modelling problems

[jaromirjakacki]

Good evening

I am writing this email because a lot of modelling experts read this list. I think the problem is linked to turbulence. 

Currently we are working on a coupled ice-ocean model that will cover the Baltic Sea region with lateral boundaries in the vicinity of Goteborg. Ocean model is ROMS (as I remember version 3.8), the ice model is CICE (version 6.*). We have just started to work on tuning up the model. I think I should add that our system is based on metroms, but currently there is only MCT in the semi-original version. The system works, we have a simulation that covers about 20 years. Horizontal resolution is about 2km, we have 30 sigma layers in the vertical direction. Our problem is probably linked to turbulence closure parameters. We use MY2.5. 

The problem is visible in the structure of the water column. We made the sections from our ship from the Gulf of Gdansk up to the Arkona Basin, but usually it is finished in the Bornholm Basin.

I will attach the results. There will be a marked section as a line and two pairs of profiles. One will be salinity, the second one is temperature. It is clean visible which one is from the model. It is recalculated into the same horizontal and vertical resolution.

The comparison of vertical profiles were already presented and we think the results were better (it is already published). It looks like we have a problem with extensive vertical mixing. However because there are a lot of parameters in the model we can not be sure. Thus I am asking you for any advice that will redirect us to the proper place. 

We have lateral boundaries from the Copernicus system and we compared the salt content and heat content of the whole Baltic Sea (I believe it is east of 13 degrees (I am not sure, I will be able to provide an exact line if it will be needed)) . It does not look like it is wrong and this is the main reason why we think there is a problem with closure.

Two pictures represent a comparison of salt content and heat content of the whole Baltic sea - the red one is from copernicus (NEMO model), the black one is our result. All drawings are draft so it is possible there is a mistake.

Thank you in advance for your help,

all the best

Jaromir

[Helmut Z. Baumert]

Dear colleagues,

yesterday I answered a mail by Jaromir using only his gmail address,
forgot to use the gotm-...@googlegroups.com address.
Here is the correction.
Helmut
-------- Weitergeleitete Nachricht --------
Betreff: Re: [gotm-users:2773] modelling problems
Datum: Fri, 17 Feb 2023 19:01:42 +0100
Von: Helmut Z. Baumert <bau...@iamaris.de>
Organisation: IAMARIS
An: jaromirjakacki <jaromir...@gmail.com>

Good evening dear Jaromir & dear GOTM friends!

Congrats to your nice simulations of the Baltic and
for your courage to raise questions about turbulence!

You are not the first one, as Richard Feyneman (Nobel phys 1965)
already noticed that t. is “the most important unsolved problem of
classical physics”, because a description of the phenomenon from first
principles did not exist. It was still regarded as one of the six most
important problems in mathematics, too.

Also Lew Landau (Nobel phys 1962) was skepttic and stated e.g. that
universal constants like the von-Karman constant (turbulent boundary
layers ...) can NEVER be derived from first principles, need to be
taken like the velocity of light, c, as phenomenological data which
only can be measured.

I attach some info about my own work "as is", which possibly
helps solving these century problems. Any questions about the
science may be discussed in detail with me by direct e-Mail:
bau...@iamaris.DE (please don't use bau...@iamaris.ORG!).

Finally, I would like to draw your attention to important
conferences touching t. and taking place in 2023.
The first one takes place in Bertinoro, Italy:
> https://www.fdy.tu-darmstadt.de/iti/itihome_2.en.jsp

The second one will happen in Sapporo/Japan:
> https://www.bitcongress.com/wco2023/

Best wishes,
Helmut Z. Baumert, CEO IAMARIS
---

> --
> You received this message because you are subscribed to the Google
> Groups "GOTM-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gotm-users+...@googlegroups.com
> <mailto:gotm-users+...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/gotm-users/CANBPw4VLpNMEg8D1hjzxd97BRBfkPXHYofCHoB7fuhD4%3DkjCpA%40mail.gmail.com <https://groups.google.com/d/msgid/gotm-users/CANBPw4VLpNMEg8D1hjzxd97BRBfkPXHYofCHoB7fuhD4%3DkjCpA%40mail.gmail.com?utm_medium=email&utm_source=footer>.

--
Helmut Z. Baumert
IAMARIS - Dr. Dr. Baumert
Bgm.-Jantzen-Str. 3
D-19288 Ludwigslust, Germany
www.iamaris.de
Tel +49 3874 6197 120
Mob +49 179 4034 240

[Ramsey R. Harcourt]

Hi Jaromir,

This is maybe a question better posted on a ROMS board, but what are your lateral diffusivties; viscosities, I think called TNU2, TNU4, VISC2, VISC4 for Laplacian, Hyperviscos/diffusivities?

