[Spice_discussion] how are the long-term predicts made for ITRF93?

Goodson, Troy D (US 392C) troy.d.goodson at jpl.nasa.gov
Wed Jul 13 15:51:34 PDT 2022


Nat,

Thank you. That's the information I was looking for.  I should've thought to get the comment section with "commnt -r earth_200101_990628_predict.bpc"  :/

Troy.

https://naif.jpl.nasa.gov/pub/naif/toolkit_docs/C/ug/commnt.html



> On Jul 13, 2022, at 1:49 PM, Bachman, Nathaniel J (US 392N) <nathaniel.j.bachman at jpl.nasa.gov> wrote:
> 
> 
> Hi Troy,
> 
> Here's the complete comment section of the file
> 
>    earth_200101_990628_predict.bpc 
> 
>>>> 
> 
> Low accuracy, long term predict earth PCK
> =========================================
> 
> Original file name:         earth_200101_990628_predict.bpc
> Creation date:              2020 April 11 18:17:00 PDT
> Created by:                 Nat Bachman (JPL/NAIF)
> 
> Data Sources
> 
> Input file:  extended EOP created from
> 
>   JPL EOP file latest.long
>   (Copied from WWW URL
>   http://epic.jpl.nasa.gov/nav/eop/latest.long)
> 
> Source EOP metadata:
> 
>    $  JPL Earth Orientation Parameter File
>    $  Last Data Point  6-APR-2020
>    $  Predicts to     28-JUN-2020
>    $
>    EOPLBL='EOP. LAST DATUM  6-APR-2020. PREDICTS->28-JUN-2020, UT1TYP=UT1.   '
>    EOPFNG='Enter MAKE_EOP  6-Apr-2020 11:04:43      linked 12-Jan-2017 00:15:37  '
>    EOPUT1='UT1'
>    EOPTYP='EOP'
>    EOPTIM=' 6-Apr-2020 11:04:43     '
>    EOPTRF='ITRF93'
>    EOPCRF='ICRF93'
> 
> Input file: leapseconds kernel naif0012.tls
> 
> 
> Coverage
> 
>   Start time:             2020 JAN 01 00:01:09.183 TDB
>   Stop time:              2099 JUN 28 00:01:09.182 TDB
> 
>   UTC Epoch of last datum:  6-APR-2020
> 
> Particulars
> 
>   This is low-accuracy predict PCK file giving the orientation of the Earth as
>   a function of time for the interval shown above. This PCK file is made from
>   an extended EOP file which consists of the JPL EOP file named above, with
>   four extra records appended. These records are designed to add 79 years to
>   the coverage of the EOP file without affecting interpolated data values for
>   times within the coverage interval of the original EOP file. Data values
>   of records in the extended interval are constant, with the exception of
>   the TAI-UT1 offsets. Those are adjusted so that the corresponding values
>   of TAI-UT1R are constant.
> 
>   The terrestrial frame whose orientation is given by this file is ITRF93; the
>   inertial base frame is Ecliptic of J2000. Rotational effects included are:
> 
>      -Precession (1976 IAU model)
>      -Nutation   (1980 IAU model)
>      -Nutation corrections
>      -Rotation through true sidereal time
>      -Polar motion
>      -Nutation corrections
> 
>   Euler angles have been extracted from the Ecliptic-to-ITRF93 rotation matrix,
>   and Chebyshev polynomials were fit to the Euler angles. The Chebyshev
>   expansion degree was 49; the degree of the retained polynomials is 20.
>   The coverage interval for each set of polynomials is 1 day or less.
> 
>   This file represents the same earth orientation as does the input EOP file
>   for the time range:
> 
>      01-JAN-2020 UTC
>      28-JUN-2020 UTC
> 
>   For epochs after the end of EOP coverage, polar motion and nutation
>   corrections from the EOP file are held constant at values obtained by
>   extrapolating 1 day from the last source EOP record. TAI-UT1 values are
>   set so that the corresponding TAI-UT1R values at each epoch of the
>   extrapolated records are equal. Details of the EOP extrapolation process
>   are given below.
> 
>      Let n be the number of data records in the original EOP file. Let
>      MJD(i) be the epoch of the ith record expressed as a modified
>      Julian date. The added records have the following
>      characteristics:
> 
>         Record n+1:  Epoch is MJD(n) + 1. Nutation correction and polar
>                      motion values are obtained by extrapolating to the
>                      epoch a cubic Lagrange interpolating polynomial
>                      fitted to the corresponding data in records n-3 : n.
>                      TAI-UT1 is obtained by converting to TAI-UT1R the
>                      TAI-UT1 data in records n-3 : n, extrapolating an
>                      interpolating polynomial as for the other data, then
>                      converting the extrapolated TAI-UTR1 value to TAI-UT1.
>                      UTC string is consistent with the MJD epoch.
> 
>         Record n+2:  Epoch is MJD(n) + 2. Nutation correction and polar
>                      motion values match those of record n+1. TAI-UT1 is
>                      set so that the corresponding value of TAI-UT1R at
>                      this epoch is that corresponding to TAI-UT1 at the
>                      epoch of record n+1. UTC string is consistent with the
>                      MJD epoch.
> 
>         Record n+3:  Epoch is MJD(n) + 3. Data values are as described for
>                      record n+2, adjusted for this epoch.
> 
>         Record n+4:  Epoch is MJD(n) + 28854. Data values are as described
>                      for record n+2, adjusted for this epoch.
> 
>      In all of the four records described above, the values of TAI-UT1
>      have been selected so the corresponding values of TAI-UT1R are
>      constant and equal to the value resulting from extrapolating the cubic
>      interpolating polynomial for TAI-UT1R, fitted to the final four values
>      of TAI-UT1R derived from the original EOP file, to the epoch of
>      record n+1.
> 
