Learn space weather by watching real storms

Most space-weather guides are articles. This one is a simulator with a curriculum: five short modules, each built around the live map — real eruptions, replays of the great storms, and exercises where you make the forecast before reality is revealed. No account, no cost, works on a phone. Plan on roughly two hours end to end, or dip into any module alone.

1The Sun has weather

Start with the one idea everything else hangs on: the Sun doesn't just shine, it erupts — throwing billion-ton clouds of magnetized plasma called coronal mass ejections into space at hundreds to thousands of kilometers per second. A CME is matter, not light: it takes one to three days to reach Earth, and that travel time is why storms can be forecast at all. Its cousin the solar flare is a flash of light — here in 8 minutes, no outrunning it.

Read the ten-minute explainer, then look at today's real Sun:

📖 What is a CME? (the explainer) ☀ What did the Sun launch in the last 48 hours?

2Watch a storm happen

Theory is done; now watch the real thing. In May 2024, giant sunspot region AR 3664 fired a barrage of CMEs at Earth and produced the first G5 (extreme) geomagnetic storm in two decades — aurora reached the tropics. The replay below runs on NASA's measured data for those days.

▶ Replay the Gannon superstorm ▶ The same storm in 3D

• Watch the timeline at the bottom: each triangle is a launch, each dot an arrival. Notice the CMEs leaving in groups — later ones plowing into earlier ones. Stacked arrivals are what made this storm extreme.
• Scrub time by dragging the timeline: run the impacts in slow motion, then let it play.
• In the 3D view, notice each cone's tilt — a CME can look Earth-aimed from above yet fire over the poles. Latitude matters.

Then two case studies with a question each — scrub the replay and answer before reading the story:

▶ Québec 1989 — when the storm hit, the provincial grid collapsed in 92 seconds. Question: scrub back from impact — how much warning did operators have between the CME's launch and its arrival? (then read the story)

▶ Starlink 2022 — a minor G1 storm destroyed 38 of 49 freshly launched satellites. Question: if a G1 can do that, what's the actual lesson — storm strength, or timing? (the story)

3How forecasting actually works

Three instruments, in order of warning time:

• Coronagraphs (days of warning): space telescopes that blot out the Sun's disk so the faint CME shows. A CME aimed straight at us appears as a halo expanding around the whole Sun. This is where every arrival forecast starts.
• The drag model (the forecast): a fast CME plows into the slower solar wind ahead of it and decelerates, like a boat in water. Feed in launch speed and direction, and physics gives an arrival estimate. Honestly stated: it's good to about ±12 hours — validated against hundreds of real arrivals on our science page, misses included.
• The L1 tripwire (30–60 minutes): spacecraft sit 1.5 million km sunward of Earth. When a shock front sweeps past them, arrival is now measured, not modeled. That last hour is the only near-certain part of space weather forecasting.

▦ Open the tracker in monitoring mode ∑ The actual math & its error bars

4Predict it yourself

Forecasters commit to numbers before reality arrives. Your turn — no peeking before you've made your call.

Exercise A — the July 2012 monster

A CME leaves the Sun at ~2,500–3,000 km/s — one of the fastest ever recorded. Earth's orbit is about 150 million km out. Your call: roughly how many hours from launch to Earth's distance? (Distance ÷ speed. Ignore deceleration for the estimate.)

Exercise B — the St. Patrick's Day surprise

March 2015: a thoroughly ordinary CME leaves the Sun — moderate speed, taking a leisurely ~2 days to arrive. Nothing about the launch looks special. Your call: what storm level would you forecast — minor (G1), moderate (G2), strong (G3) or worse?

5Tonight, for real

Everything above ran on history. The Sun is doing something right now — so end by making a real call with live data:

1. Open CME activity today. Anything inbound? How fast?
2. Check the live tracker: what's the solar wind speed and Kp at this moment?
3. Make your forecast: is aurora plausible at your latitude in the next 48 hours? Commit to yes or no.
4. Then check back tomorrow and score yourself — that's the whole discipline of forecasting in one habit.

Want the answer pushed to you instead? The tracker has a free mailing list — storm alerts only, no noise. We publish our own misses.

6For teachers

Everything here is free, account-less, and runs on Chromebooks and phones. Three ready-to-teach outlines, each ~45 minutes:

Lesson 1 — "The Sun threw something at us." Warm-up (5 min): today's Sun on the today page — count the CMEs. Direct instruction (10 min): CME vs. flare, using the explainer's cannonball/flash framing. Guided activity (20 min): the Gannon replay on the projector — students narrate what they see; scrub on request. Exit ticket (10 min): "Why does a CME give days of warning but a flare gives none?"

Lesson 2 — Forecasting with honest error bars. Run Module 4's two exercises as a class: students commit predictions on paper before each reveal. Discussion: why was Exercise A predictable and Exercise B not? Connects directly to NGSS science practices — analyzing data, using models, and evaluating uncertainty — and to Earth–Sun system content in the ESS standards.

Lesson 3 — The great storms (jigsaw). Small groups each take one storm story (Carrington 1859, Québec 1989, Halloween 2003, Starlink 2022…), replay it, and report back: what happened, what broke, what would break today? Closing question: what's different about our infrastructure now vs. 1859?

• Projector setup: open the tracker with the clean preset — panels collapsed, just the map. Full screen, scrub live.
• Classroom dashboard: the free embed widgets put the live map on any class webpage or smartboard.
• Orientation: the 90-second video tour on the help page gets a class using the interface fast; the help page's glossary covers the jargon.
• Accuracy questions: our science page publishes the model's real validation — including the misses — if students push on "how do you know?"

Using this in a classroom? Tell us how it went (feedback box at the bottom of Help) — it directly shapes what we build for teachers next.