gamit配套的平差程序：globk的使用幫助信息，如下：
[zhao@zhaoZhongHai ~]$ globk
+++++++++++++++++++++++++++++++++++++++++++++
+ GLOBAL KALMAN FILTER Ver 5.33 +
+++++++++++++++++++++++++++++++++++++++++++++
GLOBK: Global Kalman filter analysis program
GLOBK Ver 5.32: Global Kalman filter analysis program.
Runtsring:
% GLOBK <std out> <print file> <log file> <exper. list> <command file> <OPTION>
where <std out> is a numerical value (if 6 is typed then output will be
sent to current window, any other numerical value will
send output to a file fort.nn)
<print file> is the name for the output print file with the
solution in it. If the print file already exists, then
the new solution will be appended to it.
<log file> is a log file which contains the running time for the
program and the pre-fit chi**2 value for each input
covariance matrix file. If the log file already exists,
then the new solution will be appended to it.
<exper. list> is a list of covariance files to process. (This
list is most easily generated with re-direction of
and ls command e.g., % ls ../gotex/g*.glc > gotex.gdl).
Any file name may be used but convention is to end it
with .gdl.
<command file> is a list of commands (see specific GLOBK
documentation file). The command file controls the
options in the program.
<OPTION> string that can appear at the starts of command file lines
which will be executed. (Normally these lines are ignored
since they do not start with a blank. (New Ver 5.08)
See also glist, glred, glorg, htoglb, plot, ensum, blsum, bcsum.
See bottom of help for basic commands. Newer features are added at the
top of this file (Add 171230).
GLOBK Version 5.33: 200222
--------------------------
APP_PTIDE command modified to allow explicit selection of the mean or
secular pole to used. IERS10 is the current mean pole and IERS20 is the
linear secular pole to be used for ITRF2020.
Updated routine to carry ground antenna azimuth information.
GLOBK Version 5.32: 190509
--------------------------
Increased number of parameters to handle 40000 parameters. curr_com_ver
updated and hence glorg can not be used with old com_files. Modified
mar_ut1 behavior to decouple UT1 from LOD. Addded frame stabilization to
glbak when glorg invoked in globk run. Implemented changes to accommadate
the new GAMIT ECOMC model. This change does affect the apr_svs and apr_rad
commnd ordering of the radiaton parameters. In particular, the constant
Z and X terms that were never. The place holders (0 nomrally) for these
paraneters should removed from command files. (kf/includes/glb_hdr_def.h
for comments un SV orbits codes). This change also affects the format of
globk satellite ephemeris files (SVS files).
Introduced new wild card features that allow forms such as Week<WWWW><D>_B2UM.GLX
to be used where WWWW and D will be replaced by GPS week and day of week
during run. Poptions include
<WWWW> GPS week
<D> GPS day of week
<Y4> 4-digit year
<Y2> 2-digit year
<Y1> 1-digit year (arc)
<DOY> Day of year
<MM> Month of year
<DD> Day of month
<HR> Hour of day
<MIN> Minute of hour
<SEC> Seconds of minute
NOTE: Only one option per-extent can be used (i.e., once a xxx_file
command using this option has been issued, all subsequent commands will
not in invoked. The - and @ wild card subsitions are made before the
wild-date replacement.
Back solutions were extended to allow frame realization for each h-file
in the back solution. The out_glb command in the back solutiokn was
implemented to allow frame realized binary h-files to be created during
the back solution.
GLOBK Version 5.31: 180402
--------------------------
Updated to support GNSS with new satallite naming scheme. For GPS
the default is to retain the PRN_<pn><sv> form where pn is 2-digit
PRN, <sv> is 2-digit satellite vehicle number. The new scheme is
<sys><sv3>_<pn> where sys is G--GPS, R-Glonass, E-Galileo, C-Beidou
and the sv number has 3-digits. The "use_prnn N" command will invoke
using the new scheme for all satellites.
New command
@USE_PRNN <Y/N>
where option Y (default) will use the legacy PRN_XXXX form for GPS satellites
(Other GNSS system use the new format). Option N will convert the GPS
names to the new format. Command must be given near top of command file.
GLOBK Version 5.30: 161128
--------------------------
Incread max_rn from 65536 to 262144, max_ss from 4096 to 8192. GLORG and
curr_glb_com_ver also increases.
GLOBK Version 5.29: 150825
--------------------------
Added the satellite vechicle number to the PRN number so that different
satellites using the same PRN can be distinquished. Information based
on satellite information in svnav.dat at the time the original ascii h-files
were generated.
GLOBK Version 5.28: 150526
--------------------------
Modified memory allocation so that >20,000 in the state vectors of the
solution and input binary h-files can be used analyzed. Increased
number of digits in output parameter number to handle more than 9,999 parameters
GLOBK Version 5.27: 140809
--------------------------
Size increase to number of parameters and process noise values to handle
GAGE velocity analysis. Added arrays for apriori coordinate file sigmas.
Added -PER (case sensitive) to apriori file name will cause the periodic
terms not to be used.
GLOBK Version 5.26: 140405
--------------------------
Added +OPT or -OPT option as first argument of APP_PTIDE command to either
add or remove (if already applied) the ocean pole tide contribution. Option
reqires file opoleloadcoefcmcor.txt with ocean pole tide factor grid be
available in ~/gg/tables. Also added ttracking of atmospheric and 2nd order
ionospheric delay models in h-file.
GLOBK Version 5.25: 140327
--------------------------
Added passing of earth radiation and antenna thrust models so that gfiles
are correctly saved from glbak and globk.
GLOBK Version 5.24: 131106
--------------------------
Increased maximum number of renames allowed (max_rn). The com_files from this
version are not compatiable with earlier versions of globk (glsave.5.23 is
compatable).
GLOBK Version 5.23: 130716
--------------------------
Added white noise process noise to mar_scale command so that the process
noise does not depend on time difference. Allows processing of hfiles on
the same day with different scales (ITRF05 versus ITRF08 scales). com_file
version change.
GLOBK Version 5.22: 130419
--------------------------
Added command APP_MODL which allows load models to be applied or removed
from analysis if they have been applied in gamit processing or added with
hfupd. Current load types are ATMLOAD and HYDROLD. Using a - infront of
name will remove the correction is alread applied. Example:
app_modl -atmload -hydrold
would remove load corrections already applied in the solution.
Added RESET command to eq_file commands which resets all the renames in hfiles
back to _GPS names. Care should be used with combined files where multiple
renamed sites could end up with the same name (positions will be combined
in globk run). Using the name RESET for the eq_file will do the name thing.
GLOBK Version 5.21: 120411
--------------------------
No version number change. Small change to MUL_PMU command to allow
START <offset (days)> as option. Offset in normally 0.5 for 24-hr sessions.
GLOBK Version 5.21: 110815
--------------------------
Added feature which allows multiple OPTIONS to be used by separated by
+ signs in the runstring (i.e., OPT1+OPT2+OPT3 would interpret all lines
that start with either OPT1, OPT2 or OPT3. Lines are processed in the order
of the entries (i.e., Lines starting with OPT1OPT2 will be processed).
GLOBK Version 5.20: 110512
--------------------------
Implemented IERS2010 mean pole postition for pole tide correction. Update
is applied when app_ptide command is used. Caution is needed for SINEX files
since the status of the pole-tide is not reported there. (Best is not to
use app_ptide unless status is known).
GLOBK Version 5.19: 101015
--------------------------
Passed full antex file name into files so that it is available in sinex
files (com file version changed)
GLOBK Version 5.18: 090930
--------------------------
Added BREAK Command to commands in earth quake files, and added feature
that allows multiple earthquake files to be specified
Information about the BREAK command is given in the EQ_FILE section below.
NOTE: The comfiles from this verions of globk are not compatable with
earlier version and so the new glorg/glsave/glout versions can not be used
with older solutions. Keeping copies of the old programs will allow the
older files to be continued to be used.
- Added feature to allow multiple earthquake files. Using the name
NONE or RESET (case sensitive) will reset the list of files to zero.
- Added feature to use NONE or RESET for aproiri coordinate files to
reset this list to no entries. (Useful with OPTION feature in runstring).
- Added RENM to output options which will cause the full rename list
to output the log files or screen. Useful when the break command is used
to get the names of the sites generated by the breaks.
- Added negative value to free_log to turn off logs (useful when constructing
monthly solutions far from earthquake.
NOTES ON HFILE DEPENDENT RENAMES:
Care needs to be exercised when hfile dependent renames are blended
with other non-specific renames to ensure that the desired renames
happens. Renames are processed in globk by first sorting all the
renames by the 4-char code of the original site name and then for
each site, sorting in time by the beginning time. Open ended renames
(i.e., one with end times of 2100 1 1) have their end times adjusted
to match the start time of the next rename. The hfile specification
is ignored in setting these end times. The renames are then applied
in time order according to the file name and start and end times.
Three can be problems when a site is being completely renamed (to a
different 4-char code) when there are later renames (non-hfile specific)
involving the original 4-char code are in the eq-file. One solution
to this is specify an end date for the rename before 2100; doing this
will stop the automatic changing of the end date. When 4-char codes
are not changed and the original name in the rename uses only 4
characters then the start time of the hfile dependent may need to
changed to ensure that is applied after other non-hfile dependent
renames.
