Hi all,
I'm working on my research about conducting ideal simulation with real terrain.
However, I found an issue with the boundary condition. There is unrealistic convection at the boundary.
I used symmetric because periodic didn't make sense.
This is the land use index of my domain and if I set the boundary condition as periodic, the model thinks there is ocean in the north and east boundary side and there is land in west and south side.

However, it is not true. The blue and purple part (land) is a part of my country so there should be land in the onrth and east side.
So I chose symmetric and after running wrf I got the result below.

This is hourly rainfall and you can see there is rainfall on an upper-left side.
I also tried modifying share/module_bc.F like below.
I applied this in the all side of the boundary in the code.
This means the u or v wind will maintain their direction at the boundary, unlike the original symmetric condtion.
However it didn't work, and the result was quite the same.
If you have any ideas to reduce the convergence at the boundary, please let me know.
I'm working on my research about conducting ideal simulation with real terrain.
However, I found an issue with the boundary condition. There is unrealistic convection at the boundary.
I used symmetric because periodic didn't make sense.
This is the land use index of my domain and if I set the boundary condition as periodic, the model thinks there is ocean in the north and east boundary side and there is land in west and south side.

However, it is not true. The blue and purple part (land) is a part of my country so there should be land in the onrth and east side.
So I chose symmetric and after running wrf I got the result below.

This is hourly rainfall and you can see there is rainfall on an upper-left side.
I also tried modifying share/module_bc.F like below.
IF( variable == 'u' ) THEN
DO j = MAX(jds,jts-1), MIN(jte+1,jde+jstag)
DO i = 0, bdyzone-1
!dat(ids-i,j) = - dat(ids+i,j) ! here, u(0) = - u(2), etc
dat(ids-i,j) = dat(ids,j)
ENDDO ! normal b.c symmetry at u(1)
ENDDO
I applied this in the all side of the boundary in the code.
This means the u or v wind will maintain their direction at the boundary, unlike the original symmetric condtion.
However it didn't work, and the result was quite the same.
If you have any ideas to reduce the convergence at the boundary, please let me know.
Attachments
Last edited: