C. GSI Namelist: Name, Default Value, Explanation¶
The following are lists and explanations of the GSI namelist variables. You can also find them in the source code gsimod.F90.
Variable name | Default value | Description |
---|---|---|
SETUP | General control namelist | |
gencode | 80 | source generation code |
factqmin | 1 | weighting factor for negative moisture constraint |
factqmax | 1 | weighting factor for supersaturated moisture constraint |
clip_supersaturation | .false. | flag to remove supersaturation during each outer loop |
factv | 1 | weighting factor for negative visibility constraint |
factl | ||
factp | ||
factg | ||
factw10m | ||
facthowv | ||
deltim | 1200 | model timestep |
dtphys | 3600 | physics timestep |
biascor | -1 | background error bias correction coefficient |
bcoption | 1 | 0=ibc (no bias correction to bkg); |
1= sbc(original implementation) | ||
diurnalbc | 0 | 1= diurnal bias; 0= persistent bias |
Variable name | Default value | Description |
---|---|---|
SETUP | General control namelist | |
niter(0:50) | 0,… | Maximum number of inner loop iterations for each outer loop |
niter_no_qc(0:50) | 1000000 | Inner loop iteration at which to turn on variational quality control |
miter | 1 | number of outer loops |
qoption | 1 | option for moisture analysis variable; 1:q/qsatg 2:normalized RH |
cwoption | ||
pseudo_q2 | .false. | breed between q1/q2 options, that is, (q1/sig(q)) |
nhr_assimilation | 6 | assimilation time interval (currently 6 hours for global, 3 hours for regional ) |
min_offset | 3 | time of analysis in assimilation window |
iout_iter | 220 | output file number for iteration information |
npredp | 6 | number of predictors for precipitation bias correction |
retrieval | .false. | logical to turn off or on the SST physical retrieval |
nst_gsi | 0 | indicator to control the Tr Analysis mode: |
0 = no nst info ingsi at all; | ||
1 = input nst info, but used for monitoring only | ||
2 = input nst info, and used in CRTM simulation, but no Tr analysis | ||
3 = input nst info, and used in CRTM simulation and Tr analysis is on | ||
nst_tzr | 0 | indicator to control the Tzr_QC mode: |
0 = no Tz retrieval; | ||
1 = Do Tz retrieval and applied to QC | ||
nstinfo | 0 | number of nst variables |
fac_dtl | 0 | index to apply diurnal thermocline layer or not: 0 = no; 1 = yes |
fac_tsl | 0 | index to apply thermal skin layer or not: 0 = no; 1 = yes. |
nst_tzr | ||
tzr_bufrsave | .false. | logical to turn off or on the bufr Tz retrieval file true=on |
diag_rad | .true. | logical to turn off or on the diagnostic radiance file (true=on) |
diag_pcp | .true. | logical to turn off or on the diagnostic precipitation file (true=on) |
diag_conv | .true. | logical to turn off or on the diagnostic conventional file (true=on) |
diag_ozone | .true. | logical to turn off or on the diagnostic ozone file (true=on) |
diag_aero | .false. | logical to turn off or on the diagnostic aerosol file (true=on) |
diag_co | .false. | logical to turn off or on the diagnostic carbon monoxide file (true=on) |
iguess | 1 | flag for guess solution (currently not working) |
-1 do not use guess file | ||
0 write only guess file | ||
1 read and write guess file | ||
2 read only guess file | ||
write_diag | .false., … | logical to write out diagnostic files for outer iteration |
reduce_diag | .false. | namelist logical to produce reduced radiance diagnostic files |
Variable name | Default value | Description |
---|---|---|
SETUP | General control namelist | |
oneobtest | .false. | one observation test flag true=on |
sfcmodel | .false. | if true, then use boundary layer forward model for surface temperature data. |
dtbduv_on | .true. | logical for switching on (.true.) sensitivity of uv winds to microwave brightness temperatures. if true, use d(microwave brightness temperature)/d(uv wind) in inner loop |
ifact10 | 0 | flag for recomputing 10m wind factor |
= 1 compute using GFS surface physics | ||
= 2 compute using MM5 surface physics | ||
= 0 or any other value - DO NOT recompute - use value from guess file | ||
l_foto | .false. | option for First-Order Time extrapolation to observation |
offtime_data | .false. | if true, then allow use of obs files with ref time different from analysis time. default value = .false., in which case analysis fails if observation file reference time is different from analysis time. |
npred_conv_max | 0 | maximum number of conventional observation bias correction coefficients |
id_bias_ps | 0 | prepbufr id to have conv_bias added for testing |
id_bias_t | 0 | prepbufr id to have conv_bias added for testing |
id_bias_spd | 120 | prepbufr id to have conv_bias added for testing |
conv_bias_ps | 0 | magnitude of ps bias(mb) |
conv_bias_t | 0 | magnitude of t bias(deg K) |
conv_bias_spd | 0 | magnitude of spd bias(m/sec) |
id_bias_pm2_5 | ||
conv_bias_pm2_5 | ||
id_bias_pm10 | ||
conv_bias_pm10 | ||
stndev_conv_ps | 1.0 | |
stndev_conv_t | 1.0 | |
stndev_conv_spd | 1.0 | |
use_pbl | .false. | Logical flag to include PBL effects in tendency model. |
use_compress | .false. | option to turn on the use of compressibility factors in geopotential heights |
nsig_ext | 13 | number of layers above the model top which are necessary to compute the bending angle for gpsro |
gpstop | 30.0 | maximum height for gpsro data assimilation. Reject anything above this height. (km) |
perturb_obs | .false. | logical flag to perturb observation (true=on) |
perturb_fact | 1 | magnitude factor for observation perturbation |
oberror_tune | .false. | logical to tune (=true) oberror |
preserve_restart_date | .false. | if true, then do not update regional restart file date. |
crtm_coeffs_path | ./ | path of directory w/ CRTM coeffs files |
berror_stats | berror_stats | filename if other than “berror_stats” |
newpc4pred | .false. | option for additional preconditioning for pred coeff |
adp_anglebc | .false. | option to perform variational angle bias correction |
angord | 0 | order of polynomial for variational angle bias correction |
Variable name | Default value | Description |
---|---|---|
SETUP | General control namelist | |
passive_bc | .false. | option to turn on bias correction for passive (monitored) channels |
use_edges | .true. | option to exclude radiance data on scan edges |
biaspredvar | 0.1 | set background error variance for radiance bias coeffs |
lobsdiagsave | .false. | write out additional observation diagnostics |
l4dvar | .false. | turn 4D-Var on/off (default=off=3D-Var) |
lbicg | .false. | use B-precond w/ bi-conjugate gradient for minimization |
lsqrtb | .false. | Use sqrt(B) preconditioning |
lcongrad | .false. | Use conjugate gradient/Lanczos minimizer |
lbfgsmin | .false. | Use L-BFGS minimizer |
ltlint | .false. | Use TL inner loop (ie TL intall) |
nhr_obsbin | -1 | length of observation bins |
nhr_subwin | -1 | length of weak constraint 4d-Var sub-window intervals |
nwrvecs | -1 | Number of precond vectors (Lanczos) or pairs of vectors (QN) being saved |
iorthomax | 0 | max number of vectors used for orthogonalization of various CG options |
ladtest | .false. | Run adjoint test |
ladtest_obs | .false. | if true, doing the adjoint check for the observation operators |
lgrtest | .false. | Run gradient test |
lobskeep | .false. | keep obs from first outer loop for subsequent OL |
lsensrecompute | .false. | does adjoint by recomputing forward solution |
jsiga | -1 | calculate approximate analysis errors from lanczos for jiter=jsiga |
ltcost | .false. | calculate true cost when using Lanczos (this is very expensive) |
lobsensfc | .false. | compute forecast sensitivity to observations |
lobsensjb | .false. | compute Jb sensitivity to observations |
lobsensincr | .false. | compute increment sensitivity to observations |
lobsensadj | .false. | use adjoint of approx. Hessian to compute obs sensitivity |
lobsensmin | .false. | use minimisation to compute obs sensitivity |
iobsconv | 0 | compute convergence test in observation space |
=1 at final point, =2 at every iteration | ||
idmodel | .false. | uses identity model when running 4D-Var (test purposes) |
iwrtinc | .false. | when .t., writes out increments instead of analysis |
jiterstart | 1 | first outloop iteration number |
jiterend | 1 | last outloop iteration number |
lobserver | .false. | when .t., calculate departure vectors only |
lanczosave | .false. | save lanczos vectors for forecast sensitivity computation |
llancdone | .false. | use to tell adjoint that Lanczos vecs have been pre-computed |
lferrscale | .false. | Something related to forecast error |
print_diag_pcg | .false. | logical turn on of printing of GMAO diagnostics in pcgsoi.f90 |
tsensible | .false. | option to use sensible temperature as the analysis variable. Works only for twodvar_regional=.true. |
lgschmidt | .false. | option for re-biorthogonalization of the gradx and grady set from pcgsoi when twodvar_regional=.true. |
Variable name | Default value | Description |
---|---|---|
SETUP | General control namelist | |
lread_obs_save | .false. | option to write out collective obs selection info |
lread_obs_skip | .false. | option to read in collective obs selection info |
