Properly rendered 3d plot_surface in matplotlib

Question

I am trying to render a surface, but I do not manage to get a pretty visualisation. The plot_surface function from matplotlib gives me the following figure:

produced by the code below. How do I get rid of this transparency and the wireframe that is still visible if you look carefully?

import numpy as np
import matplotlib.pyplot as pl
from mpl_toolkits.mplot3d import Axes3D
pl.ion()

nx = 512
ny = 512

Lx = 2.e6
Ly = 2.e6

x = np.linspace(0., Lx, nx)
y = np.linspace(0., Ly, ny)

xx, yy = np.meshgrid(x,y)

Ld = 6.e4
h = np.exp(-( (xx - 0.5*Lx)**2 + (yy - 0.5*Ly)**2) / Ld**2 )

pl.figure()
ax = pl.subplot(111, projection='3d')
ax.plot_surface(xx/1000., yy/1000., h, alpha=1., cstride=1, rstride=1, linewidth=0)
ax.set_zlim3d(-0.2, 1.)

Answer 1

It's only a workaround, but this works for most matplotlib routines like eg contourf (where I had the same problem before); calling the plot routine (in this case plot_surface ) twice solves both problems:

The left figure is with calling plot_surface once, the right one calling it twice.

For a non-transparent surface, setting antialiased=False helps (left figure below), with transparency antialiased=True produces very thin lines at the polygon edges (I suspect because the polygons slightly overlap), but they are hardly visible (right figure below).

fig = pl.figure()
ax = pl.subplot(121, projection='3d')
surf = ax.plot_surface(xx/1000., yy/1000., h, alpha=1.0, cstride=1, rstride=1, linewidth=0, antialiased=False)

ax = pl.subplot(122, projection='3d')
surf = ax.plot_surface(xx/1000., yy/1000., h, alpha=0.3, cstride=1, rstride=1, linewidth=0, antialiased=True)