Recenter ens. members around the deterministic background.

[guillaumevernieres]

The code/application exists (under utils) but is not implemented in the workflow yet. We need to pass recentering increments to the ensemble members.

[guillaumevernieres]

Notes:
1 - We only need the initial conditions (MOM.res.nc and cice.res.nc), not need to stage f006 & f009
2 - Create soca_recenter.yaml based off of soca_ensb.yaml
3 - We need to turn off most of the options of the re-centering ... (@guillaumevernieres 's job)
4 - move the ensemble of increment to the comroot/model_data

[AndrewEichmann-NOAA]

in feature/ocnanalecen

[AndrewEichmann-NOAA]

@guillaumevernieres Which of I presume the following should be copied back?

ocn.ssh_recentering_error.incr.2018-04-15T09:00:00Z.nc
ice.ssh_recentering_error.incr.2018-04-15T09:00:00Z.nc
ocn.steric_explained_variance.incr.2018-04-15T09:00:00Z.nc
ocn.ssh_unbal_stddev.incr.2018-04-15T09:00:00Z.nc
ocn.ssh_total_stddev.incr.2018-04-15T09:00:00Z.nc
ocn.ssh_steric_stddev.incr.2018-04-15T09:00:00Z.nc
ocn.pert.steric.4.nc
ocn.pert.steric.3.nc
ocn.pert.steric.2.nc
ocn.pert.steric.1.nc
ocn.bkgerr_stddev.incr.2018-04-15T09:00:00Z.nc
ice.trash.incr.2018-04-15T09:00:00Z.nc
ice.bkgerr_stddev.incr.2018-04-15T09:00:00Z.nc

[guillaumevernieres]

I forgot to change the name of the increments ... and we're not outputting sea-ice.
We should copy the ocn.pert.steric*, but they also should be name differently.

[AndrewEichmann-NOAA]

@CoryMartin-NOAA Do you know what the increment files should be called when copied back to COM after recentering? Started with (for example) enkfgdas.20180415/12/mem001/model_data/ocean/restart/gdas.ocean.t12z.incr.nc

[CoryMartin-NOAA]

that's a good start, they shouldn't be in the restart/ directory though, but rather the analysis/ directory. I'm not sure what the 'final' naming would be, but that seems good enough for now to me.