Search code examples
python-3.xmatplotlibaxis-labelscartopy

Hide right-side axis (latitude) labels for high-latitude non-rectangular projection in Cartopy

I am trying to hide the right-side labels (aka latitude) in this Cartopy map using right_labels = False but only works for some values. However, the top/bottom labels can be hidden using top_labels = False and bottom_label=False.

To reproduce:


 import numpy as np
    import matplotlib.pyplot as plt
    import cartopy.crs as ccrs
    import matplotlib.ticker as mticker
    import matplotlib.path as mpath
    import cartopy.feature as cf

    """
    Plot Alaska
    """

    # Map View Using Cartopy
    fig = plt.figure(figsize=(8,6))

    xmin=-163
    xmax=-120
    ymin=50
    ymax=71

    proj = ccrs.LambertConformal(central_longitude=(xmin+xmax)/2, central_latitude=(ymin+ymax)/2)
    ax = fig.add_subplot(1, 1, 1, projection=proj)
    n = 20
    aoi = mpath.Path(
        list(zip(np.linspace(xmin, xmax, n), np.full(n, ymax))) + \
        list(zip(np.full(n, xmax), np.linspace(ymax, ymin, n))) + \
        list(zip(np.linspace(xmax, xmin, n), np.full(n, ymin))) + \
        list(zip(np.full(n, xmin), np.linspace(ymin, ymax, n)))
    )
    ax.set_boundary(aoi, transform=ccrs.PlateCarree())

    # Plot Ocean Borders
    ocean = cf.NaturalEarthFeature('physical', 'ocean', scale='50m', edgecolor='k', facecolor='lightblue', lw=1,
                                   linestyle='-')
    ax.add_feature(ocean)
    # Colored Land Background
    land = cf.NaturalEarthFeature('physical', 'land', scale='50m', facecolor='snow', lw=1, linestyle='--')
    ax.add_feature(land)

    ax.set_extent([xmin, xmax, ymin, ymax], crs=ccrs.PlateCarree())
    # Set gridlines to variable so you can manipulate them
    gl = ax.gridlines(draw_labels=True, crs=ccrs.PlateCarree(), x_inline=False, y_inline=False)
    gl.xlocator = mticker.FixedLocator([-160, -150, -140, -130, -120])
    gl.ylocator = mticker.FixedLocator([50, 55, 60, 65, 70])
    gl.top_labels = False
    gl.right_labels = False

    plt.show()

And this is what I obtain:

funky axes Alaska

I have a suspicion it has to do with using ax.set_boundary() to make the map but after looking online I couldn't find anything except this GitHub issue that mentions the issue but it is closed so I suppose they fixed it. I am using Cartopy version 0.21.1.

The question arose from looking into another SO question: Cartopy labels not appearing for high-latitude non-rectangular projection (all code credits to @kbiegseismic).

Solution

  • As it is obvious that ax.set_boundary() confuses the process within .gridlines(), especially, when it needs to hide some of the y-labels in this particular case. Consequently, some of the labels' visibility are set falsely.

    We will need to wait for next version of Cartopy to see better gridlines() funtioning. If you need a workaround, you can try the code snippets below. Just place it towards the end of existing code.

    # Generate the plot to enable access to the labels' attributes
plt.draw()

# Iterate for the y-labels
# The right labels have x coordinates > 0
# The left labels < 0
for ea in gl.ylabel_artists:
    right_label = ea.get_position()[0] > 0
    # print(ea, ea.get_position()[0], ea.get_visible())
    if right_label:
        ea.set_visible(False)

plt.show()

    no-r-labels

    If you modify the map boundaries to have small extra buffer like these:-

    pg = 0.006
#xmin=-163 -pg
xmax=-120 +pg
ymin=50 -pg
#ymax=71 +pg

    the final plot will include the missing labels at 120°W longitude and 50°N latitude.

    plot2