[Mitgcm-support] integrate_for_w.F

[mitgcm-support at dev.mitgcm.org]

Chris,

I've just submitted a change to integrate_for_w.F
that fixes the DEC problem in a different way
and keeps W valid in overlap region.

It seems the calc_gt and calc_gs changes aren't
needed to make it work? The changes are different
to calc_gs and calc_gt are different for each
which is strange so I'm wondering if they should
be kept?

I've left calc_gt and calc_gs unchanged.

A.