Small errors in the calculation of Gulf Stream separation and North Atlantic latitude

[oceandie]

I believe the code to calculate the Gulf Stream separation and North Atlantic latitude has two errors:

1. It assumes that the horizontal grid is a geographical grid (i.e., lon and lat are 1D), while it is a curvilinear grid - see https://github.com/JMMP-Group/MARINE_VAL/blob/main/SCRIPT/cal_GS_sep_NAC_lat.py#L28 -> I don't think this is creating a massive error (especially for the GS), although strictly speaking is not correct.

2. is using deptht (see https://github.com/JMMP-Group/MARINE_VAL/blob/main/SCRIPT/cal_GS_sep_NAC_lat.py#L17), which is fine for standard z-levels with partial steps but is not correct with generalised vertical coordinates (e.g., MEs).