Wednesday 30 September 2020

A very unusual fireball over NW Europe on 22 September 2020 (that went in and out of the atmosphere again)


The fireball of 22 sept 2020, ~3:53:40 UT. Image (c) Cees Bassa (stack of 2 images)

In the early morning of 22 September 2020, around 3:53:40 UT (5:53:40 local time), a very unusual long duration fireball appeared in the skies of NW Europe. It had a duration of over 20 seconds, and for several Dutch all-sky meteor camera's that captured it, it was a horizon-to-horizon event.

In this blog post, I provide a preliminary analysis of (mostly) Dutch camera records of this fireball. As it turns out, the meteoroid survived its brief passage through the upper atmosphere and came out unscathed at the other end!

The image above (which is a stack of two images, each showing a part of the trail) was captured by the all-sky camera of Cees Bassa in Dwingeloo, the Netherlands. The image below is one of several images captured by the cameras of Klaas Jobse in Oostkapelle. Other Dutch photographic meteor stations that caught it were that of the Bussloo VST (Jaap van 't Leven), Twisk (Marco Verstraaten) and Utrecht (Felix Bettonvil). Oostkapelle delivered both the sectored all-sky image below, and additional widefield images. A wide-angle image taken from Over in the UK by Paul Haworth was also kindly made available for analysis.


Click to enlarge

This was a fireball that entered the Earth's atmosphere under a very shallow, grazing angle: a so-called 'earthgrazer'. Because of the horizon-to-horizon aspect, I immediately suspected that this could be a very rare subcategory of 'earthgrazer': for a few of these have been known to enter the atmosphere, reach a lowest point above earth surface, and then leave the atmosphere at the other side again! 

 In other words, the situation of the schematic below:


The most famous case of this kind is the 1972 Grand Tetons daylight fireball (the first with instrumental records), but there have been a handful more since.

Analysis shows that the fireball from 22 September 2020 indeed belongs to this rare class of objects. The meteoroid approached the earth surface to a minimum distance of 91.7 km and then left the atmosphere again, on an altered orbit.

The Dutch photographic images plus Paul Haworths' image from the UK document some 745 km of ground-projected trajectory. The fireball moved from East-Northeast to West-Southwest, over Germany, the Netherlands, the southern North Sea basin and Britain. 

AOS from the photographic images was at 101 km altitude over Germany, around 53o.26 N 10o.22 E near L√ľneburg just south of Hamburg. LOS was at 105 km altitude around 51o.98 N 0o.60 W, between Luton and Milton Keynes in the UK. 

The point of closest approach (indicated by a cross in the map below) was near 52o.80 N 5o.23 E at 91.7 km altitude above the geoid, over Lake IJssel in the Netherlands, not too far from the Twisk camera station which had it nearly overhead.

Click to enlarge

As this was a horizon-to-horizon event, it is likely that the actual trajectory started a bit more eastwards, and ended a bit more  westwards (although Paul Haworth's image shows that by the time it left view of his camera in the UK, the object was rapidly fading).

The plot below shows the atmospheric altitude of the fireball along its ground track. It reached a lowest point at 91.7 km (where it was moving parallel with the earth surface), and then moved away again, surviving the close encounter:


Click diagram to enlarge

Note that the trajectory was, of course, not as 'curved' as the diagram might suggest: the fireball was moving along a nearly straight path and the 'curve' in the diagram is in reality due to the curved earth surface below it (incidently proving again that Flat Earthers are wrong)!

The Twisk, Oostkapelle and Utrecht camera's had an electronic periodic "shutter" in front of the sensor, providing speed data for this fireball. The fireball entered the atmosphere with an initial speed of 33.6 km/s. It barely slowed down during it's grazing encounter with our atmosphere, leaving it again at a speed of ~30 km/s. It hence was too fast to be captured by the Earth: it moved on in a heliocentric orbit after the encounter.

The object was likely not particularly big. Some first quick calculations suggest something in the 20-40 cm range for the initial pre-atmospheric size (but this will need more study). The object was not very bright (Klaas Jobse, who saw it visually, estimated a brightness of magnitude -5) and it did not penetrate deep into the atmosphere. There will obviously have been some mass ablation, but probably limited: a sizable part of the original mass should have survived and moved along into space again.

