The SWAN setup is combining wind and swell, so this should be a representative case. It has two files, representing two steps of nesting (100 m and 50 m resolutions). Therefore, we should run n1 first and only run n2 after n1 is finished.

Get the bathymetry file: scp ekman:/work/bast/lfs/wave-models/Hellerflesa.bot hellerflesa/swan/Hellerflesa.bot
In hellerflesa/swan/Hellerflesa_n1 submit run.sh.
In hellerflesa/swan/Hellerflesa_n2 submit run.sh (this nests into niva_1.nst created by the first run).

REEF3D

The REEF3D setup takes the Hs (significant wave height) from the SWAN simulation above, and we insert the offshore wave peak period (Tp) for showcasing the diffraction pattern.

When one needs to use a shorter peak period of waves (shorter waves as well), one needs to refine the grid, as we would like to keep 30-50 cells per wavelength (corresponding to Tp).

In hellerflesa/reef3d:

Get the bathymetry file: scp ekman:/work/bast/lfs/wave-models/geo.dat hellerflesa/reef3d/geo.dat
First submit 01-run-divemesh.sh (takes only seconds).
Then submit 02-run-reef3d.sh.

The two steps are in separate job scripts because DIVEMesh seems to use OpenMP and REEF3D uses MPI.

If you change number of cores, also adapt "M 10" in ctrl.txt and control.txt.