4. GSI Diagnostics and Tuning¶
The guidance in this chapter will help users understand how and where to check output from GSI to determine whether a run was successful. Properly checking the GSI output will also provide useful information to diagnose potential errors in the system. This chapter starts with an introduction to the content and structure of the GSI standard output file: (stdout). It continues with the use of a single observation to check the features of the GSI analysis. Then, observation usage control, analysis domain partitioning, fit files, and the optimization process will all be presented from information within the GSI output files (including stdout).
This chapter follows a case example for 2014061700. This case uses a WRF-ARW NetCDF file as the background and analyzes several observations typical for operations, including most conventional observation data, several radiance data sets (AMSU-A, HIRS4, and MHS), and GPSRO data. The case was run on a Linux cluster supercomputer, using four processors. Users can execute this test to reproduce the following results by visiting:
http://www.dtcenter.org/com-GSI/users/tutorial/index.php
Understanding Standard Output (stdout)¶
In Section [sec3.3], we listed the files present in the GSI run directory following a successful GSI analysis and briefly introduced the contents of several important files. Of these, stdout is the most useful because critical information about the GSI analysis can be obtained from the file. From stdout, users can check if the GSI has successfully completed, if optimal iterations look correct, and if the background and analysis fields are reasonable. Understanding the content of this file can also be very helpful for users to find where and why the GSI failed if it crashes.
The structure of stdout follows the typical steps of a meteorological data analysis system:
- Read in all data and prepare analysis:
- Read in configuration (namelist)
- Read in the background
- Read in observations
- Partition domain and data for parallel analysis
- Read in constant fields (fixed files)
- Calculate observation innovations
- Optimal iteration (analysis)
- Save analysis results
In this section, the detailed structure and content of stdout are explained using the online example case: 2014061700. To keep the output concise and make it more readable, most repeated content was deleted (shown with a dotted line). For the same reason, the precision of some numbers has been reduced to avoid line breaks in stdout.
The following indicates the start of the GSI analysis. It shows the date and time that GSI started running:
* . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * .
PROGRAM GSI_ANL HAS BEGUN. COMPILED 1999232.55 ORG: NP23
STARTING DATE-TIME JUL 02,2016 20:36:21.760 184 SAT 2457572
The following shows the content of anavinfo, a list of state and control variables:
gsi_metguess_mod*init_: 2D-MET STATE VARIABLES:
ps
z
gsi_metguess_mod*init_: 3D-MET STATE VARIABLES:
u
v
div
vor
tv
q
oz
cw
gsi_metguess_mod*init_: ALL MET STATE VARIABLES:
u
v
div
vor
tv
q
oz
cw
ps
z
state_vectors*init_anasv: 2D-STATE VARIABLES ps
sst
state_vectors*init_anasv: 3D-STATE VARIABLES u
v tv
tsen q
oz cw
prse
state_vectors*init_anasv: ALL STATE VARIABLES u
v tv
tsen q
oz cw
prse ps
sst
control_vectors*init_anacv: 2D-CONTROL VARIABLES ARE
ps sst
control_vectors*init_anacv: 3D-CONTROL VARIABLES ARE
sf vp
t q
oz cw
control_vectors*init_anacv: MOTLEY CONTROL VARIABLES
stl sti
control_vectors*init_anacv: ALL CONTROL VARIABLES
sf vp
ps t
q oz
sst cw
stl sti
Next is the content of all namelist variables used in this analysis. The
1st part shows 4DVAR setup information. Please note that while this
version of the GSI includes a 4DVAR option, it remains untested. The
general setup for the GSI analysis (3DVAR) is located in the &SETUP
section of the GSI namelist. Please check Appendix B for definitions and
default values of each namelist variable.
GSI_4DVAR: nobs_bins = 1
SETUP_4DVAR: l4dvar= F
SETUP_4DVAR: l4densvar= F
SETUP_4DVAR: winlen= 3.00000000000000
SETUP_4DVAR: winoff= 3.00000000000000
SETUP_4DVAR: hr_obsbin= 3.00000000000000
SETUP_4DVAR: nobs_bins= 1
SETUP_4DVAR: ntlevs_ens= 1
SETUP_4DVAR: nsubwin,nhr_subwin= 1 3
SETUP_4DVAR: lsqrtb= F
SETUP_4DVAR: lbicg= F
SETUP_4DVAR: lcongrad= F
SETUP_4DVAR: lbfgsmin= F
SETUP_4DVAR: ltlint= F
SETUP_4DVAR: ladtest,ladtest_obs,lgrtest= F F F
SETUP_4DVAR: iwrtinc= -1
SETUP_4DVAR: lanczosave= F
SETUP_4DVAR: ltcost= F
SETUP_4DVAR: jsiga= -1
SETUP_4DVAR: nwrvecs= -1
SETUP_4DVAR: iorthomax= 0
SETUP_4DVAR: liauon= F
SETUP_4DVAR: ljc4tlevs= F
SETUP_4DVAR: ibin_anl= 1
in gsimod: use_gfs_stratosphere,nems_nmmb_regional,wrf_nmm_regional= F F F
GSIMOD: ***WARNING*** set l_cloud_analysis=false
INIT_OBSMOD_VARS: reset time window for one or more OBS_INPUT entries to
1.50000000000000
INIT_OBSMOD_VARS: ndat_times,ndat_types,ndat= 1 81
81
INIT_OBSMOD_VARS: nhr_assimilation= 3
GSIMOD: ***WARNING*** reset oberrflg= T
calling gsisub with following input parameters:
&SETUP
GENCODE = 78.0000000000000 ,
FACTQMIN = 0.000000000000000E+000,
FACTQMAX = 0.000000000000000E+000,
CLIP_SUPERSATURATION = F,
FACTV = 1.00000000000000 ,
FACTL = 1.00000000000000 ,
FACTP = 1.00000000000000 ,
FACTG = 1.00000000000000 ,
FACTW10M = 1.00000000000000 ,
FACTHOWV = 1.00000000000000 ,
R_OPTION = F,
DELTIM = 1200.00000000000 ,
DTPHYS = 3600.00000000000 ,
BIASCOR = -1.00000000000000 ,
BCOPTION = 1,
DIURNALBC = 0.000000000000000E+000,
NITER = 0, 2*50, 48*0,
NITER_NO_QC = 51*1000000,
MITER = 2,
QOPTION = 2,
CWOPTION = 0,
NHR_ASSIMILATION = 3,
MIN_OFFSET = 180,
PSEUDO_Q2 = F,
IOUT_ITER = 220,
NPREDP = 6,
/
&GRIDOPTS
...
&BKGERR
...
&ANBKGERR
...
&JCOPTS
...
&STRONGOPTS
...
&OBSQC
...
&SUPEROB_RADAR
...
&LAG_DATA
...
&HYBRID_ENSEMBLE
...
&RAPIDREFRESH_CLDSURF
...
&CHEM
...
This version of GSI attempts to read multiple time level backgrounds for option FGAT (First Guess at Appropriate Time), however we only have provided one time level in this test case. Therefore, there is an error while reading background information:
CONVERT_NETCDF_MASS: problem with flnm1 = wrf_inou1, Status = -1021
We can ignore errors for missing files wrf_inou1, wrf_inou2, …, and wrf_inou9, because we are only running 3DVAR with one background.
Next, the background fields for the analysis are read in, and the
maximum, minimum, and median values of the fields at each vertical level
are displayed. Here, only part of the variables ZNU and T are shown,
with all other variables read by the GSI listed solely as the variable
name in the NetCDF file(rmse_var = T). Maximum and minimum values are
useful for a quick verification that the background fields have been
read successfully. From this section, we also know the time
(iy,m,d,h,m,s) and dimension (nlon,lat,sig_regional) of the
background field.
dh1 = 3
iy,m,d,h,m,s= 2014 6 17 0 0
0
dh1 = 3
rmse_var = SMOIS
ndim1 = 3
ordering = XYZ
staggering = N/A
start_index = 1 1 1 0
end_index = 332 215 4 0
WrfType = 104
ierr = 0
rmse_var = T ndim1 = 3 dh1 = 3
WrfType = 104 ierr = 0
ordering = XYZ staggering = N/A
start_index = 1 1 1 0
end_index = 332 215 50 0
nlon,lat,sig_regional= 332 215 50
rmse_var = P_TOP ndim1= 0
WrfType = 104 WRF_REAL= 104 ierr = 0
ordering = 0 staggering = N/A
start_index = 1 1 1 0
end_index = 332 215 50 0
p_top= 2000.000
...
...
rmse_var = ZNU ndim1= 1
WrfType = 104 WRF_REAL= 104 ierr = 0
ordering = Z staggering = N/A
start_index = 1 1 1 0
end_index = 50 215 50 0
k,znu(k)= 1 0.9990000
k,znu(k)= 2 0.9960001
k,znu(k)= 3 0.9905000
...
...
k,znu(k)= 49 7.1999999E-03
k,znu(k)= 50 2.3500000E-03
rmse_var = ZNW ndim1= 1
...
rmse_var = RDX ndim1= 0
...
rmse_var = RDY ndim1= 0
...
rmse_var = MAPFAC_M ndim1= 2
...
rmse_var = XLAT ndim1= 2
...
rmse_var = XLONG ndim1= 2
...
rmse_var = MUB ndim1= 2
...
rmse_var = MU ndim1= 2
...
rmse_var = PHB ndim1= 3
...
rmse_var = T ndim1= 3
WrfType = 104 WRF_REAL= 104 ierr = 0
ordering = XYZ staggering = N/A
start_index = 1 1 1 0
end_index = 332 215 50 0
k,max,min,mid T= 1 321.5280 270.7682 309.0504
k,max,min,mid T= 2 321.6272 270.9064 309.1002
k,max,min,mid T= 3 321.4596 271.1610 309.1918
k,max,min,mid T= 4 321.2505 271.6038 309.3501
k,max,min,mid T= 5 321.6713 272.2668 309.4191
...
...
k,max,min,mid T= 48 632.2557 567.8249 596.6701
k,max,min,mid T= 49 659.2219 604.4777 630.4330
k,max,min,mid T= 50 689.7565 646.8995 668.5146
rmse_var = QVAPOR ndim1= 3
...
rmse_var = U ndim1= 3
...
rmse_var = V ndim1= 3
...
rmse_var = XLAND ndim1= 2
...
rmse_var = SEAICE ndim1= 2
...
rmse_var = SST ndim1= 2
...
rmse_var = IVGTYP ndim1= 2
...
rmse_var = ISLTYP ndim1= 2
...
rmse_var = VEGFRA ndim1= 2
...
rmse_var = SNOW ndim1= 2
...
rmse_var = U10 ndim1= 2
...
rmse_var = V10 ndim1= 2
...
rmse_var = SMOIS ndim1= 3
...
rmse_var = TSLB ndim1= 3
...
rmse_var = TSK ndim1= 2
...
rmse_var = Q2 ndim1= 2
...
rmse_var = QCLOUD ndim1= 3
...
rmse_var = QRAIN ndim1= 3
...
rmse_var = QSNOW ndim1= 3
...
rmse_var = QICE ndim1= 3
...
rmse_var = QGRAUP ndim1= 3
...
rmse_var = QNRAIN ndim1= 3
...
rmse_var = RAD_TTEN_DFI ndim1= 3
...
For some variables, the following NETCDF error information might show up when they are not in the background fields. These errors dont affect the GSI run so you can ignore them.
rmse_var = QSNOW ndim1= 3
WrfType = 104 WRF_REAL= 104 ierr = -1021
ordering = XYZ staggering = N/A
start_index = 1 1 1 0
end_index = 332 215 50 0
NetCDF error: NetCDF: Variable not found
NetCDF error: NetCDF: Variable not found
NetCDF error in wrf_io.F90, line 2842 Varname QSNOW
NetCDF error in wrf_io.F90, line 2842 Varname QSNOW
Again, some error information on missing background files shows up. Ignore if you are not doing FGAT.
