open TELEMAC-MASCARET The mathematically superior suite of solvers

Hay All

I am currently working with Telemac-2D model on the Belgium coastal areas. I encountered a problem in adjusting the bottom friction coefficient. For the optimum(from sensitivity analysis)value of the coefficient(manning's), there exists high underestimation of amplitudes and to a smaller extent shift in phase values. The model is forced at the open boundaries by Europian Shelf dataset from TOPX. I am also using this same dataset as a reference in the calibration process. I have checked for possible errors in the open boundary and found none. I have tested the model performance by changing the open boundary data forcing.I used the Atlantic Ocean dataset from TOPX and found similar errors as before. what shall I do?

I have attached the steering file and grid files.

thank you
Biniyam S.

Hello,

You should be very careful in your case with numerical parameters, such as implicitation coefficients which have been set at 1 in your case, which is diffusive for waves. You can try to get closer to 0.5 and if you have stability problems, act on the free surface gradient compatibility, by reducing it. Another point is that you have calibration coefficients to prescribe tides, this will be another answer that we are preparing.

With best regards,

Jean-Michel Hervouet

Hello
in order to calibrate the wave features you have the following keywords that could help:
-COEFFICIENT TO CALIBRATE TIDAL RANGE: which calibrates the range of the wave height
-COEFFICIENT TO CALIBRATE TIDAL VELOCITIES which calibrates the velocities
-COEFFICIENT TO CALIBRATE SEA LEVEL

please have a look on the new user manual of the v7p0 release where all these features are well explained.

with my best regards
Riadh

Hi everyone,

I am supervising the work of Biniyam.

About the calibration coefficients. I have the impression that the idea would be to amplify the tidal signal provided by the OTIS dataset. However, we checked our model results close to the open boundary, and they are very good. It is when we check them in the area of interest (further from the open boundary) that the tidal signal seems dumped.

Therefore, we have the impression that our model is too diffusive and we would be glad to hear your suggestions to solve that problem.

Thank you for your help.

Hello,

Then the implicitation coefficients are the first thing to try, as they are rather important for wave propagation.

I would advise also to avoid the Smagorinski model, that must give very big diffusion coefficients if you have large elements, and to ask for more accuracy of the solver. At least 1.E-6 rather than 1.E-4.

With best regards,

JMH