Space debris and traffic management

Orbital debris is the only environmental constraint that could plausibly close a region of space to use. It is worth understanding precisely, because the mechanism is counterintuitive: the danger is not that orbits are crowded, but that the energy of a collision creates far more debris than it destroys.

The population

Total mass in Earth orbit is roughly 13 000 tonnes. Most of it is in a small number of large objects: spent upper stages and defunct satellites, each weighing tonnes. Those large intact objects are simultaneously the least likely to be hit and the most consequential if they are, because each is a reservoir of thousands of future fragments.

The band between 1 and 10 cm is the dangerous one, and its danger is structural rather than statistical. Objects that size are too small to track reliably and too large to shield against. A 1 cm aluminium sphere at 10 km/s carries kinetic energy comparable to a small car at motorway speed, concentrated into a point. No practical shield stops it, and no sensor reliably catalogues it, so it can neither be avoided nor survived.

Why hypervelocity changes everything

Collision velocities in low Earth orbit average about 10 km/s, and a head-on encounter between satellites in opposing orbits reaches 15 km/s. At those speeds the physics is not mechanical.

The impact stress vastly exceeds the strength of any material, so both projectile and a region of target behave as fluids and then as vapour. The projectile does not penetrate in the everyday sense; it and part of the target vaporise and partially ionise on contact, and the expanding cloud does the damage.

Kinetic energy scales as $v^2$, so a 10 km/s impact carries roughly a thousand times the energy per unit mass of a rifle bullet. This is why making the wall thicker is such an inefficient defence, and why the Whipple shield works by deliberately disintegrating the projectile at a standoff distance instead of resisting it.

The Kessler cascade

Donald Kessler and Burton Cour-Palais described the mechanism in 1978. It is a population dynamics argument, not a crowding argument.

The critical property is in the branch condition. A collision converts two objects into thousands, so above a threshold density the population grows from its own collisions faster than atmospheric drag removes it. Past that point, halting all launches would not stop the growth. The cascade is self-sustaining.

This threshold behaviour is why the debris problem is not analogous to ordinary pollution. There is no gradual degradation with proportional cost; there is a regime change. Current modelling suggests parts of the 800 to 1000 km band may already be near or past the threshold, which is precisely the altitude range where drag is too weak to clear fragments within a human lifetime.

Where it came from

Three events produced a large fraction of the tracked population.

Two observations follow.

The altitudes are the problem. Fengyun-1C and Iridium-Cosmos both occurred near 800 km, where orbital lifetimes are measured in centuries. Cosmos 1408 was at 480 km, where drag will clear the fragments in years to decades. The same event at different altitudes has consequences differing by orders of magnitude in duration.

The Iridium-Cosmos collision was accidental and involved a tracked object. Cosmos 2251 was a defunct satellite in the catalogue, and the conjunction was not acted upon. It demonstrated that the tracked population alone is sufficient to produce cascading events, without invoking the untracked fraction at all.

Beyond deliberate and accidental collisions, the largest historical source is explosions of spent upper stages with residual propellant or pressurised tanks. This is entirely preventable by passivation, and requiring it is one of the clearest successes of debris mitigation policy.

Conjunction assessment

Operational collision avoidance is a probabilistic decision problem under uncertainty, not a geometry problem.

Space surveillance networks track catalogued objects and propagate their orbits forward. Where two objects approach within a screening volume, a conjunction data message is issued giving predicted miss distance, time of closest approach, and the covariance of each object’s position.

The decision quantity is the probability of collision, computed by integrating the combined position uncertainty distribution over the volume representing the two objects’ combined size. Typical practice manoeuvres when that probability exceeds $10^{-4}$, the threshold the ISS uses.

There is a genuinely counterintuitive property here worth spelling out. Probability of collision does not rise monotonically as uncertainty grows. With very large uncertainty, the probability density is spread so thinly that the chance of it landing on the tiny target volume is small. Probability peaks at intermediate uncertainty, where the distribution is concentrated enough to matter but not precise enough to rule the collision out.

The practical consequence is perverse: better tracking data can raise the computed probability, because it concentrates the distribution. An operator who sees probability increase after a tracking update has not become more endangered; they have become better informed.

Screening volume

A box around the predicted position, typically hundreds of metres to kilometres, within which a conjunction is reported.

Manoeuvre threshold

Commonly a collision probability above 1e-4. Lower thresholds mean more manoeuvres, each costing propellant and interrupting the mission.

Lead time

Manoeuvres are typically decided a day or more ahead, because a small burn early is far cheaper than a large one late.

The coordination gap

When both objects are active and manoeuvrable, both operators may manoeuvre. Without coordination they can manoeuvre into each other. There is no binding international mechanism requiring coordination.

The ISS has performed more than thirty debris avoidance manoeuvres since 1999, and large constellation operators now perform automated avoidance manoeuvres in the tens of thousands per year, which is itself an indicator of how the traffic environment has changed.

Mitigation

The rules have tightened substantially, and the direction is clear.

The move from 25 years to 5 years is a substantial engineering change, not a paperwork one. It effectively requires active deorbit for anything above about 600 km, which means the propulsion system must still function at end of life, after the design life has already been exceeded. That is a reliability requirement on the least reliable moment of the mission, and it puts a floor under how cheap a satellite can be.

Disposal options in order of cost:

Active debris removal

Mitigation slows the growth. It does not reduce the existing population, and if parts of the environment are already past the cascade threshold, mitigation alone is insufficient. Modelling consistently indicates that removing roughly five large objects per year from the most congested bands is needed to stabilise the population.

No such capability is operational. The technical difficulties are specific:

That last row is often decisive. There is no legal mechanism to clean up an abandoned object, and the same technology that removes a derelict could remove a functioning satellite, which makes any deployed capability dual-use and politically sensitive.

Demonstrations have made real progress: RemoveDEBRIS tested net and harpoon capture, Astroscale’s ELSA-d demonstrated magnetic capture of a prepared target and ADRAS-J performed close approach and inspection of a genuine derelict upper stage, and ClearSpace-1 is developing a capture mission. All are proving technology rather than operating a service, and the economics of a service remain unsolved because the beneficiary is everyone and the payer is nobody.

Traffic management

The deeper problem is institutional. There is no air traffic control for space. Conjunction warnings are issued as advisories, operators decide independently whether to act, and no rules of the road assign right of way.

With a few hundred active satellites this was workable. With tens of thousands, including multiple commercial constellations each performing automated manoeuvres, it is not obviously stable. The specific gaps are: no binding requirement to coordinate manoeuvres, no shared authoritative catalogue with covariances, no assignment of responsibility when two manoeuvrable objects conjunct, and no enforcement of mitigation rules against operators licensed in permissive jurisdictions.

Proposals exist for civil space traffic coordination, and several national efforts are developing capabilities. The problem is fundamentally international, and international regimes take decades, while the population is growing now. See Policy and regulation.

References

• Kessler, D. J. and Cour-Palais, B. G. “Collision frequency of artificial satellites: The creation of a debris belt”, Journal of Geophysical Research, 83:2637, 1978.
• IADC-02-01, IADC Space Debris Mitigation Guidelines, revision 3, 2021.
• ESA Space Debris Office, Annual Space Environment Report, published yearly. The primary public source for population statistics.
• Liou, J.-C. and Johnson, N. L. “Risks in Space from Orbiting Debris”, Science, 311:340, 2006.
• Liou, J.-C. “An active debris removal parametric study for LEO environment remediation”, Advances in Space Research, 47:1865, 2011.
• Alfano, S. “Review of Conjunction Probability Methods for Short-term Encounters”, AAS 07-148.