GeoPandas Coordinate System Transformation (to_crs) inverts Shapefile

Question

Converting the 10m resolution ocean shapefile from Natural Earth to a different projection "inverts" the shapefile ("drying up the ocean"):

import geopandas as gpd
oceans = gpd.read_file('./ne_10m_ocean')

oceans.plot()

Ok, here the ocean is nice and blue 🐟

But converting to another projection, centered in Nigeria (ESRI:102022), we get this:

oceans.to_crs("ESRI:102022").plot()

What is going on here? How can I re-fill the ocean?

Answer 1

Thanks to the hint by Ian Turton, I found a simple solution that uses GeoPandas: I clipped the oceans shapefile to the area I want to plot, thereby avoiding polygons being turned "inside-out":

GeoDataFrame.clip

target_projection = "ESRI:102022"

lower_left = gpd.points_from_xy(
    x = [2.6], # longitude
    y = [5.725311], # latitude
    crs='EPSG:4326' # = WGS 84
)

upper_right = gpd.points_from_xy(
    x = [7.7], # longitude
    y = [9.3], # latitude
    crs='EPSG:4326' # = WGS 84
)

oceans = gpd.read_file('./ne_10m_ocean')

oceans_clipped = oceans.clip(
    mask=[
        lower_left.x[0],
        lower_left.y[0],
        upper_right.x[0],
        upper_right.x[0]
    ]
)