vievs user workshop 14 – 16 september, 2011 vienna vie_glob hana spicakova
Post on 20-Dec-2015
216 views
TRANSCRIPT
2VieVS User Workshop 2011
VIE_GLOB_V1d
compatible with VieVS Version 1d
parameters which can be estimated - from combination of more sessions
station coordinates and velocities: TRFsource coordinates: CRFEarth orientation parameters
- session-wise as reduced parameterszenith wet delaytroposphere gradientsEarth orientation parametersstation and source coordinates (not suitable for global estimation)
input datadatum free N-matrices (ATPA) and b-vectors (ATPl)VieVS/DATA/LEVEL2created within a „normal“ run of VieVS
VieVS User Workshop 2011 4
How to run VIE_GLOBVieVS/COMPILE/VIE_GLOB_V1d/vie_glob.m
path to the directory TEST_LEVEL2 with LEVEL2 data
the subdirectory TEST_OUT will be created in VieVS/OUT/GLOB/_ESTIMATES /_PLOTSeventually in: /CRF /TRF
only sessions with a lower RMS than chosen (in this case RMS < 2) will be used in the global adjustment
cun
PvvT
0
if ticked, N matrices and b vectors with reduced parameters from each session will be saved into VieVS/OUT/GLOB/BACKWARD_DATA
INPUT
OUTPUT
VieVS User Workshop 2011 6
GUI for TRF
VieVS/DATA/GLOB/TRF/
DATUM/
REDUCE/
DISCONT/
VELOC/
VELOC/TIES/
the original VLBI-DISCONT.txt file is provided at the web site http://vlbi.geod.uni-bonn.de/IVS-AC/data/VLBI-DISCONT.txt
TRF_datum1.txt
few_observations.txt
const_velocity_8410.txt test.txt
choose between NNT/NNR or NNT/ NNR /NNS
VieVS User Workshop 2011 7
GUI for CRF
VieVS/DATA/GLOB/CRF/
DATUM/
FIXED_SOURCES/ REDUCE/
choose between NNR or NNR + translation in declination condition
CRF_datum1.txt
test.txt special_handling_39.txt
VieVS User Workshop 2011 8
Theoretical background
Sort parameters in the N-matrix and b-vectorReduction of parameters
always reduced: clock parameters, zwd and troposphere gradientscan be reduced: EOP, stations and sources not suitable for global solution
Stacking of the reduced normal equation systems
2
1
2
1
2221
1211
b
b
x
x
NN
NN
globally estimated p.
reduced p.
211221211 NNNNNreduc
2122121 bNNbbreduc
nsereducreducreduc NNNN _2_1_ ...
nsereducreducreduc bbbb _2_1_ ...
9
Constraints
zwd, troposphere gradients, EOPconstraints are chosen in VIE_LSM
station coordinates and velocitiesrelative constraints for velocity
same velocity for all coordinate intervals at one station (e.g. antenna repair)
same velocity at stations in the same area (e.g. Hobart26 and Hobart12)
datum definition
source coordinatesdatum definition
VieVS User Workshop 2011
10
Datum definition
Station coordinates and velocitiesNNT/NNR condition
constrained 12 parameters Helmert transformation(3 translation and 3 rotation components + their time derivatives w.r.t. time)
NNT/NNR/NNS condition constrained 14 parameters Helmert transformation
(3 translation , 3 rotation components and 1scale component + their time derivatives w.r.t. time)
Source coordinatesNNR conditionNNR condition + 1 translation parameterparameter dz accouns for a global translation of the source coordinates in declination, caused e.g. by inaccuracy of the tropospheric propagation correction for sources observed at low elevations) (Feissel-Vernier et al. (2006))
fixing selected sources to a priori coordinates
VieVS User Workshop 2011
11VieVS User Workshop 2011
Final solution
applying of the constraints
final solution for global parameters
estimates are stored as a structure array in Matlab format and as a txt file
VieVS/OUT/GLOB/_ESTIMATES/TEST_OUT/globsol_TEST_LEVEL2.matglob_results_TEST_LEVEL2.txt
0
T
B
BNN REDUCCREDUC wbb REDUC
CREDUC
REDUCCREDUC bNdx
11
12
Estimates - Example
VieVS User Workshop 2011
VieVS/OUT/GLOB/_ESTIMATES/TEST_OUT/glob_results_TEST_LEVEL2.txtnames of sessions
number of sessionscorrections to station a priori coordinates and their standard deviations [cm]
corrections to station a priori velocities and their standard deviations [cm/y]
stations which were used for NNT/NNR
epoch of the coordinates [mjd]
start, end [mjd] of the interval (time between breaks)
sessions with RMS>maxRMS
VieVS User Workshop 2011 13
corrections to source a priori coordinates (RA, De) and their standard deviations [mas]
