How NASA Tracks Near Earth Objects: Radar and Data (2026)

Short version: NASA tracks near Earth objects by photographing the same patch of sky over and over, spotting the faint dots that shift position between frames, turning those positions into an orbit, and then checking that orbit against Earth’s own path. When the two paths line up badly, an impact monitoring system at JPL recalculates the odds every time new data arrives.

If you have seen an asteroid headline this month, that pipeline is what produced it. Most of the time it ends quietly: the object passes, the number of people killed by nothing is zero. Here is how the machinery actually works, in the order it runs.

Table of Contents

• What are near Earth objects?
• How NASA tracks near Earth objects from discovery
• How telescopes find objects that move against the stars
• What happens after a near-Earth object is discovered
• How NASA calculates an asteroid’s orbit
• How radar helps NASA study near Earth objects
• How NASA assesses impact risk
• Why an impact probability spikes and then falls back to zero
• What the Torino Scale actually means
• Why NASA tracks some objects more closely than others
• Frequently Asked Questions
• Has an asteroid ever hit Earth with humans on it?
• What happens if an asteroid on a collision course hits Earth?
• What will happen on March 16, 2880?
• What is the Torino Scale used for?
• Why is a 2029 asteroid approach described as scary?
• Can NASA stop an asteroid from hitting Earth?
• Key Takeaways

What are near Earth objects?

A near-Earth object is any comet or asteroid whose orbit swings close enough to Earth’s that a collision is worth taking seriously. The formal cutoff is a closest approach to the Sun of less than 1.3 astronomical units, which is roughly the distance at which an orbit crosses Earth’s lane around the Sun.

The category is really three families wearing one label. Asteroids are rocky or metallic leftovers from planet formation, most of them in the main belt between Mars and Jupiter. Comets are icy bodies whose volatiles boil off near the Sun and produce the tail. Meteoroids are simply the small grains of dust and grit that burn up in our atmosphere every night.

Asteroids that meet the near-Earth criteria get the abbreviation NEA, and comets qualify too, which is why NASA uses the broader NEO term.

Within that group sits the potentially hazardous asteroid, or PHA. A PHA is an NEO larger than one kilometre across whose orbit passes within about 0.05 AU of Earth’s. Both conditions have to be met. A two-kilometre asteroid on a path that never comes near us is not a PHA, and a 100-metre rock on a path straight at us is not one either, because it would burn up long before it reached the ground.

That single label carries most of the public alarm. It is a screening category, not a forecast.

How NASA tracks near Earth objects from discovery

The direct answer is that automated survey telescopes take wide-field photographs of the sky repeatedly, software compares the images to find objects that moved while the stars stayed put, and the measured positions feed orbit-determination software that produces a three-dimensional trajectory. Everything after that is arithmetic run forward in time.

Nobody sits at a desk watching a video feed for moving dots. The scale is the problem. There are far too many asteroids for human eyeballs, so the discovery step is a software problem wrapped around a telescope problem.

The division of labour matters here, because it is widely misunderstood. NASA’s Near-Earth Object Observations Program funds and coordinates the work. The Jet Propulsion Laboratory, a federally funded research centre managed for NASA, runs the Center for Near Earth Object Studies, or CNEOS, which is the office that holds the orbit catalogue, computes close approaches and operates the impact monitoring system called Sentry.

But JPL does not find most of the objects. That job belongs to a spread of ground-based surveys run by universities, national observatories and agencies in other countries, plus a steady contribution from amateur astronomers with serious telescopes. Several of the discoveries that get news coverage start with a hobbyist in a backyard observatory somewhere.

How telescopes find objects that move against the stars

Detection relies on repeat imaging of the same sky region, typically minutes apart. Stars hold still in the frame. An asteroid drifting a few arcseconds across the sky lands somewhere slightly different each time. Stack two images and subtract them, and the stars vanish while the moving thing stays.

The main ground surveys each handle this differently. Catalina Sky Survey in Arizona scans for objects that move fast, including ones too faint for deeper surveys. Pan-STARRS in Hawaii runs wide-field cameras on a scheduled nightly search. The Asteroid Terrestrial-impact Last Alert System, or ATLAS, was built around the fireball: it looks for the bright short-lived flashes that incoming objects make, and it can catch objects only days out from arrival.

