open TELEMAC-MASCARET The mathematically superior suite of solvers

hi everybody,
I am trying to calibrate my 3d model with the salinity, I do have in situs measurements along a river (I don't have the exact bathymetry of the river so I used deductions from maps and charts), I use as input at the sea border a monthly salinity. the salinity profil in the river globaly respect the form of the measurement but there is a difference of magnitude.




In my model surface concentrations are too high and bottom concentrations are too low. (I use recommended and default values)
eg: at surface : measurements <1 g/l but calculated >10g/l.
at bottom : measurement can sometimes be >36 but calculated barely exceed <31

in fact also there is an intrusion of "coin sale" at the outlet of the river

here my configuration for tracer

//
TRACER TRANSPORT
/

---

NUMBER OF TRACERS = 1

NAMES OF TRACERS = 'SALINITY KG/M3'

INITIAL VALUES OF TRACERS = 0

PRESCRIBED TRACERS VALUES = 33.5; 0

/TRACERS VERTICAL PROFILES =????

/ACTIVE TRACER (SAL)

DENSITY LAW : 2

//example de tracer-dry
COEFFICIENT FOR HORIZONTAL DIFFUSION OF TRACERS = 1.E-6
COEFFICIENT FOR VERTICAL DIFFUSION OF TRACERS = 1.E-6
/ACCURACY FOR DIFFUSION OF TRACERS = 1.E-8
/PRECONDITIONING FOR DIFFUSION OF TRACERS = 34
TREATMENT OF FLUXES AT THE BOUNDARIES = 1;2
//

what I tried:
-changing the turbulent model from K-E to Spallat-Almarat => crash "Not A Number" error,
-playing with the coefficient of diffusion of Tracer and the coefficient of diffusion of velocity => not expected results
-changing bathymetry of the river, my first deduction of the general bathymetry of the river was 3.8m I modified it gradually till 10m but it become worst.

thx for your help

Josias

Hi Josias,

Not familiar with salinity modelling in TELEMAC, but do you account for density differences? It looks like your salinity limit is fairly 'vertical', I would expect a salt wedge near the bottom.

Hope that helps,
Patrick

Hi Patrick,
thanks for your answer and suggestion.

No I didn't take account "density differences". How can I do it ?

"sorry for not translating "coin sale" (salt wedge)."
Yes I expect that salt wedge with more salt concentration near the bottom.

bests
Josias

Hi Josias,

Have a look at the Lock-Exchange example. I haven't done any of this simulation, it was just something I asked myself when I saw your figure.

Good luck,
Patrick

hi Patrick,
very interesting example and very detailed one, but after testing the different configurations proposed, I still don't find my way.
I am trying other configurations for tuburlence model this time , even if in Lock-exchange example it has been said not having a significant impact,.., but I still seeing my model too diffusive horizontally and not enough vertically ...