The observed radiant of the fireball was near RA 163o.7, DEC +6o.4. It's geocentric radiant was near RA 165o.8, DEC +3o.5. The fireball hence came out of the direction of the sun (the sun was at RA 179o.4, DEC +0o.2 at that moment). 

click map to enlarge

The orbit calculated from the 33.6 km/s initial speed and the geocentric radiant of the fireball using METORB 10, is a short-period cometary orbit of the Jupiter family type (Tisserand 2.8) close to the 13:4 orbital resonance with Jupiter. The descending node of the orbit is close to Mercury, so it could have had close encounters to this planet in the past. Perihelion was at 0.30 AU, aphelion at 4.45 AU with an orbital inclination of 3o.4 and orbital eccentricity of 0.87. The object passed perihelion on August 12.


click to enlarge

These results are preliminary, although probably close to the eventual values. The standard way of reconstructing meteor trajectories (the intersecting planes method) which I used here works fine for regular meteors, but for meteors with these extremely long, very shallow trajectories, the trajectory can get a  non-negligible curvature due to gravity. This effect is small, but I nevertheless want to re-analyze the trajectory the coming month, splitting it up in parts, so that I can account for this curvature. It will be interesting to see what the effect is on the position of perigee (the point of closest approach to earth), and on the radiant position.

Added note:  

Jelle Assink of the Royal Dutch Meteorological Institute (KNMI) reported on Twitter that infrasound from this fireball has been detected.

(a few small edits and additions have been made after this blogpost was originally posted)


I thank Paul Haworth, Cees Bassa, Klaas Jobse, Marco Verstraaten, Jaap van 't Leven and Felix Bettonvil for making their imagery available for analysis.

Sunday 20 September 2020

Observing the mysterious "Object A" (2020-063G), left in orbit by China's 'Spaceplane'


Earlier this month I wrote a post about China's brand new, recently launched and landed 'Reusable Test Spacecraft' (2020-063A), probably a 'Spaceplane' similar to the US  X-37B. It was launched on September 4 from Jiuquan, and landed on September 6 at Lop Nor, after two days on orbit (see a previous post).

As I noted near the end of that post, it left something in orbit: an object of unknown character, which the US Military tracking network now calls 'Object A' (a bit confusing I think, as the COSPAR code is 2020-063G - so I'd called it 'Object G'). It is in a 347 x 331 km orbit.

click diagram to enlarge

This does not appear to be just a piece of debris - e.g. some discarded cover. Radio observers discovered that it sends a signal in the L-band near 2280 MHz, something debris doesn't do. So, this appears to be an interesting object that had or has some function, including a radio data signal downlink. It does not appear to have manoeuvered so far, and if it is tumbling (see below) it isn't likely to do so..

I initially thought that it might be a cubesat, but it appears to be rather large for that. At maximum brightness it reaches magnitude +4, i.e. it is visible to the naked eye. Speculation is that it is either an inspector satellite used to inspect the outside of the Chinese spaceplane before landing: or maybe some jettisoned support module. The ejection from the 'Reusable Test Spacecraft' appears to have taken place some two revolutions before landing, or perhaps even earlier (see brief analysis at the bottom of a previous post).

I filmed the object this morning with the WATEC 902H equipped with a 1.8/50 mm lens - see the movie above. The mysterious object showed slow but marked brightness variations, between magnitude +4 and invisible (= fainter than +7). This confirms reports by radio observers of periodic fading in the signal.

Below is the brightness curve that I extracted from my video, using LiMovie. I was handtracking the object, and halfway lost it for over half a minute when it became too faint for the WATEC 902H (equipped with a 1.8/50 mm lens): hence the half-minute gap in the curve. The other, smaller gaps in the curve are moments that I repositioned the camera. One of these days, I really have to start using a motorized mount tracking on the satellite for this kind of endeavours.

The curve shows two brightness peaks, and two major fading episodes. Peak-to-peak period is about 80 seconds, so if this is due to a tumble, it is a slow tumble.

click diagram to enlarge

When I first picked it up (it had just come out of earth shadow), it initially was very bright and steady (see the movie in top of this post). But then it started to get fainter, untill I momentarily lost it. When I picked it up again, it was becoming brighter again, and after a slow peak, it faded again to invisibility. The fades are faster than the brightening phase and brightest phase.

