GSI Fundamentals (2) Run and Namelist

Transcription

1 24 GSI Community Tutorial July 4-6, 24, NCAR, Boulder GSI Fundamentals (2) Run and Namelist Ming Hu and Hui Shao Developmental Testbed Center

2 Outlines GSI fundamentals (): Setup and Compilation GSI fundamentals (2): Run and Namelist Input files needed by GSI Run GSI with a sample run script Configure GSI with GSI namelist Set up data usage GSI fundamentals (3): Diagnostics GSI fundamentals (4): Applications This talk is tailored based on the Chapter 3 of the GSI User s Guide 2

3 GSI input files Run GSI with a sample run script Configure GSI with GSI namelist Set up data usage Before You Run GSI User s Guide (section 3.) 3

4 GSI input and output files Analysis results; stdout and diagnostic files, J(x) = (x-xb)tb-(x-xb)+(y-h[x])tr-(y-h[x]) + Jc Background file Observations files Background covariance file CRTM coefficients Observation error file Control and state variable info file Observation data control info files Fixed files Namelist Radiance bias correction coefficient files 4

5 Background file Types of background fields a) b) c) d) e) f) g) WRF NMM input fields in binary format WRF NMM input fields in NetCDF format WRF ARW input fields in binary format WRF ARW input fields in NetCDF format GFS input fields in binary format NEMS-NMMB input fields RTMA input filed (2-dimensional binary format) DTC mainly support (a)-(d) Cold start: WPS + REAL : wrfinput_d Cycle: forecast file from WRF model : wrfout_d_(time) 5

6 Sample Observation files (table 3.) GSI Name prepbufr satwnd amsuabufr amsubbufr radarbufr gpsrobufr ssmirrbufr tmirrbufr sbuvbufr hirs2bufr hirs3bufr hirs4bufr msubufr airsbufr mhsbufr ssmitbufr amsrebufr ssmisbufr gsndbufr l2rwbufr gsndrbufr gimgrbufr omibufr iasibufr gomebufr mlsbufr tcvitl seviribufr atmsbufr crisbufr modisbufr Content Conventional observations, including ps, t, q, pw, uv, spd, dw, sst, from observation platforms such as METAR, sounding, et al. satellite winds observations AMSU-A b radiance (brightness temperatures) from satellites NOAA-5, 6, 7,8, 9 and METOP-A/B AMSU-B b radiance (brightness temperatures) from satellites NOAA-5, 6,7 Radar radial velocity Level 2.5 data GPS radio occultation and bending angle observation Precipitation rate observations from SSM/I Precipitation rate observations from TMI SBUV/2 ozone observations from satellite NOAA-6, 7, 8, 9 HIRS2 b radiance from satellite NOAA-4 HIRS3 b radiance observations from satellite NOAA-6, 7 HIRS4 b radiance observation from satellite NOAA-8, 9 and METOP-A/B MSU observation from satellite NOAA 4 AMSU-A and AIRS radiances from satellite AQUA Microwave Humidity Sounder observation from NOAA-8, 9 and METOP-A/B SSMI observation from satellite f3, f4, f5 AMSR-E radiance from satellite AQUA SSMIS radiances from satellite f6 GOES sounder radiance (sndrd, sndrd2, sndrd3 sndrd4) from GOES-, 2, 3, 4, 5. NEXRAD Level 2 radial velocity GOES sounder radiance from GOES-, 2 GOES imager radiance from GOE-, 2 Ozone Monitoring Instrument (OMI) observation NASA Aura Infrared Atmospheric Sounding Interfero-meter sounder observations from METOP-A/B The Global Ozone Monitoring Experiment (GOME) ozone observation from METOP-A/B Aura MLS stratospheric ozone data from Aura Synthetic Tropic Cyclone-MSLP observation SEVIRI radiance from MET-8,9, ATMS radiance from Suomi NPP CRIS radiance from Suomi NPP MODIS aerosol total column AOD observations from AQUA and TERRA Example file names gdas.t2z.prepbufr gdas.t2z.satwnd.tm.bufr_d gdas.t2z.bamua.tm.bufr_d gdas.t2z.bamub.tm.bufr_d ndas.t2z. radwnd. tm2.bufr_d gdas.t2z.gpsro.tm.bufr_d gdas.t2z.spssmi.tm.bufr_d gdas.t2z.sptrmm.tm.bufr_d gdas.t2z.osbuv8.tm.bufr_d gdas.t2z.bhrs2.tm.bufr_d gdas.t2z.bhrs3.tm.bufr_d gdas.t2z.bhrs4.tm.bufr_d gdas.t2z.bmsu.tm.bufr_d gdas.t2z.airsev.tm.bufr_d gdas.t2z.bmhs.tm.bufr_d gdas.t2z.ssmit.tm.bufr_d gdas.t2z.amsre.tm.bufr_d gdas.t2z.ssmis.tm.bufr_d gdas.t2z.goesfv.tm.bufr_d ndas.t2z.nexrad. tm2.bufr_d gdas.t2z.goesnd.tm.bufr_d gdas.t2z.omi.tm.bufr_d gdas.t2z.mtiasi. tm.bufr_d gdas.t2z.gome.tm.bufr_d gdas.t2z. mlsbufr.tm.bufr_d gdas.t2z.syndata.tcvitals.tm gdas.t2z. sevcsr.tm.bufr_d gdas.t2z. atms.tm.bufr_d gdas.t2z. cris.tm.bufr_d 6

