The re-entry of IGS 1B on 26 July 2012

While I was away on holidays, the Japanese spy satellite IGS 1B (03-009B) that malfunctioned in March 2007, re-entered on 26 July 2012, over Polynesia.

The demise of this satellite was covered for over a year on this blog: as the satellite was weighing 1.2 tons and as it had an unknown amount of remnant fuel onboard, the uncontrolled reentry raised some questions and initial concerns (see coverage here).

The last amateur observations of the object were done by Mike Waterman (USA) on July 24th and Alain Figer (France) in evening twilight of July 25th. The last amateur orbital update by Ted Molczan based on a.o. Mike Waterman's observations, showed it to have descended to a 211 x 213 km orbit on July 24th and analysis of this dataset by this author using Alan Pickup's SatEvo suggests reentry on July 26, somewhere between approximately 9:50 and 10:50 UTC.

USSTRATCOM published a final TIP for IGS 1B on July 26th (that they did so for a classified object is unusual), placing re-entry at 26 July 2012, 09:52 +/- 2 min UTC, near 25 S, 186 E, which is near New Zealand. This is at the start of the reentry window given above and hence seems very reasonable even though the reentry coordinates are a verbatim copy (down to one decimal) of a pre-decay prediction issued at 7:34 UTC (only the uncertainty value has changed, from 2 hours to 2 minutes). No details on the orbital development in the final few revolutions were given.

The map below shows the USSTRATCOM determined reentry location and final trajectory. In principle, the re-entry could have been observed from the northern islands of New Zealand and potentially the Fiji-Tonga area. Note that only half a revolution later (about 30 minutes later) it would have passed over NW Europe and next west Africa.

click map to enlarge

The diagram below shows the orbital evolution in terms of apogee and perigee altitudes, from malfunction early 2007 to decay on 26 July 2012. It is based on orbital element sets calculated by Mike McCants and Ted Molczan from amateur observations, including mine:

click diagram to enlarge

2

IGS 1B was a nice object to observe over the past years: it was bright, and it was interesting to follow its orbital evolution towards decay. The observation that remains the most vivid imprint in my memory is the one that resulted in the picture below: on 2 September 2011, while I was watching and photographing a pass in a slightly hazy sky, the satellite brightly flared to at least magnitude -8 if not more: the brightest satellite flare I have ever seen. I was jumping up and down and yelling "WOOOOOWWWW!!!!" when this happened. It resulted in this wonderful, eerie picture:

click image to enlarge

Update on IGS 1B (10 June 2012)

We are slowly getting closer to the uncontrolled re-entry of the 1.2 tons malfunctioned Japanese spy satellite IGS 1B (2003-009B). The satellite's orbit has by now dropped below 340 km altitude (see second diagram below, perigee/apogee values from orbits calculated by Mike McCants based on amateur observations including mine). That is well below the ISS orbital altitude (ISS is currently in a 392 x 406 km altitude orbit). A mere month ago it was still considerably higher, in a 366 x 368 km orbit: it lost over 25 km or orbital altitude since. It is dropping fast, and the rate increases (see diagram below).

Using the latest amateur orbital elements for the object and Alan Pickup's SatEvo software with the current 10.7 cm solar flux, re-entry is currently estimated to occur between mid-July and early August 2012.

An update on IGS 1B - a spy satellite about to reenter this summer

(updated reentry prediction at the end of this post)

Last year I wrote extensively about the malfunctioned Japanese spy satellite IGS 1B (2003-009B) which is about to reenter, uncontrolled, into the atmosphere soon. The above footage of the object was shot by me last April 13 and 14 and shows two passes over Leiden.

This radar satellite, launched to keep an eye on North-Korea in 2003, malfunctioned in March 2007, halfway though its mission. Since then, it has been steadily coming down (see diagrams below) in a way that clearly shows that the satellite operators do no longer have control over it.

Last year, I pointed to the fact that the 1.2 tons  satellite will reenter in 2012 (so this year), and likely still has some remnant fuel onboard. A subsequent assessment by high-end amateur satellite tracker Ted Molczan showed that this amount of fuel is limited - probably about 14 to 50 kg, an order of a magnitude less than the infamous case of USA 193 in 2008. This assessment is important, as an uncontrolled reentry of a satellite with fuel onboard is a potential hazard (reason why I wrote about it last year) and authorities were (and are) very quiet about it. Ted's assessment, the only public one to date, helped to put the potential risks involved into proper context.

In the autumn of 2011 we temporarily lost track of IGS 1B because it entered winter invisibility for the Northern hemisphere (where most of our observers are located). Early April this year, it emerged from this winter blackout again. I did a failed attempt to recover it on the evening of April 2, and then Russell Eberst successfully recovered it a day later on the evening of April 3. Since then, I observed it on April 13, 14 and 22 (see video footage above of the April 13 and 14 passes) and other amateurs have observed it as well.