>      The expected behavior of the interpolation algorithm is as follows:
> 
>         Time range               Interpolation characteristics
>         ============             =============================
> 
>         Epoch of record 1 to
>         epoch of record n-1:     Interpolation is identical to standard EOP
>                                  file.
> 
>                                  Orientation error relative to that provided
>                                  by a binary PCK derived from the source EOP
>                                  file is non-zero due to differences in the
>                                  Chebyshev fitting process. Error is at the
>                                  1.1 nanoradian level.
> 
>         Epoch of record n-1 to
>         epoch of record n:       Nearly identical to standard EOP file;
>                                  small round-off error will occur.
> 
>                                  Orientation error is at the 1.1 nanoradian
>                                  level.
> 
>         Epoch of record n to
>         epoch of record n+1:     Some ringing of interpolated data
>                                  will occur. This interval has duration 1 day.
> 
>                                  Orientation error relative to that provided
>                                  by a binary PCK derived from a JPL EOP
>                                  released 4 days after the source JPL EOP file
>                                  (thereby providing accurate predicted data
>                                  4 days further into the future--up to 2020
>                                  July 2) is at the 225 nanoradian level.
> 
>         Epoch of record n+1 to
>         epoch of record n+2:     Some ringing of interpolated data will occur.
>                                  This interval has duration 1 day.
> 
>                                  Orientation error measured as for the
>                                  interval described above is at the
>                                  245 nanoradian level.
> 
>         Epoch of record n+2 to
>         epoch of record n+3:     Ideally nutation corrections and polar
>                                  motion will be constant, as will TAI-UT1R
>                                  derived from TAI-UT1 values in the file.
>                                  Some round-off error should be expected.
>                                  This interval has duration 1 day.
> 
>                                  Orientation error measured as for the
>                                  interval described above is at the
>                                  265 nanoradian level.
> 
>         Epoch of record n+3 to
>         epoch of record n+4:     Data are as described for records n+2 to
>                                  n+3. This interval has duration 79 years.
> 
>                                  Orientation difference relative to the
>                                  IAU_EARTH reference frame's rotation model
>                                  ranges from the 1.25 milliradian level at
>                                  the interval start to the 4.75 milliradian
>                                  level at the end.
> 
>>>> 
> 
> So, as you see, the only data source is the cited JPL EOP file. 
> 
> Since TAI-UT1 is frozen at the end of the EOP coverage interval, the orientation
> from the PCK for times after that epoch will be off by, on average, about 1 arcsecond for 
> each future leapsecond. Nutation and polar motion will be off too, but the magnitudes
> of those errors are dominated by that of the spin error.
> 
> This file enables operation of applications that require Earth binary PCKs
> having very long coverage, but that require only very low accuracy. For example, 
> using this file, view periods can be computed with an accuracy of a few minutes. 
> 
> Please let me know if you have further questions.
> 
> Best regards,
> 
>  -Nat Bachman (JPL/NAIF) 
> 
> Nathaniel.Bachman at jpl.nasa.gov
> 
> 
> 
> 
> ________________________________________
> From: Spice_discussion <spice_discussion-bounces at naif.jpl.nasa.gov> on behalf of Goodson, Troy D (US 392C) <troy.d.goodson at jpl.nasa.gov>
> Sent: Wednesday, July 13, 2022 12:23 PM
> To: spice_discussion at naif.jpl.nasa.gov
> Subject: [Spice_discussion] how are the long-term predicts made for ITRF93?
> 
> The tutorial https://naif.jpl.nasa.gov/pub/naif/toolkit_docs/Tutorials/pdf/individual_docs/23_lunar-earth_pck-fk.pdf
> as well as https://naif.jpl.nasa.gov/pub/naif/generic_kernels/pck/aareadme.txt
> describe long-term predicts for ITRF93:
> 
>   Low accuracy, long term predict kernel. The extended predict region of this
>   kernel---the time interval following the end of the predict region of the
>   input EOP file---does not estimate changes in UT1-TAI. The dates in the
>   file name are the file's coverage start and stop times:
> 
>      earth_200101_990628_predict.bpc
> 
> 
> But, I believe IERS only provides predicts for 1 year. See https://www.iers.org/IERS/EN/DataProducts/EarthOrientationData/eop.html
> and EOP2 files from https://eop2-external.jpl.nasa.gov are only predicted for about 78 days.
> 
> What is the source for the long-term prediction in a BPC kernel like earth_200101_990628_predict.bpc?
> 
> Thanks,
> 
> Troy.
> _______________________________________________
> Spice_discussion mailing list
> Spice_discussion at naif.jpl.nasa.gov
> http://naif.jpl.nasa.gov/mailman/listinfo/spice_discussion




More information about the Spice_discussion mailing list