This is adopted from this example to visualize sweeps of radar data as a 3D volume. The 3D visualization here help us gather insight into the 3D structure storms.
import xarray as xr
import xradar as xd
import pyvista as pv
import pvxarray # Import to register accessor
from open_radar_data import DATASETS
pv.set_plot_theme("document")
pv.set_jupyter_backend("server")
# Temporary manual mode override; waiting for pyvista release including changes from https://github.com/pyvista/pyvista/pull/5798
pv.global_theme.trame.server_proxy_enabled = False
pv.global_theme.trame.jupyter_extension_enabled = True
Fetching CfRadial1 radar data file from open-radar-data repository.
filename = DATASETS.fetch("cfrad.20080604_002217_000_SPOL_v36_SUR.nc")
radar = xd.io.open_cfradial1_datatree(filename)
radar = radar.xradar.georeference() # Add georeferencing to plot in 3D
list(radar.children)
Now we need to stack all of the sweeps on top of each other to create a 3D data volume
ds_list = []
for key in list(radar.children):
if "sweep" in key:
ds_list.append(radar[key].ds.drop_duplicates(dim="azimuth"))
ds = xr.concat(ds_list, dim="sweep")
ds
Use pyvista-xarray's accessor to create a PyVista mesh from the data volume. This will build a curvilinear grid (pyvista.StructuredGrid)
mesh = ds["DBZ"].pyvista.mesh(x="x", y="y", z="z")
mesh
Preview the data volume
mesh.plot(clim=(-20, 60))
Pseudo-volume rendering with transparent contours
pl = pv.Plotter()
pl.add_mesh(mesh.contour(), opacity=0.9, clim=(-20, 60), lighting=False)
pl.show(jupyter_backend="static")