Set missing O-A correctly

[Ian Culverwell]

ropp_1dvar_diagnostics.f90 incorrectly sets missing values of diag % OmA to -999.0. It should set them to ropp_MDFV (=-9.9999e7), as, for example ropp_qc_OmB.f90 does with diag % OmB.

[Ian Culverwell]

Done at r4204.

[Ian Culverwell]

Actually there are a few of these 900.0/999.0 funnies in ropp_1dvar. Repair them all at r4205.

[Ian Culverwell]

Note that there's a long-standing (since ROPP5.0) 'feature' of the test_1dvar.sh script, in that it compares ropp_1dvar_refrac -comp to ropp_1dvar_refrac, and therefore says

   Found max difference [%] in observation scenarios:      0.093733434
   Found max difference [K] in temperature scenarios:       0.12276346
   Found max difference [g/kg] in specific humidity scenarios:     -0.081683540
   Found max difference [hPa] in pressure scenarios:       0.11046826
IDL plot_1dvar_01 Error: (32) Deviations in 1DVar between measurement and used background exceeds 0.01%.

In other words, these are the expected differences arising from the use of the compressibility factors.

Since it's a very artificial problem anyway (analysis=background because it uses simulated obs), the simplest way to avoid the warning is to replace

if strmid(file_id, 0, 2) eq '01' and max([max(tempdiff,/absolute), max(shumdiff,/absolute), max(presdiff,/absolute)]) gt 0.01 then begin
   print,'IDL plot_1dvar_01 Error: (32) Deviations in 1DVar between measurement and used background exceeds 0.01%.'
endif

by

if file_id eq '01' and max([max(tempdiff,/absolute), max(shumdiff,/absolute), max(presdiff,/absolute)]) gt 0.01 then begin
   print,'IDL plot_1dvar_01 Error: (32) Deviations in 1DVar between measurement and used background exceeds 0.01%.'
endif

in plot_1dvar.pro. (So that the error message isn't produced if file_id='01comp'.) This has been done at r4206.

[Ian Culverwell]

No other problems when running through ropp_1dvar make tests, so closing ticket.