use_gfs_ozone | .false. | option to read in gfs ozone and interpolate to regional model domain |
check_gfs_ozone_date | .false. | option to date check gfs ozone before interpolating to regional model domain |
regional_ozone | .false. | option to turn on ozone in regional analysis |
lwrite_predterms | .false. | option to write out actual predictor terms instead of predicted bias to the radiance diagnostic files |
lwrite_peakwt | .false. | option to writ out the approximate pressure of the peak of the weighting function for satellite data to the radiance diagnostic files |
use_gfs_nemsio | .false. | option to use nemsio to read global model NEMS/GFS first guess |
liauon | .false. | treat 4dvar CV as tendency perturbation (default=false) |
use_prepb_satwnd | .false. | allow using satwnd’s from prepbufr (historical) file |
l4densvar | .false. | logical to turn on ensemble 4dvar |
ens4d_nstarthr | 3 | start hour for ensemble perturbations (generally should match min_offset) |
use_gfs_stratosphere | When true, a guess gfs valid at the same time as the nems-nmmb guess is used to replace the upper levels with gfs values. The purpose of this is to allow direct use of gdas derived sat radiance bias correction coefs. | |
pblend0 | 152 | The nems-nmmb vertical coordinate is smoothly merged with gfs above this level. Below this level, is original nems-nmmb. |
pblend1 | 79.0 | The nems-nmmb vertical coordinate is smoothly merged with gfs below this level. Above this level,is gfs. |
step_start | 1.e-4 | initial stepsize in minimization |
diag_precon | .false. | if true do preconditioning |
lrun_subdirs | .false. | logical to toggle use of subdirectires at runtime for pe specific files |
emiss_bc | .false. | option to turn on emissivity bias predictor |
upd_pred | 1 | bias update indicator for radiance bias correction; 1.0=bias correction coefficients evolve |
use_reflectivity | .false. | option of using reflectivity |
lnested_loops | .false. | allow for nested resolution outer/inner loops |
lwrite4danl | .false. | logical to write out 4d analysis states if 4dvar or 4denvar mode |
lsingleradob | .false. | logical for single radiance observation assimilation. Uses existing bufr file and rejects all radiances that don’t fall within a tight threshold around oblat/oblon (SINGLEOB_TEST) |
ssmis_method | 1 | choose method for SSMIS noise reduction 0=no smoothing 1=default |
ssmis_precond | 0.01 | weighting factor for SSMIS preconditioning (if not using newpc4pred) |
R_option | .false. | Option to use variable correlation length for lcbas based on data density - follows Hayden and Purser (1995) (twodvar_regional only) |
thin4d |
Variable name | Default value | Description |
---|---|---|
GRIDOPTS | Grid setup variables, including regional specific variables | |
jcap | 62 | spectral resolution of the analysis |
jcap_b | 62 | spectral resolution of background (model guess field) |
nsig | 42 | number of sigma levels |
nlat | 96 | number of latitudes |
nlon | 384 | number of longitudes |
hybrid | logical | hybrid data file flag true=hybrid |
nlat_regional | 0 | Number of y grid point in whole regional domain |
nlon_regional | 0 | Number of x grid point in whole regional domain |
diagnostic_reg | .false. | logical for regional debugging |
update_regsfc | .false. | logical to write out updated surface fields to the regional analysis file (default = false) |
netcdf | .false. | if true, then wrf files are in netcdf format, otherwise wrf files are in binary format. |
regional | .false. | logical for regional GSI run |
wrf_nmm_regional | .false. | logical for input from WRF NMM |
nems_nmmb_regional | .false. | logical for input from NEMS NMMB |
wrf_mass_regional | .false. | logical for input from WRF MASS-CORE (ARW) |
twodvar_regional | .false. | logical for regional 2d-var analysis |
filled_grid | .false. | logical to fill in points on WRF-NMM E-grid |
half_grid | .false. | logical to use every other row of WRF-NMM E-Grid |
nvege_type | 24 | number of types of vegetation; old=24, IGBP=20 |
nlayers(100) | 1 | number of sub-layers to break indicated model layer into prior to calling radiative transfer model |
cmaq_regional | .false. | Background input is from CMAQ model |
nmmb_reference_grid | H | =’H’, then analysis grid covers H grid domain |
= ’V’, then analysis grid covers V grid domain | ||
grid_ratio_nmmb | sqrt(2) | ratio of analysis grid to nmmb model grid in nmmb model grid units. |
grid_ratio_wrfmass | 1.0 | ratio of analysis grid to wrf mass grid in wrf grid units |
