Velocity Fields From Geodesy to Geodynamics G. Stangl Federal Office of Metrology and Surveying (BEV)‏ + Space Research Institute, Austrian Academy of Science

Structure  The Principle of GPS  From Measurements to Coordinates  Coordinates and other Products  Time Series of Coordinates  Time Series Products  Station Velocities and Velocity Fields  Interpretation of Velocity Fields  Linking Geosciences Seminar Uni Wien

GPS - System  Satellites  Orbits designed to have at least 4 in view every time (excluded polar regions)  Signals:  2 frequencies (L MHz, L MHz)  2 codes (C/A, P) submitting time of transmission t 0  Arrival time t 1 recorded  c(t 1 - t 0 ) = (pseudo-)range  Phase of arriving signal ϕ 1 can be measured, but is ambiguous (ϕ 0 and nλ unknown) © Wikimedia Commons Seminar Uni Wien

GPS - Basics Positions: Solution of intersecting (at least) 3 spheres Accuracy depending on geometry, accuracy of signals and choice of models © Wikimedia Commons Seminar Uni Wien GPS concept is based on real time positioning !!! (but may be ‘misused’ for post processing

GPS - Meters to Millimeters  Geometric limits:  Code resolution 1 m  Phase resolution 1 mm  Problem phase ambiguities  Positioning error Environmental Influences: Ionosphere, Troposphere, Stability of equipment and Marker Order ~ 20 m © Wikimedia Commons Seminar Uni Wien

Coordinates from Measurements  Station coordinates by intersection of pseudoranges, phase measurements for mm-precision  PPP (Precise Point Positioning):  Apply best models of orbits, clocks, environment for single positioning  + independent absolute values  - full impact of model errors  Baseline Strategy:  Reduce model errors by differencing synchronous measurements of 2 points  + model errors decreased by factor 100  - position relative to a priori values Seminar Uni Wien

Reference of Coordinates  3D Cartesian coordinates X, Y, Z  Referring to reference frame of precise orbits (ITRF2005/IGS05, ITRF2008/IGS08)  Underlying reference system has connection to real Earth (centre, pole, meridian) –> Earth-fixed system  Markers relate coordinates to solid Earth  Tectonic plates are defined by a cloud of markers ->pole + rotation © Wikimedia Commons Seminar Uni Wien

Analysis of Measurements  Designed to parameters wanted:  Positioning: Coordinates and troposphere estimated, orbits fixed  Weather forecast: Tropospheric zenith delays estimated, coordinates fixed  Ionosphere: TEC (Total Electron Content) estimated, all other variables removed (2 frequencies)  Orbit determination: Markers fixed  Always a dozen of standard models applied! Seminar Uni Wien

Main Products  Coordinates  Zenith delays  TEC Maps AREF09 COMBINED SOLUTION TO EUREF AUSTRIA ETRF FEB-10 14: LOCAL GEODETIC DATUM: ETRF2000 EPOCH: :00:00 NUM STATION NAME X (M) Y (M) Z (M) FLAG 829 AMST 00000S A Seminar Uni Wien

Time Series of Coordinates  Daily results condensed to weekly ones  Velocities estimated as slopes of regression  Time varying reference frame, constraints at least to 3 reference points (positions + velocities, Helmert transformation 3-7 parameters)  First run ‘Raw Time Series’ -> outliers (=erratic residuals above a certain threshold), discontinuities (=sudden jumps within time series), non-linear trends (=seasonal variations, offsets camouflaged)  ‘Cleaned Time Series’ after removal of outliers  Handling of discontinuities not yet standardized, mostly time slices with common velocities assumed Seminar Uni Wien

11 Raw Time Series Time Series Analysis Space Domain Cleaned Time Series Model change GPS week 1400 Overlayed by receiver malfunction weeks Outliers Seminar Uni Wien

Not Cleaned - False Velocities Raw v X = mm/y v Y = 17.1 mm/y v Z = 8.8 mm/y Cleaned v X = mm/y v Y = 10.7 mm/y v Z = 9.9 mm/y Seminar Uni Wien

