open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello

I would like to simulate sea surge events over the Atlantique coast. Actually, I succeed in running basic tide simulations with telemac2d, but when I tried to run the program with pressure and wind after to have modified the meteo.f subroutine and the keywords, the results are exactly the same. Apparently, the model reads the meteo.f modified file has the comments appear for each time steps.

I would be glad for any help or advice.

Best regards

naulinjp wrote:

Edit :

I found the problem with the wind as the keyword “coefficient d'entrainement du vent” was not set. A coefficient equal to 2.1E-6 has permitted to obtain more satisfying results. However, I still don’t solve the problem with the atmospheric pressure.

Hello

To activate the wind influence, you need first to put the keywork WIND=TRUE otherwise it will not be considered. Then, you can edit the coefficient of influence through the keyword COEFFICIENT OF WIND INFLUENCE. A final hint for the wind, if your model has tidal flat regions with drying areas, the wind will generate unphsysical velocities, thus you neeed to use the keyword THRESHOLD DEPTH FOR WIND (PROFONDEUR LIMITE POUR LE VENT) to tell to telemac to apply the wind influence only for wet areas which have water depth greater than the this theshold.

For the atmospheric pressure, also you need to activate it through the keyword ATMOSPHERIC PRESSURE (PRESSION ATMOSPHERIQUE) =TRUE.
The way you have implemented the routine meteo is correct however the conditions you have imposed for wind and pressure do not vary in time and space, which is at least unrealistic for surge events. If you have real measures (or data) for wind and pressure you can edit the subroutine meteo and then you can see the influence of these two parameters in time.

I hope that this helps.

King regards

Riadh ATA

Thank you for your help.

In effect, as you said, the use of homogeneous parameters is not realistic but it was just tested to take in hand the model.

Since, I have modified the meteo.f file with distributed wind and pressure field (homogeneous in time for the moment) and it seems to be functional at first glance. In return, the time of running has passed from 6 min to 3 hours, but I suppose it's normal.

I was just asking on which criteria the COEFFICIENT OF WIND INFLUENCE should be chosen. Is it only a fitting parameter?

Thank you again

Hello

The theory behind the implementation of wind in Telemac can be found in the user manual (english) page 41.
Try to use parallel version to reduce CPU time.

With my best regards,

Riadh

Hello

I come to you again as I still have problems with the implementation of atmospheric pressure. The model is actually applied in an area of about 10000km² near the city of La Rochelle.

It appears that the results obtained with pressure forcing (purple curve in the joined figure) are not coherent with the input pressures. These inputs are atmospheric pressure fields at the mean sea level and are distributed in space and time. Normally, the results should be similar with the orange curve obtained after the application of a classical relationship of 1cm of elevation for 1 mbar under the reference pressure. I equally find that homogeneous in space pressure fields drive the model to produce the same results that without pressure.

So I have the following questions:
- What is the pressure model used / How is implemented the atmospheric pressure in the model?
- Why homogeneous in space pressure fields have no impact over the simulations?
- In the manual user, it is explained that, by default, the pressure is fixed at 100000 Pa. Is this value defined in a subroutine? Is it possible to change this value (by 101325 Pa for example)?
- To improve the simulations, I think it is possible to adjust the “COEFFICIENT DE CALAGE DU MARNAGE” and “COEFFICIENT DE CALAGE DU NIVEAU DE LA MER. Do you have any advice about these two parameters or other that could improve the simulations?

With my best regards,

Jean-Philippe

Hello,

In some sense, the equivalent of atmospheric pressure translated in water height (hydrostatic theory, with proper sign) is added to the free surface, and then we take the gradient of all this. So only gradients of atmospheric pressure will do something, which answers your question :

- Why homogeneous in space pressure fields have no impact over the simulations?

(think of a swimming pool, the level does not change, whatever the atmospheric pressure)

So only gradients in your computation domain will trigger some difference.

If I understand well, you expect that if you increase the atmospheric pressure on a domain with open boundaries the water should flow away and the level go down of a quantity corresponding to hydrostatic theory. We all intuitively expect this but I do not think that it is a well posed problem, because the atmospheric pressure outside your computation domain is not given: it is a missing data that should probably be given through a specific boundary condition.

Last remark: I find strange that your model skips from 6 mn to 3 hours just by adding wind or pressure, unless they are the only forcing terms.

With best regards,

Jean-Michel Hervouet

Hello

Thank you very much for your answer. I understand better now the results obtained by the model. As you are implying, the most rigorous solution is probably to integrate the atmospheric pressure into the boundary condition. However, given the final objective to apply the model in a consistent number of sector and events, I would like to avoid as much as possible the generation of complex boundary conditions. That’s why I tried to introduce the pressure-water level relationship mentioned before as a post-treatment step in order to correct roughly the water height estimates without using the gradient method. The result seems more coherent with the observations even if the method is less satisfying from a physical point of view.

As you have noticed, the simulations present also, for the moment, a delay with the observation that could reach 3 hours. I still don’t know exactly what the problem could be; even if the wind and pressure are actually the only forcing terms. Is it possible that the problem comes from the TPXO harmonic constant? Or that it comes from the geographic projection system in Lambert II for the geometry file and WGS84 for the TPXO database?

Best regards

Hello,

For the question on TPXO, we'll have to wait for Chi-Tuan Pham's return from holidays. As for the delay between applying wind and the response in the basin, it should be somewhat related to or even larger than the travel time across the domain of a long wave with celerity ?

With best regards,

Jean-Miche Hervouet