[MITgcm-support] Advection_schemes

Thu Aug 19 09:15:41 EDT 2004

Petter,

You mentioned Tr1 - that isn't the ptracer package that we use for passive
tracers but a deprecated single passive tracer (which I thought we had
deleted). I would not be surprised if any Tr1 code that exists is
inconsistent with the full suite of advection schemes. Try using ptracer and
your problems should go away.

Martin's tests are worth doing to convince yourself it's all working...

A.

-----Original Message-----
From: mitgcm-support-bounces at mitgcm.org
[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Martin Losch
Sent: Thursday, August 19, 2004 2:20 AM
To: mitgcm-support at mitgcm.org
Subject: Re: Re: Re: [MITgcm-support] Advection_schemes

Petter,
I am a little puzzled by your results, I have been using DST-3 with 
flux limiter (scheme 33) for quite some time now also with older 
checkpoints and (17) passive tracers (with the ptracers-pkg), and I 
haven't had any of your problems, yet. Let's not hope that you have 
discovered something that will make my last year's work worthless 
(don't worry, it's worthless anyway (o:)

Another test: use a linear equation of state (eosType='LINEAR') and set 
sBeta=0., so that theta is basically density/buoyancy and salinity is 
passive. Then salinity and your explicit passive tracer should give the 
same results (provided that ptracers_forcing uses the salinity forcing, 
this is the default). In all of my experiments this was the case.

Martin

On Aug 19, 2004, at 12:20 AM, Petter Stenström wrote:

> Hi Patrick,
>
> Thanks for your reply. Noop, I did not forget to specify
> tracerAdvScheme=33. The problem seems rather to be that multiDimAdv 
> has to be false when using scheme no 33. I looked again in the 
> documentation and
> found a note on that.
>
> The strange thing is that this seems to be required only for tr1, not 
> for T and S. I ran a simple experiment in a simple domain with exactly 
> the same
> initial distribution and concentration of T and tr1, and the same 
> advection
> scheme (33). tr1 goes to infinity, whereas T stays on realistic values.
> Strange since tr1 seems to use the same diffusivity constants as T in 
> the
> default configuration. I will look closer in the code and try to 
> figure out
> what the difference between the two is.There are slight differences 
> also
> when using other schemes (I've tried 3rd order upwind biased).
>
> /Petter
>
>
>
> ------- Original message -------
> From: heimbach at MIT.EDU
> Date: Wed, 18 Aug 2004 15:27:19 -0400
>
> Hi Petter,
>
> sorry I oversaw your reply
> (I am CC-ing to mitgcm-support and suggest
> that you always send to that list).
>
> Re. your question I think you forgot to also specify  tracerAdvScheme
> = 33 in your data file.
> Can you try and see whether that gives the
> desired results?
>
> -Patrick
>
>
>
> Quoting Petter StenstrXm <petters at kth.se>:
>
>> Hi Patrick,
>>
>> thanks. 33 is apparently a flux limited 3rd order DST. It works fine 
>> with
> S
>> and T, but not with a passive tracer. I cannot understand why, but 
>> false extrema appear in the TR1 field, but not in the S and T fields, 
>> when I use
>> 33 for all of them. Is there an obvious explanation for this that I 
>> don't
>> see?
>>
>> Petter
>>
>> ------- Original message -------
>> From: heimbach at MIT.EDU
>> Date: Tue, 17 Aug 2004 14:35:19 -0400
>>
>> Hi Petter,
>>
>> they are described in sections 2.16, 2.17
>> http://mitgcm.org/sealion/online_documents/node71.html
>> http://mitgcm.org/sealion/online_documents/node76.html
>> of the online documentation,
>> but for the actual number associated with each sheme
>> you'll have to look in the header file pkg/generic_advdiff/GAD.h
>>
>> -Patrick
>>
>>
>>
>> Quoting Petter StenstrXm <petters at kth.se>:
>>
>>> Hi!
>>>
>>> Is there a list of all the different advection schemes available and
> their
>>> corresponding numbers? I cannot find any in the documentation.
>>>
>>> I have been using advection scheme 33 for some time (which I thought
> was a
>>> flux limiter scheme), but when I look now, I cannot see that the 
>>> value
> 33
>>> is interpreted anywhere in the code.
>>>
>>> Regards,
>>>
>>> Petter Stenström
>>> Dept of Land and Water Res Eng
>>> Royal Institute of Technology
>>> SE-100 44 Stockholm
>>> Phone: +46 8 7908657
>>> Email: petters at kth.se
>>>
>>> _______________________________________________
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>>> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
>>>
>>
>>
>> --------------------------------------------------------
>> Patrick Heimbach   Massachusetts Institute of Technology
>> FON: +1/617/253-5259                  EAPS, Room 54-1518
>> FAX: +1/617/253-4464             77 Massachusetts Avenue
>> mailto:heimbach at mit.edu               Cambridge MA 02139
>> http://www.mit.edu/~heimbach/                        USA
>>
>>
>>
>>
>
>
> --------------------------------------------------------
> Patrick Heimbach   Massachusetts Institute of Technology
> FON: +1/617/253-5259                  EAPS, Room 54-1518
> FAX: +1/617/253-4464             77 Massachusetts Avenue
> mailto:heimbach at mit.edu               Cambridge MA 02139
> http://www.mit.edu/~heimbach/                        USA
>
>
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support

_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org
http://dev.mitgcm.org/mailman/listinfo/mitgcm-support