The idea is that Metop data generated by UCAR (in files with names like atmPhs_MTPA..._nc}}) should not suffer any mission-specific pre-processing or open loop processing. The former has been in ROPP since ROPP-9.1, but the latter is new. It has been implemented at r6830, which is very close to DMI's r5097.

Initially this caused a failure in the occ core test:

Running t_pp_occ_1 (PP occ; default options) ...

----------------------------------------------------------------------
                     ROPP PP File Comparison Tool
----------------------------------------------------------------------

INFO (from ropp_pp_compare):  Comparing ropp_pp_test_2m.nc and ../data/ropp_pp_test_2m_reference.nc:
   the results of running test t_pp_occ_1 (PP occ; default options)
 
ERROR (from ropp_io_fields_compare):  Profile 1 Lev1b%Npoints values ARE significantly different: |diff| =                  0001 >  0000
INFO (from ropp_pp_compare):  Differences occurred in profile 1
INFO (from ropp_pp_compare):  1 elements of ropp_pp_test_2m.nc and ../data/ropp_pp_test_2m_reference.nc differ significantly
****************************
********** *FAIL* **********
****************************

This was because the input file to the test (../data/ropp_pp_test.nc) has

         :processing_centre = "UNKNOWN" ;

and this gets copied across to ro_data%processing_centre. This results in:

WARNING (from ropp_pp_preprocess):  Processing centre UNKNOWN not recognised.
   No open loop pre-processing conducted.

rather than the desired

INFO (from ropp_pp_preprocess):  COSMIC data: openloop preprocessing
... (from ropp_pp_openloop):  Navbit correllation area:          144.4           -0.6
... (from ropp_pp_openloop):  Max. navbit correlation:    0.000 at  -10

and this causes the failure. (The test file has

	dim_lev1b = 1198 ;
	dim_lev2a = 969 ;

while the reference file has

	dim_lev1b = 1197 ;
	dim_lev2a = 968 ;

.)

The problem is easily remedied by setting

         :processing_centre = "UCAR" ;

in the input file. Then the occ (and other ) core tests work fine.

We might need to do an analogous thing in the test folder.

ROPP PP user guide updated to reflect this change at r6829.

Checks out OK, so closing ticket.

For the record, the only differences between r6829 of the ROPP-11 code and r5097 of dmi_trunk_8.1 are:

• In ROPP-11 the CHAMP pre-processing

  IF ( ANY(ro_data%Lev1a%snr_L2p == 0.0_wp) ) THEN
     ro_data%Lev1a%snr_L2p(:) = ro_data%Lev1a%snr_L1ca(:)
  ENDIF
  

is done in subroutine ropp_pp_preprocess_CHAMP;

• ROPP-11 includes some GNOS-preprocessing in a call to ropp_pp_preprocess_GNOS;
• ROPP-11 allows for ro_data%processing_centre(1:3) == 'DMI' as well as 'EUM', even though the former is unlikely to hold useful data.
• ROPP-11 includes some processing for ro_data%processing_centre(1:4)='UNKN', so that ROPP can still process legacy datasets; and
• ROPP-11 only corrects for navigation bits in the open loop processing of Metop data if there are any external navigation bits (i.e. if ANY(BTEST(LCF(:),1)) is .TRUE.).