User Tools

Site Tools


public:vievs_manual:important_files

Important files

Parameter files

The parameter files are of type .mat (Matlab internal binary format) and contain the VieVS processing options (e.g. minimum elevation angle, quality limit, EOP models, estimation intervals and constraints). These files can be used both in the processing (VIE_BATCH) as well as in the interface (VIE_SETUP) for loading the parameters into the GUI.

The parametrization for the 'main' VieVS (except parameters for Simulation, Scheduling and Global solution) is saved in a parameter file. After every modification (e.g. button click, checkbox value change), the current parametrization is saved automatically as auto_save_fromGUI_yyyymmdd.mat in /WORK/PARAMETERS/ . This 'last' parametrization can be loaded by selecting File - Parameter files - Load current. This might be useful when you want to use the same parametrization in several runs (but keep in mind that this file is changed every time the GUI is changed!).

It is also possible to save a desired parametrization in an own parameter file. Simply go to File - Parameter files - Save parameters as…, select a filename and save it. Later you can load this parametrization (File - Parameter files - Load parameters).

Please note: The parameters used for a specific data processing run by the VieVS modules are automatically stored in the following directories along with the processing results (VieVS data structures): /DATA/LEVEL1\

VieVS data structures

The output (basically the processing results and the used parameters) of the different VieVS modules is written the following directories (considering sub-directories defined in the Run options):

  • /DATA/LEVEL0: output of VIE_INIT
  • /DATA/LEVEL1: output of VIE_MOD (output structures from VIE_INIT, updated with VIE_MOD processing results (partial derivatives, computed delays, meteorologic data, etc.))
  • /DATA/LEVEL2: output of the VIE_GLOB
  • /DATA/LEVEL3: output of VIE_LSM (estimates plus formal errors, residuals, etc.)
  • /DATA/LEVEL4: output of VIE_SIM
  • /DATA/LEVEL5: output of VIE_SCHED

The VieVS data structures written to these LEVELx directories are described in detail below.

LEVEL0

LEVEL1

LEVEL2

Containing .mat-structures of global parameter estimation vie_glob.

LEVEL3

The actual estimation results calculated in VIE_LSM are stored here.

  • atpa_.mat: contains the normal equation matrix of the least-squares adjustment
  • atpl_.mat: right hand side vector of the least-squares adjustment
  • opt_.mat: structure containing all estimation options of vie_lsm
  • res.mat: structure containing the residuals of the least-squares adjustment

LEVEL4

Containing .mat-structures of simulation data.

LEVEL5

Containing .mat-structures of scheduling data.

runp

The file runp.mat is one of the main files in a VieVS processing run. This Matlab struct is created in the interface (VIE_SETUP) and contains the sub-directories for all three modules (init_path for VIE_INIT - LEVEL0, mod path for VIE_MOD - LEVEL1, lsm_path for VIE_LSM - LEVEL3), as well as a logical variable for all six modules (sched, sim, init, mod, lsm, glob), de fining whether or not this module should be run. The file must be in /WORK/ and is overwritten for every single processing run.

Process lists

The process list is a .mat file, which contains a list of sessions to be processed (multiple sessions can be selected). The file process_list in /WORK/ is loaded by vie batch and the sessions written in that file are processed. User defined process lists (e.g. CONT campaigns, Intensives in 2010,…) can be loaded and saved in the interface of VieVS.

File - Parameter files - Save process_list as saves the selected sessions from the listbox in File - Set input fi les to a process list.

This process list can be loaded in File - Set input files by selecting Browse for process_lists. All the containing sessions should appear in the listbox of File - Set input files. A function for automatically generating process lists is mk_list.

Outlier files

Outlier fi les are simple ASCII fi les containing line-per-line observations which have been (automatically or manually) set as outliers. This is usually done for observations with large residuals. One line contains the two participating stations and the mjd of the observation. Three example lines of such a file are:

NYALES20 SVETLOE  56009.261099537034
SVETLOE  WETTZELL 56009.261099537034
KOKEE    NYALES20 56009.256249999999

If an outlier file does already exist, all new outliers will be appended (the old ones will not be removed!). Diff erent folders may exist in the /DATA/OUTLIER/ directory. One can select the folder where outlier fi les are written to, in the File - Set input files part of the GUI. The outlier file of a session can be opened by right-clicking on the session in the listbox and selecting the Outlier file.

Define outliers

They can either be defi ned automatically or manually:

Automatically

In Run - VieVS estimation settings the user can select one of two outlier tests.

  • The 'simple' outlier test sets observations as outliers which have larger residuals than $c \cdot m_0$ where $c$ is a user-defi nable number (e.g. $5$) and $m_0$ is the a posteriori standard deviation of unit weight.
  • The 'normal' outlier test has a outlier-threshold of $c \cdot m_0 \cdot \sqrt{Q_{vv}}$ where $Q_{vv}$ is the variance-covariance matrix a posteriori of the observations.

If no automatic outlier selection should be used, the two checkboxes should be unclicked!

Manually

Outliers can be selected manually in the plotting of residuals. The button Select outliers lets the user drag open a box (left mouse-button) and select observations. A click on Remove outliers writes the selected observations to the outlier file.

Treatment of outliers

VieVS removes outliers always at the beginning. If no outlier file is specifi ed before the processing starts no outliers will be removed in this run. This means that if you want to have a solution without outliers, you have to process the session twice: One for the detection (automatic or manual) of them, the second with those outliers removed!

The checkbox Eliminate outliers in File - Set input files tells the program to remove observations which are written in the outlier file specifi ed by the directory.

