We are looking into this, but for the majority of cases it might be that there is already a significant improvement related to this with the new "Clip meshing" which was released with version 11.0 of the software, but in a prototype state, please ensure outputs are reviewed before use. The reason it provides an improvement is that it does a coarse mesh first (which is likely to remain unchanged if only small amendments are made) and follows this with a more detailed mesh after. It would be great to get any feedback you might have on your experiences with the Clip meshing and consistency of results.
The clip mesher is okay. But it creates very poorly conditioned cells in general, but particularly around mesh zones, i.e. cells that are not triangles, tiny 'slither' type cells, cells with concave faces etc.
It doesn't really seem to properly incorporate the mesh zone faces in the mesh - rather than making a mesh that smoothly incorporates mesh and BL vertices it just creates a base mesh and then adds new faces along the boundaries of mesh zones/BL, which leads to really poor mesh quality where you most want it to good (i.e. at culvert inlets and outlets).
This particular idea, of local mesh adjustment has a lot of merit, especially as our 2D models get larger and larger and is already common in other modelling packages.
In general, the mesher in ICM is considerably behind competitor packages (even HECRAS in many respects) and would really benefit from additional investment/refinement.
Great to see SubGridSampling added recently :)
Thanks for the idea! The Clip meshing option introduced in version 11 (May 2020) is intended to only adjust the mesh edges required to accommodate new changes and the performance of meshing has been significantly improved over the past few years. These features were included as an alternative to local remeshing.