jcap_gfs | spectral truncation used to transform high wavenumber spectral coefficients to a coarser resolution grid,when use_gfs_ozone = .true. or use_gfs_stratosphere = .true. | |
jcap_cut | ||
use_sp_eqspac | .false. | if .true., then ensemble grid is equal spaced, staggered 1/2 grid unit off poles. if .false., then gaussian grid assumed for ensemble (global only) |
Variable name | Default value | Description |
---|---|---|
BKGERR | Background error related variables | |
vs | 1/1.5 | scale factor for vertical correlation lengths for background error |
nhscrf | 3 | number of horizontal scales for recursive filter |
hzscl(3) | 1, 1, 1 | scale factor for horizontal smoothing, n=1,number of scales (3 for now) |
specifies factor by which to reduce horizontal scales (i.e. 2 would then apply 1/2 of the horizontal scale) | ||
hswgt(3) | 1/3, 1/3, 1/3 | empirical weights to apply to each horizontal scale |
norh | 2 | order of interpolation in smoothing |
ndeg | 4 | degree of smoothing in recursive filters |
noq | 3 | 1/4 of accuracy in compact finite differencing |
bw | 0 | factor in background error calculation |
norsp | 0 | order of interpolation for smooth polar cascade routine default is norsp=0, in which case norh is used with original polar cascade interpolation (global only). |
fstat | .false. | logical to separate f from balance projection |
pert_berr | .false. | logical to turn on random inflation/deflation of background error tuning parameters |
pert_berr_fct | 0 | factor for increasing/decreasing berror parameters, this is multiplied by random number |
bkgv_flowdep | .false. | flag to turn on flow dependence to background error variances |
bkgv_rewgtfct | 0 | factor used to perform flow dependent reweighting of error variances |
bkgv_write | .false. | flag to turn on=.true. /off=.false. generation of binary file with reweighted variances |
fpsproj | .true. | controls full nsig projection to surface pressure |
fut2ps | controls the projection from unbalance T to surface pressure | |
adjustozvar | adjusts ozone variances in the stratosphere based on guess field | |
cwcoveqqcov | sets cw Bcov to be the same as B-cov(q) (presently glb default) |
Variable name | Default value | Description |
---|---|---|
ANBKGERR | Anisotropic background error related variables | |
anisotropic | .false. | if true, then use anisotropic background error covariance |
ancovmdl | 0 | covariance model settings - 0: pt-based, 1: ensemble based |
triad4 | .true. | for 2d variables, if true, use blended triad algorithm |
ifilt_ord | 4 | filter order for anisotropic filters |
npass | 1 | 2×npass = number of factors in background error |
normal | 200 | number of random vectors to use for filter normalization ( if < 0 then slightly slower, but results independent of number of processors) |
binom | .true. | if true, weight correlation lengths of factors using binomial distribution, with shortest scales on outside, longest scales on inside. This can help to produce smoother correlations in the presence of strong anisotropy |
ngauss | 3 | number of Gaussians to add together in each factor |
rgauss | 0 | multipliers on reference aspect tensor for each Gaussian factor |
anhswgt | 1.0 | empirical weights to apply to each gaussian |
an_vs | 1 | scale factor for background error vertical scales (temporary carry over from isotropic inhomogeneous option) |
grid_ratio | 2.0 | ratio of coarse to fine grid in fine grid units |
grid_ratio_p | 0 | ratio of coarse to fine grid in fine grid units for polar patches |
nord_f2a | 4 | order of interpolation for transfer operators between filter grid and analysis grid |
an_flen_u | 1 | coupling parameter for connecting horizontal wind to background error |
an_flen_t | 1 | coupling parameter for connecting grad(potential temperature) to background error |
an_flen_z | 1 | coupling parameter for connecting grad(terrain) to background error |
rtma_subdomain_option | .false. | if true, then call alternative code which calls recursive filter directly from subdomain mode, bypassing transition to/from horizontal slabs. This is mainly to improve efficiency for 2d rtma analysis. at the moment, this only works for twodvar_regional=.true. rtma_subdomain_option will be forced to false when twodvar_regional=.false. |
lreadnorm | .false. | if true, then read normalization from fixed files |
nsmooth | 0 | number of 1-2-1 smoothing passes before and after background error application |
nsmooth_shapiro | 0 | number of 2nd moment preserving (shapiro) smoothing passes before and after background error application. |
