I am running ROMS on a semi-enclosed embayment with result output at each time step (10 seconds), and the result shows that volume is not conserved as I compute the volume fluxes at the four boundaries of a water column, of which the grid size is 555 m * 567 m. The sea surface level of the water column is controlled by the M2 tide with amplitude roughly 2 m, but the total boundary volume flux stays positive all the time, indicating that the volume of the water column is continuously increasing (see the image below).
![Image](https://www.dropbox.com/s/62pwk6ekykbvrts/zeta_Q.png?dl=0)
Is the grid too big to test the volume conservation, or there is some other issue missed? Any suggestion is welcomed and appreciated.
Cheers.
Wencai