[MITgcm-devel] seaice_lsr

[Martin Losch]

Hi Patrick,

I would like to rearrange seaice_dynsolver and seaice_lsr a little  
bit. In particular I'd like to move the copy of uice(i,j,1,bi,bj) to  
uice(i,j,2,bi,bj) and uiceC into seaice_lsr. All that's left in  
seaice_dynsolver would be two calles to seaice_lsr like this:
DO K=1,2
   CALL SEAICE_LSR(k,mytime,myiter,mythid)
ENDDO
Whye to I want to do this? Because then it would be easy and neater to  
implement pseudo time stepping (increase the number of calles to LSR).

Now I am afraid that this will interfere with the adjoint because I am  
not sure what to do with store directives before the first call of  
seaice_lsr. Naively I would put them into seaice_lsr right after the  
"predictor step".
You can have a look at what I want to do here: faulks.csail.mit.edu:/u/ 
u0/mlosch/scratch/MITgcm/pkg/seaice

Any comments?

Martin