Third ILRS AWG Meeting (Matera 2000)
Minutes of ILRS Analysis Working Group Workshop
Matera
9-10 November 2000
(written by C.Luceri, R.Noomen, R.Devoti)
List of Participants:
- Graham Appleby, NERC, [email protected]
- Richard Biancale, CNES, [email protected]
- Roberto Devoti, Telespazio(ASI/CGS), [email protected]
- Maurice Dube, Raytheon ITSS, [email protected]
- Ramesh Govind, AUSLIG, [email protected]
- Van Husson, HTSI, [email protected]
- Rainer Kelm, DGFI, [email protected]
- Cinzia Luceri, Telespazio(ASI/CGS), [email protected]
- Maria Mareyen, BKG, [email protected]
- Phil Moore, University of Newcastle, [email protected]
- Ron Noomen, DEOS, [email protected]
- Toshimichi Otsubo, CRL/NERC, [email protected]
- Matti Paunonen, Finnish Geodetic Institute, [email protected]
- Erricos Pavlis, JCET, [email protected]
- Mike Pearlman (Friday), CfA, [email protected]
- Paolo Rutigliano, Telespazio(ASI/CGS), [email protected]
- Pete Shelus (Friday), McDonald Observatory, [email protected]
- Mark Torrence, Raytheon ITSS, [email protected]
THURDAY, November 9, 2000
(note: the paragraph numbering of the minutes follows the agenda items numbering in appendix 1)
1. Opening
Brief introduction of the participants.
Noomen reports on the interest of the outside community and underlines the importance of the workshop. The agenda is accepted (appendix 1).
2. Role of ILRS within IERS
Noomen gives an introduction on the ILRS products (coordinates, velocities, EOPs, geocenter, etc.) and the ILRS clients (IERS, the geophysical community and ourselves) (all material presented is included in appendix 2).Noomen illustrates the structure of the new IERS organization and underlines the official status of the ILRS participation within IERS. The new IERS organization is very well structured, with specific Product Centers and Technique Centers (of which ILRS is one). ILRS will be responsible for the provision, in the future, of one official product to the respective IERS Product Centers; alternative solutions can be submitted to the Combination Research Center. This can only be effective after the ILRS has shown that its official product is of the best possible quality. The new organization will probably be operational by the summer of next year. Mareyen expresses her doubts about the connection between the users and the IERS Directing Board. Kelm remarks that IGN has been accepted as Product Center for the Terrestrial Reference; IGN and DGFI have been assigned to Analysis Centers for Terrestrial Reference. A call for participation to be a further Analysis Center will be issued for others.
2.1. ITRF2000
Noomen summarizes the ITRF2000 workshop held in Paris (2-3 November, 2000) underlining the fundamental role of SLR for the definition of origin (all SLR solutions) and scale (all SLR and VLBI, each technique contributing for 50%) for ITRF2000 and the large number of contributions. He illustrates some results of a preliminary ITRF2000 solution.
2.2. Operational products
The reasons for developing official products are discussed. They are:
- - ILRS term of reference
- - Quality control (internal, external)
- - Combination should be better (lower uncertainty, more reliable) than individual solutions (no degradation by poor solutions)
- - Stimulus for improvements (quality, competition) and for operations (deadlines)
- - Request of clients: Bulletin A (preferably now), other IERS products (in the future)
- - Other services are providing (IGS) or about to provide (IVS, within a few days) official products
- - Symbol of maturity
3. Action since AWG Delft
Noomen gives a summary of the main activities of the Analysis Working Group from the very beginning (September 1998) until now (appendix 3).
4. Minutes AWG Delft
An overview of the action items from the minutes of the Delft meeting (appendix 4). All have been satisfied, except for 2:
- 1) Husson will get in contact with Pavlis for the eccentricity files.
- 2) Husson will check if the description of the pilot projects is on the web pages. They are put on the list of action items of the Matera AWG meeting.
5. Pilot Project "Positioning and Earth Orientation"
5.1. Presentations of the contributions (station coordinates, earth orientation) to the pilot project (L1, L2 and L1&L2 solutions) are given:
Luceri (ASI) gives a general description of the analysis procedure and a comparison of the results with ITRF97 (appendix 5).