7 Sample Observation files (table 3.) GSI Name prepbufr satwnd amsuabufr amsubbufr radarbufr gpsrobufr ssmirrbufr tmirrbufr sbuvbufr hirs2bufr hirs3bufr hirs4bufr msubufr airsbufr mhsbufr ssmitbufr amsrebufr ssmisbufr gsndbufr l2rwbufr gsndrbufr gimgrbufr omibufr Content Conventional observations, including ps, t, q, pw, uv, spd, dw, sst, from observation platforms such as METAR, sounding, et al. satellite winds observations AMSU-A b radiance (brightness temperatures) from satellites NOAA-5, 6, 7,8, 9 and METOP-A/B AMSU-B b radiance (brightness temperatures) from satellites NOAA-5, 6,7 Radar radial velocity Level 2.5 data GPS radio occultation and bending angle observation Precipitation rate observations from SSM/I Precipitation rate observations from TMI SBUV/2 ozone observations from satellite NOAA-6, 7, 8, 9 HIRS2 b radiance from satellite NOAA-4 HIRS3 b radiance observations from satellite NOAA-6, 7 HIRS4 b radiance observation from satellite NOAA-8, 9 and METOP-A/B MSU observation from satellite NOAA 4 AMSU-A and AIRS radiances from satellite AQUA Microwave Humidity Sounder observation from NOAA-8, 9 and METOP-A/B SSMI observation from satellite f3, f4, f5 AMSR-E radiance from satellite AQUA SSMIS radiances from satellite f6 GOES sounder radiance (sndrd, sndrd2, sndrd3 sndrd4) from GOES-, 2, 3, 4, 5. NEXRAD Level 2 radial velocity GOES sounder radiance from GOES-, 2 GOES imager radiance from GOE-, 2 Ozone Monitoring Instrument (OMI) observation NASA Aura Infrared Atmospheric Sounding Interfero-meter sounder observations from METOP-A/B The Global Ozone Monitoring Experiment (GOME) ozone observation from METOP-A/B Aura MLS stratospheric ozone data from Aura Synthetic Tropic Cyclone-MSLP observation SEVIRI radiance from MET-8,9, ATMS radiance from Suomi NPP CRIS radiance from Suomi NPP MODIS aerosol total column AOD observations from AQUA and TERRA Example file names gdas.t2z.prepbufr gdas.t2z.satwnd.tm.bufr_d gdas.t2z.bamua.tm.bufr_d gdas.t2z.bamub.tm.bufr_d ndas.t2z. radwnd. tm2.bufr_d gdas.t2z.gpsro.tm.bufr_d gdas.t2z.spssmi.tm.bufr_d gdas.t2z.sptrmm.tm.bufr_d gdas.t2z.osbuv8.tm.bufr_d gdas.t2z.bhrs2.tm.bufr_d gdas.t2z.bhrs3.tm.bufr_d gdas.t2z.bhrs4.tm.bufr_d gdas.t2z.bmsu.tm.bufr_d gdas.t2z.airsev.tm.bufr_d gdas.t2z.bmhs.tm.bufr_d gdas.t2z.ssmit.tm.bufr_d gdas.t2z.amsre.tm.bufr_d gdas.t2z.ssmis.tm.bufr_d gdas.t2z.goesfv.tm.bufr_d ndas.t2z.nexrad. tm2.bufr_d gdas.t2z.goesnd.tm.bufr_d Only part of the observation files available in real application All observation files are in BUFR format Details on observation data process, BUFR file process will be introduced in lecture: iasibufr gomebufr mlsbufr tcvitl seviribufr atmsbufr crisbufr modisbufr gdas.t2z.omi.tm.bufr_d gdas.t2z.mtiasi. tm.bufr_d gdas.t2z.gome.tm.bufr_d gdas.t2z. mlsbufr.tm.bufr_d gdas.t2z.syndata.tcvitals.tm gdas.t2z. sevcsr.tm.bufr_d gdas.t2z. atms.tm.bufr_d gdas.t2z. cris.tm.bufr_d Data Processing: Observation process and BUFR/PrepBUFR 7

