[MITgcm-devel] ICE workshop - recap and homework!

[Hong Zhang]

Menemenlis, Dimitris (3248) wrote:
> Martin, Gael, Ian, and Patrick, we also find that the new ice code (checkpoint63j) creates more ice than our previous integrations (also with non-legacy code but circa checkpoint63c).  We are trying to understand what is principal cause of this additional sea ice, so that we don't have to readjust everything from scratch.  Any advice on where to look would be appreciated.
>
> Hong has summarized his results here:
> http://ecco2.jpl.nasa.gov/data1/meeting/2012/03_28/hzhang_SO_seaice.ppt
> http://ecco2.jpl.nasa.gov/data1/meeting/2012/03_28/SO_seaice.tar
> Compare only JSO24 and JSO30 - ignore JSO29, which attempted to use SItracer instead of HSALT for salinity.
>
> The differences between JSO24 and JSO30 are as follows:
>
> JSO24:
> SEAICE_OPTIONS.h
> #define MCPHEE_OCEAN_ICE_HEAT_FLUX
> #define GRADIENT_MIXED_LAYER_TURBULENCE_FACTOR
> #define SEAICE_DO_OPEN_WATER_GROWTH
> #define SEAICE_DO_OPEN_WATER_MELT
> #define FENTY_AREA_EXPANSION_CONTRACTION
> #define SEAICE_ADD_SUBLIMATION_TO_FWBUDGET
> #define ALLOW_SEAICE_FLOODING
> #define SEAICE_CAP_HEFF
>
> JSO30:
> EXF_OPTIONS.h
> # define EXF_LWDOWN_WITH_EMISSIVITY
> data.sea	ice
>   SEAICE_no_slip     = .TRUE.,
>   SEAICE_wetAlbTemp  = 0.0,
>   SIsal0             = 4.0,
>   SEAICE_mcPheePiston= 0.0008333,
>   SEAICE_frazilFrac  = 0.1,
>   SEAICE_mcPheeTaper=0.92,
>   
Hi all,
We tried some more experiments to find the cause for the more seaice 
problem.
By switching off EXF_LWDOWN_WITH_EMISSIVITY  (in EXF_OPTIONS.h )
we can basically recover the result of old run,
while other combination of run parameters make minor change.
Will look into this flag of EXF_LWDOWN_WITH_EMISSIVITY.
In the meanwhile, anybody has experience with EXF_LWDOWN_WITH_EMISSIVITY?
thanks
hong