Time series analysis of GNSS-SLR co-located stations International GNSS Service Workshop 2012 23 - 27 July 2012, Olsztyn, Poland K. So nica, D. Thaller , R. Dach, L. Ostini, A. Jäggi, S. Lutz, G. Beutler C. Rodriguez-Solano, P. Steigenberger, U. Hugentobler M. Fritsche, R. Dietrich K. Wang, M. Rothacher ś Astronomical Institute, University of Bern, Switzerland Institut für Astronomische und Physikalische Geodäsie, TU München, Germany Institut für Planetare Geodäsie, Technische Universität Dresden, Germany Institute of Geodesy and Photogrammetry, ETH Zürich, Switzerland Fig. 2 Assessment of the impact of Blue-Sky effect on SLR stations using mean APL vertical correction applied on SLR stations (the size of circles denotes number of weekly solutions). Units: mm Contact address K Sośnica rzysztof Astronomical Institute, University of Bern Sidlerstrasse 5 3012 Bern (Switzerland) krzysztof.sosnica@aiub.unibe.ch 4. Impact of APL corrections Is the consistency between SLR and GNSS solutions improved by applying the APL corrections at the observation level? Currently, the APL corrections are not recommended for inclusion in operational space geodetic solutions and they were not used for deriving the latest realisation of ITRF. The Blue-Sky effect is, however, a limiting factor for the consistency between SLR and GNSS solutions. Table 1 compares the local ties at the co-located stations with the GNSS-SLR solutions whenAPL corrections are applied or omitted. Two co-locations do not seem to have proper tie measurements, namely Riga and San Fernando. APL corrections improve the consistency between estimated and measured ties by approximately 0.2 mm. But for the stations with moderate impact of APL the improvement is larger, e.g. from 9.0 mm to 8.1 mm for Borowiec, from 4.2 mm to 3.8 mm for Zimmerwald, and from 4.0 mm to 2.8 mm for Beijing. The SLR stations with the largest APL impact have unfortunately no co-locations with GNSS. 7. Summary ! ! The Blue-Sky effect may assume values up to 4.4 mm for in-land stations, Applying APL corrections improves the consistency of SLR and GNSS solutions, eliminates the impact of the Blue-Sky effect, and reduces the amplitudes of the annual signals of station and geocenter coordinates in both, SLR and GNSS solutions. Poster compiled by K. So , July 2012 Astronomical Institute, University of Bern, Bern krzysztof.sosnica@aiub.unibe.ch śnica 1. Introduction We analyse the 17-year time series of GNSS and SLR station coordinates stemming from the reprocessing project „ ” jointly carried out by four universities: Technische Universität Dresden (TUD), Eidgenössische Technische Hochschule Zürich (ETHZ), Universität Bern, Astronomisches Institut (AIUB) andTechnische Universität München (TUM). The positions of 70 SLR stations are derived using laser measurements to LAGEOS-1 and LAGEOS-2 for the time span 1994.0-2011.0. The positions of 340 GNSS stations are derived from GPS-only solutions for 1994.0-2002.0 and combined GPS-GLONASS solutions for 2002.0- 2011.0. Geodätische und geodynamische Nutzung reprozessierter GPS-, GLONASS- und SLR-Daten 3. Blue-Sky effect The omission of Atmospheric Pressure Loading (APL) may lead to inconsistencies between optical (SLR) and microwave (GNSS, VLBI, DORIS) techniques. SLR observations can be carried out only during good weather conditions, whereas microwave observations are weather- independent. Weather dependence of the optical observations causes the so-called Blue-Sky effect, i.e., a systematic shift of the station height. Applying APL corrections compensate the Blue- Sky effect, to some extent. We estimate the impact of the Blue-Sky effect on all SLR stations (see Fig. 2). The Blue-Sky effect has the largest impact on the continental stations, i.e. on Golosiv, Ukraine (4.4 mm), Wuhan, China (3.2 mm), Beijing, China (2.5 mm), Altay, Russia (2.3 mm). For the best performing stations, the Blue-Sky effect exceeds 1 mm, e.g., for Zimmerwald, Switzerland (1.2 mm), Wettzell, Germany (1.2 mm), Hartebeesthoek, South Africa (1.1 mm). Thus, the reduction of the Blue-Sky effect is of crucial importance, e.g., for the Global Geodetic Observing System (GGOS), in which the required accuracy of station coordinates is 1 mm for all techniques of space geodesy. 2. FODITS analysis The time series of GNSS and SLR station coordinates are analysed using a new program of the Bernese GNSS Software, called FODITS (Find Outliers and Discontinuities In Time Series). In this program all statistically significant station events are detected, i.e., station discontinuities induced by technical and environmental sources, velocity changes (caused typically by earthquakes), outliers, and periodicities (annual and semi-annual signals of coordinate time series, see Section 4). Figure 1 shows the analysis of three co-located stations. Most of the detected events are system-specific. Fig. 1 Time series of GNSS-SLR co-located stations: Zimmerwald, Switzerland (top), Borowiec, Poland (middle), and Monument Peak, California (bottom). Daily GNSS station coordinates are shown on the left, whereas weekly SLR