open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello

I am running a simulation using TPXO (h_tpxo7.2 u_tpxo7.2) as a BINARY DATABASE from volkov.oce.orst.edu/tides/global.html. However my results seems like

1. I have not obtained the amplitude and phase once I compared the model simulation results again the global Model OTIS

2. The results seems as there are certain instability.

Action I have taken until now

a) I have reduced the number of time steps from 60 to 30
b) I have refined a little bit my mesh because originally I had a very coarse resolution.

My questions:

1) How can I confirm that the boundary conditions I used are correct
2) Do I need use some subroutine to set up the boundaries
3) How can I confirm that the model is taken the forcing file correctly (I meant the TPXO)

Thank you
Regards
Joel

Hello,

I am not very familiar with the area you want to model in Mexico, but I am not sure that your geometry file is in the good geographic system. For a point located at (110°W, 23°N) close to Cabo San Lucas, it seems that its coordinates should be close to 602500 m E and 2543900 m N in UTM12N, not -12200000 m E and 2600000 m N does not it? You may have a shift with your East axis.

Could you confirm?

Chi-Tuan

Hello again,

Looking quickly at your steering file, you should use the regional solution "Atlantic Ocean".

The value you write for the CORIOLIS COEFFICIENT is not the one expected for 23°N (see the user or reference manual or the dictionnary).

The value of FRICTION COEFFICIENT = 0.6 with a Manning'law seems rather big.

You can also try to use both elevations and velocities files (with the 566 code rather than 544 code in the boundary conditions file + the option OPTION FOR LIQUID BOUNDARIES=2;2).

Hope this helps,

Chi-Tuan

Hello

With reference to your first answer

I am using the following Matlab code to convert lat lon to UTM

%% convert to UTM

R_earth=6371000.;

X=A*pi*R_earth/180;

Y=R_earth*log(tan(((B*pi/180)+pi/2)/2));


Where A y longitude and B latitude (both are incomes)

Secondly, I am using the Regional TPXO model (Pacific Ocean)and the results appeared similar as when I used the global model.

I will follow you recommendations about the boundaries conditions.

Finally the value I set up for Manning's law help me to stabilize the model a little bit (see attached figure)






I will contact you shortly to see what happened


Thank you
Regards
Joel

Hello,

What you use is Mercator's projection (GEOGRAPHIC SYSTEM = 5), not UTM, with keywords LONGITUDE OF ORIGIN POINT = 0. (default value) and LATITUDE OF ORIGIN POINT = 0 (you can read the subroutine conv_mercator_to_degdec to get the inverse formulae).

A small mistake in a previous answer, I wanted to talk about "Pacific Ocean" of course, but you have understood.
In the ocean, the results with the global solution TPXO or the regional solution Pacific Ocean should be similar. Differences may occur close to the coast if there are tidal flats in the TPXO-like solutions for the location where you want to interpolate tidal Boundary Conditions.

Chi-Tuan

Hello Chi-Tuan

I set up the simulation differently however the results seems worst than my previous simulations.

Attached you can see the steering file.

I did the follow

I changed my boundaries conditions
I changed the geographic system
I set up LATITUDE OF ORIGIN POINT = 0 and finally I used LONGITUDE OF ORIGIN POINT = 0

Any other recommendations to obtain better results.
Thank you
Regards
Joel

Hello,

In your area (North), the CORIOLIS COEFFICIENT is positive and the value corresponding to 23°N is not correct. Anyway, with so large areas, you should use SPHERICAL COORDINATES = YES.
I would recommend you to generate a new mesh with the origin of ordinates axis around the center of your domain (around 20°N) and change the keyword LATITUDE OF ORIGIN POINT to this value (around 20°N). I would also change the origin for longitude and also change the corresponding keyword to have a local origin in the domain.

For the OPTION FOR LIQUID BOUNDARIES, Thompson method (=2) + 566 code in the BOUNDARY CONDITION FILE may improve the results, but not always.

With so large area to model, you should use
INITIAL CONDITIONS = 'TPXO SATELLITE ALTIMETRY'.

Hope this helps,

Chi-Tuan

Hello Chi-Tuan

Before setting up again the simulations I would like to be sure that the geo file is in the correct coordinates system. I know that I can shift the axis using BK (Blue Kenue) Edit>Shift in X/Y> and write the coordinates of the axis. My questions are

How can I know which coordinates need to set up in the shift axis.

Do I need to select one point to set up the coordinates of it in order to move the image in BK.

How can I confirm whether I am working in the right axis position.

Thank you
Regards
Joel

Hello,

Sorry but I cannot use BK. I would recommend to have a Mercator projection with an origin around the location (240°W;20°N), around the center of your model, and use the two keywords LONGITUDE OF ORIGIN POINT and LATITUDE OF ORIGIN POINT with the values -240. and 20.

To check that your transformation is OK for TPXO module to compute the boundary conditions, you can write the LON and LAT arrays after the calls of CONV_MERCATOR_TO_DEGDEC (in CONDI_TPXO or BORD_TIDE_TPXO depending on the INITIAL CONDITIONS you choose) in a serial run (--ncsize=1) for a 1 time step calculation. The coordinates have to be in the area of your model.

Hope this helps,

Chi-Tuan