CODEWITHSUNDEEP
matplotlibpython-xarraycartopyfacet-grid
dodohjk
dodohjk

Reputation: 198

Why does subplot_kws projection hides the lat and lon?

With a xarray dataset, how can I use FacetGrid with a specific cartopy projection and keep the latitude and longitude values along the axis?

I was trying to use faceting as described in the user-guide here. As soon as I try to give the subplots a specific projection, I lose the lat and lon values on the axis

here is an example:

import xarray as xr
import cartopy.crs as ccrs

airtemps = xr.tutorial.open_dataset("air_temperature")
air = airtemps.air - 273.15
t = air.isel(time=slice(0, 365 * 4, 250) )
g_simple = t.plot.contourf(x="lon", y="lat", col="time", col_wrap=3,
                           transform = ccrs.PlateCarree(),
                           subplot_kws={"projection": ccrs.PlateCarree()},
                           levels=51,
)

enter image description here

but if I comment out the projection line

import xarray as xr
import cartopy.crs as ccrs

airtemps = xr.tutorial.open_dataset("air_temperature")
air = airtemps.air - 273.15
t = air.isel(time=slice(0, 365 * 4, 250) )
g_simple = t.plot.contourf(x="lon", y="lat", col="time", col_wrap=3,
                           #  subplot_kws={"projection": ccrs.PlateCarree()},
                           levels=51,
)

enter image description here

Why does it happen? and how can i fix this?

Upvotes: 3

Views: 183

Answers (1)

RuthC
RuthC

Reputation: 3896

Cartopy turns the axis labels off by default because they do not make sense for all of Cartopy's projections. You can turn them on again by using the set_visible method on each axis. I also here added coastlines because I think it helps make it obvious where we are looking.

It seems that xarray is using tight_layout underneath somewhere. I couldn't figure out how to make space for the labels and keep the colorbar in the right place with tight_layout so I switched to using the more modern compressed layout. It turns out you can't switch the layout engine in this way if you already have a colorbar, so I turned the colorbar off initially in the call to contour and added back after switching layout engines.

import xarray as xr
import cartopy.crs as ccrs
import matplotlib.pyplot as plt

airtemps = xr.tutorial.open_dataset("air_temperature")
air = airtemps.air - 273.15
t = air.isel(time=slice(0, 365 * 4, 250) )
g_simple = t.plot.contourf(x="lon", y="lat", col="time", col_wrap=3,
                           transform = ccrs.PlateCarree(),
                           subplot_kws={"projection": ccrs.PlateCarree()},
                           levels=51, add_colorbar=False
)

for ax in g_simple.axs.flat:  # loop through the map axes
    subplotspec = ax.get_subplotspec()
    if subplotspec.is_last_row():
        ax.xaxis.set_visible(True)
    if subplotspec.is_first_col():
        ax.yaxis.set_visible(True)
    ax.coastlines()
    
g_simple.fig.set_layout_engine("compressed")
g_simple.fig.get_layout_engine().set(h_pad=0.2)  # More vertical space between maps
g_simple.add_colorbar()

plt.show()

[image upload seems to be currently broken across StackOverflow]

Upvotes: 3

Related Questions