GLOBK Version 5.17: 080724:
---------------------------
Added use of +REFERENCE_FRAME in apriori coordinate files to allow the
name of the reference frame to be passed to globk output file. Line must
start with +REFERENCE_FRAME e.g.
+REFERENCE_FRAME SNARF 1.0
Added FREE_LOG <sigma (m)> command to allow the sigma of the log estimates
to be increased. Default if sigma not passed is 0.1 m. Used during muliple
combinations so that earthquake log terms are not constrained multiple times.
Ver 5.18: Addition: setting the sigma to a negative value will turn off the
log estimates even when specified in earthquake file.
GLOBK Version 5.16: 080416:
---------------------------
Added feature that all file names can have wild cards in the name.
GLOBK Version 5.15: 080129:
---------------------------
Updated to allow orbits to be rotated when different nutation models
have been used in the gamit runs.
GLOBK Version 5.14: 080103:
---------------------------
Updated to pass nutation and gravity field information through h-files
into solution (Release 10.33 and later)
GLOBK Version 5.13: 070824
--------------------------
Added feature to allow solutions and saved binary hfiles to be
referenced to the center epoch of data being combined. The new
feature is invoked with the MIDP output option in the CRT_OPT,
PRT_OPT or ORG_OPT commands. If OUT_GLB is included in the same
globk command file and either the PRT_OPT or CRT_OPT has MIDP set,
the output binary h-file will also be referenced to the center
epoch. When GLSAVE is run stand-alone, the -M option will generate
a binary hfile referenced to the central epoch. If the MIDP
option was used in the globk command file, the -N option in GLSAVE
will generate a binary hfile referred to the end epoch (the
conventional output for GLOBK).
Cautions: The MIDP option should not be used if large stochastic
variations in station positions are present because the estimated
stochastic variation will be applied incorrectly to the midpoint
of the data span.
The polar motion/UT1 (PMU) estimates are always output at the
end epoch because they do have large stochastic variations. To
preserve the PMU estimates, the MUL_PMU command should be used.
When MIDP hfiles are saved with only a single PMU value estimated,
it is saved as a multiday PMU estimate with 1 value. To recover
the estimates in later globk runs, MUL_PMU representation of the
PMU estimates must be used and the table of values must contain
the epoch of PMU estimate in the Hfile. Converting the binary
hfile to SINEX will shows the estimates and their epoch in the
binary hfile.
GLOBK Version 5.12: 060518
--------------------------
Increased dimensions for stations and parameters. Added fatal
for sourcing a file inside a file being sources (See SOURCE command)
MOD 050524:
-----------
Added DEL_SCRA YES command to delete scratch files (com, sol and srt) at
end of run. NOTE: When this commands is used, glorg, glout, glsave can
not be run after globk. Feature added because of use of wild cards in the
names of com_file, sol_file and srt_file for parallel runs in the one
directory.
GLOBK Version 5.11: 040801
--------------------------
Changed the memory allocation method to 64-bit compatable. Fortran compiler
must support integer*8 in this version.
GLOBK Version 5.10: 040703
--------------------------
Added atmospheric delay estimates into globk. (Needs solve version 10.11
or higher to be useful). Command structure is
Aprori sigma definition
APR_ATM <Site/ALL/COMMON> <Apriori Sigma (m)>
where Site is specific site names, ALL is for all sites, COMMON restricts
site list to just those sites that are common between networks. (Feature
meant to be used when multiple networks on the same day are combined to
a single network).
Apriori Sigma for average zenith delay (m)
MAR_ATM <Site/ALL/COMMON> <RW process noise (m^2/yr)>
Defines Markov process noise for zenith delays (Random walk model with
units m^2/year
GLOBK Version 5.09: 040220
--------------------------
Fixed problem in wild card mar_neu, added SMAR to output optiions (org_opt)
so that the site dependent Markov process noise can be reported in the
org and prt files.
Fixed treatment of satellite antenna offsets so that user input is used
by globk
Changed 'all' return in get_command from 999 to 999999 so there in no
overlap with site numbers.
GLOBK Version 5.08: 030610
--------------------------
- The major change in this version is the addition of logarithmic function
estimation after an earthquake (instead of using process noise to model
postseismic behavior). Feature is invoked in the earthquake file using the
EQ_LOG definition line for an earthquake. (see eq_def beflow). When only
one day of data is processed, the log terms are not estimated. This update
changes the globk common file version so this version of globk will not be
able to read old com_files.
- Introduced the <OPTION> feature in the runstring for globk, glorg, and glred.
- Introduced the use of wild card character (@) in site name specifications
for the use_site, stab_site and mar_neu commands. When this feature is used
as name such as CHIL_@ will match all site names that start with CHIL_.
Fearure is useful for ensuring that all renamed version of site are used (or
not used) and have the same process noise given to them.
- Introduced "decimate" command in globk forward runs. Command must be given
in first group of commands in the command file. Form: decimate N M when every
N'th Hfile will be used starting at the Mth. Useful for quick solutins with
continuous data. Fatal error results if command is out of order (applied to
all commands in the group that must appear before all other commands).
Example
DECIMATE 7 3
would process one in seven files starting the third file in the gdl file.
- Added new type of rename. Site renamed to end in _XPS will not be used when
more than one day of data are processed. When one-day of data processed they
will be used. Feature useful for getting time series of deleted data (glred
run) while not have the site effect a globk forward solution.
- Introduced output of full precision geodetic coordinates (WGS-84 system) and
UTM coordinates. Increased precision by one significant digit of nearly all
outputs. (Positions and velocities now given to 0.01 mm and 0.01 mm/yr).
- The pole tide is now applied to the mean-pole defined by the IERS 2000
conventions. (Update is automatically handled when pole tide is applied).
- Introduced -RS option (realistic sigma) to ensum and enfit. Invoking this
option, generates sigmas that account for correlations in the time series (see
enfit help).
GLOBK Version 5.07: 030129
--------------------------
Most changes made to glorg. Small addition of LOG functions in
extended model (see enfit). Added option to list parameters to
be estimated at start of run.
Changes to PoleTide algorithm: Pole tide correction changed to
be consisent with IERS standards with the reference pole linearly
drifting with time. This correction is automatically applied to
any hfile with the pole tide correction applied.
GLOBK Version 5.06a: 011126:
----------------------------
Small modification to stop sites renamed to end in _XCL from
being renamed during earthquakes (i.e., they will retain the
_XCL name and no be used in the solution).
GLOBK Version 5.06: 010519:
---------------------------
Modifications made to the treatment of the EXTENDED model in apriori
coordinate files. Explicit external statements were added because on
some system the block data was not loaded correctly which causes incorrect
results when extended features were used in glorg apriori files.
GLOBK Version 5.05: 000901:
---------------------------
NOTE: The com_file from this version of globk are not backwards
compatable. You should save copies of glsave and glorg if
you want to read old globk com_files.
Minor enhancements of the program. Features added include:
(a) Renaming a site to end in _XCL will automatically exclude the site
from being used in the globk/glred analyses (equivalent to use_site -site_XCL)
(b) New command added which allows noise to added to the coordinate
components of indivual stations. The format of the new SIG_NEU
command is:
SIG_NEU <Site> [HF code] <NEU Sigmas (m)> [Start date] [End date]
where
<site> is the name of a site with the following card features:
- Full 8-character to match a site exactly (name is the current
name after all renames and earthquakes have been applied.
- Leading charaters of a name and allow sites with those leading
characters will have the noise applied.
- @[string] will match all sites whose name end with string
Examples: AUCK would match all names such as AUCK_GPS or AUCK_GEQ
@_BHI would match all sites renamed to end in _BHI
@ (by itself) would match all sites.
ALL will also match all sites (050429)
[HF code]is optional string that must occurr exactly in the hfile
name for the additional noise to be added
<NEU Sigmas (m)> are the 3 sigmas for Norh, East and Up (m)
[Start date] is an optional calender date of the form YYYY MM DD HR MIN
which sets the earliest time that the additional sigmas will be
applied (check is done on the reference epoch of the hfile)
[End date] is an optional end date of the above form.
All entries which apply to a site are added (in a variance sum).
The sigmas being added are reported to the screen and log files.
Example full command
sig_neu AUCK_GPS nzvt 0.000 0.000 0.1000 2000 02 02 12 0 2000 02 02 12 0
should add noise to the height for the experimen on 2000 02 02 (a
1-minute tolerance is applied to the times and so this should apply
to hfiles which contain nzvt in their names and were from data centered
on 12:00 hrs UTC.
GLOBK Version 5.04: 000404:
--------------------------
Added feature that when the out_sol file command is used in glorg,
and an out_glb command has been given in the globk command file, then
two binary hfiles are written. The loose globk hfile is still written
to the file named in the out_glb command, and the glorg constrained version
of the hfile is written to a file with .CON appended to the name of the
file in the out_glb comamnd.
GLOBK Version 5.03: 991110:
---------------------------
New features added: Non-secular terms added to the apriori coordinate
files. These terms allow periodic, exponential, logarithmic and offset
and rate changes to be introduced into the apriori coordinate files.
