | Geo Points |
Netcdf Latitude Variable | latitude |
Netcdf Longitude Variable | longitude |
Netcdf Value Variable | temperature |
Visualise it to get a default plot. In the solutions folder is a script called gallery_example_nc_maritime_obs.py, which converts from Kelvin to Celcius and adds some stying to the plot (generated as PDF - simply remove the mv.setoutput()
command to get an on-screen visualisation).
Extract data and convert to pandas
Have a look in the solutions folder and edit and run the script netcdf_to_pandas.py. This shows how to extract some metadata from the previous netCDF file, and also some value arrays and convert into a pandas dataframe. The code is also here: