Visualization with FV3-JEDI

The FV3-JEDI system works with cubed-sphere fields and reads only cubed-sphere data from GFS and GEOS output, as outlined in IO. Visualization of cubed-sphere data is nontrivial since most software for scientific visualization expects the field to be arranged on a latitude-longitude grid.

Data that is output in the GEOS format, and specifically by the file output under the configuration key filename_bkgd is readable by Panoply, available from https://www.giss.nasa.gov/tools/panoply/. Panoply can interpret and create plots of cubed-sphere fields directly. It has a simple GUI and a number of useful features.

The GFS data that is output in tile files is not straightforward to visualize and no software for doing so is provided with FV3-JEDI or the FV3 models. In order to plot fields output by GFS it has to first be converted to another format. That format can either be GEOS, and then use Panoply, or to a longitude-latitude grid using the lonlat type of output. FV3-JEDI is provided with a convert state application in the bin directory. An example configuration (convertstate_gfs_c2ll.yaml) for driving the convert state application and converting GFS output to longitude-latitude is provided in the testinput directory of the repository. In this example the file gfs.bkg.lonlat.20180415_000000z.nc4 is output with all the fields on the lon-lat grid. The executable fv3jedi_plot_field.x, provided in the bin directory can be used to plot the field at a certain level. The following example plots the temperature field T at layer 64.

fv3jedi_plot_field.x --inputfile gfs.bkg.lonlat.20180415_000000z.nc4 \
                     --fieldname T --layer 64

If working with a system that has X-forwarding or on a local system, the optional argument --showfig=true can be applied. This will display the figure after saving it.

Note that the above program requires Cartopy, Matplotlib, Click, and netCDF4 Python packages.