 Are you using any compiler options to rotate them to geopotential or isopycnal directions. Looks like mixing along sigma coordinates over your bathymetry with large lateral diffusivities (I’ve seen TNU2 ~ 1-5 m^2/s with different groups lately) could generate specious vertical mixing much larger than some error in MY2.5 that it’s not particularly known for. 

I’m looking for causes of too much mixing (at much greater depths than yours over a slope) in a regional ROMS model too, but it’s the Canuto-A k-epsilon using GLS that is baked into ROMS doing parameterized  vertical mixing in our case. 

There was a bug in an earlier GOTM version of MY2.5 several years ago but this doesn’t sound like it  wouldn’t expect it in a ROMS distribution.

-Ramsey Harcourt

UW/APL Seattle

PS not following all your figures and arguments, but maybe it just helps to confirm with Dr Baumert that turbulence is not the easiest job but someone’s gotta do it. Hang in there and keep pushing.

On Thu, Feb 16, 2023 at 11:56 AM jaromirjakacki <jaromir...@gmail.com> wrote:

Good evening I am writing this email because a lot of modelling experts read this list. I think the problem is linked to turbulence.   Currently we are working on a coupled ice-ocean model that will cover the Baltic Sea region with lateral boundaries

ZjQcmQRYFpfptBannerStart

This Message Is From an Untrusted Sender

You have not previously corresponded with this sender.

See https://itconnect.uw.edu/email-tags for additional information. Please contact the UW-IT Service Center, he...@uw.edu 206.221.5000, for assistance.

 

ZjQcmQRYFpfptBannerEnd

--
You received this message because you are subscribed to the Google Groups "GOTM-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to gotm-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gotm-users/CANBPw4VLpNMEg8D1hjzxd97BRBfkPXHYofCHoB7fuhD4%3DkjCpA%40mail.gmail.com.

[Helmut Z. Baumert]

Hi Ramsey & Jaromir & GOTM:

Only two comments on the ongoing debate:

1. to see CLEARLY the effect of numerical diffusion
due to discretization of advection it is helpful to
set the physical (turbulent) diffusivity equal to ZERO.

2. I have applied this trick in my approach to turbulence
when I considered the EULER equations rather than
Navier-Stokes. The latter contains friction and Euler
not. Applied to turbulence it means that Euler represents
Re = \infty, in practice: VERY high, at least much higher
than in the Princeton/ONR superpipe or in competing
experimental systems (attached: Bailey et al 2014
and Durst et al 2002).

This trick helped me to derive ALL relevant fundamental
constants of turbulence theory from first principles on
for the case Re=\infty. E.g. my value of Karman's
constant fits nicely in the Princeton range
\kappa = 0.40 +/- 0.02.

What did I learn? Sometimes it's more userful
to simplify things from the beginning and look how
far one gets.

Ciao,
Helmut Z.(iegfeld) Baumert
---

> __

> This Message Is From an Untrusted Sender
> You have not previously corresponded with this sender.
> See https://itconnect.uw.edu/email-tags
> <https://itconnect.uw.edu/email-tags> for additional information.
> Please contact the UW-IT Service Center, he...@uw.edu

> <mailto:he...@uw.edu> 206.221.5000, for assistance.
> __

> <mailto:gotm-users+...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/gotm-users/CANBPw4VLpNMEg8D1hjzxd97BRBfkPXHYofCHoB7fuhD4%3DkjCpA%40mail.gmail.com <https://urldefense.com/v3/__https://groups.google.com/d/msgid/gotm-users/CANBPw4VLpNMEg8D1hjzxd97BRBfkPXHYofCHoB7fuhD4*3DkjCpA*40mail.gmail.com?utm_medium=email&utm_source=footer__;JSU!!K-Hz7m0Vt54!i5Od11tWe80P5GTHJCfRnok2v1dRQZ8vWiovBjPxRGuyTElNef2MO_ChB9Us16ChImQOooSgBBdGR8e_KXrNp7I$>.

>
> --
> You received this message because you are subscribed to the Google
> Groups "GOTM-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to gotm-users+...@googlegroups.com

> <mailto:gotm-users+...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/gotm-users/CANc59GD-o2hAMXR%3DMNYcFDga%2B8uk5hQoZpt7mFd4r_HzLaszjA%40mail.gmail.com <https://groups.google.com/d/msgid/gotm-users/CANc59GD-o2hAMXR%3DMNYcFDga%2B8uk5hQoZpt7mFd4r_HzLaszjA%40mail.gmail.com?utm_medium=email&utm_source=footer>.