Suggested improvements to EUMETSAT data readers

[Ian Culverwell]

The ROPP-10.0 beta reviewers say

EUMETSAT is currently working on rolling out a new version of the operational
GRAS processing. Part of the development is an evolution of the existing level
1b data format, making it compatible with radio occultation data formats for
future missions such as Sentinel-6/Jason-CS and EPS-SG. In particular, the
updated data format supports data from multiple GNSS systems and will be used
throughout EUMETSAT's radio occultation data portfolio. We therefore explored
how difficult the adaptation of the EUMETSAT readers in ROPP to the upcoming
data format changes will be.

We found that the required changes - at least for the converters of EUMETSAT
level 1b bending angle to ROPP bending angle data formats - are relatively
straightforward. To start with, the support for the new data format version in
ropp_io requires support for variable length strings in netCDF4. A corresponding
patch has been available to the ROM SAF for some time
(https://gitlab.eumetsat.int/ro/ropp/commit/9e94f97260d0a6096e21eef97219b9f4a6617cb9).

Once the above is in place, the actual support for the upcoming EUMETSAT EPS-SG
(v13) and generic RO (v14) level 1b data formats is straightforward, at least
for the bending angle part: we are currently testing
https://gitlab.eumetsat.int/ro/ropp/commit/04c00b790a3340c3346399e8c19400ee4748e7e2
and
https://gitlab.eumetsat.int/ro/ropp/commit/e6933b77945620d4ed01d462eada317f461772bd
which cover the eum2ropp tool. As far as we understand, additional modification
are, however, still required to also implement support for the conversion of
carrier phase and amplitude data which are used by ropp_pp . Given that the
modifications for supporting the upcoming data format changes are minor, we
recommend that they should be included in the final ROPP v10 release.

Since EUM and DMI are probably the only users of this bit of code, it seems that there would be a small risk in doing this after the official review. But it requires further input from EUM and DMI.

[Ian Culverwell]

This is a duplicate (triplicate!) of #652 and #682. All done. Closing ticket.