Why Space Junk Doesn't Just Fall Down

The short version

Everything in low Earth orbit is falling. That is what an orbit is — a fall that keeps missing the ground. The reason a satellite does not simply come down is that it is moving sideways fast enough to stay ahead of the curve of the planet.

What eventually ends that arrangement is air. Not much of it, but not zero either. Even at several hundred kilometres up there is a thin haze of atmosphere, and every pass through it steals a little energy. The orbit sags, the object drops into slightly thicker air, which slows it more, which drops it further. The end is quick and hot.

So debris does clear itself. The catch is how long that takes, and the answer is almost entirely about how high the debris started.

Air runs out fast

Atmospheric density does not taper off gently with height. It falls away roughly exponentially, which means each step up removes far more of it than the last.

The practical consequence, as a rough guide:

• Below ~400 km — months to a few years. This is why the Space Station needs regular reboosts; left alone it would come down.
• Around 600 km — decades.
• Around 800 km and up — centuries.

Two identical fragments, one placed 400 km higher than the other, are not facing similar fates. They are facing fates separated by a factor of a hundred or more.

Mass and shape matter too. A dense, compact fragment punches through the thin air more easily than a light, flat one with the same weight, so it lasts longer. Solar activity matters as well: over the eleven-year solar cycle the upper atmosphere heats and swells, reaching higher and pulling debris down faster during solar maximum, then subsiding again.

What the catalogue actually shows

This is not a theoretical argument. The public satellite catalogue has been tracking two natural experiments for years, and the contrast between them is stark.

In January 2007, China destroyed one of its own defunct weather satellites, Fengyun-1C, with a ground-launched missile. The intercept happened at roughly 865 km — high in the crowded shell. It produced the largest debris cloud in the history of spaceflight.

In November 2021, Russia destroyed a defunct Soviet-era satellite, Cosmos 1408, in a similar test. That intercept happened at roughly 480 km — low. It also produced a large cloud, and it forced the crews aboard the International Space Station and the Chinese station to shelter in their escape capsules as the fragments swept past.

Nineteen years have passed since the first event and only five since the second. If time were what cleared orbit, the older cloud would be the one that had faded.

It is the other way around. As of a catalogue read on 26 July 2026:

• Fengyun-1C — 3,537 fragments catalogued, 2,325 still on orbit. Roughly a third have come down in nineteen years.
• Cosmos 1408 — 1,808 fragments catalogued, 5 still on orbit. Over 99% have come down in under five.

The younger cloud has essentially vanished. The older one is still largely intact and will be for a very long time. The only meaningful difference between them is a few hundred kilometres of altitude.

Why the high shell is the one to worry about

Put those two facts together — that the busiest orbital shells sit between roughly 700 and 900 km, and that debris there clears on a timescale of centuries — and you have the reason orbital debris is treated as a serious long-term problem rather than a housekeeping annoyance.

Anything that breaks up down low is a temporary hazard. Anything that breaks up high is, for practical purposes, permanent. A collision at 800 km does not just create fragments; it creates fragments that will still be there long after everyone involved has retired, sweeping through a shell that other satellites need to use.

That feedback possibility is what Kessler syndrome describes. In 1978, NASA scientist Donald Kessler pointed out that once enough material is in orbit, collisions between existing objects can generate new fragments faster than the atmosphere removes them — a process that continues on its own momentum even if nothing else is ever launched. It is a description of a runaway loop, not a prediction of a specific date, and the high, slow-clearing shells are where the ingredients sit.

What is actually being done

Cleaning up existing debris is hard in an unglamorous way. The physics is manageable — several missions have demonstrated capturing a target or docking with a dead satellite — but each fragment has to be dealt with individually, there are tens of thousands of tracked ones, and there are vastly more too small to track and too numerous to chase.

There is also a legal knot: under space law, an object continues to belong to the country that launched it. Removing someone else's dead satellite is not simply a matter of deciding to.

So most effort goes into prevention. Modern satellites are increasingly designed to remove themselves — carrying enough propellant to lower their orbit at end of life, or operating low enough that drag finishes the job within a defined period. Rocket upper stages are vented and passivated so leftover fuel and pressurised tanks cannot explode years later, which historically caused a great deal of the debris now in the catalogue.

None of that helps with what is already up high. That part is a waiting game measured in centuries.

Watching it happen

The live map plots the tracked fragments of each of these events as they move, coloured by which cloud they came from. It is worth watching for a minute simply to see the shape: the debris does not sit in a neat ring where the event happened. Each fragment left the breakup with a slightly different velocity, so the cloud spreads along the orbit, and Earth turns underneath it. What starts as a point becomes a band, and eventually a shell wrapped around the whole planet.

That spreading is exactly why a single event nineteen years ago is still something every satellite operator in that altitude range has to think about.