open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

Recently I have been attempting to model meteorological induced surges over the Arabian Gulf. The model domain is setup as shown in the attached figure with the open model boundaries extending quite far into the Gulf of Oman.

Tidal boundaries are forced with TPXO with free velocities and non-stationary wind and atmospheric pressure fields are applied from ERA5. I am having some difficulty in correctly treating the boundaries of the model as prescribing the water levels with boundary code 5 seems to lead to an incompatibility with wind induced surges. Consequently, modelled surges tend to resonate (likely at the natural period of the basin) which I believe is due to reflection at the boundaries.

Although some resonance may be natural due to the basin characteristics, the modelled resonance appears to be unphysical and is not entirely reproduced in results obtained from the Deltares GTSM global storm surge model.

So far I have attempted:
- Applying Thompson boundary conditions: No significant improvement and sometimes leads to unphysical velocities
- Applying 10 minute water level residuals at the boundary of the model domain without tides and only modelling atmospheric processes with no tides: Have not had any success as the value of the surge is assumed constant over the boundary which leads to the formation of velocity jets. Thompson boundaries are unable to suppress the issue.

Attached is a figure showing the surge residual obtained from subtracting water levels from two simulations; 1: Tides + Winds/Pressure and 2: Tides Only. Also plotted against the results is the GTSM modelled surge values at a similar location and the GTSM water level extracted at the boundary of the domain. There is a bit of an offset in the water levels due to slight differences in MSL between the two models (as shown in the GTSM water level at the T2D boundary included in the plot).

I was wondering if anybody has come across this issue before or has some tips that might help. I was hoping that non-reflective boundaries might be the answer, however at present it seems that Telemac doesn't have this functionality

Thanks in advance,
Toby

Hello Toby,

You can try to add an inverse barotropic effect to the predicted water level at the open boundary nodes as you have atmospheric pressure fields.

Hope this helps,

Chi-Tuan

Hi Chi-Tuan,

Thanks for the feedback.

I also stumbled across this as a common boundary treatment in the literature and plan on trying to implement it during an upcoming surge hindcast study. Will report back on the outcomes.

Regards,
Toby