Vertical Langrangian Coordinates?

The FV3 core uses a sort of vertical Langrangian coordinates.
For the smallest time steps, the model uses a vertical lagrangian
coordinates.  After N time steps, the layers are mapped to those
of the hybrid sigma-presseure coordinates.  After "N*M" time steps, 
the physics are run.

Output is usually (necessarily?) done at a multple of the "physics" 
time step. So the user sees that the model output uses hybrid 
sigma-pressure coordinates.  So seeing is believing?