Minutes of the EUREF LAC Workshop (Warsaw, Nov. 18-19, 2010)
by Heinz Habrich, Bundesamt für Kartographie und Geodäsie, Germany
Summary of technical sessions
The 7th EPN Local Analysis Centre (LAC) Workshop was held on November 18-19, 2010 in Warsaw, Poland. It was hosted by the Military University of Technology, Faculty of Civil Engineering and Geodesy, Centre of Applied Geomatics. 37 registered workshop participants and the representation of 15 from currently 17 LACs demonstrated the broad interest in this workshop and enabled profound discussion and approved conclusions. The second day of the event has been reserved for EPN re-processing issues. The workshop program is available at the EPN-CB website.
The technical sessions were opened by reports from the EUREF Technical Working Group (TWG). A general survey of current EPN products as combined station coordinates and troposphere delays has been presented by Heinz Habrich and Wolfgang Soehne. Carine Bruyninx pointed to modernization aspects of the EPN tracking network and summarized the tracking and analysis of GLONASS satellite signals within the EPN. She identified differences of station coordinates up to 4 mm in horizontal and 8 mm in vertical direction by changing the receiver phase center variations from numbers according to the IGS05.ATX file to the new preliminary IGS08.ATX file.
Christof Völksen informed in the second session: "Reports from LACs" about the inclusion of a rapid analysis of the BEK subnetwork and the ongoing re-organization of BEK joining a consortium of several geodetic research institutes in Munich. Günter Stangl presented a project about regional TEC maps for the improvement of EGNOS positions and mentioned a reduction of the personal at OLG for GNSS analyses. Eric Pottiaux presented the wide research spectrum at ROB that includes the LAC activity, EPN re-processing, dense velocity field contribution, regional vs. global network effect studies, and ionosphere and troposphere monitoring developments. Alessandro Caporali gave an overview about LAC UPA and introduced a proposed COST action about plans for a European GNSS observatory. Rosa Pacione informed about the plans to change the software from Microcosm to GIPSY/OASIS for the routine processing. Elmar Brockmann found an increase of the weighted rms values within the daily solutions what he assigned to possible atmosphere mismodeling during the summer. Miroslawa Igondova showed the status of the LAC at SUT, followed by Tomas Liwosz for LAC WUT who pointed to the different ambiguity resolution options used, depending on the baseline lengths. Mariusz Figurski showed the various contributions from MUT to EUREF in an overview, for example MUT's daily EPN analysis and the re-processing preformed with two different software packages.
The second day of the workshop started with reports from the pilot re-processing. Christof Völksen summarized the history of the re-processing activities and presented three alternative options to proceed for the next steps. These options were covered during the round table discussion. Günter Stangl found from its re-processing an improved estimation of the troposphere yielding in smaller repeatabilities for the height component. Rosa Pacione showed results from three slightly different PPP solutions with the GIPSY software, where the troposphere mapping function and the handling of higher order ionosphere terms has been modified between each run. The minimal error for coordinate repeatability was obtained by using the newest mapping function (VMF1) and considering higher order ionosphere terms. Karolina Szafranek compared re-processed DD-network against PPP solutions from MUT that were computed with the Bernese GNSS software. While some stations showed small coordinate differences, there occured significant differences at some other sites. The reprocessing of LAC GOP by Jan Dousa was presented by Christof Völksen. GOP has already reprocessed the period 1996 to 2007, including their EPN network plus 16 additional sites, which were requested by the EPN CB. The re-processed MUT solutions computed with GAMIT software were presented by Andrzej Araszkiewicz. He divided the EPN network into seven subnetworks and processed in total 202 stations. The GLOBK program was used to combine the sub-networks.
Rolf Dach announced the new release 5.2 of the Bernese GNSS software. The analysis model will be extended for SLR processing, handling of grid files such as used for the VMF1 troposphere mapping function and others. It is planned to provide this version until end of 2011 to the users. Ambrus Kenyeres showed differences in phase shift and amplitude of annual signals in EPN coordinate time series in the case the EPN tracking network will be extended for global sites. He is using the approach of stacking normal equations from regional and global solutions and did not use the approach combining regional and global sites on the observation level. The EPN coordinate results fit better into global solutions if EPN and global stations are combined.
