Hacker News new | past | comments | ask | show | jobs | submit login
Kessler Syndrome (wikipedia.org)
57 points by misthop 6 days ago | hide | past | web | favorite | 57 comments

The US government's High Frequency Active Auroral Research Program (HAARP)(1), a favorite subject of conspiracy theorists, uses a high frequency, high-power transmitter to temporarily excite a small part of the ionosphere. One theoretical use for HAARP is to produce an ion outflow that drags oxygen up to 800km and could be used to enhance the drag on orbiting satellites (and presumably debris as well), de-orbiting them faster.(2)

1) https://en.wikipedia.org/wiki/High_Frequency_Active_Auroral_...

2) https://www.nap.edu/read/18620/chapter/6

When I first read about SpaceX's plan to launch up to 42,000 satellites, I actually wondered if we'd end up with a night sky filled with artificial objects. Then I looked at the size of the earth: 197,000,000 square miles. Which is basically one satellite for 4,690 square miles. And that's at ground level. Add in more distance for low earth orbit height, and it becomes even bigger. The cubic space between two of these objects (which I'm not sure where to begin to calculate) must be pretty massive.

The sky is huge. And though it's probable that one thing will hit another up there, it doesn't seem very probable, even if something were to explode into a billion tiny bullet -like projectiles.

The issue isn't that there isn't a lot of space up there (as the name would imply, there's a lot of room). The issue is that nothing up there stays still: everything is orbiting the earth, moving at velocities in excess of 7 km/sec (sometimes well in excess at very high altitudes).

Different space missions have different orbit requirements, so there is a wide range of inclinations, altitudes, eccentricities, and ascending nodes. Take a look: http://stuffin.space/. It's a very dynamic system with a dizzying amount of pieces in it.

The issue isn't so much mega-constellations in and of themselves, since presumably with good operations automation and standardized communications between operators we can thrust and avoid potential collisions as they arise, even with very large constellations. The issue is reliability: applying a small burn to nudge out of the way isn't an issue when one or both of the two spacecraft in a potential collision is operating and has propellant (although accidents still happen [1]), but if they're both dead, that's a recipe for a hypervelocity trainwreck.

That's why it's important to ensure reliable fail-safe de-orbiting of end-of-life spacecraft. The FCC is moving in this direction, but unfortunately the rulemaking is still pending.

[1] https://en.wikipedia.org/wiki/2009_satellite_collision

While visualizations like stuffin.space are fun to look at, I think they add to the perception problem. The GP's comment was that space is huge compared to the number of human-made satellites. This cannot be refuted by pointing to a visualization that at the default zoom level represents small communications satellites to be the size of Rhode Island.

> sometimes well in excess at very high altitudes

High circular orbits move a lot slower than lower circular orbits (eg: geostationary orbits basically don't move at all relative to the ground). It's the highly elliptical orbits that have massive velocities.

FWIW, geostationary orbital velocity is like 3km per second.

I'm not sure that being stationary relative to the ground is very relevant since we are talking about stuff colliding in space. The more stable orbits seem like they might be less prone to issues, though.

Yep, I was referring to highly eccentric orbits like Molniyas... apologies, I should have been more clear!

This is the sort of thinking that made us first pollute our cities, then the seas, and then the atmosphere.

Humans are great at saying, "Look at all this amount of space. Let's dump junk in it!" until it becomes untenable.

By the time its untenable commercial enterprises have already made their profit, and it's up to the public to foot the bill for cleanup.

