Closed gmao-msienkie closed 1 year ago
As far as I know 'time_launch' is only passed to the setupt and setupw routines to be written out the the nc4 files. Thus I don't believe it has any affect on the assimilation. In this case the additional code only sets time_launch for aircraft observations. The other observation types (besides radiosonde) continue to have a missing value for 'time_launch' . I believe we need 'time_launch' for use with Jedi so that the ascending/descending profiles for aircraft can be distinguished for the ObsErrorFactorConventional filter. In the GSI this filter is applied separately for each ascending or descending profile (via errormod_aircraft) in a call from read_prepbufr, however the Jedi QC deals with all the observations at once and needs the information to distinguish the profiles.
By the way, sorry about all the differences. I have code set to strip out excess whitespace when I save files from 'vim'.
Thanks Meta.
Just to make sure - tested for zero diff ... good.
results should be zero diff except for modifying the nc4 diag to include 'launch time' for aircraft profiles.