How to remove the baroclinic effect

[zhaoyun]

Dear all,
I try to remove the baroclinic effect(remove the baroclinic pressure gradient), that is, I want to run the model in barotropic. I define the 'DJ_GRADPS', so the code correspond to 'prsgrd32.h'.

ru(i,j,k,nrhs)=on_u(i,j)*0.5_r8* &
& (Hz(i,j,k)+Hz(i-1,j,k))* &
& (P(i-1,j,k)-P(i,j,k)- &
& HalfGRho* &
& ((rho(i,j,k)+rho(i-1,j,k))* &
& (z_r(i,j,k)-z_r(i-1,j,k))- &
& OneFifth* &
& ((dRx(i,j)-dRx(i-1,j))* &
& (z_r(i,j,k)-z_r(i-1,j,k)- &
& OneTwelfth* &
& (dZx(i,j)+dZx(i-1,j)))- &
& (dZx(i,j)-dZx(i-1,j))* &
& (rho(i,j,k)-rho(i-1,j,k)- &
& OneTwelfth* &
& (dRx(i,j)+dRx(i-1,j))))))

My question is: Is it right to set the ru=0, rv=0, so that the baroclinic pressure gradient is removed? My model is OK, but after I set the ru=0 and rv=0, the model blow up. Thank you for your attentions.

[kate]

If you want a barotropic solution, you turn off SOLVE_3D and set ndtfast to 1. What more are you looking for?

[zhaoyun]

Thank you for your quick response, Kate. I need to keep the 3D structure of temperature, salinity and velocity. I mean that:
The density gradient term is omitted from the momentum equation, though the density field still evolves with time, which can affect the mixing. The ‘barotropic’
solution is, strictly speaking, ‘zero density gradient’ solution (i.e., the portion of the circulation driven by forces other than the density gradient).
I do not know how to change the model code. May it be 'ru=0, rv=0' and ndtfast=1?

[kate]

You want the variations in T and S but not density? You can use a linear equation of state in which rho is uniform.

[zhaoyun]

Thank you, professor Kate. If the rho is set to be constant, the vertical mixing will be changed. This is not the case I wish.
I want to delete the horizontal pressure gradient induced by density gradient.
For example, upwelling can change the horizontal density gradient. How could I remove the baroclinic effect by the upwelling in the case with the normal T,S and Rho?