Thursday 17 September 2020

The structure of Space (2): formation of a geosynchronous ring


A year ago, I published a post on 'The structure of Space'. In that contribution, I discussed structure in orbital element space, identifying spacecraft orbital 'families'. But how about 'structure' in a more traditional spatial sense?

A few days ago, I had some light Twitter-banter with the fantastic Alice Gorman ( aka 'Dr Spacejunk'). She asked

"why is it that space junk has not turned into rings around Earth, just like Saturn or Neptune? Kessler and Cour-Palais (1978) argued it was because the smaller particles were removed by atmospheric drag before enough mass could accumulate"

I pointed out that our Earth by all means does have an artificial ring of space debris and functional payloads: the geosynchronous ring. Which, as we will see, is more of a geosynchronous torus, actually. It is not so dense (yet) with objects as the rings of Saturn or Neptune (but give it time!), but it will be long lasting, outliving us humans.

This sparked a few creative days (I had to get my mind off a few things). I wrote a small .NET application that calculates ECEF coordinates of satellites and fed it with all objects from the CSpOC and our classified catalogues: 20058 tracked objects ranging from operational payloads to small debris particles. Next, I used QGIS to make plots of these ECEF coordinates (I know: I often tend to use software for things other than what they originally were developed for).

The results are in the images below: each image pair shows a polar view at left looking down at the north pole, and a side view at right, looking in the equatorial plane. The plots are for 15 September 2020 at 0h UT.

Here is a zoomed in view, showing the objects in lower orbits (left a polar view, right an equatorial view: thickmarks are in units of 2000 km):

Click image to enlarge

As an aside: notice the small circular area with lower object density at the pole, rimmed by a higher density ring (it is also visible in the wider plot below). This is due to the fact that objects in polar orbits tend to have orbital inclinations  a few degrees higher than 90 degrees: notably so to achieve a sun-synchronous orbit (typical orbital inclinations for such orbits are 97-98 degrees).

In the wider, zoomed out plots below that show the higher objects, you can clearly see the geosynchronous ring at ~35785 km in the polar view at left (thickmarks are in units of 10 000 km). It is made up of geostationary and geosynchronous satellites and debris.These are the satellites that bring you satellite television, satellite telephony, and that bring SIGINT and early warning data to the militaries of various governments. The objects inside the outer ring are objects in MEO (e.g. GPS satellites) and GTO (old rocket stages form launches to GEO and other debris):

Click image to enlarge

If you look at the equatorial view at right, you'll note that the geosynchronous 'ring' is actually more a geosynchronous torus. You see a thin line of actual geostationary objects (mostly operational or untill recently operational payloads) in the Earth's equatorial plane with orbital inclination ~0: and a wider band of geosynchronous objects, that have orbital inclinations between roughly 0 and 15 degrees (both operational payloads, defunct payloads including some in a graveyard orbit, and debris).

The latter torus is situated slightly slanted with respect to the Earth's equatorial plane. The orientation of this slant shows a daily cycle, causing a funny 'wave' like behaviour of these satellites over a full day, when we look at their geographic positions in the equatorial plane, as can be seen in this mesmerizing animation that I created:

click animated map to enlarge

In this animation, the colours represent the object density plotted as a kernel density heatmap: red areas are most dense with objects. The small white dots are the actual geosynchronous satellites (plotted for 15 September 2020). There is a thin line of objects at latitude ~0 that are truely geostationary due to stationkeeping: these hardly move. But the geosynchronous objects with inclinations > 0 show a wave-like pattern of movement over the day!

This movement is a tidal effect, created by solilunar perturbations: gravitational perturbations by the sun and (notably) the moon. These tug on these objects a little, so unless you do frequent stationkeeping manoeuvers that keep the orbital inclination near zero, these objects will see their orbital inclinations start to oscillate, between 0 and 15 degrees over a period of roughly 54 years (53-55 years: it depends on the exact altitude of the satellite). This causes the torus, and the slant. The daily 'wave' (wobble) of this torus is caused by a combination of these tidal effects and the daily rotation of the earth, similar to ocean tides.