Govind (AUSLIG) presents his strategy, which includes estimation of range and time biases on a pass-by-pass basis. The results for the rototranslation parameters and the coordinate residuals are identical to ASI's (appendix 6).
Otsubo (CRL) presents his analysis. Here, a non-standard GM value was adopted. Differences in scale factors of the submitted solutions are highlighted (appendix 7).
Kelm (DGFI) presents his analysis. Here, a range bias is estimated or modeled on a station-by-station basis. The internal consistency between the L1 and L2 solutions is shown (appendix 8).
Biancale (GRGS) presents the solution: the range biases are adjusted, the GRIM station coordinate standard errors are used to weight the corresponding SLR observations. No a priori constraints are adopted, even if the a priori matrix exists and shows significant constraints: the SINEX file is probably wrong and needs to be checked (appendix 9).
Pavlis (JCET) weighs the SLR observations equally (apart from a few exceptions); the range biases are adjusted (appendix 10).
Appleby (NERC) presents his solution. A non-standard GM value was adopted (appendix 11).
Due to time constraints (too late) the meeting continues with the SINEX format issues while a list of questions on the future of POS+EOP is distributed to the audience to be discussed tomorrow (appendix 12).
5.3. SINE format
The ILRS interpretation of the SINEX format is discussed (appendix 13).
FILE/COMMENT: This section must include information on the analysis. As a starting point, technical description file for submission to IERS/ITRF can be used but, to fulfil the function properly, more information must be added (appendix 14).
Information to be included in the FILE/COMMENT of the SINEX: the one which is provided in the IERS technical description, plus:
- - gravity field (static and time dependent) order & degree
- - ocean tides
- - solar radiation, thermal forces, albedo (visible) + earthshine (infrared) nutation, precession
- - a priori EOP
- - a priori station (position, velocity)
- - ocean loading
- - pressure loading
- - geocenter (motion, tidal frequencies)
- - range biases
- - arc length
- - estimated parameters
- - empirical accelerations
- - frequency of EOP solutions
- - reference point for coordinate solutions
- - wavelength of observations
- - satellite used dynamic polar motion
- - data weighting
- - GM
- - relativistic effects
- - third body attraction
- - atmospheric and earth tides
- - tropospheric bias
- - time bias
- - minimum elevation
- - site velocity
- - constraints
- - pole tide
- - center of mass
- - model for tropospheric delay
In order to properly represent biases, the participants agree to also put the solved-for bias parameters in the SOLUTION parts of the SINEX file. The range biases will be expressed in meters, time biases in milliseconds, scale biases in part-per-billion and tropospheric biases in meters. As far as the sign of the biases is concerned, the following conventions apply: +1 m range bias means that the observation is 1 meter long, time bias +1 sec. means that the observation is 1 sec. late. The biases will be introduced in the SOLUTION/ESTIMATE, SOLUTION/APRIORI and in a new BIAS/EPOCHS block. In the SOLUTION/ESTIMATE, and SOLUTION/APRIORI blocks, the "parameter type" (A6) will indicate the type of bias (RBIAS for range bias, TBIAS for time bias, SBIAS for scale bias, ZBIAS for tropospheric bias, respectively), the "site code" (A4) will provide the ID of the station, the "point code" (A2) will provide the satellite (for now: L1 and L2 for LAGEOS-1 and LAGEOS-2, respectively), the "solution id" (A4) will be used as the indicator of the bias for this specific station (i.e. will be 1 if only 1 bias is estimated) and "epoch" will refer to the beginning date of the relevant time interval. To properly specify the actual time interval of the bias solution, the new block BIAS/EPOCHS (similar to SOLUTION/EPOCHS) will be used (appendix 14). Fields are to be interpreted as described above; the "observation code" (A1) is to be used to denote the type of bias (i.e. R, T, S or Z).
Regarding the SITE/ECCENTRICITY block, the analysis centers must pay attention to the values specified here; eccentricities must be subtracted if different from zero.