station coordinates are shown on the right. For the Zimmerwald no discontinuities and velocity changes are detected. Only annual signal and a few outliers are significant for the Zimmerwald GNSS station. For the Borowiec GNSS station one discontinuity in 1998 is detected (due to the interference with a cell phone relay), and a significant annual amplitude is detected, as well. There are no statistically significant events for Borowiec SLR station. For the Monument Peak GNSS station, a discontinuity in 2010 is detected (of unknown reason) and for the SLR station two discontinuities and one velocity change are detected (caused by range biases due to a time interval counter). The annual signal is significant only for the SLR Monument Peak station. Posters and other publications from the AIUB´s Satellite Geodesy Group: http://www.bernese.unibe.ch/publist Fig. 3. Comparison between GNSS-SLR local ties and station coordinate differences derived from solutions with and without APL corrections. Units: mm Fig. 4 Annual amplitudes of height components for SLR- GNSS co-located stations for the solutions with and without APL corrections. Units: mm Fig. 5. Spectral analysis of geocenter Y coordinate time series in the SLR and GNSS solutions with APL corrections and without APL corrections. Fig. 6. Differences of geocenter coordinate estimates in SLR and GNSS solutions due to APL corrections. Units: mm 5. Annual amplitudes of station coordinates Figure 4 shows the annual amplitudes of station heights for SLR and GNSS co-located stations. Solutions with APL and without APL corrections are presented. For some co-located stations the agreement between the GNSS- and SLR-derived amplitudes is rather poor (e.g., for Graz, McDonald and Monument Peak), implying that the amplitudes are influenced by technique- specific problems and data processing issues, and they do not show any geophysical or environmental effects. On the other hand, for stations Greenbelt, Tahiti, San Fernando, and Hartebeesthoek the agreement between the amplitudes is at the sub-mm level. The amplitudes of the height component are usually smaller for the SLR stations (on averrage 2.6 mm and 2.3 mm for the solution without APL and with APL corrections, respectively) than for the GNSS stations (3.5 mm and 2.8 mm for the solution without APL and with APL corrections, respectively). Larger variations of the vertical component in GNSS can be explained by correlations between the height component and other estimated parameters, e.g., station clock corrections and troposphere delay. 6. Geocenter coordinates Figure 5 shows that the APL corrections reduce the annual amplitude of geocenter coordinates. The impact of APL corrections is, however, different for the X and Y components for SLR and GNSS (see Fig. 6). This is caused by the global distribution of SLR stations. The network of SLR stations is unbalanced; the majority of high performing SLR stations is located along the X axis. SLR stations located along the Y axis are either coastal stations with minor impact of APL or low performing in-land stations. The GNSS network is to a large extent well balanced. Co-location Local tie Difference of position between local tie and the solution Station GNSS SLR Weeks dX dY dZ Without APL With APL Graz, Austria GRAZ 7839 513 -2.558 8.516 -1.321 12.1 11.9 McDonald, Texas MDO1 7080 496 22.394 8.467 23.408 9.4 9.4 Monument Peak, California MONP 7110 482 31.365 -5.456 20.526 9.1 9.7 Zimmerwald, Switzerland ZIMM 7810 470 13.506 5.986 -6.420 4.2 3.8 Yarragadee, Australia YAR2 7090 467 -18.612 -12.467 -5.841 4.5 4.9 Greenbelt, Maryland GODE 7105 456 54.230 97.009 93.863 4.1 3.7 Wettzell, Germany WTZR 8834 415 3.824 68.202 -15.518 6.7 5.9 Matera, Italy MATE 7941 346 -29.157 -22.201 37.912 10.2 10.4 Hartebeesthoek, South Africa HARB 7501 345 -743.471 1994.877 207.587 3.7 3.8 San Fernando, Spain SFER 7824 345 45.041 -35.273 -89.594 97.8 97.9 Concepcion, Chile CONZ 7405 286 -25.449 35.349 -74.042 7.2 8.1 Grasse, France GRAS 7845 233 -1.173 -81.348 5.620 4.8 5.0 Borowiec, Poland BOR1 7811 217 25.767 -72.908 -0.324 9.0 8.1 Mt Stromlo, Australia STR1 7825 217 -38.054 4.584 58.108 12.2 11.7 Beijing, China BJFS 7249 199 16.517 -118.317 146.279 4.0 2.8 Tahiti, French Polynesia THTI 7124 184 -8.456 24.551 -28.299 23.8 23.8 Riga, Latvia RIGA 1884 162 3.401 -18.661 6.963 51.7 50.0 Arequipa, Peru AREQ 7403 153 18.614 -0.547 21.499 3.0 2.7 Potsdam, Germany POTS 7836 141 50.091 95.219 -40.438 3.9 4.4 MEAN 14.8 14.6 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 Graz, Austria McDonald, Texas Monument Peak, California Zimmerwald, Switzerland Yarragadee, Australia Greenbelt, Maryland Matera, Italy Hartebeesthoek, South Africa San Fernando, Spain Concepcion, Chile Grasse, France Borowiec, Poland Mt Stromlo, Australia Beijing, China Tahiti, French Polynesia Riga, Latvia Arequipa, Peru Potsdam, Germany Wettzell, Germany GNSS w/o APL SLR w/o APL GNSS with APL SLR with APL Acknowledgements We would like to thank Swiss National Science Foundation (SNF) for the financial support within the SNF Project 200021E-131228