I think you're vastly underestimating the amount of surface area covered by a satelite:

A quick check of Wikipedia[1] told me that the orbit altitude of Starlink satelites is 550km. The earth has a radius of 6371 km. Based on some trig calculation (which can be done on this[2] little php web applet I found), the visible percentage of the Earth from 550km is 4%! That's a lot. Of course, that's a significant overestimate, since it counts surface area which is just barely technically visible at a nearly 180 degree angle, and assumes a spherical Earth. Let's be generous, and say that only 5% of that area is both reasonably visible from StarLink (I apologize for this crudeness, I'm too lazy to do the math) and above trees/buildings/mountains. So that's .2% of the Earth visible from each satelite at an azimuth which is possibly visible for an average human observer on the ground. That gives an average number of 84 satelites observable, assuming a normal distribution, and likely many more for an average observer in an above-average-density area. Obviously these aren't going to be as visible and bright as the ISS or Iridium flares, but they aren't gonna be sparse.



SpaceX's satellites will be in relatively low orbits, which means that atmospheric drag will pull them out of orbit, like in a decade or two.

Even better, SpaceX's new Starship super rocket will make it vastly cheaper to put mass into orbit. That means it will become possible to put up a bunch of satellites that will collect dead satellites and other space debris and return them to the ground.

For the higher orbital shells they're using, it can easily take 100+ years to deorbit from atmospheric drag alone

I doubt it will be affordable to collect space debris. Space is huge and orbital maneuvers need a lot of fuel.

Orbital debris covers stuff large enough to be monitored by U.S. Combined Space Operations Center’s (CSpOC) Space Surveillance Network (SSN). It also covers stuff too small (from about 5mm to about 30cm) to be monitored by CSpOC but large enough to pose risk to human and robotic space missions. This smaller debris would be monitored by ODPO.


> NASA's main source of data for debris in the size range of approximately 5 mm to 30 cm is the Haystack Ultrawideband Satellite Imaging Radar (HUSIR). The HUSIR radar, operated by the Massachusetts Institute of Technology’s Lincoln Laboratory, has been collecting orbital debris data for the ODPO since 1990 under an agreement with the U.S. Department of Defense. HUSIR statistically samples the debris population by "staring" at selected pointing angles and detecting debris that fly through its field-of-view.

> The data are used to characterize the debris population by size, altitude, and inclination. From these measurements, scientists have concluded that there are approximately 500,000 debris fragments in orbit with sizes down to one centimeter.


> 3. How much orbital debris is currently in Earth orbit?

> More than 23,000 orbital debris larger than 10 cm are known to exist. The estimated population of particles between 1 and 10 cm in diameter is approximately 500,000. The number of particles larger than 1 mm exceeds 100 million.

Orbital debris is already causing damage. Here's some examples from Hubble: https://orbitaldebris.jsc.nasa.gov/measurements/in-situ.html

The photo gallery shows other examples of impact damage as well as examples of debris: https://orbitaldebris.jsc.nasa.gov/photo-gallery/

Even if there is a fair amount of sky (solid angle per satellite?), there isn't that much space in the useful areas e.g. If you want a gazillion geostationary satellites or in molniya orbits then there's much less orbit per satellite.

The initial ideas for geostationary satellites dating back to their invention by Arthur C Clark envisioned basically manned space stations with communication enquipment attached. Only technological advances later on enabled the current ring of large independent satellite that are long lived yet still eventually expandable.

Given that the GEO ring space is indeed finite & should be kept free of debris (even crossing one) at all costs I can imagine eventually reverting to the idea of large communication stations instead of individual satellites.

The stations will likely still not be manned will support automated & possibly even manned servising. It could look likey large truss, kinda similar to the one used on the ISS. Inside it would host power & fuel interconnects. Attached to it would be the communication payloads, solar power modules, engines and fuel tanks. Periodically a resuply/repair craft would visit, replenishing fuel & swapping any failed modules. Depending on your launch costs broken mpdules would either be tossed to a safe graveyard orbit or brought back (possibly to a LEO space station) for refurbishment.

The first steps of this might be already happening with the recently launched Mission Extension Vehicle:


The MEV will grapple a satellite that is low on fuel and will act as it's replacement propulsion unit. In this way it can serve at multiple satellites and ppssibly even serve as a tug for satellites that have failed & are are likely to fail soon, to make sure they reach the designated graveyard orbit.

