Planetary defence: near-Earth objects, impact risk and deflection

At a glance

Planetary defence means finding the asteroids and comets that could hit Earth, estimating the odds of an impact and preparing to prevent it or limit the damage, from deflecting an object in space to evacuating an area on the ground.

Counts change daily: the NEOCC figures were last updated on 30 October 2026 (12:43 UTC) and NASA’s size breakdown in December 2025, and catalogues can differ by a few hundred objects. Objects spotted before impact: ESA’s past impactors list.

Sources: [1] [2]

What counts as a near-Earth object?

A near-Earth object (NEO) is an asteroid or comet with its perihelion within 1.3 au of the Sun; near-Earth comets must also have periods under 200 years. NASA’s Center for Near-Earth Object Studies (CNEOS) divides near-Earth asteroids into Atiras (orbits inside Earth’s), Atens and Apollos (Earth-crossers with semi-major axes smaller or larger than Earth’s) and Amors (approaching from outside). See also our small bodies page.

A potentially hazardous asteroid (PHA) comes within 0.05 au (about 7.5 million km) of Earth’s orbit and has an absolute magnitude H of 22.0 or brighter, equivalent to about 140 m for an assumed reflectivity of 14%. Darker objects of the same brightness are larger, so such sizes are uncertain. The label flags an orbit to watch, not an expected impact.

Sources: [3]

Finding them: the survey telescopes

Most NEOs are found by wide-field telescopes that image the same sky several times a night and look for moving points of light. NASA’s Near-Earth Object Observations programme funds the Catalina Sky Survey (University of Arizona) and Pan-STARRS and ATLAS (University of Hawaii). ATLAS scans the whole sky several times a night with four telescopes in Hawaii, Chile and South Africa, and in 2025 found the interstellar comet 3I/ATLAS (see interstellar objects). Caltech’s Zwicky Transient Facility has scanned the northern sky every two days since 2018.

Reports go to the IAU Minor Planet Center; CNEOS and ESA’s NEO Coordination Centre (NEOCC) compute orbits independently, and planetary radar refines them (see observatories and radio astronomy). Congress asked NASA in 2005 to find more than 90% of NEOs larger than 140 m; by December 2025 about 95% of near-Earth asteroids above 1 km were known, but fewer than half of those above 140 m.

The Vera C. Rubin Observatory in Chile officially began its ten-year Legacy Survey of Space and Time in 2026, after early operations had found more than 11,000 new asteroids, 33 of them NEOs, in about six weeks (NSF). NASA’s NEO Surveyor, a 50 cm infrared telescope designed to find at least two-thirds of NEOs larger than 140 m in five years, is due to launch no earlier than September 2027; the solar arrays on its sunshade finished testing in October 2026 (NASA).

Sources: [2] [4]

Size, frequency and consequence

An impactor’s size sets both how often such impacts happen and what they do (Figure 1). NASA estimates that 10 m objects, which make a very bright fireball and a strong sonic boom, strike about once a decade; 50 m objects, capable of local devastation, about every 1,000 years; and 1 km bodies, capable of global devastation, about every 700,000 years.

On 30 June 1908 an explosion over Tunguska, Siberia, flattened about 2,000 km² of forest; a 2019 NASA-led reassessment favoured a 50–80 m stony asteroid releasing 10–30 megatons 10–15 km up, and an average interval for such events of millennia rather than centuries (NASA). On 15 February 2013 an unknown asteroid about 19 m across burst over Chelyabinsk, Russia, with about 500 kilotons, injuring some 1,500 people, mostly with flying glass (ESA); Peter Brown and colleagues concluded that impacts by objects tens of metres across may be up to ten times more frequent than previously estimated. The Chicxulub impact of about 66 million years ago (University of Texas), by a body roughly 10 km across (Catalina Sky Survey), left a crater 180–200 km wide in Mexico and is linked to the extinction of more than 70% of species, including the dinosaurs.

Sources: [5] [6]

Interactive lab: impact energy and frequency

Set a diameter, density, speed and entry angle, or start from the Chelyabinsk values loaded below. The lab returns the mass, the energy in TNT equivalent, the average interval between such impacts and a rough crater size.

For Chelyabinsk (19 m, 3,300 kg/m³, 19 km/s, 18° above the horizon) it gives about 510 kilotons, within the measured 500 ± 100, and predicts an airburst. The Tunguska preset (50 m, 15 km/s) gives about 5 megatons; try 60–80 m at 15–18 km/s to reach the 10–30 megatons of NASA’s 2019 study. As the notes explain, atmospheric break-up is not modelled.

