How are
asteroids discovered?

All ground-based asteroid surveys share the same core technique. A telescope photographs a region of sky. Twenty minutes or so later, it photographs the same region again. Then again. Planets, stars, and galaxies remain fixed relative to each other between exposures. An asteroid moves - shifted slightly in position by its own motion combined with Earth's. Software compares the frames and flags any moving source.

The flagged candidate goes through automated filters designed to reject satellite trails, cosmic ray hits, and optical artefacts before being passed to a human reviewer or a more sophisticated classifier. If it passes, the observation is submitted to the Minor Planet Center within hours. From that point, global networks of follow-up observers can be tasked to take additional measurements.

The technique itself is straightforward. The engineering challenge is doing it at scale - covering large fractions of the sky repeatedly, every clear night, while processing the resulting torrent of data fast enough to catch objects that may approach Earth within days.

A confirmed detection goes to the Minor Planet Center (MPC), the IAU body responsible for cataloguing all solar system small bodies. The MPC assigns a provisional designation based on the year and a letter-code system that encodes when in the year the discovery was made and its order within that period. For example, "2024 BX1" indicates an object found in the second half of January 2024.

The MPC circulates the discovery to follow-up networks. Other observatories take additional measurements over subsequent nights, extending the arc of data from which the orbit is computed. A longer data arc produces a tighter orbital solution - reducing the uncertainty in the asteroid's future position. Once the orbit is determined with sufficient precision and the object has been observed across multiple oppositions, it receives a permanent number and, eventually, a name.

Objects whose orbits fall within 1.3 AU of the Sun - making them near-Earth objects (NEOs) - are automatically assessed by NASA's automated Sentry system for potential close approaches with Earth.

Optical surveys detect asteroids; radar characterises them. Once an asteroid is identified and its approximate position is known, planetary radar can bounce a signal off its surface and time the return - measuring distance to within metres and velocity to within millimetres per second. This precision far exceeds what optical astrometry can achieve.

NASA's Goldstone Solar System Radar in the Mojave Desert, California, is the primary facility currently active for this work. The now-decommissioned Arecibo Observatory in Puerto Rico - until its collapse in 2020 - was the world's most powerful planetary radar and produced detailed shape models and surface maps for hundreds of near-Earth asteroids. Goldstone has filled much of that role since.

Radar observations are particularly valuable when an asteroid is newly discovered and its orbit carries large uncertainties. A single radar session during a close approach can tighten the orbital solution by orders of magnitude - sometimes eliminating a provisional Sentry risk entry entirely.

Ground-based surveys have a fundamental limitation: they cannot look too close to the Sun. Asteroids on orbits interior to Earth's - Atiras - and those approaching from the sunward direction are largely invisible to ground-based telescopes during the day. This was dramatically illustrated by the Chelyabinsk meteor of 2013, which arrived from the Sun's direction without any warning.

Space telescopes operating in the infrared have a further advantage: they detect asteroids by their thermal emission rather than reflected sunlight. This means they can find dark, low-reflectivity asteroids that optical surveys miss. NASA's WISE/NEOWISE mission operated for over a decade, detecting thousands of asteroids by their heat signatures before its decommissioning in 2024.

The successor mission is NEO Surveyor, a space telescope designed specifically for near-Earth object detection. Operating at the L1 Sun-Earth gravitational point, it will survey regions of the sky inaccessible to ground observatories and aims to discover 90% of all PHAs (potentially hazardous asteroids, meaning those 140 metres or wider with orbits passing within 0.05 AU of Earth's orbit) larger than 140 metres within a decade of operation.

Most asteroid surveys prioritise completeness over speed - cataloguing every object systematically, however slowly. ATLAS was designed with a different mission: detect objects approaching Earth from any direction and issue an alert as quickly as possible, even if that means warning arrives only days before a potential impact.

Operating from Hawaii, Chile, and South Africa, ATLAS covers the whole accessible sky every two nights at a depth sufficient to detect any object above roughly 50 metres in diameter within 1 LD (lunar distance, approximately 384,400 km) of Earth. Its multi-site design guards against cloud cover at any single location. The system has successfully detected objects just days before their closest approach - sufficient time for civil emergency planning even if not for deflection.

The Chelyabinsk event demonstrated the cost of gaps in coverage. A 20-metre object that reached the atmosphere undetected caused extensive glass damage across a large area of Russia and injured over 1,400 people. ATLAS and similar programmes represent a systematic response to that lesson.