Estimates - Example
estimates of EOP:
dx, dy pole [mas]
dut1 [ms]
dX, dY [mas]
and their standard deviations
VieVS User Workshop 2011 14
Results
„new catalogues“ (TRF and CRF) are created automatically in VieVS/OUT/GLOB/
TRF/TEST_OUT/trf_TEST_LEVEL2.txt
and
CRF/TEST_OUT/trf_TEST_LEVEL2.txt
this new TRF catalogue can be used as input catalogue in VieVS
VieVS User Workshop 2011 15
PlotsVieVS/OUT/GLOB/_PLOTS/TEST_OUT/
ant_activity_TEST_LEVEL2.eps
ant_map_TEST_LEVEL2.eps
stations in sessions included in the global adjustment
map of stations,
blue circles: station included in NNT/NNR
red circles: station excluded from NNT/NNR
VieVS User Workshop 2011 16
Plots
VieVS/OUT/GLOB/_PLOTS/TEST_OUT/
sou_activity_TEST_LEVEL2.eps
sou_map_TEST_LEVEL2.eps
sources in sessions included in the global adjustment
map of sources,
blue circles: source included in NNR
red circles: source excluded from NNR
17
Processing of a long time span of data (20 years or more)
Problems
if e.g. EOP would be considered as global parameters (stacking at the overlapping time steps), the final N matrix would be too huge – not possible to handle it at a normal PC
position of some sources is not stable enough to handle it as a constant value
unequal observation time at the stations – at some stations the observation interval is not sufficient to estimate reliable velocity
VieVS User Workshop 2011
18
Processing of a long time span of data (20 years or more)
Problems
if e.g. EOP would be considered as global parameters (stacking at the overlapping time steps), the final N matrix would be too huge – not possible to handle it at a normal PC
position of some sources is not stable enough to handle it as a constant value
unequal observation time at the stations – at some stations the observation interval is not sufficient to estimate reliable velocity
VieVS User Workshop 2011
19
Processing of a long time span of data (20 years or more)
Problems
if e.g. EOP would be considered as global parameters (stacking at the overlapping time steps), the final N matrix would be too huge – not possible to handle it at a normal PC
position of some sources is not stable enough to handle it as a constant value
unequal observation time at the stations – at some stations the observation interval is not sufficient to estimate reliable velocity
VieVS User Workshop 2011
VieVS User Workshop 2011 20
In these cases the concerned parameters have to be excluded from the global adjustment by a session-wise reduction.
The vector x2 we get from a so-called back solution for each session:
Covariance matrix
Reduction
2
1
2
1
2221
1211
b
b
x
x
NN
NN
globally estimated p.
reduced p.
After running VIE_GLOB we know the vector x1 (global parameters)
1211222
1222 xNNbNx
122121121
122
12222
NNQNNNQ
VieVS User Workshop 2011 21
Reduction
If the button in the 1st GUI of VIE_GLOB is ticked, all needed parameters for a back solution will be stored in: VieVS/OUT/GLOB/BACKWARD_DATA/TEST_OUT/
i.e., N21, N22 and b2 for each session + some other information, e.g. description of the columns
After finishing of VIE_GLOB run VieVS/OUT/GLOB/backward_solution.min the source code you have to specify the name of the TEST_OUT and TEST_LEVEL2 directories and which reduced parameters you want to estimate
The estimated values will be written into .txt files in VieVS/OUT/GLOB/BACKWARD_SOLUTION/TEST_OUT/ as
ant_TEST_LEVEL2.txt
eop_TEST_LEVEL2.txt
sou_TEST_LEVEL2.txt
tgr_TEST_LEVEL2.txt
zwd_TEST_LEVEL2.txt
VieVS User Workshop 2011 22
Backward solution: zwd
The estimates for each station can be plotted with:
VieVS/OUT/GLOB/plot_backward_zwd.m
VieVS User Workshop 2011 23
Backward solution: tgr
The estimates for each station can be plotted with:
VieVS/OUT/GLOB/plot_backward_tgr.m
VieVS User Workshop 2011 24
Backward solution: EOP
The estimates for each EOP can be plotted with:
VieVS/OUT/GLOB/plot_backward_eop.m
VieVS User Workshop 2011 25
Backward solution: antenna coordinates
The estimates for each station can be plotted with:
VieVS/OUT/GLOB/plot_backward_ant.m
VieVS User Workshop 2011 26
Backward solution: source coordinates
The estimates for each source can be plotted with:
VieVS/OUT/GLOB/plot_backward_sou.m