[MITgcm-devel] Re: [MITgcm-support] obcs_apply_ptracer

[Dimitris Menemenlis]

Jean-Michel, your little tst_2+2 script is a gem ... but I still struggling with 
seaice_obcs.  So far I have determined that seaice_obsc 2+2 remains broken for 
following conditions:

1. tempAdvScheme=2,
    saltAdvScheme=2,

2. StaggerTimeStep=.TRUE.,

3. StaggerTimeStep=.TRUE.,
    useSALT_PLUME  = .FALSE.,

4. StaggerTimeStep=.TRUE.,
    useSALT_PLUME  = .FALSE.,
    useOBCS        = .FALSE.,

It is fixed with following:

1. useSEAICE = .FALSE.

2. usedynamics=.FALSE.,

3. adding land at SouthWest edge

Difference starts appearing in C-LSR solver, the U-iteration, first time that it 
is called by the 1iB test:

> [dimitri at skylla run]$ grep lsr std_outp.1iB
>  U lsr iters, error =     14  5.33052768592057E-13 differ <<<<
>  V lsr iters, error =     28  3.83693077310454E-13 same

> [dimitri at skylla run]$ grep lsr std_outp.2it
>  ...
>  U lsr iters, error =     14  5.33059707485961E-13 differ <<<<
>  V lsr iters, error =     28  3.83693077310454E-13 same

Will delve more into this later today.

D.