Anyway.... I am trying to use and compile the GOTM library following the instruction on gotm.net/software/linux/
after this cloning
Cloning into 'code'...
remote: Enumerating objects: 20714, done.
remote: Counting objects: 100% (849/849), done.
remote: Compressing objects: 100% (318/318), done.
remote: Total 20714 (delta 568), reused 785 (delta 517), pack-reused 19865
Receiving objects: 100% (20714/20714), 8.96 MiB | 2.69 MiB/s, done.
Resolving deltas: 100% (15188/15188), done.
Submodule 'extern/CVMix-src' (github.com/CVMix/CVMix-src.git) registered for path 'extern/CVMix-src'
Submodule 'extern/fabm' (github.com/fabm-model/fabm) registered for path 'extern/fabm'
Submodule 'extern/flexout' (github.com/BoldingBruggeman/flexout) registered for path 'extern/flexout'
Submodule 'extern/gsw-fortran' (github.com/TEOS-10/GSW-Fortran.git) registered for path 'extern/gsw'
Submodule 'extern/stim' (github.com/BoldingBruggeman/stim) registered for path 'extern/stim'
Cloning into '/home/ben/code/extern/CVMix-src'...
remote: Enumerating objects: 3334, done.
remote: Counting objects: 100% (10/10), done.
remote: Compressing objects: 100% (8/8), done.
remote: Total 3334 (delta 2), reused 9 (delta 2), pack-reused 3324
Receiving objects: 100% (3334/3334), 5.15 MiB | 3.87 MiB/s, done.
Resolving deltas: 100% (2399/2399), done.
Cloning into '/home/ben/code/extern/fabm'...
remote: Enumerating objects: 16022, done.
remote: Counting objects: 100% (790/790), done.
remote: Compressing objects: 100% (315/315), done.
remote: Total 16022 (delta 495), reused 712 (delta 422), pack-reused 15232
Receiving objects: 100% (16022/16022), 17.60 MiB | 2.25 MiB/s, done.
Resolving deltas: 100% (11250/11250), done.
Cloning into '/home/ben/code/extern/flexout'...
remote: Enumerating objects: 1038, done.
remote: Counting objects: 100% (124/124), done.
remote: Compressing objects: 100% (79/79), done.
remote: Total 1038 (delta 70), reused 87 (delta 36), pack-reused 914
Receiving objects: 100% (1038/1038), 245.44 KiB | 367.00 KiB/s, done.
Resolving deltas: 100% (493/493), done.
Cloning into '/home/ben/code/extern/gsw'...
fatal: unable to access 'github.com/TEOS-10/GSW-Fortran.git/': Failed to connect to github.com port 443: Connection timed out
fatal: clone of 'github.com/TEOS-10/GSW-Fortran.git' into submodule path '/home/ben/code/extern/gsw' failed
Failed to clone 'extern/gsw'. Retry scheduled
Cloning into '/home/ben/code/extern/stim'...
fatal: unable to access 'github.com/BoldingBruggeman/stim/': Failed to connect to github.com port 443: Connection timed out
fatal: clone of 'github.com/BoldingBruggeman/stim' into submodule path '/home/ben/code/extern/stim' failed
Failed to clone 'extern/stim'. Retry scheduled
Cloning into '/home/ben/code/extern/gsw'...
remote: Enumerating objects: 1058, done.
remote: Counting objects: 100% (79/79), done.
remote: Compressing objects: 100% (35/35), done.
remote: Total 1058 (delta 44), reused 75 (delta 44), pack-reused 979
Receiving objects: 100% (1058/1058), 10.86 MiB | 593.00 KiB/s, done.
Resolving deltas: 100% (725/725), done.
Cloning into '/home/ben/code/extern/stim'...
remote: Enumerating objects: 314, done.
remote: Counting objects: 100% (53/53), done.
remote: Compressing objects: 100% (13/13), done.
remote: Total 314 (delta 45), reused 40 (delta 40), pack-reused 261
Receiving objects: 100% (314/314), 163.34 KiB | 724.00 KiB/s, done.
Resolving deltas: 100% (186/186), done.
Submodule path 'extern/CVMix-src': checked out '365591d2269cbffe4e5873d7a7254b3d092fbb68'
Submodule path 'extern/fabm': checked out 'a0d56f0ce590dc35168e75a2ac1ca7bffd911fb6'
Submodule path 'extern/flexout': checked out 'c8fcca8375e08ed5f68f8a0f06aa23a9f528860a'
Submodule 'lib/fortran-yaml' (github.com/BoldingBruggeman/fortran-yaml) registered for path 'extern/flexout/extern/fortran-yaml'
Cloning into '/home/ben/code/extern/flexout/extern/fortran-yaml'...
remote: Enumerating objects: 286, done.
remote: Counting objects: 100% (34/34), done.
remote: Compressing objects: 100% (24/24), done.
remote: Total 286 (delta 20), reused 24 (delta 10), pack-reused 252
Receiving objects: 100% (286/286), 84.32 KiB | 385.00 KiB/s, done.
Resolving deltas: 100% (143/143), done.
Submodule path 'extern/flexout/extern/fortran-yaml': checked out '2d203312f4ee057673e6d5db94542ae770813048'
Submodule path 'extern/gsw': checked out '5d2271fc526033e68e297b080e2aff920db347bd'
Submodule path 'extern/stim': checked out 'c5537cebe0c85375e5e8aed1bb38c35ffb2f3c60'
Submodule path 'extern/gsw': checked out '5d2271fc526033e68e297b080e2aff920db347bd'
Submodule path 'extern/stim': checked out 'c5537cebe0c85375e5e8aed1bb38c35ffb2f3c60'

I got the error message following this command
"git submodule update --init --recursive "

error
"fatal: not a git repository (or any of the parent directories): .git"

anyway,..., following instructions for using Cmake for compiling/building the GOTM, I got this other error I don't know how to manage it





thx for your help
bests
Josias

Hello Josias,

I wonder:
- which advection schemes you use for velocities, tracers (and k-epsilon if using k-epsilon),
- how many horizontal planes and which mesh transformation you use over the vertical,
- if using non-hydrostatic version or not.

Can you share your whole steering file please?

What are the mean, maximum and minimum water depth along the river?