NOTE: default for nsmooth and nsmooth_shapiro is 0. | ||
if both are > 0, then nsmooth will be forced to zero. | ||
afact0 | 0.0 | anistropy effect parameter, the range must be in 0.0-1.0. |
covmap | .false. | if true, covariance map would be drawn |
Variable name | Default value | Description |
---|---|---|
JCOPTS | Constraint term in cost function (Jc) | |
ljcdfi | .false. | if .false., uses original formulation based on wind, temp, and ps tends |
when .t. uses digital filter initialization of increments (4dvar) | ||
alphajc | 10.0 | parameter for digital filter |
switch_on_derivatives | .false., … | if true, then compute horizontal derivatives of all state variables (to be used eventually for time derivatives, dynamic constraints and observation forward models that need horizontal derivatives) |
tendsflag | .false. | if true, compute time tendencies |
ljcpdry | .false. | when .t. uses dry pressure constraint on increment |
bamp_jcpdry | 0.0 | parameter for pdry_jc |
eps_eer | -1.0 | Errico-Ehrendofer parameter for q-term in energy norm |
ljc4tlevs | .false. | when true and in 4D mode, apply any weak constraints over all time levels instead of just at a single time |
Variable name | Default value | Description |
---|---|---|
STRONGOPTS | Strong dynamic constraint | |
reg_tlnmc_type | 1 | =1 for 1st version of regional strong constraint |
=2 for 2nd version of regional strong constraint | ||
tlnmc_option | 0 | integer flag for strong constraint (various capabilities for hybrid): |
=0: no TLNMC | ||
=1: TLNMC for 3DVAR mode | ||
=2: TLNMC on total increment for single time level only (for 3D EnVar) or if 4D EnVar mode, TLNMC applied | ||
to increment in center of window | ||
=3: TLNMC on total increment over all time levels (if in 4D EnVar mode) | ||
=4: TLNMC on static contribution to increment ONLY for any EnVar mode | ||
nstrong | 0 | if > 0, then number of iterations of implicit normal mode initialization to apply for each inner loop iteration |
period_max | 1000000.0 | cutoff period for gravity waves included in implicit normal mode initialization (units = hours) |
period_width | 1.0 | defines width of transition zone from included to excluded gravity waves |
nvmodes_keep | 0 | number of vertical modes to use in implicit normal mode initialization |
baldiag_full | .false. | flag to toggle balance diagnostics for the full fields |
baldiag_inc | .false. | flag to toggle balance diagnostics for the analysis increment |
Variable name | Default value | Description |
---|---|---|
OBSQC | Observation quality control variables | |
Parameters used for gross error checks are set in file convinfo (ermin, ermax, | ||
ratio) Parameters below used for nonlinear (variational) quality control | ||
dfact | 0 | factor for duplicate observation at same location for conventional data |
dfact1 | 3.0 | time factor for duplicate observation at same location for conventional data |
erradar_inflate | 1 | radar error inflation factor |
tdrerr_inflate | .false. | logical for tdr obs error inflation |
tdrgross_fact | 1 | factor applied to tdr gross error |
oberrflg | .false. | logical for reading in new observation error table (if set to true) |
vadfile | ’none’ | character(10) variable holding name of VAD wind bufr file |
noiqc | .false. | logical flag to bypass OI QC (if set to true) |
c_varqc | 1 | constant number to control variance qc turning on speed |
blacklst | .false. | logical for reading in raob blacklist (if set to true) |
use_poq7 | .false. | Logical to toggle accept (.true.) or reject (.false.) SBUV/2 ozone observations flagged with profile ozone quality mark |
hilbert_curve | .false. | option for hilbert-curve based cross-validation. works only with twodvar_regional=.true. |
tcp_refps | 1000.0 | reference pressure for tcps oberr calculation (mb) |
tcp_width | 50.0 | parameter for tcps oberr inflation (width, mb) |
tcp_ermin | 0.75 | parameter for tcps oberr inflation (minimum oberr, mb) |
tcp_ermax | 5.0 | parameter for tcps oberr inflation (maximum oberr, mb) |
qc_noirjaco3 | .false. | controls whether to use O3 Jac from IR instruments |
qc_noirjaco3_pole | .false. | controls wheter to use O3 Jac from IR instruments near poles |
qc_satwnds | .true. | allow bypass sat-winds qc normally removing lots of mid-tropo obs |
njqc | ||
vqc | ||
aircraft_t_bc_pof | .false. | logical for aircraft temperature bias correction, pof is used for predictor |
aircraft_t_bc | .false. | logical for aircraft temperature bias correction |
aircraft_t_bc_ext | .false. | logical for reading aircraft temperature bias correction from external file |