8 CRTM coefficients. Fixed files Each fix file has many choices from different GSI applications Table 3.2 lists fixed file names required in a GSI run, the content of the files, and corresponding example files from the regional, global and RTMA applications: Table 3.2 GSI fixed files, content, and examples File name used in GSI Content Example files in fix/ anavinfo Information file to set control and analysis variables berror_stats background error covariance errtable Observation error table anavinfo_arw_netcdf anavinfo_ndas_netcdf global_anavinfo.l64.txt anavinfo_rtma_gust_vis_7vars nam_nmmstat_na.gcv nam_glb_berror.f77.gcv global_berror.l64y386.f77 new_rtma_regional_nmm_berror.f77.gcv nam_errtable.r3dv prepobs_errtable.global Observation data control file (more detailed explanation in Section 4.3) convinfo Conventional observation information file satinfo pcpinfo satellite channel information file precipitation rate observation information file ozone observation information file ozinfo global_convinfo.txt nam_regional_convinfo.txt new_rtma_regional_convinfo.txt global_satinfo.txt global_pcpinfo.txt global_ozinfo.txt Bias correction and Rejection list satbias_angle satbias_in t_rejectlist, w_rejectlist,.. satellite scan angle dependent bias correction file satellite mass bias correction coefficient file Rejetion list for T, wind, et al. in RTMA global_satangbias.txt sample.satbias new_rtma_t_rejectlist new_rtma_w_rejectlist 8

9 Big Endian BE files Big_Endian with the release version. Global BE files and ch4, co, co2, n2o history files global Allfor./fix fixed files are linked run directory Fix GSIdirectory RTMA application rtma To files make easy to manage, this to release version created 4 sub-d Fix for GSI RAP of application rap holdfiles special group fix files, whichon are introduced in table 3.3. by run script based user s option Please note released GSI 3.3 tar files doesn t include./fix/global and./fix/rtma for Table 3.3 List of sub-directories space saving. Please downloand comgsi_v3.3_fix_global.tar.gz if you needintofix rundirectory Directory name global case and comgsi_v3.3_fix_rtma.tar.gz to runcontent RTMA case from the GSI user s Little Endian Bacground Error covariance (BE) files Little_Endian Manywebpage. subdirectories Big Endian BE files Big_Endian to make fix directory Each release version GSI calls certain version of CRTM library Global BE filesand andneeds ch4, co,the co2, n2o history files global data This Fix files for GSIassimilation. RTMA application rtma coefficients to do radiance clean corresponding version of CRTM version of GSI uses CRTM are listed table 3.4. Fix files for GSIin RAP application rap The coefficients files Please note released GSI 3.3 tar files doesn t include./fix/global and./fix/r Table 3.4 List ofsaving. radiance coefficients used by CRTM space Please downloand comgsi_v3.3_fix_global.tar.gz if you nee File name used in GSI global case Content Example files to run RTMA case from the and comgsi_v3.3_fix_rtma.tar.gz Nalli.IRwater.EmisCoeff.bin webpage. IR surface emissivity Nalli.IRwater.EmisCoeff.bin Fixed files NPOESS.IRice.EmisCoeff.bin coefficients NPOESS.IRice.EmisCoeff.bin NPOESS.IRsnow.EmisCoeff.bin NPOESS.IRsnow.EmisCoeff.bin Each release version GSI calls certain version of CRTM library and needs NPOESS.IRland.EmisCoeff.bincorresponding version of CRTM NPOESS.IRland.EmisCoeff.bin coefficients to do radiance data assimilati NPOESS.VISice.EmisCoeff.binversion of GSI uses CRTM NPOESS.VISice.EmisCoeff.bin The coefficients files are listed in table 3 NPOESS.VISland.EmisCoeff.bin NPOESS.VISland.EmisCoeff.bin NPOESS.VISsnow.EmisCoeff.bin NPOESS.VISsnow.EmisCoeff.bin Table 3.4 List of radiance coefficients used by CRTM NPOESS.VISwater.EmisCoeff.bin NPOESS.VISwater.EmisCoeff.bin File name used in GSI Content Example files FASTEM5.MWwater.EmisCoeff.bin FASTEM5.MWwater.EmisCoeff.bin Aerosol coefficients AerosolCoeff.bin AerosolCoeff.bin Nalli.IRwater.EmisCoeff.bin IR surface emissivity Nalli.IRwater.EmisC NPOESS.IRice.EmisCoeff.bin coefficients NPOESS.IRice.Emis Cloud scattering and CloudCoeff.bin CloudCoeff.bin NPOESS.IRsnow.EmisCoeff.bin NPOESS.IRsnow.Em emission coefficients NPOESS.IRland.EmisCoeff.bin Sensor spectral response ${satsen}.spccoeff.bin NPOESS.IRland.Emi ${satsen}.spccoeff.bin NPOESS.VISice.EmisCoeff.bin NPOESS.VISice.Emi characteristics NPOESS.VISland.EmisCoeff.bin Transmittance ${satsen}.taucoeff.bin ${satsen}.taucoeff.bin NPOESS.VISland.Em 9 NPOESS.VISsnow.EmisCoeff.bin NPOESS.VISsnow.E coefficients NPOESS.VISwater.EmisCoeff.bin NPOESS.VISwater.E

