RE: Decay date of a KH 8-52 SRV revealed

From: Ted Molczan (ssl3molcz@rogers.com)
Date: Sun Jan 13 2013 - 21:58:42 UTC


Jonathan McDowell wrote:

> I have now analysed the orbital data in the declassified document
> 54.pdf in this collection.
> It confirms previous rumours that the saved TLEs for the A object at
> http://www.planet4589.org/space/elements/13000/S13040
> are NOT CORRECT - they are fine through Apr 28 but from Apr 28 to May
> 23 they refer to a DIFFERENT OBJECT
> (or maybe something invented). On Apr 28 1982-06A moved to a low 150 x
> 308 km orbit and seems to have
> been untracked by NORAD for almost a month.  More details at
> http://planet4589.org/space/jsr/latest.html
> (permalink will eventually be  planet4589.org/space/jsr/back/news.673)
>
> I solicit analysis of the Apr 28-May 23 published TLEs for 13040
> (which I obtained from Allen Thomson's collection).
> What is this slowly decaying object? Does it later get retagged as
> something else?

This is highly interesting.

The apparent transition from reliable to doubtful TLEs occurs between epochs 82117.177 and 82118.130. Here they are as
recorded in Jonathan's archive:

1 13040U          82117.17705877  .00013321           14937-2    1601
2 13040  97.2339 195.5378 0008193 314.9290  45.0548 14.86887207 14136

1 13040U          82118.13027880  .00010471           14937-2    1613
2 13040  97.2339 196.3828 0008194 311.8286  48.2226 14.76834731 14273

At face value, the orbit was raised about 32 km in the interim, which would have required two manoeuvres. The time of
closest approach between the two orbits should be the approximate mid-point between the manoeuvres, but it occurs one
day too early, near 11:54 UTC (epoch 82116.4958).

Also of interest is that the TLEs spanning epoch days 82118 through 82142 appear to represent the identical orbit
propagated to different epochs. Note the constant value of inclination and B* throughout, which was carried over from
the epoch 82117 TLE. I tested this hypothesis by propagating the epoch 82118.130 TLE to the epoch of four subsequent
official TLEs, using a routine that extracts the mean elements generated internally within SGP4:

Archived TLE
1 13040U          82124.63437970  .00010505           14937-2    1652
2 13040  97.2339 202.2153 0008201 290.4308  69.6023 14.76971116 15235

Propagated 82118.13027880 TLE:
1 13040U 82006A   82124.63437500  .00010510  00000-0  14937-2 0    05
2 13040  97.2339 202.2152 0008201 290.4193  69.5920 14.76971112    06

Archived TLE
1 13040U          82130.52820390  .00010536           14937-2    1675
2 13040  97.2339 207.5015 0008207 271.0654  88.9617 14.77095082 16105

Propagated 82118.13027880 TLE:
1 13040U 82006A   82130.52820602  .00010540  00000-0  14937-2 0    02
2 13040  97.2339 207.5015 0008207 271.0189  89.0330 14.77095078    09

Archived TLE
1 13040U          82137.91172475  .00010564           14937-2    1715
2 13040  97.2339 214.1254 0008215 246.8489 113.1858 14.77250892 17199

Propagated 82118.13027880 TLE:
1 13040U 82006A   82137.91172454  .00010579  00000-0  14937-2 0    09
2 13040  97.2339 214.1254 0008215 246.7153 113.3563 14.77250890    01

Archived TLE
1 13040U 82006A   82142.04344872  .00010596  00000-0  14937-2 0  1734
2 13040  97.2339 217.8327 0008220 233.3163 126.7324 14.77338330 17806

Propagated 82118.13027880 TLE:
1 13040U 82006A   82142.04344907  .00010601  00000-0  14937-2 0    09
2 13040  97.2339 217.8327 0008220 233.1156 126.9921 14.77338332    07

The propagation employed a version of SGP4 that closely conforms to the Center for Space Standards & Innovation version
as of Fall 2008. Propagation was to the nearest second of time of the epoch of the official TLE.

Given the vagaries inherent in the natural evolution of orbits, and the imperfections in the orbital model and tracking
data, the negligible differences between the propagated and purported actual elements would be virtually impossible.

As another test, I computed the separation between the official epoch 82118 and 82142 TLEs, and found that in the
interim any difference was almost entirely along-track, ranging between zero and 8.45 km, which is improbably small.
Moreover, the difference evolved gradually, suggesting that it resulted mainly due to a slight difference between the
propagators underlying the TLEs, and the one used for the test.

> Is it really possible that NORAD failed to notice a 4.5 tonne polar
> LEO satellite for several weeks, or is this the first known instance
> of deliberate TLE disinformation?

It is highly unlikely for the GAMBIT satellite to have gone untracked for so long. The orbits in question do not
correlate with the four known debris objects, based on comparison with their near-contemporaneous TLEs. Deliberate
disinformation seems a possibility.

I offer the following speculation for discussion.

From launch through epoch 82117, the apogee and perigee of 82006A did not resemble that of any previous Gambit or
Hexagon mission, which could have been at least somewhat confusing to the Soviets. They might have wondered whether it
was an imaging or a SIGINT satellite. Certainly, they would not have known that the object was intended to eventually
lower its orbit to the standard Gambit dimensions.

I recall (source escapes me at the moment) that one motivation behind the later stealthy Misty IMINT satellite was the
Soviet ability to track the KeyHoles, which facilitated deception and denial. The sudden manoeuvre by 82006A to the
standard Gambit orbit might have served as a test of how quickly the Soviets would detect the change, as well as an
opportunity to collect some imagery in the interim, against targets (hopefully) not practicing the usual deception and
denial.

I do not know how long it took U.S. TLEs to reach the Soviets in those days, but perhaps the bogus TLEs were intended to
prolong their hoped for confusion. The results could have informed the development and argument in support of Misty,
which reportedly was approved in Spring 1983.

Alternative ideas welcome.

Ted Molczan


_______________________________________________
Seesat-l mailing list
http://mailman.satobs.org/mailman/listinfo/seesat-l



This archive was generated by hypermail 2b29 : Sun Jan 13 2013 - 21:59:13 UTC