create a 13km conus domain 3dVar run using static background error covariance from Bump

[TingLei-NOAA]

On top of existing 13km conus domain pure Ens3Dvar run, we want to add in the bump based static background covariance component. Hence, we will first create a static 3Dvar.
This issue will document how it is done and provide a place for comments/discussions.

[HuiLiu-NOAA]

Ting,

Can you tell me how to start this work? Thanks.

hui

On Thu, Nov 9, 2023 at 2:28 PM TingLei-NOAA @.***> wrote:

Assigned #5 https://github.com/NOAA-EMC/RDASApp/issues/5 to @HuiLiu-NOAA https://github.com/HuiLiu-NOAA.

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/RDASApp/issues/5#event-10917181570, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOF252GODVMXKRNO566HRPLYDUVGPAVCNFSM6AAAAAA7FCNP2WVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJQHEYTOMJYGE2TOMA . You are receiving this because you were assigned.Message ID: @.***>

[TingLei-NOAA]

Hui, I am trying to to sort out how to do this. We can discuss this on the issue webpage, that is one benefit opening an issue. Regards, Ting

---

Ting Lei

Physical Scientist, Contractor with Lynker in support of

EMC/NCEP/NWS/NOAA

5830 University Research Ct., Cubicle 2765

College Park, MD 20740

@.***

301-683-3624

On Mon, Nov 13, 2023 at 11:57 AM Hui Liu @.***> wrote:

Ting,

Can you tell me how to start this work? Thanks.

hui

On Thu, Nov 9, 2023 at 2:28 PM TingLei-NOAA @.***> wrote:

Assigned #5 https://github.com/NOAA-EMC/RDASApp/issues/5 to @HuiLiu-NOAA https://github.com/HuiLiu-NOAA.

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/RDASApp/issues/5#event-10917181570, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AOF252GODVMXKRNO566HRPLYDUVGPAVCNFSM6AAAAAA7FCNP2WVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJQHEYTOMJYGE2TOMA>

. You are receiving this because you were assigned.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/RDASApp/issues/5#issuecomment-1808581227, or unsubscribe https://github.com/notifications/unsubscribe-auth/APEFS7BMC5WYUYF46ARN2GDYEJGP7AVCNFSM6AAAAAA7FCNP2WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMBYGU4DCMRSG4 . You are receiving this because you authored the thread.Message ID: @.***>

[TingLei-NOAA]

Bump official documentation from JCSDA: https://jointcenterforsatellitedataassimilation-jedi-docs.readthedocs-hosted.com/en/latest/inside/jedi-components/saber/BUMP.html. Bump doc from the Bump User Training at 2023 : https://drive.google.com/file/d/1AOeECej4-45wpTuAa4O5zXcQ07-_Wdcc/view?usp=sharing and https://drive.google.com/file/d/1A3KjQfQHu_hyQUcK4539WQu2Dawp5w_D/view?usp=sharing (thanks to Benjemin and Anna) The current test fv3jedi_test_tier1_3dvar_lam_cmaq could be used as a good example .

[andytangborn]

I think what you are looking for is the yaml that computes stddev, right? I use the following:

/scratch1/NCEPDEV/da/Andrew.Tangborn/JEDI/gdas_app_july28_2023/work_aero_c96/staticb_var_aero.yaml for aerosols fv3-jedi (so it might be a bit different from what you need). This step is after the staticb_prep_YYYYMMDD.yaml step.

I have examples of staticb_prep in the same directory. It is a bit confusing, but it creates two kinds of files: var and m4 (probably moment files), that are needed for computing the stddev files. The hardest part is that you need to have 1. An ensemble of background files at each time and 2. Three different times. I'm not running an ensemble, so I got around this by using 3-4 different experiments that have different physics or emissions, etc and call that my ensemble.

Have a look at my staticb_prep yaml files in that directory, and also notice the Data directories below them. You'll need to construct something very similar.

[HuiLiu-NOAA]

thanks.