FYI, as I wrote in
opentelemac.org/index.php/kunena/21-tele...es-and-tracers#41260
and
gitlab.pam-retd.fr/otm/telemac-mascaret/-/issues/838
there was a bug when using Spalart Allmaras turbulence model with tracers. It has already been fixed on the main and will be available for next release (v8p4).

Chi-Tuan

Hi Chi-Tuan,
thx for your reply:
- SCHEME FOR ADVECTION OF VELOCITIES and TRACERS: MURD-PSI with SCHEME OPTION FOR ADVECTION OF VELOCITIES AND TRACERS 2 predictor-corrector of first order
I mainly tried those turbulence models
HORIZONTAL TURBULENCE MODEL = 1, 3
VERTICAL TURBULENCE MODEL = 1,2,3
I didn't tried Sparllat-Allmaras because very awkward results (excessively high values) when coupling T3D with TOMAWAC

- horizontal plane I tried
NUMBER OF HORIZONTAL LEVELS =3 [most of time] ,4[1 time),8[1 time] (nearly same results so I stay with 3)

MESH TRANSFORMATION =1 (never changed it)

-NON-HYDROSTATIC version : never changed because I apply variation of the pressure in the meteo_file, I think it may not take the pressure into account....

sharing of the steering file .... the last one ... I try to integrate the temperature to see its influence over the salinity or the density (not significant change).



this is an other steering file (one of previous Steering file) without temperature, I used the Lock-exchange steering file configurations (main one) I modified it, trying to use the PRANDTL number of the see water in the literature (7.2) and not only the value 1. I don't remember the results,
but I think it was not satisfying enough or maybe lack of perseverance... too tired of fails




I use constant depth of 3.8 or 4: values from salinity charts, because bathymetry not available. but I also tried 5,6,7 and 10 (not interesting results) .

Now I am trying to isolate the river for adding more nodes and planes and also having shorter computation time for being able to explore more configurations and combinations....
I want to integrate the flowrate of the sediment transport in the water column...

thx for your help
Josias

Hello Josias,

My first idea is to increase the number of horizontal planes. If you want to see something variing over the vertical, 3 planes will generally not be enough (1 plane for the bottom, another one for the free surface, that means you only have 1 intermediate plane to describe a variation).
Perhaps you can already try 11, 21 perhaps more? I do not know the behaviour in this area.

Looking at the provided steering files, I would recommend a few things:
- use values for IMPLICITATION FOR DEPTH and IMPLICITATION FOR VELOCITIES a little bit greater than but not equal to 0.5 (e.g. default values = 0.55) otherwise you may get unstability issues,
- try the LIPS scheme (which is now the recommended one for TELEMAC-3D and the default value) in particular for tracers: SCHEME FOR ADVECTION OF TRACERS = 5 + SCHEME OPTION FOR ADVECTION OF TRACERS = 4. It is less diffusive and works for dry areas contrary to Predictor-Corrector schemes,
- use TREATMENT OF NEGATIVE DEPTHS = 2 otherwise you may have negative depths for some dry nodes,
- if using k-epsilon model, it is not recommended to change the associated COEFFICIENT FOR HORIZONTAL/VERTICAL DIFFUSION OF VELOCITIES,
- MINOR CONSTITUENTS INFERENCE = YES is not recommended. Why have chosen to activate this keyword?
- for the diffusion solving, the default one (conjugate gradient) is efficient (combined with preconditioning for the operation = 34).

And keep NON-HYDROSTATIC VERSION = YES.

Is there a stratification to capture? Classical sigma transformation (MESH TRANSFORMATION = 1) is not always the best choice to model such phenomena.
You can try to change the distribution over the vertical (MESH TRANSFORMATION = 2) or use a z-layer with 1 or several fixed planes).

Hope this helps,

Chi-Tuan

Hi chi-tuan
Thx , I do what you said, and I give you back the results .
Josias

hi,
I modified my previous steering file, isolating the river, using water level from a previous computation in the liquid boundary instead of tpxo,.....

I got this error message when the computation is processing the steering file

utils.exceptions.TelemacException: Unknown keyword TRACERS VALUES in steering file CAS_FLVSP_SALTM73Q_H3V2TM_F4.txt

the fact is the word is well written in the steering file
PRESCRIBED TRACERS VALUES

I moved the keyword to an upper place, adding space before it and now,
the new error is about the keyword who followed the PRESCRIBED TRACERS VALUES , I mean the DENSITY lAW (well written) with

Unknown keyword lAW in steering file CAS_FLVSP_SALTM73Q_H3V2TM_F4.txt

thx for your help

josias