Why are Starlink "Racetrack" Flares [Mostly] Reported from Planes?

Mick West

Administrator
Staff member
I've been thinking about this and why the particular Starlink satellite is flaring and how we can further demonstrate why it is flaring at that moment in time. Simply - they are flaring by reflecting the light from the the sun using the satellite chassis as a mirror, not their solar array. The chassis always points towards directly down towards the earth as it orbits in order to broadcast its radio signal (internet) to the closest subscribers

Source: https://phys.org/news/2020-05-spacex-theyre-starlink-satellites-visible.html
1665433732947.png


Bearing in mind reflection theory, the suns's position which should therefore be directly on the other side of the earth of the satellite as the axis. (This was initially proposed by Reddit user danse-macabre-haunt). I have checked this on In-The-Sky.org using the example above and in the stitched image below which lets us see through the ground it confirms that the sun is almost directly below the horizon in line with the flaring satellites.

1665433108946.png
https://in-the-sky.org/satmap_planetarium.php?latitude=+35.179267&longitude=-93.969778&timezone=+00:00
(you'll need to set the date & time & view on that link if you wanna see an accurate prediction)

The image here shows the "On Station" satellite with its chassis parallel to the surface of the Earth, visible because it is scattering light. 2022-10-25_10-53-13.jpg
However to get a flare or glint, the light has to reflect off a surface (not simply scatter). Something I did not immediately recognize is that this means such glints will generally not be visible from the ground.

Consider this simplified representation:
2022-10-25_11-06-11.jpg


The green disk is the Earth (viewed from the side, with the North Pole at the top).
The Sun's rays come in essentially parallel from the right.
The Satellite orbits with the chassis parallel to the surface to the surface of the Earth.
So, if the Earth were perfectly spherical, and all the angles are perfect, then any ray from the sun reflecting off the chassis would miss the Earth.
However for a viewpoint above the surface, there's specific latitudes where you can get a direct reflection off the bottom the chassis, but only at certain latitudes.
These latitudes will change with the seasons.


Interactive:
https://www.geogebra.org/calculator/fjpzsmws

It gets a little more complicated with three dimensions. Things to consider:
  • The Earth is an oblate spheroid - i.e. it's flatter at the poles.
  • The Earth's surface is not smooth, it has mountains.
  • The Earth's atmosphere curves light down and around the horizon
  • The above three all gives a bit more potential for visibility from the ground.
  • The sun also needs to be roughly in line with the satellite
This all suggests that for a plane at a given altitude there's a fairly narrow window where they will see the flares.

It also suggests that the ideal view latitude is 45* plus the tilt of the Earth's axis towards the sun. This is complicated though by the elliptical (oblate spheroid) nature of the Earth. The window will be wider with the altitude of the plane.
 

Attachments

  • 2022-10-25_11-12-29.mp4
    1.2 MB
There was a thread earlier this year on sightings from south or Central America; I can't recall just where, so I'm having trouble finding it. Does that fall into the zone of places where starlink is visible from the ground?
 
What I can't exactly figure out is why there would be a concentration of reports at the North. Are the reflections just more noticeable there? Are they brighter or longer? Just where the region happens to intersect the Pacific tracks?

Consider all the Starlink satellites. At any given point their chassis are parallel to the ground, so we can imagine instead a fixed sphere of mirrors encircling the globe.

2022-10-25_13-45-38.jpg

So, opposite the sun, there should be a ring of locations where the glints are visible, and in every direction, I think.

Hmm.....
 
It also suggests that the ideal view latitude is 45* plus the tilt of the Earth's axis towards the sun. This is complicated though by the elliptical (oblate spheroid) nature of the Earth. The window will be wider with the altitude of the plane.
Very interesting model. I've been trying to recreate this myself, hadn't thought of using GeoGebra. My only comment re this is, are you assuming that the sun's ray are directly over the North pole? Most of the sightings have been just off North towards Northwest. Not sure what effect that would have, I have a feeling it may amplify the effect of the oblate bulge around the equator.
 
are you assuming that the sun's ray are directly over the North pole? Most of the sightings have been just off North towards Northwest. Not sure what effect that would have, I have a feeling it may amplify the effect of the oblate bulge around the equator.
I wasn't really assuming anything beyond the 2D sim. It's a little confusing in 3D. I don't really have the time to build a proper model. So many factors to consider - including the actual orbits. Check this out:
https://heavens-above.com/Starlink.aspx

Click on one of the satellites to see the orbit line.



