****************************************************************************** DORIS Electronic Mail Tue May 26 11:31:23 WETDST 1998 Message Number 0042 ****************************************************************************** Author: Chopo Ma (cma@virgo.gsfc.nasa.gov) Subject: Study of EOP from defined TRF(/CRF) frames To: DORIS, GPS, SLR and VLBI Analysis Centers There was considerable discussion of the separate evolution of EOP and the TRF at the various IERS workshops and meetings in Paris in October 1996. In order to bring about consistency for the immediate future, changes were made in the EOP series at the beginning of 1997. An analysis study was proposed to provide the IERS with some more information to improve consistency in the future. I am writing this memo as the volunteer study coordinator to help define the study and to encourage participation by interested analysis centers, which is the prerequisite for any activity. The motivation for this study is to understand the causes for the divergence over time between the pole/UT1 defined by the ITRF and the pole/UT1 defined by the continuity of historical EOP series. Using several types of space geodetic observations, various analysis software systems, and a number of methods for attaching the results to the TRF or the underlying Earth, multiple analysis centers estimate EOP on a continuing basis. The IERS makes use of the available EOP results to produce both rapid service and long term EOP series. In equally diverse ways analysis centers also generate TRF results, which are used by the IERS to make the ITRF. Since the EOP series and the ITRF are each created separately, it is perhaps not surprising that inconsistencies arise, but the magnitude of the divergence that has been seen is undesirable. The analysis used to generate EOP and TRF results at each analysis center involves a number of steps and assumptions. In order to try to identify the factors that lead to divergence, this study will simplify the process. The method of the study is to generate EOP series using ITRF positions and velocities (and ICRF95 source positions for VLBI) as fixed, a priori values. The resulting EOP from each analysis will thus be forced into consistency with the same ITRF. It is possible that systematic differences will be seen between techniques or within techniques caused by software or data distribution even thought the identical ITRF is used. The EOP section of the IERS Central Bureau (and any other interested parties) will examine the trends between the EOP series from different techniques and centers and between these series and the current IERS series to suggest how the EOP and ITRF can be kept consistent in the future. The proper execution of the study involves several steps: 1) understanding the scope and purpose of the study, 2) convenient access to the ITRF (and ICRF95), 3) identification of software limitations, 4) agreement on the use of ITRF/ICRF information, 5) initial trials, and 6) a delivery date for EOP. 1) This consistency study is meant to be a cooperative research effort to provide the greatest commonality within the limitations of each analysis. The EOP results will not represent the "best" analysis of any technique or center since the data and analysis inconsistencies inherent in the ITRF used will be incorporated. The goal is to understand how the space geodetic data and analysis respond to the frame definition. Since the study is aimed at improving the relationship of the ITRF and EOP in the future, it is proposed to use the new ITRF96. This realization incorporates the results from the latest IERS annual submissions from the spring of 1997. There are several reasons for suggesting the use of a new realization. First, the latest IERS submissions extend the time interval for determining velocities, important for more recent techniques such as GPS and DORIS. Second, this realization used the full variance/covariance matrices provided. This is made possible by the capabilities of the GEOMIX software. Third, but somewhat problematical, the new realization attempted to generate a single velocity for all monuments at a site with collocated techniques, unlike previous ITRF versions. 2) The ITRF section of the IERS CB will provide the ITRF96 information to the participants in a convenient form. There will be positions and velocities for all the sites with data to be included in the analysis. Values are defined for sites without real velocity estimates. A simple ASCII file of positions and velocities may be sufficient. A common format such as SINEX with zero sigmas might be usable, but tailored formats for individual analysis systems may be very helpful. In addition ICRF95 source positions will be used for the VLBI analyses. Since the goal is to study the effect of ITRF definition on EOP biases and rates, the uncertainties of the ITRF positions and velocities (and of the ICRF95 positions) are not important. The geometry of the ITRF on a given day, defined by perfect positions and velocities, permits the real data to determine EOP. Examining the EOP biases and rates from individual solutions does not depend critically on having meaningful EOP sigmas. The VLBI analyses will also produce a time series of daily nutation offsets with similarly idealized sigmas. 3) To achieve commonality of analysis and to have a reasonable level of participation we will need to find out what is possible for each volunteering center. The minimum common configuration would fix a priori positions and velocities for all sites. VLBI radio source positions would be fixed at the ICRF95 values for defining sources. I assume that no system lacks these capabilities. The resulting EOP/nutation series would reflect perfect frames with only data and estimation errors. Another type of configuration would fix some sites and adjust the positions and velocities of the others. Again I would assume that this is generally possible. Some data from certain techniques may need to be excluded, related to discrepancies at particular sites. It will be necessary to know if such data editing is feasible at the relevant analysis centers. 4) Once the available options for the various analysis systems are known, it should be possible to agree on the level of information to use. I would hope that at least three configurations can be done by several techniques and analysis centers. The first configuration would fix all site positions and velocities. A second would fix the complete set of collocation sites. For these two configurations all techniques would use the same velocities at collocated sites. The TRF section of the IERS CB will provide the list of collocated sites. For the third, each analysis center would fix only those sites that would normally be used to tie its solution to the ITRF. A last solution might use the same data set but with site positions and velocities adjusted if that is the usual analysis mode. In all other aspects the solutions should use the analysis center's normal configuration for EOP analysis. 5) It will undoubtedly be necessary to have some trial runs since the study will not use the usual modes of analysis. These trial EOP series will be examined by the EOP section of the IERS CB to identify any problems or anomalies so they can be resolved (if possible) before the final runs. I propose that the year of 1997 be used as a trial with EOP values generated at least weekly. How many years of data to use for the final solutions is another question. The span should probably go as far as possible, i.e., from the inception of the technique through the data used for the ITRF96 realization. Data later than the ITRF96 realization should probably not be used to avoid the accidental introduction of new sites. Since it is the long term trends that are of interest, it is not necessary to have results every day. Every five to seven days should be sufficient. 6) A date for delivery of the study EOP series needs to be set, at least tentatively. I would suggest disconnecting the annual submissions and the study solutions, but some analysis centers might find it more convenient to do some of them at the same time. I would ask people to respond to several points: 1) Is the study definition reasonable? 2) Is it feasible for you to set up study solutions fixing all/some stations? 3) What is the most convenient input format for ITRF positions and velocities? 4) What is the most reasonable schedule to do the trial and final study solutions? Chopo Ma [Mailed From: Pascal Willis ]