Incorporate VAD wind DA into JEDI for RRFS

[delippi]

Description

The UFO in JEDI is the component that not only computes model-simulated observations, but also houses filters and methods for observation QC, ob errors, and bias correction. The GSI observer is the equivalent. Many forward operators for various observations have been developed for the UFO. These operators can be utilized in RDAS. However these operators still must be tested for RDAS. The steps for transition by observation platform are as follows:

1. Establish IODA processing
2. Establish operator in UFO
3. Establish YAML input file
4. Validate with test assimilation experiments
5. Create ctest, if necessary
6. Sonde data is bufr obtype=224

Requirements

To create yaml configurations for assimilating VAD wind data.

Acceptance Criteria (Definition of Done)

• Phase 1 validation (hofx validation using GSI-IODA and GSI-geovals as necessary; GSI vs FV3-JEDI)

• Phase 2 validation (QC validation; no reliance on GSI except to be used as the baseline; GSI vs FV3-JEDI)

• Phase 3 validation (FV3-JEDI and GSI vs MPAS-JEDI)

• Link any relevant pull requests here:

Dependencies

None

[delippi]

Phase 1 results
These results use the following settings

1. GSI-IODA
2. GSI-geovals
3. GSI-QC filters (by using GsiUseFlag)
4. JEDI OEFC (off) (use GsiFinalObsError)

1. GSI-IODA
2. JEDI-geovals
3. JEDI-QC filters
4. JEDI OEFC (off)
5. GSI OEFC (off)

1. GSI-IODA
2. JEDI-geovals
3. JEDI-QC filters
4. JEDI OEFC (on): Obsgrouping: stationIdentification
5. GSI OEFC (on)

There's something not quite working correct in ObsErrorFactorConventional (OEFC). Some obs are only inflated by sqrt(2) when most are inflated by about 1.8.

[SamuelDegelia-NOAA]

@delippi Do you know if the profiles are split up similarly between JEDI and GSI? The error inflation from ObsErrorFactorConventional should be based on the profile density (in terms of pressure). So comparing that between JEDI and GSI could be helpful.

[delippi]

@SamuelDegelia-NOAA, I need to look into it more and I'm going to compare notes with people from the global team on this. I'm thinking the obsgrouping is (part of) the problem but I'm not sure how yet. It could also be differences in calculations in the actual ObsErrorFactorConventional code, but that is hard to debug if I'm uncertain about the grouping and I also can't use a single ob test because it is based on ob density... I might be able to simplify some with a single profile test. I also don't know if MPI plays any role in the grouping--it shouldn't but I haven't confirmed this. This is going to be fairly tricky to debug.