Marine DA candidate for GFSv17 prototype

[JohnSteffen-NOAA]

An updated marine DA cycling experiment is developed as a potential GFSv17 prototype using a global-workflow build on Hera from August 29, 2024:

Experiment directory:
/scratch1/NCEPDEV/da/John.Steffen/expts/marine_candidate_092024

Marine DA updates for this experiment include the following:

1. Refactoring of the diagb and implementation of the diffusion operators, along with inflating the in situ obs errors and satellite SST & ADT obs errors

• The in situ obs have an error coefficient multiplier of 1000x. The original coefs were set to 0.02/0.01 for T/S, respectively.
• The LEO satellites instruments have an error coefficient of 1.0, which was an increase from the original value of 0.05.
• ADT error coefs increased from 0.1/1.0 to 0.2/2.0 for GeoVals/mesoscale_representation_error and ObsError/absoluteDynamicTopography

2. A temporary (while we scratch our heads …) box filter for ADT observations to blacklist obs within the western Arabian Sea, Gulf of Aden and Red Sea. This alleviates instabilities in this region leading to gravity wave generation and propagation.

3. Monitoring SSTs from the GEO satellites to compute diagnostics but does not assimilate the observations

Files modified in GDASapp:
parm/soca/obs/config/adt_rads_all.yaml
parm/soca/obs/config/sst_abi_g16_l3c.yaml
parm/soca/obs/config/sst_abi_g17_l3c.yaml
parm/soca/obs/config/sst_ahi_h08_l3c.yaml