Where the Sun shoots

Every coronal mass ejection NASA has catalogued since 2016, placed on an unrolled Sun at the point it launched from. Nobody told the picture where the active belts are. The eruptions draw them, and they leave the poles bare.

under 500 km/s 500 to 1,000 1,000 to 2,000 over 2,000 km/sdot size = speed · brightest dots: Earth was inside the cone

Data: NASA DONKI CME catalog, cone-model fits, through …. Longitude is measured from the Sun-Earth line at the moment of launch, so the lit middle is the hemisphere that faced Earth each time. The far side depends on STEREO-A and on limb events, so it is thinner than the near side for reasons of coverage, not physics.

1How to read it

Each dot is one CME at the latitude and longitude NASA's cone-model fit gives its direction, sized by speed. Ten years of them fill a broad band across the middle of the Sun and leave the poles almost bare: only a few hundred of 6,500 launches come from above 60°. CMEs erupt from the twisted magnetic fields of active regions, and active regions live in the low and middle latitudes, so the eruption map and the sunspot map cover the same ground. The empty poles are the cleanest thing on the page: the Sun does not launch storms from there.

The dots that matter to us are the bright ones near the centre: cases where Earth sat inside the CME's cone at launch. Even those are a geometric test, not a hit. Many arrive as glancing blows, and whether one becomes a storm depends on the magnetic field it carries, which nothing on this page can show. That is the whole reason the live tracker exists.

The same launches as rays: each begins at its launch point and follows the direction the CME moved across the sky, length by speed, with rings for the ones aimed straight at or away from the viewer. A CME launched near disk centre heads toward us, so on the sky it seems to shoot toward a pole. Pretty, but read the belts from the dots above, not from these.

2The butterfly, drawn by eruptions

Plot the same launches by date and latitude and you get a cousin of the butterfly diagram astronomers have drawn from sunspots since 1904. Sunspots start each cycle near 30° and slide toward the equator. The CMEs do something the sunspot chart cannot show. Through the quiet years of 2018 to 2020, when the Sun launched only a few dozen eruptions a year, almost all of them were aimed within 10° of the equator, even though the few spots of the time sat higher. The Sun's global field is a strong, simple dipole at minimum, and it herds outgoing CMEs toward the equatorial streamer belt. As the cycle rose the herding weakened and the launches spread: by 2024 the median direction sat 28° from the equator and only one in five was equatorial.

One dot per CME, coloured by speed, with the dashed line tracing each hemisphere's median launch latitude per half year. The thin years of 2018 to 2020 are the gap between cycle 24 and cycle 25. Directions come from NASA's cone-model fits; only CMEs with a fit are shown, and a direction is where the CME went, which is not always exactly where its spots were.

3One Sun per year

The whole Sun, one year at a time. The minimum years are nearly bald, and what little erupted hugs the equator. Then the launches multiply and spread to 30° and beyond as the cycle climbs to its 2024 to 2025 peak. The dashed globes are the next five years, a forecast of the decline to the coming minimum built from NOAA's predicted sunspot numbers.

Forecast globes: the yearly launch count follows NOAA's predicted sunspot number through a straight-line fit to the observed years 2020 to 2025, and each forecast year's dots are a resample of the past year whose sunspot number is closest, with a little jitter. NOAA's forecast ends in December 2030, so 2031 carries its final months forward as the approaching minimum. A picture of the expected pace and spread, not a prediction of any eruption.

4What this is not

Not a hit list. "Earth inside the cone" is a launch-time geometry test. The tracker's arrival model and, above all, the measured field at L1 decide what actually happens.
Not evenly sampled. The near side is watched by SOHO and SDO around the clock. The far side depends on STEREO-A's position and on limb events, so its count is lower for reasons of coverage, not physics.
Not exact positions. Cone-model fits carry uncertainties of several degrees in direction and a few hundred km/s in speed, and the direction a CME flies can differ from where its source region sat, because the Sun's field deflects it. Individual dots are approximate; the bands and the empty poles are the robust part.

5Sources and further reading

NASA CCMC DONKI: the CME catalog and its cone-model analyses (latitude, longitude, half-angle, speed).
Maunder (1904), the first butterfly diagram, drawn from Greenwich sunspot positions. Spörer's law describes the equatorward drift it shows.
Cremades and Bothmer (2004) and later work on CME deflection: eruptions are steered toward the heliospheric current sheet, most strongly near solar minimum, which is why the quiet years here sit on the equator.
See the same belts live on the tracker, where the Sunspots readout unrolls today's Sun, and the cycle's timing on the solar cycle page.

← Back to the tracker