CONVERT_NETCDF_MASS: problem with flnm1 = wrf_inou4, Status = -1021
Following this is information on the byte order of the binary background files. Since we used a NetCDF file, there is no need to be concerned with byte order. When using a binary format background, byte-order can be a problem. Beginning with the release version v3.2, GSI can automatically check the background byte-order and read it in the right order:
in convert_regional_guess, for wrf arw binary input, byte_swap= F
Information on setting the grid related variables, and the beginning and ending indices for thread one:
INIT_GRID_VARS: number of threads 1
INIT_GRID_VARS: for thread 1 jtstart,jtstop = 1
168
Information on the initial pointer location for each variable in the Jacobian for the use of satellite radiance data:
Vars in Rad-Jacobian (dims)
--------------------------
sst 0
u 1
v 2
tv 3
q 53
oz 103
Starting subroutine gsisub (major GSI control subroutine) and displaying the analysis and background file time (they should be the same):
[000]gsisub(): : starting ...
READ_wrf_mass_FILES: analysis date,minutes 2014 6
17 0 0 19175040
READ_wrf_mass_FILES: sigma guess file, nming2 0.000000000000000E+000
2014 6 17 0 0 19175040
READ_wrf_mass_FILES: sigma fcst files used in analysis : 3
3.00000000000000 1
READ_wrf_mass_FILES: surface fcst files used in analysis: 3
3.00000000000000 1
GESINFO: Guess date is 0 6 17 2014
0.000000000000000E+000
GESINFO: Analysis date is 2014 6 17 0
0 2014061700 3.00000000000000
using restart file date = 2014 6 17 0
Read in radar location information and generate superobs for radar level-II radial velocity. This case didnt have radar level-II velocity data linked, therefore there is warning about when opening the file, but this will not impact the rest of the GSI analysis.
RADAR_BUFR_READ_ALL: analysis time is 2014 6 17
0
RADAR_BUFR_READ_ALL: NO RADARS KEPT IN radar_bufr_read_all,
continue without level 2 data
Read in information from fix file scaninfo (see table 3.2) and pcpinfo (see table 3.2).
***WARNING file scaninfo not found, use default
CREATE_PCP_RANDOM: iseed= 2014061700
PCPINFO_READ: no pcpbias file. set predxp=0.0
Read in and show the content of the conventional observation information file (convinfo; see Section 1.3 for details). Here is the part of the stdout file showing information from convinfo:
READ_CONVINFO: tcp 112 0 1 3.00000 0 0 0 75.0000 5.00000 1.00000 75.0000
0.00000 0 0.00000 0.00000 0 0.00000 0.00000 2
READ_CONVINFO: ps 120 0 1 3.00000 0 0 0 4.00000 3.00000 1.00000 4.00000
0.300000E-03 0 0.00000 0.00000 0 0.00000 0.00000 2
...
READ_CONVINFO: t 120 0 1 3.00000 0 0 0 8.00000 5.60000 1.30000 8.00000
0.100000E-05 0 0.00000 0.00000 0 0.00000 0.00000 2
READ_CONVINFO: t 126 0 -1 3.00000 0 0 0 8.00000 5.60000 1.30000 8.00000
0.100000E-02 0 0.00000 0.00000 0 0.00000 0.00000 2
...
READ_CONVINFO: gps 729 0 -1 3.00000 0 0 0 10.0000 10.0000 1.00000 10.0000
0.00000 0 0.00000 0.00000 0 0.00000 0.00000 2
READ_CONVINFO: gps 44 0 -1 3.00000 0 0 0 10.0000 10.0000 1.00000 10.0000
0.00000 0 0.00000 0.00000 0 0.00000 0.00000 2
Starting subroutine glbsoi with information on reading in background fields from the intermediate binary file sigf03 and partitioning the whole 2D field into subdomains for parallel analysis:
glbsoi: starting ...
gsi_metguess_mod*create_: alloc() for met-guess done
guess_grids*create_chemges_grids: trouble getting number of chem/gases
at 0 in read_wrf_mass_guess
at 0.1 in read_wrf_mass_guess
at 1 in read_wrf_mass_guess, lm = 50
at 1 in read_wrf_mass_guess, num_mass_fields= 215
at 1 in read_wrf_mass_guess, nfldsig = 1
at 1 in read_wrf_mass_guess, num_all_fields= 215
at 1 in read_wrf_mass_guess, npe = 4
at 1 in read_wrf_mass_guess, num_loc_groups= 53
at 1 in read_wrf_mass_guess, num_all_pad = 216
at 1 in read_wrf_mass_guess, num_loc_groups= 54
READ_WRF_MASS_GUESS: open lendian_in= 15 to file=sigf03
READ_WRF_MASS_GUESS: open lendian_in= 15 to file=sigf03
in read_wrf_mass_guess, num_doubtful_sfct_all = 0
in read_wrf_mass_guess, num_doubtful_sfct_all = 0
Show observation observer as successfully initialized and inquire about the control vectors (space for analysis variables).
observer_init: successfully initialized
control_vectors: length= 5613648
control_vectors: currently allocated= 0
control_vectors: maximum allocated= 0
control_vectors: number of allocates= 0
control_vectors: number of deallocates= 0
control_vectors: Estimated max memory used= 0.0 Mb
Show the source of observation error used in the analysis (details see Section 1.7.1):
CONVERR: using observation errors from user provided table
The following information is related to the observation ingest processes, which is distributed over all the processors with each processor reading in at least one observation type. To speed up the reading process, some of the large datasets will use more than one (ntasks) processor for the ingest process.
Before reading in data from BUFR files, GSI checks the file status to insure the observation time matches the analysis time and whether the namelist option offtime_data is set (can be used to turn off the time consistency check between observation and analysis time). This step also checks for consistency between the satellite radiance data types in the BUFR files and the usage setups in the satinfo files. The following shows stdout information from this step:
read_obs_check: bufr file date is 2014061700 prepbufr ps
read_obs_check: bufr file uv not available satwndbufr
read_obs_check: bufr file rw not available radarbufr
read_obs_check: bufr file pcp_tmi trmm not available tmirrbufr
read_obs_check: bufr file hirs3 n17 not available hirs3bufr
read_obs_check: bufr file goes_img g11 not available gimgrbufr
read_obs_check: bufr file date is 2014061700 prepbufr q
read_obs_check: bufr file date is 2014061700 prepbufr t
read_obs_check: bufr file date is 2014061700 amsuabufr amsua n18
read_obs_check: bufr file amsua n18 not available amsuabufrears
...
...
read_obs_check: bufr file sndrd3 g15 not available gsnd1bufr
read_obs_check: bufr file omi aura not available omibufr
read_obs_check: bufr file seviri m09 not available seviribufr
read_obs_check: bufr file atms npp not available atmsbufr
read_obs_check: bufr file date is 2014061700 prepbufr sst
...
...
read_obs_check: bufr file seviri m08 not available seviribufr
read_obs_check: bufr file gome metop-b not available gomebufr
read_obs_check: bufr file uv not available oscatbufr
data type mta_cld not used in info file -- do not read file
prepbufr
data type gos_ctp not used in info file -- do not read file
prepbufr
data type rad_ref not used in info file -- do not read file
refInGSI
data type lghtn not used in info file -- do not read file
lghtInGSI
data type larccld not used in info file -- do not read file
larcInGSI
The list of observation types that were read in and processors used to read them:
number of extra processors 1
READ_OBS: read 33 mhs mhs_n18 using ntasks= 2 0 131 0
READ_OBS: read 34 mhs mhs_n19 using ntasks= 2 2 153 0
READ_OBS: read 35 mhs mhs_metop-a using ntasks= 2 0 563 0
READ_OBS: read 36 mhs mhs_metop-b using ntasks= 2 2 2 0
READ_OBS: read 1 ps ps using ntasks= 1 0 0 0
READ_OBS: read 2 t t using ntasks= 1 1 0 0
READ_OBS: read 3 q q using ntasks= 1 2 0 0
READ_OBS: read 4 pw pw using ntasks= 1 3 839 0
READ_OBS: read 6 uv uv using ntasks= 1 0 0 0
READ_OBS: read 10 sst sst using ntasks= 1 1 504 0
READ_OBS: read 11 gps_ref gps using ntasks= 1 2 0 0
READ_OBS: read 19 hirs4 hirs4_metop-a using ntasks= 2 3 277 0
READ_OBS: read 21 hirs4 hirs4_n19 using ntasks= 1 0 75 0
READ_OBS: read 22 hirs4 hirs4_metop-b using ntasks= 1 1 2 0
READ_OBS: read 26 amsua amsua_n15 using ntasks= 1 2 27 0
READ_OBS: read 27 amsua amsua_n18 using ntasks= 1 3 45 0
READ_OBS: read 28 amsua amsua_n19 using ntasks= 1 0 47 0
READ_OBS: read 29 amsua amsua_metop-a using ntasks= 1 1 124 0
READ_OBS: read 30 amsua amsua_metop-b using ntasks= 1 2 2 0
Display basic statistics for full horizontal surface fields (If radiance BUFR files are not linked, this section will not be in the stdout file):
GETSFC: enter with nlat_sfc,nlon_sfc= 0 0 and nlat,nlon=
215 332
GETSFC: set nlat_sfc,nlon_sfc= 215 332
================================================================================
Status Var Mean Min Max
sfcges2 FC10 1.000000000000E+00 1.000000000000E+00 1.000000000000E+00
sfcges2 SNOW 8.137817211798E-02 0.000000000000E+00 9.510296630859E+01
sfcges2 VFRC 1.701514588514E-01 0.000000000000E+00 9.899999499321E-01
sfcges2 SRGH 5.003234230527E-02 5.000000074506E-02 5.000000074506E-02
sfcges2 STMP 2.936729335948E+02 2.643117675781E+02 3.229424743652E+02
sfcges2 SMST 7.664003944557E-01 6.047149747610E-02 1.000000000000E+00
sfcges2 SST 2.942266741384E+02 2.688000183105E+02 3.240092468262E+02
sfcges2 VTYP 1.463281031101E+01 1.000000000000E+00 2.400000000000E+01
sfcges2 ISLI 3.405295601009E-01 0.000000000000E+00 2.000000000000E+00
sfcges2 STYP 1.137135051835E+01 1.000000000000E+00 1.600000000000E+01
================================================================================
Loop over all data files to read in observations, also read in rejection list for surface observations and show GPS observations outside the time window:
READ_BUFRTOVS : file=mhsbufr type=mhs sis=mhs_n18 nread= 248485 ithin= 2
rmesh= 60.000000 isfcalc= 0 ndata= 26765 ntask= 2
READ_BUFRTOVS : file=mhsbufr type=mhs sis=mhs_n19 nread= 60900 ithin= 2
rmesh= 60.000000 isfcalc= 0 ndata= 6725 ntask= 2
READ_BUFRTOVS : file=mhsbufr type=mhs sis=mhs_metop-a nread= 142555 ithin= 2
rmesh= 60.000000 isfcalc= 0 ndata= 15145 ntask= 2
READ_BUFRTOVS : file=mhsbufr type=mhs sis=mhs_metop-b nread= 113590 ithin= 2
rmesh= 60.000000 isfcalc= 0 ndata= 12185 ntask= 2
READ_PREPBUFR: messages/reports = 142 / 20925 ntread =
1
new vad flag:: F
READ_PREPBUFR : file=prepbufr type=pw sis=pw nread= 252 ithin= 0 rmesh=
120.000000 isfcalc= 0 ndata= 252 ntask= 1
READ_PREPBUFR: messages/reports = 682 / 67083 ntread =
1
READ_PREPBUFR: time offset is 3.00000000000000 hours.
new vad flag:: F
READ_PREPBUFR: messages/reports = 682 / 67083 ntread =
1
mesonetuselist: listexist,nprov= F 0
w_rejectlist: wlistexist,nwrjs= F 0
t_rejectlist: tlistexist,ntrjs= F 0
...
READ_PREPBUFR : file=prepbufr type=ps sis=ps nread= 23868 ithin= 0 rmesh=
120.000000 isfcalc= 0 ndata= 15819 ntask= 1
READ_PREPBUFR : file=prepbufr type=t sis=t nread= 26296 ithin= 0 rmesh=
120.000000 isfcalc= 0 ndata= 25686 ntask= 1
READ_PREPBUFR : file=prepbufr type=sst sis=sst nread= 0 ithin= 0 rmesh=
120.000000 isfcalc= 0 ndata= 0 ntask= 1
READ_PREPBUFR : file=prepbufr type=q sis=q nread= 24461 ithin= 0 rmesh=
120.000000 isfcalc= 0 ndata= 20989 ntask= 1
..
