Release notes for GRACE L1B-data version 00 and version 01 post date: 2006-05-02 by Gerhard L.H. Kruizinga >From 2005-01-01 onwards release 01 of the GRACE Level-1B data will only be available from PO.DAAC and GFZ's ISDC. Essentially release 00 and release 01 are identical except for the following: 1) Improved raw telemetry processing resulting in a larger recovery rate of science and housekeeping data 2) Improved lighttime corrections in the KBR1B product. Especially the lighttime range acceleration in release 00 was found to have periodic spikes caused by an inferior interpolation algorithm. 3) The meaning of two quality flags in the KBR1B product has changed Old meaning:(in release 00 bit1 and bit2 were never set to 1) bit 1 = extrapolated states for lighttime_corr bit 2 = interpolated SCA attitudes for ant_centr_corr New meaning: bit 1 = unreliable PCI data for ant_centr_corr bit 2 = interpolated PCI data for ant_centr_corr -------------------------------------------------------------------------------- post date: 2004-08-19 by Gerhard L.H. Kruizinga The following periods of data are now available: February 21-28, 2003 May 1-31, 2003 August 18-20, 2003 September 14, 2003 -------------------------------------------------------------------------------- post date: 2004-08-06 by Gerhard L.H. Kruizinga All dates between July 31, 2002 to December 1, 2002 and January 31, 2003 to June 1, 2004 are available, with the exception of the following periods which are excluded: February 21-28, 2003 May 1-31, 2003 August 18-20, 2003 September 14, 2003 The aforementioned dates will be released in the upcoming weeks or as soon as possible. -------------------------------------------------------------------------------- post date: 2004-08-06 by Gerhard L.H. Kruizinga The following alignment and vector offset products are distributed in the tar file for 2004-05-31: QKS1B_2004-05-31_A_00.dat QKS1B_2004-05-31_B_00.dat QSA1B_2004-05-31_A_00.dat QSA1B_2004-05-31_B_00.dat VGB1B_2004-05-31_A_00.dat VGB1B_2004-05-31_B_00.dat VGN1B_2004-05-31_A_00.dat VGN1B_2004-05-31_B_00.dat VGO1B_2004-05-31_A_00.dat VGO1B_2004-05-31_B_00.dat VKB1B_2004-05-31_A_00.dat VKB1B_2004-05-31_B_00.dat VSL1B_2004-05-31_A_00.dat VSL1B_2004-05-31_B_00.dat -------------------------------------------------------------------------------- post date: 2004-08-04 by Gerhard L.H. Kruizinga Ulrich Fackler and Bjorn Frommknecht of the Technical University of Munich discovered periodic patterns in the fill flag of the ACC1B data. The Fill flag is set when a data gap is detected in the raw ACC1A data. Gaps < 100.0 seconds are filled by interpolation, the interpolation is required for the CRN filter used to create the ACC1B data. For more details see Algorithm Theoretical Basis Document for GRACE level-1 process by Wu et al. It was found that a periodic time tag error in the ACC1A data, caused about 300 times a day, a data gap of about 2 seconds. This resulted in the fill flag being set in a periodic fashion in the ACC1B data. After fixing the ACC1A time tag error, no more data gaps were found in the ACC1A data and hence the periodic setting of the ACC1B fill flag. At the time of this writing it is not clear how much the removal of 2 seconds of raw data at about 300 times a day, affects the gravity field solution. Some large differences are found with release 00 when a thrust is eliminated in the raw data and which is then not accounted for in the ACC1B data. Analysis is ongoing to quantify the effect of this error. All of release 00 ACC1B data have this error. This problem will be fixed in release 01 -------------------------------------------------------------------------------- post date: 2004-08-04 by Gerhard L.H. Kruizinga Willy Bertiger from JPL discovered that in the TDP1B file the value of the clock solution had the wrong sign compared to the definition of the estimated clock solution in GIPSY-OASIS. This problem can be fixed by changing the sign of the value of the clock solution when reading the file. All TDP1B files for release 00 have this problem. This problem will be fixed for release 01. -------------------------------------------------------------------------------- post date: 2004-03-03 by Gerhard L.H. Kruizinga Shin Chan Han of Ohio State University discovered an error in the GNV1B product related to the reported formal errors of the position and velocity in GNV1B. Shin Chan noticed that the formal errors for z-components of position and velocity 30 minutes after starting midnight and 30 minutes prior to end midnight were inconsistent with formal error reported for the remainder of the day. These periods are the exact time windows for which cosine smoothing is applied to ensure continuous orbits at midnight. Due to a bug in the L1B software the formal errors were not correctly computed in these intervals for the z-components of position and velocity. The z-components of the position and velocity were not affected by this software bug. Caution is warranted when using the formal errors of the z-component of position and velocity in the GNV1B product This problem will be fixed in release 01 -------------------------------------------------------------------------------- post date: 2004-01-26 by Gerhard L.H. Kruizinga The header record count is in error in the following data products: AHK1B IHK1B MAG1B MAS1B THR1B TNK1B The actual header record is under counting the number of header records by two in the above products. Example: NUMBER OF HEADER RECORDS : 23 and should be NUMBER OF HEADER RECORDS : 25 In general when reading the header one should continue to read until the header record "END OF HEADER" is reached. After this record the data will start. This problem will be fixed in release 01 --------------------------------------------------------------------------------