Blow up due to stratification

[staalstrom]

Hei

I am running a fjord application, with one open boundary in the west.
The horisontal grid size is 100m, an the model has 20 layers.

One TS profile is used as initial condition for salt and temp (the stratification is homogeneous in the horisontal). u, v, ubar, vbar and zeta is set to zero in the inifile.
The clm file is exactly equal to the ini file but with two timesteps.
The bry file has the same TS profile and u_west, v_west, ubar_west, vbar_west and zeta_west is set to zero.
The forcing is two tidal components and rivers.
The model run well with a TS profile with a thin fresh water layer.
The time step is DT=15 seconds and NDTFAST=60 (fast timestep is 0.25 sec).

But when I do exactly the same with a profile with weaker stratification the model blows up.
A depression wave is propagation from the boundary and into the modelarea.
The maximum depth in the model is 435m.

Andre Staalstrom