Looking north, the densest areas are the part of the great circle closest to the poles (ignoring the tracks that go over the poles). They are also going left to right



So maybe, like, here:

2022-10-25_17-04-27.jpg
 
I wasn't really assuming anything beyond the 2D sim. It's a little confusing in 3D. I don't really have the time to build a proper model.
Yes, the 3d trig starts to get a bit confusing and complicated. I had started a scale 2d drawing in Visio but didn't have the time to complete it either. I like what you've done in GeoGebra - it is visually simple but the dynamic nature of the model is great. my visio sketch tried to use the lat long and altitude for the aircraft and satellite - (I converted this into a straight line polar distance) although the aircraft altitude is comparatively zero when compared to the satellite altitude (550km Vs 35000ft).

I'd like to find out more about the orientation of the Starlink satellites. I had initially theorised that the chassis was reflecting the sun's light, but maybe it is the solar array.
 
I have found numerous references to SpaceX adjusting the orientation to try and reduce brightness.

https://newatlas.com/space/spacex-sun-visor-darken-starlink-satellites/

This week, SpaceX announced a new set of measures that it's taking to mitigate the light pollution risk posed by its swarm.

During orbit raising, and while the satellites are in their parking orbit prior to reaching their operational orbits, the solar array is the main source of light pollution. This is because the satellite is in what SpaceX calls an "open book" configuration, wherein the array is laid out flat in front of the spacecraft relative to Earth’s surface in order to reduce atmospheric drag. During this phase, sunlight reflects off both the body of the spacecraft and its solar array, making it visible from Earth.

SpaceX is currently exploring the potential of rolling the satellites as they make their way to their operational orbit, in order to keep the solar array edge-on to the Sun. This would prevent light glinting off of the panel like an enormous mirror, but the technique has its own set of issues that would prevent it from being employed all the time.

It's possible that certain orientation changes make it more likely they will momentarily flare more brightly from some observer positions, and possible they orientate towards/away from the sun to try and provide power / prevent overheating depending on orbit etc. But I am unsure about how they orientate, reaction wheel? This thread seems to agree.

https://forum.nasaspaceflight.com/index.php?topic=56119.0
 
I had initially theorised that the chassis was reflecting the sun's light, but maybe it is the solar array.
Yes, I'm also lean towards that. As the earth-aligned chassis would not seem to be just limited to being "in-line" with the sun just in the North.
 
I'd like to find out more about the orientation of the Starlink satellites. I had initially theorised that the chassis was reflecting the sun's light, but maybe it is the solar array.
They're in line, there's no difference.

From the other thread:
The SpaceX update "APRIL 28, 2020 ASTRONOMY DISCUSSION WITH NATIONAL ACADEMY OF SCIENCES" seems to have moved to https://www.spacex.com/updates/ where it lacks the images that originally went with it. Thankfully, it has been republished by various space websites.
Article:
There will be a small percentage of instances when the satellites cannot roll all the way to true knife edge to the Sun due to one of the aforementioned constraints. This could result in the occasional set of Starlink satellites in the orbit raise of flight that are temporarily visible for one part of an orbit.

This simple diagram highlights why satellites in orbit raise are so much brighter than the satellites that are on-station.
spacex_starlink_visibility_945.jpg
orientationroll.pngorientationroll2.png
It may be worth reading the copy in full, it's SpaceX content.

Basically, if we're seeing a "train", the satellites are not on station yet.
If they're not on station, the "hinge" is flat ("open-book"), and the chassis is aligned with the solar array (and the visors are not in place).
The hinge is angled ("shark-fin") when the satellite is on station, because the antenna needs to point down and the solar array needs to point sunward.
sharkfinopenbook.png

As remarked in other quotes above (as well as the SpaceX update), sometimes the satellites are not quite knife-edge; and of course we don't know how reflective that knife edge still is.
 
in that case, the solar array shouldn't be flaring since it'd be aligned towards the sun.
Yes, this was my reasoning for suggesting it was the chassis that was reflecting. And that it was more or less parallel with the surface of the earth. But the critical thing is by how much 'more or less'. This little variation could have a big effect over the thousands of miles between satellite and observer. I'm still open to other ideas about this.
 
Thinking about why "racetrack" ufo reports might be expected to be reported from planes... am I correct that "racetrack" is a term used in aviation to describe a particular holding pattern? Based on Googling, I think so, but pilots or airplane folks correct me if I am wrong.

If so, it is reasonable that pilots and flight crews alike be primed to see these repeatingy tracking lights as such a pattern being flown by... something, and to use the term to describe it, where us groundlings would see just a repeating sequence of lights fading in and out.

And of course once it attracts some atttention through one pilot reporting it, others are primed to keep an eye open for it.
 