OPT files

The OPT-fi le contains several informations related to one session, usually sessions which have large residuals or generate some other 'problem'. In the following lines an example of the content of an OPT file is given:

CLOCK REFERENCE:
WETTZELL

CLOCK BREAKS: 2
BADARY   55454.4
WETTZELL 55372.369

# stations down time (optional)
STATIONS TO BE EXCLUDED: 1
MATERA   YYMMDDhhmm-YYMMDDhhmm
WETTZELL YYMMDDhhmm-YYMMDDhhmm

# if no down time:
STATIONS TO BE EXCLUDED: 1
MATERA
WETTZELL

BASELINES TO BE EXCLUDED: 3
NYALES20  ZELENCHK
SVETLOE   ZELENCHK
BADARY    ZELENCHK

SOURCES TO BE EXCLUDED: 2
1936+095
2356+385 YYMMDDhhmm-YYMMDDhhmm

STATIONS TO BE DOWN-WEIGHTED (NOISE IN [m]): 2
KOKEE    0.04
WESTFORD 0.04

NO CABLE CAL: 2
WETTZELL
MATERA  

Notes on the OPT file syntax

  • The OPT file is very sensitive in terms of blanks and characters. Please stick to the format. Remember: a station name MUST have 8 characters!
  • Comment lines begin with '#'

Clock reference

Clock parameters in VieVS are estimated relative to a reference clock, which is one of the station clocks (the clock of the first station in the antenna fi le is chosen by default). This reference clock can be changed in the OPT fi le as shown in the example lines (the clock of station WETTZELL is set as reference clock in this example). This is usually important when there is a problem (e.g. a clock break) at exactly this station. Hence the reference clock should be a clock without problems.

:!: Do not use a station with clock breaks as reference!

Clock break

Station clocks may have discontinuities ('jumps') which - if they are improperly modeled - degrade the estimates. The clock might be modeled as one (e.g. linear) function with respect to the reference clock. If a clock break is speci fied in the OPT fi le a separate linear function is estimated before and after the event. Those two functions will model the actual behavior of the clock much more accurately. It is possible to specify more than one clock break per station.

The first-solution residuals are well-suited to identify clock breaks (station and epoch) since there is only one zenith wet delay per station (atmospheric influence is large) and one linear clock function estimated per station: The residuals due to a clock break do not go into the estimated parameters in this solution.

The residuals can be visualized and clock breaks added interactively in the plotting tool.

Exclude stations, baselines and sources

If stations, baselines or sources should not take part in the processing of a session, they have to be excluded in the OPT file. Then they are removed right in the beginning of a VieVS run. It is also possible to deselect the observations of a specifi c station for a certain time interval.

mk_list

The function mk_list is a very useful tool that helps the user to select certain sessions with specified attributes (e.g. all IVS-R1 sessions from 2012 where Wettzell contributed). It can be accessed by typing

 [process_list, sess] = mk_list(s1, s2, :::)

in the command window (make sure that the Current Folder in Matlab is set to the WORK directory and that VieVS has been started before, but it does not need to be opened). s1, s2, etc. are strings containing the parameters to be set, process_list is the list of sessions and sess contains the session names. To make the process list available in the VieVS GUI, save it in the WORK/PROCESSLIST directory:

 save PROCESSLIST/test_processlist.mat process_list 

mk_list.m requires that the master files of all affected years are available in the directory DATA/MASTER and that the specified sessions exist in the DATA/NGS directory.

Parameters that can be set (you also find this information in the header of the MATLAB function; can be shown in the command window by typing 'help mk_list'):

  • use VGOSDB or NGS files: process_list=mk_list('VGOSDB',…) (default) or process_list=mk_list('NGS',…)
  • include only a specific type of session: process_list=mk_list('all',…). Possible:
    • all: use all sessions
    • R1: only R1 sessions
    • R4: only R4 sessions
    • EURO: only sessions of the European VLBI network (EUR)
    • AUA: only sessions of the AUSTRAL VLBI network (AUA)
    • I1: only INT1 (XU) sessions
    • Q1: only INT2 (XK) and INT3 (XK) sessions
    • ..
  • include only sessions from specific years: process_list=mk_list(…,'YEARS',yrs), where yrs is an array with the years you want to include
  • include only a specific station: process_list=mk_list(…,'REQSTAT',sta), where sta is the two-letter ns-code of the station
  • include only intensive sessions: process_list=mk_list(…,'ONLYINT')
  • include only non-intensive sessions: process_list=mk_list(…,'NOINT')
  • include only sessions with minimum number of stations: process_list = mk_list(…,'MINSTANUM',num), where num is the desired minimum
  • include all versions in your NGS directory : process_list=mk_list('…,'ALLVERSIONS')
  • list of sessions to be excluded: process_list=mk_list(…,'EXCLUDE',excllist) where excllist is the file name of the list containing the sessions to be excluded.

Please use the following format for the exclude list:

$14APR28XA
$14AUG18XA
$14AUG25XA
$14SEP02XA

In the following there are some examples on how to use the script:

  • Create a list of all R1/R4 between 2003 and 2007: process_list=mk_list('R1','R4','YEARS',2003:2007)
  • All sessions which METSAHOVI has participated in: process_list=mk_list('all','REQSTAT','Mh')
  • All R1/R4 in 2009, except those listed in exclude.txt: process_list=mk_list('R1','R4','YEARS',2009,'EXCLUDE','exclude.txt')
  • All CONT02 sessions, including all versions of the NGS files (in case there are more than one): process_list=mk_list('C02','ALLVERSIONS')

| ⇐ Previous section | Overview | Next section ⇒ |

public/vievs_manual/important_files.txt · Last modified: 2018/12/07 10:01 by dlandskr