Below is a 35-second integration of video frames from the April 13 video (upper right are tail stars of the Big Dipper):

click image to enlarge

Orbital evolution over the winter blackout

When IGS 1B was lost in the winter blackout in the autumn of 2011, it was in a 453 x 455 km orbit. Since then, it has come down considerably: as of 2012 May 1 it is in a 366 x 368 km orbit, almost 100 km lower (and now below the orbital altitude of the ISS). It is coming down at an increasingly fast speed, as the diagrams below show (based on orbital calculations by Mike McCants, derived from amateur observations which include my observations):

click diagrams to enlarge

 

Current Decay Prediction

Using Alan Pickup's SatEvo software with the current orbit and solar activity, I expect the reentry of IGS 1B to occur somewhere during a window that spans from June until August. As the orbit is evolving fast, it is pertinent that we keep close track of the object in order not to lose it (a few days old elements already results in several minutes uncertainty in pass time).

April 20 Brazilian fireball was NOT the reentry of an Atlas Centaur r/b

A beautiful, slow, long duration, fragmenting fireball was filmed from Brazil in the evening of April 20 (20 April near 20-21h UT). Movies can be seen here and here.

Soon after the apparition, the question came up: meteoric fireball, or a reentry of "space junk"? Based on the visuals of the two movies available, I noted that it looked like a meteoric fireball, not a reentry.

Subsequently, a piece appeared in Universe Today, claiming it was a reentry: that of an Atlas Centaur rocket, #16102, 85-087B, which launched Intelsat 512 in 1985.

That conclusion is simply wrong however. According to USSTRATCOM the object in question decayed 18.5 hours earlier than the fireball, over the western Pacific at 1:23 UTC (20 April) near 18N, 161 E. It reentered nowhere near Brazil. It did pass over northern Brasil on its final pass an hour before reentry, but that would have been at 00:30 UTC, not 20-21 UTC. Moreover, even that pass would have been too much north to see it from the southern Brasilian location where it was filmed from: it would have passed below the horizon as seen from there.

And even if pieces would have survived longer (highly unlikely in this case), these could not have caused a fireball over Brasil 18.5 hours later: the orbital plane of the r/b was wrongly oriented for that. Around the time of the fireball, any surviving objects in the Centaur orbit would not pass over Brasil but much more north, over Mexico and the Caribean (see below. Note: the object was no longer in orbit at that time!).

So, the object that was filmed was most likely a piece of asteroidal debris, a very nice, very slow and very long duration meteor grazing through the upper atmosphere and breaking to pieces.

Space debris lands in Brazil village

Through Carlos Bella on the satobs mailing list, news broke today that an object which almost certainly is space-debris crashed in the Brazilian village of Anapurus on February 22, 2012, near 6 am local time (9 UTC). It landed about 6 yards from a house and damaged trees upon impact.

Photo's of the object can be see here.They show a spherical object that strongly resembles a spherical rocket fuel cell (tank) or a Helium pressurization tank. These are the most resistent objects among space debris and often involved in reported cases of space-debris reaching earth surface.

Ted Molczan quickly noted that date, time and location correspond well to the re-entry of 1997-016C, an Ariane 44L rocket stage from the launch of two geostationary satellites, Thaicom 3 and Bsat-1A, on 17 April 1997.

The Ariane 44L r/b in question re-entered at 9:09 UTC +/- 1 min on 22 February 2012, near 4 S, 312 E. This corresponds well with the time and location of the Anapurus event (3.7 S, 317 E). Anapurus is located right on the re-entry track and was passed within a minute of the estimated re-entry time (movement of the r/b was from West to East, i.e. to the right in the map):

click map to enlarge

Phobos-Grunt: a final TIP on where it came down

USSTRATCOM today (Jan 23rd) has finally released a traditional format final TIP for Phobos-Grunt. It yields similar values to the ones initially circulating through Russian press releases (see earlier post here): 17:46 GMT (+/- 1 m) and a location in the southern Pacific, near 46 S, 87 W.

click map to enlarge

More thoughts on the Phobos-Grunt reentry

It appears that the Russian news bulletins claiming that Phobos-Grunt reentered over the southern Pacific at 17:45 UT yesterday (see my previous post) rather were model predictions than reentry times based on actual final track detections. The Russians (nor the US, for that matter) actually do not have much in terms of tracking facilities in the indicated area (S-Pacific, S-America and S-Atlantic).

The southern Oceans: a blind spot

This highlights the problem, and the similarity between the UARS case last September and the current Phobos-Grunt case. In both cases, determining where it came down was likely hampered by the final revolution and final half hour or more of its trajectory being largely over remote and empty territory. Specifically, in both cases: the southern Oceans.

A lack of tracking facilities (and humans in general) in these areas, mean that they represent a large blind spot for those who's profession it is to track these objects. Once a satellite near decay starts to make its final passes and significant parts of that happen to be over these remote locations, it basically disappears into a black hole. That's what happened with UARS in September, and what now happened with Phobos-Grunt.

Not as in the movies

While some in the media and public have expressed frustration about the lack of published information (and the contradictory information) right after Phobos-Grunt presumably came down, I feel those people lack a realistic outlook on these matters. These people apparently expect that the military is able to determine a clearcut point of reentry within minutes after the satellite has reentered.

