open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi again,

I use chart datum as a reference.If you want to see the real tide simulation go to www.previmer.org/previsions/niveaux/mode...zoneid%29/cot#appTop to see the surface elevation and thus Springs and Leaps in the zone where I do my simulation. I'll attache the geometric file so that you can take a look at the zone (it's a batymetry file).Afterwards,when the simulation works correctly,I'll extract the calculated data on two node points (79268 and 120086) because I have ADCP data to compare the results.The first point is located in UTM coordinates:zone 30,Easting 562750,Northing 550715 and the second zone 30,Easting 563076,Northing 5505052.Hope this is what you asked.
Regards,



Hanna

Hi Hanna,

The extent of your model seems to be 2.4 km x 4 km rather than 10 times more in each direction. Moreover, one digit is missing for your 1st point with coordinate North, but with one single location, I succeeded in seeing what you said.

You seem to raise a potential bug. I did some tests by changing the hour with your model and one model of mine, and like you, high waters and low waters seem to appear around the same duration with respect to time 0 s.
Nevertheless, I have the feeling that if you consider your simulation to start at midnight, you can compare your results with predictions.

Anyway, I am not used to modelling tides over so small areas. I would suggest you to enlarge the extent of your model, say 24 km x 40 km. I would also suggest you to include the shoreline in your model, I do not have any experience in modelling tide with only open boundaries.

Furthermore, some parameters seem exotic in your parameters file. Where do they come from? (e.g. STOCKAGE DES MATRICES = 1 - option 3 is now advised -, SOLVEUR = 6 - try 1 -, Finite elements are rather used to model tides ==> comment SCHEMA EN VOLUMES FINIS).
About Coriolis force, do not use COORDONNEES SPHERIQUES that work with SYSTEME GEOGRAPHIQUE = 5; then, you have to give a value for COEFFICIENT DE CORIOLIS = 1.11E-4 in your case (for latitude = 49.7 degrees N), otherwise the default value will be taken.
And in a next future, do not forget to use a LOI DE FROTTEMENT SUR LE FOND and give a value for COEFFICIENT DE FROTTEMENT SUR LE FOND.
Do not forget to give a realistic value for COTE INITIALE. The one you have taken prevent the solver from converging (see the listing printouts).
You should also give a PERIODE POUR LES SORTIES LISTING to avoid too many printouts.

At last, in your Fortran file, the subroutines come from a version 6.1 of TELEMAC-2D. BORD_TIDE and BORD_TIDAL_BC are useless in your case, you only need to start from TIDAL_MODEL_T2D (in particular, the BORD_TIDAL_BC you have taken was specific to the test case in version 6.1).

I enclose some first corrections for your parameter file and Fortran file.

To conclude for today, I suggest you to start temporarily all your simulations at midnight and let me investigate to find the bug (I do not when I have time in the next few days).

Hope this helps,

Chi-Tuan

Hi Chi-Tuan,

Thank you very much for your help and indications!
Unfortunately on my computer the simulation doesn't work that well.I'll tried yours without any modification,but it doesn't correspond to real tide data.At least it does 6h between the tide changing.Thus I recognized when searching for possible problems that my pc doesn't write anything in the "maree". I know that your FORTRAN file ask to do so and I dont now why it doesn't work.I'll try to find out the problem.
I'll equally try to enlarge the zone and include the shoreline.Perhabs this will help to get a good simulation.
Thank you very much another time!
Regards
Hanna

Hi,Chi-Tuan!

I just wanted to tell you that I can run the simulation correctly now with your file.Thank you very much!
I tested several days and found out that you can start at midnight or 12 o' clock in order to get the right simulation which correspond to the real data.You just have to put the right initial conditions to get a proper simulation.The other times seem,nevertheless,impossible to use if you want to observe the real phenomena of tide changes.
Thank you a lot!
Regards
Hanna

Hi,everybody!

As you have seen I try to simulate a tide phenomena.It works,starting at midnight or miday.I use a Telemac2d simulation using the data base TPXO for the tide simulation.
Nevertheless, when I compared the calculated results with real data on two of my mesh nodes, I don't get a corresponding response on the velocity. I've tested that's not an calculating problem of my PC by running the same simulation on another PC. They are both corresponding.
Concerning the velocity, I'm quite satisfied of the V component but the U component has a too small amplitude. I have no idea, why it is like this because normaly I should find the same difference between the real and the calculated data for U and V as they are linked.
In order for you to better understand what I mean,I'll attach the two figures. They are representing on the x axis the time scale and on th y axis the velocity in m/s. The blue line is what I have calculated and the red one ist the real data.
Thanks for all help.
Regards
Hanna

Hi Hannah,

Sorry for not having answered to your 3 weeks ago post, but I have not had so much time to investigate. I did some test for the old database, but I still have the same problem with taking into account time (except starting from midnight).