Scaling of solution
After a lengthy and detailed discussion, it was decided to allow solutions in the SINEX files to be scaled by an internal variance factor (all SOLUTION blocks consistently). This must be done such that the standard deviation represent the actual uncertainty (i.e. for the pilot project: 1 meter for coordinates). The variance factor in the SOLUTION/STATISTICS block is for information only.
FRIDAY, November 10, 2000
5.2. Presentation of comparison and combinations:
Pavlis (JCET) shows his comparison/combination results: the estimation of rototranslation parameters, coordinate residuals of the solutions among the L1&L2 solutions (appendix 10). Geodesics show a scatter of about 10 mm, whereas variations in scale and origin of 1-2 ppb and 10 mm, respectively are found.
Kelm (DGFI) presents his comparison/combination results, emphasizing that loose constraints must indeed be loose (appendix 15).
Luceri & Devoti (ASI) give a comparison (Helmert parameters based on coordinates only and on a combination of coordinates and EOPs) and DOP analysis of EOPs. They recommend that no parameter must be deleted from the normal equations because this is a constraint not recoverable from the SINEX file. Also they notice the poor quality of daily UT estimates using only LAGEOS-1 observations (appendix 16).
Husson (HTSI) gives extensive comparison results, addressing various aspects of the dataset and the solutions. The network comparison is done in terms of baselines, which appears to be a reliable tool for assessing quality and scale. One comment from the audience is that the variance/covariance should be considered (appendix 17).
Torrence (GSFC) provides a comparison in terms of baselines and uses a NASA/GSFC solution as reference (appendix 18).
Moore (University of Newcastle) presents a combination of solutions and a check on the agreement among the geocenter of the individual solutions, of the combined solution and of ITRF97 (appendix 19).
5.5. Future of POS+EOP:
To illustrate the variety of problems that are related to comparisons and combinations, Noomen gives an overview of the experiences and suggestions by others, which is the result of a questionnaire sent to various combination centers (AIUB, MIT, IGN, USNO, NRCan) (appendix 20). Noomen shows a list of questions on the future the "POS+EOP" project that must be answered by the audience (appendix 21) on products, combination and comparison.
Products:
Product(s): official status? Not yet, must await for a good quality first
-Product(s): what? TRF: time series? Pos+Vel? EOP: time series
One of the objectives is to assess (and improve) the consistency of the solutions. With this aim we should limit our products to time series of EOP and station coordinates as done until now; velocities will be considered at a later stage. The estimation of geocenter, as suggested by Biancale, will be considered in the future.
Satellites?
A combined solution for LAGEOS-1 and LAGEOS-2 is required, optional for LAGEOS-1 and LAGEOS-2 separately.
Total interval?
Noomen suggests one year of data and it is accepted. The time span will go from January 3rd , 1999 to January 1st, 2000 (inclusive). Husson and Dube will prepare the dataset by November 20th .
Frequency for TRF
One set of coordinates for intervals of 4 weeks. This is a problem for GRGS,they need modification to the SW. The AWG recommends to do these modifications.
EOP: length? product? epoch?
Pavlis underlines that daily solution with LAGEOS-1 and LAGEOS-2 are made since many years and that the community already has experience on that argument (cf. the products of the pilot projects until now), that we should generate a LOD product instead of UT and that we need to estimate rates for EOP. In order not to burden analysts with too much work (SW modification), Noomen suggests to continue with the estimation of UT and consider to make a change to LOD at the next stage. Recommendation: daily EOP and UT at the midpoint of the day as before. The equivalent of 1 meter as constraint.
Station: position? velocity?
ITRF97 a priori coordinates and velocities, Nuvel-1 velocity for those stations not included in ITRF97. 1 meter constraints for coordinates.
Timeline?
The analysis contribution must be submitted by February 1st, 2001. The next meeting will be held on March 22nd and 23rd, 2001, the week before the next EGS meeting.
Combinations/Comparisons:
Who? Backups?