In reality, this is not how it works. The military has intermittent detections when the object moves over tracking facilities, spread wide and far over a number of places around the globe. They do not have continuous coverage. They cannot track where they have no tracking facilities. And the big and empty southern oceans represent a large swath of Earth where that is the case.

In Hollywood movies and TV-series, the military (or "NASA", even though in reality it is not NASA doing the tracking) are portrayed as having a second-to-second real-time tracking opportunity of objects, with a moving dot on the screen that disappears in real-time as soon as the satellite reenters. That however, as most things in movies, is a highly unrealistic view which has little resemblance to reality.

It will take the professional analysts at USSTRATCOM and elsewhere some time to ponder the last tracking data, detections and non-detections, and maybe even then there will remain uncertainty about where Phobos-Grunt came down. Such is life, and reality.

Space-Based detections?

In the case of UARS last September, some of us have had some suspicion that Space-Based observations (Infra-Red detections by the early warning satellites of Missile Defense) were perhaps involved in the final determination of the point of reentry. These resort under another part of the military, and comments by those "in the loop" have indicated that normally there is little data exchange between these guys and the groundbased tracking guys. There is also the open question whether reentry fireballs are bright enough for these space-based systems (the DSP and SBIRS satellites) to detect them. As the specifications of these systems are of course classified, little is known about this.

[Updated] Phobos-Grunt down over the southeast Pacific?

click map to enlarge

Early reports in the Russian media report that Phobos-Grunt has come down over the southern Pacific Ocean at 17:45 UT (15 Jan).

These same Russian media have misinterpreted decay bulletins in the (recent) past however, so this is all under some caveat until clear data from non-media sources (e.g.a USSTRATCOM TIP message) appears.

The map above is a preliminary map based on orbital elements from 16:35 UT, an hour before the reentry, and shows the approximate position for Phobos-Grunt for 17:45 UT. There  might be some small discrepancy between the pictured position and real position as it is based on orbital elements from one hour before the reentry.

Note added 20:30 GMT:  Confusingly, Ria-Novosti in a new press release mentions another time and position: 17:59 UT and the Atlantic, 18S and 311 E (49 W). This exemplifies what I mean above with the caveats....

Update 20:40 GMT: USSTRATCOM released a TIP message at 20:34 UT simply saying: "Object Decayed Inside Predicted Window".
That probably means: within the window of their earlier TIP message: 16:59-17:47 UT. This could be taken to support the 17:45 UT value of the Russians and reentry over the S-Pacific.

Summary of current Phobos-Grunt reentry predictions (15 jan 11:00 UT)

Summary of current Fobos-Grunt reentry predictions, round-up of predictions compiled at 15 Jan 11:00 UTC:

USSTRATCOM (14 Jan 20:30): interval 14:52 - 19:40 UTC
Harro Zimmer (15 Jan 07:10): 17:49 +/- 45 m UTC
Ted Molczan (15 Jan 04:59): 20:39 +/- 3 hrs UTC
Aerospace Corp. (15 Jan 06:25): 19:21 +/- 3 hrs UTC 
Roscosmos (14 Jan):  18:36 UTC (uncertainty window not stated)

Listed are: source; time prediction was issued (UTC); predicted time and uncertainty interval, or predicted interval.

My own current prediction, using Alan Pickup's SatEvo, current F10.7cm flux and 09:21 UT epoch orbital elements:  20:07 +/- 2 hrs.

PLEASE take note of the still large uncertainty intervals. As yet, it is still impossible to realistically say where it will come down.

[Updated] Summary of Phobos-Grunt reentry predictions as off 14 Jan, 13:00 UTC

Current predictions by various sources for the reentry of the failed Russian Mars probe Fobos-Grunt all still have a very large uncertainty window. Therefore, it is still impossible to provide any realistic statements about the when and especially where the reentry will happen. Suggestions in the press such as these only demonstrate a continuing and thorough misunderstanding of the whole issue among some journalists.

Perhaps as a result of this, USSTRATCOM has changed the way it presents it's Fobos-Grunt reentry predictions. While for other objects they still use the classical TIP format (time with uncertainty interval, and position at the nominal time), they altered the presentation for F-G: they only list the uncertainty interval, without positions or nominal time. Given the way that TIP messages were taken completely out of context and misinterpreted in the press recently (e.g. the notorious and utterly misinformed Novosti-RIA publication here), this is all a very understandable precaution.

