ROPP returning azimuths between -180 deg and 180 deg

[Ian Culverwell]

ROPP seems to be returning azimuths between -180 deg and 180 deg, eg running

/data/nwp1/idculv/ROPP/ropp_src/branches/dev/Share/ROPP81_prototype/ropp_pp/tools/ropp_pp_occ_tool IT-PP-03_prof19.nc -o IT-PP-03_prof19_cntl.nc -mfile MSIS_coeff.nc -c gras_pp.cf -ellipsoid -d -m MSIS -no-ranchk

on the attached gives

ncks -H -vazimuth IT-PP-03_prof19_cntl.nc                                        
dim_unlim[0] azimuth[0]=-7.26265 

If run without the -no-ranchk option, we get

idculv@eld037:> ncks -H -vazimuth IT-PP-03_prof19_cntl.nc                                              
dim_unlim[0] azimuth[0]=-9.9999e+07

This is because the valid range of azimuths in ROPP is [0, 360]. This is easy enough to fix, but the consequences need careful checking.

[Ian Culverwell]

IT-PP-03_prof19.nc

[Ian Culverwell]

Note that azimuth_tp is OK, even when we run without -no-ranchk:

ncks -H -vazimuth_tp -ddim_lev1b,0,4 IT-PP-03_prof19_cntl.nc
dim_unlim[0] dim_lev1b[0] azimuth_tp[0]=352.417 
dim_unlim[0] dim_lev1b[1] azimuth_tp[1]=352.42 
dim_unlim[0] dim_lev1b[2] azimuth_tp[2]=352.423 
dim_unlim[0] dim_lev1b[3] azimuth_tp[3]=352.426 
dim_unlim[0] dim_lev1b[4] azimuth_tp[4]=352.429 

This is because azimuth_tp is calculated by ropp_utils/coordinates/tangent_point.f90, which includes the relevant line:

if (azimuth_tp(i) < 0.0_wp ) azimuth_tp(i) = azimuth_tp(i) + 360.0_wp 

ropp_utils/coordinates/occ_point.f90, which calculates the georef%azimuth (among other things), does not have this line.

The only problem with just inserting this check on negative azimuths in occ_point.f90 is that it's called by ropp_pp_preprocess.f90 and ropp_pp_spectrum_tool.f90, although I don't think either uses the azimuth that it calculates.

[Ian Culverwell]

Checked that when I rerun with the attached occ_point.f90 I get (with or without the -no-ranchk option)

ncks -H -vazimuth IT-PP-03_prof19_test.nc
dim_unlim[0] azimuth[0]=352.737 

which looks fine.

There are no other differences in the output files.

I think this is probably grounds enough for making the change.

[Ian Culverwell]

occ_point.f90

[Ian Culverwell]

Stig and Chris B agree this is a sensible fix. CB points out that UCAR azimuths lie between -180 deg and 180 deg.

See #325 and #233 for earlier work on this.

[Ian Culverwell]

Introduced at r4753.

Closing ticket.