10 Namelist and Bias Coefficient Namelist is generated in GSI run script gsiparm.anl Build the GSI namelist on-the-fly cat << EOF > gsiparm.anl &SETUP miter=2,niter()=,niter(2)=, write_diag()=.true.,write_diag(2)=.false.,write_diag(3)=.true., gencode=78,qoption=2, factqmin=.,factqmax=., ndat=87,iguess=-, oneobtest=.false.,retrieval=.false., nhr_assimilation=3,l_foto=.false., use_pbl=.false., / &GRIDOPTS JCAP=62,JCAP_B=62,NLAT=6,NLON=6,nsig=6,regional=.true., wrf_nmm_regional=${bk_core_nmm},wrf_mass_regional=${bk_core_arw}, diagnostic_reg=.false., filled_grid=.false.,half_grid=.true.,netcdf=.true., / Bias coefficient files Only sample coefficient files available: Sample.satbias, global_satangbias.txt rap_satbias_in_enhanced, rap_satbias_pc_enhanced

11 Before run GSI, you should know Location and name of background file Location and names of observation files Location of the Fix directory Location of the CRTM coefficients directory Have compiled GSI successfully: gsi.exe Where do you want to run GSI (run directory) Analysis time and configurations Which background error convariance file to use Observation data types used Too Much! No problem, let GSI run script help Computer resources you can use and how to use

12 GSI input files Run GSI with a sample run script Configure GSI with GSI namelist Set up data usage We do provide several sample run scripts for you to start with. The major one for regional is: ~/comgsi_v3.3/run/run_gsi.ksh 2 GSI User s Guide (section 3.2)

13 GSI run script steps Request computer resources to run GSI. setup Set environmental variables for the machine architecture. Set experimental variables. Check the definitions of required variables. Check setups Prepare run Generate a run directory (working directory). Copy the GSI executable to the run directory. directory Copy background file and link ensemble members if running the hybrid to the run directory. Link observations to the run directory. Link fixed files (statistic, control, and coefficient files) to the run directory. Generate namelist for GSI in the run directory. Run the GSI executable. Post-process: save analysis results, generate diagnostic files, clean run directory. 3

14 GSI run script steps setup Request computer resources to run GSI. Set environmental variables for the machine architecture.!/bin/ksh machine set up (users should change this part) GSIPROC = processor number used for GSI analysis GSIPROC= ARCH='LINUX' Supported configurations: IBM_LSF LINUX, LINUX_LSF, LINUX_PBS, DARWIN_PGI 4

15 GSI run script steps setup Request computer resources to run GSI. Set environmental variables for the machine architecture.!/bin/ksh IBM with LSF : BSUB -P???????? machine set up (users should change this part) BSUB -a poe BSUB -x BSUB -n 2 BSUB -R "span[ptile=2] GSIPROC = processor number used for GSI analysis BSUB -J gsi GSIPROC= Linux Cluster with PBS: BSUB -e gsi.err ARCH='LINUX' $ -S /bin/ksh BSUB -W :2 Supported configurations: $ -N GSI_test BSUB -q regular IBM_LSF, IBM_LoadLevel $ -cwd LINUX, LINUX_LSF, LINUX_PBS, Linux workstation: $ -r y set -x DARWIN_PGI $ -pe comp 64 $ -l h_rt=:2: $ -A?????? Set environment variables for IBM export MP_SHARED_MEMORY=yes 5 export MEMORY_AFFINITY=MCM

16 GSI run script steps setup Request computer resources to run GSI. Set environmental variables for the machine architecture.!/bin/ksh IBM with LSF : machine set up (users should change this part) GSIPROC=2 ARCH= IBM_LSF Linux Cluster with PBS: GSIPROC = processor number used for GSI analysis GSIPROC= ARCH='LINUX_PBS GSIPROC= ARCH='LINUX' Supported configurations: Linux workstation: IBM_LSF, IBM_LoadLevel LINUX, LINUX_LSF, LINUX_PBS, GSIPROC= ARCH='LINUX DARWIN_PGI ARCH is used to decide the RUN_COMMAND : IBM with LSF: RUN_COMMAND="mpirun.lsf " Linux Cluster with PBS: RUN_COMMAND="mpirun -np ${GSIPROC} 6

17 GSI run script : setup case Set experimental variables. setup case set up (users should change this part) ANAL_TIME= analysis time (YYYYMMDDHH) WORK_ROOT= working directory, where GSI runs BK_FILE = path and name of background file Where are OBS_ROOT = path of observations files input data PREPBURF = path of PreBUFR conventional obs CRTM_ROOT= path of crtm coefficients files Where is GSI FIX_ROOT = path of fix files GSI_EXE = path and name of the gsi executable system ANAL_TIME=23222 WORK_ROOT=gsi/code/comGSI_v3.3/run/test_guide OBS_ROOT=gsi/data/23222/obs PREPBUFR=${OBS_ROOT}/gdas.t2z.prepbufr.nr BK_FILE=gsi/data/23222/wrfinput_d_ARW_2-3-22_2 FIX_ROOT=gsi/code/comGSI_v3.3/fix CRTM_ROOT=gsi/CRTM_REL-2..3 GSI_EXE=gsi/code/comGSI_v3.3/run/gsi.exe 7

