[MITgcm-support] "gad_diff_r.F" and mixing

[Chris Hill]

Hi matt,

 in the situation you describe gad_diff_r is not used or needed. With gm
active the vertical k can become large because it includes a project of the
along iso surface horizontal diffusivity and bolus transport. Therefore in
the mode where gm is active implicit vertical diffususion should/is always
be used.

 Does that make sense?

chris 

> -----Original Message-----
> From: mitgcm-support-bounces at mitgcm.org 
> [mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Matt Palmer
> Sent: Tuesday, August 03, 2004 12:17 PM
> To: mitgcm-support at mitgcm.org
> Subject: [MITgcm-support] "gad_diff_r.F" and mixing
> 
> Hi there.
> 
> I've been doing model runs using the GMRedi package and 
> combining a 3-D vertical mixing field. When I compare the 
> model overturning (the Indian
> Ocean) using GMRedi and vertical mixing, the result is much smaller (2
> Sv) than for the same vertical mixing field and 
> horizontal/vertical mixing (16 Sv).
> 
> While looking into this discrepency I found that 
> "gad_diff_r.F" was not being called from "gad_calc_rhs.F". It 
> seems that if "gad_diff_r.F" is not called, it's because the 
> model is using implicit diffusion. Can someone tell me 
> whether "gad_diff_r.F" should be routinely called when using 
> vertical mixing?
> 
> Many thanks,
> 
> Matt Palmer
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://dev.mitgcm.org/mailman/listinfo/mitgcm-support
>