Two tricks make the search practical. One is how much sky a camera can cover per exposure. The other is how long the telescope can stare at that sky, because faint objects surface in longer exposures and a moving object smears across the frame if it shifts too far while the shutter is open.

That trade-off is why the field is not evenly watched. Object brightness, speed across the sky and how long a telescope can commit to one field all matter, and so does geography. Sky coverage from the southern hemisphere is thinner than from the north, and objects that arrive from directions the surveys are not watching tend to give shorter warning. Professional and amateur observers both flag this gap as the weakest part of the current system, and it is one of the reasons NASA is building a space-based survey.

What happens after a near-Earth object is discovered

First sighting is only a candidate. The discovery goes out for confirmation by other observatories, because a satellite streak, a reflection artefact or a software glitch can produce a convincing moving dot that is not a rock.

Once confirmed, the object gets a provisional designation: the year it was found, a letter and a number, written as something like 2026 AB12. A comet picks up a separate comet designation based on its discovery order that half-year. It is an identifier, not a name, and there is nothing ominous about it.

If an object is confirmed and later located well enough to earn a permanent number, that number can replace the designation. Track enough, and eventually you may see the object’s catalog name instead, in the tradition of naming near-Earth asteroids after observatories, survey teams and notable figures once a numbered object has had enough observations to be well characterized.

Then comes orbit improvement, and this is where most of the wall-clock time goes. The provisional orbit carries a large uncertainty region. Each new measurement from any telescope, anywhere in the world, gets fed back in and the error shrinks. Objects are often picked up repeatedly on later rounds of the same survey, which adds arc length and tightens the picture.

Once the orbit is well constrained, the object enters the catalogues that CNEOS screens for close approaches and potential impacts. Those catalogues are public, and anyone can check them.

How NASA calculates an asteroid’s orbit

The starting point is astrometry, which is simply the careful measurement of an object’s position on the sky at a specific time. Each telescope image yields a dot in a field of stars; the star positions are known from an earlier star catalogue, so the dot can be converted into a precise direction from Earth.

The hard part is the distance. A telescope sees a direction, not a depth, and an asteroid that appears tiny could be a small rock nearby or a large one far away. Radar resolves that ambiguity by timing the round-trip of a pulse, which gives range as well as direction. Until radar or a parallax measurement is available, orbit fitting has to treat distance as a range of possibilities and use the object’s observed motion across months to narrow it.

Those positions and times are then fitted to a Keplerian orbit: a shape, a size, an orientation, a tilt and where the object is along that path at a given moment. Six orbital elements describe a complete orbit, and from them you can compute where the object will be, at any time, to whatever precision the input measurements justify.

Orbit determination is iterative and probabilistic. The software proposes a candidate orbit, checks it against every observation, adjusts it, and repeats. What comes out is not a single line but a cloud of possible trajectories, and the honest statement about where an object will be in 2026 is a percentage probability across that cloud rather than a date and time.

Two factors decide how tight that cloud is. The length of the observation arc — how long scientists have been able to watch — matters enormously. A short arc leaves the object’s full path wide open, and some near-Earth objects travel paths that cross themselves, meaning a short arc can point to two or more possible paths that look nearly identical from Earth’s view. The second factor is observation quality, which is why radar follow-up on a passing object matters so much.

How radar helps NASA study near Earth objects

When an object gets reasonably close, ground-based radar is by far the most precise tool available. Planetary radar systems transmit a short pulse at an asteroid and listen for the echo. The delay gives distance; the strength and shape of the returned signal give size, surface roughness and rotation period.

Radar imaging can resolve features a few tens of metres across on an object that telescopes only show as a blob. During a close pass, the echo is processed into a set of delay-Doppler images, essentially a crude radar photograph, and those images are often the first genuine look at the surface of a small asteroid.

Lightcurve analysis fills in some of the same ground optically. A photometer measures the brightness of an object repeatedly as it tumbles, and the variation gives its shape and rotation rate without needing radar at all. Laser ranging is the rarest and most precise tool of the three, used on a very small number of targets.

