In early October, NASA published its preliminary scientific report on Artemis II, and tucked inside it was a scene that made everything inside me jump. On April 6, 2026, Orion slipped behind the far side of the Moon, the Sun vanished behind the lunar disk, and for nearly an hour the spacecraft was surrounded by a darkness unlike anything on any orbit around Earth. In that darkness, three astronauts — Reid Wiseman, Victor Glover, and Jeremy Hansen — watched rocks slam into the Moon from a distance of forty-six thousand miles. Five flashes. Barely visible “pinpricks” of light.
The flashes themselves had been predicted — scientists had asked the crew to watch for them. But when Mission Control in Houston heard how many there had been, “audible screams of delight,” in the words of lunar science lead Kelsey Young, broke out in the room. No one had expected a result like that: “we were not prepared to see that many impacts.”
And that is where the thing I set out to untangle begins. The Moon looks like the most frozen body in the sky — gray, motionless, “always the same.” In fact, it is the nearest shooting range to us, shelled every day, with the firing never stopping for billions of years. We have simply grown used to not looking. The question I decided to test was: what do we actually know about this bombardment — who has seen it, how much is falling, and why did nearly an hour of stellar darkness prove more valuable than any telescope?
Let me start with what it looked like, because there are details here that are easy to lose in the retelling. The eclipse lasted nearly an hour — the Sun, the Moon, and Orion lined up in such a way that the spacecraft found itself in the longest shadow possible in the Moon’s vicinity. In that darkness, the crew followed the plan: watch the dark side and log anything that flashed. None of the three saw everything alone — Wiseman used a blue marker, Glover a green one, Hansen a red one, and when it was over, five numbered dots had appeared in OneNote on the tablet. Wiseman marked three of them with a separate note: “about ten minutes after the eclipse began.” The fifth came with the note “closer to the end.”
Then the scientists did the skeptics’ routine work: ruling out cosmic rays, reflections from the spacecraft, everything that might explain flickers before tired eyes. Nothing fit — the flashes lay on the lunar disk, in different places, and all three observers had seen them independently. Their brightness made it possible to estimate the size of the bombs: each rock was “no bigger than a couple of inches” across. A couple of inches. Pebbles no bigger than golf balls, transformed at thousands of miles per hour into pinpoint explosions visible from seventy-four thousand kilometers away.
And here is the most delicious part of the report. Apollo astronauts saw similar flashes from lunar orbit — but Artemis, it seems, caught the faintest glow the human eye has ever managed to see. The Apollos flew through darkness in their own way. No one had deliberately tuned their eyes to hunt for such dim flashes from forty-six thousand miles away.
Why is a rock visible at all? The physics here is generous with paradoxes. On Earth, a meteoroid slows in the atmosphere, heats up from friction, and leaves behind a bolide — the fiery trail people admire. The Moon has no atmosphere, so the projectile does not slow down at all: it hits the surface at tens of kilometers per second. The slowing happens in a single beat — against solid ground.
The share of energy that goes into light is laughable: less than one percent. Ninety-nine goes into shaking the soil, into ejecta, into the crater. But that percent is enough: at the moment of impact, the temperature at the point of contact spikes so high that the flash can be seen from Earth, never mind from orbit. The irony is that NASA runs an official citizen-science project for this very thing, and its explanation is phrased with merciless precision: “rocks hit the Moon all the time.” You just have to be looking in the right direction — and the right direction, as it turns out, is the dark one.
One useful contrast, for the full picture: on Earth, the same ~1-meter rocks most often burn up before they get here. Sometimes spectacularly — like Chelyabinsk in February 2013: nineteen meters, half a million tons of TNT, a shock wave that circled the planet twice. But between the Moon and Earth there is an abyss in this regard: we have a shield we do not even think about until we start comparing.
The numbers on this bombardment were compiled by Bill Cooke, head of NASA’s Meteoroid Environment Office. They are the kind that make you look at lunar landscapes differently afterward.
The finest dust, under a millimeter, falls on the Moon at a rate of eleven to eleven hundred tons per day — a hundredfold range, because no one honestly knows how to count the smallest fraction. Rocks the size of ping-pong balls — roughly a hundred a day, thirty-three thousand a year, each striking with the energy of three kilograms of TNT. A five-kilogram rock leaves a crater nine meters across and throws up seventy-five tons of soil. An object two and a half meters wide — that is already once every four years, and a full kiloton of TNT — energy in the class of the smallest nuclear weapons. And only at the very end does Cooke offer the number that reassures you by exactly as much as necessary: a ping-pong-ball rock falls on any given square kilometer once every thousand years.
Notice how this fits together. The Moon takes thirty-three thousand “ping-pong balls” a year, but it has an area of thirty-eight million square kilometers, and the rare rock lands where it is welcome. Hence the whole paradox of future bases: statistically, you are almost certainly not going to get hit — but “almost certainly” is calculated across the whole surface, while a base occupies a single point, and all around it a secondary squall of ejecta is constantly boiling up.
Lunar impact flashes are a young science, and their birth smells more of an amateur club than a laboratory. A Leonid meteor storm was expected in November 1999. Someone dropped a thought on NASA’s website: what if meteors were hitting the Moon, too? David Dunham read that line. He was president of IOTA — the International Occultation Timing Association, an organization made up mostly of amateur astronomers who spend their lives learning to catch the moments when the Moon covers a star. Dunham sent out an online call: friends, watch the Moon’s dark side.
