OSEs for SST retrievals (IR)

[ShastriPaturi]

The ombs for SST with VIIRS-NPP are larger with 6hr DA window are found to be larger than the 24hr DA window (HAT10). OSEs for SST retrievals (IR) with 6hr DA window for a 1-2 wk period:

• [ ] VIIRS_NPP
• [ ] VIIRS_N20
• [ ] AVHRR_MA
• [ ] AVHRR_MB
• [ ] AVHRR_MC
• [ ] GOES_16
• [ ] GOES_17

Deliverables

• [ ] ombs against SST drifters (see issue #468)

[liuxiao37k]

A set of 4 SST OSEs were conducted for JCSDA's golden period in July 2021:

1. CFSR, Hybrid EnVAR, assimilating VIIRS NPP SST retrieval only + adt, 24h DA window. (gaea path: /lustre/f2/dev/ncep/Xiao.Liu/sandbox/05jan2023/test_hybrid_v2)
2. CFSR, Hybrid EnVAR, assimilating all SST retrievals + fnmoc profiles + adt, 24h DA window. (gaea path: /lustre/f2/dev/ncep/Xiao.Liu/sandbox/05jan2023/test_hybrid_v2_allobs)
3. CFSR, Hybrid EnVAR, assimilating VIIRS NPP SST retrieval only + adt, 6h DA window. (gaea path: /lustre/f2/dev/ncep/Xiao.Liu/sandbox/05jan2023/test_hybrid_v2_6h)
4. CFSR, Hybrid EnVAR, assimilating all SST retrievals + fnmoc profiles + adt, 6h DA window. (gaea path: /lustre/f2/dev/ncep/Xiao.Liu/sandbox/05jan2023/test_hybrid_v2_allobs_6h)

Evaluation:

1. comparing daily mean bkg against OSTIA.
2. comparing H(x) against drifter at 00Z for all OSEs.

Conclusion: Assimilating all SST retrievals and fnmoc profiles yields better results than assimilating a single SST retrieval product (VIIRS NPP).

Supporting figures:

[liuxiao37k]

@ShastriPaturi @guillaumevernieres Please see a summary report above and let me know if there are remaining tasks to do before the issue can be closed.

[ShastriPaturi]

Thanks @liuxiao37k