[MITgcm-devel] update of AIM test exp.

[Jean-Michel Campin]

Hi Chris, 

I would like to update the 3 aim.5l* verifications exp:
a) aim LatLon & EqChannel are still using the old aim pkg, and 
 since we know that aim_v23 works better than the old aim 
 implementation, we can decide to move to aim_v23 for those 2 exp.
 And for the LatLon, keep the same forcing fields (from NCEP: monthly 
 mean ground T, soil moisture and albedo, no fractional land-sea mask).
 Unless there is a good reason to keep the aim LaLon exp as it is ?
b) there are 2 ways of imposing surface BC with aim_v23 
 but right now, only 1 type of surface BC is tested (since only
 aim CS is using aim_v23). aim_LatLon could be used to test aim_v23 
 with the "NCEP" surface BC and then aim_CS can move to Franco M. 
 forcing fields in order to use (and test) the fractional land/sea mask.
 This is presently not tested, not even in the coupled set-up where
 the land-sea fraction is always 0 or 1.
 I can take, for instance, the same set of parameters that I was using 
 to produce the results for the isomorphism paper.
 This would also bring this exp closer to what I am generally running.

I would prefer to update those exp before making modifications in 
the aim code, so that we can check and follow precisely what affects 
the results.

Once this is done, we can start to think of removing the old aim pkg.

Do you agree with this plan ?

See you,

Jean-Michel