Sources: [6]

Measuring risk: Torino, Palermo and shifting odds

While an impact cannot be ruled out, NASA’s Sentry system and its ESA counterpart compute an impact probability for each possible date. The Torino scale (IAU, 1999) rates the hazard for the public from 0 to 10 by combining probability and energy: 0 (white) no hazard; 1 (green) routine; 2–4 (yellow) meriting astronomers’ attention, level 3 meaning at least a 1% chance of a collision capable of localized destruction; 5–7 (orange) threatening; 8–10 (red) certain collisions. Levels 1–4 are most likely reset to 0 by further observations. The logarithmic Palermo scale (Chesley and colleagues, 2002) compares the probability with the background chance of a similar impact before that date; values below −2 have no likely consequences, values above 0 merit concern (CNEOS).

Probabilities often rise before they fall: each observation shrinks an asteroid’s uncertainty region, so while Earth stays inside it covers a growing share, until Earth usually drops out and the probability falls to zero. ESA calls this rise an expected result of improving orbit knowledge (ESA).

Sources: [7]

2024 YR4: a case study

2024 YR4, 40–90 m across, was found on 27 December 2024 by ATLAS in Chile and rated Torino level 3 (ESA). On 27 January 2025 its chance of hitting Earth on 22 December 2032 reached 1.2%, above the 1% threshold for an IAWN notification. It peaked at 3.1% on 18 February 2025 (ESA’s own estimate peaked at 2.8%), fell below 1% on 21 February and reached 0.004% by 23 February. The chance of a lunar impact then rose to 4.3%, until JWST observations on 18 and 26 February 2026 ruled it out: the asteroid will pass about 21,200 km from the Moon’s surface. The episode rehearsed the whole warning chain, from discovery to an all-clear. See also our Moon and James Webb pages.

Sources: [8] [9]

Ways to deflect an asteroid

A threatening asteroid need not be destroyed, only made to reach the crossing point slightly early or late; because a small velocity change grows with time, warning time matters most. The EU-funded NEOShield project (2012–2015) studied the main concepts; only the first has been tested on an asteroid.

• Kinetic impactor. A spacecraft rams the asteroid and the recoil of crater debris adds to the push. Citing the US National Research Council, NEOShield judged it adequate for moderately sized objects, given decades of warning.
• Gravity tractor. A spacecraft hovering nearby tugs the asteroid with its own gravity for years, without touching it.
• Ion beam shepherd. The exhaust of an ion engine is aimed at the asteroid for a gentle, continuous push. NEOShield rated it, like the gravity tractor and laser ablation, potentially useful.
• Nuclear standoff explosion. A device detonated at a distance from the asteroid, studied only theoretically: it may be the only option when warning is very short, but military and political issues preclude optimised designs and space tests.
• Civil defence. For objects of about 50 m or less, evacuating the predicted impact area may be the best response.

Sources: [10]

DART and Hera: the first real test

NASA’s Double Asteroid Redirection Test (DART) targeted Dimorphos, a 151 m moonlet circling the 760–780 m asteroid Didymos every 11 h 55 min, about 1.19 km from its centre (Figure 2).

At 23:14 UTC on 26 September 2022 the 579 kg spacecraft hit Dimorphos at 6.1 km/s, within 25 m of its centre, watched by the Italian CubeSat LICIACube. Against a success threshold of 73 seconds, the period shrank by 32 ± 2 minutes (Johns Hopkins APL); a 2024 JPL analysis refined this to 33 minutes 15 seconds, with the orbit 37 m tighter and slightly elongated and Dimorphos reshaped (JPL).

Andrew Cheng and colleagues derived an along-track velocity drop of 2.70 ± 0.10 mm/s and a momentum enhancement factor β ≈ 3.6 for a density of 2,400 kg/m³ (2.2–4.9 over plausible densities): the ejecta recoil delivered more than twice the spacecraft’s own momentum. In 2026 a study led by Rahil Makadia used 22 stellar occultations to show that the pair’s 770-day orbit around the Sun had shortened by about 0.15 s, the first measurable change to a celestial body’s path around the Sun made by a human-made object (NASA).

ESA’s Hera, launched on 7 October 2024, flew past Mars in March 2025 (ESA) and braked in October 2026 to about 20 m/s relative to the asteroids, resolving both from 1,280 km. Detailed images of Dimorphos are due on 25 November 2026; Hera will then study the crash site, release two CubeSats and eventually attempt a landing. Dimorphos’s density, which sets β, remains an open question.

Sources: [11] [12] [13]

Interactive lab: nudging an asteroid

The calculator opens with DART’s values (580 kg at 6.14 km/s into a 151 m target of 2,400 kg/m³, β = 3.61); change the warning time to see how far the asteroid would shift along its orbit around the Sun.