I have visualized the discussed ~54-year oscillation by plotting the evolution of the orbital inclination against time of Intelsat 1 (1965-028A), the first commercial geostationary satellite that was launched in 1965. It has just completed a full cycle of this moon-and-sun induced oscillation since it's launch:

click diagram to enlarge

In the absense of active stationkeeping (operational payloads make stationkeeping-manoeuvers roughly each two weeks), there is an oscillation in longitude too, induced by the J2 resonance: due to the uneven mass distribution of the Earth (it isn't a perfect sphere but rather a slightly deformed, bulgy egg), geosynchronous objects without stationkeeping start to oscillate in longitude around one of two "stable" points. These points are at ~75 E and ~105 W longitude: the white crosses in the plot below.

click map to enlarge

This oscillation in longitude about one of the stable points is well visible in 55 years of Intelsat 1 orbital data. Below I have plotted the position of the satellite in longitude from 1965 to 2020. You clearly see it oscillate around one of the equilibrium points (the 105 W point, marked by the dashed line in the diagram), with a periodicity of about 3.1 years:

Click diagram to enlarge

Over time this effect will also tend to concentrate space debris at geosynchronous altitudes around these two points. This effect can be seen in the kernel density heatmap (the coloured band) above the diagram, and in the histogram below (two peaks in the distribution, near the first equilibrium point at 75 E and the second equilibrium point near 255 E = 105 W) although it is to some extend masked by a preference of operational payloads to be at the longitudes of either Asia or the USA, where the biggest commercial markets for satellite tv and satellite telephony are.

click diagram to enlarge

Objects at geosynchronous altitude will not decay in millions to perhaps billions of years to come: so the geostationary ring that formed since 1965 will be here to stay, well after we humans are gone. It will be one of the clearest, longest lasting archaeological signatures of the Anthropocene.

Of course the character of the ring will change. Breakups will fragment the larger objects, decreasing the particle size distribution and increasing the number of objects in the ring even when human launches have stopped. Solar Radiation Pressure will more strongly act on smaller particles, so orbital eccentricities (and presumably also inclinations) will change, causing the ring to get more diffuse in time. I do not know of really long-term simulations (the longest I have been able to find was over a mere 200 years period), so cannot put exact figures on this.

The geosynchronous ring is a remarkable form of planetary change: untill quite recently our planet did not have a ring, but now it has, and it is completely artificial. It formed in a short time. In the animation below, I have broken down the current distribution of objects in the ring (for 15 September 2020) into launch timeframes of 5 years, starting 1960 (i.e. just before the first geostationary launches started) and ending at present:

This shows the gradual, but in terms of geological time nevertheless extremely rapid formation of our planet's artificial ring over the past 55 years. This ring will be a long-lasting, visible human footprint in space, probably outlasting all others (including footprints on the moon, that will be wiped out over time by meteorite impacts).

If you have a telescope or a good camera, you can see this ring of objects every night. Here is a photograph of a small part of it:

Click image to enlarge

Friday 11 September 2020

Important update of my SkyTrack software [UPDATED]


In the context of the probable landing site of the recent Chinese Experimental Spacecraft (see previous post), I discovered a serious bug in my SkyTrack software.

As it turns out, when your chosen observing site has an altitude well over MSL, the sky positions in RA/DEC and AZIMUTH/ELEV are significantly off in version 2.5. I had not noticed this before, as my own observing location is at MSL.

The reason for this error is stupid: the SGP4 DLL needs an altitude in km, while in my code I failed to convert the site altitude from meters to kilometers....hence, a site at 995 meter was treated as if was a site at 995 km...oopsie!

Anyway: I have corrected the error and the new version 2.6 is now for download at my software website.


UPDATE 12 Sep 2020:

A second important bug fix was made, leading to version 2.7 now downloadable.

It corrects an error where, when adding a new site, southern latitudes and western longitudes were incorrectly written to the database...


Starting version 2.6, I also included a few lines of code that should solve another problem and forces the software to recognize the dot as the decimal separator during runtime of the program, no matter what your regional windows setting is. This likewise avoids output errors.

For those unfamiliar with the software, there is an earlier blogpost detailing it.

