RELEASE NOTESgfs.v13.0.2

v13.0.2 - released August 24, 2016

Dropsonde bug fix:
This fix address a known bug in global_gsi. When in the cycle has drops, this new fix will pickup the drop data. 

Code change:
/nwprod2/global_shared.v13.0.2/sorc/gsi.fd/read_prepbufr.f90

v13.0.3 - released 09/09/2016
Switch off Metop-A HIRS 

fix file change:
/nwprod2/global_shared.v13.0.3/fix/global_satinfo.txt

v13.0.4 - Released 10/27/2016
Release notes for AMSU-A bug fix

Background

Channel 15 on MetOp-B AMSU-A failed on 2016101703.  This resulted in the rejection of all data from this instrument by the GSI including the ability to monitor the data.  This CDF is a request for a minor GSI code change (small changes to four subroutines) to allow MetOp-B AMSU-A channels to be monitored and to restore to active assimilation channels 7-13 where quality control and emissivity estimated using channel 15 is not required.  Note that this is not the addition of extra data but the reinstatement of data that was being used prior to 2016101703.

The changes to the code were originally developed on top of the GSI trunk and are documented in the GSI ticket 612:
https://svnemc.ncep.noaa.gov/trac/gsi/ticket/612

Location of changes

The operational GFS (/nwprod2/global_shared.v13.0.3) was copied into the GFS branch https://svnemc.ncep.noaa.gov/trac/gfs/browser/branches/global_shared.v13.0.4.  The required changes were then applied to this branch and committed.

Only five files are changed:
sorc/gsi.fd/read_atms.f90
sorc/gsi.fd/read_bufrtovs.f90
sorc/gsi.fd/qcmod.f90
sorc/gsi.fd/setuprad.f90
fix/global_satinfo.txt

These same changes are also on WCOSS in:
/sss/emc/da/save/Andrew.Collard/RFC/AMSUA/global_shared.v13.0.4

Implementation

Copy the global_shared.v13.0.4 branch into /nwprod2/global_shared.v13.0.4 and follow standard GFS implementation procedures.

Testing
The code has been compiled and run and the expected behaviour was seen: channels 1-6 and 14-15 are monitored while 7-13 are assimilated.

It is suggested that NCO run two full GFS and GDAS cycles to assume the changes have been implemented correctly before assimilation.  These changes do not require a full parallel production test.  Testing the GFS and GDAS is sufficient.