- the top level routine is loop_1x1_bulk.m. It does the following:
	[loops over years and 6-hourly records]
	load fields (e.g. by calling ncep_load_fields.m)
	compute bulks (exf_bulk_largeyeager04.m or gmaze_bulk_coare.m)
	compute net fluxes
	compute time averages (averagesFields.m)
	write to disk (averagesFields.m)

- loop_1x1_flux.m and loop_1x1_flux_noicemask.m
  simply do a similar formatting of flux data (no bulk 
  involved) to allow easy comparisons.

-a few more scripts:
	gpcp_load_atlas.m 	loads/averages GPCP precip data set
	quickcow_load_atlas.m 	loads quickcow wind stress atlas
	plot_bulk.m		a sample script to plot results
	domaine_global_def.m, domaine.m, ecmwf_grid.m and ncep_grid.m
				handle data and computational grids

-the way scripts are setup right now:
	do not account leap years (allways do 365 days), output are daily in local dir.
	fields are interpolated to the 1x1 degree computational grid.
	atm. state/fluxes data is NCEP, ECMWF/ERA40 or CORE/LargeYeager.
	sst is Reynolds, runoff is Large and Yeager, Ice coverage is Hadley Center.
	links to data sets (from Charmaine directories and mines) are hardcoded.
	data sets are under /net/ross/raid* and /net/altix3700/raid*.