18 GSI run script : setup case setup Set experimental variables bk_core= which WRF core is used as background (NMM or ARW) bkcv_option= which background error covariance and parameter will be used (GLOBAL or NAM) if_clean = clean :delete temporal files in working directory (default) no : leave running directory as is (this is for debug only) bk_core=arw bk_core =ARW : ARW NetCDF bkcv_option=nam =NMM : NMM NetCDF if_clean=clean bkcv_option =NAM : background error file based on NAM (north H) =GLOBAL : background error file based on GFS (global) Users should NOT change script after this point In comgsi_v3.3, all binary files are Big_Endian files BUT BYTE_ORDER=Big_Endian BYTE_ORDER=Little_Endian If compiled with gfortran, set: BYTE_ORDER=Little_Endian 8

19 GSI run script : setup hybrid case GSI Applications 5.4 Introduction to GSI hybrid analysis hybrid ensemble-3dvar analysis isv3.3 an important No scripts arethe available in comgsi foranalysis this.option in the GSI system that has been used by operations. It provides the ability to bring the flow dependent background error covariance into the analysis based on ensemble forecasts. If ensemble please note that we the ensemble are located and forecasts haveassume been generated, setting up GSI to do amember hybrid analysis is straightforward only requires twowith changes in the run wrfout_d_${iiimem}. script in addition to the current 3DVAR run script: under ${mempath} name If this is not the case,members please change the following lines: Change : Link the ensemble to the GSI run directory This change is to link the ensemble members to the GSI run if [ -r ${mempath}/wrfout_d_${iiimem} ];directory thenand assign each ensemble member a name that GSI recognizes. The release version GSI can ln -sf ${mempath}/wrfout_d_${iiimem}./wrf_en${iiimem} accept 4 kinds of ensemble forecasts, which is controlled by the namelist variable regional_ensemble_option. Table 5. gives a list of options for else regional_ensemble_option and the naming convention for linking the ensemble to echo "member ${mempath}/wrfout_d_${iiimem} is not exit" GSI recognized names fi Users needs to setup it based on table 5. Table 5. the list of ensemble forecasts that can be read by GSI hybrid regional_ explanation Function called GSI recognized ensemble_ ensemble file names option GFS ensemble internally get_gefs_for_regional filelist : a text file 2 interpolated to hybrid grid ensembles are WRF NMM (HWRF) format get_wrf_nmm_ensperts 3 ensembles are ARW netcdf format get_wrf_mass_ensperts_ netcdf 4 ensembles are NEMS NMMB format get_nmmb_ensperts include path and name of ensemble files d_en, d_en2, wrf_en, wrf_en2, nmmb_ens_mem, nmmb_ens_mem2, 9

20 GSI run script: setup run directory Generate a run directory (working directory). Prepare run Copy the GSI executable to the run directory. directory Copy background file and link ensemble members if running the hybrid to the run directory. Link observations to the run directory. Link fixed files (statistic, control, and coefficient files) to the run directory. Generate namelist for GSI in the run directory. Bring over background field (it's modified by GSI so we can't link to it) cp ${BK_FILE}./wrf_inout Link ensember members if use hybrid if [ ${if_hybrid} =.true. ] ; then if [ -r ${mempath}/wrfout_d_${iiimem} ]; then ln -sf ${mempath}/wrfout_d_${iiimem}./wrf_en${iiimem} else echo "member ${mempath}/wrfout_d_${iiimem} is not exit" exit fi 2

21 GSI run script: setup run directory Generate a run directory (working directory). Prepare run Copy the GSI executable to the run directory. directory Copy background file and link ensemble members if running the hybrid to the run directory. Link observations to the run directory. Link fixed files (statistic, control, and coefficient files) to the run directory. Generate namelist for GSI in the run directory. Link to the prepbufr data ln -s ${PREPBUFR}./prepbufr Byte-order: Doesn t Matter Anymore New BUFRLIB in comgsi_v3.3 can automatically identify the byte-order and handle both Little/Big_Endian BUFR/PrepBUFR files Link to the radiance data ln -s ${OBS_ROOT}/gdas.t2z.bamua.tm.bufr_d ln -s ${OBS_ROOT}/gdas.t2z.bamub.tm.bufr_d ln -s ${OBS_ROOT}/gdas.t2z.bhrs3.tm.bufr_d ln -s ${OBS_ROOT}/gdas.t2z.bhrs4.tm.bufr_d ln -s ${OBS_ROOT}/gdas.t2z.bmhs.tm.bufr_d ln -s ${OBS_ROOT}/gdas.t2z.gpsro.tm.bufr_d amsuabufr amsubbufr hirs3bufr hirs4bufr mhsbufr gpsrobufr 2