The new lines are detected by EXTENDED being the first token on the
line in the aprori coordinate files. The general format of the lines
is:
EXTENDED <Site Name> <Type> <YY MM DD HR MN> <Parameter> <Coefficients for NEU>
where EXTENDED should be preceeded by at least one blank. Line is not
case sensitive,
The following <Types> are allowed
TYPE Format and meaning
EXP Exponential variations: The exponential starts at the time given by
<YY MM DD HR MN> and the <Parameter> is the decay time, in days, for the
exponential (ie. exp[-dtime/Parameter] where dtime is time after
the start time. The coefficients are the amplitudes for North, East and
Up. Units meters. Example:
EXTENDED JPLM_GPS EXP 1992 6 28 0 0 30.0 0.010 0.005 0.00
Results in 30 day exponential with amplitude of 10 mm North, 5 mm East and
zero for the height
LOG Logarithm variations: The logarithmic function is applied after date
<YY MM DD HR MN> and the <Parameter> is time normalization value in
days (ie., log(dtime/Paramter) where dtime is time from the start time.
The coefficients are the amplitudes for North, East and Up. Units meters.
Note: No data should be included that is within the normalization
parameter of the start time (i.e., log(0) = -infinity). Example
EXTENDED JPLM_GPS LOG 1992 6 28 0 0 300.0 0.010 0.005 0.00
Results in 300 day logarithmic with amplitude of 10 mm North, 5 mm East and
zero for the height
PERIODIC periodic variations: Applied to all dates. The <YY MM DD HR MN> is
the zero phase time of the periodic signals and the <Parameter> is
the period in days (i.e., cos(2*pi*dtime/Parameter) where dtime is time
from <YY MM DD HR MN>. The coefficients are paired as the cosine and
sine coefficients for North, East and Up. Units meters. Example
EXTENDED JPLM_GPS Periodic 2000 1 1 0 0 365.25 0.0 0.0 0.0 0.0 0.001 0.009
Results in Annual signal in height with cosine amplitude of 1 mm and
Sine amplitude of 9 mm.
OFFSET Episodic change in position and velocity. Applied after <YY MM DD HR MN>
<Parameter> in this case is not used (value of 0.0 should be given).
The coefficients are paired as offset and rate changes applied after
the start date (i.e., offset + dtime*rate where dtime is time after
start time. Units: offset meters, rate meters/year. Example:
EXTENDED JPLM_GPS Offset 1992 6 28 0 0 0.0 0.003 0.001 0.004 0.004 0.000 0.000
Results in North position and velocity change of 3 mm and 1 mm/yr,
East position and velocity change of 4 mm and 4 mm/yr. No change in height.
NOTES:
(a) Each of the terms given should be unique. If duplicates (in terms of epoch,
type and parameter) are given, the lastest one read will be used.
(b) This extended model is meant for use with time series analysis. The non-secular
terms are saved in any out binary hfile.
(c) The position estimates and adjustment output by globk and glorg include these
terms evaluated at the epoch of the output solution.
(d) Differences between position changes with the eq_file rename feature, extended
offset and hfupd. The eq_file rename changes the "observed" site position in the
input hfile where as the extended offset feature changes the "modeled" position of
the site. As such the signs are opposite to acheive the same effect. To correct
erroneous heights in hfiles, hfupd is now the preferred means. In 5.03 globk,
extended offset should only be used for repeatability runs to account for
coseismic offsets and rate changes.
Other changes at 5.03
---------------------
Post-5.02 globk requires that the gamit leap.sec file be read if polar motion
UT1 values are changed. If this file is not found in the local directory, it
is searched for $HOME/gs/tables. A warning message is now printed if the
leap.sec can not be found. In some cases, not finding the leap.sec file can
cause large rotations to be introduced in the solution.
GLOBK Version 5.02: 990226:
---------------------------
This is version is mainly bug fixs to 5.01 associated with multiday
polar motion, generating IERS table entries, and the itegrated random
walk model used in globk. The Pole-tide code was corrected.
New command:
IRW_MOD <OLD/NEW> -- By using the OLD option the pre-globk 5.02
integrated random walk stochastic model can be invoked. This
old model, couples to position estimates and rate estimates too
strongly and was original introduced in globk to allow direct
integration of rates to get position (primarily for polar motion,
UT1 estimates). The new model is more statistically correct but
still seems not to be complete. The IRW model used with multi-day
polar motion/UT1 estimates seems to be correct.
GLOBK Version 5.01: 981215:
---------------------------
New commands
------------
APP_PTID : Corrects input hfiles for pole tide if needed.
MOD 140405: Added ocean pole tide option as optional first argument
The command syntax is
app_ptid [opt options] <list of h-file codes/ALL>
where [opt options] specifies options
IERS96, IERS10, IERS20 sets the mean or secular pole model to use.
For ITRF2014 IERS10 should be used. For ITRF2020, IERS20 should
be used. IERS96 is no longer used but older h-files may have used
this model.
+OPT/-OPT is optional entry to apply or remove ocean pole tide
Form is +OPT or OPT to apply, -OPT to remove if already applied.
+SEPT/-SEPT is option to apply or remove the solid-earth pole
tide. Removal of pole-tide is not recommended
These models can only be applied or change if the status of the models
is known i.e., when SINEX files are used, the status is not known.
<list of h-file codes> is a list of strings that should appear
in the hfile name for the pole correction to be made. The reason
for this construct is that sinex files do not contain information
about the application of the pole tide, and yet some centers do
apply these correction. The list in this would be those centers
that do not apply the pole tide.
Example: app_tide sio SIO
would apply the correction to hfiles whose names include either
sio or SIO but not to other hfiles.
If ALL is given then the correction is applied to all hfiles that
do not internally indicate the correction has already been made.
..........................................................................
APR_ROT and MAR_ROT : Allows rotations and rotation rates to be estimated
independently of polar motion and UT1. These commands should not normally
be needed, but can be used as a last resort when the rotation inconsistencies
between h-files than can not be removed with the in_pmu command. Examples
could be sinex files that do not have EOP values in them or VLBI or DORIS
files than are inconsistent with other h-files.
Command syntax is:
apr_rot <sig X> <sig Y> <sig Z> <sig Xrate> <sig Yrate> <sig Zrate>
where <sig X> <sig Y> and <sig Z> are the apriori sigmas for the rotations
about the X, Y and Z axes. Units mas.
<sig Xrate> <sig Yrate> and <sig Zrate> are the apriori sigmas for
the rotation rates about the same axes. Units mas/yr.
mar_rot <RW X> <RW Y> <RW Z> <IRW X> <IRW Y> <IRW Z> <WN X> <WN Y> <WN Z>
where <RW X> <RW Y> and <RW Z> are the random walk process noise for the
rotation values. Units mas**2/yr.
<IRW X> <IRW Y> and <IRW Z> are the integrated random walk process
noise for the rotations (ie. IRW in rotation, and random walk in rotation
rate). Units: (mas/yr)**2/yr.
and
<WN X> <WN Y> and <WN Z> white noise processes for the rotations, i.e.
these variances are directly applied to the process noise independent
of the time separation of the points. Units mas**2.
GLOBK Version 5.00: 981020:
---------------------------
MUL_PMU : Sets up estimation of multi-day polar motion/UT1 estimates
The command syntax is
mul_pmu <number> <spacing> <start> <Option>
where <number> is the number of entries to introduce
<spacing> is the spacing in days between the values
<start> is the start epoch for the values which can be entered in
in a number of different ways:
(a) Actual date given as year month day hr min
(b) GPSW <offset> where the start of the GPS week containing
the hfiles will be used, offset by the value given in
<offset (days)> e.g, GPSW 0.5 would start the values at
12:00 hrs UTC on the first day of the GPS week
(c) START <offset> where START is JD of start of first h-file
and offset (days) is a applied from the start time of that
h-file. Since most files are 0-24 hrs GPST, the offset
is normally 0.5 days.
<option> is an option on treating the values with in the multiday
estimates. The options supported are:
(a) IND -- Treat each day independently i.e., the apr_wob
values are applied to each day and the markov
process is not used to correlate the days
(b) NOUT -- Ignore the UT1 entries in the input hfiles.
(c) WARN -- Print warning messages for PMU epochs that do
not fall on the mul_pmu boundaries
(d) PROP -- Propagate the mul-pmu block forward if a hfile
time falls outside the current bounds.
The apr_wob, mar_wob, apr_ut1 and mar_ut1 commands are used to set the
statistics of the values. NOTE: In version 5.0 it is not possible to
have an integrated random walk process noise model that has zero variance
at the first epoch (Specifically, UT1 can not be given zero variance on
the first day. This will be corrected in future versions).
..........................................................................
EP_TOL : Allows uses to the time tolerances for
(a) how closely the epochs of the EOP estimates need to match
the assigned in mul_pmu (Default 0.01 days)
(b) how closely ephemeris elments need to match in time (see
make_svs and svs_file below). (Default 0.1 days)
Command syntax
ep_tol <eop> <svs> ! Units are days
If a value of -1 is given for either than the default is used.
..........................................................................
APR_SVAN : Allows user to set the apriori standard deviations for the
satellite antenna offsets. (These may also be entered in
apr_svs command).
Command syntax
apr_svan <prn/all> <sig X> <sig Y> <sig Z>
where <prn/all> can be a prn name (e.g. PRN_01), or ALL to apply to all
satellites
<sig X> <sig Y> <sig Z> are the apriori sigmas in meters.
F can be used to fix the value.
Any values not given are assumed zero (i.e., no further
constraints will be applied.
The appending or prepending A to a value will cause all
remaining entries in the list to be set to this value.
..........................................................................
MAR_SVAN : Allows the user to set the random walk process noise on the
antenna offset estimates. If these values are not specified
the antenna offsets are "averaged" over all input values.