Proposals and plans for the future
The following key aspects have been discussed during the round table discussion and future steps were agreed upon:
EPN LAC Analysis in General
As soon as IGS provides the IGS08 SINEX file the corresponding station coordinates will be compared to those of the ITRF2008. The impact of using IGS08 and ITRF2008 on EPN analysis results will be investigated. Based on this comparison it will be decided, which reference coordinates will be applied for EPN computations.
As soon as the IGS08 ANTEX file will be released by IGS the impact of using this antenna information in the analysis on resulting EPN products will be evaluated.
It is expected that IGS will change from ITRF2005/IGS05 to ITRF2008/IGS08 in April 2011. At the same time the EPN will change too. Information about IGS combination results is available at ftp://igs-rf.ign.fr/pub.
RINEX format version 3:
RINEX version 3 observation files are needed for handling Galileo observations and will also support the new GPS satellite signals. EPN station operators are asked to submit observation files in the new format to the BKG data center in addition to the RINEX v2 files.
Today converter software for creating RINEX files evokes many discrepancies in the resulting RINEX header and observation data blocks by not applying the same standards. It is expected that the "teqc" converter will not be updated for RINEX v3 and thus other programs need to be checked carefully.
Summary for EPN rapid daily combination:
Rapid daily solutions are optionally computed and submitted by several LACs. The AC computes a combined rapid daily solution from available submissions. A summary of the EPN rapid daily combination shall be provided as feedback for the contributing LACs. Daily summary information will be archived at the BKG data center and will be available for download. In case of problems in using a particular LAC contribution this LAC will be informed. The recent summary file will be posted to the BKG data center website.
Towards GNSS (GPS, GLONASS, Galileo):
Taking into account that 57% of the EPN stations track GLONASS in addition to GPS, LACs are urgently asked to use GLONASS data for routine EPN analysis. LACs are also asked to update the LAC description form at the EPN CB to correctly indicate which satellite system's observation they process. This should also be indicated in the SINEX header.
Improved analysis models have been presented on various symposia recently, which are currently not yet applied in the EPN analysis by the majority of the LACs. This is e.g. the case for using new troposphere mapping functions, such as VMF1, and considering higher order ionosphere terms. Obviously LACs couldn't apply the new options as long as the used analysis software doesn't support it. It is accepted that LACs using the Bernese GNSS software may use the recent available version, i.e. version 5.0, and the supported models. The above mentioned new models may not be applied in this case.
Deadline for submission of re-processed solutions by LACs is March 31, 2011. The AC (coordinates and troposphere) will frequently update the combination and thus give feedback to the LACs. The focus of the first re-processing is on the period of the years 1996 to 2006. LACs will submit daily and weekly solutions. It is recommended to start the re-processing in reversed order (2005, 2004, ... , 1996)
The combination (EPN-Repro1) should be ready before the EUREF 2011 symposium.
Bernese GNSS Software users may use version 5.0 or newer versions. GAMIT 10.35 or 10.40 and GIPSY/OASIS 5.0 and the new version 6.0 are also used for the re-processing.
Products to be used:
Any IGS repro1 products (e.g. satellite orbits and clocks) will be accepted.
LACs shall use EPN_05.atx for receiver and satellite PCV in the first re-processing.
30 EPN sites will be analyzed by all LACs (7 days) in order to verify model consistency between all LACs. A list of stations and the dedicated observations epochs will be selected by the AC and the re-processing WG chair and send to the LACs in December 2010.
Second EPN re-processing:
The necessity of a second re-processing (EPN-Repro2) has been recognized and is intended to be performed in 2011. The selection of a consistent set of reference station coordinates, satellite and receiver antenna PCVs and satellite orbits will be more stringent as for the EPN-Repro1 analysis. Mainly the availability of these products at the IGS level determine the beginning the second re-processing.
EPN-Repro1 may be computed without GLONASS observations due to the lack of suitable re-processed products for GLONASS orbits.