On Tue, Nov 14, 2023 at 1:56 PM Andy Tangborn @.***> wrote:

I think what you are looking for is the yaml that computes stddev, right? I use the following:

/scratch1/NCEPDEV/da/Andrew.Tangborn/JEDI/gdas_app_july28_2023/work_aero_c96/staticb_var_aero.yaml for aerosols fv3-jedi (so it might be a bit different from what you need). This step is after the staticb_prep_YYYYMMDD.yaml step.

I have examples of staticb_prep in the same directory. It is a bit confusing, but it creates two kinds of files: var and m4 (probably moment files), that are needed for computing the stddev files. The hardest part is that you need to have 1. An ensemble of background files at each time and

2. Three different times. I'm not running an ensemble, so I got around this by using 3-4 different experiments that have different physics or emissions, etc and call that my ensemble.

Have a look at my staticb_prep yaml files in that directory, and also notice the Data directories below them. You'll need to construct something very similar.

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/RDASApp/issues/5#issuecomment-1810966688, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOF252B2DZ5GXRG6E6TR62LYEO5FXAVCNFSM6AAAAAA7FCNP2WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMJQHE3DMNRYHA . You are receiving this because you were mentioned.Message ID: @.***>

[TingLei-daprediction]

Another helpful link on the static B training with bump https://drive.google.com/file/d/1jiflsG-p5g-_g-AsBuvhIxdKQTkJPX-9/view?usp=sharing.

[TingLei-NOAA]

Following the regional aero 's procedure as @andytangborn listed/explained (Thanks again), I have finished a set of runs for our 13km conus test case . Those runs succeed to completion and without apparent problems. Further verification /validation , especially by comparison with GSI is to be done in the future. In the following , scripts to run each step are listed, all information needed to repeat those runs could be found through tracing those scripts (and yaml pointed in them). On hera in /scratch2/NCEPDEV/fv3-cam/Ting.Lei/dr-emc-regional_jedi/dr-rrfs_rundir/hyb-jedianl_tm05_00_gsidiag_obs_new_jedi, 1) d-static-prep.sh ( to run the prep step to get raw estimates of stddev and other moments). 2) d-static-sdtdev.sh ( to get the calibrated estimates of stddev ( only one cycle raw data (generated in the previous step ) is used). 3) d-3dvar.sh ( 3dvar run) 4) d-hyb3dvar.sh ( hybrid 3dvar run) .

[HuiLiu-NOAA]

Awesome! It will be great if we can do this for the 00Z cycle as well.

On Wed, Nov 15, 2023 at 10:43 AM TingLei-NOAA @.***> wrote:

Following the regional aero 's procedure as @andytangborn https://github.com/andytangborn listed/explained (Thanks again), I have finished a set of runs for our 13km conus run . Those runs succeed to completion and without apparent problems. Further verification /validation , especially by comparison with GSI is to be done in the future. In the following , scripts to run each step are listed, all information needed to repeat those runs could be found through tracing those scripts (and yaml pointed in them). On hera in /scratch2/NCEPDEV/fv3-cam/Ting.Lei/dr-emc-regional_jedi/dr-rrfs_rundir/hyb-jedianl_tm05_00_gsidiag_obs_new_jedi,

1. d-static-prep.sh ( to run the prep step to get raw estimates of stddev and other moments).
2. d-static-sdtdev.sh ( to get the calibrated estimates of stddev ( only one cycle raw data (generated in the previous step ) is used). 3)d-3dvar.sh ( 3dvar run)
3. d-hyb3dvar.sh ( hybrid 3dvar run) .

— Reply to this email directly, view it on GitHub https://github.com/NOAA-EMC/RDASApp/issues/5#issuecomment-1812776568, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOF252HFMICMNFUMYCOCKF3YETPLNAVCNFSM6AAAAAA7FCNP2WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMJSG43TMNJWHA . You are receiving this because you were mentioned.Message ID: @.***>

[TingLei-NOAA]

It is completed and, a rrfs ctest for the final hybrid analysis has been added to RDAS rrfs-test