Marius Comper

Shooting stars · Global Meteor Network cameras, 2025

Where shooting stars come from: the paths of 854,755 grains through the solar system

A shooting star is a grain of dust, usually lighter than a gram, that hits the air at tens of kilometres a second and burns up about 100 km high. When two cameras some way apart film it at the same moment, the two angles give its path through the air and, from that, the orbit it was following around the Sun. In 2025, 1,841 cameras of the Global Meteor Network, in 40 countries, managed this 854,755 times. Every thread on the screen is one of those orbits.

The screen shows 60,000 orbits drawn at random from the 854,755 (16,000 on a phone). Drag the picture to turn it.

19 October 2025, 02:46 UTC

A grain of Halley’s Comet over Romania

Five cameras filmed it, two of them in Romania. It began to burn 117 km above the village of Vințu de Jos in Alba county and went out 0.55 seconds later at 91 km, above Ighiu. It had entered the air at 67 km/s. Seen from 100 km it would have outshone Venus (absolute magnitude −5.2), and from its light the cameras estimate it weighed 1.3 grams.

The highlighted thread is the orbit it came on: a long ellipse much like Halley’s, crossing the Earth’s path at the point where the planet was that morning. Its far end, out beyond the planets, is measured with a wide margin of error; the part near the Earth is secure.

Every thread touches the Earth’s path

A grain is seen only if the Earth runs into it. So every orbit passes through the circle the Earth travels in a year, at the point where the planet was on the night of the meeting. The dense ring where all the threads gather is 2025 seen from above, night after night.

Threads gathered in bundles are the showers

Some grains travel together, on nearly the same orbit, because they came off the same body, most often a comet that sheds dust when it passes the Sun. The Earth crosses these trails on the same days every year, and their grains seem to come from one point in the sky, which gives the shower its name: the Perseids from Perseus, the Geminids from Gemini.

The network placed 27% of the orbits in known showers; 163 of those showers have at least 150 members.

The Perseids, dust of Comet Swift–Tuttle

The cameras recorded 24,702 Perseids in 2025. Their threads and the orbit of Comet 109P/Swift–Tuttle (the thicker line) almost overlap. On the scale astronomers use for such comparisons (the Southworth–Hawkins D criterion) the distance between them is 0.07; below 0.2, orbits are usually taken as a sign of a common origin.

The comet climbs to 51 times the Earth–Sun distance, nearly twice as far as Neptune, and was last closest to the Sun in December 1992. It returns in 2126. Perseids hit the air at 59 km/s.

The Geminids come on an asteroid’s orbit

The median orbit of the 36,415 Geminids of 2025 is almost identical to that of the asteroid (3200) Phaethon (D = 0.03), a rock a few kilometres across that dips to 0.14 of the Earth–Sun distance, nearly three times closer to the Sun than Mercury. Almost all the big showers come from comets; the Geminids are the best-known exception.

Their orbit is small and barely tilted, so Geminids reach the Earth slowly: 34 km/s, against 59 km/s for Perseids. They also go out lower: half of them below 84 km, against 95 km for Perseids.

Most belong to no shower

73% of the orbits of 2025 were placed in no shower. Most are probably grains from old trails, scattered by the pull of the planets; some may belong to faint showers the method does not recognise. A measure of the orbit relative to Jupiter (the Tisserand parameter) still shows what kind of body it resembles: 49% move like long-period comets, 25% like the comets steered by Jupiter’s pull and 26% like asteroids (the colours are in the picture’s key).

The classes describe the orbit, not the body the grain left: by this measure the Geminids too would be “asteroidal”. Phaethon at times shows comet-like activity, and how the Geminid stream formed is still being studied.

More arrive in the morning. In January and December, between 35 and 60 degrees north, the cameras recorded 3.1 times as many unassigned shooting stars between 4 and 6 am as between 7 and 9 pm (local solar time). The usual explanation: in the morning your side of the Earth faces forward, in the direction the planet moves along its orbit, and runs into grains like a car’s windscreen.

Your day: which dust the Earth runs through

Pick a day of the year. The Earth moves to where it is then, and the threads that light up are the orbits of grains recorded around that day in 2025. In 2025, on about 226 of the 365 days of the year, no shower reached even five in 100 shooting stars.

