one question about boundary condition in an idealized test

[ywang152]

Dear all,

I am trying to run an idealized test, only having free surface as the boundary condition to drive the water current inside of my domain. The initial conditions are constant all over the domain:

Code: Select all

u0=0;
v0=0;
temp0=10;
salt0=10;
zeta0=0;

However, I found that the simulated 'zeta' inside of my domain are always around 0 (~10^(-18)). It seems the boundary free surface didn't influence the inner water level at all... I suspected that my CPP option in the header file may have problem, so I used UPWELLING case CPP options ( but #undef ANA_GRID/ #undef ANA_INITIAL/ #define GLS_MIXING). Unfortunately, this doesn't work. I plotted the time-series of simulated free surface on boundary cell and inner cell:
(sorry for the inconvenience. It seems there is a problem of PHP installation on the forum. So I attached URL instead)
https://drive.google.com/open?id=1lNY4v ... exzJFHo18_

The boundary and initial condition files I used are:
https://drive.google.com/open?id=15EFJ6 ... DxaxW9QOWV
https://drive.google.com/open?id=1KmIGy ... q591sauSE2


I have thought the barotropic item ∂η/∂x would work at least....Please let me know if you have any idea. Thanks for your attention!

Best regards
Yifan

[kate]

Dynamically, the zeta outside is not playing a direct role, though it can be used in Flather type open boundary condition. You don't say what open boundary options you have chosen. In the model zeta is a second-order field in that it is computed from lateral convergences and divergences - the barotropic velocities are the first-order fields that go into it. For all the people wanting to drive tides with elevation only, you will do far better by driving with velocities.

[ywang152]

Thank you Kate!

The boundary type for free surface I used before was 'Clamp'. If I understand right, the Flather type is for 2D momentum ubar/vbar. So I tested to use real ubar/vbar/uvel/vvel to be my boundary condition, with boundary type: Fla/ Fla/ Cla/ Cla. The initial u, v, ubar, vbar, zeta are all interpolated from the same data source as boundary conditions. Initial temperature and salinity conditions are constant value 10.

Code: Select all

!                   W       S       E       N
!                   e       o       a       o
!                   s       u       s       r
!                   t       t       t       t
!                           h               h
!
!                   1       2       3       4

   LBC(isFsur) ==   Clo     Clo     Clo     Clo         ! free-surface
   LBC(isUbar) ==   Clo     Clo     Clo     Fla         ! 2D U-momentum
   LBC(isVbar) ==   Clo     Clo     Clo     Fla         ! 2D V-momentum
   LBC(isUvel) ==   Clo     Clo     Clo     Cla         ! 3D U-momentum
   LBC(isVvel) ==   Clo     Clo     Clo     Cla         ! 3D V-momentum
   LBC(isMtke) ==   Clo     Clo     Clo     Clo         ! mixing TKE

   LBC(isTvar) ==   Clo     Clo     Clo     Clo \       ! temperature
                    Clo     Clo     Clo     Clo         ! salinity

This time I found the simulated free surface on boundary cell and inner cell looks close to each other, which is ok, while the another thing that I don't understand is the simulated free surface has a high frequency oscillation:
https://drive.google.com/open?id=1DtaaG ... RrDvCvLKG_

I checked the surface u/v and ubar/vbar boundary condition, but didn't find this kind of noise.
https://drive.google.com/open?id=1QHXjH ... vkNnaZhuRq
https://drive.google.com/open?id=1bZloN ... jY33V-jlTt
https://drive.google.com/open?id=1wSS3H ... vEQ6QLh_9U
https://drive.google.com/open?id=1i51vV ... O61vV5JSLM
The noise also appeared in the previous test. I attached the figure with only the free surface time series of inner cell. Even though the magnitude is really small, the noise did exist.
https://drive.google.com/open?id=1642ke ... Pill06Ct97

Could you please give me some advice about how to remove the noise? Thank you!

Best regards
Yifan