A summary of the current predictions (round-up of predictions at 14 Jan, 13:45 UTC. Listed are: source; time/date prediction was issued; predicted time plus uncertainty interval, or uncertainty interval only):

USSTRATCOM (13  Jan 15:30 UTC): 15 Jan 08:25 - 16 Jan 03:37 UTC
Harro Zimmer (13 Jan 11:51 UTC): 15 Jan 17:05 +/- 3 hrs UTC
Ted Molczan (14 Jan 08:35 UTC): 15 Jan 22:31 +/- 8 hrs UTC
Celestrak (T.S. Kelso, 14 Jan 11:00 UTC): 17 Jan 10:04 +/- 1 day UTC
AeroSpace Corp. (13 Jan 10:40 UTC): 15 jan 17:52 +/- 14 hrs UTC
Roscosmos (14 Jan UTC):  15 Jan  21:51 17:51 UTC (no uncertainty window listed)

[added 15:45 UT, 14 Jan] My own estimate, using Alan Pickup's SatEvo and the current space weather (F10.7 cm flux) and the Jan 14.54 orbit, is for 15 Jan, 21:50 +/- 7 hrs UTC.

PLEASE take note of the large uncertainty windows on all these predictions! The uncertainty amounts to many revolutions of F-G around the Earth, so it is impossible to even indicate a Continent or Ocean (as some media sources do) where F-G will end up, at this moment.

Most of the predictions favour the 2nd part of Sunday Jan 15th.

Gearing up for the Phobos-Grunt reentry

We are in for an interesting weekend, as the failed Russian Mars probe Phobos-Grunt is experiencing it's last days of existence. Gradually having come down over the past two months, it is expected to re-enter and burn up (but perhaps not completely) in the Earth atmosphere on Sunday or Monday.

Several days before the re-entry date, it is still not possible (whatever some news outlets erroneously write) to pinpoint when and where it will come down. At the moment of writing (early Friday), the SSC prediction amounts to a still over a day wide window between 15 Jan 02:40 and 16 jan 07:40 UTC. Harro Zimmer's latest prediction is for Jan 15 between 9:00 - 15:00 UTC. Both predictions encompass multiple revolutions around the earth. Please note: all these time windows can still shift, depending on actual developments in space weather (solar activity)  and other factors.

Below video has been posted here before, and shows a Fobos-Grunt pass filmed by me from Leiden, the Netherlands, on November 28:

Further confusion on Saturday's Soyuz r/b reentry

I earlier wrote about the confusion reigning in the press concerning the sky event over Europe of last Saturday evening. Initial confusion was over wether it was a meteor, "comet" or (and that was the correct explanation, but many Dutch and German news outlets failed to properly pick that up): the reentry of  a Soyuz rocket.

Now a new confusion has arrisen: some news outlets and weblogs, e.g. that of Physorg, mistakenly link the event to last Friday's failed Russian launch of the Meridian satellite. Due to a rocket failure, this never reached earth orbit but crashed in Siberia within minutes after the launch.

As I wrote earlier, what reentered and was seen over France, Germany and the Netherlands last Saturday evening, was the 3rd stage of last Wednesday's Soyuz launch to the ISS.

The confusion probably comes from the fact that both launches used a Soyuz rocket. The failed launch that crashed in Siberia on Friday got some press attention because fragments hit a house in Russia (see a.o. here (English), with pictures of a recovered fuel tank here (Russian)).

But again: that failed launch had nothing to do with Saturdays sky event over Europe. The reentry over Europe was the 3rd stage of the earlier Wednesday launch to the ISS, that included Dutch astronaut Andre Kuipers.

(More on Last Saturday's Soyuz reentry over Europe: here and here)

11-078B Soyuz 3rd stage reentry: answers to some frequently asked questions

In the wake of the spectacular reentry over NW Europe of the Soyuz 3rd stage 2011-078B on Saturday 24 December 2011, several common questions popped up in comments, e-mails, on Twitter and in newspaper discussions. I will answer a few below.

Frequently Asked Questions:

(Q1): Are these things predictable and who makes such predictions?
(Q2) Does it really take a Soyuz rocket 3rd stage three days to fall back to earth?
(Q3) Why doesn't this happen with each Soyuz launch? Or: why not over the same location on Earth?
(Q4) has anything of the rocket stage survived to earth surface?

Answers:

(Q1): Are these things predictable and who makes such predictions?

(A): It is "sort of" predictable. Using computer models which take into account many factors of influence, one can make a prediction yielding an indication of the time a rocket stage or satellite will re-enter the atmosphere. However, even very close to that actual time of reentry, the uncertainty in these predictions is still very large. Exactly when a rocket stage will start to burn up depends on many factors, including the exact condition of the atmosphere at that moment, the shape of the rocket stage, and whether it is tumbling or not. In practise,  this turns out to be very difficult to model, even with the best computer models.

Several organisations and individuals do such predictions (and you can even find software for it on the internet). However, one of the most authoritive sources of such predictions is USSTRATCOM, the American military organisation that tracks manmade objects in space (many people think NASA does that job. But that is incorrect: it is USSTRATCOM, better known as 'NORAD'). They publish these predictions as 'TIP' messages. Their first prediction is published 2 months in advance. These still have a very large uncertainty (think of: weeks). In the days close to decay, they publish new estimates as new TIP messages that gradually become more exact. But even these can have uncertainties of several hours, even for predictions made on the day of the reentry itself.

For example, the last pre-reentry TIP message issued only 2 hours before the Soyuz 3rd stage came down, still had an uncertainty window of six hours.....