buddycheck_t | .false. | When true, run buddy check algorithm on temperature observations |
buddydiag_save | .false. | When true, output files containing buddy check QC info for all obs run through the buddy check |
biaspredt | 1 | berror var for temperature bias correction coefficients |
upd_aircraft | .true. | indicator if update bias at 06Z & 18Z |
cleanup_tail | .false. | logical to remove tail number no longer used |
Variable name | Default value | Description |
---|---|---|
OBS_INPUT | Controls input data | |
dfile | ' ' |
input observation file name |
dtype | ' ' |
observation type |
dplat | ' ' |
satellite (platform) id (for satellite data) |
dsis | ' ' |
sensor/instrument/satellite flag from satinfo files |
dthin | ' ' |
satellite group |
dval | ' ' |
relative value of each profile within group relative weight for observation = dval/sum(dval) within grid box |
dmesh(max(dthin)) | thinning mesh for each group | |
mesh size (km) for radiance thinning grid (used in satthin) | ||
dsfcalc | ' ' |
specifies method to determine surface fields within a FOV. when equal to one, integrate model fields over FOV. when not one, bilinearly interpolate model fields to FOV center. |
time_window_max | 3 | upper limit on time window for all input data |
ext_sonde | .false. | logical for extended forward model on sonde data |
l_foreaft_thin | .false. | separate TDR fore/aft scan for thinning |
Variable name | Default value | Description |
---|---|---|
SINGLEOB_TEST | Single observation test case setup | |
maginnov | 1 | magnitude of innovation for one observation |
magoberr | 1 | magnitude of observational error |
oneob_type | ' ' |
observation type (t, u, v, etc.) |
oblat | 0 | observation latitude |
oblon | 0 | observation longitude |
obpres | 1000.0 | observation pressure (hPa) |
obdattim | 2000010100 | observation date (YYYYMMDDHH) |
obhourset | 0 | observation delta time from analysis time |
pctswitch | .false. | if .true. innovation & oberr are relative (%) of background value (level ozone only) |
obchan | 0 | if > 0, selects the channel number. If <= zero, it will use all channels that pass qc in setuprad. |
Variable name | Default value | Description |
---|---|---|
SUPEROB_RADAR | Level 2 bufr file to radar wind superobs | |
del_azimuth | 5.0 | azimuth range for superob box (default 5 degrees) |
del_elev | 0.25 | elevation angle range for superob box (default .05 degrees) |
del_range | 5000.0 | radial range for superob box (default 5 km) |
del_time | 0.5 | 1/2 time range for superob box (default .5 hours) |
elev_angle_max | 5.0 | max elevation angle (default of 5 deg) minnum 50 minimum number of samples needed to make a superob |
range_max | 100000.0 | max radial range in meters to use in constructing superobs (default 100km) |
l2superob_only | .false. | if true, then process level 2 data creating superobs, then quit. (added for easier retrospective testing, since level 2 bufr files are very large and hard to work with) |
Variable name | Default value | Description |
---|---|---|
LAG_DATA| | Lagrangian data assimilation related variables | |
lag_accur | 1.0e-6 | Accuracy used to decide whether or not a balloon is on the grid |
infile_lag | inistate_lag.dat | File containing the initial position of the balloon |
lag_stepduration | 900.0 | Duration of one time step for the propagation model |
lag_nmax_bal | 1000 | Maximum number of balloons at starting time |
lag_vorcore_stderr_a | 2.0e3 | Observation error for vorcore balloon |
lag_vorcore_stderr_b | 0.0 | error = b + a*timestep(in hours) |
Variable name | Default value | Description |
---|---|---|
HYBRID_ENSEMBLE | Parameters for use with hybrid ensemble option | |
l_hyb_ens | .false. | if true, then turn on hybrid ensemble option |
uv_hyb_ens | .false. | if true, then ensemble perturbation wind variables are u,v, otherwise, ensemble |
perturbation wind variables are stream, pot. Functions. | ||
q_hyb_ens | .false. | if true, then use specific humidity ensemble perturbations, otherwise, use relative humidity |
aniso_a_en | .false. | if true, then use anisotropic localization of hybrid ensemble control variable a_en. |
generate_ens | .true. | if true, then generate internal ensemble based on existing background error |
n_ens | 0 | number of ensemble members. |
nlon_ens | 0 | number of longitudes on ensemble grid (may be different from analysis grid nlon) |
nlat_ens | 0 | number of latitudes on ensemble grid (may be different from analysis grid nlat) |
jcap_ens | 0 | for global spectral model, spectral truncation |