In short, the software takes as input a set of orbital elements in TLE-format, and allows you to calculate, for a given custom time interval in custom time steps:

- The latitude, longitude and altitude of the satellite;

- An indication whether it is sun-illuminated or not;

- the Right Ascension (RA) and Declination in the sky, for your observing site;

- the Azimuth and Elevation in the sky, for your observing site;

- the Range (in km) to your observing site;

...and optionally also:

- a KML file of the trajectory that you can load into Google Earth;

- the Doppler-corrected radio frequency (for a given central frequency);

- EFG (ECEF) x, y, z coordinates of the satellite;

- format the data as .csv so it can easily be imported into mapping applications like QGIS.

- choice between various output formats and Datums for the Lat/Lon and RA/DEC data

The software also has options to restrict the output to a certain minimum elevation as sen from your observing site, and/or output only when the satellite is sun-illuminated (and hence visible).

Saturday 5 September 2020

China launches a 'Reusable Experimental Spacecraft' - a Space Plane? [UPDATED MULTIPLE TIMES]

Early September 2020, the space tracking community was in nervous anticipation of a rather mysterious Chinese launch. Amidst tight security measures, a Changzeng-2F (CZ-2F) rocket was readied at SLS-1 of Jiuquan's Launch Area 4. Chinese tracking ships were taking up positions near South America and in the Arabian Sea. Two NOTAM's appeared suggesting a launch between 5:20 and 6:00 UT on September 4. Something was afoot! Speculation was, that this was the long anticipated inaugural launch of a robottic Space Plane, a version of the Shenlong, China's answer to the American Air Force's X-37B robottic Space Plane.

Then, on September 4th, the Chinese news agency Xinhua published a very brief news item announcing that a CZ-2F from Jiuquan had launched a 'Reusable Experimental Spacecraft' earlier that day. 

The bulletin was scarce in information but stated that "after a period of in-orbit operation, the spacecraft will return to the scheduled landing site in China. It will test reusable technologies during its flight, providing technological support for the peaceful use of space".

No further details were given on launch time, orbit or character of the spacecraft. The description of the spacecraft is a bit ambiguous. Instead of a space plane, a 'reusable spacecraft' could in theory also be some sort of capsule (e.g. like the SpaceX Dragon): but most analysts think this indeed refers to the long rumoured space plane, China's answer to the US X-37B.

Pre-launch, and based on the positions of the hazard zones from the two NOTAM's, I calculated a launch into an orbital inclination of ~45 degrees, incidentally similar to the orbital inclination of the X-37B OTV 6 mission currently on-orbit. What's more, the launch window given (the NOTAM windows were from 5:23 to 6:05 UT) indicated the possibility of a launch into the orbital plane of OTV 6! The orbital plane of OTV 6 passed over Jiuquan at 6:00 UT - near the end of the launch window.

I published the following expected track for a launch into a 45 degree inclined orbit (which we now know is wrong):

Initial pre-launch trajectory guess. Click map to enlarge

When later that day the first orbital elements by the US military tracking network appeared on the CSpOC portal, it turned out that the orbital inclination was not ~45 degrees, but 50.2 degrees, 5 degrees higher than I anticipated. The reason for the mismatch, is that the rocket apparently did a dog-leg manoeuvre during ascend. This is very clear when we plot the orbital ground track in relation to the launch site and hazard zones from the two NOTAM's: it passes obliquely between them rather than lining up.

Actual orbital track. Click map to enlarge

 A 'dog-leg' manoeuvre is usually done for safety reasons, to avoid overflying a particular area downrange (e.g. a city or a foreign nation), but can also be done to insert the spacecraft into an orbital inclination that otherwise cannot be reached from the launch site. The latter is however not the case here - [editted] the orbital inclination is higher than the launch site latitude (you cannot reach an orbital inclination that is lower than your launch site latitude without a dog-leg, but higher you can.). So the reason must be range safety.

It is clear that the launch occurred well outside the NOTAM time window (why, is not clear). My analysis, based on a proximity analysis using the orbits of the spacecraft, the upper stage of the CZ-2F rocket, and that of four engine covers ejected upon spacecraft separation, indicate spacecraft separation and insertion into orbit around 7:41 UT on September 4th, over the Chinese coast with the orbital plane lining up with Jiuquan (see image below which depicts the orbital position at orbit insertion). The launch itself then should have occured some 8-10 minutes earlier i.e. around 7:30 UT, give or take a few minutes.