13 Time Series Analysis Frequency Domain  Identical sites PATK -> PAT2‏  Model change at week 1400  Decrease of annual amplitude by 50% without physical meaning  Seasonal signals remain  Related to troposphere  What’s left? Annual Term: Up Component PATK: 12.1mm | PAT2: 5.9 mm Seminar Uni Wien

Global Velocity Field Seminar Uni Wien No-net rotation, details from GPS and DORIS Used to derive plate rotations Horizontal (2D) velocities 14

EUREF Network EPN Main purpose stability of coordinates Rotation of Eurasian Plate removed System ETRS89 Attached to ITRS Realization from ITRFyy as densification Established 1996 Definition 1989 based on idea of ‘Stable Eurasian Plate’ Seminar Uni Wien

EPN Time Series Solution No V N [mm/y] RMS VN [mm/y] V E [mm/y] RMS VE [mm/y] V U [mm/y] RMS VU [mm/y] Solution Epoch [DDD/YY] / / /99 EPN CUMULATIVE STATION VELOCITIES NORTH/EAST/UP VELOCITY COMPONENTS IN ITRF2005 NORTH/EAST/UP VELOCITY COMPONENTS IN ETRF2000 Izmit Earthquake <50 km apart Seminar Uni Wien

17 OLG Network Analysis MON OLG = Observatory Lustbuehel Graz (joint international label for AAS and BEV) Monitoring Oriental Network Transition zone Eurasia ->South + Arabian Plate Seminar Uni Wien

OLG Network Analysis CERGOP CERGOP = Central European Research Geodynamics Project since 1994 Mining District Adriatic Microplate Seminar Uni Wien

19 OLG Network Analysis ALBPOS Started in 2010 Transition zone between Dinarides and Adriatic Microplate No Velocities yet Seminar Uni Wien

20 Boundary Zone MON+CERGOP More stations at Bulgaria, however: transition remains unclear Seminar Uni Wien Aegean Plate not covered Boundary of large pull-back force? Big change in strain rate

21 Typical Earthquake Offsets Seminar Uni Wien  Post-seismic deformation at stations near (<50 km) the epicentre  Hypocentre shallow or medium (<40 km)  Co-seismic movements detectable with resolution 1 second  Deformation compliant to those in Hollenstein (2006) for the same region

22 Questions and Problems (1)  Numerical Problems:  Precision of products not standardized, unlikely increase by simply dividing by sqrt(n), velocities’ formal error presently at ±0.1 mm/year, but in reality of ± mm/year (systematic errors, non-Gaussian distribution)  Interpolation of station velocities to velocity fields at a regular grid? No, because of abrupt changes (no geophysical meaning anymore)!  How to compute strain rates without regular grid? Like (Reilinger et al. 2006), forming polygons? Between neighbouring stations? Seminar Uni Wien

23 Questions and Problems (2)  Interdisciplinary Problems:  Direction of movement, amount of coseismic offset and focal mechanism compliant at this example region, but what’s about the representativeness? How to combine velocities west and east of the Ionian Basin to one field (Microplate)?  Contradiction between interseismic (stable?) movements extrapolated for a whole earthquake cycle and energy release during major earthquakes in the order of up to 100 (all velocities too fast), how to resolve it (start/stop, other ways of transmission of energy?  How to separate geokinematic provinces with given station velocities, pure mathematics or geology or seismology or combined? Seminar Uni Wien

24 OLG Network Analysis AMON (1) Local Outliers Adriatic Microplate Escapement ? Seminar Uni Wien

25 OLG Network Analysis AMON (2) Local Outlier Mixture of Alpine uplift, sinking water level and model errors Seminar Uni Wien

26 Questions and Problems of AMON Seminar Uni Wien  Interpretation of AMON results:  NE horizontal movement between Alps and Pannonian Basin, but Hungarian stations point to NW  Northern Forelands typically Eurasian, Adriatic Microplate pressing in the South, but what’s going on at the Tauern mountains?  Can we split up velocities by components with help of gravity, hydrology and meteorology?

More Questions? Thank you