READ_BUFRTOVS : file=hirs4bufr type=hirs4 sis=hirs4_n19 nread= 55613 ithin= 2 rmesh=
60.000000 isfcalc= 0 ndata= 23408 ntask= 1
READ_BUFRTOVS : file=amsuabufr type=amsua sis=amsua_n19 nread= 20370 ithin= 2
rmesh= 60.000000 isfcalc= 0 ndata= 16912 ntask= 1
Using the above output information, many details on the observations can
be obtained. For example, the last line indicates that subroutine
READ_BUFRTOVS was called to read in NOAA-19 AMSU-A (sis=amsua_n19)
data from the BUFR file amsuabufr (file=amsuabufr). Furthermore,
there are 20370 observations in this file (nread=20370) and 16912 in
the analysis domain and within the time window (ndata=16912). The
data was thinned on a 60 km coarse grid (rmesh=60.000000).
The next step partitions observations into subdomains. The observation distribution is summarized below by listing the number of observations for each variable per subdomain (see Section 1.4 for more information):
OBS_PARA: ps 1429 3190 4655 6774
OBS_PARA: t 2564 5200 7057 11128
OBS_PARA: q 2346 4626 6148 8128
OBS_PARA: pw 65 80 63 49
OBS_PARA: uv 3358 6453 8091 11998
OBS_PARA: gps_ref 1799 1368 2664 3520
OBS_PARA: hirs4 metop-a 0 0 1146 1661
OBS_PARA: hirs4 n19 213 1020 0 0
OBS_PARA: hirs4 metop-b 0 0 85 555
OBS_PARA: amsua n15 1458 2026 830 234
OBS_PARA: amsua n18 2223 2318 108 0
OBS_PARA: amsua n19 176 960 0 0
OBS_PARA: amsua metop-a 0 0 1077 1559
OBS_PARA: amsua metop-b 0 0 265 1829
OBS_PARA: mhs n18 2550 2695 160 0
OBS_PARA: mhs n19 246 1103 0 0
OBS_PARA: mhs metop-a 0 0 1237 1809
OBS_PARA: mhs metop-b 0 0 321 2128
Information on ingesting background error statistics:
m_berror_stats_reg::berror_read_bal_reg(PREBAL_REG): get balance variables"
berror_stats". mype,nsigstat,nlatstat = 0 60 93
m_berror_stats_reg::berror_read_wgt_reg(PREWGT_REG): read error amplitudes "
berror_stats". mype,nsigstat,nlatstat = 0 60 93
Assigned default statistics to variable oz
Assigned default statistics to variable cw
From this point forward in the stdout file, the output shows many repeated entries. This is because the information is written from inside the outer loop. Typically the outer loop is run twice.
For each outer loop, the work begins with the calculation of the observation innovation. This calculation is done by the subroutine setuprhsall, which sets up the right hand side (rhs) of the analysis equation. This information is contained within the stdout file, which is shown in the following sections:
Start the first outer analysis loop:
GLBSOI: jiter,jiterstart,jiterlast,jiterend= 1 1
2 1
Calculate observation innovation for each data type in the first outer loop:
SETUPALL:,obstype,isis,nreal,nchanl= ps ps 20 0
SETUPALL:,obstype,isis,nreal,nchanl= t t 25 0
SETUPALL:,obstype,isis,nreal,nchanl= q q 26 0
SETUPALL:,obstype,isis,nreal,nchanl= pw pw 20 0
SETUPALL:,obstype,isis,nreal,nchanl= uv uv 25 0
SETUPALL:,obstype,isis,nreal,nchanl= gps_ref gps 16 0
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_n19 33 19
0 setuprad: passive obs 21 hirs4_n19
crtm_interface*init_crtm: crtm_init() on path "./"
SpcCoeff_ReadFile(Binary)(INFORMATION) : FILE: ./hirs4_n19.SpcCoeff.bin; ^M
SpcCoeff RELEASE.VERSION: 8.03^M
N_CHANNELS=19
Read_ODPS_Binary(INFORMATION) : FILE: ./hirs4_n19.TauCoeff.bin; ^M
ODPS RELEASE.VERSION: 2.01 N_LAYERS=100 N_COMPONENTS=5 N_ABSORBERS=3 N_CHANNELS=19 N_COEFFS=82000
SEcategory_ReadFile(INFORMATION) : FILE: ./NPOESS.IRland.EmisCoeff.bin; ^M
SEcategory RELEASE.VERSION: 3.01^M
CLASSIFICATION: NPOESS, N_FREQUENCIES=20 N_SURFACE_TYPES=20
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_metop-a 33 19
crtm_interface*init_crtm: crtm_init() on path "./"
...
...
In the above section, when computing the radiance observation innovation, information on reading in CRTM coefficients follows SETUPALL information. In the stdout file, only information related to available radiance data are printed. The complete innovation information can be found in the diagnostic files for each observation (for details see Appendix A.2):
...
...
FitCoeff_ReadFile(INFORMATION) : FILE: ./FASTEM6.MWwater.EmisCoeff.bin; ^M
FitCoeff RELEASE.VERSION : 1.6; DIMENSIONS= 3, 6, 2
MWwaterCoeff_ReadFile(INFORMATION) : FILE: ./FASTEM6.MWwater.EmisCoeff.bin; ^M
MWwaterCoeff RELEASE.VERSION: 1.6
SETUPRAD: write header record for mhs_n19 12 30
8 0 0 22 4 30303
to file pe0000.mhs_n19_01 2014061700
The inner iteration of the first outer loop is discussed in the example below. In this example, the maximum number of iterations is 50.
Print cost function values for each inner iteration (see section 1.6 for more details):
GLBSOI: START pcgsoi jiter= 1
pcgsoi: gnorm(1:2),b= 2.767403469782257162E+03 2.767403469782257162E+03 0.000000000000000000E+00
Begin J table inner/outer loop 0 1
J term J
surface pressure 5.7012207042385944E+03
temperature 6.4242087278840627E+03
wind 1.6782607330525603E+04
moisture 3.5878183830232451E+03
gps 7.8814883785376896E+03
radiance 3.4334884315701471E+04
-----------------------------------------------------
J Global 7.4712227839910673E+04
End Jo table inner/outer loop 0 1
Initial cost function = 7.471222783991067263E+04
Initial gradient norm = 5.260611627731377382E+01
cost,grad,step,b,step? = 1 0 7.471222783991067263E+04 5.260611627731377382E+01 1.717817994849075269E+00 0.000000000000000000E+00 good
pcgsoi: gnorm(1:2),b= 1.754232612149755596E+03 1.754232612149755141E+03 6.338911659627933792E-01
cost,grad,step,b,step? = 1 1 6.995833236051093263E+04 4.188356016565158058E+01 4.106937422100393142E+00 6.338911659627933792E-01 good
pcgsoi: gnorm(1:2),b= 1.216588309725912268E+03 1.216588309725912268E+03 6.935159575188962755E-01
cost,grad,step,b,step? = 1 2 6.275380879860417917E+04 3.487962599750622417E+01 2.174716042085542700E+00 6.935159575188962755E-01 good
pcgsoi: gnorm(1:2),b= 1.156766558917323891E+03 1.156766558917324346E+03 9.508282708864222998E-01
cost,grad,step,b,step? = 1 3 6.010807468482949480E+04 3.401127105706759579E+01 2.916832102067935306E+00 9.508282708864222998E-01 good
pcgsoi: gnorm(1:2),b= 6.945724726018979709E+02 6.945724726018985393E+02 6.004430775142600707E-01
...
...
cost,grad,step,b,step? = 1 49 4.142785387197384262E+04 1.680503228207865574E+00 2.338314294416948602E+00 1.076393393015242506E+00 good
pcgsoi: gnorm(1:2),b= 1.980029522220628557E+00 1.980029522219752591E+00 7.011209809087933786E-01
cost,grad,step,b,step? = 1 50 4.142125025938593171E+04 1.407135218172236746E+00 5.458012252072157899E+00 7.011209809087933786E-01 good
update_guess: successfully complete
At the end of the 1st outer loop, print some diagnostics about the analysis increments as well as information on the guess fields after adding the analysis increments to the background:
================================================================================
Status Var Mean Min Max
analysis U 3.027810174754E+00 -4.616646796505E+01 6.874148210358E+01
analysis V -2.783966384966E-02 -6.673607446514E+01 6.206906140999E+01
analysis TV 2.466648731614E+02 1.909849532362E+02 3.159577451606E+02
analysis Q 2.789588139811E-03 1.000000000000E-07 2.260955460480E-02
analysis TSEN 2.461750146062E+02 1.909846857229E+02 3.153599236074E+02
analysis OZ 1.000000000007E-15 1.000000000000E-15 1.000000000000E-15
analysis CW 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis DIV 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis VOR 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis PRSL 4.154470108570E+01 2.152367800892E+00 1.028272918117E+02
analysis PS 9.910751025141E+01 6.684714489139E+01 1.029767184368E+02
analysis SST 2.942451749464E+02 2.688000183105E+02 3.240092468262E+02
analysis radb 6.939468963354E-02 -1.373884240000E+02 1.230549030000E+02
analysis pcpb 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis aftb 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
================================================================================
increment u 6.699567621553E-04 -1.360159370531E+01 7.008598936474E+00
increment v 1.598741317770E-03 -9.984198525101E+00 8.688133965521E+00
increment tv -5.436012894801E-04 -2.969908758852E+00 4.753382796517E+00
increment tsen 4.740094380224E-04 -2.966625119631E+00 4.955252532399E+00
increment q -5.666454694731E-06 -4.783507458114E-03 4.607810408495E-03
increment oz 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
increment cw 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
increment prse 4.684195247526E-03 -8.941125614721E-02 1.065972973935E-01
increment ps 1.145022012350E-02 -8.941125614721E-02 1.065972973935E-01
increment sst 2.337322205329E-02 -5.017646098146E-01 1.017498439243E+00
Start the second outer loop.
GLBSOI: jiter,jiterstart,jiterlast,jiterend= 2 1
2 1
Calculate observation innovations for each data type in the second outer loop:
SETUPALL:,obstype,isis,nreal,nchanl= ps ps 20 0
SETUPALL:,obstype,isis,nreal,nchanl= t t 25 0
SETUPALL:,obstype,isis,nreal,nchanl= q q 26 0
...
When calculating the radiance data innovation, there is no need to read in CRTM coefficients again because they were already read in during the first outer loop:
SETUPALL:,obstype,isis,nreal,nchanl= ps ps 20 0
SETUPALL:,obstype,isis,nreal,nchanl= t t 25 0
SETUPALL:,obstype,isis,nreal,nchanl= q q 26 0
SETUPALL:,obstype,isis,nreal,nchanl= pw pw 20 0
SETUPALL:,obstype,isis,nreal,nchanl= uv uv 25 0
SETUPALL:,obstype,isis,nreal,nchanl= gps_ref gps 16 0
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_n19 33 19
0 setuprad: passive obs 21 hirs4_n19
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_metop-a 33 19
SETUPALL:,obstype,isis,nreal,nchanl= amsua amsua_n15 33 15
SETUPALL:,obstype,isis,nreal,nchanl= amsua amsua_n18 33 15
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_metop-b 33 19
SETUPALL:,obstype,isis,nreal,nchanl= amsua amsua_metop-a 33 15
SETUPALL:,obstype,isis,nreal,nchanl= amsua amsua_n19 33 15
SETUPALL:,obstype,isis,nreal,nchanl= amsua amsua_metop-b 33 15
SETUPALL:,obstype,isis,nreal,nchanl= mhs mhs_n18 33 5
SETUPALL:,obstype,isis,nreal,nchanl= mhs mhs_metop-a 33 5
SETUPALL:,obstype,isis,nreal,nchanl= mhs mhs_metop-b 33 5
SETUPALL:,obstype,isis,nreal,nchanl= mhs mhs_n19 33 5
The output from the inner iterations in the second outer loop is shown below. In this example, the maximum number of iterations is 50.
