General announcements

  • Next meeting in two weeks (no special topic for next week's time slot)
  • Send in any special topics you are interested in presenting and discussing

Interfaces & Optimization - Francois H.

  • Raji:
    • Created PR for updating timedate format to epoch, and height conversion for surface temperature operator
    • Working on integrating C++ surface temperature operator
  • Steve V. working on design for the Distribute step of the new Load, Filter, Distribute process for creating an ObsSpace with OSDF container
  • Steve H.
    • cleaned up some IODA test dependency issues
    • New reader refactoring
    • The BUFR parallel reader was merged. We now have parallel reading capability for both BUFR and ODB files in develop! Thanks to our partners for these contributions.
  • Liam is continuing work on GPU offloading in Atlas, Ectrans for various steps including adjoint, inverse, etc.


Algo - Christian

Nate has created PRs for doing the lat-lon writing for MPAS forecasts, but has a small bug to fix with those.
Nate is also working on PRs for other small bug fixes related to the multi-scale B.
Christian is working on continuous DA to shift the beginning of the window as well as work flow to test the changes in Skylab
Clementine recently returned from vacation and is working on previous PR's for weather bench scores and PR's reviews as well.


SOCA - Travis

No report


COMPO - Jerome

  • Mariam advancing first tempo paper
  • Jerome working on 2nd tempo paper
  • Shih-Wei - pace


OBS - Francois V.

  • Ben is presenting at the ITSC Conference in Goa, India this week.

    OBS processing:

    PR #1630 was merged: Feature/add gnssro bufr subsatid.


    Handling of the presence or absence of  GNSS-RO satelliteSubIdentifier. Needed for commercial data.


    PR #1649 under review: Fix reader in bufr2nc_fortran.x converters for Himawari binary observation files available from NOAA AWS S3 explorer. 

    IODA converter for Himawari observations provided by NOAA on AWS.


    PR #1647 under review: Converter for COWVR for the NASA STPH8 mission ocean surface wind product.



  • UFO


    3 PRs to reduce geovals memory footprint and speed up the process:


    PR #3694 is approved: Skip requesting Cloud_Fraction from GeoVals when it is calculated in ufo_crtm_untils. 


    Skip requesting Cloud_Fraction from GeoVals when it is going to be calculated, e.g., by the Thompson method within the FOVs, in UFO. This avoids failures in UFO when there is no Cloud_Fraction data in GeoVals. Thanks Jianjun.


    PR #3657 is under review: Observation operator that reduces the spatial density of requested GeoVaLs. 

    Select one GeoVaL for several observation locations. This algorithm draws a GeoVaL at every k-th observation location and copies it to the other k - 1 GeoVaLs. 


    PR #3685 is under review. Lat-lon grid algorithm for Density Reduction operator:. 

    This algorithm divides the assimilation domain into a regular lat-lon grid of user-specified spacing and extent. It draws one GeoVaL from each lat-lon grid box and uses it to represent the values of any other GeoVaLs in that box. Thanks Chris Thomas.

    PR #3666 Fix a bug: Add a geometric to geopotential height conversion for Met Office refractivity: 

    The RefMetOffice GNSSRO forward operator was previously reading geometric height from the MetaData/height variable, but treating it as a geopotential height when comparing to model data. Thanks Neill

    To be detailed during Tuesday’s OBSPROC meeting


    PR #3682 is going to be merged: Adding C++ code for Pressure Surface operator for UKMO and GSL scheme.

    Converging pressure surface operator along with temperature surface operator. Thanks Rajis



CRTM - Cheng

  • - Potential bug fixes:
       - Oxygen-affected MW bands.
       - K-matrix calculation for azimuthal dependence.
       - CRTM results binary versus netCDF instrument coefficients (v.viirs-mj2).
       - Pending hotfix PR review: surface emissivity Jacobians.
    - Generating CRTM unit test benchmark data.

    - OSN CRTM anonymous buckets ready for data access, needs more discussion with the infrastructure team on how to move forward.


NOAA-EMC - Dan

E2 Compliance:

Working toward realtime retros

Found a workaround for the need for static libraries, so don’t need that right now.

Need to eliminate using forks of several repos

Found workaround for building SOCA


Development:

Regional background error modeling.

Now have GSIbec

Can reproduce localization for GSI


GMAO

Ron-Atmosphere

Expanding use of JEDI in GMAO. Had tutorials.

Can run a full cycling system

Testing system at C720 for resource assessment

Dorook-Ocean

Testing with swat data.

Switching land mask grid to more standard one.


MMM

Jake

Recently completed tests with higher resolution than previous. Evaluating results.

Chris S.

  • Consider switching to using virtual temperature instead of temperature for analysis variable

    Discuss enabling switching of analysis variables in general.

    Yannick says this is mostly a SABER thing.


Met Office - MikeC

    • UFO PR (#3666) gone in which add geopotential height conversion to GNSSRO refractivity forward operator as part of work with PlanetIQ.
    • Obs operator wrapper for geoval reduction.
    • IODA CI testing to be added to IODA.

     
    • ETKF (transform Kalman filters) refactoring: there are two PRs under review (linear algebra and computeHofX). Both of this simplify the logic of the local volume solvers and enable easier extension in the future.
    • Ensemble splitter PR merged. Next step integration into refactored solvers.
    • Replacement of IncrementEnsemble4D with IncrementSet PR merged (including coordinated PRs in other repositories).
    • Proposal on treatment of correlated R in development.
    • PRs improving the QC masking have been merged.
    • Further scientific improvements in development.
    • Moving towards deciding between the various flavours of EnKF.


OAR - Ming

Using MPAS for regional cycling

Tested 3DVar and 3DEnVar

Also ran GSI 3DVar.

Have tuned static B so that Seeing positive impact with 3DVar


ROMS - HernanA

Frustrated with local vs. global file in BUMP

Feels debugging BUMP is hard due to so many flags/switches and difficult variable names


NRL - Dom

Working on 4DVar


  • No labels