In science, no matter how confident we are in our theories, there’s no substitute for interrogating the natural world by asking it questions about itself directly: through observation and experiment. Sometimes, that requires setting up conditions in a laboratory to create certain events whose outcomes we can measure to whatever precision we desire. At other times, however, it requires looking out into space — at the natural laboratory of the Universe — to observe how nature behaves. No matter what our expectations were beforehand, there’s no substitute for actual data in figuring out how things actually are.
Although Pluto was the first object ever discovered out beyond Neptune, spotted way back in 1930 for the very first time, its atmosphere was only directly discovered and measured in 1988: when an observatory in Earth’s southern hemisphere observed it occulting, or passing in front of, a background star. While many other Kuiper belt occultations have occurred since, only Pluto had ever been shown to have an atmosphere, rendering it unique among the known trans-Neptunian objects.
Until now. In an all new 2026 study led by Ko Arimatsu, one of the smaller Kuiper belt objects known, 2002 XV93, was precisely observed from three separate locations during a 2024 stellar occultation, and was determined to have an atmosphere, after all. This makes it the second known Kuiper belt object, after Pluto, to have an atmosphere. Here’s how we figured it out, and what it just might mean for the fields of astronomy and planetary science.
There are a large number of comets with periods between 20 and 200 years, originating from beyond Jupiter but before the end of the Kuiper belt and scattered disk in our Solar System. Beyond that is another population of objects with orbital periods in the many thousands-of-years range, suggestive of an even more distant reservoir of objects: the inner Oort cloud. The majority of Kuiper belt and Oort cloud objects remain unperturbed, and can still be found, with a good enough set of observatories, well beyond the orbit of Neptune.
Credit: William Crochot and NASA
When we think about the bodies in our Solar System, what we view today is the end result of a 4.5+ billion year tale of survival. Early on, when our planets, moons, asteroids, and Kuiper belt objects were forming out of the protoplanetary material that surrounded our young Sun, there were many volatiles present: molecules that would eventually be vaporized, sublimated, or otherwise evaporated away by too much energetic radiation, such as direct sunlight. Around the planets and moons of the inner Solar System, volatiles are rare, with materials such as:
- hydrogen ice,
- helium gas,
- nitrogen ice,
- methane ice,
- carbon monoxide ice,
- carbon dioxide ice,
- and even water ice,
only rarely found. They’re most abundant in permanently shadowed craters, deep beneath a planetary surface, or — in the case of water ice — as part of a temporary configuration on a world with liquid and gas-phase water as well.
The farther out in the Solar System you go, farther away from the Sun, the longer these volatiles can persist. In the case of Pluto, the largest body presently known in the Kuiper belt, the New Horizons flyby in 2015 showed a surface rich in many of these different ice species, an atmosphere rich in clouds and hazes containing these volatile molecules, and even snows where these ices precipitate. The atmosphere was thin but substantial, with a pressure of around 10 microbars (about 1 Pa), and was thought to be generated by the routine sublimation of those surface ices as Pluto regularly reaches perihelion, bringing it closer to the Sun than Neptune.
As New Horizons flew by Pluto, it acquired views of this distant world from several different angles. Here, 15 minutes after closest approach, New Horizons looked back and captured this image, revealing layers of atmosphere and planetary hazes, along with shadows cast along the planetary surface by towering ice mountains reaching up to 11,000 feet (3500 meters) high.
Credit: NASA/JHUAPL/SwRI
In the many years since the first discovery of Pluto’s atmosphere, an enormous number of objects from out beyond Neptune — primarily, but not exclusively, populating the Kuiper belt — have been discovered. Starting in the 1990s and continuing into the 21st century, these worlds are actually quite abundant, with many of them ranging from hundreds of kilometers up to a thousand or more kilometers in size. Many of them are now classified as dwarf planets, and it’s theoretically possible that many of the smaller objects just a few hundred kilometers in size, dominated by ices, have already pulled themselves into hydrostatic equilibrium as well.
