ROMS Wish List

[arango]

We introduced this new forum thread, which can be used to propose new features to be implemented in ROMS. We do not expect to develop every feature proposed here or every ROMS user's wish list. However, we will consider those features that make a lot of sense. We have several advanced ROMS users out there that can help us to implement and test such new features. In past, we have received several good requests about new features and mostly all of them have been implemented.

My personal wish and to do lists are long. Some features are either trivial or difficult. Some interact with the integrity of the full code. Recall that we need to propagate all the new developments in the kernel to four models: nonlinear (NLM), perturbation tangent linear (TLM), finite amplitude tangent linear (RPM), and adjoint (ADM). Andy and I were in Toulouse (France) last month working with our colleagues at CERFACS (Anthony Weaver's group). We found that ROMS is now one of the most advanced variational data assimilation (4D-Var) systems and adjoint-based ocean frameworks in the world. This has been achieved because we have maintained the full integrity of the NLM, TLM, RPM, ADM codes in every change to the repository for years. On the other hand, this has slowed down further developments in the NLM. In retrospect, this was a very wise decision. We have several new exciting algorithms to be released in the near future that make ROMS really awesome with cutting-edge capabilities.

This forum thread will help us to document and discuss all new proposed features. Please use this thread specifically for new features. Any other discussion or bug reports will be deleted. Please help us to keep this forum thread clean

[linzhenhua]

Thanks for the great work from ROMS dev team. I think non-hydrostatic feature deserves to be on the wist list.

[lanerolle]

I think ROMS is a great model to employ for simulations and I suggest the following to be added to the wish list :

1. 2-way nesting/coupling
2. Multiply connnected domains
3. Implicit time-stepping (to allow for bigger dt when grids are very fine)
4. An improved bulk flux formulation (applicable to a greater array of regions around the world and also valid for extreme phenomena such as storms)
5. Wet-point only computations (to allow for smaller grid sizes and higher grid resolutions)
6. Adaptive grids (in time, in both the horizontal and vertical) to better capture hydraulic jumps and fronts
7. Fully 3D and complete reduced boundary conditions for tidal simulations (current BCs are valid for barotropic simulations only)
8. Ability to include stuctures and obstructions in ROMS (for engineering-type simulations, etc.)

Lyon.

[kate]

Let's go all the way and ask for an unstructured mesh version.

[m.hadfield]

Perhaps the word "wish" isn't quite what we need here!

Here's a wish, in the sense of something I have thought about and may well implement but would appreciate feedback on: an interior tracer point-source option (M_SOURCE?), each source to be defined in terms of its position, vertical shape function and flux in <tracer units> m^-3 s^-1. The code would adapt the existing point source infrastructure.

[ebjorkstedt]

How about code that does what I want, not what I wrote?

But seriously, a wee wish: FOUT that works like AOUT and HOUT. For FLOATS, that is.

[rduran]

Along-path relative and potential vorticity saved as output into the FLT00.nc file for any and all kinds of floats (just as temp, salt, lon, etc are saved). Planetary is straightforward from lat so no need for that one.