open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,

I have a tidal model that has high resolution (25 m) in a tidal channel where we are interested in the tidal currents. I use TPXO boundary conditions and I have calibrated the model to give reasonably good tidal elevations w.r.t to tide gauges (with some phase errors). However, within the high resolution tidal channel the velocity profiles become very jagged and are not smooth at all. I have attached some images to show that the velocity profile is smooth at the boundaries and outside the channel, but within the channel it becomes very unstable. I also attach the free surface at the same point in the channel which is relatively smooth but also has some small instabilities. I also attach one of the model setups (cas file, geometry file and conlim file).

Things I have tried:

- smaller time-step (5 secs)
- simplified smoother bathymetry within channel
- free surface gradient compatibility (0, 0.5 and 0.9)
- including propagation + option for diffusion of velocities = 1
- bottom smoothings = 2
- thompson/non-thompson boundary conditions

None of the above have made much difference to the stability of the velocity profile within the channel. As the profiles are smooth elsewhere is seems it is something to do with the higher resolution here. Every other tidal cycle seems to be very jagged in the profile.

Any advice would be much appreciated as I have run out of ideas to get a better solution!

Kind regards,
John

Images and model setup attached.

Hi
Some thoughts about your model...

Your using TPXO v9, did you update the subroutines according to recent posts dealing with a problem of S1 wave?

If I'm not wrong, the location of your model, in the north of scotland, where there is an anphidromic point. Maybe this could have an impact on the boundary but not fully sure

The shoreline of the island looks like a vectorization of a raster. In the high resolution part of your model, this kind of limit could probably have an impact.
Maybe you could use another shoreline and see the impact...

Regards

Thanks for your reply Christophe,

I assumed that in the version I am using (v8p0r2) the warning message w.r.t S1 wasn't having an impact and that the constituent was simply ignored. Further to this the velocity profiles are smooth at my boundaries and elsewhere within the model domain so I didn't think it was the route cause of the instabilities. However, I will update based on the recent information if this is now considered best practice.

You may have a point regarding the shoreline! It is from the GSHSS coastline data but this resolution is very stepped. I will try and create a better, higher resolution coastline.

I still feel like it may be something to do with the solution when it transfers from low to high resolution. Maybe the gradient of change in the resolution is too high?

About the Shoreline, I suspected the use of GSHHG data.
On the NOAA web site, you could find those data with a higher resolution.
Look at Noaa shoreline

I agree on the fact you could smooth the transition between low and high resolution in the interesting channel...

regards