It gives Δv ≈ 3.0 mm/s, slightly above the measured 2.70 mm/s because it assumes a perfectly along-track hit. Ten years ahead, that would shift a lone asteroid by about 2,800 km, less than half an Earth radius, so three such impacts would be needed. The 300 m presets need about ten impactors with ten years’ warning but about a hundred with one. The notes explain the drift rule and the “keyholes” a real deflection must avoid.

Sources: [12]

Apophis 2029: a rare close pass

(99942) Apophis, about 340 m across, briefly had a 2.7% chance of hitting Earth in 2029 after its discovery in 2004, the highest Torino rating then given (ESA). Later observations and radar measurements in March 2021 excluded any impact for at least 100 years. On 13 April 2029 it will pass less than 32,000 km above Earth’s surface, closer than geosynchronous satellites, visible to the naked eye from the Eastern Hemisphere; asteroids this large pass so close only every few thousand years.

ESA and JAXA’s Ramses, fully approved at ESA’s Ministerial Council in November 2025, is to launch on an H3 rocket in April 2028 and arrive in February 2029 (ESA); its spacecraft was first switched on in September 2026 (ESA). NASA’s OSIRIS-APEX (formerly OSIRIS-REx) will arrive shortly after the flyby. Earth’s tides may trigger quakes and landslides on Apophis and alter its spin, and will move it from the Aten to the Apollo group.

Sources: [14]

Who sounds the alarm? International coordination

Agencies such as NASA’s Planetary Defense Coordination Office and ESA’s NEOCC do the daily work. Two UN-endorsed groups coordinate internationally: the International Asteroid Warning Network (IAWN) issues warnings, and the Space Mission Planning Advisory Group (SMPAG), endorsed by the UN General Assembly in December 2013, brings space agencies together to prepare mitigation options and briefs the UN Committee on the Peaceful Uses of Outer Space (SMPAG).

IAWN warns when an object larger than about 10 m has an impact probability above 1%. SMPAG’s criteria start mission studies for objects above about 50 m with more than a 1% chance of impact within 50 years, and terrestrial preparedness for objects above about 20 m with more than 10% within 20 years (SMPAG). 2024 YR4 triggered the first formal IAWN notification; SMPAG judged it too early for mission activities (IAWN), and an all-clear followed on 24 February 2025.

What a warning would look like. For metre-sized objects, hours: 2024 XA1 fell in a remote part of Russia’s Sakha Republic about 10 hours after discovery (ESA). The ATLAS team expects days of warning for a 20 m object and weeks for a 100 m one. For a larger object found years ahead, the 2024 YR4 sequence would continue: orbit refinement, IAWN notification, characterisation, SMPAG mission studies and, if needed, deflection or civil defence along the predicted impact corridor.

Sources: [5] [8]

Current research (2024–2026)

• 2024 YR4 (2025–2026). Earth impact ruled out in February 2025, lunar impact in March 2026.
• DART and the Sun (March 2026). The binary’s orbital speed around the Sun changed by about 11.7 µm/s.
• Hera at Didymos (October–November 2026). Both asteroids resolved; detailed images of Dimorphos due on 25 November.
• Rubin (2026). The ten-year Legacy Survey of Space and Time has started.
• Impactors caught in time. Six of the thirteen were spotted between January 2024 and September 2026, the latest 2026 RW1 on 6 September.
• Next missions. NEO Surveyor (no earlier than September 2027), Ramses (April 2028) and OSIRIS-APEX (Apophis, 2029).

Sources: [4] [8] [9] [13]

How to go further: data and tools

CNEOS publishes the Sentry impact-risk table and discovery statistics; ESA’s NEOCC portal keeps its own risk list and the list of past impactors. The IAU Minor Planet Center holds the observations, the JPL Small-Body Database gives orbits with uncertainties, and IAWN posts its notifications. For the underlying physics, see our orbital mechanics guide and orbit lab.

Known vs open questions

Established

• About 95% of near-Earth asteroids larger than 1 km are known.
• Kinetic impact works: DART shortened Dimorphos’s orbit by about 33 minutes, and ejecta multiplied its push (β ≈ 2.2–4.9).
• Metre-sized impactors can be spotted hours ahead; 13 have been.
• Neither 2024 YR4 nor Apophis poses a significant impact risk to Earth over the coming century.

Open

• More than half of near-Earth asteroids larger than 140 m are still undiscovered.
• How often do Chelyabinsk- and Tunguska-class impacts really occur?
• Dimorphos’s density and internal structure, and how a kinetic impactor would work on a larger body.
• How quickly could a real deflection mission be launched, and who would decide?
• How will Earth’s tides change Apophis in 2029?

Related on this site

• Research dossier: Instruments and measurement
• Small bodies: asteroids, comets and meteorites
• Interstellar objects
• Launch vehicles and spaceports
• The Moon
• Space guide: Orbital mechanics
• Orbit lab