pseudo_hybens | .false. | if true, turn on pseudo ensemble hybrid for HWRF |
merge_two_grid_ensperts | .false. | if true, merge ensemble perturbations from two forecast domains to analysis domain (one way |
to deal with hybrid DA for HWRF moving nest) | ||
regional_ensemble_option | 0 | integer, used to select type of ensemble to read in for regional application. Currently takes |
values from 1 to 4 | ||
=1: use GEFS internally interpolated to ensemble grid. | ||
=2: ensembles are WRF NMM format | ||
=3: ensembles are ARW netcdf format. | ||
=4: ensembles are NEMS NMMB format. | ||
full_ensemble | .false. | if true, first ensemble perturbation on first guess istead of on ens mean |
betaflg | .false. | if true, use vertical weighting on beta1_inv and beta2_inv, for regional |
coef_bw | 0.9 | fraction of weight given to the vertical boundaries when betaflg is true |
pwgtflg | .false. | if true, use vertical integration function on ensemble contribution of Psfc |
jcap_ens_test | 0 | for global spectral model, test spectral truncation (to test dual resolution) |
beta1_inv | 1 | 1/beta1, the default weight given to static background error covariance if (.not. readin_beta) |
0 <= beta1_inv <= 1, tuned for optimal performance | ||
=1, then ensemble information turned off | ||
=0, then static background turned off the weights are applied per vertical level such that : | ||
betas_inv(:) = beta1_inv , vertically varying weights given to static ; | ||
betae_inv(:) = 1 - beta1_inv , vertically varying weights given ensemble derived covariance. | ||
If (readin_beta) then betas_inv and betae_inv are read from a file and beta1_inv is not used. |
Variable name | Default value | Description |
---|---|---|
HYBRID_ENSEMBLE | Parameters for use with hybrid ensemble option | |
s_ens_h | 2828 | homogeneous isotropic horizontal ensemble localization scale (km) |
s_ens_v | 30 | vertical localization scale (grid units for now) s_ens_h, s_ens_v, and beta1_inv are tunable parameters. |
use_gfs_ens | .true. | controls use of global ensemble: .t. use GFS (default); .f. uses user-defined ens |
readin_localization | .false. | flag to read (.true.)external localization information file |
readin_beta | .false. | flag to read (.true.) the vertically varying beta parameters betas_inv and betae_inv from a file. |
eqspace_ensgrid | .false. | if .true., then ensemble grid is equal spaced, staggered 1/2 grid unit off ploes. |
if .false., then gaussian grid assumed for ensemble (global only) | ||
use_localization_grid | .false. | if true, then use extra lower res gaussian grid for horizontal localization (global runs only–allows |
possiblity for non-gaussian ensemble grid) | ||
grid_ratio_ens | 1 | for regional runs, ratio of ensemble grid resolution to analysis grid resolution |
default value = 1 (dual resolution off) | ||
oz_univ_static | .false. | if true, decouple ozone from other variables and defaults to static B (ozone only) |
write_ens_sprd | .false. | writing global ensemble spread in byte addressable format for plotting with grads |
enspreproc | .false. | flag to read(.true.) pre-processed ensemble data already |
i_en_perts_io | 0 | flag to read in ensemble perturbations in ensemble grid. |
This is to speed up RAP/HRRR hybrid runs because the same ensemble perturbations are used in 6 cycles | ||
=0: No ensemble perturbations IO (default) | ||
=2: skip get_gefs_for_regional and read in ensemble perturbations from saved files. | ||
l_ens_in_diff_time | .false. | if use ensembles that are available at different time from analysis time. |
=false: only ensembles available at analysis time can be used for hybrid. (default) | ||
=true: ensembles available time can be different from analysis time in hybrid analysis | ||
ensemble_path |
Variable name | Default value | Description |
---|---|---|
rapidrefresh_cldsurf | Options for cloud analysis and surface enhancement for RR application | |
dfi_radar_latent_heat_time_period | 30.0 | DFI forward integration window in minutes |
metar_impact_radius | 10.0 | metar cloud observation impact radius in grid number |
metar_impact_radius_lowCloud | 4.0 | impact radius for METAR cloud observation that indicate low cloud base |
l_gsd_terrain_match_surfTobs | .false. | if .true., GSD terrain match for surface temperature observation |
l_sfcobserror_ramp_t | .false. | namelist logical for adjusting surface temperature observation error |
l_sfcobserror_ramp_q | .false. | namelist logical for adjusting surface moisture observation error |
l_PBL_pseudo_SurfobsT | .false. | if .true. produce pseudo-obs in PBL layer based on surface obs T |
