Wrong-boundary index in Northern OBC depth normalization

ze miscellaneous
1

Hi,
Can anyone help me to short it out ?
In the Northern edge gradient BC shows some Wrong-boundary index (possibly).
Its doesn’t match with Southern BC. South is OK.
I think it will:

            cff2=0.5*(h(i  ,Jend+1)+zeta(i  ,Jend+1,knew)
     &               +h(i-1,Jstr-1)+zeta(i-1,Jend+1,knew))

to

            cff2=0.5*(h(i  ,Jend+1)+zeta(i  ,Jend+1,knew)
     &               +h(i-1,Jend+1)+zeta(i-1,Jend+1,knew))

The Northern boundary depth correction reads h(i-1, Jstr-1)
the Southern boundary row instead of h(i-1, Jend+1),
so cff2 (the water column depth used to rescale momentum) is
computed from the wrong location, corrupting
the barotropic correction at the Northern OBC.

Thank you for your kind response.
Best regards

2

Hi Swapnil,

Thanks for spotting this bug. I’ll make the fix in the codebase.

Enzo

3

Hi Enzo,

Thank you for your kind words. can you check this also…

missing clamp (negative salinity)

I think it will be:

          Ts=t(i,j,k,nstp,isalt)
          sqrtTs=sqrt(Ts)
to
          Ts=max(0.,t(i,j,k,nstp,isalt))   ! clamp negative salinity
          sqrtTs=sqrt(Ts)

I also read rho_eos.F and alfabeta.F both are correct.
It's very useful because in models it will be used everywhere (SALINITY).

Thank you and best regards,

4

Hi Enzo, @eleboued

Can you please check this also. I think:


qdmu_nbq_west(j,k,1)=qdmu_nbq(Istr  ,j,k)
qdmu_nbq_west(j,k,2)=qdmu_nbq(Istr+1,j,k)

to

qdmu_nbq_west(j,k,1)=qdmu_nbq(Istr-1,j,k) !ghost
qdmu_nbq_west(j,k,2)=qdmu_nbq(Istr ,j,k)  !interior

Can you also check south (qdmv).
Standard C-grid ghost-cell treatments (Shchepetkin & McWilliams, 2005)
Thank you with best regards.