Re: TLEs with future epoch/very high-altitude artsats

From: Bill Gray via Seesat-l <>
Date: Sun, 28 May 2017 09:57:31 -0400
Hello all,

    I'm an occasional lurker on this list,  and missed this thread
from about four weeks ago.  In it,  there was some discussion of
the accuracy (or lack of it) of TLEs for high-altitude objects.

    I compute and post TLEs for about a dozen such objects,  with
periods ranging from two to 80 days,  at

    These are objects that are high enough,  and move slowly enough,
that the folks tracking near-earth asteroids have stumbled across
them.  They regard artsats as vermin contaminating their data;
their main interest is in identifying them quickly and moving on
to rocks that might clobber us.  But orbits for these artsats can
be computed from their observations,  and I do that.  (Almost all
of these objects are not tracked by JSpOC.  Or if they are,  they
aren't posting TLEs for them.  Which is,  in part,  why I'm doing
it;  we use these TLEs in identifying "asteroids" as actually being

    In reply to the questions about accuracy:  because the SDP4 model
used with TLEs was not really intended for this sort of high orbit,
it's hard to fit more than a day or two at a time.  If you look at
any of the files at the above URL,  you'll see that for any given
object,  I provide a series of TLEs,  each good for one day,  and
also give the "worst residual" (maximum error of the TLE during
that day,  relative to the numerically integrated ephemeris).

    Usually,  that error is a few dozen kilometers.  Sometimes,
it's a lot worse.  If the object just went past the moon,  for
example,  or if its eccentricity is really high,  the SDP4 model
breaks down abysmally.  For some days,  you may not get a TLE
that fits at all.

    You'll note that some of these are predictive TLEs with epochs
a year or so in the future.  Some of these objects go a long time
between sightings;  the predictions won't be spot-on,  but should
allow us to say,  "That interesting 'asteroid' you just spotted is
actually space junk thus-and-such."

    JSpOC doesn't provide such error estimates with their TLEs,
of course.  They actually do provide decent TLEs for a couple of
high-flyers (1983-020A and 1983-020D,  for example,  which are
in four-day orbits).  They also provide TLEs that are totally
wrong,  for IBEX and the Chang'e boosters,  for example.   Mostly,
we're on our own for tracking and computing orbits for these guys.

-- Bill
Seesat-l mailing list
Received on Sun May 28 2017 - 08:58:10 UTC

This archive was generated by hypermail 2.3.0 : Sun May 28 2017 - 13:58:11 UTC