Print cost function values for each inner iteration (see section 1.6 for more details):
GLBSOI: START pcgsoi jiter= 2
pcgsoi: gnorm(1:2),b= 9.125529304049867960E+02 9.125529304049867960E+02 0.000000000000000000E+00
Begin J table inner/outer loop 0 2
J term J
background 5.1520678195203345E+03
surface pressure 4.1180289830866377E+03
temperature 4.3079774551559522E+03
wind 1.0714401927194920E+04
moisture 1.3696062777723114E+03
gps 3.1175680587783127E+03
radiance 2.4255186183427002E+04
-----------------------------------------------------
J Global 5.3034836704935471E+04
End Jo table inner/outer loop 0 2
Initial cost function = 5.303483670493547106E+04
Initial gradient norm = 3.020849103157896565E+01
cost,grad,step,b,step? = 2 0 5.303483670493547106E+04 3.020849103157896565E+01 1.417696886759607366E+00 0.000000000000000000E+00 good
pcgsoi: gnorm(1:2),b= 3.752399307247351885E+02 3.752399307247339380E+02 4.111979899710630493E-01
cost,grad,step,b,step? = 2 1 5.174111325649695937E+04 1.937111072511680021E+01 5.251908998246061167E+00 4.111979899710630493E-01 good
pcgsoi: gnorm(1:2),b= 2.403892651142143109E+02 2.403892651142157320E+02 6.406281566301591512E-01
cost,grad,step,b,step? = 2 2 4.977038728782249382E+04 1.550449177219989672E+01 3.447718424869901543E+00 6.406281566301591512E-01 good
pcgsoi: gnorm(1:2),b= 3.513995903418547186E+02 3.513995903418544913E+02 1.461794020522907633E+00
cost,grad,step,b,step? = 2 3 4.894159278934728354E+04 1.874565523906419173E+01 1.888694950411589302E+00 1.461794020522907633E+00 good
...
...
pcgsoi: gnorm(1:2),b= 4.240269047847287087E-01 4.240269047846388362E-01 8.928029496575288215E-01
cost,grad,step,b,step? = 2 49 4.549093632176067331E+04 6.511734828636134287E-01 3.659140330083644699E+00 8.928029496575288215E-01 good
pcgsoi: gnorm(1:2),b= 3.162643690270069974E-01 3.162643690267226138E-01 7.458592024656659492E-01
cost,grad,step,b,step? = 2 50 4.548938474781232799E+04 5.623738694383008108E-01 3.375651104905403432E+00 7.458592024656659492E-01 good
update_guess: successfully complete
Diagnostics of the analysis results after adding the analysis increment to the guess, as well as diagnostics about the analysis increments:
================================================================================
Status Var Mean Min Max
analysis U 3.031191508676E+00 -4.617266089077E+01 6.889816661368E+01
analysis V -2.943556460524E-02 -6.653467898266E+01 6.166408410383E+01
analysis TV 2.467118072154E+02 1.909576489123E+02 3.159594232825E+02
analysis Q 2.792097480151E-03 1.000000000000E-07 2.263794691793E-02
analysis TSEN 2.462214978004E+02 1.909573814527E+02 3.153808896427E+02
analysis OZ 1.000000000007E-15 1.000000000000E-15 1.000000000000E-15
analysis CW 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis DIV 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis VOR 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis PRSL 4.154936446270E+01 2.152390788296E+00 1.028757430184E+02
analysis PS 9.910817481416E+01 6.684789651716E+01 1.029802303066E+02
analysis SST 2.942451749464E+02 2.688000183105E+02 3.240092468262E+02
analysis radb 6.938254104671E-02 -1.373884240000E+02 1.230549030000E+02
analysis pcpb 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
analysis aftb 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
================================================================================
increment u 3.381333921427E-03 -1.617060550844E+00 3.067216672606E+00
increment v -1.595900755559E-03 -2.932441477540E+00 1.626847973396E+00
increment tv 4.693405394733E-02 -9.818098783002E-01 2.215784626624E+00
increment tsen 4.648264962860E-02 -9.818069362856E-01 2.215777805989E+00
increment q 2.508348182621E-06 -2.410389935149E-03 1.663695364258E-03
increment oz 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
increment cw 0.000000000000E+00 0.000000000000E+00 0.000000000000E+00
increment prse 2.718674106775E-04 -5.539352084836E-02 4.239997318097E-02
increment ps 6.645627545749E-04 -5.539352084836E-02 4.239997318097E-02
increment sst -2.077555444927E-03 -3.656869699698E-01 7.433384932582E-01
Because the outer loop is set to two, the completion of the 2nd
outer loop marks the end of the analysis. The next step is to save the
analysis results. Again, only a portion of variable “T” is shown and all
other variables are listed according to variable name in the NetCDF file
(rmse_var = T). The maximum and minimum values are useful
information for a quick sanity check of the analysis:
at 2 in wrwrfmassa
update sigf03
at 3 in wrwrfmassa
at 6 in wrwrfmassa
at 10.11 in wrwrfmassa,max,min(temp1)= 2.1931874E-02 1.3461057E-03
at 10.12 in wrwrfmassa,max,min(tempa)= 0.0000000E+00 0.0000000E+00
at 10.13 in wrwrfmassa,max,min(tempa)= 0.0000000E+00 -2.1931874E-02
at 10.14 in wrwrfmassa,max,min(temp1)= 0.0000000E+00 0.0000000E+00
iy,m,d,h,m,s= 2014 6 17 0 0
0
nlon,lat,sig_regional= 332 215 50
rmse_var=P_TOP
ordering=0
WrfType,WRF_REAL= 104 104
ndim1= 0
staggering= N/A
start_index= 1 1 1 0
end_index1= 332 215 50 0
p_top= 2000.000
rmse_var=MUB
ordering=XY
WrfType,WRF_REAL= 104 104
ndim1= 2
staggering= N/A
start_index= 1 1 1 0
end_index1= 332 215 50 0
max,min MUB= 98672.59 63425.52
max,min psfc= 102799.7 66795.70
max,min MU= 2799.734 -1187.844
rmse_var=MU
ordering=XY
WrfType,WRF_REAL= 104 104
ndim1= 2
staggering= N/A
start_index= 1 1 1 0
end_index1= 332 215 50 0
k,max,min,mid T= 1 321.6379 270.7839 309.3401
k,max,min,mid T= 2 321.7433 270.9335 309.3846
k,max,min,mid T= 3 321.4780 271.1794 309.3658
...
k,max,min,mid T= 49 662.7022 604.6494 637.0358
k,max,min,mid T= 50 693.4219 647.1161 675.5701
rmse_var=T
...
rmse_var=QVAPOR
...
rmse_var=U
...
rmse_var=V
...
rmse_var=SEAICE
...
rmse_var=SST
...
rmse_var=TSK
...
rmse_var=Q2
...
After completion of the analysis, the subroutine “setuprhsall” is called
again if write_diag(3)=.true., to calculate analysis and O-A
information (this marks the third time this information is presented):
SETUPALL:,obstype,isis,nreal,nchanl= ps ps 20 0
SETUPALL:,obstype,isis,nreal,nchanl= t t 25 0
SETUPALL:,obstype,isis,nreal,nchanl= q q 26 0
SETUPALL:,obstype,isis,nreal,nchanl= pw pw 20 0
SETUPALL:,obstype,isis,nreal,nchanl= uv uv 25 0
SETUPALL:,obstype,isis,nreal,nchanl= gps_ref gps 16 0
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_n19 33 19
0 setuprad: passive obs 21 hirs4_n19
SETUPRAD: write header record for hirs4_n19 12 30
8 0 0 22 4 30303
to file pe0000.hirs4_n19_03 2014061700
SETUPALL:,obstype,isis,nreal,nchanl= hirs4 hirs4_metop-a 33 19
SETUPRAD: write header record for hirs4_metop-a 12 30
8 0 0 22 4 30303
Deallocate the data arrays and finalize the GSI run:
gsi_metguess_mod*destroy_: dealloc() for met-guess done
observer_final: successfully finalized
glbsoi: complete
[000]gsisub(): : complete.
The end of the GSI analysis (reaching this point does not necessarily guarantee a successful analysis), which shows the date and time when GSI finished and some additional resource statistics:
ENDING DATE-TIME JUL 02,2016 20:43:40.422 184 SAT 2457572
PROGRAM GSI_ANL HAS ENDED.
* . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * .
*****************RESOURCE STATISTICS*******************************
The total amount of wall time = 438.663534
The total amount of time in user mode = 427.578998
The total amount of time in sys mode = 9.457562
The maximum resident set size (KB) = 2020132
Number of page faults without I/O activity = 312762
Number of page faults with I/O activity = 0
Number of times filesystem performed INPUT = 0
Number of times filesystem performed OUTPUT = 0
Number of Voluntary Context Switches = 7641
Number of InVoluntary Context Switches = 851
*****************END OF RESOURCE STATISTICS*************************
Different GSI applications may write out slightly different stdout file information but the major flow and information are the same. A good knowledge of the stdout file gives users a clear picture of how GSI runs and the key information provided during a GSI run like data distribution and inner iterations.
Single Observation Test¶
A single observation test is a GSI where only one (pseudo) observation is assimilated from a specific time and location within the analysis domain. By examining the analysis increments from a single observation test, one can visualize the important features of the analysis, such as the ratio of background error and observation error variance and the pattern of the background error covariance. Therefore, the single observation test is the first thing that users should run after successfully installing the GSI.
Setup a Single Observation Test¶
To perform the single observation test with the GSI, the following GSI namelist variables need to be set, which should be done through editing the script ush/comgsi_run_regional.ksh and ush/comgsi_namelist.sh:
In ush/comgsi_run_regional.ksh, turn on the single observation test:
if_oneob=Yes # Yes, or, No -- case sensitive !
In ush/comgsi_namelist.sh, set up single observation features like:
maginnov=1.0,
magoberr=0.8,
oneob_type='t',
oblat=38.,
oblon=262.,
obpres=500.,
obdattim= 2014061700,
obhourset=0.,
Note:
Please check Appendix C in the Users Guide for the explanation of each parameter. From these parameters, we can see that a useful observation in the analysis should include information like the observation type (
oneob_type), value (maginnov), observation error (magoberr), location (oblat,oblong,obpres), and time (obdattim,obhourset). Users can dump out (use ncdump) the global attributes from the NetCDF background file and setoblat=CEN_LAT,oblong=360-CEN_LON to have the observation at the center of the domain.In the analysis, the GSI first generates a prepbufr file including only one observation based on the information given in the namelist
&SINGLEOB_TESTsection. To generate this prepbufr file, the GSI needs to read in a PrepBUFR table, which is not needed when running a GSI analysis with real observations. The BUFR table is in the fix/ directory and needs to be copied to the run directory. We have put the following lines in the GSI run script for the single observation test:bufrtable=${FIX_ROOT}/prepobs_prep.bufrtable cp $bufrtable ./prepobs_prep.bufrtable
Examples of Single Observation Tests for GSI¶
Figure 1.1 is a single observation test that has a
temperature observation (oneob_type='t') with a one degree
innovation (maginnov=1.0) and a 0.8 degree observation error
(magoberr=0.8). The background error covariance converted from
global (GFS) BE was picked to provide for better illustration.
Horizontal cross sections (left column) and vertical cross sections (right column) of analysis increment of T, U, V, and Q from a single T observation
This single observation was located at the center of the domain. The results are shown with figures of the horizontal and vertical cross sections through the point of maximum analysis increment. The Figure 1.1 was generated using NCL scripts, which can be found in the util/Analysis_Utilities/plots_ncl directory, introduced in Section A.4.
Control Data Usage¶
Observation data used in the GSI analysis can be controlled through three parts of the GSI system:
- In the GSI run script, by linking observation BUFR files to the working directory
- In section
&OBS_INPUTof the GSI namelist (inside comgsi_namelist.sh) - Through parameters in info files (e.g.: convinfo, satinfo, etc.)