Command syntax
mar_svan <prn/all> <mar X> <mar Y> <mar Z>
where <prn/all> can be a prn name (e.g. PRN_01), or ALL to apply to all
satellites
<mar X> <mar Y> <mar Z> are the process noise values in m**2/yr.
Version 5.32 and above: Negative value of the process nouse will make
the antenna offset estimates white noise, with the apr_svant value applied
each day.
..........................................................................
VAL_SVAN : Allows user to change the apriori values of the antenna offsets
Command syntax
val_svan <prn/all> <X> <Y> <Z>
where <prn/all> can be a prn name (e.g. PRN_01), or ALL to apply to all
satellites
<X> <Y> <Z> are the new antenna offsets in meters. If DEF is
used for a numeric value then the default value (ie.
the value used in GAMIT will be retained)
e.g. val_svan prn_13 def def 2.0
would change just the Z component to 2.0 meters.
..........................................................................
Modifications to existing commands
----------------------------------
All Commands:
Comments may be added to the end of all commands by using the ! or #
character in the line. This feature used to in general work but
now has been explicitly coded e.g.,
apr_neu bad_site 100 100 100 0 0 0 ! This site is bad
is processed as
apr_neu bad_site 100 100 100 0 0 0
APR_SVS and MAR_SVS
-------------------
Additional flexibility has been added to the input for these commands.
The allow the statistics of the satellite orbital elements, radiation
parameters, and satellite antenna offsets. The basic format remains
the same but extended by 6 (making 23 in all) arguments for the new
radiation parameters and antenna offsets. However to aviod entering
all these values, the following options are allowed:
(a) Numeric value with R pre- or appended to enter radiation parameters
values that have not been enter yet.
(b) Numeric value with A pre- or appended to enter satellite antenna
statistics that have not been entered.
When this form is used all entries that have not been entered will be
set the numeric value.
Values not entered will be set to zero.
Examples:
* X Y Z Xd Yd Zd Dir Others SvAnt
apr_svs prn_09 10 10 10 1 1 1 1 0.01R 5A
would set the apriori sigma for the direct radiation parmater to 100%,
all other radiation parameters are 1%. The antenna offset parameters
would all be set to 5 meters. If the 5A was not given, then the antenna
offset would be zero (thus not estimated in globk).
GLOBK Version 4.17 Added features: 980518
----------------------------------
Two new features were added and several output features were changed.
There have also been signficant changes in GLORG which are explained
in the GLORG help file.
Changed output:
(1) The use of rename and earthquake entries are now tracked and only
those that are actually used in the solution are output to the prt
and org files. (The numbering of entries reflects their absolute
number in the list given be the user.
(2) An additional argument has been added to the gdl files to allow
the diagonal of input covariance matrices to be scaling. The reading
of the gdl file has also been made more flexible. The general
format of the gdl file is now:
[glb file name] <scale> <diagonal scale> <+>
where [glb file name] is the binary h-file name where lines starting
with * or # are ignored
<scale> is an optional scale for the whole covariance matrix
<diagonal scale> an optional scale for the diagonal of the
covariance matrix in parts-per-million (ppm) to be
added to 1.0. <scale> must be given to if <diagonal scale>
is to be used. The diagonal scale is often useful for
unstable sinex files which have been deconstrained.
<+> is an optional argument for GLRED which will cause the
next binary hfile to be combined with this one (and
so on until a line is encounter without a + at the
end of the line. The scale factors no longer need
be given for the + sign to be interpretted by GLRED.
A typical gdl file line may now look like
h980126_stan.glx 3.2 2 +
which would scale the covariance matrix by 3.2 and scale the diagonal
by (1.0+2*1.d-6).
(3) The prefit chi**2/f values written to the log and status files are
now saved for the forward and backward runs and output when the
GDL prt option in specified. If a back solution is not run, or
an experiment was not used in the back solution because it was
rejected in the forward run, its chi**2/f will be -1.
The new form of the output when the GDL option is specified in the
prt_opt or org_opt commands now looks like.
EXPERIMENT LIST from cont.srt
# Name SCALE Diag PPM Forw Chi2 Back Chi2 Status
1 /users/tah/SIO_GLX/H98/H980419_SIO.GLX 1.000 .000 1.134 .809 USED
2 /users/tah/SIO_GLX/H98/H980420_SIO.GLX 1.000 .000 2.113 2.104 USED
3 /users/tah/SIO_GLX/H98/H980421_SIO.GLX 1.000 .000 .717 2.338 USED
4 /users/tah/SIO_GLX/H98/H980422_SIO.GLX 1.000 1.000 3.421 1.471 USED
5 /users/tah/SIO_GLX/H98/H980423_SIO.GLX 4.000 .000 2.083 1.307 USED
6 /users/tah/SIO_GLX/H98/H980424_SIO.GLX 2.000 3.000 .693 .245 USED
7 /users/tah/SIO_GLX/H98/H980425_SIO.GLX 1.000 .000 2.433 .008 USED
where <SCALE> is the overall scaling of the covariance matrix
<Diag PPM> is the diagonal scaling in parts-per-million difference from 1.0
<Forw Chi2> is the forward solution Pre-fit Chi**2/degree-of-freedom
<Back Chi2> is the backward solution Pre-fit Chi**2/degree-of-freedom
<Status> indicates is the file was used in the solution.
New commands added:
(1) SOURCE command was added which allows globk and glorg commands to
be read from another file inside the primary globk command file.
The command can only be used in the primary globk command file.
Format:
SOURCE <file name>
where <file name> is a file containing globk commands (can not include
another source command).
This command is useful for having a single file with e.g. use_site
commands which are shared among multiple globk command files.
(2) RAD_RESE command allows the user to specify that the radiation
parameters should not be reset when a new IC epoch is encountered.
The default (compatible with pre-4.17 verions of globk) is to reset
the radiation parameters when a new IC epoch is encountered.
Format:
RAD_RESE No
will cause the radiation parameters not to be reset thus allowing
the average value to be computed when multiple days of data are
combined. Markov process noise (mar_svs command) can be used allow
the radiation parameters to change with time when this command is
invoked. It is reccommended the Z option be specified in the
make_svs command when this option is used. (If the Z option,
which sets the direct-radiation parameter to 1.0 and the other
parameters to 0.0, is not used the average difference from the
(most-likely) time variable radiational parameters will be estimated
by globk.
GLOBK Version 4.16 Added features: 980216
----------------------------------
Fixed long standing bug in XYZ_to_GEOD which computed the rotation matrix
between XYZ and NEU slightly wrong. Effect is neglible on results but
could cause some North-Up correlation when height sigmas are large.
Fixed a long standing bug in getting pivot element during matrix inversions.
Seems to have no effect on results but might have affected numerical stability.
GLOBK Version 4.15 Added features: 971112
----------------------------------
Added (optional) inclusion of h-file name string check to the rename command
in the eq-files. (see RENAME below).
GLOBK Version 4.14 Added features: 970909
----------------------------------
New output features added (see glorg help for list of all options)
ERAS -- Erases the output prt or org files before printing (should
NOT be used in glred runs because this will erase the output
before each day of data is processed).
NOPR -- Can be used to stop the output of the prt, crt or org files.
GLOBK Version 4.13 Added features: 970607
----------------------------------
A new output option was added at this version plus we modified the
arguments to the MAX_CHII command. We also changed some of the internal
book-keeping in globk to record better the start and end times of the
h-files used in the solutions.
The new output option is GDLF which can be added to any of the prt_opt,
crt_opt or org_opt commands. When this option is used a new block is added
to the globk output files which list the experiments used. The form is
EXPERIMENT LIST from sio.srt
# Name SCALE Diag Scale Status
1 raw97/h9704201159_dgga.glx 1.000 1.00000000 USED
2 raw97/h9704201159_igs1.glr .001 1.00000000 USED
3 raw97/h9704211159_dgga.glx 1.000 1.00000100 USED
4 raw97/h9704211159_igs1.glr 1.000 1.00000000 USED
5 ../SIO_96/raw97/h9704221159_dgga.glr 1.000 1.00000000 NOT USED
6 raw97/h9704221159_dgga.glx 1.000 1.00000000 NOT USED
where SCALE is the overall variance scaling of the input globk hfiles,
Diag Scale is the scaling applied to the diagonal, and
Status reports if the hfile was actually used in the solution.
The SCALE value is input following the h-file name in the gdl list. The diagonal
scaling is not yet documented and is mainly useful for reducing the effects of numerical
instabilites especially with IGS SINEX files from some centers.
The changes to the MAX_CHII command are given below where the command is
described.
Finally, in version 4.13, the back solution code was modified to ensure that h-files
not used in the forward solution were not used in the back solution either.
GLOBK Version 4.12 Added features: 961127
----------------------------------
New commands have been added with 4.12 and some internal changes have
been made to the operation of globk. We also extended the options in
the max_chii command that was introduced in version 4.11 of globk.
(We extended the documention under 4.11 to show this new option).
The other major change made was to allow the direct copying of the initial
covariance matrix from the globk input files when a loose solution
is run and there are no additional computed parameters (such as earth
rotation or translation parameters). A "loose" solution is defined by
all apriori variances being greater than 100 (i.e. 10 meters for positions
and 10 m/yr for velocities). A new command has been introduced which
allows this option to be turned off. The command is simply NO_DIRCP
with no options. (This is useful if a large rotation is to be removed
in the globk run). There is a small impact on the chi**2 computation
when a direct copy is made becuase the parameters in the direct copy
matrix are included in the sum of the number of degrees of freedom in
the solution).
