geodetic reference frames in presence of crustal deformations

Geodetic Reference Frames In Presence of Crustal Deformations

Martin Lidberg1,2, Maaria Nordman3,

Jan M. Johansson1,4, Glenn A Milne5,

Hans-Georg Scherneck1,Jim L Davis6

1Onsala Space Observatory, Chalmers, Sweden2Lantmäteriet (National Land Survey of Sweden)

3Finnish Geodetic Institute, Finland4Swedish National Testing and Research Institute

5University of Durham, UK6Harvard-Smithsonian Centre for Astrophysics

EGU 2008, 14-18 April, Vienna, Austria

Effects of the land uplift

First determination of the land uplift related to sea level apparent land uplift

Determination of horizontal rates and absolute land uplift values;the purpose of BIFROST

Outline

• The BIFROST GPS network• GPS analysis• Reference frame realization• Time series analysis and data

editing• Evaluation of the velocity field• Address noticed problems and

possible causes• Conclusions

The analysis includes:

• Public sites from IGS and EPN (EUREF Permanent Network) (blue dots )

• Sites not in the public domain (yellow diamonds )

Totally: 83 sites

Period of analysis:Aug. 1993 – Oct. 2006

The extended BIFROST network

GAMIT / GLOBKGAMIT (GPS analysis)• Traditional analysis strategy• 10° elevation cut off angle• Trop. zenith delay & gradients

• the Niell 1996 mapping functions

• Relative antenna PCV values (“absolute” PCV not used so far)

• a priori orbits from SCRIPPS

GLOBK (combination & ref. frame)

• combination of sub-networks• reference frame realization.• Combine the regional

analysis with “complete IGS analysis” from SCRIPPS (GAMIT h-files).

• Satellite orbits are given loose constraints in the quasi-observations.

GPS analysis strategy

GIPSY• Precise Point Positioning (PPP) using JPL products• And ambiguity fixing• Used to validate the GAMIT/GLOBK solution

ITRF2000:43 “good” sites as candidates for the daily stabilization. in GLOBK.

ITRF2005:78 sites as candidates.- Include breaks from ITRF2005 coordinate list

Combination for “global solution” & reference frame realization

(ITRF2000) Results from stabilization (ITRF2005)

Sca

le (

ppb)

post

fit R

MS

(m

m)

No

used

site

s

Example of time series of GPS positions

De-trended position time series from Vilhelmina (64°N) for the complete period Aug. 1993 – Oct 2006

1993-1996: - some “bad” antenna radoms

PROBLEMS !!!???Non-linear time-series in the vertical:- Bent “banana”-shape ???- Or rate change by 2003 ???

Tide model problem??

Watson, Tregoning, Coleman (2006) “Impact of solid earth tide models on GPS coord. and trop. time series”, GRL.

The global network:35 selected sites. (black squares )- Cover the globe - Connect regional & global analysis- Include “good” sites for reference frame realization

Reference frame sites:23 sites as candidates (yellow circles )

New combination - with my own global analysis

Combined Regional BIFROST + “35-site” global

- Vertical “banana-shape” heavily reduced!!! (maybe not eliminated..)

- In the analysis we have “stable” sites with +10 yr observations, and “new” sites (< 5 yr).

Two step reference frame approach1. Determine pos+vel for 27 “good sites” (Swe+Fin+some EPN) from “stable” period 1998 to 2004 (7yr). 2. Apply 6-par transf. (no scale) of all daily solutions to the “regional” frame defined above.

After common mode reduction using daily transformations

Derived velocity field relative to Eurasia

Red: ITRF2000 (eura)Green: ITRF2005 (eura)

ITRF2000: removed the Eurasia plate tectonic motion using the ITRF2000 Euler pole for Europe

ITRF2005: transformed (rotated) to the ITRF2000_eura velocities

RMS of velocity at 7 European IGS sites: 0.5 mm/yr level

-> suggest a successful reference frame realization.

For POTS and METS; “my” velocities and official ITRF2005 agree by: North: 0.2 mm/yrEast: 0.3 mm/yrUp: 0.3 mm/yr

Glacial Isostatic Adjustment (GIA)

Glaciations cause change of load from ice and ocean to the earth. This load cause deformation of the earth shape and mass re-distribution.

Updated GIA model (Milne et al 2001)Ice history model from Lambeck

120 km lithosphere, upper mantle visc. 51020 Pas lower mantle visc. 51021 Pas

Thanks Glenn Milne, Maaria Nordman, Pippa Whitehouse!!

Vertical velocity

ITRF2005ITRF2000GIA model Ekman (1998) based on:• mareographs and

levellings,•1.2 mm/yr eustatic sea

level rise•change of the geoid based

on Ekman & Mäkinen (1998)

Compared to the ITRF2000values: mean Std

(mm/yr)ITRF2005 0.4 0.1GIA model -0.4 0.5 Ekman -0.4 0.6

The newGAMIT/GLOBK solution (in ITRF2005)And GIA model

The new station velocities

Validation using GIPSY (0.1, 0.1, 0.2) (n,e,u) mm/yr

New GIA model minus GPS,“best sites”(0.3, 0.2, 0.3) (n,e,u) mm/yr

0

45

90

135

180

225

270

315

0

45

90

135

180

225

270

315

Borås: Lat. 57.7° Kiruna: Lat. 67.9°

Satellite geometry at two locationsWill this cause error in height, if e.g. elevation dependence (antenna models) and mapping functions are imperfect?

The “hole” in the sky plot at high latitude sites!

Vilhelmina, 64°N

BIFROST+global SOPACIERS1996 tide modelRelative antenna models

BIFROST+Own global 35 site netMixed IERS1996+IERS2003 tide modelRelative antenna models

45 global+35 EPN sitesIERS2003 tide modelAbsolute antenna modelsGMF mapping function(Only every 10 day)

GPS-velocities and GIA-model agree at- 0.4 mm/yr level (1) horizontal - 0.5 mm/yr level (1) vertically

GIA model explain observed velocities to 0.5 mm/yr!!

OK for coordinate transformation and “geo-referencing” purposes; but for sea level work (exploring GPS and Tide gauge obs.) we must be careful and continue the work regarding: - “long term stability in GNSS results” - improvements in the reference frame!in order to reached the 0.1 mm/yr goal!

Proper velocity estimation need correct models in the GPS analysis!-Therefore the re-processing of IGS, with “best” models is very important (abs. antenna models, tide model, mapping function, ionosphere..)- Will be a base for the next improved ITRF200x!

Conclusion and outlook