[Python] Contourf bug outside extent with cartopy Orthographic

I'm trying to plot this xarray data array (file here) :


file_cyclic :

<xarray.DataArray (latitude: 26, longitude: 361)>
array([[66.66666667, 61.53846154, 52.94117647, ..., 66.66666667,
68.75 , 66.66666667],
[58.82352941, 57.14285714, 50. , ..., 82.35294118,
75. , 58.82352941],
[66.66666667, 66.66666667, 64.28571429, ..., 75. ,
72.22222222, 66.66666667],
...,
[85.91549296, 87.14285714, 85.07462687, ..., 86.95652174,
87.67123288, 85.91549296],
[92.06349206, 91.80327869, 88.70967742, ..., 93.5483871 ,
90.76923077, 92.06349206],
[92.59259259, 92.85714286, 91.07142857, ..., 90.56603774,
90.19607843, 92.59259259]])
Coordinates:
* longitude (longitude) float32 0.0 1.0 2.0 3.0 ... 357.0 358.0 359.0 360.0
* latitude (latitude) float32 -60.0 -61.0 -62.0 -63.0 ... -83.0 -84.0 -85.0


using this code :

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

file_cyclic = xr.open_dataset('./my_data.nc') #If you want to try with my data
file_cyclic= file_cyclic['__xarray_dataarray_variable__'] #If you want to try with my data


plt.close('all')


fig, ax = plt.subplots(1, 1, figsize=(12, 12),
subplot_kw={'projection': ccrs.Orthographic(central_latitude=-90, central_longitude=0),
'transform': ccrs.PlateCarree()})
cont = file_cyclic.plot.contourf(levels=np.arange(0,100.1,5),cmap='YlGnBu', transform=ccrs.PlateCarree(), vmin=0, vmax=100, ax=ax,add_colorbar=True, extend='neither')

ax.coastlines()
ax.set_extent([file_cyclic.longitude.min(), file_cyclic.longitude.max(), file_cyclic.latitude.min(), file_cyclic.latitude.max()], crs=ccrs.PlateCarree())


I get a bug : my colorscale is off and it fills my no-data area (where I have 0, nan or no data actually) with color...

Bug

However when I change the levels to :

levels=np.arange(1,100.1,5)


it doesn't produce a bug, but this is not the levels I want.

no bug, but not the levels I want

I don't understand where this is coming from since, when I manually put

levels=[ 1, 5., 10., 15., 20., 25., 30., 35., 40., 45., 50.,
55., 60., 65., 70., 75., 80., 85., 90., 95., 100.]


the bug is still here, so this is not coming from the lowest level. Just bottom level changed, still bugging

The bug is also here if I change the higher level...

This is not just filling my no-data area, otherwise I would have cropped this area from the plot, this is also messing with my colorscale, so that's a huge problem.

How can I handle that?

Thank you,

Continue reading...