Another small change is that when the DESCRIPT command is used, the
description string is written at the top of the print and glorg files.
New site selection commands:
----------------------------
We have added two new commands to help limit the sites that are included
in a globk analysis. We have also "cleaned" up the code which keeps
track of the sites actually used in an analysis so that only sites with
data should appear in the final solution (if a site is deleted during the
run due to bad prefits, it will still appear in the output (see max_chii
command below).
USE_POS <+/-> <Lat LL> <Long LL> <Lat UR> <Long UR>
where <+/-> indicates to use (+) or not to used (-) the sites in
the box specified by the latitude and longitude range.
<Lat LL> <Long LL> <Lat UR> <Long UR> are the bounding regions
of the area covered with LL--lower left, and UR upper right. The
latitude and longitude are given in degrees. Longitude are positive
east and can be given as positive or negative values.
For example: To exclude all sites in Southern California
use_pos - 30.0 -125 35.0 -115.
This command is used in conjunction with the standard use_site command.
Specifically is use_site clear is initially specified, then use_pos + ..
command can be used to add sites in different regions, or alternatively
if use_site all is initally given (default) then use_pos - .. can be
used to remove regions. The use_site command can be used after use_pos
to add or remove specfic sites by name.
USE_NUM <Number>
Where <Number> is an integer value that specifies the number of times a site
must be used for it to included in the analysis. The command is applied
after all other use options have been processed, and will result only
in site being removed from the analysis (i.e., if use_site -NAME has
been used to remove site NAME, then it will not be included in the analysis
even if it is used more than <Number> times. In the same way if use_site +NAME
had been given, but site NAME was used less than <Number> times, it will
NOT be used in the analysis.
New parameter estimate commands
-------------------------------
New parameters have in introduced in globk vers 4.12 to allow for
translation rates of change, and scale changes. For the translations
the new commands replace the amr_tran command (although amr_tran can
still be used). The new commands are:
apr_tran <sig X> <sig Y> <sig Z> <sig dX/dt> <sig dY/dt> <sig dZ/dt>
where <sig X> <sig Y> <sig Z> are the apriori sigmas for the X, Y, and Z
translation parameters (meters) and
<sig dX/dt> <sig dY/dt> <sig dZ/dt> are the apriori sigmas for the
dX/dt, dY/dt and dZ/dt translation rate parameters (meters/year). These
are the new parameters which were not in earlier versions of globk.
mar_tran <mar X> <mar Y> <mar Z> <mar dX/dt> <mar dY/dt> <mar dZ/dt> <WN X> <WN Y> <WN Z>
were <mar X> <mar Y> <mar Z> are the process noise for the translations
(m**2/yr) and
<mar dX/dt> <mar dY/dt> <mar dZ/dt> are the process noise for the
translation rates of the changes (m/yr)**2/year.
<WN X> <WN Y> <WN Z> white noise process applied independent of time
step (optional, m**2)
apr_scale <sig scale> <sig scale rate>
where <sig scale> is the apriori sigma for the scale (parts-per-billion, ppb)
and <sig scale rate> is the apriori sigma for the scale rate of change (ppb/yr).
* MOD TAH 130716: Added white noise variance as optional third argument
mar_scale <mar scale> <mar scale rate> [scale white noise variance]
where <mar scale> is the process noise for the scale (ppb**2/yr) and
<mar scale rate> is the process noise for the scale rate of change
(ppb/yr)**2/yr.
[scale white noise variance] white noise variance added to all data
steps indenpendent of time separation (ppb**2).
GLORG run from within GLOBK
---------------------------
A new feature has been added which allows glorg to be run directly
from with in globk. The feature is invoked using the org_cmd command
(given below) which gives the name of glorg command file. The following
commands are used with this feature:
ORG_CMD <glorg command file name>
where <glorg command file name> is the name of the command file with
the glorg commands in it (see glorg.hlp)
ORG_OPT <options list>
where <options list> is the list of options to constrol the glorg
output. The most commonly used are BLEN--output baseline lengths
and components,BRAT--Output baseline rates of the change components,
PSUM--Position summary, VSUM--Velocity summary, CMDS--globk command file
summary.
ORG_OUT <File name>
where <File name> is the name of the output file. If this command is not
given, then the output goes to a file name generated by replacing the characters
after the last period (.) in the GLOBK print file name with 'org'. If there is no
last period, then .org is appended to the print file name.
GLOBK Version 4.11 Added features:
----------------------------------
(1) The change to version 4.10 was simply one of increasing the
maximum number of stations and parameters that could be processed.
(2) With 4.11 (modified in version 4.13), a new command MAX_CHII was introduced
MAX_CHII <max chi**2 Increment> <max prefit difference> <max rotation>
where the three arguments give tolerances on the chi-square increment, change
in a priori parameter values, and allowable pre-solution rotations of a network
before combining a new h-file.
<max chi**2 Increment> gives the maximum allowable increment in the chi**2 when
a new h-file is combined in the solution. In this version of globk these data
will not be added to the solution and the solution will continue to run. The same
procedure will happen with negative chi**2 increments (which result from numerical
insatabilities and can be solved often by tightening the apriori constraints and
reducing the magnitudes of process noise on the stochastic parameters. More detail
diagonistic information is now output to the log file if a negative chi**2 increment
occurrs. The default value of max_chii is 100.0.
<max prefit difference> is maximum difference in the prefit residual for station
coordinates. This value also sets the limits for other parameters in globk but with
the input value internally scaled to provide comparably reasonable tolerances for the
other parameters. For EOP, the prefit tolerance is 10 times the surface rotation
implied by the station coordinates, and for orbital initial position and velocity,
1000 times the station tolerance. If the prefit difference exceeds this limit, then
the estimate for this parameter is not included (row and column removed from the
covariance matrix), or in the case of a station coordinate, all three coordinates
are excluded. The default is 10,000, corrsponding to 10 km for a station position.
When the aprioris are well known, this value can be set small (e.g., 0.1 for global
networks with good orbits, corresponding to 10 cm for station position, 32 mas for EOP,
and 100 in orbital initial position).
<max rotation> sets the tolerance for a rotation of the prefit station coordinates
before they are compared with the current solution. This check complements the new
feature of globk in which an orientation difference between the stations of the input
h-file and the current is computed. If this value exceeds <max rotation>, then the
rotation is removed from the station coordinates and added to the EOP parameter estimates
and a message is written to the screen. The purpose of this feature is to avoid having
to set the Markov values of the EOP parameters inordinately large to handle a few h-files
for which the EOPs of the phase processing had large errors. For global networks, the
rotation can be determined well from data, so setting a small <max rotation> value
(e.g. 20 mas) is useful to maintain small Markov values. For regional networks, the
rotation is poorly determined, so the tolerance should be kept large to avoid erroneous
rotations that can cause numerical problems. The default is 10,000 mas.
GLOBK Version 4.03 Added features:
----------------------------------
(1) Report_stat calls added to support automatic monitoring scripts
for globk
(2) In version 4.03, J2000 and B1950 hfiles can be mixed but the
make_svs command must be used to ensure that a single ephemeris file
is generated with the B1950 and J2000 ephemerides differing by only
the rotational terms between J2000 and B1950.
(3) With this release, there is an important, but hopefully transparent
change, in that globk is now a single program rather than the original
from of several linked programs (i.e., globk scheduled glfor, glbak, glout
glsave and glinit to run). The "segmented" version is still available in
the globk directory and is called globs. The combined version, pointed
to from the kf/bin directory is saved in the globc directory. When any
module of globk is remade, it is necessary to re-make in the kf/globc
directory. The effect of this change is the globk common file (given
with the com_file command) is not written unless the name is given in
the command file. If the com_file is not written then glorg can not be
run after globk finishes.
GLOBK Version 4.0 New Features:
-------------------------------
With version 4.0 the binary file structure for globk changed considerably
to support carrying the information needed for SINEX files. This
additional information increases the size of binary hfiles by about 5%.
As far as known, the new file structure is both foward and backward
compatability and GLOBK Ver 4.0 is backward compatible with the old
binary file structure.
Several new features and commands have been added with this release.
In Globk, new commands are:
MAKE_SVS <file name> <Z>
automatically makes the satellite ephemeris file for this solution.
If the optional Z is specified after the file name then the radiation
parameters are set such that the direct radiation is 1.0 and all
the other radiation paramters are zero. This can have consequences
when small constraints are placed on the radiation parameters. The
default is not to reset the radiation paramaters (Vers. 4.11). Use of
make_svs is the recommended method for getting the ephemeris file.
(It is much more robust than concatinating the ephemeris files from
htoglb and using unify_svs). The ephemeris elements will be the same
as those in the gfiles used in the original gamit runs (For post-3.2
htoglb generated binary hfiles; for eariler files the estimated orbital
elements will be used. When hfiles from the same day are combined,
the elements from the first hfile read will used.
NOTE: <file name> is overwritten with this command if it already exists.
NOTE: This command must be given at the top of the command file. It falls
in the same class as com_file, srt_file, eq_file and srt_dir commands.
