Question: without stretching, does the grid need to rotate?

[weiyuan-jiang]

The "stretched_grid" is set to true when stretch factor is not closed to 1 and do_schmidt is true.

GFDL_atmos_cubed_sphere/tools/fv_grid_tools.F90

Lines 578 to 579 in add33b9

	if ( (Atm%flagstruct%do_schmidt .or. Atm%flagstruct%do_cube_transform) .and. abs(atm%flagstruct%stretch_fac-1.) > 1.E-5 ) then
	stretched_grid = .true.

But in "direct_transform" or "cube_transform", even the stretch factor is closed to 1 and the grid does not stretch, the grid still rotate any way.

Is this intentional ?

[lharris4]

We can always rotate the grid even in the absence of stretching, which is useful for rotating the parent grid and inserting nests. If you wish to run with a standard uniform-resolution global domain, I do strongly recommend setting do_cube_transform = .F.. We are also encouraging people to move away from do_schmidt since that creates a tile 6 that is oriented with south at the "top" of the grid, unlike do_cube_transform for which tile 6 has south at the bottom as we expect.