And other similar & more ambitious missions are already planned and being built.

It has happened already so it’s a pretty realistic event.

I don't understand why you got downvoted. NASA release photographs of damage caused by space debris.

Serious comment: I've thought about this idea a lot since I heard about it, and my mind immediately goes to "can't we deflect these things?"

Assuming that the dangerous orbiting debris is in a few "layers" (like a shell at a certain distance from the earth) which conflict with satellites operating at the same layer, it should only come in at any of 360° angles on a semi-2D plane. Like, it's not going to strike a satellite from much of an angle above or below, because then it would by definition not be in orbit, correct?

Could we not reasonably launch a series of deflective plates or shields in the same plane through the "danger zones" encircling important satellites, or creating a safe zone for passage with shuttles?

Obviously we can't create a ring around the entire globe, nor protect all satellites, but it seems that a few well-placed shields in the worst areas would accumulate a growing "cleanup score" over time.

> Like, it's not going to strike a satellite from much of an angle above or below, because then it would by definition not be in orbit, correct?

It sounds like you're assuming that orbits are typically close to circular. I have no idea how true that is, but I imagine that a Kessler-type chain reaction would tend to make it less true.

1) it’s not true, and it becomes less true over time. Gravitational perturbations change the orbital planes of the debris.

2) It’s hard to get your head around the relative velocities you can get from orbital debris. A bullet travels at around 750 miles per hour. A satellite in low-earth orbit travels at 17,500 miles per hour. Orbital debris could impact a spacecraft at something like a cosine loss of that number, which depending on the exact inclination could be on the order of 10,000 miles per hour. So a factor of 15 faster than a bullet.

Shielding takes mass. Mass is expensive. It costs $5000/pound to launch mass into low earth orbit;

> 1) it’s not true, and it becomes less true over time. Gravitational perturbations change the orbital planes of the debris.

I can't imagine gravitational perturbations in LEO are going to significantly alter eccentricity. I'm trying to find some data that supports this, will report back.

Well, to be clear, when I say "gravitational perturbations" I don't mean GRAVITATIONAL perturbations ;) I mean any terms other than the strict 2-body problem dynamics: gravitational perturbations, third body gravitational effects, solar pressure, atmospheric drag, especially acting on non-spherical bodies, etc.



So, my imagination is lacking, so let me make sure I've got this right: The main gravitational perturbation you're thinking of isn't interaction with other small satellites or Earth's varying density; a bigger effect is going to be because the Moon, which is also moving relative to everything else is also a part of the equation.

Out of curiosity, wouldn't atmospheric drag have a circularizing effect on eccentric orbits? I'd have guessed that, since drag is greatest when the satellite is at its perigee, it's going to have a much larger relative effect on the apogee.

I am making no claim about what effects would be bigger (other than atmospheric drag, which is always the biggest effect in LEO). All I’m saying is that in general, all those nonlinear effects we tend to ignore, when integrated over long periods of time, tend to make orbits change their eccentricity and inclination, and that makes them more of a collision hazard because the relative velocity is higher than it would be if they were in more similar orbits to the thing they collide with.

When something hits a deflective plate, it could ablate part of the plate, thus increasing the amount of debris in that particular orbit.

> Could we not reasonably launch a series of deflective plates or shields in the same plane through the "danger zones" encircling important satellites, or creating a safe zone for passage with shuttles?

No, the problem is, while the probability to collide with any one debris particle is low, there are just a lot of them. So when your plate collides with some, there are still a lot of them.

Shouldn't most of the pieces from a collision in LEO crash into the earth? If you consider a plane tangent to the altitude of the object, most pieces ejected below that plane should reenter and burn up. Any piece ejected outward will be in an orbit that will return to the collision point from underneath and must therefore hit the earth. Only parts ejected nearly in-plane will have lasting orbits. That's not to say it wouldn't be a problem.

That's right, collisions in LEO are not a problem. All objects in LEO are being slowed by atmosphere, the smaller and less dense the object the faster.

