AGU Fall Meeting, San Francisco, USA, 15 19 December 2008 Influence of the Reference Frame Alignment on Station Positions and Velocities: Global or Regional? J. Legrand(1), N. Bergeot(1), C. Bruyninx(1), G. Woppelmann(2), M.-N. Bouin(3), and Z. Altamimi(4) (1) Royal Observatory of Belgium (2) UMR LIENSS, Université de La Rochelle-CNRS, France (3) CNRM / Centre de Météo Marine, France (4) LAREG/IGN, France
Motivations Is it necessary to process a global GNSS network in order to estimate reliable site positions & velocities? What is the impact of processing a global network instead of a regional network? on positions on velocities What is the accuracy limit of a regional solution? Goal: Highlight and quantify the network effect in a regional network
Previous Work on Station Positions
Regional Network Impact on Station Positions (1) ~ 1 year of data 42 EUREF stations (24 also IGS reference stations) Global Network regional network + 47 global IGS reference stations
Impact on Station Positions (2) Regional or Global positions expressed in ITRF2005 under minimal constraints using different selections of reference stations Different regional position solutions can show biases (up to the cm-level in horizontal and vertical components) with respect to each other Global position solutions: differences < 3 mm, < 1 mm in Europe Differences between regional and global position solutions can reach 1 cm in horizontal and 2 cm in vertical Regional Network: very sensitive to the network effect Global Network: much more stable
Velocity Fields
Data and Network ULR reprocessing: contribution to TIGA project [Wöppelmann et al., 2007] Weekly global SINEXs from 1997-2006.9 220 continuous GPS stations 205 stations with >3.5 years of data Weekly SINEXs expressed in ITRF2005
Extraction of Regional Weekly SINEXs Extraction Global weekly SINEXs Regional weekly SINEXs Same coordinates and covariance information for the common stations!!
Velocity Field Estimations Regional and Global weekly SINEXs Stacked with C A T RE F Software [Altamimi] Reject outliers and properly handle discontinuities Regional and Global cumulative solutions (positions & velocities) Datum definition : expressed in ITRF2005 under minimal constraints approach using a selection of reference stations Good agreement between the solution and the ITRF2005 at least 3 years of data in the ITRF2005 and in the ULR timeseries
Initial Results Several sets of reference stations were tested Confirms the results for the station positions Global network: differences < 0.2 mm/yr => behave in a stable way Regional network: more sensitive to the set of reference stations (outliers and geometry) Quantify the network effect with 3 representative solutions
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional Network: # 60 stations Datum: 23 reference stations Regional solution B: Regional Network: # 60 stations Datum: 14 stations located only on the European continent
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional Network: # 60 stations Datum: 23 reference stations Regional solution B: Regional Network: # 60 stations Datum: 14 stations located only on the European continent
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional Network: # 60 stations Datum: 23 reference stations Regional solution B: Regional Network: # 60 stations Datum: 14 stations located only on the European continent
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional Network: # 60 stations Datum: 23 reference stations Regional solution B: Regional Network: # 60 stations Datum: 14 stations located only on the European continent
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional Network: # 60 stations Datum: 23 reference stations Regional solution B: Regional Network: # 60 stations Datum: 14 stations located only on the European continent
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional solution B: Regional Network: # 60 stations Datum: 23 reference stations Regional Network: # 60 stations Datum: 14 stations located only on the European continent AGU Fall Meeting, San Francisco, USA, 15 19 December 2008
3 Representative Solutions Definition of 3 different solutions: 1 Global solution: Global Network: # 220 stations Datum: 83 reference stations geographically well-distributed 2 Regional solutions: Regional solution A: Regional solution B: Regional Network: # 60 stations Datum: 23 reference stations Regional Network: # 60 stations Datum: 14 stations located only on the European continent AGU Fall Meeting, San Francisco, USA, 15 19 December 2008
3 Representative Solutions VGLOB VREGA VREGB AGU Fall Meeting, San Francisco, USA, 15 19 December 2008
Agreement with ITRF2005 V GLOB V REGA RMS of the differences between ITRF2005 and the solution for All the reference stations All the regional IGS05 reference stations Number of Stations V REGB E (mm/yr) N (mm/yr) U (mm/yr) GLOB 83 0.50 0.32 1.43 REG A 23 0.36 0.35 0.89 REG B 14 0.41 0.20 0.70 GLOB 26 0.62 0.30 1.54 REG A 26 0.58 0.31 1.50 REG B 26 0.66 0.49 1.64
Comparison Between Global and Regional Velocity Fields Comparison of the 3 velocity fields V GLOB V REGA and V REGB Direct comparison Helmert transformation Euler pole
Difference Between Global and Regional Velocities(1) V GLOB -V REGA : Horizontal 0.3 ± 0.4 mm/yr max: 0.9 mm/yr
Difference Between Global and Regional Velocities(2) V GLOB -V REGB : Horizontal 0.6 ± 0.7 mm/yr max: 1.3 mm/yr
Difference Between Global and Regional Velocities(3) V GLOB -V REGA : Vertical 0.3 ± 0.5 mm/yr max: 1.2 mm/yr
Difference Between Global and Regional Velocities(4) V GLOB -V REGB : Vertical 0.1 ± 1.0 mm/yr max: 2.9 mm/yr
Helmert Transformation Helmert transformation between global and regional velocities: Allows to explain almost all the differences Residuals depend only on the size of the network Transformation parameters depend only on the reference stations used Horizontal Vertical RMS: 0.01 mm/yr MAX: 0.2 mm/yr RMS: 0.1 mm/yr MAX: 0.6 mm/yr
Euler Pole Estimation Euler pole estimation of the Western part of Europe 40 stations: continuously observed during at least 3 years; formal error < 1.5 mm/y; post-fit velocity residual <1.5 mm/yr Residual velocity fields
Impact on the Relative Velocity Fields VR GLOB -VR REGA VR GLOB -VR REGB RMS: 0.17 mm/yr MAX: 0.5 mm/yr RMS: 0.21 mm/yr MAX: 0.8 mm/yr Systematic effect due to Translation rates
Impact on the Relative Velocity Fields VR GLOB -VR REGA VR GLOB -VR REGB Residuals after Helmert transformation RMS: 0.17 mm/yr MAX: 0.5 mm/yr RMS: 0.21 mm/yr MAX: 0.8 mm/yr Systematic effect due to Translation rates RMS: 0.01 mm/yr AGU Fall Meeting, San Francisco, USA, 15 19 December 2008 MAX: 0.2 mm/yr
Summary Network effect causes discrepancies between two regional solutions or between regional and global solutions This effect can reach: Positions: cm level Velocities: mm/yr level Horizontal relative velocities: 0.5 mm/yr Regional and global networks can lead to different geodynamic interpretations