l_PBL_pseudo_SurfobsQ | .false. | if .true. produce pseudo-obs in PBL layer based on surface obs Q |
l_PBL_pseudo_SurfobsUV | .false. | if .true. produce pseudo-obs in PBL layer based on surface obs UV |
pblH_ration | 0.75 | percent of the PBL height within which to add pseudo-obs |
pps_press_incr | 30hPa | pressure increase for each additional pseudo-obs on top of previous level |
l_gsd_limit_ocean_q | .false. | if .true. do GSD limitation of Q over ocean |
l_pw_hgt_adjust | .false. | if .true. do GSD PW adjustment for model vs. obs station height |
l_limit_pw_innov | .false. | if .true. do GSD limitation of PW obs |
max_innov_pct | 0.1 | sets limit of PW ob to a percent of the background value (0-1) |
l_cleanSnow_WarmTs | .false. | if .true. do GSD limitation of using retrieved snow over warn area |
(Ts > r_cleanSnow_WarmTs_threshold) | ||
l_conserve_thetaV | .false. | if .true. conserve thetaV during moisture adjustment in cloud analysis |
r_cleanSnow_WarmTs_threshold | 8.0 | threshold for using retrieved snow over warn area |
i_conserve_thetaV_iternum | 3 | iteration number for conserving thetaV during moisture adjustment |
l_gsd_soilTQ_nudge | .false. | if .true. do GSD GOES cloud building |
l_cld_bld | .false. | if .true. do GSD soil T and Q nudging based on the lowest t analysis increment |
cld_bld_hgt | 1200m | sets limit below which GOES cloud building occurs |
build_cloud_frac_p | 0.95 | sets the threshold for building clouds from satellite |
clear_cloud_frac_p | 0.1 | sets the threshold for clearing clouds from satellite |
nesdis_npts_rad | 1 | NESDIS cloud product impact radiu (grid points) |
iclean_hydro_withRef | 1 | if =1, then clean hydrometeors if the grid point has no echo and maxref=0 |
iclean_hydro_withRef_allcol | 0 | if =1, then clean whole column hydrometeors if the observed max ref =0 and satellite cloud shows clean |
Variable name | Default value | Description |
---|---|---|
rapidrefresh_cldsurf | Options for cloud analysis and surface enhancement for RR application | |
l_use_2mq4b | 0 | background used for calculate surface moisture observation innovation |
=0 Use Q from the 1st model level. (default) | ||
=1 use 2m Q as part of background | ||
i_use_2mt4b | 0 | background used for calculate surface temperature observation innovation |
=0 Use T from the 1st model level. (default) | ||
=1 use 2m T as part of background | ||
i_gsdcldanal_type | 0 | options for how GSD cloud analysis should be conducted |
=0. no cloud analysis (default) | ||
=1. cloud analysis after var analysis | ||
=5. skip cloud analysis and NETCDF file update | ||
i_gsdsfc_uselist | 0 | options for how to use surface observation use or rejection list |
=0 . EMC method (default) | ||
=1 . GSD method | ||
i_lightpcp | 0 | options for how to deal with light precipitation |
=0 . don’t add light precipitation (default) | ||
=1 . add light precipitation in warm section | ||
i_sfct_gross | 0 | if use extended threshold for surface T gross check |
=0 use threshold from convinfo (default) | ||
=1 for cold surface, threshold for gross check is enlarged to bring more | ||
large negative innovation into analysis. |
Variable name | Default value | Description |
---|---|---|
CHEM | Chemistry data assimilation | |
berror_chem | .false. | if berror file is supplied for chemistry |
oneobtest_chem | .false. | single observation test for chemistry |
maginnov_chem | 30.0 | if oneobtest_chem=T magnitude of innovation for chemistry |
magoberr_chem | 2.0 | if oneobtest_chem=T magnitude of observation error for chemistry |
oneob_type_chem | pm2_5 | if oneobtest_chem=T type of chemical observation |
oblat_chem | 45.0 | if oneobtest_chem=T latitude of the observation |
oblon_chem | 270.0 | if oneobtest_chem=T longitude of the observation |
obpres_chem | 1000.0 | if oneobtest_chem=T pressure of the observation |
diag_incr | .false. | if user wishes to output to a binary file increment |
elev_tolerance | 500.0 | for surface chemical observation sometimes elevation (elev_obs) of the |
measurement is available (sometimes not). | ||
tunable_error | 0.5 | tuning parameter to specify representativeness error for in-situ observations |
in_fname | cmaq_input.bin | name of background file for cmaq |
out_fname | cmaq_output.bin | name analysis file for cmaq |
incr_fname | chem_increment.bin | if diag_incr=T name of the binary dump for pm2_5 |
laeroana_gocart | .false. | when true, do chem analysis with wrfchem and modis |
l_aoderr_table | ||
aod_qa_limit | ||
luse_deepblue | ||
aero_ratios | ||
wrf_pm2_5 |