“Space is big,” the British humourist Douglas Adams once observed litotically. “I mean, you may think it's a long way down the road to the chemist, but that's just peanuts to space.” And yet, in all this vastness, there are now some regions so crowded that it is possible for a pair of satellites no bigger than compact cars to collide by purest accident, with no malice aforethought, as happened 800 kilometres above Siberia on 10 February (see page 940).

The fact that humans have managed to spread traffic hazards beyond the confines of their planet is humbling. It is also a serious problem — one that could severely hamper scientific research, weather forecasting, commerce and the national defence of various nations. It needs to be sorted out.

There are two complementary ways of doing this. One is for all satellite operators to abide by debris-minimizing rules such as those promulgated by the UN Committee on the Peaceful Uses of Outer Space: depressurize fuel tanks when you are finished with them, take steps to make sure that batteries don't explode, shut down flywheels after the mission is over and, most crucially, drive low Earth orbit satellites to a fiery atmospheric death when they have fulfilled their tasks. The fact that there is no clear way to enforce such practice does not mean that the international community should not try to insist upon it.

The other response is better tracking, which would allow satellites at risk to manoeuvre out of the way of each other. The military establishments in both the United States and Russia track objects in orbit. The Americans make some of their data available to the world at large; the Russians, to their shame, do no such thing. But the shared US data, although better than nothing, are more crude than those that the military keeps to itself. Better data would allow better decision-making by satellite operators weighing evasive action.

The US military has various reasons for not providing the very best data to all who ask for them. To do so would reveal the capabilities of the US surveillance systems — and perhaps their blind spots — in uncomfortable detail. It would also make the targeting of anti-satellite weapons easier.

One solution would be to release the data to a trusted intermediary with the analytical power to look for potential collisions and alert operators when things look bad. Another would be for the US military to do something along these lines itself. It already provides such services for some high-value missions by NASA and some allies. Expanding the service would make sense. If the US national security apparatus were to reduce the number of future collisions by letting third parties know of the risks, it would be improving the survival chances of its own spacecraft as well as everyone else's — and no one has more valuable assets in low Earth orbit than America's soldiers and spooks.

In the long run, an independent tracking system with its own data sources would be the ideal solution, and to its credit Europe has made some moves towards developing such a thing. But on the world stage this does not seem to be a priority. The problem is that by the time it becomes one — maybe two or three collisions down the line — the threat may have been ratcheted up far enough to be considerably less tractable. Every time two objects in orbit collide, they create more debris that can lead to more collisions. The way things are going, this will be one of those problems where the need for action becomes truly obvious only after it is too late.