Each part provides different levels of control for data usage in the GSI, which is introduced below:
- Link observation BUFR files to the working directory in the GSI run script:All BUFR/PrepBUFR observation files need to be linked to the working directory with GSI recognizable names before they can be used in a GSI analysis. The run script (run_gsi_regional.ksh) makes these links after locating the working directory. Turning these links on or off can control the use of all the data contained in the BUFR files. Table [tab41] provides a list of all default observation file names recognized by GSI and the corresponding examples of the observation BUFR files from NCEP. The following is the first three rows of the table as an example:
List of all default observation file names recognized by GSI¶ GSI Name Content Example file names prepbufr Conventional observations, including ps, t, q, pw, uv, spd, dw, sst, from observation platforms such as METAR, soundings, etc. gdas1.t12z.prepbufr satwndbufr satellite winds gdas1.t12z.satwnd.tm00.bufr_d amsuabufr AMSU-A 1b radiance (brightness temperatures) from satellites NOAA-15, 16, 17,18, 19, and METOP-A/B gdas1.t12z.1bamua.tm00.bufr_d gdas1.t12z.1bamua.tm00.bufr_d [tab41]
The left column is the GSI recognized name (bold) and the right column are names of BUFR files from NCEP (italic). In the run script, the following lines are used to link the BUFR files in the right column to the working directory using the GSI recognized names shown in the left column:
# Link to the prepbufr data ln -s ${PREPBUFR} ./prepbufr # Link to the radiance data ln -s ${OBS_ROOT}/gdas1.t12z.1bamua.tm00.bufr_d amsuabufrThe GSI recognized default observation filenames are set up in the namelist section&OBS_INPUT, which can be changed based on application needs (see below for details). - In the GSI namelist (inside comgsi_namelist.sh), section
&OBS_INPUT:In this namelist section, observation files (“dfile” column) are tied to the observation variables used inside the GSI code (“dsis” column). For example, part of sectionOBS_INPUTshows:&OBS_INPUT dmesh(1)=120.0,dmesh(2)=60.0,dmesh(3)=30,time_window_max=1.5,ext_sonde=.true., / OBS_INPUT:: ! dfile dtype dplat dsis dval dthin dsfcalc prepbufr ps null ps 1.0 0 0 prepbufr t null t 1.0 0 0 prepbufr q null q 1.0 0 0 prepbufr pw null pw 1.0 0 0 satwndbufr uv null uv 1.0 0 0 prepbufr uv null uv 1.0 0 0 prepbufr spd null spd 1.0 0 0 prepbufr dw null dw 1.0 0 0 radarbufr rw null rw 1.0 0 0 prepbufr sst null sst 1.0 0 0 gpsrobufr gps_ref null gps 1.0 0 0 ssmirrbufr pcp_ssmi dmsp pcp_ssmi 1.0 -1 0 ... amsuabufr amsua n15 amsua_n15 10.0 2 0 amsuabufr amsua n18 amsua_n18 10.0 2 0 ...
This setup tells GSI that conventional observation variables
ps,t, andqshould be read in from the file prepbufr, while AMSU-A radiances from NOAA-15 and NOAA-18 satellites should be read in from the file amsuabufr. Deleting a particular line in&OBS_INPUTwill turn off the use of the observation variable presented by the line in the GSI analysis but other variables under the same type can still be used. For example, if we delete:amsuabufr amsua n15 amsua_n15 10.0 2 0
Then, the AMSU-A observation from NOAA-15 will not be used in the analysis but the AMSU-A observations from NOAA-18 will still be used.The observation filename in “dfile” can be different from the sample script (comgsi_namelist.ksh). If the filename in “dfile” has been changed, the link from the BUFR files to the GSI recognized name in the run script also needs to be changed correspondingly. For example, if we change the “dfile” in amsuabufr for NOAA-15 to beamsuabufr_n15,amsuabufr_n15 amsua n15 amsua_n15 10.0 2 0 amsuabufr amsua n18 amsua_n18 10.0 2 0
Then a new link needs to be added in the run script:
# Link to the radiance data ln -s ${OBS_ROOT}/le_gdas1.t00z.1bamua.tm00.bufr_d amsuabufr ln -s ${OBS_ROOT}/le_gdas1.t00z.1bamua.tm00.bufr_d amsuabufr_n15The GSI will read NOAA-18 AMSU-A observations from file amsuabufr and NOAA-15 AMSU-A observations from fileamsuabufr_n15based on the above changes to the run scripts and namelist. In this example, both amsuabufr andamsuabufr_15are linked to the same BUFR file and NOAA-15 AMSU-A and NOAA-18 AMSU-A observations are still read in from the same BUFR file. If amsuabufr andamsuabufr_15link to different BUFR files, then NOAA-15 AMSU-A and NOAA-18 AMSU-A will be read in from different BUFR files. The changeable filename in dfile gives GSI more flexibility to handle multiple data resources. - Use info files to control data usageFor each variable, observations can come from multiple platforms (data types or observation instruments). For example, surface pressure (ps) can come from METAR observation stations (data type 187) and rawinsonde (data type 120). There are several files named *info in the GSI system (located in ./fix) to control the usage of observations based on the observation platform. Table [tab42] is a list of info files and their function:
The content of info files¶ GSI File name Function and Content convinfo Control the usage of conventional data, including tcp, ps, t, q, pw, sst, uv, spd, dw, radial wind (Level 2 rw and 2.5 srw), gps, and pm2_5 satinfo Control the usage of satellite data. Instruments include AMSU-A/B, HIRS3/4, MHS, ssmi, ssmis, iasi, airs, sndr, cris, amsre, imgr, seviri, atms, avhrr3, etc., and satellites include NOAA 15, 17, 18, 19, aqua, GOES 11, 12, 13, METOP-A/B, NPP, DMSP 15,16,17,18,19,20, M08, M09, M10, etc. ozinfo Control the usage of ozone data, including sbuv6, 8 from NOAA 14, 16, 17, 18, 19. omi_aura, gome_metop-a, and mls_aura pcpinfo Control the usage of precipitation data, including pcp_ssmi and pcp_tmi aeroinfo Control the usage of aerosol data, including modis_aqua and modis_terra [tab42]
The header of each info file includes an explanation of the content of the file. Here we discuss the two most commonly used info files:
- convinfoThe convinfo file controls the usage of conventional data. The following is part of the convinfo file:
!otype type sub iuse twindow numgrp ngroup nmiter gross ermax ermin var_b var_pg ithin rmesh pmesh npred pmot ptime tcp 112 0 1 3.0 0 0 0 75.0 5.0 1.0 75.0 0.000000 0 0. 0. 0 0. 0. ps 120 0 1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. ps 132 0 -1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. ps 180 0 1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. ps 180 01 1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. ps 181 0 1 3.0 0 0 0 3.6 3.0 1.0 3.6 0.000300 0 0. 0. 0 0. 0. ps 182 0 1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. ps 183 0 -1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. ps 187 0 1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0 0. 0. 0 0. 0. t 120 0 1 3.0 0 0 0 8.0 5.6 1.3 8.0 0.000001 0 0. 0. 0 0. 0. t 126 0 -1 3.0 0 0 0 8.0 5.6 1.3 8.0 0.001000 0 0. 0. 0 0. 0. t 130 0 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.001000 0 0. 0. 0 0. 0. t 131 0 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.001000 0 0. 0. 0 0. 0. t 132 0 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.001000 0 0. 0. 0 0. 0. t 133 0 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0. t 134 0 -1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0. t 135 0 -1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0. t 180 0 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0. t 180 01 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0. t 181 0 -1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0. t 182 0 1 3.0 0 0 0 7.0 5.6 1.3 7.0 0.004000 0 0. 0. 0 0. 0.
The meaning of each column is explained in the header of the file and is listed in Table [tab43].list of the content for each convinfo column¶ Column Name Content of the column otype observation variables (t, uv, q, etc.) type prepbufr observation type (if available) sub prepbufr subtype (not yet available) iuse flag if to use/not use / monitor data; = 1, use data, the data type will be read and used in the analysis after quality control; = 0, read in and process data, use for quality control, but do NOT assimilate; = -1, monitor data. This data type will be read in and monitored but not be used in the GSI analysis. twindow time window (+/- hours) for data used in the analysis numgrp cross validation parameter - number of groups ngroup cross validation parameter - group to remove from data use nmiter cross validation parameter - external iteration to introduce removed data gross gross error parameter - gross error ermax gross error parameter - maximum error ermin gross error parameter - minimum error var_b variational quality control parameter - b parameter var_pg variational quality control parameter - pg parameter ithin Flag to turn on thinning (0, no thinning, 1 - thinning) rmesh size of horizontal thinning mesh (in kilometers) pmesh size of vertical thinning mesh npred Number of bias correction predictors pmot Option to keep thinned data as monitored, 0: do not keep, other values: keep ptime time interval for thinning, 0, no temporal thinning, other values define time interval (less than six) [tab43]
From this table, we can see that parameter “iuse” is used to control the usage of data and parameter “twindow” is used to control the time window of data usage. Parameters gross, ermax, and ermin are for gross quality control. Through these parameters, GSI can control how to use certain types of data in the analysis. - satinfoThe satinfo file contains information about the channels, sensors, and satellites. It specifies observation error (cloudy or clear) for each channel, how to use the channels (assimilate, monitor, etc), and other useful information. The following is part of the content of satinfo. The meaning of each column is explained in Table [tab44].
!sensor/instr/sat chan iuse error error_cld ermax var_b var_pg cld_det amsua_n15 1 1 3.000 20.000 4.500 10.000 0.000 -2 amsua_n15 2 1 2.200 18.000 4.500 10.000 0.000 -2 amsua_n15 3 1 2.000 12.000 4.500 10.000 0.000 -2 amsua_n15 4 1 0.600 3.000 2.500 10.000 0.000 -2 amsua_n15 5 1 0.300 0.500 2.000 10.000 0.000 -2 amsua_n15 6 -1 0.230 0.300 2.000 10.000 0.000 -2 amsua_n15 7 1 0.250 0.250 2.000 10.000 0.000 -2 amsua_n15 8 1 0.275 0.275 2.000 10.000 0.000 -2 amsua_n15 9 1 0.340 0.340 2.000 10.000 0.000 -2 amsua_n15 10 1 0.400 0.400 2.000 10.000 0.000 -2 amsua_n15 11 -1 0.600 0.600 2.500 10.000 0.000 -2 amsua_n15 12 1 1.000 1.000 3.500 10.000 0.000 -2 amsua_n15 13 1 1.500 1.500 4.500 10.000 0.000 -2 amsua_n15 14 -1 2.000 2.000 4.500 10.000 0.000 -2 amsua_n15 15 1 3.500 15.000 4.500 10.000 0.000 -2 hirs3_n17 1 -1 2.000 0.000 4.500 10.000 0.000 -1 hirs3_n17 2 -1 0.600 0.000 2.500 10.000 0.000 1 hirs3_n17 3 -1 0.530 0.000 2.500 10.000 0.000 1
list of the content for each satinfo column¶ Column Name Content of the column sensor/instr/sat Sensor, instrument, and satellite name chan Channel number for certain sensor iuse = 1, use this channel data; =-1, dont use this channel data error Variance for each satellite channel error_cld Variance for each satellite channel if cloudy ermax Error maximum for gross check to observations var_b Possible range of variable for gross errors var_pg Probability of gross error icld_det Use this channel in cloud detection if > 0 [tab44]
Domain Partition for Parallelization and Observation Distribution¶
The standard output file (stdout) has an information block that shows
the distribution of different kinds of observations in each sub-domain.
This block follows the observation input section. The following is the
observation distribution from the case shown in Section
1.1. From the introduction, we know the prepbufr
(conventional data), radiance BUFR files, and GPS BUFR files were used.
In this list, the conventional observations (ps, t, q,
pw, and uv), GPSRO (gps_ref), and radiance data (amusa,
hirs4, and mhs from Metop-a, Metop-b, NOAA 15, and
18) were distributed among four sub-domains:
OBS_PARA: ps 1429 3190 4655 6774
OBS_PARA: t 2564 5200 7057 11128
OBS_PARA: q 2346 4626 6148 8128
OBS_PARA: pw 65 80 63 49
OBS_PARA: uv 3358 6453 8091 11998
OBS_PARA: gps_ref 1799 1368 2664 3520
OBS_PARA: hirs4 metop-a 0 0 1146 1661
OBS_PARA: hirs4 n19 213 1020 0 0
OBS_PARA: hirs4 metop-b 0 0 85 555
OBS_PARA: amsua n15 1458 2026 830 234
OBS_PARA: amsua n18 2223 2318 108 0
OBS_PARA: amsua n19 176 960 0 0
OBS_PARA: amsua metop-a 0 0 1077 1559
OBS_PARA: amsua metop-b 0 0 265 1829
OBS_PARA: mhs n18 2550 2695 160 0
OBS_PARA: mhs n19 246 1103 0 0
OBS_PARA: mhs metop-a 0 0 1237 1809
OBS_PARA: mhs metop-b 0 0 321 2128
This list is a good way to quickly check which kinds of data are used in the analysis and how they are distributed in the analysis domain.