Once an object has reentered, USSTRATCOM does a post-analysis of the last orbital information, and publishes a "final' TIP message mentioning when and where the object came down (so this is done "after the fact"). These can be (but are not always, as it depends on how well the object was tracked during it's last hours of existence) very accurate. Sometimes, as was the case with this reentry of the Soyuz 3rd stage, they provide a time with an uncertainty of only minutes, plus a quite accurate position. In other cases, where less recent tracking data is available, the final uncertainty is much larger.

Note that a re-entering satellite or rocket booster has a speed of 7.5 km per second (4.7 miles per second)! So even if the predicted time has an uncertainty of just 15 minutes, this amounts to an uncertainty of 13,500 km (8,400 miles) in the position of the object when it reenters! This is why it is impossible to pinpoint the expected point of re-entry beforehand, when it is not a "controlled" re-entry. (in a "controlled" re-entry, the satellite operators send a command to the satellite to make a rocket burn at a precise time, kicking it down over a designated spot, usually the Pacific ocean. This Soyuz reentry was however not such a "controlled" reentry).

Many people mistakenly think that in this day and age of supercomputers, scientists (or the military) can predict everything. In reality, satellite/rocket reentries like this are so complex that even the best computer models can only give rough indications untill just minutes before the actual re-entry.


(Q2) Does it really take a Soyuz rocket 3rd stage three days to fall back to earth?

(A) Yes, it does. That last rocket stage is jettisoned that high above earth surface, that it does not just rapidly fall back on a ballistic trajectory (such as the 1st and 2nd stages do) but actually reaches Low Earth Orbit, and stays in orbit around the earth for several days. In effect, it becomes a satellite for a while in a very low orbit around Earth. Under influence of gravity and drag from the outer atmosphere, the orbit slowly evolves and becomes smaller and smaller. On the first day only gradually, but as it slowly comes down, this gradually goes faster and faster.

The influence of our atmosphere reaches several hundreds of kilometers up: even the International Space Station experiences some atmospheric drag, and would fall down within a year if its orbit was not regurlarly raised using the rocket engines of the Progress spacecraft docked to the ISS.

It takes about 3-4 days for a Soyuz 3rd stage from a launch to the ISS to come down. The exact amount of time is variable and different in each new case, as it depends on many factors. Our atmosphere is variable in extent and density, notably under the influence of solar activity. When the sun is active and many charged particles from solar outbursts reach earth, these interact with our atmosphere and the atmosphere slightly expands as a result of this. This means that objects at the altitude of the Space Station and below that (such as the Soyuz rocket stage) will experience a "thicker atmosphere", i.e. more drag from the atmosphere's outermost layers, and as a result they will come down faster. When it reaches at altitude of only 120 km (75 miles) it goes very quick: within minutes the rocket stage has dropped tens of kilometers, slowed down considerably, and finaly plunges straight down from that moment onwards.

The exact moment this happens, is highly dependant on these variations in extend of our atmosphere due to variations in solar activity. This is another reason why a satellite or rocket re-entry is so difficult to predict: one short but intense outburst occuring on the sun will next make a rocket stage fall back much quicker than expected.

Below diagram shows the orbital evolution of the Soyuz 3rd rocket stage that decayed last Saturday. It had to make 52 full orbits (full circles) around the Earth before it burned up. It's orbit was a bit "eccentric", which means that it was not a perfect circle but an ellipse. So on each revolution around the earth, there is a point where it is a bit higher above earth (called the "apogee") and a point where it is closest to the earth (called "perigee"). In the diagram, the values for these altitudes have been plotted as a red and blue line. Note how fast these altitudes change in the final hours before re-entry.

(Q3) Why doesn't this happen with each Soyuz launch? Or: why not over the same location on Earth?

(A) It does happen with each Soyuz launch to the ISS. The Soyuz 3rd stage always comes down some 3-4 days after the launch.

That reentry however is never over the same location on earth. The reasons for this, have already been outlined as part of the answer to question (2) above. An important factor of influence on how quickly a rocket stage comes down, is the variable earth's atmosphere, under influence of variability in solar activity. These factors are different for each new case. This is why the 3rd stages of Soyuz launches to the ISS never fall down near the same spot twice.


(Q4) has anything of the rocket stage survived to earth surface?

(A) Not that we so far know of. Usually, the rocket stage almost completely burns up in the atmosphere. Sometimes, a few smaller bits survive (quite often spherical fuel tanks). For example, an object that is likely a rocket fuel tank came down in Namibia in November and might be part of a rocket stage used in a Russian November  launch to the International Space Station.

[Updated] Breaking News: Decay of Soyuz r/b stage from André Kuipers' launch to ISS observed from the Netherlands!

UPDATE - the final TIP for Soyuz r/b 38037 / 2011-078B has been released by USSTRATCOM near 18h GMT and it indeed shows that this was the Soyuz r/b: reentry time is quoted as 16:25 +/- 1 minute GMT at 49 deg N, 7 deg E. This fits the observations well.

