Coastal aquifer: 3D density variability

Hello all,

I am working on a coastal aquifer 3D model, for wich before getting to the model itself, I am doing some previous simpler models in order to gradually understand the process before getting to a more complex model.

I am having problems to model a very simple variable density 3D model. It consists of a rectangular mesh with 6 Layers (1000x2400x45; length perpendicular to the coastline x length along the coastline x depth). Everything is fine if I run the simulation with a density ratio of 0 (i.e. without considering density variability), but once I apply a value of 0.025, the head and concentration distribution calculated are senseless (i.e. orders of magnitude without any sense like e+12 for the head ).

The BCs I apply are:

- Constant Saltwater Head at the seaside boundary
- Constant Flow (Neumann) at the inland boundary.
- Constant concentration (35g/L) at the seaside boundary. I apply a min-mass flow constraint of 0, in order to desactivate the BC where water flows out of the model (i.e. the shallow part of the boundary)
- Constant concentration (0 g/L) at the inland boundary.


Thank you

Freshwater-saltwater interface involves very sharp transition between different concentrations that cannot be modelled with a coarse mesh (not dissimilar to "High head error in Richard's steady case" topic on this forum).
So to model the problem you need to either:
1) increase dicretisation (may not be possible)
2) smooth transition from freshwater to saltwater by increasing dispersivity and/or switching on "streamline upwinding" option.

Additional notes:
- generally, you don't need constant concentration of zero on landward side (mass cannot appear from  recharge BC anyway)
- it advisable to use "saltwater head" BC type if model top elevation changes below the sea level (in the data panel right click on "fluid flow" => "add parameter")

Thank you Durchlassigkeitsbeiwert,

I am trying the ideas you proposed I'll be updating on my advancings.

Meanwhile, does anyone know of other seawater intrusion 3D models that may serve as a reference to build mine? Any documentation on numerical criteria for this density-driven type models?

Also thanks for the additional notes. I want to clarify that I used the "saltwater head" BC option on the seaside.