Observation Innovation Statistics¶
The GSI analysis provides a set of files named fort.2* to summarize observations fit to the current solution in each outer loop (except for fort.220, see explanation in the next section). The content of some of these files is listed in Table [tab45]:
| File Name | Variables in file | Ranges/units |
|---|---|---|
| fort.201 or fit_p1.analysis_time | fit of surface pressure data | mb |
| fort.202 or fit_w1.analysis_time | fit of u, v wind data | m/s |
| fort.203 or fit_t1.analysis_time | fit of temperature data | K |
| fort.204 or fit_q1.analysis_time | fit of moisture data | percent of qsaturation guess |
| fort.205 | fit of precipitation water data | mm |
| fort.206 | fit of ozone observations from sbuv6_n14 (, _n16, _n17, _n18), sbuv8_n16 (, _n17, _n18, _n19), omi_aura, gome_metop-a/b, and mls_aura | |
| fort.207 or fit_rad1.analysis_time | fit of satellite radiance data, such as: amsua_n15(, n16, n17, n18, metop-a, aqua, n19), amsub_n17, hirs3_n17, hirs4_n19 (, metop-a), etc. | |
| fort.208 | fit of precipitation rate (pcp_ssmi and pcp_tmi) | |
| fort.209 | fit of radar radial wind (rw) | |
| fort.210 | fit of lidar wind (dw) | |
| fort.211 | fit of radar superob wind data (srw) | |
| fort.212 | fit of GPS data (refractivity or bending angle) | fractional difference |
| fort.213 | fit of conventional sst data | C |
| fort.214 | fit of tropical cyclone central pressure | |
| fort.215 | fit of Lagrangian tracer data | |
| fort.217 | fit of aerosol product (aod) | |
| fort.218 | fit of wind gust | |
| Fort.219 | fit of visibility |
[tab45]
To help users understand the information inside these files, some examples are given in the following sub-sections with corresponding explanations.
Conventional observations¶
Example of files, including single level data (fort.201, fort.205, and fort.213)
current fit of surface pressure data, ranges in mb
--------------------------------------------------
pressure levels (hPa)= 0.0 2000.0
it obs type stype count bias rms cpen qcpen
o-g 01 ps 120 0000 101 0.1813 0.6089 0.4711 0.4711
o-g 01 ps 180 0000 601 0.0378 0.6354 0.7171 0.7171
o-g 01 ps 180 0001 1428 0.1857 0.8555 0.7164 0.7164
o-g 01 ps 181 0000 610 0.1959 0.8991 0.7387 0.7387
o-g 01 ps 182 0000 1 0.9173 0.9173 2.8976 2.8976
o-g 01 ps 187 0000 11149 0.1999 0.7877 0.3360 0.3360
o-g 01 all 13890 0.1912 0.7931 0.4105 0.4105
o-g 01 ps rej 120 0000 1 -3.5799 3.5799 0.0000 0.0000
o-g 01 ps rej 180 0001 7 -0.3349 4.9273 0.0000 0.0000
o-g 01 ps rej 181 0000 47 1.1146 54.8539 0.0000 0.0000
o-g 01 ps rej 183 0000 3 10.4797 10.4836 0.0000 0.0000
o-g 01 ps rej 187 0000 52 4.2528 7.4112 0.0000 0.0000
o-g 01 rej all 110 2.7186 36.2804 0.0000 0.0000
o-g 01 ps mon 132 0000 1 0.9173 0.9173 0.2104 0.2104
o-g 01 ps mon 180 0000 113 0.0447 0.5158 0.8709 0.8709
o-g 01 ps mon 180 0001 24 0.2122 0.4050 0.3559 0.3559
o-g 01 ps mon 181 0000 207 0.0910 0.9492 1.2514 1.2514
o-g 01 ps mon 183 0000 1386 0.3974 1.0861 0.0000 0.0000
o-g 01 ps mon 187 0000 88 -0.0100 0.5290 0.7534 0.7534
o-g 01 mon all 1819 0.3188 1.0169 0.2378 0.2378
Example of files including multiple level data (fort.202, fort.203, and fort.204)
ptop 1000.0 900.0 800.0 600.0 100.0 50.0 0.0
it obs type styp pbot 1200.0 1000.0 900.0 800.0 150.0 100.0 2000.0
------------------------------------------------------------------------------------------
o-g 01 uv 220 0000 count 44 223 223 478 437 519 4231
o-g 01 uv 220 0000 bias 0.26 0.51 0.04 0.35 0.99 0.97 0.59
o-g 01 uv 220 0000 rms 2.21 2.48 2.51 2.84 5.88 4.79 4.18
o-g 01 uv 220 0000 cpen 0.31 0.38 0.44 0.61 1.64 1.18 0.90
o-g 01 uv 220 0000 qcpen 0.31 0.38 0.44 0.60 1.63 1.18 0.89
o-g 01 uv 223 0000 count 0 6 16 88 32 8 331
o-g 01 uv 223 0000 bias 0.00 0.23 -0.29 0.85 -0.76 -4.89 0.27
o-g 01 uv 223 0000 rms 0.00 1.28 1.32 2.86 5.21 6.63 3.76
o-g 01 uv 223 0000 cpen 0.00 0.06 0.07 0.45 0.88 1.31 0.73
...
o-g 01 all count 1799 1726 2001 3050 469 527 13124
o-g 01 all bias 0.05 0.91 0.78 0.63 0.87 0.88 0.62
o-g 01 all rms 2.46 2.58 2.65 3.21 5.83 4.82 3.54
o-g 01 all cpen 0.24 0.24 0.30 0.45 1.58 1.19 0.52
o-g 01 all qcpen 0.23 0.24 0.30 0.44 1.58 1.19 0.52
o-g 01 uv rej 220 0000 count 0 0 0 0 0 0 800
o-g 01 uv rej 220 0000 bias 0.00 0.00 0.00 0.00 0.00 0.00 3.11
o-g 01 uv rej 220 0000 rms 0.00 0.00 0.00 0.00 0.00 0.00 6.41
...
o-g 01 rej all count 23 108 21 41 2 0 1008
o-g 01 rej all bias 44.72 6.40 -0.30 -7.38 11.46 0.00 3.84
o-g 01 rej all rms 56.08 16.60 6.87 10.47 43.15 0.00 12.28
o-g 01 rej all cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 01 rej all qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 01 uv mon 220 0000 count 0 4 0 1 8 9 29
o-g 01 uv mon 220 0000 bias 0.00 1.07 0.00 6.04 -1.03 0.23 -2.21
o-g 01 uv mon 220 0000 rms 0.00 20.49 0.00 17.78 9.06 8.51 12.52
o-g 01 uv mon 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 01 uv mon 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
...
o-g 01 mon all count 3573 7736 1004 563 8 9 13052
o-g 01 mon all bias -0.01 0.40 0.05 -0.30 -1.03 0.23 0.23
o-g 01 mon all rms 2.55 3.25 4.67 5.98 9.06 8.51 3.60
o-g 01 mon all cpen 0.81 1.21 1.67 1.28 0.00 0.00 1.12
o-g 01 mon all qcpen 0.80 1.15 1.53 1.16 0.00 0.00 1.07
...
Please note that five layers from 600 to 150 hPa have been deleted to make each row fit into one line. Only observation type 220 and 223 are shown as an example.
Table [tab46] lists the meaning of each item in file fort.201-213 except file fort.207:
| Name | Explanation |
|---|---|
| it | outer loop number = 01: observation - background = 02: observation - analysis (after 1st outer loop) = 03: observation - analysis (after 2nd outer loop) |
| obs | observation variable type (such as uv or ps) and usage, which includes: blank: used in GSI analysis mon: monitored (read in but not assimilated by GSI) rej: rejected because of quality control in GSI |
| type | prepbufr observation type (see the BUFR Users Guide for details) |
| styp | prepbufr observation subtype (not used) |
| ptop | for multiple level data: pressure at the top of the layer |
| pbot | for multiple level data: pressure at the bottom of the layer |
| count | the number of observations summarized under observation types and vertical layers |
| bias | bias of observation departure for each outer loop (it) |
| rms | root mean square error of observation departure for each outer loop (it) |
| cpen | observation part of penalty (cost function) |
| qcpen | nonlinear qc penalty |
[tab46]
The contents of the fit files are calculated based on O-B or O-A for each observation. The detailed departure information about each observation is saved in the diagnostic files. For the content of the diagnostic files, please check the content of the array “rdiagbuf” in one of the setup subroutines for conventional data (for example, setupt.f90). We provide a tool in appendix A.2 to help users read in the information from the diagnostic files.
These fit files give lots of useful information on how data are analyzed by the GSI, such as how many observations are used and rejected, the bias and root mean squared (RMS) error for certain data types or for all observations, and how analysis results fit to the observation before and after analysis. Again, we use observation type 220 in fort.202 (fit_w1.2014061700) as an example to illustrate how to read this information. The fit information for observation type 220 (soundings) is listed below. Like the previous example, five layers from 600 to 150 hPa were deleted to make each row fit into one line. All fit information of observation type 220 is shown.
ptop 1000.0 900.0 800.0 600.0 100.0 50.0 0.0
it obs type styp pbot 1200.0 1000.0 900.0 800.0 150.0 100.0 2000.0
-------------------------------------------------------------------------------------
o-g 01 uv 220 0000 count 44 223 223 478 437 519 4231
o-g 01 uv 220 0000 bias 0.26 0.51 0.04 0.35 0.99 0.97 0.59
o-g 01 uv 220 0000 rms 2.21 2.48 2.51 2.84 5.88 4.79 4.18
o-g 01 uv 220 0000 cpen 0.31 0.38 0.44 0.61 1.64 1.18 0.90
o-g 01 uv 220 0000 qcpen 0.31 0.38 0.44 0.60 1.63 1.18 0.89
...
o-g 01 uv rej 220 0000 count 0 0 0 0 0 0 800
o-g 01 uv rej 220 0000 bias 0.00 0.00 0.00 0.00 0.00 0.00 3.11
o-g 01 uv rej 220 0000 rms 0.00 0.00 0.00 0.00 0.00 0.00 6.41
o-g 01 uv rej 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 01 uv rej 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
...
o-g 01 uv mon 220 0000 count 0 4 0 1 8 9 29
o-g 01 uv mon 220 0000 bias 0.00 1.07 0.00 6.04 -1.03 0.23 -2.21
o-g 01 uv mon 220 0000 rms 0.00 20.49 0.00 17.78 9.06 8.51 12.52
o-g 01 uv mon 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 01 uv mon 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
...
ptop 1000.0 900.0 800.0 600.0 100.0 50.0 0.0
it obs type styp pbot 1200.0 1000.0 900.0 800.0 150.0 100.0 2000.0
-------------------------------------------------------------------------------------
o-g 02 uv 220 0000 count 44 223 223 478 437 519 4231
o-g 02 uv 220 0000 bias 0.21 0.05 -0.43 0.03 0.69 0.96 0.39
o-g 02 uv 220 0000 rms 2.13 2.26 2.29 2.56 5.19 4.55 3.73
o-g 02 uv 220 0000 cpen 0.32 0.31 0.37 0.50 1.27 1.07 0.71
o-g 02 uv 220 0000 qcpen 0.32 0.31 0.37 0.50 1.27 1.07 0.71
...
o-g 02 uv rej 220 0000 count 0 0 0 0 0 0 800
o-g 02 uv rej 220 0000 bias 0.00 0.00 0.00 0.00 0.00 0.00 2.96
o-g 02 uv rej 220 0000 rms 0.00 0.00 0.00 0.00 0.00 0.00 6.33
o-g 02 uv rej 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 02 uv rej 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
...
