On 1 September 1859, Richard Carrington saw the Sun do something nobody had seen before. Two patches of intensely bright white light flared over the sunspot group he was sketching, bright enough to rival direct sunlight. He ran to fetch a witness. When he came back, within sixty seconds, it was already much changed and enfeebled. Minutes later it was gone. Richard Hodgson, watching independently from London, saw the same thing. Neither man knew what he had watched. The term solar flare did not exist yet.
Carrington also noted that a great magnetic storm began about seventeen hours and forty minutes after the flash, early on 2 September. He refused to claim the link. "One swallow does not make a summer," he wrote. Seventeen hours. That is the number that stayed with me: a billion-ton cloud of magnetized plasma, 150 million kilometers of space, and the Victorian world's most advanced machine running on the sky. The CME made the crossing in about 17.6 hours, far faster than a normal eruption, possible then only because an earlier one had cleared the way.
The storm hit a world that had just wired itself together. Compasses spun. Telegraph lines across Europe and North America sparked, shocked their operators, and set paper alight. On the Boston to Portland line, two operators cut their batteries out of the circuit and kept working on nothing but the current the sky was pushing through the wire. "Mine is disconnected, and we are working with the auroral current," Boston said. "Better than with our batteries on," Portland answered. They ran that line for two hours.
Above them the sky turned red and green as far south as Cuba, Jamaica, Hawaii, and Colombia. Gold miners in the Rocky Mountains woke at one in the morning, lit the stove, and started breakfast, convinced the Sun had risen behind a cloud. People in the northeastern United States read newspapers outside by aurora light. Nobody knew what the aurora was, or that the Sun powered it. They just looked up.
It remains the benchmark: the strongest geomagnetic storm we can point to and say this actually happened. The Colaba magnetogram suggests a Dst near −1760 nT, two to three times the 1989 storm that blacked out Quebec in ninety seconds. But the record is thinner than the folklore. The May 1921 storm was comparable by some measures, and the February 1872 storm ranks among the three largest. Tree rings record events in 774 and 993 that dwarfed this one. The Sun has done worse. It just has not done it while we were watching.
A repeat today would not burn telegraph paper. It would stress transformers that take months to replace, degrade the GPS timing the grid now depends on, and threaten satellites. The Carrington-class CME of July 2012 missed us by nine days. Lloyd's put the US bill for a direct hit at 0.6 to 2.6 trillion dollars.
Editorial
The part I keep coming back to
The operators. Nobody knew the aurora was powered by the Sun, let alone that a storm could push current through a wire. They cut the batteries anyway and kept working, because the line was live either way. That is the honest response to a phenomenon you cannot yet explain, and it is the response I hope I would have: not panic, not reverence, just keep the messages moving.
Our version of that night would be more expensive, not more surprising. The Sun does not keep a calendar, and the 2012 near-miss is a reminder that another Carrington-class event will happen. The grid was not built for it. The transformers that would fail are not stockpiled. The systems we rely on now, forecasting and shutdown, were both invented since then. The question I cannot answer is whether the people who run them would cut the battery and keep working.