Using sink option in MIKE21 and introducing Rainfall

I am trying to do a simple simulation for a small area with MIKE 21. The model area has two boundaries east and west.  Upstream boundary condition is a water level time series and a constant water level is given at downstream boundary. The model runs but still I am facing few problems.
1.  When I increase the US water level than a certain value, it gives an error says “ very high water level, near …..” with abnormal run completion.
2.  To eliminate the above problem, I introduce “Sink” option at some grids giving high discharge, but there are no effects to the outputs which I had before introducing sink discharge.
3.  Also, I add Precipitation (constant then I tried for a time series), but in this case also I could not find any effects to the results due to the rainfall.
I will be much grateful, if anyone could give some hints on these problems… I have got struck for couple of days due to these problems….
Thank you

Dear Moufar,
A couple of ideas that you could look into:
1. As you maintain a constant water level on the downstream boundary while increasing the water level on the upstream boundary, you may provoke very high velocities at the downstream boundary and/or a water level gradient just upstream this boundary. All dependent on the actual values of course, but this might cause the instability.
2. You could also inspect the upstream boundary in order to see if there are important irregularities in the first points or rows of the model
3. The introduction of a sink in order to lower the water level is not a good way to solve the described problem. The problem must have its origin in the actual bathymetry or in the boundary conditions. The sink will only complicate further the setup.
4. Regarding the precipitation, have you considered whether the amount of water that you add to the model through precipitation is significant and thus estimate how large the effect could be (for example in rise of water level)?

Best regards,
Poul Kronborg