Pre 5.32
APR_RAD <DRAD> <YRAD> <ZRAD> <BRAD> <XRAD> <DCOS> <DSIN> <YCOS> <YSIN> <BCOS> <BSIN>
Post 5.32
APR_RAD <DRAD> <YRAD> <BRAD> <DCOS> <DSIN> <YCOS> <YSIN> <BCOS> <BSIN> <2UD COS> <2UD SIN> <4UD COS
where the 11 values are the apriori sigmas, all dimensionless (i.e, 1.0
is the total solar radiation value), for each of the types of
radiations parameters. GLOBK checks which radiation parameters are
available in the binary hfiles and will estimate only those that are
available. If different sets of parameters are used in different
binary hfiles, then all the types used will be set. These 11 values
can also be added to the APR_SVS command line (after the 6 for position
and velocity of the satellite).
Pre 5.32
MAR_RAD <DRAD> <YRAD> <ZRAD> <BRAD> <XRAD> <DCOS> <DSIN> <YCOS> <YSIN> <BCOS> <BSIN>
Post 5.32
MAR_RAD <DRAD> <YRAD> <BRAD> <DCOS> <DSIN> <YCOS> <YSIN> <BCOS> <BSIN> <2UD COS> <2UD SIN> <4UD COS
Similar to the APR_RAD command except for the process noise. The
process noise units are /year.
A warning message will be printed if the additional 11 elements
are not added to either the apr_svs or mar_svs commands. If the
apr_rad and mar_rad commands are used, then this warning may be
ignored.
NEW PRT_OPT/CRT_OPT/BAK_OPT forms:
---------------------------------
With version 4.0, these options may be entered in string form. The
GLORG help gives the 4-character codes for each of the options. The
bit-mapped integer value form still works. When the string option
is used in a command file, the codes are blank separated. When used
in a runstring (for glorg for example), the codes should be separated
by : or = signs, i.e., in a command file the usage would be
prt_opt blen psum
In a runstring
glorg test.org blen:psum:cmds glorg.cmd test.com
A new print option has been introduced which allows adjustments to
positions to be printed in a form similar to the velocity summary
(i.e., one line per site with laitude longitude and position adjustments
and sigma.
CHANGES TO OTHER PROGRAMS:
--------------------------
(1) htoglb has been modified to read SINEX files (see help)
(2) glorg has a new "constrain" command which is similar to equate
but the equate is not absolute.
(3) autcln has new features including SNR editing and normal pointing
(see autcln help).
(4) GLRED now allows automatic concatination of binary hfiles using
at + sign at the ends of lines (see glred help).
GLOBK Version 3.3 New Features:
-------------------------------
Version 3.3 has mainly enhancements of existing features and some
improvements to glorg. One new program have been added to aid in
making apr coordinate files be internally consistent. Changes in
globk are:
(1) The underscore character can be added at the ends of names to
force the recognition of blanks in a name. This is mainly for VLBI
station names where, for example, DSS15 could not be separated from
DSS15_LA. With the enhancement, DSS15_ can be separated from DSS15_LA
because the trailing _ in DSS15_ is forced to be a blank.
(2) Apriori coordinates for sites and satellites are now written into
the binary hfiles and so if no new apriori coordinates are given in
the apr_file, the apriori from the GAMIT solution will be used. In
previous versions of globk, the estimated position of the site would
have been used.
(3) The updating of the covariance matrix when a new satellite ephemeris
epoch is encounted has been changed. In early versions of globk, the
apr_svs variances were simply added to the covariance matrix. In 3.3
the rows and columns are cleared before the apr_svs variances are
added to the covariance matrix. This change means that orbital elements
on satellites can not grow to large values when there is no or poor data
determinations of orbist from previous orbital arcs.
(4) A problem associated with estimated only some components of pole
position and UT1 has been fixed to that h-file with only polar motion
estimated for example but not UT1 are now processed correctly. It is
still recommended that all or no orientation parameters be estimated
rather than mixing with only some of the components estimated.
(5) Glorg has been enhanced to allow easier specifcation of coordinate
system constraints and to estimate poles of rotation using selected
site's velocities. See GLORG help for more details.
(6) A new program, unify_apr has been added to help make coordinates
of sites with different names have the same apriori coordinates and
velocity. See UNIFY_APR help for more details.
(7) The wild-card feature which uses @ to replace parts of the names
of file with the corresponding characters from the gdl file name has
been extented to allow the com_file, srt_file and sol_file to use
the wild-card. This feature is useful for saving the com_file and
sol_file (that are needed by glorg) for different analysis strategies
where the gdl file is changed between solutions. NOTE: When glred is
run, the gdl file name passed to globk is different for each day of
data and so if a wild card is used, then the com_file and sol_file
will be saved for each day. This could use up a considerable amount
of disk space, and so unless these files are wanted, the wild card
feature should not be used wth glred runs.
(8) The use of F/f in the apr_neu command to fix a coordinate is
now implemented consistently.
GLOBK Version 3.2 New Features:
-------------------------------
Two new commands have been added to Globk in version 3.2 and there has
been several bug fixes, improvements to the memory management, and
enhancements to the output.
New Commands:
=============
BAK_PRTS <CLEAR ALL List of sites ...>
The command allows the user to specify which sites should always be
printed in the back solution file. The CLEAR is optional and should
normally not be used since it will override the internal rules for
outputting a site. ALL will select all sites. The following rules
apply
for outputting:
--Site positions--:
All and only sites that are Markov (i.e. mar_neu or mar_site command
used) or are effected by an earthquake (see EQ_FILE command) will be
output unless CLEAR is used in the BAK_PRTS commands in which case only
those sites listed in the BAK_PRT command will be printed independent of
whether they are Markov or effected by an earthquake. In all cases,
only sites that were observed on the day being output will be printed.
--Baselines and baseline components--:
In general, for a baseline to be printed, both sites in the baseline
need to pass the criteria to be output. For baselines to be printed
bak_opt bit 1 (decimal value 2) must be set. If only bak_opts 2 is set
all baselines between sites observed on the day will be printed. (This
can become a long list). Restrictions can be placed on the baselines
printed with the following options:
bak_opt 2050 (bits 2 and 11 set): Restricts the baselines to only those
that are effected by an earthquake (including pre and post-seismic
intervals extended by 2 days so that part of "deterministic" period) or
are Markov. Use of BAK_PRTS allows additional sites to be printed that
do not pass the earthquake or markov rules. Also the bak_prts sites are
printed even when they do not occurr on a given day. (All other
baselines must appear on the day).
.....................................................................
EQ_FILE <file name>
The new file contains commands that can be used to tell globk about
earthquakes and what to do when one is encountered. There is also a
general command included for renaming sites (useful for un-doing htoh
runs used in the past to account for earthquakes.
NOTE: This command falls in the group of commands that must be issued at
the top of the globk command file because it effects the names of the
sites that will be used in interpreting subsequent commands.
With in the EQ_FILE the following commands may be used: (They can not
be used in the globk command file). An example of the EQ_FILE is given
below. In the standard version (parameter max_eq in kalman_param.h),
parameters for 128 earthquakes can be specified in a single run.
Using the name NONE will clear the entries from previous eq_files (useful
when the options feature is used to select different types of runs).
A " eq_file RESET" command will act like a RESET command in an eq_file .
EQ_DEF <Code> <Lat> <Long> <Radius> <Depth> <epoch>
This command defines the Earthquake. The arguments are:
<Code> a two letter code which is used (optionally) in renaming sites
and identifying this earthquake in subsequent commands. NOTE: the code
can not be the same as the last two letters of any site effected by the
earthquake. Specially PS can never be used as a code. NOTE: This
command must be issued before others referring to the Earthquake and the
same code may not be used twice.
<Lat> and <Long> are the geodetic latitude and longitude of the center
of rupture plane (approximately) in decimal degrees for the WGS84
ellipsoid.
<Radius> is the radius over which the Earthquake is assumed to have some
effect. Radii are computed as the chord distance from the lat, long and
depth of the Earthquake to sites. NOTE: For an earthquake to have an
effect, the radius must be greater than the depth. [Units: km]
<Depth> is the "depth" to the earthquake. No station will (in general)
be closer to the earthquake than its depth. The depth is the negative
of the ellipsoid height of the Earthquake (i.e, the value is positive).
The Depth is used for scaling the spatially dependent quantities
discussed below. [Units: km]
<epoch> Time of the earthquake entered as year month day hour min (all
separated by blanks. The experiment epoch of the data is used to test
if the data is before or after the earthquake. (For one-day hfiles, the
experiment epoch is the middle of the data; for combined global files it
is usually the middle of the last day of data in the combined file. If
srt_dir -1 was used in the globk run generating the combined file, the
experiment epoch will be at the start of the data set.)
EQ_RENAM <Code>
This command tells globk to rename the sites effected by the earthquake
by replacing the last two letters of the site name by the code for the
Earthquake. This command can be used in conjunction with the stochastic
options discussed below. For generating apriori coordinate files, this
is the preferred option. NOTE: If no apriori coordinate and velocity is
in the apr_file for globk for the renamed site, the values for the
original site will be used.
EQ_COSEI <Code> <Static Sigmas NEU> <Spatially dependent Sigmas NEU>
The command specifies the stochastic treatment of sites effected by the
earthquake. (If the command is not issued there will be no stochastic
variations at the time of the Earthquake. (If the sites are renamed
then this command is not needed, but can still be used to achieve the
effect discussed below. The arguments are:
<Code> for the earthquake (to get the parameters for earthquake in the
eq_def command.