22 GSI run script: create GSI namelist Generate namelist for GSI in the run directory. Build the GSI namelist on-the-fly cat << EOF > gsiparm.anl &SETUP miter=2,niter()=,niter(2)=, write_diag()=.true.,write_diag(2)=.false.,write_diag(3)=.true., gencode=78,qoption=2, factqmin=.,factqmax=., ndat=87,iguess=-, oneobtest=.false.,retrieval=.false., nhr_assimilation=3,l_foto=.false., use_pbl=.false., / &GRIDOPTS JCAP=62,JCAP_B=62,NLAT=6,NLON=6,nsig=6, regional=.true., wrf_nmm_regional=${bk_core_nmm},wrf_mass_regional=${bk_core_arw}, diagnostic_reg=.false., filled_grid=.false.,half_grid=.true.,netcdf=.true., / &BKGERR vs=${vs_op} hzscl=${hzscl_op} Complete list is in Advanced User s Guide Appendix. bw=.,fstat=.true., Most of options should stick with default values / 22

23 GSI run script: run and submit Run the GSI executable case $ARCH in 'IBM_LSF' 'IBM_LoadLevel') ${RUN_COMMAND}./gsi.exe < gsiparm.anl > stdout 2>& ;; * ) ${RUN_COMMAND}./gsi.exe > stdout 2>& ;; esac To submit the run script IBM supercomputer LSF: IBM supercomputer LoadLevel: Linux cluster PBS: Linux workstation:./bsub < run_gsi.ksh./llsubmit run_gsi.ksh./qsub run_gsi.ksh./run_gsi.ksh No difference from other job submitting 23

24 GSI run results: file in run directory Examples of GSI run directory after clean amsuabufr amsubbufr anavinfo berror_stats convinfo diag_amsua_n5_anl.2852 diag_amsua_n5_ges.2852 diag_amsua_n8_anl.2852 diag_amsua_n8_ges.2852 diag_conv_anl.2852 diag_conv_ges.2852 diag_mhs_n8_anl.2852 diag_mhs_n8_ges.2852 errtable fit_p.2852 fit_q.2852 fit_rad.2852 fit_t.2852 fit_w.2852 fort.2 fort.22 fort.23 fort.24 fort.25 fort.26 fort.27 fort.28 fort.29 fort.2 fort.2 fort.22 fort.23 fort.24 fort.25 fort.27 fort.28 fort.29 fort.22 fort.22 gpsrobufr repeat the GSI run:. gsi.exe gsiparm.anl hirs3bufr hirs4bufr l2rwbufr 2. list_run_directory mhsbufr 3. ozinfo 4. pcpbias_out pcpinfo prepbufr prepobs_prep.bufrtable satbias_angle satbias_in satbias_out satinfo stdout stdout.anl.2852 wrfanl.2852 wrf_inout Turn off clean option or copy CRTM coefficients to run directory Get in run directory Replace wrf_inout Run gsi.exe Please reference to GSI Fundamentals (3): Diagnostics and GSI Fundamentals (4): Applications for more details of the contents 24

25 GSI run script: more options The comgsi_v3.3 has more than one sample run script./run/run_gsi.ksh : For regional GSI applications we just introduced in detail The default format of all binary files are in Big_Endian format For regional GSI applications when the GSI compiled with gfortran: Need to pick byte-order based on platform./run/run_gsi_global.ksh For global GSI applications (GFS background) All binary files are in Big_Endian format./util/rtma/run_gsi_rtma.ksh For RTMA GSI applications All binary files are in Big_Endian format 25

26 GSI input files Run GSI with a sample run script Configure GSI with GSI namelist Set up data usage GSI User s Guide (section 3.4) GPS ARW Global Radiance 26

27 GSI Namelist sections SETUP: General control variables. GRIDOPTS: Grid setup variable, including global regional specific namelist variables. BKGERR: Background error related variables. ANBKGERR: Anisotropic background error related variables. JCOPTS: Constraint term in cost function (Jc) STRONGOPTS: Strong dynamic constraint. OBSQC: Observation quality control variables. OBS_INPUT: Input data control variables. SUPEROB_RADAR: Level 2 BUFR file to radar wind superobs. LAG_DATA: Lagrangian data assimilation related variables. HYBRID_ENSEMBLE: Parameters for use with hybrid ensemble option RAPIDREFRESH_CLDSURF: Options for cloud analysis and surface enhancement for RR application 3. CHEM: Chemistry data assimilation 4. SINGLEOB_TEST: Pseudo single observation test setup Complete list is in Advanced User s Guide Appendix: Over 3 namelist variables 27