Thinking about why "racetrack" ufo reports might be expected to be reported from planes... am I correct that "racetrack" is a term used in aviation to describe a particular holding pattern? Based on Googling, I think so, but pilots or airplane folks correct me if I am wrong.

If so, it is reasonable that pilots and flight crews alike be primed to see these repeatingy tracking lights as such a pattern being flown by... something, and to use the term to describe it, where us groundlings would see just a repeating sequence of lights fading in and out.

And of course once it attracts some atttention through one pilot reporting it, others are primed to keep an eye open for it.
All correct. But this thread is more about why the pilots are seeing and reporting them. i.e., why are they particularly visible from planes.

If they were also as visible from the ground, then I think we'd be getting a lot more video, and better video (as it lasts for quite a while.)
 
I found the attached paper describing the reflectivity of the Starlink and it seems to suggest that they will be particularly visible at low elevation levels, even with the sun visor attached...

https://arxiv.org › astro-ph
[2107.06026] A Sky Brightness Model for the Starlink 'Visorsat' Spacecraft

https://arxiv.org/abs/2107.06026

Screenshot_20221026_192530.jpg
 

Attachments

  • A_Sky_Brightness_Model_for_the_Starlink_Visorsat.pdf
    5.4 MB · Views: 131
This is Hulsey's video, overlaid on Stellarium, showing the position relative to the sun. Nothing really new, but clear the flaring happening "above" the sun is a big clue as to what is going on.
 
All correct. But this thread is more about why the pilots are seeing and reporting them. i.e., why are they particularly visible from planes.
Sure, but I am wondering if that assumption is valid -- ARE they more likely to be visible from a plane? Or is it just a case of they are more likely to be interpreted in this way by flight crews, and/or more likely to be reported now, with the word "racetrack" attached?

Not sure how to test for that... but putting it out there because lots of folks here are quite good at figuring out ways to rule things out, or in.
 
https://arxiv.org/abs/2107.06026

2022-10-26_12-50-27.jpg

This really seems like the prime contender. It really seems to fit in terms of where the plane is and where the sun is. My main puzzle is why it's (seemingly) always in the North. The Sun grazes the Earth all around the terminator, and there are satellites everywhere, so this type of reflection should be everywhere.

Perhaps the simple answer is that it is everywhere, but just a lot more common and sustained in the north because of

A) the clustering of orbits moving parallel to the terminator in the north
B) The terminator isn't moving very fast. i.e you stay the "sweet spot" distance to it for longer.
C) Planes going West fly WITH the terminator, so they stay in a sweet spot for longer

Which all suggest there would be other Starlink chassis flares that are both less frequent and shorter in duration in other areas. Occasional short flashes, not in any particular grouping or direction, that the pilot would not pay much attention to.
 
Perhaps they only happened in the North in this recent 'flap' in August and September because that was Northern hemispheric summer. Maybe in the coming months we'll see the same in the southern hemisphere?

(This is of course speculation)
 
There is, of course, far less air traffic in the southern hemisphere, so maybe not.
Could it be as simple as most of the worlds humans are in the northern hemisphere? 88% of humanity lives in the Northern Hemisphere.

That site also has brilliant map which shows that the half of earth facing from roughly the same position of Micks map in #4, contains 93% of all humans.

1E283220-443E-44DE-B1CD-EAF0694641AC.pnghttp://www.radicalcartography.net/human-hemisphere.png

It almost has to be people-related as the satellites spend half their time over each hemisphere.
 
It's not that it's in the Northern hemisphere, it's that it's in the North, as in they have to look north to see them.
but wouldn't it work by looking South in the southern hemisphere? the point is to look towards the set sun, isn't it?
 
My main puzzle is why it's (seemingly) always in the North. The Sun grazes the Earth all around the terminator, and there are satellites everywhere, so this type of reflection should be everywhere.

Perhaps the simple answer is that it is everywhere, but just a lot more common and sustained in the north because of

A) the clustering of orbits moving parallel to the terminator in the north
B) The terminator isn't moving very fast. i.e you stay the "sweet spot" distance to it for longer.

C) Planes going West fly WITH the terminator, so they stay in a sweet spot for longer

Which all suggest there would be other Starlink chassis flares that are both less frequent and shorter in duration in other areas. Occasional short flashes, not in any particular grouping or direction, that the pilot would not pay much attention to.
A) and B) are quite interesting.
The two videos I took a closer look at just now would support this hypothesis.

I figured MUFON case 124374 (https://www.metabunk.org/threads/mu...ing-starlink-flares-racetrack-illusion.12586/) might lend itself to this since we have exact positions and times for the observations.