<Static Sigmas NEU> these are three values separated by spaces that give
the standard deviation of the displacements (in meters) expected at the
time of the earthquake (i.e., apriori standard deviation of the
coseismic displacements.) These values will be applied to all sites.
<Spatially dependent Sigmas NEU> these are three values separated by
spaces that give the spatially dependent standard deviations. The
standard deviation a site is computed using:
Sig = SigSpatial*(depth/dist)^2
where Sig is the computed sigma, SigSpatial is the sigma given in the
command, depth is the depth of the Earthquake and dist is the distance
the site is from the Earthquake.
Units: on all sigmas is meters.
NOTE: When the rename and coseismic features are used together, the
covariance matrix elements and solution for the site name before the
earthquake (which itself may have renamed from previous earthquakes) are
copied to the new site name before the coseismic standard deviations are
added. The coseismic sigmas are not applied to the old site name.
WARNING. When the combination of rename and coseismic are used, the
apriori coordinates and velocities of the two site names should be the
same unless you are want to constrain the incremental adjustment to a
dislocation model for example. It the renamed site does not appear in
the apriori file, then globk will automatically do this.
NOTE: When the rename and coseismic features are used together, a
**NOTE** may be printed during the globk run saying that the original
site name of a renamed site could not be found. This will occur for all
sites that are used only after an earthquake. The Note can be ignored
if this is the case.
EQ_PRE <Code> <dur> <Static Markov NEU> <Spatially dependent Markov NEU>
This command allows the specification of random walk parameters to be
applied before the earthquake.
<Code> is the earthquake code.
<dur> is the number of days before the earthquake to start the process
noise. The estimates will be output for up to 2 days before the start
of the process when a back solution is run. (Units: days)
<Static Markov NEU> three values that give the process noise parameters
in North, East and Up to be applied equally to all stations. The units
are mm^2/day and the process is a random walk, i.e., after n days the
process noise
<Spatially dependent Markov NEU> three values hat give the process noise
parameters in North, East and Up to be applied in a spatially dependent
fashion. (Same spatial dependence as the coseismic displacements except
squared since the variance is specified in the process noise).
NOTE: When the rename feature is used, the preseismic process noise is
not applied to the new site name (i.e, the site name used after the
earthquake).
EQ_POST <Code> <dur> <Static Markov NEU> <Spatially dependent Markov
NEU>
This command allows the specification of random walk parameters to be
applied after the earthquake. The command and specifications are
exactly the same as EQ_PRE except this process is applied after the
earthquake. If the rename option is used then the process noise will
not be applied to the old site name.
EQ_LOG <Code> <tau (days)> <static NEU sigma (m)> <Spatial sigma NEU (m)>
This command allows the specification of a log parametric fit to the
position estimates after an earthquake with code <code>. The model
used is the same as the one in enfit and tsview. The parametric
form is ln(1+dt/tau) where dt is the time after the earthquake. The
interpretation of the sigmas is the same as for eq_cosei. The positions
output by globk have the log contribution included (and therefore give
the instantaneous position). Log terms are extracted as part of the
apriori model when sh_exglk is used with the -apr option. Normally, eq_post
would not be used when eq_log is used. The log terms can be extracted
with sh_exglk with the -log option. (The log term lines in the globk
outout all start with the string LOG).
RENAME <Orig Site name> <New site name> [hfile code] <epoch range> <Position change>
The command, while not directly related to earthquakes, appears in the
eq_file because it involves renaming sites and is useful for un-doing
previous htoh runs used previously to treat earthquakes. The command
arguments are:
<Orig Site name> Original site name that appears in the binary h-files,
In most cases, it is recommended that just the first 4-characters be
given so that all previously renamed sites will be matched. (The @ or *
wildcard also be used for specify wild ends of names).
<New site name> New name for the site (Should not conflict with other
names in the binary h-files).
The following arguments are optional.
[hfile code] a string that must appear in the name of the h-file
for the rename to be applied. This argument may be omitted since
globk can distinguish whether the third argument is a string or a
number (epoch range); however, the string may not be entirely numerical.
More than one set of h-files can be included by issuing the command
multiple times with the same site names and epoch range.
It is useful for renaming sites during combinations of experiments on
the same day when there is a name conflict between sites in the
individual h-files [Feature Added at 4.15 11/12/97].
<epoch range> range of time over which the rename will be applied.
Specified as a pair of year month day hour min values (all separated by
spaces). Any hfile in which the start date is after the first date and
the end date in before the second date given will have the site renamed.
<Position change> is an optional position change to accompany the name
change. The position change is either XYZ or NEU (three values)
followed by a type declaration (XYZ or NEU). If no type is given, XYZ
is assumed. Units are meters and the change should move the site from
the original position to the new position. North, East and Up (NEU) are
defined as North along (the ellipsoidal) meridian direction at the
apriori coordinate of the new site name, East along the East longitude
direction, and ellipsoidal height. The rotation matrix is defined by
these directions and the NEU are rotated to XYZ using this rotation
matrix.
NOTE: The renames are applied before any earthquake processing, so names
generated from earthquakes use the new site name.
BREAK <Site Name> [hfile code] <Epoch> [optional end name]
The break command allows times of breaks to be specified.
<Site Name> is name of site with first break (4-char code only is normally
recommended).
[hfile code] is an optional code that acts the same as in the RENAME
command
<Epoch> is Year, month, day, hr and min of break. (With 24-hr Hfiles the
hr and min will not be meaningful).
[optional end name] is an optional last three characters of the name
for the site name after the break (e.g. APS)
When no end name is specified, names starting at 1PS-9PS and APS-ZPS will
be generated. Numbers and letters will be skipped if a duplicate name
would be generated.
RESET
Command added 130221: Resets all the site names back to _GPS extent except
_XPS and _XCL sites. Caution should be used in combined files where site
names could change within the same binary hfile. In these case, use hfupd
with RESET command in eq_file. Hfupd should handles this case succesfully.
Use glist to verify site name change as expected.
EXAMPLE of EQ_FILE:
+++++++++++++++++++
*----------------------------------------------------------------------
* Example of earthquakes and renames
* The following renames undo the htoh runs for Landers.
rename mojave1b mojave12
rename ds1b_gps ds10_gps
rename gold_gps ds10_gps
rename jplb_gps jplm_gps
rename sioc_gps sio2_gps
rename vndo_gps vndp_gps
rename pinb_gps pin1_gps
* Example of moving a GPS site to a VLBI site location. The tie is from
* ITRF-1990.
rename aron_gps algopark 92 5 1 0 0 92 8 30 0 0 -94.7630 -61.0170 -6.6660 XYZ
* Example to show a height change. The height change is dual frequency
* combination. (This correction would move from the phase center to the
* ground mark, i.e, zero height was used in the original processing.)
rename hart_gps hart_gps 90 1 1 0 0 92 8 30 0 0 0.0 0.0 -9.8013 NEU
* Correct for differences in antenna phase-enter models between regional
* and global processing:
rename math_gps math_aps igs 93 4 14 0 0 94 11 10 00
rename math_gps math_aps pgga 93 4 14 0 0 94 11 10 00
* Allow a break at site MATH
break math 2003 12 26 00 00
* The Lander's earthquake. Here are using the spatially dependent
* versions. The coseismic displacement sigma 100 km from the earthquake
* origin is 72 mm horizontal and 28 mm vertical in this case. For the
* Markov process the values correspond to 1.6 mm^2/day for all three
* components. (The parameters here were computed approximately from
* dislocation models of Landers). Landers/Big Bear has a moment of
* 1x10^20 N-m, and Ms 7.5)
*
eq_def LA 34.45 -116.50 500 20 92 6 28 12 0
eq_renam LA
eq_cosei LA 0.000 0.000 0.000 1.8 1.8 0.7
eq_pre LA 30 0.000 0.000 0.000 1000.0 1000.0 1000.0
eq_post LA 30 0.000 0.000 0.000 1000.0 1000.0 1000.0
* Other earthquakes could be defined here. If pre-seismic or
* post-seismic process noise overlap from different earthquakes
* then both sets of process noise are added.
*----------------------------------------------------------------------
.....................................................................
ENHANCEMENT OF EXISTING FEATURES IN GLOBK Ver 3.2
=================================================
(1) Experiment variance specification. The .gdl file format has been
extended to allow the specification of a variance scaling factor on
individual h-files. If not value is given, then a scaling of 1 is used.
Example:
/data11/tah/glbg/h92060111a.gld 1.000
/data11/tah/glbg/h92060211a.gld 3.076
/data11/tah/glbg/h92060611a.gld 4.328
/data11/tah/glbg/h92060711a.gld 7.072
In this case the scales were derived from the pre-fit chi**2 increments
for a forward solution. (grep and awk were used to generate the file
from the globk log file.
(2) Use of default file names for the com_file, srt_file and sol_file
now works correctly for both globk and globc. Therefore if these files
are not specified the defaults will be used.
(3) Globk now used full dynamical memory allocation so it uses only the
memory it needs. The memory is allocated after the list of global files
has been printed (glinit run). If there is sufficient memory at that
time, then the rest of run will have enough memory. A small message is
printed saying the amount of memory allocated.
(4) Warning messages are now printed at the start of the globk run if
there is no apriori site coordinate for a site in the apriori file.
(This message is printed if globk could construct an apriori coordinate
from renamed and earthquake sites.) The apriori coordinate used when
there is none in the apriori file, is the estimated position from the
last binary h-file in which the site was used.
(5) Reading of the svs_file has been modified so that the whole file is
read. Consequently, when there are duplicate entries, the last entry
found will be used. (In Ver 3.1 the first entry would have been used).
This change was made so that when regional and global h-file that
contain different satellites are used and the svs_file has not had
unify_svs run on it, all the satellite elements should be found.
(6) Multiple apriori coordinate files can now be specified with each one
possibly overwriting the results from previous ones. This feature has
been added so that all earthquake effected sites can be in one file so
that if we don't want to use the updated coordinates, the second file
need not be specified. Giving the name NONE or RESET (both upper case)
will reset the apriori file list to have zero entries.
e.g.,
apr_file old.apr
apr_file NONE
apr_file new.apr
apr_file extended.apr-PER ! do not use periodic terms in file.
no entries from the old.apr file will be used.
(max_apr_files sets the maximum number, standard value is 10 files.)
-------------------
BASE GLOBK COMMANDS
-------------------
Command at start of command file. All file names can contain
wild-card characters (Normally @ included to use gdl file name root.
Necessary for parallel runs in same directory).
Below @ at start of line means one or more spaces, no tabs.
COMMAND AT START OF COMMAND FILE
--------------------------------
@COM_FILE <file name>
Specifies file name for saved information about solution.
Needed if GLORG/GLSAVE/GLOUT to be run separately.
@SRT_FILE <file name>
Internal file with time sorted list of input h-files. Wild-card
should be used if run in parallel.
@EQ_FILE <file name>
Name of file with renames and earthquakes.
@MAKE_SVS <file name> <Z/A>
Name of file for satellite ephemerides (only needed if orbits
are being estimated. <Z/A> options specifies whether the radiation
parameters should set to zero, expect 1 for direct force) (Z) or
if apriori values from GAMIT run should be used (A).
@SRT_DIR <+/-1>
Direction for sorting +1 increasing time, -1 decreasing. Final
position estimates will be at the end or start of the data span
depending on direction. +1 is default. Normally not changed.
@USE_PRNN <Y/N>
where option Y (default) will use the legacy PRN_XXXX form for GPS satellites
(Other GNSS system use the new format). Option N will convert the GPS
names to the new format.
Command that can appear in any order and multiple times. Later
commands replace results of earlier ones.
FILE NAME COMMANDS
------------------
(specifying the name often invokes an action to happen i.e.,
specifying bak_file invokes a back smoothing solution to run.
@APR_FILE <file name>
Name of file with apriori coordinates (m), velocity (m/yr) and
epoch (decimal years) of sites. Site name first entry on line.
May also contain EXTENDED entries for offsets, logs, exponential
and periodic terms in the motion. Added -PER to file name will
stop periodic terms being applied. Command may be issued
multiple times.
@SOL_FILE <file name>
Internal binary file with Kalman filter state vector and covariance
matrix. Must be keep to run GLORG/GLSAVE/GLOUT as external programs.
Name should be unique for parallel runs (use @ wild card).
@BAK_FILE <file name>
Invokes running the Kalman filter smoothing run. Normally not
needed unless smoothed values are needed.
@OUT_GLB <file name>
Name of file to be written containing solution in GLOBK H-file
format which can be used in later solutions.
@IN_PMU <file name>
Specifies input file with apriori values of polar motion and UT1
(Earth Orientation Parameters, EOP). Files generated with
sh_update_eop and have file names starting with pmu.<series>.
Name is arbitrary. Useful if EOP values are to be studied.
@SOURCE <file name>
Name of file containing GLOBK commands that are read when the
SOURCE command is read. Most GLOBK commands can be included
in the files except the SOURCE command itself i.e., only one
layer of nesting is allowed).
OUTPUT CONTROL COMMANDS
-----------------------
These commands specify the types of information to output from
the solution.
@DESCRIPT <String containing description of solution>
General way to have a description put at the beginning of
output text files. Defaults if not specified.
@CRT_OPT <options>
@PRT_OPT <options>
@BAK_OPT <options>
@ORG_OPT <options>
General output and control options (see descriptions above and
in the GLORG help file).
@USE_SITE <ALL CLEAR list of sites>
Specifies sites to be used in the solution with ALL resetting
all sites to be used, CLEAR specifying no sites (command is
then used to add sites). The list of sites are separated by
spaces and can have + or nothing to add a site and - to remove it.
Non-unique starting strings maybe specified (i.e., -P056 or -P056@
will remove all sites that have starting characters P056).
Command can be issued multiple time.
@USE_POS <+/-> <Lat LL> <Long LL> <Lat UR> <Long UR>
Specifies region defined by lower-left (LL) and upper-right (UR)
latitude and longitude in degrees. + - at start either adds or
removes sites in bounding box.
@USE_NUM <number>
<number> is an integer value that specifies the number of times
a site must be used for it to included in the analysis.
@MAX_CHII <Incement> <Prefit diff> <Max Rotation>
Set conditions for solutions and sites to be included (see above)
@DEL_SCRA <YES>
Deletes the GLOBK intermediate scratch files. Useful for time
series runs but should not be used for post-processing analyses
with GLORG.
PARAMETER PROCESS NOISE COMMANDS
--------------------------------
These commands effectively specify which parameters GLOBK should
estimate and how the stochastic process noise should be handled.
Commands come in pairs with APR_xxx specifying the apriori sigma
(Sig) and MAR_xxx process noise values for a random walk process.
The random walk process noise values are <unit>^2/year were unit
is the unit of the specific parameter.
@APR_NEU <ALL/SITE> <N Sig> <E Sig> <U Sig> <dN/dT Sig> <dE/dT Sig> <dU/dT Sig>
@MAR_NEU <ALL/SITE> <N> <E> <U> <dN/dT> <dE/dT> <dU/dT>
ALL will apply to all sites, SITE can be a non-unique starting
name applied to specific sites.Units are meters and meters/yr.
Specifying 0 will cause the parameter not be estimated (state remains
unchanged from input H-files). To force a zero variance use F for
the Sig (sigma) values. Zero process noise or not using the MAR_NEU
command will make the parameter deterministic (standard weighted
least squares).
@APR_WOB <Xp Sig> <Yp Sig> <dXp/dT Sig> <dYp/dT Sig>
@MAR_WOB <Xp RW> <Yp RW> <dXp/dT RW> <dYp/dT RW>
Units are milli-arc-sec (mas) and mas/yr. Markov process values are
mas^2/yr and (mas/yr)^/yr. These commands will align the values of
the EOP parameters in the input H-files and this can be problematic
with some SINEX derived H-files (see APR_ROT command).
@APR_UT1 <UT1 Sig> <dUT1/dT Sig>
@MAR_UT1 <UT1 RW> <dUT1/dT RW>
Units are milli-arc-sec (mas) and mas/yr. Output converts to
milli-time-seconds (ms) and ms/day.
Version 5.32: Added negative value of <dUT1/dT RW> will decouple LOD
from UT1 i.e., integrating LOD will not generate UT1 values. Used
to keep UT1 at aproiri values when GNSS orbits are estimated each day.
@APR_ROT <X Sig> <Y Sig> <Z Sig> <dX/dt Sig> <dY/dT Sig> <dZ/dT Sig>
@MAR_ROT <X RW> <Y RW> <Z RW> <dX/dt RW> <dY/dT RW> <dZ/dT RW> <X WN> <Z WN> <Z WN>
Allows estimation of general rotations around the XYZ axes without
direct being tied to the EOP parameters. Useful for some SINEX files
(see above). The MAR_ROT has three additional values which allow
process noise, independent of the time separation between H-files
(Needed if SINEX files at the same time have different EOP values
that are not included in the SINEX files). Units mas and mas/yr.
@APR_SVS <ALL/PRN> <IC and radiation Sigs>
@MAR_SVS <ALL/PRN> <IC and radiation RWs>
Values for treating orbits (see detailed discussions above). See
APR_RAD list of radioation paramters above and change at version 5.32.
@APR_RAD <ALL/PRN> <Radiation Sig>
@MAR_RAD <ALL/PRN> <Radiation RW>
Values for radiation parameters only (ee detailed discussions above).
@APR_SVAN <ALL/PRN> <X Y Z offset Sig>
@MAR_SVAN <ALL/PRN> <X Y Z offset RW>
Values for satellite antenna phase center offset estimates (see above).
Should be used when MIT global H-files included in solution.
@APR_TRAN <X Sig> <Y Sig> <Z Sig> <dX/dT Sig> <dY/dT Sig> <dZ/dt Sig>
@MAR_TRAN <X RW> <Y RW> <Z RW> <dX/dT RW> <dY/dT RW> <dZ/dt RW>
Allows estimation of translations of the reference frame. When using
BASELINE solutions from GAMIT and some SINEX files, these commands
should be included to explicitly allow reference frame transition.
In BASELINE mode, should be used with GLORG pos_org command includes
translation. Units are meters and m/yr. See discussions above.
@APR_SCAL <Scale Sig> <Scale rate Sig>
@MAR_SCAL <Scale RW> <Scale rate RW>
Allows estimation of scale and scale rate changes (See above).
MUST be used if scale included in GLORG pos_org command. Units ppb and ppb/yr.
PROGRAM globk terminating