[MITgcm-support] stagger time step

[Martin Losch]

Hi Marco,

all of the so-called “multiDim” advection schemes (and 33 is one of them) should be used with staggerTimeStep=.TRUE., 

I have had the experience, that if you don’t do it, you can end up with a totally overturned stratification during the initial adjustment process. The reason (internal waves affecting the numerical stability) is explained in the documentation and in other posts on this list.

Btw, type “staggertimestep” in the search window of http://mitgcm.org/public/search.html already gives a bunch of entries in the mailing list and in the manual (I am surprised it actually works (o:, although you get many more with a google search like this: “site:mitgcm.org staggertimestep”, without the quotes!)

Martin

> On 09 Mar 2015, at 16:15, marco reale <reale.marco82 at gmail.com> wrote:
> 
> dear mitgcm users,
> 
> I’m doing some experiments using MITGCM with advective scheme 33 for temperature and salinity and KPP implemented. 
> 
> Despite the model doesn’t blow up after the end of simulation checking the STDERR file I found this error reported:
> 
> 
> (PID.TID 0000.0001) ** WARNING ** GAD_CHECK: potentially unstable time-stepping (Internal Wave)
> (PID.TID 0000.0001) ** WARNING ** GAD_CHECK: need "staggerTimeStep=.TRUE." in "data", nml PARM01
> 
> So I’m repeating the same experiment using staggerTimeStep = .TRUE. again but I would like to know how this error may affect the outputs of my experiment or if I have to expect changes between the two experiment outputs 
> 
> Cheers
> 
> Marco
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support