Moment of orbital insertion. click to enlarge

The spacecraft was inserted into a 50.2 degree inclined, initially 332 x 348 km orbit. During the hours after launch, the spacecraft made small orbital manoeuvres (see diagram below). At the time of writing (5 September 20:45 UT) it is in a 331 x 347 km orbit.

click diagram to enlarge

The later than initially expected launch time and, through a dog-leg manoeuvre, insertion into a 50.2 degree inclined orbit moved the orbital plane away from that of the X-37B OTV 6, although the two orbital planes are still near. Igor Lissov has pointed out some resemblance to the orbital plane of another US classified payload, USA 276, which has a similar orbital inclination to the Chinese spacecraft (but 50 km higher orbital altitude). The RAAN difference is 8 degrees:

click to enlarge

Based on the current orbits of all three spacecraft, there will be no close approaches of the Chinese spacecraft to either of these classified US payloads over the coming two weeks.

OTV 6 is currently in a 383 x 391, 45.0 degree inclined orbit. The difference in RAAN with respect to the Chinese spacecraft is 13.4 degrees, with a 5.2 degree difference in inclination and about 40-50 km difference in orbital altitude.

USA 276, the mysterious spacecraft that made a close approach to the ISS in May 2017 (see my July 2017 article in The Space Review), is currently in a 397 x 395, 50.0 degree inclined orbit. The difference in RAAN with respect to the Chinese spacecraft is 7.9 degrees, with a 0.2 degree difference in inclination and about 50-60 km difference in orbital altitude.

The Chinese 'reusable' spacecraft was launched from SLS-1, one of two launch platforms at Launch Area 4 of the Jiuquan Space Launch Center. Below is a Copernicus Sentinel 2B image of the launch complex, taken on September 2nd, two days before the launch. The two launch platforms are indicated: the southernmost one is the platform used for this launch.

click image to enlarge

It will be interesting to see where the 'reusable spacecraft' will eventually land. One likely candidate is a military airfield, the Dingxin Test and Training Base, that is located some 75 km southwest of the launch site. I have indicated both the launch site (A) and the potential landing site (B) in the Copernicus Sentinel 2B image below. The second image gives a more detailed look on the airbase.

Click image to enlarge

Click image to enlarge

We have no clue how long the spacecraft will stay in orbit. It will be interesting to see when and where it lands.

The 'reusable spacecraft' has the CSpOC catalogue entry #46389 (COSPAR ID 2020-063A). The CZ-2F upper stage is object #46390 (2020-063B). The four ejected engine covers (with apogees in the 458 to 566 km range), have numbers 46391-46394 (2020-063A to 202-063F).

UPDATE 6 Sept 2020 8:45 UT:

Xinhua reports on Sept 6 that the spacecraft has landed after 2 days on-orbit. Depending on the landing site, landing should have been (based on orbital overpass) either around 1:55 UT at Lop Nor (an alternative landing site suggested), or 6:45 UT at Dingxin Airbase.

UPDATE 2, 9:30 UT:
As the Chinese version of the Xinhua bulletin dates to an hour after the first option (1:55 UT), it seems that the landing was near 1:55 UT near Lop Nur in the Taklamakan desert (HT to Jonathan McDowell).

UPDATE 3, 10:30 UT:
This is the potential landing site, a triangular arrangement of 5 km long landing strips in the Taklamakan Desert. The orbital track of the spacecraft passed some 42.5 km northwest of it around 1:54 UT, more or less parallel to what appears to be the main landing strip:

Click image to enlarge

Click image to enlarge

UPDATE  4, 14:00 UT:
This is an updated diagram of the orbital evolution over the test flight. It seems no large manoeuvers were tried during this flight.

Click diagram to enlarge

UPDATE 5, 16:00 UT:

Jonathan McDowell noted that a new object related to the launch has been catalogued, object 2020-063G, #46395. My analysis suggests it was ejected from the experimental spacecraft near 22:25 UT on the 5th, two revolutions before landing. It likely is a cubesat of some sort. It is in a  332 x 348 km, 50.2 degree inclined orbit. (Update 8 Sept: on Twitter, Bob Christy has suggested that it might be a small inspector satellite, used to inspect the outside of the experimental spacecraft prior to deorbit)