Peaks of the big showers

The shares come from one year of records, mostly from the northern hemisphere, where most cameras are; showers of the southern sky appear less often than they are.

When to see the 2026 Geminids and Leonids, and how many an hour

The Geminids peak on 14 December 2026 at 14:00 UTC, in daylight over Europe, so the good nights are the ones before and after. The Moon will be a thin crescent that sets early in the evening. The much weaker Leonids peak on the night of 17 to 18 November. How many you see depends on how high the point they seem to come from stands, and how dark the sky is where you are.

The count uses the International Meteor Organization’s formula: the zenithal hourly rate (150 for the Geminids, 15 for the Leonids) times the sine of the radiant’s height, reduced for the faintest stars you can see. Each place’s sky comes from the night-light map on the page “The lost stars”. The shape of the peak is estimated from the ratio of the shower to unassigned shooting stars recorded by the cameras in 2025, so the hourly numbers are an estimate, not a measurement. They assume no moonlight; when the Moon is up the chart shows it and you will see fewer.

The showers of 2025 and the bodies they come from

ShowerOrbits in 2025SpeedParent body
Geminids36,41534 km/s(3200) Phaethon (D = 0.03)
Perseids24,70259 km/s109P/Swift–Tuttle (D = 0.07)
Southern delta Aquariids13,66740 km/snot certain
Orionids10,44466 km/s1P/Halley (D = 0.17)
eta Aquariids8,22165 km/s1P/Halley (D = 0.10)
Southern Taurids5,48427 km/sproposed by the IAU list: 2P/Encke; D = 1.50, today’s orbit does not match
Quadrantids5,40641 km/s2003 EH1 (probable; D = 0.21)
shower with IAU code STS4,93429 km/snot certain
alpha Capricornids4,59622 km/s169P/NEAT (D = 0.15)
shower with IAU code FTR3,49928 km/snot certain
nu Eridanids3,00766 km/snot certain
Leonids3,00570 km/s55P/Tempel–Tuttle (D = 0.07)
Northern Taurids2,76827 km/s2P/Encke (probable; D = 0.23)
sigma Hydrids2,63358 km/snot certain
Comae Berenicids2,50963 km/snot certain
shower with IAU code NZC2,31439 km/snot certain

Where the data come from and what they do not say

The Global Meteor Network (GMN) is a network of cheap video cameras built and run by volunteers. When two or more cameras film the same meteor, the network’s software computes its path through the air, its speed and its orbit around the Sun. The page uses the network’s monthly files for 2025 (966,072 trajectories).

854,755 remain: 78,316 orbits with an eccentricity of 1 or more were removed, nearly all measurement errors rather than grains from outside the solar system, and 33,001 trajectories seen from angles too close together (under 10 degrees) for a reliable orbit.

Shower membership is set by the network, from the radiant’s position, the speed and the date of the meeting. A shower is tied to a parent body on the page only if the median orbit of its members matches that body’s orbit (D criterion under 0.2); between 0.2 and 0.3 it says “probably”, and above that, that the parent is not certain. Matching orbits support a common origin but do not prove it for each grain.

Mass is an estimate from brightness, with a fixed luminous efficiency of 0.7%; the median, 0.065 grams, describes the shooting stars cameras catch, not all of them.

Related: Halley’s Comet · The lost stars · The sky on your birthday (Romanian)

Sources

  1. Global Meteor Network, trajectory summary files 2025 (CC BY 4.0) globalmeteornetwork.org
  2. Vida D. et al. (2021), The Global Meteor Network – methodology and first results, MNRAS 506, 5046 doi.org
  3. IAU Meteor Data Center, list of all meteor showers (version of 3 Nov 2026) www.ta3.sk
  4. NASA JPL Small-Body Database: 109P, 55P, 1P, 3200 Phaethon and other parent bodies ssd.jpl.nasa.gov
  5. Rendtel J. (ed.), IMO Meteor Shower Calendar 2026 www.imo.net
  6. Southworth R. B., Hawkins G. S. (1963), Statistics of meteor streams, Smithsonian Contributions to Astrophysics 7, 261 adsabs.harvard.edu
  7. Astronomy Engine (Don Cross, MIT licence): Sun, Moon and radiant heights github.com