Orbital debris sounds like a slow accumulation of litter. It is not — it arrives in bursts. Two deliberate missile tests and one accidental collision account for 24.6% of everything currently tracked as debris, and those are the clouds plotted on the globe above. Each has a date, a cause, and a very different fate.
12,522pieces of debris on orbit right now · plus 2,416 spent rocket stages
Cosmos 1408 was destroyed at 480 km and is 99.7% gone. Fengyun-1C was destroyed at 865 km — 15 years earlier — and is still only 34.3% gone. The lower cloud is younger and has all but vanished; the higher one is older and largely intact. Time is not what cleans orbit. Air is, and there is almost none of it above about 600 km.
China destroyed its own defunct Fengyun-1C weather satellite with a ground-launched missile, creating the largest debris cloud in the history of spaceflight.
A dead Russian military communications satellite ran into the working Iridium 33 over Siberia — the first accidental hypervelocity collision between two intact satellites.
Objects the tracking networks can see and follow but have not tied to a known launch. Most are fragments whose parent was never established.
Counted in the catalogue but not attributed to a launch, so no before-and-after figures exist for it. Shown on the map, left out of the comparison above.
The other half of the 2009 collision: a working American communications satellite, shattered without warning while in service.
Russia destroyed a defunct Soviet-era signals-intelligence satellite in a missile test, forcing the crews of the ISS and the Chinese station into their escape capsules. Because it happened low, air drag has almost finished cleaning it up.
The four clouds above are the ones the public catalogue publishes as named groups, which is why they are the ones plotted. They are not the whole problem. The remaining debris traces back to 843 separate launches, and the two single biggest sources are not missile tests at all — they are the same rocket, twice:
Broke up after delivering its payload, leaving 623 tracked fragments still on orbit — more than the Iridium 33 cloud and the Cosmos 1408 cloud put together.
Broke up after delivering its payload, leaving 579 tracked fragments still on orbit — more than the Iridium 33 cloud and the Cosmos 1408 cloud put together.
Both are reachable from the same free catalogue and are not drawn here yet. Saying so is the point: a map that showed four clouds and implied that was all of it would be accurate about its dots and wrong about the sky.
The figures on this page come from two different counts, and they are not interchangeable. The four named clouds above are counted from CelesTrak’s own curated debris group for this event — the fragments CelesTrak attributes to it by name, which is the same list this map propagates. The two launch-designator sources below are counted by launch designator — every catalogued fragment sharing this launch’s ID, which is how the catalogue groups breakups that have no curated group of their own.
The two methods disagree slightly — by one to three fragments per cloud — because a single launch can carry several satellites, and the parent CelesTrak attributes a fragment to is not always the launch it rode. Neither count is wrong; they are answers to slightly different questions. Every figure below is labelled with the method that produced it.
Per-cloud figures derived from CelesTrak SATCAT — the public satellite catalogue — read July 26, 2026. Live positions come from CelesTrak GP element sets, propagated in your browser. “On orbit” means no decay date on file; “burned up” means a decay date has been recorded. Counts shift as fragments reenter, so these are a snapshot of one day, not a running total. Live catalogue read July 26, 2026. 3,076 fragments across the four dated clouds.