Modifications to eum2ropp

[Stig Syndergaard]

eum2ropp.f90 needs revision in many places, in particular in the level1a part. Comments are in the attached file. It also needs update to be able to read updated group structure in EUMETSATs NetCDF4 files. It currently does not work on recent nc files.

[Stig Syndergaard]

eum2ropp now works on recent NetCDf4 files from EUMETSAT (back compatible) and with a switch to choose either cl, cl+ol, cl+rs, or none.

The code can be used for GPAC 0.4.0. See ​http://websat.dmi.dk/trac/ticket/730

Still there are further revisions to do before it is ready for GPAC 2.2.0 and 2.3.0 etc, so the ticket stays open.

[Stig Syndergaard]

When merging the dmi_trunk and ROPP 7.1, we took the Met Office solution, which is not back compatible. Thus, the current version in dmi_trunk_7.1 is not back compatible (the version in dmi_trunk_6.1 is). This is okay. We don't expect to use the back compatibility.

[Stig Syndergaard]

List of code revisions:

1. Issue when cl and ol/rs records overlap. How is LCF used later on?

2. Why is the phase accumulated before navbits are removed. Is this a bug?

3. Why use ATAN2 instead of 'ATAN3' as recommended by MDID document.

4. Do we want a switch to read EUMETSATs ol/rs excess phase or do we want to construct it from I's and Q's (as is hardcoded now)?

5. Revise the use of internal navigation bits when external ones are missing. What happens in the occ_tool pre-processing?

6. Is removing data with tracking state = 2 okay? Check MDID.