如何使用matplotlib在Cartopy极地立体图中获得平滑的网格线？

Question

When plotting with matplotlib and cartopy's NorthPolarStereo projection, the constant-latitude gridlines (parallels) are not smooth circles.

Here is a minimum working example:

from matplotlib import pyplot
import cartopy.crs

ax = pyplot.axes(projection=cartopy.crs.NorthPolarStereo())
ax.set_extent((-180, 180, 60, 90), crs=cartopy.crs.PlateCarree())
ax.gridlines()

pyplot.show()

As you can see, the parallels are a series of straight-line segments between points. There aren't enough points along each line to make them appear smooth. How can I make the parallels more circular?

Answer 1

Q: How can I make the parallels more circular?

A: You need larger value of n_steps property of the gridliner object created by gridlines() . Its default value is 30. Here is the relevant code that sets the value to 90, and generates better plot.

gl = ax.gridlines(draw_labels=False, xlocs=None, ylocs=None)
gl.n_steps = 90

See reference here .

Sample output:

Answer 2

This is due to the default threshold being too large for this use case. This is not a user-controllable part of Cartopy's projection interface. You can hack around it using the private _threshold attribute, dividing it by 100 looks reasonable for your use case:

from matplotlib import pyplot
import cartopy.crs

hacked_proj = cartopy.crs.NorthPolarStereo()
hacked_proj._threshold /= 100.
ax = pyplot.axes(projection=hacked_proj)
ax.set_extent((-180, 180, 60, 90), crs=cartopy.crs.PlateCarree())
ax.gridlines()

pyplot.show()

This is not a great solution because it uses the implementation details of the Stereographic class, it would be better to have better setting of this in Cartopy. Opening a ticket on the Cartopy Github repo with this use case would be a good start https://github.com/SciTools/cartopy/issues .