I picked videos file1 and file2, took the geolocated positions from that thread and narrowed down the times using the flightpath KML from ADS-B Exchange.

Code:
file1 35.15N 77.93W 12-Aug-22 05:07:24Z
file2 35.63N 77.90W 12-Aug-22 05:11:45Z

file1-1.jpg
file1-2.jpg

This is the situation in file1. One of the flashers is circled in red in the overlay. If that overlay is correct then I think I got the individual sats sorted. I believe that the flashes were most likely caused by the four Starlink sats that moved past that area at that time, Starlinks 3917, 2736, 3208 and 1259.

As it turns out they were all clustered around the northernmost point of their orbits moving parallel with the terminator.

file1-3.jpg
file1-4.jpg

Same with file2. 3919 and 2155 were in a similar spot with similar lighting conditions.

file2-1.jpg
file2-2.jpg
file2-3.jpg
file2-4.jpg

That's a distance of 2,000 km by the way. Five degrees west off true north.

ge.jpg

By no means definite proof but intriguing all the same.
 
Hey @Easy Muffin - nice to see someone else using in-the-sky.org, I love that website. Just a two quick points why I used it in this (and other) invetigations as opposed to other websites such as Heavens-Above.com . Firstly, the Graphical User Interface great and allows the user to see the whole sky in a number of ways as observed by the witness. Secondly, and most importantly, in-the-sky.org uses the legacy satellite data (known as Two Line Elements, or TLEs) that was valid at the time set in the browser. Heavens-Above does allow you to look at historical passes but they are calculated using today's TLEs, which means they will be less accurate. But I still prefer Heavens Above for satellite predictions.

Very interesting comments you made about the clustering at the nothern most point of their orbit. It means that teh flaring satellites don't even need to be part of the same launch batch of satellites, which we see by the non sequential STARLINK-XXXX names in the screen shots.
 
Hey @Easy Muffin - nice to see someone else using in-the-sky.org, I love that website. Just a two quick points why I used it in this (and other) invetigations as opposed to other websites such as Heavens-Above.com . Firstly, the Graphical User Interface great and allows the user to see the whole sky in a number of ways as observed by the witness. Secondly, and most importantly, in-the-sky.org uses the legacy satellite data (known as Two Line Elements, or TLEs) that was valid at the time set in the browser. Heavens-Above does allow you to look at historical passes but they are calculated using today's TLEs, which means they will be less accurate. But I still prefer Heavens Above for satellite predictions.
Yeah, it's a really nice site for this purpose. Actually the only one I know that makes historic TLEs available this easily. Ideally there'd be an option to download them but beggars can't be choosers, right? I mean it still gets the job done, it's only a tad more clunky in the browser than if I did with the apps I'm used to.
Very interesting comments you made about the clustering at the nothern most point of their orbit. It means that teh flaring satellites don't even need to be part of the same launch batch of satellites, which we see by the non sequential STARLINK-XXXX names in the screen shots.
Correct, there doesn't seem to be a connection to the individual launches once they're on station. The four sats from the first video for example were sent up on four different launches over a span of more than two years.

1259 - March 18, 2020 - Mission L5
2736 - May 26, 2021 - Mission L28
3208 - Dec 2, 2021 - Launch Group 4-3
3917 - May 13, 2022 - Launch Group 4-13
 
Actually the only one I know that makes historic TLEs available this easily. Ideally there'd be an option to download them but beggars can't be choosers, right?
Is there a public source for these somewhere? I could munge one of those site, but don't they get them from JPL or somewhere?

Here's what Heavens Above data looks like:


2022-10-29_09-35-49.jpg
 

Attachments

  • www.space-track Starlink 14 15 Aug 22.txt
    1.1 MB · Views: 110
So it seems like two conditions need to met for these to become visible from a plane or at least show up in NVGs from the surface.

1 - The sun needs to be almost exactly in line with the satellite. (This one I suppose is easy enough to comprehend.)
2 - The sun needs to be roughly 40 degrees below the horizon. (I'm a little stumped on this one but every single sighting I looked at this was the case. Something to do with the geometry of sun, reflective satellite surface and observer?)

Now re: why are they always in the north, I think it really might be as simple as that's the direction where most Starlink sats will travel across the area defined by 1) and 2).