One such object is was discovered in 2002: 2002 XV93, which makes two revolutions around the Sun for every three revolutions that Neptune makes. After its 2002 discovery, astronomers searched for what we call “precovery” observations: where historical data is mined to look for any other images of this object, now that we know it exists and where it should have been. Observations dating back as far as October of 1990 were found to contain this object, allowing astronomers to reconstruct its orbit and determine a great number of its orbital properties quite precisely: its perihelion, its semimajor axis, and its eccentricity among them.
This diagram shows the reconstructed orbit of the trans-Neptunian object known most commonly as 2002 XV93, in white, which orbits out of the plane of the other planets in the Solar System slightly. Its orbital distance from the Sun varies from a minimum of 34.4 AU to a maximum of 44.12 AU. While the object was discovered in 2002, precovery data exists going all the way back to 1990, giving us a 36 year baseline of data from which we can reconstruct its orbit.
Credit: JPL/NASA/Caltech
There are many objects known that are extremely similar to 2002 XV93: they form a class known as plutinos. Pluto, Charon, Orcus, Achlys, and Ixion are the largest known plutinos, and hence they all have names related to mythological creatures associated with the underworld. 2002 XV93, being smaller and having no other known remarkable properties, hasn’t received any such designation: at least, not yet. In fact, prior to this latest study, the most remarkable thing that happened to 2002 XV93 was that it was observed spectroscopically by JWST back in 2022, allowing us to acquire a series of spectra of ice-rich plutinos, including this one.
Those JWST observations, which also looked several other Kuiper belt objects, including numerous plutinos (at right, below), revealed crystalline water ice at 3.1 microns, which provides strong indirect present for amorphous water ice on those bodies (with dips at 1.5 and 2.0 microns), as well as carbon dioxide ices (with a dip at 4.27 microns) on 2002 XV93. However, there were no signs of what one might call “hyper-volatile” compounds that can sublimate into the gas phase at temperatures typically achieved at these roughly 40 AU distances: methane, nitrogen, or carbon monoxide.
This graph shows the JWST-obtained spectra of a variety of Kuiper belt objects, including cold classical KBOs (at left) and plutinos (at right). Included among the plutinos is the once-unremarkable object 2002 XV93, with signatures of amorphous water-ice, crystalline water-ice, and carbon dioxide ice leaving characteristic signatures in its spectrum.
Credit: A.C. Souza-Feliciano et al., Astronomy & Astrophysics, Letter to the Editor, 2024
These JWST observations painted a very similar picture for 2002 XV93 as for other similar plutinos found in the Kuiper belt: suggesting that if it ever had a substantial atmosphere, it sublimated away and escaped into space long ago. Since the temperature of such a world should be determined by its distance from the Sun, and its ability to “hold onto” volatile compounds should be determined by its surface gravity, 2002 XV93 was expected to be an airless world, like all the other known Kuiper belt objects, including Pluto’s giant (and much larger than this object) moon Charon, had previously been determined to be.
But in science, we don’t simply take the observations we have and our best theories for how we predict things should be and call it a day; we seek to probe, measure, and test the Universe directly, wherever possible. Only by looking directly, and by making direct observations and measurements of the quantities we’re seeking to understand, can we truly uncover how the Universe works. That’s part of the key process of what science is at its core: putting questions about “what’s going on in the Universe” to the Universe itself, and interrogating it in such a way that it reveals the answers to our questions directly, through experimentation, observation, and measurement. That’s what led to the key 2024 occultation measurements, which revealed the existence of 2002 XV93‘s atmosphere.
When one astronomical object occupies the same line-of-sight as another, an occultation will occur, as the “closer” object blocks the light that would otherwise be visible from the “farther” object. The Moon occults all of the other planets; the Moon and planets and other planetary bodies occult background stars, revealing the relative distances between them. If a planetary body has a substantial atmosphere, occultation studies can reveal that as well, which is how Pluto’s atmosphere was discovered back in the late 1980s.
Credit: Bob King/Stellarium/Sky & Telescope
The only way to have an occultation is where, from the perspective of any given observer, a foreground object lines up with and, from that perspective, passes in front of a luminous, background object, obscuring its light. We normally approximate:
- Earth as a point,
- distant objects in our Solar System as a point,
- and the even-more distant stars as a point,
but when it comes to occultations, that’s not a good enough approximation. These distant objects don’t behave as truly point-like objects, but rather as small disks in relative motion to the background of the fixed, more point-like stars, where there’s a time that the occultation begins, a duration for the span of the occultation, and a time that the occultation ends.
Moreover, Earth can’t be treated like point-like either. Just as your left eye and right eye see different perspectives when you switch between them, particularly when you’re viewing nearby objects relative to more distant ones, the location of where you are on Earth matters tremendously for whether you get an occultation or not, and what such an occultation would look like. Based on the best data we had as of January 10, 2024 — the date of the occultation of 2002 XV93 — what you see below is the predicted path of the occultation relative to three locations in Japan, where the red line was the predicted center-line of the occultation, the blue solid lines surrounding it represents the predicted size of the shadow, and the blue dotted lines represent the uncertainty on the prediction.
This diagram shows the expected path of the January 10, 2024 occultation of a background star by Kuiper belt object 2002 XV93, where the red line represents the predicted center-line, the blue solid lines represent the expected limits of the shadow’s reach given a diameter of 500 km, and the blue dotted lines represent the uncertainty in the location of the center line. In actuality, the center line passed just south of Kyoto, the shadow intersected both Kyoto and Kiso, and Fukushima was just out of the shadow’s range, but was sensitive to any atmosphere that 2002 XV93 possessed.
Credit: Ko Arimatsu et al., Nature Astronomy, 2026
When the occultation event occurred, there were three sites with telescopes set up to attempt to measure it: in Kyoto, in Kiso, and in Fukushima, all in Japan. Based on the uncertainties, scientists were hoping to get at least one observation of the occultation, but recognized it was possible that the shadow would fall outside of the predicted region, and that no detection of the object would be seen at all. While Fukushima was right along the predicted center-line, the uncertainties were large enough that the occultation shadow could have missed all three sites entirely, or hit only one or two of them. After all, based on how small the object 2002 XV93 actually is, even the center-line of the occultation should create an event that would last for under 20 seconds as observed from Earth.
However, with this particular set of observations, the astronomers got lucky in a good way.
- In Kyoto, predicted to just fall outside of the occultation shadow, the longest-duration occultation of the three sites was actually observed: nearly 20 full seconds, as it wound up being located just slightly above the location of the true center-line of the occultation.
- Kiso, just slightly to the north of Kyoto, saw an occultation that was almost as long, and due to the superior observing conditions (with much less light pollution), was able to acquire even better data than at the Kyoto site.
- And Fukushima, which was the northernmost site of all, didn’t actually observe an occultation of the object in front of the background star, but saw a partial dimming/absorption of the light, followed by a re-brightening, indicating that it was located just outside of the disk of this planetary body, but where an atmosphere partially absorbed the light from the background star during the occultation event.
These three graphs show the occultation light-curves obtained from three sites in Japan during January 10, 2024, as Kuiper belt object 2002 XV93 occulted a background star. The dip in the received light seen from all three sites, but most exquisitely from Kiso, reveals the presence of an atmosphere, with the Fukushima site data notably experiencing only an occultation by the atmosphere, not by the main disk of the planetary body itself.
Credit: Ko Arimatsu et al., Nature Astronomy, 2026
Above, you can see the data acquired at all three sites, with error bars on each data point, as published in Ko Arimatsu’s team’s study in Nature Astronomy on May 04, 2026. If the Kuiper belt object doing the occulting, 2002 XV93, had lacked an atmosphere entirely, then the “walls” of the big dip in flux from the Kyoto and Kiso sites would have made straight lines. If there were no occultation from Fukushima, the observations would have yielded a straight, horizontal line, while if there were only a “partial occultation” observed, there would have been a straight line fit to the dimming appearing on either side of the peak of the event.
