Hello everyone,

I’m working on a flood modeling project using MIKE 21 HD, and I’ve encountered an issue I hope someone here might help clarify.

In my simulations, I observe that when I use a coarser triangular mesh, the model outputs (especially water depth) match reasonably well with observed data. However, when I refine the mesh (i.e., use smaller elements expecting more accurate results), the computed water depths become significantly larger than expected and diverge from the observed values.

I’ve checked the bathymetry, boundary conditions, and time step settings, and they all remain consistent across the simulations. I expected the refined mesh to improve spatial resolution and accuracy, so I’m puzzled as to why the opposite seems to be happening.

Has anyone experienced something similar with MIKE 21? Could this be related to mesh smoothing, numerical instability, or the way the software handles bathymetric interpolation at smaller scales?

Any suggestions or insights would be greatly appreciated.

Thank you!
Evans Asante.