28 Namelist: outer loop and inner loop miter: number of outer loops of analysis. niter(): max iteration number of inner iterations for the st outer loop. niter(2): max iteration number of inner iterations for the 2nd outer loop. Inner loop will stop when: reaches this maximum number reaches the convergence condition fails to converge. The niter sets for fast test in the sample script, usually set 5 for regional applications &SETUP miter=2,niter()=,niter(2)=, write_diag()=.true.,write_diag(2)=.false.,write_diag(3)=.true., gencode=78,qoption=2,factqmin=.,factqmax=., ndat=87,iguess=-,oneobtest=.false.,retrieval=.false., nhr_assimilation=3,l_foto=.false.,use_pbl=.false., / 28

29 Namelist: output of diagnostic files If following variable is true: write_diag(): write out diagnostic data before st loop (O-B) write_diag(2): write out diagnostic data in between the st and 2nd loop write_diag(3): write out diagnostic data at the end of the 2nd outer loop if the outer loop number is 2, it saves information of O-A Examples in run directory O-A: diag_conv_anl diag_amsua_n8_anl O-B: diag_conv_ges diag_amsua_n8_ges Details on the content of each file and how to read them are in: GSI Fundamentals (3): Diagnostics &SETUP miter=2,niter()=,niter(2)=, write_diag()=.true.,write_diag(2)=.false.,write_diag(3)=.true., gencode=78,qoption=2,factqmin=.,factqmax=.,deltim=$deltim, ndat=77,iguess=-,oneobtest=.false.,retrieval=.false., nhr_assimilation=3,l_foto=.false.,use_pbl=.false., 29 /

30 Namelist: single observation test oneobtest=.true In &SINGLEOB_TEST, set up Innovation Observation error Observation variable Location Section 4.2 for more details Please see Dayrl s talk: Background and Observation Error Estimation and Tuning in 2 GSI summer tutorial for example of single observation application (slides available on-line) &SETUP miter=2,niter()=,niter(2)=, write_diag()=.true.,write_diag(2)=.false.,write_diag(3)=.true., gencode=78,qoption=2,factqmin=.,factqmax=., ndat=77,iguess=-,oneobtest=.false.,retrieval=.false., nhr_assimilation=3,l_foto=.false.,use_pbl=.false., / &SINGLEOB_TEST maginnov=.,magoberr=.8,oneob_type='t', oblat=38.,oblon=279.,obpres=5.,obdattim=${anal_time}, obhourset=., / 3

31 Namelist: the background file Regional: if true, perform a regional GSI run If false, perform a global GSI analysis. wrf_nmm_regional: if true, background comes from WRF NMM. wrf_mass_regional: if true, background comes from WRF ARW. netcdf: only works for performing a regional GSI analysis. if true, WRF files are in NetCDF format otherwise WRF files are in binary format &GRIDOPTS regional=.true.,wrf_nmm_regional=.false,wrf_mass_regional=.true., diagnostic_reg=.false., filled_grid=.false.,half_grid=.true.,netcdf=.true., / 3

32 Set data time window PrepBUFR/BUFR file cut-off time (more details in data process talk) time_window_max: namelist option to set maximum half time window (hours) for all data types. &OBS_INPUT dmesh()=2.,dmesh(2)=6.,dmesh(3)=6.,time_window_max=.5, twindow: set half time window for certain data type (hours). In convinfo file, for conventional observations only Conventional observations within the smaller window of time_window_max and twindow will be kept to further process!otype ps ps t t t type sub iuse twindow numgrp ngroup nmiter gross ermax ermin var_b var_pg ithin rmesh pmesh

33 GSI input files Run GSI with a sample run script Configure GSI with GSI namelist Set up data usage. link observation files 2. set namelist 3. set conventional obs GPS ARW Global 4. Set radiance obs 33 GSI User s Guide (section 4.3) Radiance

34 (): link observation files Link the observation files into run directory Link to the prepbufr data ln -s ${PREPBUFR}./prepbufr Link to the radiance data ln -s ${OBS_ROOT}/gdas.t2z.bamua.tm.bufr_d amsuabufr ln -s ${OBS_ROOT}/gdas.t2z.bhrs4.tm.bufr_d hirs4bufr ln -s ${OBS_ROOT}/gdas.t2z.bmhs.tm.bufr_d mhsbufr GSI Name prepbufr Content Conventional observations, including ps, t, q, pw, uv, spd, dw, sst, from observation platforms such as METAR, sounding, et al. Example file names gdas.t2z.prepbufr amsuabufr amsubbufr radarbufr gpsrobufr ssmirrbufr tmirrbufr sbuvbufr hirs2bufr hirs3bufr hirs4bufr msubufr AMSU-A b radiance (brightness temperatures) from satellites NOAA-5, 6, 7,8, 9 and METOP-A AMSU-B b radiance (brightness temperatures) from satellites NOAA5, 6,7 gdas.t2z.bamua.tm.bufr_d gdas.t2z.bamub.tm.bufr_d ndas.t2z. radwnd. tm2.bufr_d gdas.t2z.gpsro.tm.bufr_d gdas.t2z.spssmi.tm.bufr_d gdas.t2z.sptrmm.tm.bufr_d gdas.t2z.osbuv8.tm.bufr_d gdas.t2z.bhrs2.tm.bufr_d gdas.t2z.bhrs3.tm.bufr_d gdas.t2z.bhrs4.tm.bufr_d gdas.t2z.bmsu.tm.bufr_d Radar radial velocity Level 2.5 data GPS radio occultation observation Precipitation rate observations from SSM/I Precipitation rate observations from TMI SBUV/2 ozone observations from satellite NOAA6, 7, 8, 9 HIRS2 b radiance from satellite NOAA4 HIRS3 b radiance observations from satellite NOAA6, 7 HIRS4 b radiance observation from satellite NOAA 8, 9 and METOP-A MSU observation from satellite NOAA 4 34

