Re: Question about the age of available TLEs

From: Bill Gray via Seesat-l <>
Date: Sun, 11 Feb 2018 20:37:26 -0500
Hi Gavin,

    Unfortunately,  there's not much documentation on how Space-Track
does what it does (as I found out when I started generating TLEs from
raw data).  They usually,  but not always,  place the epoch such that
the object is at the ascending node,  sometimes at the descending node,
and sometimes at the midpoints.  I think they see some convenience in
this;  at the nodes,  z=0,  and at the midpoints,  dz/dt=0.

    However,  since Space-Track doesn't document much,  I can't really
address the question of their specific elements very well.

    _Usually_,  accuracy for Space-Track elements is best at the epoch
and gradually decays as you get away from it.  I mostly work with the
asteroid community,  and have a program to check asteroid observations
to see if they match known artsats;  I keep old TLEs mostly so I can
match old observations,  taking whichever elset has the closest epoch
to the observations in question.

    But since Space-Track never tells you anything along the lines
of "this TLE was fitted to the motion over the following time span,
and matches to within the following accuracy",  you have to guess a lot.
The TLEs could match to within meters or to within a thousand km.

    As to other sources of TLEs:  there are,  of course,  Mike's
'classfd.tle'.  I think there are separately produced TLEs for
navigation satellites,  fitted to SP3 data,  though I'm not turning
them up right now.  I compute TLEs for several high-flying objects
in multi-day orbit,  mostly objects Space-Track doesn't track :

    In my case,  I compute a TLE to cover each day for a given time
span,  based on whatever observations I have of that object,  and I
provide a measure of how well the TLE fits the integrated ephemerides.
(Which,  usually,  is about how well the TLE would fit the actual
motion of the object,  except in some cases where the observations
aren't all that good.)

-- Bill
Seesat-l mailing list
Received on Sun Feb 11 2018 - 19:38:32 UTC

This archive was generated by hypermail 2.3.0 : Mon Feb 12 2018 - 01:38:32 UTC