Best practice for including surface waterbodies in dewatering model setup?

[andreitablan]

Hi all,
We're developing a dewatering calculation workflow where we model drawdown in groundwater due to pumping. Our current approach uses a steady model for initial conditions (natural baseline), then a transient model to simulate the dewatering effect over time.

We now want to account for surface waterbodies (rivers, lakes, canals) as boundary conditions in addition to the phreatic line, since they influence both the initial water table and the drawdown behavior in nearby areas.

We understand that we should represent these as RiverString elements in TimFlow, but we're unsure about the conceptual approach:

Should we add the surface waterbodies to the steady model (natural baseline conditions), the transient model (dewatering scenario), or both?
Which TimFlow model properties/functions would you recommend for setting fixed-head boundary conditions on surface water features?
Any guidance on the best practice would be helpful! :))

[mbakker7]

Thanks for your question. I assume you are interested in the lowering of the head because of your dewatering activities. For that, you only need a timflow.transient model. timflow.transient simulates the change in head caused by the transient stresses: in you case the change in head caused by turning on the dewatering pumps. If the water level in the surface water does not change in response to the dewatering (which is common), you can specify them with the River or RiverString element and by specifying tsandh='fixed', which means the head change is fixed to zero along the element(s).
Hope this helps.

[andreitablan]

Thanks, that helps. We already use a steady model to establish the initial state, and then a transient model for the dewatering phase. The surface-water levels are assumed constant during dewatering, so I will represent them in the transient model as fixed-head RiverStrings.
Just to confirm: would you recommend adding them only in the transient model, or also in the steady model if we want their influence included in the initial baseline?

[MattBrst]

Well, if you are only interested in the drawdown (head change) caused by the dewatering, then you do not need an initial state. The initial state can be seen as zero. This is what geohydrologists call superposition modeling.

If you have other reasons to make a steady model as well, you'd better add the river in the steady model as well.