open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,
I am currently trying to interface Telemac with the Previmer atlas of tidal constituents.
I have seen a post on the forum indicating a way to do it. It consists in interpolating phase and amplitude at the domain boundaries and then generating a text file containg those information.
Instead of building this text file (each time we create a new Telemac configuration), would it be possible to transform the entire Previmer database and create a Previmer file having the same format as TPXO ? By doing it, the use of Previmer data would be much easier (we would call the Previmer database exactly in the same way as it is done with TPXO).
Thank you.
Jérôme

Hello Jérôme,

If someone can provide a NetCDF reader of the PREVIMER atlases and a correct description of the initial files, I can have a look.
It seems that in the TPXO module, some harmonic constituents are missing at the moment like Tta1 and KQ1 and have to be added (and the implementation validated).

Anyway, you only need to generate a HARMONIC CONSTANTS FILE once if you do not change the geometry and the boundary conditions.
There are 2 main advantages of dealing with this manner:
- the size of this ASCII file is generally lighter than the TPXO tidal solutions,
- the reading and interpolation of this file is quicker than the one of TPXO for every computation.

At last, even if Previmer atlases can be transformed in the same format as TPXO tidal solutions, who can store these files and make them available to everyone and would Previmer owners agree doing that? If you have a contact, do not hesitate to give it.
I have recently discussed with Hamburg University for HAMTIDE database to do so and they agree (for release v8p5) but I think it should also be done for Previmer if we have to change the format of the input files.

Chi-Tuan

Hi Chi-Tuan,
Thank you very much for your answer. It is very useful to have your feedback on this.
I totally agree that TPXO, although very easy to use (and free), has several drawbacks (limited number of constituents, lack of comprehensive metadata, poor resolution in my zone of interest (English Channel)…). I really hope that using the Previmer atlas will improve my model performance.
Since I posted this message, I started to write a Matlab subroutine which interpolates Previmer tidal constituents at the boundary of the Telemac domain and then create an ASCII file (Actually, I wrote my own routine because I did not manage to use the tool of the LEGOS). Yet, I managed to force a Telemac model with one open boundary. However, I still have a problem (with the numbering of the nodes) when I considered a domain with multiple open boundaries. I am confident in solving this soon. When it will work, I will not need converting the Previmer files into a single file having the same format as TPXO.
You have raised important points as regards the transformation of Previmer files: who will store the files? do we have the rights to do it? I don’t have the answers to these two questions. Thus, I now doubt this is the best way to interface Previmer with Telemac.
I don’t have contact with the people in charge of Previmer database (I contributed to this project, but it was a long time ago…) but if you feel it is interesting to contact them, I can search for contact details (via colleagues working at Ifremer or CEREMA).
Last, good news for HAMTIDE!
Thank you for your help.
Cheers,
Jérôme

The last link I know about Previmer is:
marc.ifremer.fr/produits/atlas_de_composantes_harmoniques
If you have more recent information, do not hesitate to give.

Chi-Tuan

Hi Chi-Tuan,
In addition to my e-mail of yesterday, could you have a look at the enclosed files (geometry and boundary conditions files) and check that the header of the ASCII previmer file is OK.
I have a doubt on the numbering because the first node of the BC file is located in the middle of the western boundary.
Thank you.
Best regards,
Jérôme

It seems that the ASCII Previmer file was not attached in my previous reply. It should be OK now. Otherwise, these are the first lines:

2
1 2835
1496 1547
122 1 M
M2
-5.978940 48.584873 1.733689 115.080305 0.382616 98.337711 0.443496 55.518837
-5.938331 48.555702 1.741267 114.658587 0.377625 97.664706 0.452285 56.008136
-5.897722 48.526531 1.748740 114.209473 0.375898 96.814595 0.463144 56.300854
...

Hello Jérôme,

A few extensions are allowed on this forum (at least .slf, .cli, .txt, .zip), it may be your issue.

Have you read the French report:
opentelemac.org/downloads/TRAINING%20AND...ation_maree_v6p2.pdf

Appendix 4 describes the format expected for the header.

With your boundary conditions file, I think, the first lines should be:
3
1 36
1496 1547
2802 2835
122 37 M
37 if there are 37 harmonic constituents e.g.

Maritime boundaries are read from the boundary numbering: 1st is the more south-east then turn anti-clockwise. In your case, your western open boundary is split into 2 parts.

You should then order the line for every open boundary with the same numbering.

Be sure the name of the harmonic constituents is written as in the declarations_telemac2d module (in array MISC_CONSTID_BTM)

Looking quickly at your mesh, you have some issues which should be solved prior to running any computation I think:
- there are 5 overconstrained triangles (I have used the keyword CHECKING THE MESH = YES, e.g. around (1.6;52.2) in England,
- there is a small element (one edge of triangle is around 1^-6 m!),
- the growth of element size is too fast between some elements, it should not be so big (try not to increase more than 20% or even 10% from one element to a neighbour.

Hope this helps,

Chi-Tuan

Hi Chi-Tuan,

Thank you very much for your help.
With this numbering, it works well! Yet, I only considered M2 but I will apply the same procedure to the other constituents.

I will also use CHECKING MESH in the future. Actually, I did not know this option (very convenient).

Last, I will try to obtain the contact of the person in charge of Previmer data. It will take time as I will be on holiday next week and then I will travel for work. I will keep you in touch.

Best regards,
Jérôme

Hello Jérôme,

For next time (for you or other users), one way to generate the first lines of the harmonic constants file is to generate it from JMJ database, e.g. using the enclosed file (please update the name of the geometry + boundary conditions file + number of values equal to the number of open boundaries for keywords OPTION FOR LIQUID BOUNDARIES (or comment it) + OPTION FOR TIDAL BOUNDARY CONDITIONS. You have to use the modified tidal_model_t2d subroutine or change the standard tidal_model_t2d subroutine with TIDALBCGEN = .FALSE.

You only need to use CHECKING THE MESH once for 1 geometry file, no need to do it for every computation if using the same geometry file, you will save time.

Chi-Tuan

Thank you very much Chi-Tuan.