To be seen in the future, depending on the progress made during the next months.Official Request For Participation (RFP) to the future POS+EOP? It is considered too premature to release an official RFP now, since the community needs to better understand the quality of data products itself, first. Depending on the outcome of the pilot project during the next months, there may be an official RFP which is to be released by the ILRS GB. As for now, Noomen will prepare an invitation to participate (as was done before), mail it around the AWG participants for comments and then distribute it within the ILRS community only. The official RFP should not be issued until after the next meeting to give the SLR analysis centers the opportunity to increase their experience and will be opened to all the centers. The combination center(s) will be chosen on the basis of the answers to the RFP.
Demands on combination activity and results (internal, per institute internal, w.r.t. mean external, w.r.t C04, Bulletin B etc. external w.r.t. ITRF97, ITRF2000 etc. EOP alone? TRF alone? POS+EOP?)?
The request to participate in the comparison/combination should specify a minimum number of specific products that are to be generated, in order to be able to compare the various (comparison/combination) contributions and actually choose an official combination center and backups. Within these specifications, it is up to the center to make a comparison and/or combination and choose the different parameters (coordinates, baseline etc.) to consider.
6. Pilot Project "harmonization"
Husson gives a brief overview of the status. The differences have been identified, the next step should be to have these reduced.
7. Pilot Project: Benchmarking
Husson will prepare an invitation to this project by the next AWG.
8. Pilot Project: Orbits
Nothing has been made but it remains on the table.
9. Miscellaneous
9.1 Report of the IVS/IGS/ILRS: Appleby comments on this activity.
In February a first meeting was held at Ktzting. Correspondence among the participants to the WG is primarily made by e-mail. The discussion is now on the possible contribution of VLBI to check the bias and it is still going on. At the moment, no additional action is given to ILRS; an invitation to ILRS to track intensively GPS35 and 36 is expected if the IVS community decides to start a tracking campaign.
9.2 Data screening:
How to treat outliers in the observations? The audience agrees that the matter should remain to the analysts, no edited data set will be prepared for the community.
9.3 Data flow:
The labels for the solution for the new POS+EOP must be changed. As an example bkg.990103.sol5.v1.snx, where the date is the beginning of the 4-week data interval (including possible leading zeroes) (and the other elements are as in the previous definition.
- ITRF2000 methodology
Biancale reports on a request by Altamimi. He is asking the AWG to identify a set of reliable (time span, data quality, station motion) SLR stations to be used to define the origin and scale of the new ITRF2000. He would like to do some analysis to see if this is a good way to proceed toward ITRF2000. Biancale, Pavlis, Husson will make a suggestion or this.
9.4Next meeting:
March 22th and 23th, 2001, Nice, France
10. Action items
- Noomen: prepare invitation to participate in "positioning + earth orientation" (time series solutions), distribute draft among AWG/Matera participants and next distribute within the ILRS community.
- Noomen: prepare invitation to participate in "positioning + earth orientation"(comparison/combination), distribute draft among AWG/Matera participants and next distribute within the ILRS community.
- Dube/Husson: prepare dataset for "positioning + earth orientation"
- Husson: prepare plan of action for pilot project "Benchmarking"
- Biancale/Husson/Pavlis: prepare list of SLR stations for definition of origin and scale in ITRF2000
- Husson/Pavlis: prepare files with station eccentricities in NEU and XYZ and put them on the ILRS web page
- Husson: put description of pilot projects on the ILRS web page
- Noomen/Husson: update description of SINEX for ILRS pilot project and put on the ILRS web page. Inform Tom Herring
Appendices
- 1)Agenda
- 2)All material for item 2. Role of ILRS within IERS
- 3)Activities AWG
- 4)Action items from AWG Delft minutes
- 5)POS+EOP solution ASI
- 6)POS+EOP solution AUSLIG
- 7)POS+EOP solution CRL
- 8)POS+EOP solution DGFI
- 9)POS+EOP solution GRGS
- 10)POS+EOP solution and comparison/combination JCET
- 11)POS+EOP solution NERC
- 12)Questions on the future of POS+EOP project
- 13)SINEX format
- 14)Info for FILE/COMMENT and new SINEX labels
- 15)Comparison/combination DGFI
- 16)Comparison/combination ASI
- 17)Comparison/combination HTSI
- 18)Comparison/combination GSFC
- 19)Comparison/combination NCL
- 20)Comparison/combination experiences by others
- 21)Future of POS+EOP