ROMS suitable for bottle-necked strongly stratified estuary?

[alkx]

Hello ROMS users,

I believe everyone that could afford to go is in Rio by now, but I will ask anyway

Short version: can ROMS simulate strongly stratified water column in a moderately large temperate estuary that has a small mouth both horizontally and vertically, and this possibly for 1 year durations?

Detailed version: we are simulating flows in an estuary of a couple of hundreds of km2 which displays stable stratification throughout the year due to salinity: between 5 and 20 PSUs at the surface and 28-30 in deeper waters. Temperatures are stable in the bottom waters and variable at the surface. Pycnocline is quite sharp, and the flows from 2 main rivers dictates this consistent stratification. The mouth to the ocean is very small in comparison, around 300m horizontally and down to 6m vertically, whereas the estuary can get as deep as 50m in the center. Because of the sill it is a bit like a fjord in some ways. This estuary is microtidal.
- Would ROMS be capable of simulating a year long period (more if possible even) while keeping this stratification intact? Meaning that numerical dispersion would be kept to a minimum?
- Is there any special trick that is required to achieve stable stratification? Certain advection schemes, time steps, etc.
- Would grid sizes of 50m (if not smaller) or less be achievable using regular computers (not HP Cluster)? By that I mean something like a computer that does not be custom made, hence ranging from Core i7 base with 4-6 cores up to maybe dual Xeon 8 physical cores (8x2=16 in total). And obviously acceptable runtimes too (not 1 year for a year of simulation! ).

We would like to get an accurate representation of the circulation in order to calculate residence times. I must also precise that the end goal is to simulate the ecology within that environment, but I plan to create a separate topic once I have the answer to this one if that's OK.

Any comments (though preferably constructive! ) are welcome!

PS: I apologize if this is not the right place for such a topic, and apologize again if this has been discussed many times before though I did search the forum.
PPS: I have searched for previous publications in similar environments and found beginning of evidence in Juan de Fuca, Hudson and Chesapeake estuaries for instance but none is so similar that I wouldn't need to ask

[ptimko]

I would think ROMS should be able to do this with few problems. About 10 years ago I was running a modified version of POM in an fjord that had near super-critical flow over a sill that had a coincident contraction. The horizontal resolution I used was 50 metres. I had 10 layers and depth range was 20 - 100 metres. I was able to get a reasonable simulation of the tides but could never get the solution to stay stable when I applied wind stress, however. Didn't try doing a 1 year simulation though. I only ran my model for about 35 days. I've thought about trying to repeat the simulation using ROMS but I've never got around to it yet.

[alkx]

Thanks, that helps! Any more experience from others in modelling similar environments?

[tony1230]

I am now running a biogeochemical case on a well-calibrated-validated hydrodynamic model, but still not run successfully. Stratification seems to me is an important stress that exerts a critical control on biogeochemical processes, say it blocks bottom-layer nutrients be carried upward by vertical transport. So recently i was planning to do a similar simulation on water stratification.

My hydrodynamic model run successfully for one year and reproduced majority features within different time periods (apr. and sep.) for both tidal currents, temperature and salinity. I have perhaps all forcing drivers like wind,airt,airp,airq,short and long radiation and rivers.

- Shou