Data-files and results from evaluation.

GOCE HPF Test Data Analysis.

Staff: M.Veiherts and C.C.Tscherning.

Test data set, Summer 2004.


Data resides in the directory /scratch/cct and are obtainable from SRON using sftp.

Software and results.

General job for reformatting.

PVT data

Program to compare old and new version of pvt data
Result of comparison.
Job to run general reformatting job.

Version of stat2pot for the processing of pvt-data.
Job-file and output-file.

Job to compare height-anomalies with EGM96 to deg. 60
Summary of results for data not corrected for tilt.


Program to analyse common mode accelerations.
Result of analysis.


Job to compare ESA input gradients with gradients computed from EGM96 and rotated using the Rotation matrix derived from the quarternions. E-mail to ESA with results of test of input gravity gradients.
GOCE Satellite Attitude Calculated from Quarternions and Statevector by M.Veicherts, report to GOCE HPF, 2004-10-19.

Official test data set, Received 2004-12-14.

Check of data-consistency

Initially noise-free data has to be used to check that the test dat are consistent.

The state-vector data are run through the program stat2pot.f (edited version 2005-02-11 or later), and a file with
time, latitude, longitude, ellipsoidal height, and values of the anomalous potential (T) or the height anomaly (zeta) (TBD) corrected for bias and tilt are produced.

The files is used as input to the program /scratch/cct/egm96.job, where the values are compared to EGM96 values to degree 60.

The differences should be below 1 m in mean and stdv. for height anmalies and 10 m**2/s**2 for T. The histogram should show a normal distribution. If not, check the time-offset for GPS time.

The rotation matrix is produced by Martin's program.

The position file (from stat2pot) is merged with the gravity gradient file, so that we have a file with
time, lat., long., h, 6 gravity gradients.

A GEOCOL job is run, where input is this file and the rotation matrix-file. The difference should for all 6 gravity gradients be very close to zero.

The same tests are now repeated for the data in /disk3/GOCE, i.e. EGG_GGT_2C, SST_POS_2C, EGG_IAG_2C and CCD_2C, for the whole month. A meged file with positions in geodetic coordinates and 6 gravity gradients is prepared. It should be considered to create a smaller file with 5 s or larger sampling in order to reduce file size.

Statistics are made of the differences between the noise-free and the noisy data. PSD and auto-covariances are computed.

Martin prepares a new test report.

Test of computational procedures


An initial calibration is done using e.g. a GRACE field, by determining a bias for each of the 6 components.

The merged file is used in the 3 calibration areas in order to determine calibration factors (bias and tilt, and separately scale factor).

A file with the results on the required format is created.


The same data as used in WP3000 as well as the anomalous potential data (T) are gridded. The T data are gridded separately. The gravity gradients ae used to form a Trr and a Tnn-Tee grid (TBC). The grid should initially have a spacing of 0.72 deg, i.e. 500 points in longitude.

The grid is made in blocks of 10 deg. x 10 deg. using a reference potential to degree 48. A 2 deg. overlap is to be used. The covariance function is the global one scaled to the block variance.

Job-files are produced by /cct/goce/test2/gridgen.f, which must be modified accordint to the new specifications.

The gridded data are used with /cct/dgravsoft/sphgric.f to produce spherical harmonic coefficients to deg. 200., as well as the error-covariances.

The result is verified and send to Mirko, together with the gridded-values. However, Mirko wants the second order derivatives with respect to the longitude. This requires a modification of GEOCOL or a new off-line program.

Last update 2005-02-13 by cct.