Bugfix ocean ensemble recentering#1558
Merged
Merged
Conversation
shlyaeva
requested review from
CatherineThomas-NOAA,
RussTreadon-NOAA and
guillaumevernieres
guillaumevernieres
approved these changes
Mar 18, 2025
Contributor
guillaumevernieres
left a comment
guillaumevernieres
left a comment
There was a problem hiding this comment.
Thanks for the quick fix @shlyaeva !
Collaborator
Author
|
Note: I've checked and I think ice recentering doesn't have this problem, as it's done on states instead of increments, and is directly adding the recentering increment to the ensemble member. |
shlyaeva
added
hera-GW-RT
emcbot
added
hercules-GW-RT-Running
|
Automated GW-GDASApp Testing Results: |
emcbot
added
hercules-GW-RT-Passed
|
Automated GW-GDASApp Testing Results: |
emcbot
added
hera-GW-RT-Passed
jspsl
pushed a commit
that referenced
this pull request
Mar 26, 2025
# Description Ocean ensemble recentering was double-counting ensemble perturbations, effectively inflating the ensemble by a factor of 2 (almost, the extra perturbation was added with IAU, so some of it was probably damped). The accumulation of amplified perturbations over several cycles may have been the cause of ensemble forecast failures in gfsv17 trial experiments. On develop: IAU increment for ocean recentering = (det-mean) + (ensmember-mean). In this branch: IAU increment for ocean recentering = (det-mean) With this branch, as expected, IAU increments passed to the ocean model are all the same, pulling from the mean towards deterministic background. Tested with C48 one cycle, compared IAU increments and looked through the logs. # Issues Resolves #1557 # Automated CI tests to run in Global Workflow <!-- Which Global Workflow CI tests are required to adequately test this PR? --> - [ ] atm_jjob <!-- JEDI atm single cycle DA !--> - [ ] C96C48_ufs_hybatmDA <!-- JEDI atm cycled DA !--> - [ ] C96C48_hybatmaerosnowDA <!-- JEDI aero/snow cycled DA !--> - [ ] C48mx500_3DVarAOWCDA <!-- JEDI low-res marine 3DVar cycled DA !--> - [x] C48mx500_hybAOWCDA <!-- JEDI marine hybrid envar cycled DA !--> - [ ] C96C48_hybatmDA <!-- GSI atm cycled DA !--> Co-authored-by: Anna Shlyaeva <anna.v.shlyaeva@noaa.gov>
danholdaway
added a commit
that referenced
this pull request
Mar 26, 2025
* develop: added original date (#1570) Update OCEAN_BASIN_FILE_ORION path for marine ctest (#1569) Fixing window_begin and end in yaml for redating ADT (#1565) Save LETKF mean analysis and mean increment for verification (#1566) Update JEDI hashes (20250324) (#1567) Removes obsolete json files from marine bufr2ioda (#1564) Add hybrid 4denvar as a phase3 validation configuration (#1552) Bugfix/ocn ecen (#1550) Bugfix SSS obs for LETKF (#1560) Bugfix ocean ensemble recentering (#1558) apply ci path patch only when running g-w ctests (#1554) Update JEDI hashes (20250317) (#1531) add option to create_experiment.py in create_exp.sh (#1546) Update jcb hashes and support initial IODA stats capability (#1547) Fix GW-CI CTest dependencies (#1534) Converts prepoceanobs to use JCB-style obs list and config files (#1527) JEDI-based ensemble recentering and analysis calculation (#1488) tools for phase3 validation (#1526) Update JEDI hashes (20250306) (#1520) New jcb-gdas # and insitu obs prep bug fix (#1523) Marine in situ obs update (SSS, moorings, drifters) (#1511) Save marine background and analysis spread (#1517) Minor build and install updates and ctest cleanup (#1512) correct logic for upper loop limit in chem_diagb (#1516) Incorporate multi-cycle concatination and separation of variables in post-BUFR observations (#1504) Update JEDI hashes (20250225) (#1475) Add a depth dependent e-fold scale (#1494) Update jcb and da-utils hashes to support land DA. (#1509) Add options to inflate the ensemble while recentering (#1501) Add components to use jcb yamls for marine letkf job (#1486) Add GNSSRO bufr2ioda and json files (#1448) Removal of the marine verification tools (#1499) update logic to identify MSU machine (#1498) IODA Converter with In-Situ Observations Concatenation and ObsError Inflation (#1472) Add sea ice recentering (#1478) marine vrfy bugfix (#1483) add LD_LIBRARY_PATH patch for wcoss2 build (#1484) Bugfix for ensemble recenter + add a test (#1477) Add code changes for the implementation of the low-resolution B (#1441) update spack-stack path in Hera modulefiles (#1471) fixed glider output filename (#1469) add gdas_waveinit to C48mx500_3DVarAOWCDA ctest suite (#1470) update jcb-gdas hash 20250128 (#1466) Time series of omb stats and html doc (#1463) Update marine vrfy HTML codes (#1464) Automate the verify tools to submit from start pdy to end pdy (#1458) Update an offline marine vrfy tool (#1447) Update JEDI hashes (20250127) (#1460)
DavidNew-NOAA
pushed a commit
that referenced
this pull request
Jan 16, 2026
# Description Ocean ensemble recentering was double-counting ensemble perturbations, effectively inflating the ensemble by a factor of 2 (almost, the extra perturbation was added with IAU, so some of it was probably damped). The accumulation of amplified perturbations over several cycles may have been the cause of ensemble forecast failures in gfsv17 trial experiments. On develop: IAU increment for ocean recentering = (det-mean) + (ensmember-mean). In this branch: IAU increment for ocean recentering = (det-mean) With this branch, as expected, IAU increments passed to the ocean model are all the same, pulling from the mean towards deterministic background. Tested with C48 one cycle, compared IAU increments and looked through the logs. # Issues Resolves #1557 # Automated CI tests to run in Global Workflow <!-- Which Global Workflow CI tests are required to adequately test this PR? --> - [ ] atm_jjob <!-- JEDI atm single cycle DA !--> - [ ] C96C48_ufs_hybatmDA <!-- JEDI atm cycled DA !--> - [ ] C96C48_hybatmaerosnowDA <!-- JEDI aero/snow cycled DA !--> - [ ] C48mx500_3DVarAOWCDA <!-- JEDI low-res marine 3DVar cycled DA !--> - [x] C48mx500_hybAOWCDA <!-- JEDI marine hybrid envar cycled DA !--> - [ ] C96C48_hybatmDA <!-- GSI atm cycled DA !--> Co-authored-by: Anna Shlyaeva <anna.v.shlyaeva@noaa.gov>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Ocean ensemble recentering was double-counting ensemble perturbations, effectively inflating the ensemble by a factor of 2 (almost, the extra perturbation was added with IAU, so some of it was probably damped). The accumulation of amplified perturbations over several cycles may have been the cause of ensemble forecast failures in gfsv17 trial experiments.
On develop:
IAU increment for ocean recentering = (det-mean) + (ensmember-mean).
In this branch:
IAU increment for ocean recentering = (det-mean)
With this branch, as expected, IAU increments passed to the ocean model are all the same, pulling from the mean towards deterministic background.
Tested with C48 one cycle, compared IAU increments and looked through the logs.
Issues
Resolves #1557
Automated CI tests to run in Global Workflow