Question about FLAGS to introduce river sources

[isoares]

Dear all,

Does anybody knows what is the difference between the flags Q_PSOURCE and UV_PSOURCE that we use to make rivers in ROMS ?

I am running a few experiments with the Amazon River. Because the Amazon discharge is so large, I had to introduce 1000 sources, so the amount in each source is small. I made a pool that is connected to the river through a channel, and the sources are located in the pool. I am using river_Vshape to concentrate the river discharge in a few top layers.

I see that when I use the flags Q_PSOURCE, UV_PSOURCE and TS_PSOURCE I introduce a low salinity anomaly and a volume anomaly, so that SALINITY decreases and ZETA increases inside the pool. I am using 0 for salinity in the sources.

However, when I use only TS_PSOURCE and UV_PSOURCE flags I dont have the volume anomaly and ZETA is not increased inside the pool.

WIKIROMS website does not mention Q_PSOURCE. It only make reference to UV_PSOURCE and TS_PSOURCE.

I am little confused about how rivers are done in ROMS. As I understand, we need both volume (mass) and low_salinity sources to create the barotropic and the baroclinic Kelvin waves that occur when the River flows into the ocean.

Does anybody has an input in that matter ?

thanks in advance

Ivan

[wilkin]

Look carefully through step3d_t.F.

You will see that the Q_PSOURCE case adds the point source flux directly to the divergence of vertical advection. So it is as if the source miraculously appeared in the middle of the cell.

How this differs from the UV_PSOURCE case is that the UV option adds a lateral flux across the user's defined vertical face/edge of a tracer cell at the appropriate u or v point on the staggered C-grid. The user must define the u,v point face carefully - especially with respect to the sign of the river_transport because this affects whether the flow is in to or out of the cell - i.e. we allow for a "sink" being a negative SOURCE.

Sources specified this way cause the flux across a cell face to be SET TO the source flux - the flux is not ADDED to predetermined flux in the middle of the open ocean. So it really only makes sense to do this for faces that coincide with the land/sea boundary defined by the masking - hence the i,j setting is crucial.

Once the UV_PSOURCE flux through a face is set, ROMS takes the divergence of the lateral fluxes on opposing faces of a using the user's chosen advection algorithm. In principle the tracer divergence is therefore effectively the same as the Q_PSOURCE case.

Where these two differ is that the Q_PSOURCE cannot input horizontal *momentum* to a cell, whereas UV_PSOURCE can.

In truth, I don't think many users have exercised the Q_PSOURCE code aggressively.

[isoares]

Thank you very much Dr. Wilkin. I was mistakenly thinking that UV was a source of momentum, which was added in the momentum equations, while Q was a source of mass introduced in the continuity equation.

The name UV_PSOURCE is confusing, it leads us to think that it is a SOURCE, but it is not since, as you said, "the flux across a cell face is SET TO the source flux".

Your comments helped a lot.

Thank you

Ivan