See for example this animation of the Palmdale case. Note that these are not the historical TLEs so the individual sats are not necessarily in their correct places but as a swarm they do move across the sky in a similar manner. I marked the visibility region with a small circle. Due north is where orbital mechanics come together in such a way that when viewed from Palmdale, they move parallel with the horizon and like road traffic, they move through the circle one by one as they reach their respective orbit's northernmost point. As discussed in this thread earlier, the same mechanics also lead to a general clustering of sats around the northernmost point of their orbits above 53 degrees northern latitude. Lots of opportunities then to observe at least a couple of them.



For comparison, the same lighting conditions as they will happen tonight. Now the view is towards the west and over there the sats are moving all over the place. There's less of them around to begin with, and whilst some of them pass the circle most will stay invisible. Probably a much less conspicuous sight.



An unrelated one, this is tonight's sky from my place. I'm at a much higher latitude, about 50.5 N - that's just south of Calgary for North Americans. Accordingly, when the sats cluster together around their northernmost point they appear far higher in my sky and any light they reflect will go right over my head and hit people further south from me. I could therefore only hope to catch the odd stray sat that just so happens to be in the right spot.

Funnily enough there's a whole gaggle of them streaming upwards in what I can only assume would be a pretty spectacular sight. Unfortunately they are a tiny bit too far left so I reckon I'll never see them!
Although it might be an idea to keep an eye out for when one of these streamers is predicted to be in the right spot. I haven't got night vision equipment but If the weather's playing along I might pack my camera, head out and try to get them in a long exposure. I guess it would be interesting to see at what elevation they flare out.

And again, this is all assuming this strange sun at -40 thing holds true.
 
they flare out.

And again, this is all assuming this strange sun at -40 thing holds true.

I know what you're gettign at here, but I'm not sure that "The sun needs to be roughly 40 degrees below the horizon." This needs to be refined somewhat to

My very simplified model of the flaring satellite situation is as follows. It shows that the critical dimension is the 550km orbital height of the satelltie, added to the radius of the earth. This gives a incedent angle of about 67 degrees on the satellite, assuming it is parallel to the surface of the earth. Obviously the changign position of the sub and observer will make this a very dyamic model.

(Not to scale, & assumes the Sun is at infinty)
1667740496958.png

1667740560246.png
 
Aye it's not a very well presented theory yet :p
I was just wondering why so far for every single observation we find the sun at that elevation with the sats close to the horizon. Is the cone of the flare so narrow that there is no other possible constellation?
It'd be quite handy if it were, because we could then determine regions with favourable viewing conditions. With the sats 550 km up and appearing 5 degrees above the horizon they'll be roughly 2000 km away, and since the best viewing direction seems to be towards the north (or the south in the southern hemisphere), that region would be 2000 km south (or north) of the latitude at which the sats pass into orbital sunset/sunrise and at whatever longitude corresponds to a sun elevation of -40 degrees at that time. And if the time of year is such that this latitude coincides with the satellites' most extreme orbital latitude (53 N or S for the Group 1 Starlinks) then you're likely to see loads of them.
I think that's how this could work, anyway.
 
My very simplified model of the flaring satellite situation is as follows. It shows that the critical dimension is the 550km orbital height of the satelltie, added to the radius of the earth. This gives a incedent angle of about 67 degrees on the satellite, assuming it is parallel to the surface of the earth. Obviously the changign position of the sub and observer will make this a very dyamic model.

(Not to scale, & assumes the Sun is at infinty)
1667740496958.png

1667740560246.png
If your graph was accurate, the flare ought to be reddish, from the sunlight passing through a lot of atmosphere.

Does the angle at the center of the Earth correspond to the actual positions of observer and identified satellites?

What are you basing your 90⁰ sighting angle on? (perfect observer line vs. radius)

I think it is possible that the normal vector of whatever surface is reflecting the sun on the satellite may not be aligned with the center of the Earth, i.e. it might be tilted slightly north or south. Can we determine this from the data?
 
If your graph was accurate, the flare ought to be reddish, from the sunlight passing through a lot of atmosphere.
Yeah, that's a fair point. But not really relevant to the geometric points in my head.
Does the angle at the center of the Earth correspond to the actual positions of observer and identified satellites?
Yes it does, however this is a 2d model that is taken from a slice of a 3d Globe. The actual angle is a function of both the lat and long of the observer and directly below (nadir) of the satellite. Haven't quite worked out the formula for that yet.
What are you basing your 90⁰ sighting angle on? (perfect observer line vs. radius).
The 90° assumes that the observed flare is at the horizon, and therefore the vector to it will be tangential to the surface of the earth at the observers point, which is the same as 90° to the radius of the earth.

Edit - Mick's Geogebra model is pretty much the same as this, but assumes an airborne observer:

https://www.geogebra.org/calculator/fjpzsmws
 
Last edited:
Back
Top