In the Dutch press, there meanwhile appears to be a lot of confusion. The Dutch National Police claims that they talked to "NASA" who apparently said it was a "meteor" (or "comet"). So THAT is widely claimed in the press now, to the point of calling the identification with the Soyuz 3rd stage "speculation". Which it is not: it is based on factual data and now clearly confirmed by the USSTRATCOM JSpOC TIP message. What more do you want?!

I have no idea to whom (or even where: NASA is big...) the Police spoke, but for all things it could have been the JPL janitor....

At any rate: appart from my analysis below (which is already clear), the USSTRATCOM TIP message mentioned above makes unambiguously clear that this was the Soyuz 3rd stage.

Note that to access the USSTRATCOM TIP message via the link above you need an approved account. USSTRATCOM is the US military Command responsible for tracking manmade objects in space, and perhaps better known under their former name NORAD.

- end of update

Multiple reports are coming in, among others by experienced Dutch meteor observers Carl Johannink (Gronau) and Arnold Tukkers (Denekamp), of a bright and very slow fragmenting object seen low in the west-southwest near Venus at 16:26 UTC, 24 December.

From the descriptions it clearly was a reentry of an artificial object (space junk), as the event was too long in duration and too slow to be a meteoric fireball.

And it was not "just" a random bit of space debris, it turns out:

The observations fit with 2011-078B (#38037), the last stage of the Soyuz rocket that brought Dutch astronaut Andre Kuipers up to the ISS earlier this week. It was already predicted to decay near this moment by USSTRATCOM.

Below is the  predicted trajectory of the Soyuz  3rd stage for the Gronau/Enschede area (and below that, the ground trajectory). It is based on an orbit with an epoch near noon of 24 December (epoch 11358.49032868. Source: USSTRATCOM), so a few hours old, which will introduce some minor discrepancies (a few seconds in time). But it fits the descriptions very well in terms of time and trajectory in the sky.

click images to enlarge

[UPDATE 7 October 2017]:

I recently modelled the re-entry of 2011-078B in GMAT, using the MSISE90 model atmosphere with actual Spaceweather of that time. Drag surface was set at 60% of the maximum drag surface for a Soyuz upper stage: this yields a decay position and decay time well in agreement with the JSpOC TIP position and is close to what the drag value for a tumbling, fragmenting object would be.

As seen from Gronau in Germany, it yields the following sky trajectory. Compare with Carl Johannink's description below: it matches his description well.

click map to enlarge

- end of update

- continuing original post:

Some quickly translated descriptions by two experienced Dutch meteor observers (compare to the sky trajectory map above for their area):

Arnold Tukkers, Denekamp (Netherlands):

At 17:26 CET (=16:26 UTC) I looked out of the window and saw a strange phenomena just above the rooftops behind us. It looked like a very, very slow meteor fragmented in several pieces. Like Peekskill but less bright.
Multiple fragments. Because it was so low in the sky, I walked upstairs and could still see the last part from the bedroom window. So it at least took 20 seconds. [...]
What a sight! Trajectory for me (did not see initial part) southwest-southeast. Altitude maximum 20 degrees. Colour brown-red.

Carl Johannink, Gronau (Germany):

Just was looking at Venus in evening twilight.
Left of it an object appeared from behind a cloud that I first thought to be an aircraft, but next I found something was not right. The thing sometimes brightened and became fuzzy, trailing a circa 8 degree long tail. Maximum brightness about -4.
The object roughly moved from SSW to SE at an elevation of about 15 degrees. The whole phenomena took over half a minute.

To see the second part of the trajectory I had to walk to a different room. Called in Elisabeth, together we saw the object fragment into pieces (each individual piece about mag. 0 to +1) and then fade out.

The whole event looked much alike to the New Years Eve satellite decay of 1978, albeit being somewhat less bright.

Update:
A number of video's from Germany have surfaced which likely show the event. Here are a few:
video 1
video 2
Video 3 (on the Bad Astronomer's blog)
Video 4
Video 5

FAQ

Read the answers to Frequently Asked Questions for this reentry case I published later here.

If rescue fails, Fobos-Grunt will reenter soon

As new attempts to contact the probe failed, the future is looking increasingly grim for Fobos-Grunt (aka "Phobos-Grunt" and "Phobos-SOIL"), the Russian space probe launched on 8 November that should have gone to the Martian moon Phobos for a sample return mission, but instead got stuck in Low Earth Orbit.

The probe is currently stuck in a very low orbit measuring 207 x 339 kilometer after it's propulsion unit apparently failed, failing to lift it into a GTO (and from there an interplanetary trajectory):

If the probe isn't revived in due time - and the Russsian operators are still frantically trying to do so - it is doomed. With an orbit at this low an altitude, it is a short matter of time before it comes down again - another case of an imminent uncontrolled reentry of a very large satellite (over 13 metric tons, including the fuel). How much fuel is onboard is not clear to me: different media sources quote quite different amounts, but all amounts quoted are in terms of several tons, with several sources settling for 7 tons (see also here).
As pointed out here by Anatoly Zak, seizable chunks of the probe could survive reentry, and survival is certainly expected for the actual Fobos sample return capsule (which was designed for reentry).

