Use the ghost-cell spacing for the z one-sided viscous gradient #2067

s_compute_fd_gradient's one-sided z derivative at idwbuff(3)%beg
divided by z_cc(beg + 2) - z_cc(is3_viscous%beg). is3_viscous%beg is
idwbuff(3)%beg + 1 there, so the denominator spanned one cell instead
of two and the outermost ghost gradient was doubled on a uniform grid.

Divide by z_cc(beg + 2) - z_cc(beg), matching the x and y stencils.

The 27 3D Viscous tests pass against the existing goldens.

A fix for this is already written. The pull request was closed only to reduce the number of simultaneously open PRs from this author, not because the work was rejected or found incorrect.

• Closed PR: Use the ghost-cell spacing for the z one-sided viscous gradient #1991
• Branch: fix/viscous-z-ghost-spacing on the sbryngelson fork, at 2253ec7fcd
• Reopening needs no force-push and no rebase: gh pr reopen 1991 -R MFlowCode/MFC

Tracked with the pr-to-do label so the whole set is discoverable together.

Filed with Claude Code.