If two objects collide in LEO (I really mean close to Earth at this point) the perigee of any resulting fragment must be at the same distance or lower so even if resulting apogee is high the object must spend at least some time close to Earth where it will be captured by atmosphere.

The real problem are objects that collide above LEO where it is possible for collision fragments to have orbit that will never hit appreciable atmosphere and they can orbit for hundreds or thousands of years.

In the context of Kessler syndrome, objects that are doomed to deorbit after a collision can still be a "real problem." They just need to remain in orbit long enough to hit something else to contribute to the domino effect behind Kessler syndrome.

It's effectively an epidemiological transmissibility or chain-reaction question.

If a given debris item strikes, on average, <1 other objects before deorbiting, then the Kessler cloud will (eventually) disperse, absent new objects being inserted.

If a given debris item strikes, on average, 1 other object, the Kessler cloud is self-sustaining (or at least until particle size drops to some minimum).

And if a given debris item strikes, on average >1 other objects, the cloud grows and the collision rate increases with time, again until the material is sufficiently dispersed and/or eventually deorbits.

Launching new satellites sustains this equation.

Accidental or intentional debris-field creation of course greatly exacerbates it.

(What the theshold is for >1 collisions I have no idea, I suspect there's research on this.)

A simple solution to global warming, shade the planet

The number of collisons required to negatively impact satellite operations is vastly lower than that required to attenuate solar flux.

Satellite observations of Earth's ground and meterological conditions are vastly more useful than an equivalent expense of space-based solar shielding.

And if you really want to block solar flux, there are far more efficient options, either atmospheric aerosols or L1 orbiting centrifugally-stabilised nanofilm mirrors.

Though I'm dubious about the practicality of any such, the notion of a reflecterised or even simply an absorptive graphene film might be a potential application of such a material. The challenge would be to get it into position, deploy it, and maintain position and orientation over time, without disrupting the fabric itself. As an application for a very low mass-to-area substance with high opacity or reflectivity (as an absorption medium, it would re-radiate thermal energy in all directions as a blackbody, reflectorisation would tend to halve the necessary area, if I'm thinking this through correctly).

Graphene is increadibly lightweight, at 0.763 mg/m^2, or 0.763 kg/km^2. A shield capable of blocking 1% of solar flux (1% * 2 * pi * (8000 km)^2) would weigh about 3 million kg, or 3,000 tonne, if comprised of a single graphene layer. The actual mass budget would be far larger, but based off this minimum with addition of more structural members (likely: nanofilliments), guidance, station-keeping, and possibly some form of visibility enhancements. The concept isn't immediately and obviously intractable, however.


The total material required would be a minuscule fraction of that require to blanket numerous orbital levels in particles of similar blocking potential.

sounds like scientist really thought about shading the planet

Is this a solvable problem? Can you launch a steel plate of 8000 kg and attach thrusters to it, and drive it into known objects? Occasionally refuel the thruster module. I'm guessing a 1/2" steel will absorb some pretty significant energy. Make it 1" steel in the vicinity of the guidance thruster system. Have a plan to deorbit the steel in sections. It's essentially a filtration system.

Huge blocks aren't actually the best way to destroy projectiles with that much energy, monolithic shields will tend to spall material off their back side and even small pieces will deal massive damage. Wipple shields are more effective because the debris are shattered by the initial layer and as a result the energy is spread out a lot more when it hits the subsequent layers.

The problem is you need to destroy more debris than you create with your shield.


1/2 oz piece of plastic into Al at 15k mph: https://external-preview.redd.it/oi0C_a113_sSzHtSaPvzn6_0Gfg...

Im skeptical this would work for anything bigger than paint flakes and possibly some very small fragments the size of tiny bolts.

The thing is at orbital speeds colisions are VERY energetic events that don't behave likey normal low speed collisions. The thing coliding effectively turn into rapidly expanding plasma before any of the expected plastic deformation, shearing, etc. one would expect on a low speed collision (say a truck in full speed hotting a wall).

