[IGSMAIL-1816] improved polar motion predictions

Jim Jim
Fri Feb 27 07:23:48 PST 1998

IGS Electronic Mail      Fri Feb 27  7:23:48 PST 1998      Message Number 1816

Author: Jim Ray & Brian Luzum, USNO
Subject: improved polar motion predictions

Partly to address the IGS need for improved EOP predictions to use for
better GPS orbit predictions (IGP products), the IERS Bulletin A (Rapid
Service and Predictions) is being modified.  Historically, Bull. A has
concentrated on providing the best predictions of longer term EOP values
(up to one year in the future).  The change being made will significantly
improve the near-term performance of Bull. A polar motion predictions
(no change is being made to the predictions of UT1).  It will take effect
with Bull. A issue Vol. XI, No. 17, on Tuesday, 03 March 1998.

The following table compares the new and old performance of Bull. A
polar motion predictions:

    prediction interval     polar motion error (milliarcseconds)
          (days)                    old            new
    -------------------          ---------      ---------
             1                     0.442          0.206
             2                     0.882          0.515
             3                     1.322          0.893
             4                     1.749          1.296
             5                     2.143          1.682
             6                     2.497          2.044
             7                     2.820          2.374
            10                     3.712          3.282
            15                     5.186          4.705
            20                     6.436          5.965
            40                    11.026         10.755
            60                    14.757         14.651

    "polar motion error" means RMS error per component based on a
    retrospective analysis of actual polar motion variations during
    1994 - 1998.

It can be seen that the improvement is most significant for the shortest
prediction intervals (53% for 1 day) but dimishes over longer spans.

For real-time users, given two updates of Bull. A each week, the longest
prediction interval is 7 days (for Tuesday updates compared with the
previous Thursday issue which normally contains most recent data from
2 days earlier).  This means that real-time users should experience
polar motion errors no greater than about 2.4 mas (in an RMS sense
per component).

When the IGS delivery schedule for its Rapid products advances from
21:00 UTC to 16:00 UTC each day (by the end of 1998) this will allow
each Bull. A update to be performed during Washington afternoons with
a reduced data latency of 1 day rather than the current 2 days.  Thus
the maximum prediction interval for real-time users will shorten to
6 days with an associated RMS polar motion error of about 2.0 mas.
The Bull. A update schedule could be changed to Monday/Thursday from
the current Tuesday/Thursday to further reduce the maximum prediction
interval to 5 days with an RMS polar motion error of about 1.7 mas.
This has not been done in the past because most U.S. holidays occur
on Mondays.

Research is continuing into further improvements, which are likely to
be implemented later this year.

IERS Bulletin A Web site      http://maia.usno.navy.mil/bulletin-a.html

[Mailed From: Jim Ray (USNO 202-762-1444) <jimr at Maia.usno.navy.mil>]

More information about the IGSMail mailing list