Date: Wed, 13 Nov 2002 11:57:16 -0800 From: Pascal Willis To: Jean-Paul Berthias Cc: Bruce J Haines , "John C. Ries" , Yoaz Bar-Sever , nzelensk@geodesy.gsfc.nasa.gov, sluthcke@xyz.gsfc.nasa.gov, krchoi@csr.utexas.edu, Martine Feissel Subject: test of new CNES preprocessing for DORIS/JASON data Dear Jean-Paul, following our discussion in New Orleans I have tested the new JASON/DORIS file that you kindly provided on AVISO for cycle 20 (ja1_doris_new_data_cy020.willis.Z) here are my conclusions (detailed results in annex) : - the new preprocessing provides slightly better DORIS residuals and results - there is a clear offset of - 9mm in the orbit radial component that is created in the DORIS orbits between the 2 formats, the new orbit being closer to JPL/GPS reduced dynamic solution - the new solution for point positioning is slightly closer to ITRF2000 and the scale agrees better with ITRF2000 So, in my opinion, the new preprocessing that you did gives better results and should be adopted (format 2.2?). It explains 1 cm offset (out of the 3 cm offset that I saw before) between my DORIS orbit and the JPL/GPS reduced dynamic orbits. So I still have some work to do there. Do you want me to do additional tests? Best regards Pascal NB: I would like to do some phase center offset determination with DORIS, but there is no data after cycle 20. Would it be possible to get some more cycle before the AGU in order to do a complete study on that? ------------------------------------------------------------------ 1) POD DORIS residuals / fixed network solution (ITRF2000 + J. Ries complement) day CDDIS AVISO in mm/s 204 0.378 0.376 205 0.380 0.378 206 0.390 0.389 207 0.380 0.378 208 0.370 0.368 209 0.378 0.375 210 0.383 0.381 211 0.381 0.379 NB: very small improvement <0.5% 2) DORIS orbit comparison (AVISO data vs CDDIS data) RMS over 30 hour (every minute) day H C L (in mm) 204 -8.4 +/- 0.7 1.5 +/- 4.2 2.0 +/- 2.4 205 -8.5 +/- 0.9 0.8 +/- 4.9 0.6 +/- 2.5 206 -8.7 +/- 0.6 0.3 +/- 4.4 1.4 +/- 1.5 207 -9.1 +/- 0.5 -0.6 +/- 5.8 1.3 +/- 1.3 208 -8.9 +/- 0.7 1.1 +/- 4.3 1.4 +/- 1.7 209 -8.7 +/- 0.8 1.7 +/- 5.9 1.3 +/- 2.2 210 -8.7 +/- 0.7 0.3 +/- 5.0 1.4 +/- 1.9 211 -9.0 +/- 0.7 -0.1 +/- 4.1 1.2 +/- 1.9 Systematic bias in radial of -9 mm NB: When I estimate phase center offset for DORIS, amazingly I get the same orbits in the 2 cases, RMS are 1 mm in H, 2 mm in C and 2 mm in L, meaning that this 9 mm offset explain a large part of the problem 3) DORIS orbit vs JPL reduced dynamic orbit (in radial component) day CDDIS AVISO in mm 204 -34.5 +/- 24.1 -26.1 +/- 24.0 205 -37.2 +/- 13.9 -28.7 +/- 13.9 206 -31.7 +/- 17.6 -23.0 +/- 17.4 207 -35.0 +/- 12.8 -25.9 +/- 12.7 208 -36.1 +/- 15.6 -27.1 +/- 15.6 209 -40.7 +/- 17.8 -31.9 +/- 17.5 210 -34.1 +/- 15.3 -25.3 +/- 15.1 211 -40.0 +/- 14.9 -31.0 +/- 15.1 better agreement with JPL/GPS reduced dynamic orbit NB: I have used the full 30-hour arc for GPS-reduced dynamic (which may not be a good idea) 4) Point positioning DORIS residuals / free network solution day CDDIS AVISO 204 0.364 0.363 205 0.362 0.362 206 0.358 0.358 207 0.362 0.361 208 0.348 0.347 209 0.354 0.353 210 0.363 0.363 211 0.359 0.358 marginal improvement in residuals 5) comparison with ITRF 2000 + Ries complements using all stations but 3 (SODB, MORB and GALA) CDDIS AVISO scale factor (10-9) -14.0 +/- 5.5 -11.0 +/- 5.5 N residuals (mm) 40.6 39.5 E residuals (mm) 52.2 51.7 V residuals (mm) 44.6 42.4 small improvement in scale factor as well as in point positioning residuals