Solar Cycle 24's biggest storm arrived, improbably, on St Patrick's Day — and turned skies green from Alaska to the northern-tier United States while the internet made the obvious joke. It remains the textbook example of a moderate CME over-delivering.
On March 15, 2015, a partial-halo CME left the Sun at unremarkable speed. Forecasters expected a modest glancing blow. Instead, on March 17 the CME's sheath and flux rope carried a long interval of strongly southward magnetic field — the ingredient that actually opens Earth's magnetosphere — and the storm deepened all day, bottoming at Dst −223 nT1: a G4 severe storm2 and the strongest of Solar Cycle 24.
It became a case study in why storm forecasting is hard: the eruption looked ordinary; the geometry was not. The observatory's field guide covers the same lesson for live forecasting today — Kp tells you what happened, Bz tells you what's about to.
The observatory's St Patrick's scene is rebuilt through the measured pipeline with satellite-anchored oval boundaries in both hemispheres3. It's the cycle-24 benchmark to compare against the 2024–2025 storms: a severe storm from a slower solar cycle, still strong enough for aurora across the northern US.
The source event was unassuming: a partial-halo CME on March 15, 2015, launched with only a C-class flare — nothing that screamed "biggest storm of the cycle." It arrived early on March 17, St Patrick's Day, announcing itself with a sudden impulse in the morning hours UT. What made the day historic was not the arrival but what the solar wind did afterward: the interplanetary magnetic field turned strongly and persistently southward, and the storm deepened in two distinct steps through the day rather than one clean main phase.
By evening it had become the strongest geomagnetic storm of Solar Cycle 24 — a G4 event nobody had forecast at that severity. Aurora was photographed across the northern tier of the United States and from central Europe, and the southern oval lit up for observers in Tasmania and New Zealand. For forecasters it became a case study in why CME magnetic-field orientation, which cannot be measured until the cloud reaches the upstream monitors, is the hardest and most consequential unknown in space weather.
A two-step main phase changes what observers on the ground experience. Instead of one deep excursion and a slow recovery, the ring current was energized, paused, then energized again — so the auroral oval pushed equatorward, breathed back, and pushed again hours later. Nights like that reward patience: observers who gave up after the first lull missed the second act. The reconstruction preserves that rhythm, because it is driven by the measured indices of the day rather than an idealized storm curve.
The St Patrick's Day scene is built from particle-precipitation measurements taken by the POES and MetOp satellites as they crossed the auroral zones through the storm, fused onto the observatory's oval grid, with field-aligned-current data from the AMPERE constellation as a placement prior. The band's position was then re-anchored against auroral boundaries extracted from DMSP/SSUSI far-ultraviolet imagery — several hundred independent boundary observations per hemisphere — which corrected a systematic poleward bias that raw current-system fits carry. Both hemispheres are reconstructed; the storm's southern display over the Tasman Sea renders from the same fused data as the northern one. As everywhere on this site, satellite passes are measurements and the evolution between them is a fitted model, labeled as such. The data & methods page documents every step, including the parts that remain honest approximations.
Numbered references for the claims above.
More storms: Gannon May 2024 · Halloween 2003 · Carrington 1859 · October 2024
By The Space Weather Observatory · checked against the data & methods record · Updated August 21, 2026
the sky · the sun · field guide · the great storms · data & methods · about · privacy · pro · ☁️ seasonmap · ☕ support