open TELEMAC-MASCARET The mathematically superior suite of solvers

I am trying to force a tidal model of the Irish Sea using the TPXO global tidal database. My grid has two open boundaries at the southern entrance and the northern entrance. However, when I try and run the model using the TPXO tidal database no tide is imposed at the northern boundary (boundary 2). At first I thought this might be because no tidal constituents could be found in the database in this region but it seems that the routine just doesn't generate tidal boundary conditions for the second boundary. If I remove the southern boundary so that the northen boundary becomes boundary one tidal conditions are imposed on this boundary.

1. Can the TPXO tidal database be used to force two or more open boundaries with the given routines?
2. If so does anyone know why it might not be working for me when I include two open boundaries?

Many thanks,
John

Hello,

I am not the person in charge of TPXO, nor am I a specialist, but this is just to make sure that you are not only missing the hint that the keyword :

OPTION FOR TIDAL BOUNDARY CONDITIONS = 1

has become manifold, if you have several boundaries you must write e.g.:

OPTION FOR TIDAL BOUNDARY CONDITIONS = 1;0;1

my example being for 3 open boundaries, with tidal conditions on boundaries 1 and 3. It was not the case before and the tide was imposed on all boundaries, which is sometimes not wanted.

With best regards,

Jean-Michel Hervouet

Many thanks M. Hervouet. I worked it out through trial and error in the end as it was not clear in the documentation for v7p0 but very useful that it is now not universally applied.

Kind regards,
John

Hi,
I have found that the TPXO worked very well for two tidal boundaries with h,u,v specified for telemac2d V6.3 north and south with the east as a no flux boundary. However when I run the same model with version 7.1, it runs but I get incorrect results with for my case the tide flows all ways entering from the south over the entire tidal cycle, which is not the case as there is reverse on the change of the tide.

However when I specify a single boundary (h,u,v) in V7.1 that runs along the south, the east and the north the results produced are good and behave as expected. Therefore given that 6.3 works for two boundaries and V7.1 works correctly for 1 boundary only I am wondering if there is some kind of bug in the latter version when two tidal boundaries are specified. Has anybody else noticed this?

Regards
Tony

Hi
this is not a bug, just an evolution to enhance the management of boundaries when user mix tidal and non tidal boundaries.
Between 6.3 and 7.1 the format of keyword OPTION FOR TIDAL BOUNDARY CONDITIONS changed.
In 6.3, there was only one global value but now with the version 7.1, there is one value per liquid boundary

Regards

You miss understand me.
I understand the difference between 7.1 and 6.3 but I have found that latter version is not giving the correct answer when the two boundaries are applied and applied correctly (north and south opensea boundaries)
OPTION FOR TIDAL BOUNDARY CONDITIONS = 1;1
and when I apply a single boundary running along the entire opensea south-east and north I get the correct results with 7.1 and these compare well with 6.3 with two boundaries (north and south) applied (basically flow is rectilinear and the east boundary acts as streamline with velocities paralell to it)
I am just seeking clarification whether anybody else has found similar effects when using two or more tidal boundaries from TPXO.

regards
Tony

Hi Tony,

Are you the boundaries you talk about continuous liquid boundaries or are they divided by land? If they are continuous I think you have to specify them as one boundary even though they have different orientations.

w.r.t to my Irish Sea model I managed to run the model with TPXO boundary conditions and the results I get compared to tide gauges are better than my previous runs using 5 tidal constituents extracted from a wider area model. However, I created a model of the North Sea and implemented the TPXO boundary conditions. It seems that the simulated tide is significantly under predicted at all locations beyond what could be resolved with tuning the bed friction or due to the bathymetry. After an initial investigation it seems that no TPXO boundary conditions are implemented at the English Channel entrance for half the open boundary.

Is there a distance tolerance between the open boundary node and the nearest TPXO location or could there be another reason no boundary conditions are implemented along half the boundary?

If I set the keyword COEFFICIENT TO CALIBRATE TIDAL RANGE as 2.0 I get near perfect results near the open boundary but the results are less good inside the model domain with a significant phase error.

no it does not have to be a continuous boundary as the east boundary can be a streamline/ land boundary as velocities are almost completely parallel to it.
the earlier 6.3 version gave virtually the same result if I used a single continuous boundary or two boundaries (south and north) whereby the newer version gives incorrect results for tow boundaries and thus my query as to whether it is right.

From your post V7.1 does not appear to be working properly for your model, The Irish Sea is very rectilinear north and south directions, I assume your eastern boundary is almost a streamline line boundary and therefore you should be able to specify two boundaries if you wished. not sure why only half your boundary is given an elevation without seeing your model files, I have not encountered that problem before with TPXO.

regards
Tony

Hello

Is it possible to send us your model with both options. We will investigate the problem.

with my best regards

Riadh ATA

Hi Riadh,
I attach a another application which is of the IrishSea with southern and northern boundaries. I have tested this for the exact same files using V6.3 on Windows and V7.1 on UBUNTU with v6.3 giving expected results and v7.1 having a northerly flow direction throughout the simulation. I assume you have the TPXO databases as referenced in the CAS.

Regards

Tony C