Reentry estimates

Estimates of when Fobos-Grunt  will come down are a bit complicated. Ted Molczan has noted that over the past day, the orbital evolution was unusual - Ted points out that if the last two orbit releases are not faulty (a  possibility), it means that either the probe is manoeuvering (which from all the negative statements by the Russians in the press seems unlikely) or - more likely - has started to vent fuel since yesterday. As a result, it might have gotten a very mild orbital boost (the leaking fuel acts like a small rocket engine).

Before this unusual behaviour started, orbital elements 11314.14749491 to 11314.77184893 and Alan Pickup's SatEvo software with current solar activity levels suggested a nominal reentry time no later than early January 2012. SSC meanwhile predicts reentry for November 26th, 2011. The current unusual orbital evolution - if real - might change things a bit, but eventually it will come down.

This means it will likely come down somewhere over the next few weeks or months if the operators cannot revive it over the next two weeks. With an orbital inclination of 51.4 degrees, it can come down anywhere between 51 N and 51 S latitude.

There has been a call out to observers to observe the probe - it's brightness behaviour can yield clues as to whether it is starting to tumble, e.g. because of the suspected fuel venting. So far, observations by Brad Young and Michael Murphy from the US suggest the probe is stable in brightness with no sign of tumbling (see here and here).

Unfortunately, the probe is currently not visible from NW Europe where I am located: it makes passes near midnight, completely in shadow.

ROSAT down over the Indian Ocean [RENEWED UPDATE]

ROSAT is no more.....according to a TIP bulletin by SSC issued at 03:41 UTC, it reentered at 1:50 UTC +/- 7 minutes, placing it over the Indian Ocean (and far away from any eyewitnesses, bar maybe some ships and maybe Ceylon, Sumatra and Birma/Malaysia/S-China if it survived into the second half of the given window: and of course the US tracking facility at Diego Garcia).

See update 2 below for latest map version (update 1 below now deprecated)

update 1 - deprecated. 
[Below two map shows the final orbit and (thick red line) the reentry windows according two assessments:
- The first is with the last available TLE (issued over a day before the reentry) propagated with SatEvo.
- The second, is made by adjusting the time (by about 5 minutes) so that the satellite position matches the nominal position for the reentry given by SSC in the TIP bulletin: 7 N, 90 E]

click maps to enlarge

[As can be seen there is some difference between these two. This is the result of there currently being no recent TLE available: the last available TLE dates 24 hours before the reentry. In the 24 hours between that last issued TLE and reentry, the orbit evolved fast. Without the availability of more recent elements, it is difficult to assess where the satellite exactly was along its orbit. That uncertainty is no more than a few minutes in time, but that amounts to over 2000 km in position....]

UPDATE 2: SSC released a new tle, with an epoch dating to two hours before the decay (about 1.5 revolutions). This allows this map to be created, which closely tallies with map 2 above:

click map to enlarge

It shows the difference a new TLE much closer in time to the decay makes, with regard to locating the satellite in its final moments....

(note: thanks to Daniel Fischer for inquiring about the differences between my map and Simone Corbellini's map, and to Simone for communications on the why of the time offset)

ROSAT reentry update (2)

Update to my previous post: Space-Track (SSC) finally released a new elset, 11294.85810509, which is still and "old" elset (almost a day old). And they released a new TIP with a new reentry prediction.

The new TIP gives this prediction:
Space-Track (SSC): 23 Oct, 02:34 UTC +/- 7 hrs

Based on the new TLE, independant analyst Harro Zimmer now provides the following prediction:
Harro Zimmer: 23 Oct, 04:17 UTC +/- 4 hrs

Other independant analysts have not updated yet (see  my previous post for these values). I did a run of SatEvo with the newly released TLE and the current F10.7 cm solar flux, and get, for what it's worth, a projected time of 10:20 UTC (23 Oct) +/- 5 hrs. That seems a bit late compared to the SSC and Zimmer estimates (although the uncertainty windows of course overlap).

ROSAT reentry update

New independent updates on the projected moment of the ROSAT reentry are hampered by the fact that no new orbital elements have been released for 1.5 days now - the last elset released being elset 11294.06213865 (epoch time 21 Oct 01:29:29 UTC) as of this moment (22 Oct 12:10 UTC).