When you think about it, this effectively all the fuel and oxydizer that has been used to place the given object into orbit releasing all it's energy at once in a very small area.

As an example let's say your steel plate manged to hit a 700 kg satellite at ~8 km/s, which is perfectly possible with, with the worst case being a 16 km/s combined in a head on collision. How much energy will get released ?

Say the sattelite has been put in orbit using thr Soyuz 2 launcher, which can accelerate up to 7000 kg to tje orbital speed of 8 km/s. Soyuz 2 weighs about 312 tons, more than 90% of which is fuel and oxydizer. It could launch 10 of our 700 kg satellites.

This effectively means a 700 ks sattelite hittimg your 8 ton steel plate is equivalent of a point detonation of ~30 tons of perfect mixture of liquid oxygen and kerosene.

I don't tging a 8 ton steel plate will survive that in any usable shape.

This isn't right at all, a lot of the energy from launch goes into 1) gravity 2) air resistance 3) accelerating the fuels to do 1 and 2. Only the actual mass and velocity of the satellite matter. It's a lot but it is very far from being equal to the whole mass of the rocket.

Indeed, it's definitely not 100% conversion of chemical to potential energy & speed, but given that we are talking hundreds of tons of very energetic chemicals, it's still a lot & IMHO still a nice visualization.

No, it's not even conversion efficiency of burning fuels, rockets expend a lot of their energy accelerating fuel that is later burned as well. That energy is not in the satellite when it's in orbit.

Yes, it's a very solvable problem. You just vaporize all the debris in your way and clear a path.

It's only unsolvable in the current context that a nuclear bomb going off in MEO or GEO is worse than some space trash. If the situation became dire enough that we couldn't get off the planet, then you could just detonate some high yield explosives in orbit.

Detonating a lot of nuclear devices in orbit would devastate large sections of the surface with EMP bursts as gamma rays ionize huge sections of the upper atmosphere.


Shoot. Great minds think alike. One note, you'd probably want to deflect hits into space rather than absorb them.

It would be the world's largest ever game of Pong.

why not deflect them into the atmosphere to burn up?

Yeah, especially in LEO / MEO. In GEO towards space seems like a better bet though.

Though dead satellites in GEO make a lot more news.

There's a great short story that makes use of this in its setup. Lostronaut, published in the New Yorker in 2008: https://www.newyorker.com/magazine/2008/11/17/lostronaut

For lovers of anime and hard sci-fi, Planetes is a very good series exploring the issue: https://www.imdb.com/title/tt0816398/

Is there any way to watch this without buying DVDs or pirating? I saw it posted a few weeks ago and went looking and came up blank other than purchasing physical media

Check out your local library. Most have a selection of DVDs that can be rented at no cost. If your local library doesn't have it, interlibrary loans can get you pretty much anything.

This is how I get most of the TV shows and movies that I watch, now that streaming services are fragmenting into uselessness.

Sunrise Studios, who made the Anime, has a streaming agreement with b-ch.com - But the entire site is in Japanese, and I don't know if they have any regional IP blocks.


There's a book by Neal Stephenson about this, called Seveneves.

Seveneves is a great book (although a little grim for my taste), but it is not about Kessler Syndrome. It's about a scenario where (spoiler alert for the first sentence), the Moon blows up, and humanity has to deal with the consequences.

relevant xkcd - https://xkcd.com/2264/

It's interesting that inevitably we will reach a `critical-mass` of this debris and a chain reaction will result.

The fact that paint-chips travel at sufficient velocity to cause damage, I somehow doubt that even the most wild guesses as to the actual amount of debris is accurate. It must be several orders of magnitude too low.

A lot of theories have been proposed to de-orbit these hazards, personally I think our only hope is MegaMaid.

> personally I think our only hope is MegaMaid

That's just ludicrous.

Applications are open for YC Summer 2020

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact