AIUB

AGU2019: G21C-0748American Geophysical UnionFall Meeting 201909–13 December 2019, San Francisco, CA

R. Dach1, I. Selmke2, D. Arnold1, L. Prange1, S. Schaer1,3, D. Sidorov1,P. Stebler1, A. Villiger1, A. Jäggi1, U. Hugentobler2

(1) Astronomical Institute, University of Bern, Bern, Switzerland(2) IAPG, Technical University Munich, Munich, Germany(3) Swiss Federal Office for Topography, swisstopo, Wabern, Switzerland

Poster compiled by Rolf Dach, December 2019Astronomical Institute, University of Bern, BernEmail rolf.dach@aiub.unibe.ch

CODE Contribution toIGS Repro3 Campaign

IntroductionCODE (Center for Orbit Determination in Europe) is a joint venture of thefollowing four institutions:

• Astronomical Institute, University of Bern(AIUB), Bern, Switzerland

• Federal Office of Topography swisstopo,Wabern, Switzerland

• Federal Agency of Cartography and Geodesy(BKG), Frankfurt a. M., Germany

• Institut für Astronomische und PhysikalischeGeodäsie, Technische Universität München(IAPG, TUM), Munich, Germany

It is acting as a global Analysis Center of the IGS, the International GNSSService (Johnston et al. 2017) . The operational computations are per-formed at AIUB, whereas IGS–related reprocessing activities are usuallycarried out at IAPG, TUM. All solutions and products are generated withthe latest development version of the Bernese GNSS Software (Dach et al.2015).

Previous Reprocessing EffortsCODE already contributed to the previous reprocessing efforts of the IGS,e.g., repro2 (Steigenberger et al. 2014) providing GPS and GLONASS or-bits, Earth rotation parameters, and coordinate solutions up to the end of2014.In the frame of the H2020 funded project European Gravity Service forImproved Emergency Management (EGSIEM, Jäggi et al. 2019), amongothers, another reprocessing effort was performed in order to obtain thefollowing products:

GPS GLONASSGNSS satellite orbits: since 1994 since 2002GNSS satellite clocks, sampling 30 s: since 2000 since 2008GNSS satellite clocks, sampling 35 s: since 2003 since 2010

The full dataset of results (Sušnik et al. 2016) is available atftp://ftp.aiub.unibe.ch/REPRO_2015/ .The EGSIEM-repro does include additional periodic terms in the D-component of the solar radiation pressure model pointing from the satel-lite to the Sun – the so-called ECOM2 (empirical CODE orbit model,Arnold et al. 2015) whereas for the IGS-repro2 the classical ECOM withoutperiodic terms in the D-component was used.

Model Update for the Current Reprocessing EffortFigure 1: Stationnetwork as se-lected by CODEfor the currentreprocessingeffort. Stationswith antennasand radom com-binations (“COPYFROM NONE”) arenot used as far aspossible.

Some important model updates are planned to be introduced with the cur-rent reprocessing effort with respect to the current final product genera-tion at CODE:

• IERS-convention related: new mean pole (R. Ray) and high fre-quency pole model (Desai and Sibois 2016)

• Inclusion of Galileo as the third GNSS constellation:requires an update of receiver and satellite antenna corrections aswell as Galileo-specific orbit models

• Orbit modelling: shift pulses from noon to orbit midnight; requireslong-arc solution (72 hours)

Scheduling of Stochastic PulsesCODE orbits are generated by extracting the middle part of a long-arc so-lution covering 3 days. In order to compensate for potential deficiencies inthe orbit modeling, empirical, instantaneous velocity changes – so calledstochastic pulses – are added every 12 hours (Beutler et al. 1994).In Figure 2a) there are some planes with bigger, others with lower mag-

nitude of the orbit misclosures during the eclipse season. For those withthe lower values the eclise period was close to noon. For that reason an-other solution was computed where the pulses are scheduled at the biggestdistance of the satellite from the Sun (orbit midnight, Figure 2b) . As Fig-ure 2c) shows, this results in a reduction of the misclosures by about 10%.

a) stochastic pulses every 12 hours b) stochastic pulses at orbit midnight c) difference (green: better for orbit midnight)

Figure 2: Orbit misclosures for GPS satellites during year 2014. The satellites are listed according to their PRN and the orbital planes.

Earth Rotation ParametersOne-day short-arc solutions

a) setup as in CODE’s final solution(stochastic pulses every 12 hours)

b) setup as in CODE’s repro3 solution(stochastic pulses at orbit midnight)

Three-day long-arc solutions

c) setup as in CODE’s final solution(stochastic pulses every 12 hours)

d) setup as in CODE’s repro3 solution(stochastic pulses at orbit midnight)

Figure 3: Amplitude spectra of the differences between system-specific polar motion estimates and theCO4-series during the year 2014.

The Earth rotation parameters (daily offset and rates for X-and Y-component as well as Length of day, LOD) are es-timated in a global solution individually per GNSS (GPS,GLONASS, Galileo; as proposed by Scaramuzza et al. 2018).The difference with respect to the CO4-series is computedand the related spectra are shown in Figure 3 .In repro3 solution series (right hand plots) a new high-frequency pole model (Desai and Sibois 2016) is used. Inthe operational final solution (as for all other space geode-tic techniques where the CO4-series is based on) still the oldIERS2010 model (left hand plots) is applied. For that reasonthe direct interpretation of these differences is difficult.As Abraha et al. (2017) have shown, deficiencies in tidal effectmodelling may map in different ways for individual GNSSbecause of different resulting aliasing periods due to the or-bit revolution periods. To check the consistency of the Earthrotation solutions obtained with different GNSS may thus beused as a quality measure.In particular in the one-day short arc solution (top plots inFigure 3) some differences due to the model changes are vis-ible. Which of them is caused by the change in the high-frequency pole model and which by the change of the orbitmodel can be discussed in detail when adding a third solu-tion where only one of the two models is changed.Increasing the length of the orbit arcs helps to reduce thenoise and all amplitudes in the spectra significantly (alreadyreported, e.g., by Lutz et al. 2016) when comparing the topand bottom plots. The estimation of the polar motion param-eters becomes only possible from satellite techniques becauseof the oblateness of the Earth (otherwise only the rates wouldbe accessible). This fact introduces a direct relation betweenthe stability of the orbits and the quality of the obtained Earthrotation parameters.It is clearly visible in all the plots that the Earth rotation pa-rameters obtained from GLONASS show a particular peak at(nearly) eight days that is not visible for the GPS-based series.The period of eight sidereal days is known as the repetitionrate of the GLONASS ground tracks. The similar period forGPS is one sidereal day.

Station Coordinate SeriesThe datum definition of the daily solutions have been realized by No-Net-Rotation conditions (NNR) with respect to the currently used IGS14 refer-ence frame. A verified list of reference frame sites has been used for thedatum definition.The station coordinates from the daily three-day long-arc solutions aremore consistent to the reference frame because of the related smoothingeffect (see Figure 4).When the different models are enabled step by step, the consistency ofthe three-day solution with the IGS14 reference frame is improved by theupdated antenna corrections (see Figure 5).

Figure 4: Resultsof a Helmert–transformationfrom daily solutionswith respect to theIGS14 reference frameduring the year 2014from the repro3-stylesolution.

Figure 5: Resultsof a Helmert–transformationfrom a series of three-day solutions withrespect to the IGS14reference frame wherethe different modelchanges are enabledstep-by-step duringthe year 2014.

Summary• CODE will contribute to the current reprocessing effort of the IGS

with a multi-GNSS solution including GPS, GLONASS, and Galileo.• The change of the scheduling of stochastic orbit parameters is bene-

fitial for the resulting orbits. This procedure requires long-arc solu-tions (72 hours).

• The longer orbit arcs are also beneficial for the noise in the Earthrotation series. System-specific (GPS and GLONASS) series becomemore consistent than in one-day short-arc solutions.

The inclusion of Galileo requires updated receiver and antenna correctionsthat have extensively been tested within the IGS, see Villiger et al. (G12A-06) or Rebischung et al. (G11A-03) at this conference.

ReferencesK.E. Abraha, F.N. Teferle, A. Hunegnaw, and R. Dach. Effect of unmodelled tidal displace-

ments in GPS and GLONASS coordinate time series. Geophysical Journal International,2017. in press.

Daniel Arnold, Michael Meindl, Gerhard Beutler, Rolf Dach, Stefan Schaer, Simon Lutz, LarsPrange, Krzysztof Sosnica, Leos Mervart, and Adrian Jäggi. CODE’s new solar radi-ation pressure model for GNSS orbit determination. Journal of Geodesy, 89(8):775–791,August 2015. doi: 10.1007/s00190-015-0814-4.

G. Beutler, E. Brockmann, W. Gurtner, U. Hugentobler, L. Mervart, M. Rothacher, and A. Ver-dun. Extended orbit modeling techniques at the CODE processing center of the Inter-national GPS Service for Geodynamics (IGS): Theory and initial results. ManuscriptaGeodaetica, 19(6):367–386, April 1994.

R. Dach, S. Lutz, P. Walser, and P. Fridez, editors. Bernese GNSS Software, Version 5.2. Astro-nomical Institute, University of Bern, Bern, Switzerland, November 2015. ISBN 978-3-906813-05-9. doi: 10.7892/boris.72297. URL ftp://ftp.aiub.unibe.ch/BERN52/DOCU/DOCU52.pdf. User manual.

Shailen D. Desai and Aurore E. Sibois. Evaluating predicted diurnal and semidiurnal tidalvariations in polar motion with GPS-based observations. Journal of Geophysical Research,121(7):5237–5256, 2016. doi: 10.1002/2016JB013125.

A. Jäggi, M. Weigelt, F. Flechtner, A. Güntner, T. Mayer-Gürr, S. Martinis, S. Bruinsma,J. Flury, S. Bourgogne, H. Steffen, U. Meyer, Y. Jean, A. Sušnik, A. Grahsl, D. Arnold,K. Cann-Guthauser, R. Dach, Z. Li, Q. Chen, T. van Dam, C. Gruber, L. Poropat,B. Gouweleeuw, A. Kvas, B. Klinger, J-M. Lemoine, R. Biancale, H. Zwenzner,T. Bandikova, and Shabanloui A. European Gravity Service for Improved EmergencyManagement (EGSIEM) - from concept to implementation. Geophysical Journal Interna-tional, 218(3):1572–1590, 2019. doi: 10.1093/gji/ggz239.

G. Johnston, A. Riddell, and G. Hausler. The International GNSS Service. In P.J.G. Teunissenand O. Montenbruck, editors, Springer Handbook of Global Navigation Satellite Systems,pages 967–982. Cham, Switzerland: Springer International Publishing, 2017. doi: 10.1007/978-3-319-42928-1.

Simon Lutz, Michael Meindl, Peter Steigenberger, Gerhard Beutler, Krzysztof Sosnica, StefanSchaer, Rolf Dach, Daniel Arnold, Daniela Thaller, and Adrian Jäggi. Impact of the arclength on GNSS analysis results. Journal of Geodesy, 90(4):365–378, 2016. doi: 10.1007/s00190-015-0878-1.

Stefano Scaramuzza, Rolf Dach, Gerhard Beutler, Daniel Arnold, Andreja Sušnik, and AdrianJäggi. Dependency of geodynamic parameters on the GNSS constellation. Journal ofGeodesy, 92(1):93–104, 2018. doi: 10.1007/s00190-017-1047-5.

Peter Steigenberger, Simon Lutz, Rolf Dach, Stefan Schaer, and Adrian Jäggi. CODE repro2product series for the IGS, 2014. URL http://www.aiub.unibe.ch/download/REPRO_2013.

Andreja Sušnik, Rolf Dach, Arturo Villiger, Andrea Maier, Daniel Arnold, Stefan Schaer, andAdrian Jäggi. CODE reprocessing product series, 2016. URL http://www.aiub.unibe.ch/download/REPRO_2015.

Contact addressRolf DachAstronomical Institute, University of BernSidlerstrasse 53012 Bern (Switzerland)Email rolf.dach@aiub.unibe.ch