在savefig之前添加以下代码：

ax.auto_scale_xyz([0, 500], [0, 500], [0, 0.15])

如果你不想要方形轴：
编辑site-packages\mpl_toolkits\mplot 3d\axes3d.py中的get_proj函数：

xmin, xmax = np.divide(self.get_xlim3d(), self.pbaspect[0])
ymin, ymax = np.divide(self.get_ylim3d(), self.pbaspect[1])
zmin, zmax = np.divide(self.get_zlim3d(), self.pbaspect[2])

然后添加一行设置pbaspect：

ax = fig.gca(projection = '3d')
ax.pbaspect = [2.0, 0.6, 0.25]

this question的答案非常适合我。你不需要设置任何比率，它会自动完成所有事情。

在github上有一个问题：https://github.com/matplotlib/matplotlib/issues/8593
上面的解决方案似乎不再起作用了。现在必须按以下方式编辑site-packages\mpl_toolkits\mplot3d\axes3d.py中的get_proj函数：

try:
            self.localPbAspect = self.pbaspect
        except AttributeError:
            self.localPbAspect = [1,1,1]

        xmin, xmax = ( lim / self.localPbAspect[0] for lim in self.get_xlim3d() )
        ymin, ymax = ( lim / self.localPbAspect[1] for lim in self.get_ylim3d() )
        zmin, zmax = ( lim / self.localPbAspect[2] for lim in self.get_zlim3d() )

这就是我解决浪费空间问题的方法：

try: 
    self.localPbAspect=self.pbaspect
    zoom_out = (self.localPbAspect[0]+self.localPbAspect[1]+self.localPbAspect[2]) 
except AttributeError: 
    self.localPbAspect=[1,1,1]
    zoom_out = 0 
xmin, xmax = self.get_xlim3d() /  self.localPbAspect[0]
ymin, ymax = self.get_ylim3d() /  self.localPbAspect[1]
zmin, zmax = self.get_zlim3d() /  self.localPbAspect[2]

# transform to uniform world coordinates 0-1.0,0-1.0,0-1.0
worldM = proj3d.world_transformation(xmin, xmax,
                                         ymin, ymax,
                                         zmin, zmax)

# look into the middle of the new coordinates
R = np.array([0.5*self.localPbAspect[0], 0.5*self.localPbAspect[1], 0.5*self.localPbAspect[2]])
xp = R[0] + np.cos(razim) * np.cos(relev) * (self.dist+zoom_out)
yp = R[1] + np.sin(razim) * np.cos(relev) * (self.dist+zoom_out)
zp = R[2] + np.sin(relev) * (self.dist+zoom_out)
E = np.array((xp, yp, zp))

现在只需要在这里声明完整的解决方案。Axes3D.set_box_aspect

ax.set_box_aspect(aspect=(x_scale, y_scale, z_scale))

完整代码：

import matplotlib.pyplot as plt
from matplotlib import cm
import numpy as np

# Create figure.
fig = plt.figure(figsize=(16, 9))
ax = fig.add_subplot(projection='3d')

# Generate example data.
R, Y = np.meshgrid(np.arange(0, 500, 0.5), np.arange(0, 40, 0.5))
z = 0.1 * np.abs(np.sin(R/40) * np.sin(Y/6))

# Plot the data.
surf = ax.plot_surface(R, Y, z, cmap=cm.jet, linewidth=0)
fig.colorbar(surf)

# Set viewpoint.
ax.azim = -160
ax.elev = 30

# Label axes.
ax.set_xlabel('Along track (m)')
ax.set_ylabel('Range (m)')
ax.set_zlabel('Height (m)')

# padding for axis label not to overlap with axis ticks
ax.xaxis.labelpad=30    

# changing aspect ratio
ax.set_box_aspect(aspect=(4, 1, 0.5), zoom=1.3)

# Save image.
fig.savefig('data.jpg', dpi=200)

Resulting plot image
感谢@ImportanceOfBeingErnest提供问题和答案。