modeling a river as an outer boundary condition, what slices do I have to use?

Hi there,

I have a question understanding the modeling of a river. Let's say we have a river as an outer model boundary in a 3d-model with ten layers, all slices are horizontally constant. So for example the rightern model boundary is the river. River bed in reality would be located within the second layer of the model area. Top sliced is phreatic, model is unconfined. I would like to connect the river with the aquifer via fluid transfer bc. with constraint (head = river bed). To make it simple, let's say river bed and even the head in the river is constant within the model area.

--> So do I use the fluid transfer bc on all slices? Or just on slice 1 and 2? Or in general, all slices from top slice to the river bed, but not below??

I guess I shouldn't use the bc on all slices, coz results show an outflow from my model below the river bed (below layer 2, where the river bed is located) in that case. I expected a vertical flow towards the river in the lower layers, and not a horizontal model outflow in those layers.

The flowfield in cross-section corresponds to my expectation, if I use the bc only on slice 1+2.

Could someone please confirm?
Thanks a lot....
h.

Indeed I'd also typically use slices 1 and 2. In case of a thick top aquifer, it might be worth putting an extra slice at approximately the depth of the river bed. You should be aware, however, that your model assumes that the location of the river is a perfect groundwater divide in this case - which might or might not be true.

Hi,
I'm currently building a model in which I've encountered the same issue as the original poster.
One of the rivers in my model is a model boundary. So far it is represented by a cauchy-BC in slices 1 and 2. Thus the same model boundary on the remaining lower layers of the model are considered as no-flow. Yet considering the model area, I'm actually expecting a groundwater flow out of the model domain beneath the river. How could that issue be addressed?
Any help would appreciated. Regards,

In many such cases a larger model area would be a good idea, at least if the outflow occurs in the same aquifer the river is in. Putting a boundary condition below the river can of course produce a model outflow, but - no matter what the BC is - most probably not reflect the natural conditions very well. For example, the aquifer on the other river side in reality will contribute to storage etc., and you cannot consider this by just putting a BC. As a personal opinion I think that groundwater models are way too often cut off at rivers as assumed natural boundaries, while in reality there is significant flow across the river in the aquifer below, and the conditions on both sides of the river influence the groundwater conditions.