open TELEMAC-MASCARET The mathematically superior suite of solvers

Apologies, Chi-Tuan, I mean't SCHAR41 and not SCHAR_41. I am using the version v8p1r1.

Steering file and user_streamline.f attached.

I realise that this issue has been discussed elsewhere on the forum but the Z-output of my drogues file appears to be a sigma value and not the true height (see below). Has this issue been resolved?

Thanks,

Tom

TITLE = "DROGUES FILE"
VARIABLES = "LABELS","X ","Y ","ELEVATION Z M ","COLOUR"
ZONE DATAPACKING=POINT, T="G_ 60.0000 SECONDS", I= 260, SOLUTIONTIME= 60.0000
1, 362271.40192894,6174884.89298620, 0.83022239, 1
1, 362261.37296470,6174874.71334613, 0.80605801, 1
1, 362291.18030387,6174854.85181394, 0.79225369, 1
1, 362301.04926483,6174844.86514178, 0.77455626, 1
1, 362261.37536581,6174884.77849432, 0.80846297, 1
1, 362261.11444645,6174884.72497364, 0.75177167, 1
1, 362241.45004449,6174884.58704236, 0.80183265, 1
1, 362261.10329696,6174844.46350206, 0.73884877, 1
1, 362251.31596803,6174854.46907939, 0.77758765, 1
1, 362211.57365162,6174803.83599522, 0.78990378, 1
1, 362221.49157841,6174808.90584898, 0.77207934, 1
1, 362208.46034478,6174813.80393475, 0.74527809, 1
1, 362223.33265708,6174798.81993193, 0.72420345, 1
1, 362228.46395359,6174802.93391994, 0.77322680, 1
1, 362217.28263934,6174806.79456528, 0.70146637, 1
1, 362236.43790325,6174815.08927705, 0.78054909, 1
1, 362222.25859351,6174820.93548791, 0.71038979, 1

Indeed if you use a sigma transformation, you will not call SCHAR41.

Anyway, if you talk of the bug in www.opentelemac.org/index.php/kunena/21-...n-drogue-output-file
the answer is not yet.

In addition to the steering file you have sent + user_streamlines subroutine, do you use additional files that do not come from TELEMAC-3D tide example, just to reproduce the bug for your case?

Chi-Tuan

PS : looking quickly at your steering file, I would not recommend to use the Smagorinski turbulence model over the vertical, I am not sure it fully works for TELEMAC-3D, but if you have good (or bad) feedback, do not hesitate to tell. + Using characteristics for the advection of tracers is not recommended as this advection scheme is not conservative.

I have have 6 sigma levels in my 3D case. The only additional file I am using is user_flot3d.f for the drogue release positions.

There is not really a bug as the case runs but I just cannot introduce a stochastic diffusion model in 3D. I do have access to a python script for 2D stochastic diffusion (x,y, not vertical) that works well on the drogues file but I just wanted to understand how to activate stochastic diffusion (x,y) in the actual Telemac3D solver.

For turbulence in the vertical I tried k-epsilon but had divergence issues so switched to the more stable Smagorinsky. In terms of sea levels and current speeds these compare well with measured data for my case. The tracer fields appear logical too, even with turbulent diffusvity coming from Smagorinsky. Would you recommend any other turbulence option for the vertical?

I will try alternative advection schemes for the tracers.

Thanks,

Tom

Hello Tom,

My use of Smagorinski turbulence model is not big but for one TELEMAC-3D with shallows, it was not very good over the vertical (too mixed).
Anyway, if it is OK for you, it is good news (perhaps it works better in the oceans or seas?).

I generally use k-epsilon or mixing length model as turbulence model over the vertical.

Chi-Tuan

Thanks for this, Chi-Tuan. I will experiment with the various options you have presented.

Tom