allow chunk and two merged chunks for spherical velocity statistics.

[MFraters]

Since there are spherical geometries as well, it makes sense to me to allow them.

[tjhei]

Did you check that it works correctly?

[MFraters]

I have checked that the values look reasonable with the two merged chunks:

RMS, max velocity:                              0.014 m/year, 0.133 m/year
Radial RMS, tangential RMS, total RMS velocity: 0.00272 m/yr, 0.0138 m/yr, 0.014 m/yr

This is subduction model, with a small slab in a large box, with a large part of the lithosphere moving at about 2 to 5 cm/yr horizontally. The total RMS is the same, the tangantial RMS (horizontal direction I assume) is nearly the same as the RMS, which I think makes sense in this mode and the Radial RMS (depth) is much lower.