Instead, what we can infer from the data — with the best data coming from the Kiso site — is that this world must indeed have an atmosphere, and not only that, but we can use the occultation data to infer “how much” of an atmosphere it possesses. The refractive signature of the atmosphere indicates that:
- it’s thin,
- with a pressure in the 100-200 nanobar (about 0.1-0.2 Pa) range,
- likely composed of either methane or nitrogen gases,
- and extending for approximately 40 kilometers in altitude above the solid surface of the world.
From ingress in Kiso, egress at Kiso, and the full occultation light-curve at Fukushima, the two models you see below represent the best fits to the data of all.
This graph shows, based primarily on data from an observatory at Kiso in Japan but augmented by Fukushima data and supported by Kyoto data, the reconstructed atmosphere for 2002 XV93. The data is inconsistent with a no-atmosphere model, and instead supports either a methane or nitrogen gas atmosphere with a substantial 100-200 nanobar pressure atmosphere around it.
Credit: Ko Arimatsu et al., Nature Astronomy, 2026
Encapsulated in these results is a profound illustration of the core motivation behind why we do science at all: because until we look at the Universe itself, and until we allow actual data to answer the question of what a thing is and how it behaves, we can never truly be certain. Only by making the key measurements, and taking the actual data, can we ever know what’s actually occurring in Nature itself. Remember, physics, at its core — including astrophysics, geophysics, and all other applications of physics — is an experimental and observational science. Theories and models can tell you what you expect to happen based on what you know so far, but if you want the answer to how the Universe actually behaves, there is no substitute for looking for yourself.
What’s extra remarkable about this study is, with such a small size for the planetary body (estimated to have a diameter of 470 km, excluding the atmosphere) and with such a cold calculated temperature at its distance from the Sun, one can calculate the rate of escape of any theoretical atmosphere based on its composition. With either a nitrogen or methane composition to its atmosphere, that would lead to an evaporation timescale for the entire atmosphere of 2002 XV93 on timescales ranging from 100-1000 years. Since this object is billions of years old, that tells us something profound: this atmosphere must have a source that’s replenishing it, perhaps even continuously, over time.
This chart shows the measured pressures (or constraints on pressure) around a variety of trans-Neptunian objects. Of all the objects ever measured through stellar occultations, only Pluto and 2002 XV93 were found to actually have non-zero atmospheres, with all other plutinos, including plutinos much larger than 2002 XV93, exhibiting no signature of an atmosphere at all.
Credit: Ko Arimatsu et al., Nature Astronomy, 2026
Could it be via the same mechanism as Pluto, where it approaches the Sun in its orbit, causing surface ices to sublimate, putting materials into its atmosphere, and then those materials condense and precipitate as it moves farther from the Sun, lessening its atmosphere, until it returns to closest approach again just a couple of hundred years later?
That’s unlikely. There were no frozen gases on the surface seen with JWST data, pointing away from the “sublimation formed an atmosphere” scenario. Instead, ideas like:
- cryovolcanic eruptions brought interior material to the surface, forming an atmosphere,
- a recent impact from another object kicked up material, creating its atmosphere,
- or that there’s something profound about this body, in particular, that renders it uncommon (and able to hold an atmosphere) among Kuiper belt objects.
This atmosphere can’t be a universal property among similarly sized Kuiper belt objects, however. Makemake, which has a diameter of 710 km, has also experienced a stellar occultation and was found to have no atmosphere at all, down to limits much more stringent than the measured atmosphere of 2002 XV93; a similar story exists for Quaoar, also owing to the work of Ko Arimatsu, as well, with limits down to 6 nanobars of pressure.
As with a great many unexpected discoveries, this represents a jumping-off point for future studies, as we’re now compelled, as a global community, to try and make sense of this finding in the context of the full suite of what’s known. Pinpointing the origin and properties of this atmosphere, as well as the quest to find additional planetary atmospheres around bodies beyond Neptune in our Solar System, ought to intensify in the aftermath of this profound new find. It’s a cautionary tale against assuming the answer before you look. Sometimes, simply daring to conduct the science yourself is all that’s required to discover the reality of what was previously thought to be impossible!