35 (2): set namelist Set namelist section &OBS_INPUT Thinning mesh size for each satellite group upper limit on time window for all input data &OBS_INPUT dmesh()=2.,dmesh(2)=6.,dmesh(3)=6.,dmesh(4)=6.,dmesh(5)=2,time_window_max=.5, dfile()='prepbufr, dtype()='ps', dplat()=' ', dsis()='ps', dval()=., dfile(2)='prepbufr' dtype(2)='t', dplat(2)=' ', dsis(2)='t', dval(2)=., dfile(3)='prepbufr', dtype(3)='q', dplat(3)=' ', dsis(3)='q', dval(3)=., dfile(4)='prepbufr', dtype(4)='uv', dplat(4)=' ', dsis(4)='uv', dval(4)=., dfile(27)='msubufr', dtype(27)='msu', dplat(27)='n4', dsis(27)='msu_n4', dval(27)=2., dfile(28)='amsuabufr', dtype(28)='amsua', dplat(28)='n5', dsis(28)='amsua_n5, dval(28)=., dfile(29)='amsuabufr, dtype(29)='amsua', dplat(29)='n6', dsis(29)='amsua_n6, dval(29)=., Observation type Satellite (platform) id (for satellite data) Input observation file name Can be changed if need. dthin()=, dthin(2)=, dthin(3)=, dthin(4)=, dthin(27)=2, dthin(28)=2, dthin(29)=2, Satellite thinning mesh group Weighting factor for super-obs Sensor/instrument/satellite flag from satinfo files The details on data thinning are in GSI Fundamentals (4): Applications 35

36 (3): set conventional observations convinfo control the usage of conventional data (t, q, ps, wind, ) and GPS RO refractivity and bending angle based on data type. User s Guide section 4.3 for more details!otype ps ps ps ps t t t t t t type sub iuse twindow numgrp ngroup nmiter gross ermax ermin var_b iuse=: use data iuse=: do not use data iuse=-: monitor data Gross Check Parameters var_pg ithin rmesh pmesh Data thinning Parameters The last column npred has been deleted to fit the table in a page 36

37 (4): set radiance observations Satinfo (User s Guide section 4.3 for more details) Control the usage of satellite radiance data (AMSU-A, AMSU-B, HIRS3, ) based on platform and channels.!sensor/instr/sat amsua_n5 amsua_n5 amsua_n5 amsua_n5 amsua_n5 amsua_n5 hirs3_n7 chan iuse error error_cld ermax var_b.... var_pg.... icld_det iuse ermax: max error (for QC) =-2: do not use error: variance for each satellite channel =-: monitor if diagnostics produced =: monitor and use in QC only =: use data with complete quality control =2 use data with no airmass bias correction =3 use data with no angle dependent bias correction 37 =4 use data with no bias correction

38 Summary Link the observation files ln -s ${PREPBUFR}./prepbufr ln -s ${OBS_ROOT}/gdas.t2z.bamua.tm.bufr_d amsuabufr Set namelist section &OBS_INPUT &OBS_INPUT dmesh()=2.,dmesh(2)=6.,dmesh(3)=6.,dmesh(4)=6.,dmesh(5)=2,time_window_max=.5, dfile()='prepbufr, dtype()='ps', dplat()=' ', dsis()='ps', dval()=., dfile(2)='prepbufr' dtype(2)='t', dplat(2)=' ', dsis(2)='t', dval(2)=., dfile(3)='prepbufr', dtype(3)='q', dplat(3)=' ', dsis(3)='q', dval(3)=., dfile(28)='amsuabufr', dtype(28)='amsua', dplat(28)='n5', dsis(28)='amsua_n5, dval(28)=., dfile(29)='amsuabufr, dtype(29)='amsua', dplat(29)='n6', dsis(29)='amsua_n6, dval(29)=., dthin()=, dthin(2)=, dthin(3)=, dthin(28)=2, dthin(29)=2, Set info file!otype ps ps ps ps t t t t t t type sub iuse twindow numgrp ngroup nmiter gross ermax ermin var_b !sensor/instr/sat chan.3 iuse amsua_n amsua_n amsua_n amsua_n5 amsua_n5 hirs3_n var_pg ithin rmesh pmesh error error_cld ermax

39 Questions?