Radar has an obvious limit: it is a ground-based, target-acquired technique. Someone has to point the dish at an object that is already known and already close enough to be worth the time. Radar cannot survey the sky. It refines what telescopes found, and for the great majority of tracked objects, no radar data ever exists.

How NASA assesses impact risk

Once an orbit exists, it is projected forward and compared with Earth’s position. The output is a close-approach table: the dates an object comes near, the distances in astronomical units, and in lunar distances, and the relative velocities at closest approach. A close approach is a flyby, and the word close is doing a lot of work in headlines.

For potential impacts, the same orbit runs through Sentry, the impact monitoring system CNEOS operates. Sentry automatically screens for objects that could hit Earth, including on horizons far longer than any news cycle, and publishes its risk tables openly.

Why an impact probability spikes and then falls back to zero

Because early orbits are uncertain, and an uncertain orbit looks like it might hit. When an object is first found, the smallest body that fits the observations could plausibly be on a collision course. The software does not hide that; it reports the probability, and on a genuinely ambiguous object the number can look alarming on day one.

Then more data arrives. Follow-up telescopes get a look, the object’s brightness improves, sometimes radar pins down the range, and the cloud of possible trajectories narrows to something that clearly misses. The probability collapses to zero or to a fraction of a percent, and the story drops off the news.

This has happened repeatedly with objects that became briefly famous, including 2024 YR4, which drew wide coverage precisely because its early probability looked bad before further observations moved it out of the way. Community astronomy and space forums ask about this pattern more than almost anything else in the field, and the answer is the same each time: the spike is a measurement artefact, not a change in the object.

What the Torino Scale actually means

The Torino Scale is a communication tool, from zero to ten, that sorts close approaches into four broad colour-coded bands. It is not a prediction machine, and it does not update itself. Each entry is assessed and published by the NASA expert team that exists for exactly this purpose, and it carries its own caveats about how well determined the orbit is.

Levels one and two are the normal state of business. They cover an object with a close approach where collision is very unlikely, and one where close approach is certain but collision is not expected. Level one in particular means, more or less, keep watching. Most published objects that get any attention at all sit there.

Levels three and above are where media attention should ramp up, and that is rare. Nothing on the scale means an impact has happened or is booked. A high rating is a signal about the current orbit knowledge and its uncertainty, not a countdown.

There is one more subtlety worth knowing about. Even when a probability is small rather than zero, Earth’s gravity can reshape an orbit slightly on a close pass, and small changes to an object’s velocity can move the resulting impact date by days. That is why long-horizon tables list the dates an impact could occur rather than a single date. Known near-Earth asteroids that pass inside the Moon’s orbit are watched with extra care partly because of this keyhole effect.

Why NASA tracks some objects more closely than others

Follow-up time goes where it changes an outcome. A small, faint object on a comfortable orbit gets a few observations a month. A large object whose path brings it inside the Moon’s orbit, or one whose orbit is poorly determined years out, gets rapid imaging and radar time wherever it is available.

Size drives the response because energy scales with it. A few-metre object is most likely to break up entirely in the atmosphere; a large one is not, and the consequences would be regional or worse. Distance drives it too, because a distant pass gives years of warning while a late-discovered close arrival gives days.

Orbital uncertainty is the underrated factor. A comfortable, well-known rock can be tracked by the survey that found it on its next scheduled sweep. An awkward object needs someone to point a big telescope at it on short notice, and those hours are contested between many objects every week.

None of this means most objects are dangerous. It means attention is finite and gets allocated by risk, and a quiet week in the news usually means the work is going normally rather than that something is being hidden.

Frequently Asked Questions

Key Takeaways

Telescopes photograph the same sky repeatedly and find the dots that move. Those positions become an orbit, the orbit is projected forward against Earth’s path, and Sentry at CNEOS recalculates the impact probability whenever new observations land.

When you meet a new asteroid story, three questions settle most of it. Was the object confirmed by more than one observatory? Is the probability from an early, uncertain orbit, or from a refined one? And does the report describe a close approach, which is a flyby, or an actual impact probability?

The public data behind all of this is open. Close-approach tables and Sentry risk pages are published by CNEOS, and the orbital elements behind them can be checked independently. That is worth doing before reposting a warning number, since the number you screenshot has usually changed by the time you hit publish.