NOAA-EMC / GSI

Gridpoint Statistical Interpolation
GNU Lesser General Public License v3.0
66 stars 150 forks source link

Inclusion of MTG IRS data #553

Closed jderber-NOAA closed 9 months ago

jderber-NOAA commented 1 year ago

This issue is for the inclusion of Meteosat Third Generation (MTG) InfraRed Sounder (IRS) data in the GSI. The inclusion differs from the inclusion of some of the other types of data in that 1. The data comes into NCEP in the form of Principle Components (PCs). Then in the obsproc data processing, several things are done to the data (e.g., thinning, conversion to radiances, use of a Hamming-apodization) before it is output as BUFR data and then eventually fed into the GSI. While the obsproc part of the data processing is not technically part of the GSI, it does have implications for the use of the data. Also, note that there is currently no real data from this instrument because the data has not yet been launched.

We decided to use the NWP-SAF code for preprocessing the data (with some modifications). This code should not be redistributed without permission of the NWP-SAF. No redistribution of these BUFR files is planned (except for GSI research purposes). Attached to this write-up is the initial documentation from the NWP-SAF. Also attached are the release notes for version 1.1.

NWPSAF-MO-UD-053-IRSPP_User_Manual.pdf

Release_note_IRSPP_v1.2.pdf

Important Items/Issues from preprocessing.

  1. Currently preprocessing uses NETCDF, HDF5 and LAPACK. The use of eccodes has been removed. The conversion to use the libraries currently on the ncep machines of netcdf, hdf5 and LAPACK have been unsuccessful and will have to await someone else smarter than me.
  2. Current version allows selection of warmest FOV to make it more likely to select clear FOVs out of 10x10 thinning array.
  3. Output of the radiances has been modified to use the NCEP BUFR routines. Note that the PCs were not output in the current version (could be done, but could not think of any use for them). There are 2 local descriptors (the same as for the NWP-SAF) code (CSSQ and CDSQ) which are 2 quality tables.

Contents of BUFR file

! SAID 001007 satelliteIdentifier (CCT C-5) ! OGCE 001033 centre (CCT C-1) ! GSES 001034 subCentre (CCT C-12) ! SIID 002019 satelliteInstruments (CCT C-8) ! SCLF 002020 satelliteClassification (334 = MTG?) ! ! YEAR 004001 Year ! MNTH 004002 Month ! DAYS 004003 Day ! HOUR 004004 Hour ! MINU 004005 Minute ! SECO 004006 Second ! FOVN 005043 fieldOfViewNumber (1 to 25600 - 160x160) ! SLNM 005041 scanLineNumber ! FORN 005045 fieldOfRegardNumber (dwell) ! RDTF 002165 radianceTypeFlags (4=apodized, 5=unapodized) ! SELV 007001 heightOfStation (m, to nearest 100m, geostationary height range) ! CLATH 005001 High resolution latitude ! CLONH 006001 High resolution longitude ! SAZA 007024 satelliteZenithAngle ! BEARAZ 005021 bearingOrAzimuth ! SOZA 007025 solarZenithAngle ! SOLAZI 005022 solarAzimuth ! CHSF 025142 channelScaleFactor ! SMRA 014047 scaledMeanAvhrrRadiance (imager mode, 3x3 pixels) ! SSDR 014048 scaledStandardDeviationAvhrrRadiance ! CONFLG 025187 confidenceFlag (0=valid, 1=invalid, 15=missing) ! SQFA 040026 scoreQuantizationFactor ! RRIB 040016 residualRmsInBand ! CHNM 005042 channelNumber ! SRAD 014044 channelRadiance ! TOBD 008076 band (2=LW 3=MW)

! CSSQ : Quality flags: ! spatial_sample_quality, 13-bit (bit 1 defined here as most significant) ! Bit Meaning ! 1-4 reserved ! 5 solar_straylight_correction_warning ! 6 solar_straylight_warning ! 7 noisy_detector_sample_warning ! 8 undersaturated_detector_sample_warning ! 9 saturated_detector_sample_warning ! 10 dust ! 11 cloudy ! 12 limb_view ! 13 space_view ! ! CDSQ detector_sample_quality, 4-bit (bit 1 defined here as most significant) ! Bit Meaning ! 1 excluded_detector_sample ! 2 noisy_detector_sample ! 3 undersaturated_detector_sample ! 4 saturated_detector_sample

With the creation of the new BUFR files for the MTG IRS data, modifications were made to process this data within the GSI.

Two major issues.

  1. The initial CRTM coefficients are not yet available from NESDIS, so testing of the use of this data ended at the point these coefficients were necessary.
  2. Since real data is not available, the quality control is not yet developed.

In the current version the satellite is referred to as MTG1 and the instrument as IRS. Future satellites would be MTG2,...

There are currently produced 198 channels. Please see documentation from NWPSAF for details. Note that the Longwave (LW) and Mid-Wave (MW) are included together, not as separate groups as in the input PCs.

RussTreadon-NOAA commented 11 months ago

@jderber-NOAA , may we close this issue?

RussTreadon-NOAA commented 11 months ago

@jderber-NOAA and @ADCollard , may this issue be closed?

jderber-NOAA commented 9 months ago

Russ,

Last I heard they were going to wait until Jedi to work on this one. So it can be closed unless Andrew wants it kept open.

Removed myself from assignees.

RussTreadon-NOAA commented 9 months ago

Thank you, @jderber-NOAA , for the feedback.

@ADCollard , would you like this GSI issue to remain open or shall we open a new issue in GDASApp, JEDI-T20, or the appropriate JEDI repo?

ADCollard commented 9 months ago

@RussTreadon-NOAA Please close this issue.