[MITgcm-support] DST3 advection scheme

[Chris Hill]

J-M,

Who but an MITgcm user would be asking such questions on a Sunday afternoon!
However, what happens if Rj==0?

Chris 

> -----Original Message-----
> From: mitgcm-support-bounces at mitgcm.org 
> [mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of samar 
> khatiwala
> Sent: Sunday, January 11, 2004 5:10 PM
> To: Jean-Michel Campin
> Cc: mitgcm-support at mitgcm.org
> Subject: Re: [MITgcm-support] DST3 advection scheme
> 
> Hi Jean-Michel
> 
> Thanks for responding so quickly. (Who else would on a Sunday 
> afternoon!) This is exactly the answer I was looking for.
> Just to confirm, are you suggesting I change:
> 
>         thetaP=Rjm/(1.D-20+Rj)
> to
>         thetaP=Rjm/Rj
> 
> and
>         thetaM=Rjp/(1.D-20+Rj)
> to
>         thetaM=Rjp/Rj
> 
> Yes, that would reproduce the equations in the documentation 
> and make the scheme linear in the tracer concentration. Just 
> what I wanted.
> 
> Thanks, Samar
> 
> On Sun, 11 Jan 2004, Jean-Michel Campin wrote:
> 
> > Hi Samar,
> >
> > > However, looking at the source in gad_dst3_adv_x.F I am 
> no longer certain.
> >
> > Apart from the 1.d-20 (which I don't like too much because for some 
> > biological tracer, e.g. iron or others, the usual 
> concentration can be 
> > very small also, and the chance to get Rj equal -1.D-20 is not
> > zero) that I guess you can remove (because thetaM & P are 
> allways *Rj 
> > after), the scheme looks like it should, linear regarding 
> the tracer, 
> > but non linear regarding the flow field. Is this answering your 
> > question ?
> >
> > See you,
> >
> > Jean-Michel
> >
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
>