Anyway, for solutions coming from OSU (TPXO and other solutions) + NEA coming from LEGOS, time seems to be correctly taken into account.

For your current problem with TPXO (glad to see that users try to use it), can you send your steering file + Fortran file (if something concerning tides has been changed) + boundary conditions file, please? What about your harmonic constants solution, do you use the European Shelf, Atlantic Ocean or TPXO solution?

Moreover, can you tell me which geographic system you use, is it the same as your previous posts, it seems to be UTM North 30, does it? And for your ADCP data, is your North refered to magnetic North? I am not sure that it may have great influence, but it is just to be sure.

At last, can you telle me which reference you use for bottom elevation (Chart Datum or Mean Sea Level)? When modelling tides, it is advised to use Mean Sea Level reference.

Regards,

Chi-Tuan

Hi Chi Tuan,
thank you very much for responding.
Actually I changed to TPXO because I thought the problem might be because of the JMJ data base which takes only 4 harmonic constants into acount.By the way do you know the origine of the JMJ data base.It's just for documentation.
I use the TPXO for the European Shelf,as the ADCP is in the Channel.
Normaly, I didn't change anything in my steering file,nor fortran file.I only changed the data base.I'll nevertheless attach them together with the boundary conditions file, so you can take a look at it, if you want to.
I didn't try another time than midnight and miday as it was working for these with the JMJ data base.I'll try it now to confirm what you are saying

Yes I still use the geographic system UTM North 30.I don't know which North is refered to by ADCP,but I'll check it and tell you then.Nevertheless I can exclude the problem of which North I refer to because I also compared the norm of the velocity of both ADCP and TELEMAC.I just didn't post this figure.

Last but not least, my bathymetry is referenced to Chart Datum.I didn't change it to Mean Sea Level even if TELEMAC is referenced to it because I do not know how.

Regards,
Hanna

Hi Hanna,

JMJ database comes from EDF. Jean-Marc Janin (JMJ) was a former developper and user of TELEMAC a long time ago (I have never met him). You can refer to:
J.-M. Janin, X. Blanchard. “Simulation des courants de marée en Manche et Proche Atlantique,” EDF DER–LNH report HE-42/92.58, 1992.

You seem not having attended last TELEMAC-MASCARET Workshop nor Conference (no problem :)). You can find enclosed the slides I showed last week and the paper about modelling tides. It may help you. All proceedings and slides are supposed to be available on this website very soon, but if you cannot wait... :)








Moreover, I do not know if the enclosed report is available on the website, but it can also help you, even though it is only for version 6.1 (and in French only for this version, sorry). Version 6.2 is to be written in a close future (when I have time... and translated in English).




Anyway, I read your files and I have some comments to give you, the two first ones are the most important I think:
- if CORIOLIS = YES, you should give a value to 'COEFFICIENT DE CORIOLIS' (see the dictionnary to calculate it), as already told in a previous post. I think it should not be too different from 1.11E-4 (for latitude 49.7°). Otherwise, default value equal to 0 is taken (that may influence your horizontal velocities),
- you do not use any law of gottom friction 'LOI DE FROTTEMENT SUR LE FOND', so that there is no friction by default. In real cases, that must be more realistic (and also a way to calibrate your model). If using a law of friction (e.g. Strickler), do not forget to give a value to the friction coefficient 'COEFFICIENT DE FROTTEMENT', as already told in a previous post,
- with version 6.2, if you have tidal solutions coming from OSU (e.g. TPXO or European Shelf), you can initialise your computation with them (both elevation and velocities) rather than giving a uniform initial elevation. Please use keyword option 'ALTIMETRIE SATELLITE TPXO' for keyword 'CONDITION INITIALE' ('COTE INITIALE' is then useless), but it works better if bathymetry is refered to Mean Sea Level,
- if you want to prescribe both elevation and velocities, you should use 'OPTION POUR LES FRONTIERES LIQUIDES' equal to 2 (Thompson's conditions),
- if you want to calibrate your model (if needed), you can use the two coefficients 'COEFFICIENT DE CALAGE DU MARNAGE' and 'COEFFICIENT DE CALAGE DES VITESSES DE COURANT', in addition to the friction coefficient,
- if you have any wiggles for free surface or velocities, you can decrease the value of 'COMPATIBILITE DU GRADIENT DE SURFACE LIBRE' (e.g. for a case, I had to use 0.5 rather than 0.9), in particular when using Thompson's conditions,

For the reference of bathymetry, you can use CORFON subroutine to change it, by substracting half tidal range of an astronomical tide, coming from the harmonic constants database. I do not have any subroutine already implemented, but I would do like this.

Regards,

Chi-Tuan