Short version : back in 2015,

I (mis)identified this object as 2015-007B,
the second stage of the DSCOVR spacecraft. We now have good evidence
that it is actually 2014-065B, the booster for the
Chang’e 5-T1 lunar
mission. (It will, however, still hit the moon within a few kilometers
of the predicted spot on 2022 March 4 at 12:25 UTC,
within a few seconds of the predicted time.)

The full story

This morning (2022 Feb 12), I received an e-mail from Jon Giorgini
at the Jet Propulsion Laboratory
(JPL), inquiring about my statement that
“DSCOVR (and this rocket stage with it) passed close by the moon on 2015
February 13, two days after it was launched. Jon pointed out that
JPL’s Horizons system showed
that the DSCOVR spacecraft’s trajectory did not go particularly close to
the moon. It would be a little strange if the second stage went right
past the moon, while DSCOVR was in another part of the sky. There’s
always some separation, but this was suspiciously large.

Prompted by Jon’s e-mail, I dug into my e-mail archives to figure
out why I had originally identified the object as the DSCOVR stage
in the firs place, seven years ago. I did that digging in full
confidence it would prove that the object was, in fact, the DSCOVR
second stage.

How the object was originally (mis)identified as the DSCOVR second stage

On 2015 March 14, about a month after DSCOVR was launched, the
Catalina Sky Survey
found a possible near-earth asteroid. As is the usual practice,
they posted their observations on the
NEOCP
(Near-Earth Object Confirmation Page). This is where astronomers
post data about objects they’ve found that might be near-earth
asteroids or comets; the idea is that others can then try to observe
them as well and say “I found it, too, at the following location” or
“Hmmm, can’t find that one.” The discoverers, thinking they had a
rock, gave it the temporary name WE0913A.

These data are publicly available, and shortly thereafter,
an astronomer in Brazil
noted on a newsgroup that the object was orbiting the earth, not
the sun, suggesting it might be a human-made object.
I replied to
confirm this, and wrote that I thought it was either DSCOVR
or some bit of hardware associated with it, but that I was having
problems finding any data on DSCOVR’s trajectory.

Further data confirmed that yes, WE0913A had gone past the moon
two days after DSCOVR’s launch, and I and others came to accept the
identification with the second stage as correct. The object had about
the brightness we would expect, and had showed up at the expected time
and moving in a reasonable orbit.

Essentially, I had pretty good circumstantial evidence for the
identification, but nothing conclusive. That was not at all
unusual. Identifications of high-flying space junk often
require a bit of detective work, and
sometimes, we never do figure out the
ID for a bit of space junk; there are

some UOOs (Unidentified Orbiting Objects) out there.

Why I’m now sure it’s actually the Chang’e 5-T1 booster

In hindsight, I should have noticed some odd things about
WE0913A’s orbit. Assuming no maneuvers, it would have been in
a somewhat odd orbit around the earth before the lunar flyby.
At its highest point, it would be near the moon’s orbit;
at its lowest (perigee), about a third of that distance.
I’d have expected the perigee to be near the earth’s surface.
The perigee seemed quite high.

However, rocket hardware often does strange things in its
early days in space, with leftover fuel leaking
out and pushing it
around. That causes changes in the orbit, so that when you
try to figure out where the junk came from, you get a wrong (or
at least altered) answer. Such things happen routinely, but in
this case, it would have required an unusual (though possible)
sort of leakage, occurring several days after the lunar flyby. And
it would have had to be a pretty substantial amount of leakage.
Also, I didn’t have a trajectory for DSCOVR at the time, and the
lunar flyby seemed quite plausible (spacecraft often use a lunar
flyby to adjust their orbits).

So matters stood, until I received Jon’s e-mail. That prompted
me to look for earlier space missions that might account for the object.
It couldn’t have been up all that long before, because this is
a large, bright object; somebody would have seen it. So it had to
have been launched not long before March 2015, in a high orbit going
past the moon. Few objects go that high; most stay relatively near
the earth.

The preceding candidate launch was the Chang’e 5-T1 mission,
launched at 18:00 UTC on 2014 October 23.
Its booster was (we thought) never seen.

It’s unclear when the Chang’e 5-T1 booster would have gone by the
moon, but four days after launch would be a reasonable ballpark
estimate. Running the orbit for WE0913A further backward, I got a
lunar flyby on 2014 October 28 :

Orbital elements:  WE0913A = DSCOVR object = 2015-007B = NORAD 40391
   Perilune 2014 Oct 27.9200390 +/- 0.000152 TT = 22:04:51.37 (JD 2456958.4200390)
A1: 4.24e-9 +/- 4.51e-11   A2: 7.59e-11 +/- 3.01e-11
A3: -6.33e-10 +/- 4.72e-11 AU/day^2 [1/r^2]
Epoch 2014 Oct 28.0 TT = JDT 2456958.5                        Find_Orb
q  6826.0085634 +/- 99.4            (J2000 ecliptic)
H   27.4  G 0.15                    Peri.  110.9433956 +/- 0.14
                                    Node     7.0878255 +/- 0.13
e   2.054713052 +/- 0.0151          Incl.  119.1718318 +/- 0.13
71 of 85 observations 2015 Mar. 14-July 14; mean residual 0".553

…which is a quite close lunar flyby at about the right time.
But even more impressively, the orbit before the flyby has it in
the “usual” lunar transfer orbit. If we assume that perigee was
just above the atmosphere (as it usually is), then it happens
about half an hour after launch time, and the ground path starts
out near the launch site in China and proceeds almost due East.
In short, it looks exactly like a Chinese lunar mission ought
to look in every particular.

Orbital elements:  WE0913A = DSCOVR object = 2015-007B = NORAD 40391
   Perigee 2014 Oct 23.7643109 +/- 0.00152 TT;  Constraint: q=6500k
A1: 3.78e-9 +/- 2.24e-11   A2: 3.65e-10 +/- 2e-11   A3: -6.45e-10 +/- 6.51e-11 AU/day^2 [1/r^2]
Epoch 2014 Oct 24.0 TT = JDT 2456954.5                        Find_Orb
M   8.3102144356 +/- 0.053          (J2000 equator)
n  35.2592288110 +/- 0.0058         Peri.  141.9756875 +/- 0.14
a198802.40474 +/- 21.8              Node   301.7804670 +/- 0.15
e   0.967304355 +/- 4.7e-6          Incl.   27.0705663 +/- 0.16
P  10.21d                  H 27.4   G  0.15   U  9.3
q 6499.9727089 +/- 0.115    Q 391104.83678 +/- 44.1
71 of 85 observations 2015 Mar. 14-July 14; mean residual 0".848

To add to the new evidence,
Jonathan McDowell has sent
orbital elements for a LuxSpace
amateur radio cubesat that got a “ride share” with the booster,
and it’s a very close match.

In a sense, this remains “circumstantial” evidence. But I would
regard it as fairly convincing evidence. So I am persuaded that the
object about to hit the moon on 2022 Mar 4 at 12:25 UTC is actually
the Chang’e 5-T1 rocket stage.

---

* As I understand it, early rockets would
be left in orbit with leftover fuel in them, and the fuel would eventually
heat up and cause the rocket to explode, creating more junk. So a
‘pressure relief valve’ was added : pressure builds up, a valve opens,
and the leftover fuel leaks out. Home water tanks usually have a
similar valve in case pressure gets beyond what the system could probably
handle : better to have water released safely than sprayed all over
your house.