o-g 02 uv mon 220 0000 count 0 4 0 1 8 9 29
o-g 02 uv mon 220 0000 bias 0.00 2.16 0.00 5.80 -1.00 0.23 -0.82
o-g 02 uv mon 220 0000 rms 0.00 18.44 0.00 12.60 10.31 8.58 10.87
o-g 02 uv mon 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 02 uv mon 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
...
ptop 1000.0 900.0 800.0 600.0 100.0 50.0 0.0
it obs type styp pbot 1200.0 1000.0 900.0 800.0 150.0 100.0 2000.0
----------------------------------------------------------------------------------------
o-g 03 uv 220 0000 count 44 223 223 478 437 519 4231
o-g 03 uv 220 0000 bias 0.28 0.08 -0.32 0.11 0.74 1.00 0.43
o-g 03 uv 220 0000 rms 2.11 2.17 2.20 2.38 5.00 4.46 3.60
o-g 03 uv 220 0000 cpen 0.31 0.29 0.34 0.43 1.18 1.03 0.65
o-g 03 uv 220 0000 qcpen 0.31 0.29 0.34 0.43 1.18 1.03 0.65
...
o-g 03 uv rej 220 0000 count 0 0 0 0 0 0 800
o-g 03 uv rej 220 0000 bias 0.00 0.00 0.00 0.00 0.00 0.00 2.98
o-g 03 uv rej 220 0000 rms 0.00 0.00 0.00 0.00 0.00 0.00 6.35
o-g 03 uv rej 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 03 uv rej 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
...
o-g 03 uv mon 220 0000 count 0 4 0 1 8 9 29
o-g 03 uv mon 220 0000 bias 0.00 1.86 0.00 6.09 -0.98 0.07 -0.94
o-g 03 uv mon 220 0000 rms 0.00 18.76 0.00 12.59 10.34 8.69 11.01
o-g 03 uv mon 220 0000 cpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
o-g 03 uv mon 220 0000 qcpen 0.00 0.00 0.00 0.00 0.00 0.00 0.00
In loop section o-g 01, from the count line, we can see there
were 4231 sounding observations used in the analysis. Among them, 44
were within the 1000-1200 hPa layer. Also from the count lines, in
the rejection and monitoring section, there were 800 observations
rejected and 29 observations monitored. In the same loop section, from
the bias line and rms lines, we can see the total bias and RMS
error of O-B for the sounding information is 0.59 and 4.18. The bias and
RMS error for each vertical layer can also be found in this file.
Next we can see bias and RMS error values from different loops, as shown with the comparison in the following three lines:
o-g 01 uv 220 0000 rms 2.21 2.48 2.51 2.84 5.88 4.79 4.18
o-g 02 uv 220 0000 rms 2.13 2.26 2.29 2.56 5.19 4.55 3.73
o-g 03 uv 220 0000 rms 2.11 2.17 2.20 2.38 5.00 4.46 3.60
These three lines show that the RMS error reduced from 4.18 (o-g 01, which is O-B) to 3.73 (o-g 02, which is O-A after the 1st outer loop) and then to 3.60 (o-g 03, which is O-A after the 2nd outer loop, which is also the final analysis result). The reduction in the RMS error shows that observation type 220 (sounding) was used in the GSI analysis to modify the background fields to fit to the observations.
Satellite Radiance¶
The file fort.207 is the fit file for radiance data. Its content includes important information about the radiance data analysis.
The first part of the file fort.207 lists the content that corresponds to those in the file satinfo, which is the info file to control radiance data usage.
RADINFO_READ: jpch_rad= 2723
1 amsua_n15 chan= 1 var= 3.000 varch_cld= 20.000 use= 1 ermax= 4.500 b_rad= 10.00 pg_rad= 0.00 icld_det=-2
2 amsua_n15 chan= 2 var= 2.200 varch_cld= 18.000 use= 1 ermax= 4.500 b_rad= 10.00 pg_rad= 0.00 icld_det=-2
3 amsua_n15 chan= 3 var= 2.000 varch_cld= 12.000 use= 1 ermax= 4.500 b_rad= 10.00 pg_rad= 0.00 icld_det=-2
4 amsua_n15 chan= 4 var= 0.600 varch_cld= 3.000 use= 1 ermax= 2.500 b_rad= 10.00 pg_rad= 0.00 icld_det=-2
5 amsua_n15 chan= 5 var= 0.300 varch_cld= 0.500 use= 1 ermax= 2.000 b_rad= 10.00 pg_rad= 0.00 icld_det=-2
...
This shows there are 2723 channels listed in the satinfo file and the 2723 lines following this line include the details for each channel.
The second part of the file is a list of the coefficients for bias correction, after reading the satbias_in file:
RADINFO_READ: ***WARNING instrument/channel ahi_himawari8 16
not found in satbias_pc file - set to zero
RADINFO_READ: guess air mass bias correction coefficients below
1 amsua_n15 -1.965181 0.000000 39.068176 -0.923195 0.408460 0.000000 0.000000 0.001735 3.689709
2.367472 8.862940 -1.848292
2 amsua_n15 -7.415403 0.000000 80.138019 -0.818472 -1.013216 0.000000 0.000000 0.014279 2.638804
8.315096 22.041458 -1.564773
...
Each channel has 12 coefficients listed in a line. Therefore, there are 2723 lines of radiance bias correction coefficients for all channels, though some of the coefficients are zero.
The 3rd part of the fort.207 file is similar to other fit files with content repeated in three sections, providing detailed statistics about the data in stages before the 1st outer loop, between the 1st and 2nd outer loops, and after the 2nd outer loop. The results before the 1st outer loop are used here as an example to explain the content of the results:
Summaries for various statistics as a function of observation type
sat type penalty nobs iland isnoice icoast ireduce ivarl nlgross metop-a hirs4 371.46818888 385 143 0 25 0 1452 0 qcpenalty qc1 qc2 qc3 qc4 qc5 qc6 qc7 371.46818888 0 0 1651 3330 0 0 0 sat type penalty nobs iland isnoice icoast ireduce ivarl nlgross metop-b hirs4 0.00000000 34 25 0 0 0 139 0 qcpenalty qc1 qc2 qc3 qc4 qc5 qc6 qc7 0.00000000 0 0 97 279 0 0 0 6 rad total penalty_all= 13986.0933845987129 rad total qcpenalty_all= 13986.0933845987129 rad total failed nonlinqc= 0 ...
Table [tab47] outlines the meaning of each item in the above statistics:
Summary of the radiance observation fit file (fort.207)¶ Name Explanation sat satellite name type instrument type penalty contribution to cost function from this observation type nobs number of good observations used in the assimilation iland number of observations over land isnoice number of observations over sea ice and snow icoast number of observations over coast ireduce number of observations that reduce qc bounds in tropics ivarl number of observations removed by gross check nlgross number of observation removed by nonlinear qc qcpenalty nonlinear qc penalty from this data type qc1-7 number of observations whose quality control criteria has been adjusted by each qc method (1-7). For details, see the Radiance Chapter of the Advanced Users Guide rad total penalty_all summary of penalty for all radiance observation types rad total qcpenalty_all summary of qcpenalty for all radiance observation types rad total failed nonlinqc summary of observations removed by nonlinear qc for all radiance observation types [tab47]
Note: One radiance observation may include multiple channels, and not all channels are necessarily used in the analysis.
Summaries for various statistics as a function of channel
1 2 3 4 5 6 7 8 9 10 11 1 1 amsua_n15 915 247 3.000 0.6623748 1.2673027 0.2414967 2.4209147 2.0627099 2 2 amsua_n15 902 261 2.000 0.5209886 1.1773765 0.2626527 2.2705374 1.9414234 3 3 amsua_n15 1114 48 2.000 1.2349467 -1.5090660 0.3421312 1.9416738 1.2218089 4 4 amsua_n15 1162 0 0.600 -0.2114506 -0.3195250 0.5286580 0.6271317 0.5396275 5 5 amsua_n15 1162 3 0.300 -0.0888496 -0.1685199 0.4352839 0.2734131 0.2153040 6 6 amsua_n15 2093 3 -0.230 -1.8141034 -0.0695503 0.8415303 0.2646033 0.2552992 7 7 amsua_n15 2342 28 0.250 -0.0830293 0.0453835 0.6141132 0.2522587 0.2481427 8 8 amsua_n15 2311 59 0.275 0.0201868 0.1758359 0.7782391 0.3173862 0.2642267 9 9 amsua_n15 2124 246 0.340 0.1262987 0.5008846 1.4907547 0.5544190 0.2376868 10 10 amsua_n15 82 2288 0.400 0.6700157 0.8263504 2.0387089 0.8304076 0.0819868 15 15 amsua_n15 862 300 3.000 0.8244923 -2.2960147 0.3769625 2.7445692 1.5036544 ... 63 4 hirs4_metop-a 11 217 0.400 0.9717384 0.9536444 1.8583033 0.9570578 0.0807581 64 5 hirs4_metop-a 81 4 0.360 0.1655806 -0.2231640 0.3231878 0.3433454 0.2609289 65 6 hirs4_metop-a 25 27 0.460 -0.8415009 -1.1454801 2.6082742 1.1558578 0.1545404 66 7 hirs4_metop-a 20 3 0.570 -0.9563649 -1.1287970 1.4138248 1.1699278 0.3074870 67 8 hirs4_metop-a 23 0 1.000 1.3954294 0.6716651 0.1668694 0.9134028 0.6190078 ...
Table [tab48] lists the meaning of each column in the above statistics:
Content of fit statistics for each channel in the fort.207 file¶ Column # Content 1 series number of the channel in satinfo file 2 channel number for certain radiance observation type 3 radiance observation type (for example: amsua_n15) 4 number of observations (nobs) used in GSI analysis within this channel 5 number of observations (nobs) tossed by gross check within this channel 6 variance for each satellite channel 7 bias (observation-guess before bias correction) 8 bias (observation-guess after bias correction) 9 penalty contribution from this channel 10 square root of (observation-guess with bias correction)**2 11 standard deviation [tab48]
Final summary for each observation type
it satellite instrument # read # keep # assim penalty qcpnlty cpen qccpen o-g 01 rad n16 hirs3 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad n17 hirs3 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad metop-a hirs4 13832 7315 651 371.47 371.47 0.57061 0.57061 o-g 01 rad n18 hirs4 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad n19 hirs4 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad metop-b hirs4 2527 646 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad g11 goes_img 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad g12 goes_img 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad aqua airs 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad n15 amsua 43515 30810 12976 8854.4 8854.4 0.68236 0.68236 o-g 01 rad n18 amsua 30690 25575 8203 3642.7 3642.7 0.44407 0.44407 o-g 01 rad n19 amsua 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad metop-a amsua 4890 3878 1466 652.20 652.20 0.44489 0.44489 o-g 01 rad metop-b amsua 810 718 294 201.79 201.79 0.68636 0.68636 o-g 01 rad aqua amsua 0 0 0 0.0000 0.0000 0.0000 0.0000 o-g 01 rad n17 amsub 0 0 0 0.0000 0.0000 0.0000 0.0000 ...
Table [tab49] lists the meaning of each column in the above statistics:
Content of the final summary section for the fort.207 file¶ Name Explanation it stage (o-g 01 rad = before 1st outer loop for radiance data) satellite satellite name (n16=NOAA -16) instrument instrument name (HIRS-3) # read number of data (channels) read in within analysis time window and domain # keep number of data (channels) kept after data thinning # assim number of data (channels) used in analysis (passed all qc process) penalty contribution from this observation type to cost function qcpnlty nonlinear qc penalty from this data type cpen penalty divided by (the number of data assimilated) qccpen qcpnlty divided by (the number of data assimilated) [tab49]
Similar to other fit files, a comparison between results from different outer loops can give us very useful information on how much impact each channel and data type has in the GSI.
Convergence Information¶
There are two ways to check the convergence information for each iteration of the GSI:
- Standard output file (stdout):The value of the cost function and norm of the gradient for each iteration are listed in the file stdout.