Meanwhile, here is a summary of the latest available predictions at this moment of writing:
Space-Track (SSC): 23 Oct, 01:31 UTC  +/- 14 hrs
Aerospace Corp: 23 Oct, 13:24 UTC +/- 16 hrs
Harro Zimmer: 23 Oct, 05:33 UTC +/- 6 hrs
Ted Molczan (using SatEvo): 23 Oct, 05:00 +/- 10 hrs
T.S. Kelso: 23 Oct, 03:15 UTC (uncertainty not listed)
DLR: 23 Oct, ~3h UTC +/- 9hrs (midtime from window given)

The FAA has realeased a NOTAM warning for the reentry:

!FDC 1/9172 FDC SPECIAL NOTICE .. ..........EFFECTIVE IMMEDIATELY UNTIL 1110252359 UTC. AIRCRAFT ARE ADVISED THAT A POTENTIAL HAZARD MAY OCCUR DUE TO REENTRY OF THE SATELLITE ROSAT INTO THE EARTH'S ATMOSPHERE. THE FEDERAL AVIATION ADMINISTRATION (FAA) IS WORKING WITH THE DEPARTMENT OF DEFENSE (DOD) AND THE NATIONAL AERONAUTICS AND SPACE ADMINISTRATION (NASA) TO ENSURE THAT THE MOST CURRENT RE-ENTRY INFORMATION IS PROVIDED TO OPERATORS AS QUICKLY AS POSSIBLE. FURTHER NOTAMS WILL BE ISSUED IF SPECIFIC INFORMATION BECOMES AVAILABLE INDICATING A UNITED STATES (US) AIRSPACE IMPACT. IN THE INTEREST OF FLIGHT SAFETY, IT IS CRITICAL THAT ALL PILOTS/FLIGHT CREW MEMBERS REPORT ANY OBSERVED FALLING SPACE DEBRIS TO THE APPROPRIATE ATC FACILITY AND INCLUDE POSITION, ALTITUDE, TIME, AND DIRECTION OF DEBRIS OBSERVED. THE DOMESTIC EVENTS NETWORK /DEN/ TELEPHONE 202-493-5107, IS THE FAA COORDINATION FACILITY

Gearing up for ROSAT's re-entry, and an older observation of a Breeze-M tank near M31

Shortly after UARS, another satellite about to reenter is in the news: ROSAT. I last observed and photographed it about a week ago (see here and here) - since then, passes have become unfavourable for the Netherlands.

In an interesting twist, Sky & Telescope's J. Kelly Beatty reports that DLR and ESA sources confirmed to him that they expect the entire telescope mirror array - which weights 1.6 tons! - to survive reentry, impacting intact!

Various modellers now project the reentry to occur on October 23rd. Here is a short list of what various sources currently predict [editted 12:10 UTC, Oct 21, with latest Molczan update):

Space-Track (SSC):  23 Oct, 05:49 UTC (+/- 24 hrs)
Harro Zimmer: 23 Oct, 05:03 UTC (+/- 48 hrs)
Ted Molczan (using SatEvo): 23 Oct, 05:00 UTC (+/- 10 hrs) [editted]
Aerospace Corp.: 23 Oct, 13:24 UTC (+/- 16 hrs)

Since Ted uses the same software I used for my UARS predictions, and hence our results will be similar, I will not put forward my own predictions here but refer to Ted's.


A Breeze-M near M31, the Andromeda nebula

In my post of October 2nd, I featured an image I took on 29 September of a Russian Proton upper stage Breeze-M tank near the trail of USA 129. I wrote that:

These pieces of Russian space debris pop up more often on my images lately. They are the jettisonable torroidal (doughnut-shaped) fuel tanks of a Breeze-M, the upper stage of a Proton M. There are now over 40 of these spent empty tanks in space, often in highly elliptic orbits representative of a geostationary transfer.

Just a few days later, on October 2nd, I took advantage of clear skies to image M31, the Andromeda galaxy. The camera (Canon EOS 450D) with the Samyang 1.4/85mm lens was piggybacked on a Meade ETX-70 in order to use the telescope drive to follow the stars. A long series of 10 second images was taken.

Several satellites showed up on the image series, including a Breeze-M tank again, this time 2006-056B:

click image to enlarge

Here is the final image of M31, a stack of 105 individual 10 second images:

click image to enlarge

Given that this image was taken from a town center with modest equipment, I am quite happy with it! If you compare it to a single frame image (above) it shows the strong improvement in signal-to-noise ratio that comes from stacking images.The two satellite galaxies come out much better, and so does a glimpse of the spiral structure and dust bands in the Andromeda galaxy.

Another ROSAT observation

Yesterday evening was clear, and I again observed the doomed satellite ROSAT (see my previous post), in deep twilight (sun at -6 degrees). It was again bright, magnitude +1, very fast, easy to see even though the sky was still bright blue with only a few stars visible. Like my earlier observation the day before yesterday, it was steady in brightness, with no sign of brightness variations, suggesting it is not tumbling.

Photographically it was a challenge: I had to do some serious image editing tricks to pull the trail out of the bright twilight background on the image below (on the unedited image, the trail is visible but very inconspicuous):

click image to enlarge

These high elevation (near 70 degrees) twilight passes are quickly moving  earlier (and too early) in the evening for me: yesterday's was the last one I could expect to realistically observe. Passes at lower elevation (12-14 degrees) in late twilight will become visible for me after tomorrow and might allow me to observe it for a few more days later this week, until these passes move too early as well.

Using Alan Pickup's SatEvo software and the current 10.7cm solar flux, I get a projected decay at October 23. Harro Zimmer, using another model, gets October 24th. These predictions still have an uncertainty of a few days, so expect them to shift over the coming days, amongst others due to changing solar activity.