| Date: | 1996-01-11 15:00:00 | |
| Sender: | Andrew Sinclair, RGO <ats@ast.cam.ac.uk> | |
| Subject: | [SLR-Mail] No. 8: Survey of SLR data pre-processing | |
| Author: | Andrew Sinclair, RGO | |
| Content: | ******************************************************************************** SLR Electronic Mail 1996-01-11 15:00:00 UTC Message No. 8 ******************************************************************************** Author: Andrew Sinclair, RGO Subject: Survey of SLR data pre-processing SURVEY OF SLR DATA PRE-PROCESSING BY SINGLE-PHOTON STATIONS At the Eurolas meeting held in Munich in 1995 March the effects of satellite signature were reviewed, in particular as seen by stations which use detectors that detect the first returning photon. The conclusions are detailed in the report of the meeting, which has been widely distributed, and further copies are available from ats@ast.cam.ac.uk. It was established that there are several possible reference points that could be adopted from the distribution of data obtained from a satellite, and due to the effects of satellite signature these reference points may need different corrections in order to relate them to the centre of mass of the satellite. At the Laser Ranging Workshop in Herstmonceux in 1984 it was recommended that the reference point should be the mean of the data, using an iterated rejection level of 3.0 times RMS from the mean. However many stations have decided that this rejection level retains too many outliers and is too strongly affected by the skew tail of the data, and so rejection levels of 2.5 or even 2.0 times RMS have been adopted, and other stations (e.g. RGO) use their own variants on this recommendation. This situation was discussed at the CSTG SLR Subcommission meeting in Bern in 1995 December, and it was considered that it might be desirable for all stations to adopt the same processing method, but first it would be sensible to carry out a survey of what is being done at present, and to ask stations to make a trial of using some alternative reference points and compare them with their standard reference point. In order to facilitate this survey a standard subroutine DISTRIB has been written, and is available from the Eurolas Data Centre (FTP 129.187.165.3, cd pub/laser/software). The input to the subroutine is an array RES of the residuals of the satellite ranges from the smoothing function, or just the set of calibration ranges from a terrestrial ranging session. The units should be two-way light time in picosec. (The process of fitting the smoothing function will probably use a tight rejection limit, but having obtained the smooting function a wider band of residuals should be selected for input to DISTRIB, so that the full distribution can be examined.) The subroutine returns various quantities, of which the important ones for this survey are: DMEAN mean of the data using a 2.5xRMS iterated rejection, SKEW skewness of the data retained within 2.5xRMS, RKURT kurtosis of the data retained within 2.5xRMS, PEAK the location of the peak of the data distribution, LEHM the location of the leading edge half maximum of the data distribution. The quantities DMEAN, PEAK and LEHM should be converted to one-way millimetres by multiplying by 0.149896, and the differences DMEAN -LEHM and DMEAN -PEAK should be formed. Also whatever reference point the station is using at present should be continued, and the difference DMEAN -`station` should be formed. Stations should form and save these three differences and also the skewness and kurtosis, for all the passes that they track and for all their terrestrial calibration sessions, together with the name of the satellite or target and the date. I will later send out a second message giving details of arrangments for collecting summaries of this information. The aim will be to make an assessment of the data in time for the CSTG meeting to be held in Porto in June 96. The survey is aimed primarily at stations operating at single photon detection levels, but it would not do any harm if stations operating at multi-photon return levels using MCP detectors were also to collect this information. This collection of data should also help to decide what quantities are useful for inclusion in a revision of the normal point data format. --------------------------------------------------------- The table below summarises RGO data for the period 1995 May 4 to December 15, and gives an indication of the sort of results that may be expected. Average values of the data quantities have been formed. The values in brackets are the RMS scatter of the individual values about the average. Some comments on the RGO data : 1. The quantity 1*RMS is the mean using a very tight clipping of the data, and is an alternative method of deriving the peak. The RGO results show that it is fairly similar to PEAK, but has rather more scatter, and so PEAK is preferred, and it is not recommended that stations should collect the data for 1*RMS. 2. For these tests RGO was forming the mean using a 3.0xRMS rejection. RGO has now changed to using a 2.5xRMS rejection, using the standard subroutine DISTRIB. 3. For these tests RGO did not save the data for LEHM, and so no results are available. RGO has now started to save the LEHM data. 4. The quantity `RGO` is the usual reference point adopted by RGO, which is seen to be fairly close to the PEAK. 5. It is important that the data for terrestrial calibration are collected also. The RGO results show a difference between mean and peak for T/R of about 2 mm, and so the similar differences of about 2 mm for ERS-1, ERS-1, GFZ-1, GPS35 and GPS36 probably arise largely from the ranging system itself, and do not give clear evidence of additional contributions from the satellite. 6. Use of a 2.5xRMS clipping of the data as is used in DISTRIB, instead of 3.0xRMS as used in the RGO tests, will probably reduce the kurtosis values to less than 3.0. --------------------------------------------------------- Satellite No.of 3*RMS-1*RMS 3*RMS-PEAK 3*RMS-`RGO` Skewness Kurtosis or TR. passes T/R 5521 1.96 (1.38) 1.73 (0.98) 1.77 (1.05) 0.50 (0.22) 3.27(0.31) ERS-1 229 2.33 (1.97) 2.04 (1.15) 2.14 (1.51) 0.49 (0.26) 3.29(0.39) ERS-2 238 1.80 (1.91) 1.63 (1.25) 1.70 (1.37) 0.40 (0.26) 3.21(0.40) GFZ-1 197 1.86 (2.16) 1.54 (1.43) 1.79 (1.66) 0.37 (0.33) 3.17(0.50) GPS-35 25 2.25 (2.62) 2.21 (2.45) 2.30 (2.48) 0.49 (0.36) 3.43(0.57) GPS-36 8 1.05 (1.72) 1.66 (1.20) 1.29 (0.94) 0.45 (0.22) 3.19(0.41) Meteor-3 220 3.05 (1.89) 2.59 (1.58) 2.84 (1.71) 0.55 (0.27) 3.30(0.38) Starlette 254 3.72 (2.05) 2.96 (1.49) 3.57 (1.91) 0.56 (0.23) 3.22(0.36) Stella 238 3.38 (2.06) 2.72 (1.45) 3.38 (1.75) 0.52 (0.24) 3.21(0.40) GL-63 24 5.89(10.06) 3.48 (2.46) 9.58(19.71) 0.32 (0.11) 2.77(0.27) GL-67 30 6.06 (6.48) 3.52 (2.63) 5.92(15.92) 0.35 (0.17) 2.88(0.28) Lageos-1 493 7.28 (1.55) 6.43 (1.92) 8.11 (1.97) 0.75 (0.15) 3.48(0.36) Lageos-2 274 7.37 (1.44) 5.93 (1.56) 7.82 (1.98) 0.74 (0.13) 3.37(0.26) Etalon-1 30 20.61(10.25) 14.24 (7.53) 36.24(15.14) 0.58 (0.35) 3.19(0.45) Etalon-2 24 22.93 (2.84) 14.51 (3.80) 35.94 (6.53) 0.69 (0.11) 3.08(0.45) Ajisai 258 25.18 (4.78) 21.94 (7.93) 35.90(10.60) 0.89 (0.18) 3.44(0.36) Topex 381 29.86 (9.68) 32.70(12.31) 44.12(15.42) 1.02 (0.16) 3.52(0.34) Andrew Sinclair ats@ast.cam.ac.uk From: ats@ast.cam.ac.uk ******************************************************************************** |