Overbank Spilling

Hi,

I am trying to simulate a situation where a culvert goes under (through) a built-up road, but where no defined channels exist in the area.  (i.e. the road runs through a wetland.)  I want to know where water is ponding up behind the road (calculated with the overland flow processes) and if I can solve that problem using a culvert.

Therefore, I've run the MIKE SHE model with no link to a MIKE 11 model to determine a baseline condition of where ponded water occurs.

I now want to use MIKE 11 to determine the volume of water that would flow through a culvert under the road at those locations.  So, I want the ponded water to flow from the MIKE SHE model into the MIKE 11 model upstream of the road, flow through the structure (culvert) under the road, and then flow from the MIKE 11 model back into the MIKE SHE model downstream of the the road, so that MIKE SHE can inform me where the water then flows due to overland processes, and if the culvert indeed solves the problem. 

Is this possible?  It seems that using the overbank spilling option should allow this to occur.  Here's what is not clear to me from the documentation:
1. can I put in a short MIKE 11 channel in the middle of my model domain that connects two "low" spots in the MIKE SHE model?
2. what should the boundary conditions be on the ends of the MIKE 11 channel for this to occur?  Right now I've assumed two "closed" channel ends.
3. can overbank spilling be used to essentially "empty" the MIKE 11 channel back into the MIKE SHE model in this situation?

Thank you for any clarification you can provide on overbank spilling and the ability to put in a short MIKE 11 channel to connect areas of lower topography!

Janna

Dear Janna

The solution that you describe could be done in MIKE SHE, buit is a bit tricky.  I would recommend the following approach:
•              Define a MIKE 11 channel coupled with MIKE SHE using the weir option (in this case the bank elevation in the cross sections determines when spill should occur)
• Use no flow boundaries at both ends of the MIKE 11 channel (no water flowing to the branch or leaving the branch)
• The exchange with the MIKE SHE model occurs using GW exchange and flood spilling (and OL flow but this only goes from SHE to M11). If I understand you situation correct you should define the cross sections equal to the topography (use a width of one cell). The overbank spilling will only spill water when the water level is above the bank elevations in the cross sections. The remaining part of the water in the channel could be exchanged with SHE using the M11 – SZ exchange, but requires that the water table in the M11 branch is above the GW elevation. For this part you might want to adjust the leakage coefficient.
• Another solution could be to use the flood code option, as this will ensure that water from the branch is “mapped” onto the terrain, and in this case you will actually have the same water level in the river and on the surface. The flood codes should then be used for the same area as you have your cross sections, as this will map the water onto the terrain, and it should then be allowed to flow from there. This might give you a solution that is closer to what you want.

I would recommend that you try this approach on a small “dummy” model to ensure that it works, and to adjust the parameters, before implementing it into the large model.

Thanks
Oluf (DHI-DK)

Hi Oluf,
Thanks for your previous post.  I am still working on getting this to function properly.  I end up having quite significant mass error in my MIKE 11  model using the flood code option.

Can I ask a clarifying question:
1. when you say "use no flow boundaries at both ends of the MIKE 11 channel", I have interpreted that as setting both ends as being "closed".  Is this the boundary option you mean?  I guess another way of interpreting that would be setting the branch ends as "open", with "inflow" boundary type, set to a constant of "0" discharge... I can't quite work out in the documentation how to set a no flow boundary in any other way.

Thanks again,
Janna

Dear Janna

Sorry for the late reply. Yes I mean using a flow boundary with Q=0 in both ends.

BR
Oluf (DHI)