How does lightning get started?

Balloons flown into thunderstorms carry field meters, and for fifty years they have brought back the same awkward number: the strongest electric fields inside a storm are roughly ten times weaker than what it takes to break down air at that altitude. By the textbook physics of sparks, lightning should not start. It starts about a hundred times a second, somewhere on Earth.

Why it matters

Lightning kills thousands of people a year, starts a large share of the world’s wildfires, and batters aircraft, wind turbines, and launch pads. Forecasters can say a storm is likely to flash; nobody can say from first principles when or where a given cloud will spark first, because the first millisecond is the part physics has not settled. It is also a matter of pride. This is the most familiar violent event in nature, studied since Franklin flew a kite at it, and its opening act is still unexplained.

What has been tried

The charging is understood: collisions between soft hail and ice crystals in supercooled cloud water transfer charge, and updrafts sort the debris into vast charged layers. The trouble is the spark. C. T. R. Wilson suggested in 1925 that electrons accelerated to near light speed could ionize air in fields too weak for ordinary breakdown, and a 1990s version of that idea has cosmic rays seeding avalanches of runaway electrons. A rival picture needs no help from space: the tips of ice crystals concentrate the local field enough to launch filaments of ionization called streamers. Radio telescope arrays have now imaged flashes being born, metre by metre, and what they see looks like growing streamer activity, with no cosmic ray arriving first. Yet in 2025 a separate team, reading the polarization of a flash’s first radio signals, argued the opposite: that a cosmic ray shower had prepared the path. The two results sit in the literature, unreconciled.

Where the edge is

Three gaps, honestly stated. Where the missing factor of ten comes from, though balloons sample storms so sparsely that brief pockets of stronger field cannot be ruled out. How a spray of cold streamers becomes one hot, conducting channel. And what role, if any, relativistic particles play: aircraft campaigns have found tropical storms glowing in gamma rays for hours, which suggests something exotic is at work in these clouds, but not yet what.

What would count as an answer

A model that starts from measured fields and real cloud ice and predicts where and when the first spark forms, confirmed by an observation sharp enough to catch the seed itself. We can already guide lightning down a laser beam. An answer would let us say why the cloud fires at all.