It worked. On the night of November 18, North American observers recorded at least six flashes, each confirmed by at least two independent observers — either on video or with their own eyes. The first was seen by Brian Cudnik in Houston: a fourteen-inch telescope, 4:46 in the morning, a brief orange flash on the dark side, roughly twice as dim as Polaris. While he was writing Dunham an email, Dunham was already scrolling through his black-and-white video recording from a five-inch telescope — and found the exact same flash, pinning down the time to the second. Later, two more observers found another four flashes on their tapes.
The formula behind that breakthrough is worth remembering: the telescope does not have to be big, but you need the right night, the Moon’s dark side, and coordination through the internet. Nature did the rest. It was the first confirmed detection of lunar meteor impacts — and a decade and a half later, the same hunt caught the record described below.
On September 11, 2013, at 20:07 GMT, the Spanish MIDAS system — a network of telescopes in the south of the country, created specifically to hunt these flashes — caught an event that prompted José María Madiedo of the University of Huelva to put it without qualification: “the largest and brightest impact flash ever observed on the Moon.”
The numbers were these. A rock weighing around four hundred kilograms, between six-tenths of a meter and one and a half meters wide, struck the lunar surface at sixty-one thousand kilometers per hour. Energy — fifteen tons of TNT. An ordinary lunar flash lasts a fraction of a second. This one burned for more than eight seconds and approached Polaris in brightness — meaning that, in theory, it could have been seen with the naked eye from Earth, if someone had been looking at the right moment. The team’s models predicted a crater around forty meters across.
But the most important thing about this event was not the record; it was the conclusion drawn from it. Meter-scale impacts, the Spaniards decided, occur far more often than the models had shown — on the Moon and on Earth alike. It is just that here, such rocks burn up in the atmosphere, leaving a bolide, while on the Moon they leave a crater. In this case, the Moon works as a detector: it records what Earth conceals.
For a long time, lunar soil was thought to be a museum: events layered over billions of years with almost no movement. LRO, which has been in lunar orbit since 2009, shattered that picture in the dullest and therefore most convincing way possible — with pairs of “before and after” images.
A team led by Emerson Speyerer of Arizona State University compared images of the same sites taken at different times and, over the course of the mission, found more than two hundred new craters between three and forty-three meters across — a third more than the models predicted. Around the craters, they found forty-seven thousand “patches” — areas where the soil had been shaken up. The key discovery was that most of the patches formed around fresh craters: secondary ejecta, fragments flying outward from the impact site and continuing to churn the surface.
The recalculation produced a figure that makes engineers of future bases’ hair stand on end. The upper couple of centimeters of regolith turn over every eighty-one thousand years — at least a hundred times faster than earlier estimates, which accounted only for direct micrometeoroid hits and ignored the secondary squall. In other words, the lunar surface is not a museum but an actively churned roadbed. Emerson Speyerer said outright what this means for designers: equipment on the surface must be able to withstand strikes from small particles traveling at up to five hundred meters per second.
And the peak of this story came in 2024, found not by a scientific grant but by human curiosity. Robert Wagner of Intuitive Machines was scrolling through an enormous map of the Moon as part of a routine data-quality check and noticed an unusually large bright patch with a dark halo — a sure sign that the soil there had been shaken up. Comparison with images from 2011 confirmed it: between April 11 and May 22, 2024, an object the size of a three- to six-story building struck the Moon and left a crater two hundred and twenty-two meters across and forty-three meters deep. Two football fields. Three school buses stacked on top of one another. Scientists estimated that something like this happens once a century or less. The crater was named McGetchin, after a pioneer of lunar geology.
Over seventeen years, LRO has already accumulated a thousand new craters and more than a hundred thousand documented changes to the surface. The Moon, I repeat, looks frozen. It is not frozen — we just do not catch it in the act.
There is also a seismic side to all this, one that is easy to forget. From 1969 to 1977, a network of seismic stations deployed by Apollo crews operated on the Moon. Over eight years, it recorded thousands of events, with meteoroid impacts set apart as their own class of moonquakes — alongside deep and shallow ones. Modern models even speak of a “meteoroid seismic hum”: the Moon is not silent; it is continuously knocking, only the human ear has never heard it and probably never will — because sound has nothing to travel through.
This is where everything converges into one point. Artemis II observations provided what no telescope can: eyes that saw impacts in real time on the side of the Moon that no instrument on Earth could see at that moment. For the frequency of small impacts — and those are precisely what determine the long-term risk to future bases — this is an extraordinarily rare dataset. Plus a fundamental rediscovery: the statistics say a rock will almost never hit one particular point, but around that point a secondary squall is constantly boiling, and it is that squall now being measured through the flashes.
And the final irony I took away from this investigation. To see the most intense bombardment available to humans, no new telescope was needed. It took a piece of darkness — nearly an hour when the Sun hid behind the Moon — and three pairs of eyes willing to stare into the blackness and not look away. Five flashes in an hour — an impact rate instruments on earlier missions could only dream of. A couple of inches, seventy-four thousand kilometers, three people with tablets. Sometimes science is just not blinking.