The following is an example showing the iterations from the first outer loop:
GLBSOI: START pcgsoi jiter= 1 pcgsoi: gnorm(1:2),b= 9.869857497554413276E+05 9.869857497554413276E+05 0.000000000000000000E+00 Initial cost function = 3.915930707165839704E+04 Initial gradient norm = 9.934715646436194447E+02 cost,grad,step,b,step? = 1 0 3.915930707165839704E+04 9.934715646436194447E+02 4.821051367939106942E-03 0.000000000000000000E+00 good pcgsoi: gnorm(1:2),b= 4.101618851769856410E+05 4.101618851769853500E+05 4.155702200144395508E-01 cost,grad,step,b,step? = 1 1 3.440099807266351127E+04 6.404388223530688720E+02 5.837551195665082078E-03 4.155702200144395508E-01 good pcgsoi: gnorm(1:2),b= 4.679752995326447999E+05 4.679752995326457312E+05 1.140952673675940110E+00 cost,grad,step,b,step? = 1 2 3.200665706943234181E+04 6.840872017021256397E+02 3.727016244273905453E-03 1.140952673675940110E+00 good pcgsoi: gnorm(1:2),b= 2.579351145588153449E+05 2.579351145588165091E+05 5.511724973869557287E-01 ... cost,grad,step,b,step? = 1 9 2.515019683100273687E+04 2.511847728725914237E+02 5.552768045789822915E-03 9.872916643640710088E-01 good pcgsoi: gnorm(1:2),b= 5.868194961445817898E+04 5.868194961445837544E+04 9.300748853414525508E-01 cost,grad,step,b,step? = 1 10 2.479985164931967302E+04 2.422435749704379191E+02 5.481216572289848883E-03 9.300748853414525508E-01 good
The following are the iterations from the second outer loop:
Initial cost function = 2.792919782749931983E+04 Initial gradient norm = 4.241369976412337337E+02 cost,grad,step,b,step? = 2 0 2.792919782749931983E+04 4.241369976412337337E+02 4.301269527061492466E-03 0.000000000000000000E+00 good pcgsoi: gnorm(1:2),b= 1.641799966598818428E+05 1.641799966598813771E+05 9.126577097845959274E-01 cost,grad,step,b,step? = 2 1 2.715543302058953486E+04 4.051913087171068923E+02 3.965037770683364597E-03 9.126577097845959274E-01 good pcgsoi: gnorm(1:2),b= 7.369958699881075881E+04 7.369958699881142820E+04 4.488950450613589660E-01 cost,grad,step,b,step? = 2 2 2.650445313264244396E+04 2.714766785541821719E+02 6.065878225371862040E-03 4.488950450613589660E-01 good pcgsoi: gnorm(1:2),b= 5.010117245725526300E+04 5.010117245725520479E+04 6.798026216627733875E-01 ... cost,grad,step,b,step? = 2 9 2.488459155328830457E+04 1.491452772915179139E+02 6.349376870755444636E-03 9.522651754285673675E-01 good pcgsoi: gnorm(1:2),b= 2.350218829989638834E+04 2.350218829989681763E+04 1.056548139732611746E+00 cost,grad,step,b,step? = 2 10 2.474335402213209454E+04 1.533042344486817683E+02 5.304880218120413757E-03 1.056548139732611746E+00 good
Here, we can see the number of outer and inner loops (minimization iteration). The meaning of the names (bold) used in stdout are explained in the following:
cost: the cost function values, (=J)grad: inner product of gradients (norm of the gradient (Y*X))step: stepsizeb: parameter to estimate the new search direction
As a quick check, the cost function reduced from 3.915930707165839704E+04 to 2.479985164931967302E+04 in the 1st outer loop and reduced from 2.792919782749931983E+04 to 2.474335402213209454E+04 in the 2nd outer loop.
- Convergence information in file fort.220:In file fort.220, users can find more detailed minimization information about each iteration. A detailed description and example are provided in the Advanced Users Guide.
To evaluate the iteration convergence, we usually make plots based on the information from fort.220, such as the value of the cost function and the norm of the gradient. The following are example plots showing the evolution of the cost function and the norm of the gradient in different outer loops:
Evolution of the cost function (left column) and the norm of the gradient (right column) in the first outer loop (top row) and the second outer loop (bottom row). The Y-axis is the iteration number.
Scripts are available in the release code to read convergence information from fort.220 and produce the above plots. Please see Section A.3 for information on where to locate and how to run these scripts.
Conventional Observation Errors¶
Each observation type has its own observation errors. In this section, we introduce several topics related to the conventional observation error processing in GSI. The observation error for satellite radiance and its adjustment is discussed in the Advanced Users Guide.
Getting Original Observation Errors¶
For the global GSI analysis, when oberrflg (a namelist option in
section &obsqc) is true, observation errors are generated based on
an external observation error table according to the observation type.
Otherwise, observation errors are read in from the PrepBUFR file.
For regional analyses, GSI forces the use of an external observation
error table to get observation errors no matter what the oberrflg is
set to (oberrflg is forced to be true for regional runs in
gsimod.F90).
The external observation error table file, errtable, includes observation errors for all types of conventional observations. It is copied from the ./fix directory by the run script. This release package has three sample external observation error table files, nam_errtable.r3dv, prepobs_errtable.global, and rtma/new_rtma_nam_errtable.r3dv in the ./fix directory. The nam_errtable.r3dv is used in the sample run script as a default observation error table. The observation error file is a text file that can be easily edited to tune the error values. The following shows a portion of nam_errtable.r3dv file for rawinsondes and its description of each column in Table [tab410]:
Column # 1 2 3 4 5 6
120 OBSERVATION TYPE
0.11000E+04 0.12696E+01 0.60737E+00 0.10000E+10 0.76322E+00 0.10000E+10
0.10500E+04 0.13282E+01 0.66294E+00 0.10000E+10 0.76322E+00 0.10000E+10
0.10000E+04 0.13932E+01 0.74223E+00 0.10000E+10 0.76322E+00 0.10000E+10
0.95000E+03 0.14390E+01 0.83688E+00 0.10000E+10 0.79899E+00 0.10000E+10
0.90000E+03 0.14354E+01 0.94025E+00 0.10000E+10 0.83561E+00 0.10000E+10
0.85000E+03 0.13669E+01 0.10439E+01 0.10000E+10 0.87224E+00 0.10000E+10
220 OBSERVATION TYPE
0.11000E+04 0.10000E+10 0.10000E+10 0.18521E+01 0.10000E+10 0.10000E+10
0.10500E+04 0.10000E+10 0.10000E+10 0.20636E+01 0.10000E+10 0.10000E+10
0.10000E+04 0.10000E+10 0.10000E+10 0.22799E+01 0.10000E+10 0.10000E+10
0.95000E+03 0.10000E+10 0.10000E+10 0.24211E+01 0.10000E+10 0.10000E+10
0.90000E+03 0.10000E+10 0.10000E+10 0.24934E+01 0.10000E+10 0.10000E+10
0.85000E+03 0.10000E+10 0.10000E+10 0.25155E+01 0.10000E+10 0.10000E+10
| Column # | 1 | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|---|
| Content | Pressure | T | RH | UV | Ps | Pw |
| Unit | hPa | degree C | percent/10 | m/s | mb | kg/m2(or mm) |
[tab410]
For each observation type, the error table has six columns and 33 rows (levels). The 1st column prescribes 33 pressure levels, covering 1100 hPa to 0 hPa. Columns 2-6 prescribe observation errors for temperature (T), relative humidity (RH), horizontal wind component (UV), surface pressure (Ps), and the total column precipitable water (Pw). The missing value is 0.10000E+10.
The observation error table for each observation type starts with the observation type number defined for the PrepBUFR files, such as:
120 OBSERVATION TYPE
220 OBSERVATION TYPE
The PrepBUFR data type numbers 100-199 are for temperature (T), moisture (q), and surface pressure (Ps) observations, while numbers 200-299 are for horizontal wind component (UV) observations. A detailed explanation of each data type number can be found in the following table on the EMC website:
http://www.emc.ncep.noaa.gov/mmb/data_processing/prepbufr.doc/table_2.htm
For more details on PrepBUFR/BUFR, please check the BUFR/PrepBUFR Users Guide, which is freely available at the DTC BUFR/PrepBUFR website:
Observation Error Gross Error Check within GSI¶
The gross error check is an important quality control step to exclude questionable observations that degrade the analysis. Users can adjust the threshold of the gross error check for each data type within the convinfo file to make the gross error check more or less strict for a certain data type. For example, the following is a part of the convinfo file without the last five columns:
!otype type sub iuse twindow numgrp ngroup nmiter gross ermax ermin var_b var_pg ithin
ps 183 0 -1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0
ps 187 0 1 3.0 0 0 0 4.0 3.0 1.0 4.0 0.000300 0
t 120 0 1 3.0 0 0 0 8.0 5.6 1.3 8.0 0.000001 0
t 126 0 -1 3.0 0 0 0 8.0 5.6 1.3 8.0 0.001000 0
The gross check for each data type is controlled by columns “gross”, “ermax”, and “ermin.” If an observation has observation error set to “obserror,” then a gross check ratio is calculated:
ratio = (Observation-Background)/max(ermin,min(ermax,obserror))
If ratio > gross, then this observation fails the gross check and will not be used in the analysis. The unused observation is indicated as a “rejection” in the fit files.
Background Error Covariance¶
The GSI package has several files in ./fix to hold the pre-computed background error statistics for different GSI applications with different grid configurations. Within the ./fix subdirectories ./fix/Big_Endian and ./fix/Little_Endian contain the fix files corresponding to each endianness. Since the GSI code has a build-in mechanism to interpolate the input background error matrix to any desired analysis grid, the following two background error files can be used to specify the B matrix for any GSI regional application.
- nam_nmmstat_na.gcv : contains the regional background error statistics, computed using forecasts from the NCEPs NAM model covering North America. The values of this B matrix cover the northern hemisphere with 93 latitude lines, from -2.5 degrees to 89.5 degrees with 60 vertical sigma levels from 0.9975289 to 0.01364.
- nam_glb_berror.f77.gcv : contains the global background errors based on NCEPs GFS model, a global forecast model. The values of this B matrix cover the globe with 192 latitude lines from -90 degrees to 90 degrees and 42 vertical sigma levels from 0.99597 to 0.013831.
The background error matrix files listed above are in Big Endian binary form (therefore located in the Big_Endian directory). In the Little_Endian directory, nam_nmmstat_na.gcv and nam_glb_berror.f77.gcv are their Little Endian versions for certain computer platforms that cannot compile GSI with the Big Endian option. In this release version, GSI can be compiled with the Big Endian option with PGI and Intel.
Analysis Increments¶
Analysis increments are defined as the difference between analysis results and the background (A-B). A plot of analysis increments can help users understand how the analysis procedure modifies the background fields according to observations, background and observation error covariances, and other constraints. You can either calculate analysis-guess and plot the difference field or use the tools introduced in Appendix A.4 to make analysis increment figures for different analysis fields.
Running Time and Memory Usage¶
In addition to analysis increments, run time and memory usage are other important features of an analysis system, especially for operational code like the GSI.
The GSI standard output file (stdout) gives the GSI start time and end time of the analysis at the beginning and end of the file. For example:
* . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * .
PROGRAM GSI_ANL HAS BEGUN. COMPILED 1999232.55 ORG: NP23
STARTING DATE-TIME JUL 02,2016 20:36:21.760 184 SAT 2457572
...
ENDING DATE-TIME JUL 02,2016 20:43:40.422 184 SAT 2457572
PROGRAM GSI_ANL HAS ENDED.
* . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * . * .
This tells us the analysis started at 20:36:21.760 and ended at 20:43:40.422, which means GSI used 7 minutes and 19 seconds to finish.
Following the ending date-time, there is a resource statistics section at the end of the stdout file, which gives information about run time and memory usage for the analysis:
*****************RESOURCE STATISTICS*******************************
The total amount of wall time = 438.663534
The total amount of time in user mode = 427.578998
The total amount of time in sys mode = 9.457562
The maximum resident set size (KB) = 2020132
Number of page faults without I/O activity = 312762
Number of page faults with I/O activity = 0
Number of times filesystem performed INPUT = 0
Number of times filesystem performed OUTPUT = 0
Number of Voluntary Context Switches = 7641
Number of InVoluntary Context Switches = 851
*****************END OF RESOURCE STATISTICS*************************