[MITgcm-devel] dic_atmos

[Jean-Michel Campin]

Hi Patrick and Stephanie,

I made few changes in dic_atmos.F (mainly for restart & initialisation)
and the IO & mpi stuff are now hidden from TAF.

I checked the restart with multi-threaded and dic_int1=3 (atmos.box) and
it was working. I tried also the AD verification (global_ocean_biogeo)
test and it was working (for me) as before.

But may be a cleaner solution would be to split dic_atmos.F
into 3 S/R (and remove istate argument):
1) dic_ini_atmos (with most of the code for the case istate=0)
2) dic_set_atmos : called before DIC_SURFFORCING , 
         to set AtmospCO2 array
         (might be merge into dic_surfforcing.F if not too long)
3) dif_step_atmos : called after dic_biotic_forcing from gchem_forcing_sep.F
         (from the same place as current DIC_ATMOS)
         to integrate "FluxCO2" over the domain and step forward 
         "total_atmos_carbon".
but I am not going to change this anytime soon.

Cheers,
Jean-Michel