Asteroid Discovery is Outpacing Analysis

There are around 1.3 million known asteroids, and hundreds of new near-Earth asteroids are found every month as detection surveys improve. But finding an asteroid is only the first step. To judge whether one is hazardous, and how it might be deflected, we need its physical properties: rotation period, shape, size, and composition. Measuring those usually takes hours of dedicated telescope time per object, so analysis falls further behind discovery with every passing month. When the Vera C. Rubin Observatory (LSST) begins its survey, that gap will widen enormously. I built PhAst (Photometric Asteroid characterisation) to help close the gap.

What Does PhAst Do?

PhAst is an open-science pipeline that accelerates asteroid characterisation by combining two kinds of data that are normally kept apart:

• Sparse, serendipitous observations, where an asteroid happens to drift through a wide-field survey image. Each is a single data point, but there are enormous numbers of them: about 82,000 from Gaia Data Release 3 and about 63,000 from the Zwicky Transient Facility (ZTF).
• Dense, targeted observations, where a telescope follows one asteroid through a full rotation, from my own robotic-telescope proposals (Alnitak, AAVSO, Burke Gaffney, Faulkes), citizen scientists, and archives such as ALCDEF and the PDS Asteroid Photometric Data Catalog.

On their own, sparse survey points are too scattered to yield a rotation period, and dense light curves exist for only a small fraction of asteroids. PhAst merges the two so they reinforce each other, letting one pipeline characterise objects that neither dataset could handle alone.

Methodology: Six Steps

1. Identify and centroid known stars and asteroids in each image, using the Gaia DR3 star catalogue and NASA's HORIZONS ephemerides.
2. Choose the aperture that measures an asteroid's brightness with the least background noise (differential photometry).
3. Calibrate against four to seven stable comparison stars to compute the asteroid's magnitude.
4. Apply corrections for distance, phase angle, and light-travel time to build a clean, reduced light curve.
5. Merge the sparse survey photometry with the dense targeted light curves onto a common scale.
6. Fold the combined light curve at trial periods using congruent-modulo arithmetic, taking the period with the lowest root-mean-square error as the rotation (or mutual orbital) period.

I used high school mathematics to make PhAst accessible to students and citizen scientists, namely using weighted means, medians, standard deviations, modulo arithmetic, logarithms, and slopes, implemented in Python with NumPy, SciPy, AstroPy, and Matplotlib.

Test Application on Didymos: Target of NASA DART Mission

To test PhAst, I applied it to the Didymos-Dimorphos binary system, the target of NASA's DART planetary-defence mission. Merging ZTF survey photometry with citizen-scientist light curves from ALCDEF produced a phase curve giving an absolute magnitude H = 18.14, a geometric albedo of about 0.14, and a size of about 0.84 km for Didymos.

Folding the combined light curves at trial periods, the lowest root-mean-square error fell at 2.26 hours, the rotation period of Didymos. It did not change after the DART impact, exactly as expected, while the mutual orbital period of Dimorphos around Didymos, the direct signature of a successful deflection, decreased.

Population Studies: Space Mission Targets and Binary Asteroids

Didymos was a controlled test. The wider goal of PhAst is scale. By integrating around 140,000 sparse survey observations with dense data, I generated phase curves for about 2,100 asteroids, deriving each one's absolute magnitude, geometric albedo, size, and, where the data allowed, rotation period.

That sample includes asteroids that matter to space missions. Using only open survey and citizen-science data, PhAst characterised:

• Targets of NASA's Lucy mission: 3548 Eurybates (67 km, albedo 0.05) and 10253 Westerwald.
• Targets of the UAE's main-belt mission: 269 Justitia (46 km) and 15094 Polymele (27 km).
• Binary asteroids: 3378 Susanvictoria and 2825 Crosby, including their mutual orbital periods.
• Understudied asteroids: 1798 Watts, 7344 Summerfield, 2006 MG13, and 2007 AD11, several with little prior physical characterisation.

For most of these, PhAst recovered their size, albedo, rotation, and a rubble-pile structural classification. It showed that a citizen scientist can determine physical properties for a target asteroid from data that already exists, before a spacecraft arrives.

Population Study Results: What 2,100 Asteroids Reveal about the Early Solar System?

Geometric albedo is a proxy for composition: carbonaceous (C-type) asteroids are dark (albedo below 0.1), silicaceous (S) and metallic (M) types are moderate, and rare enstatite (E) types are bright. Applied across the full sample, this let PhAst map how composition varies with distance from the Sun.

The trend is clear: carbonaceous asteroids become more abundant with distance, from about 46 per cent in the inner belt (2.0 to 2.5 AU) to about 75 per cent in the outer belt (2.82 to 3.28 AU). Carbon-rich, volatile-bearing material dominates the cooler outer main belt, consistent with those bodies having formed farther from the young Sun. A tool built from open data and school-level mathematics can, at population scale, recover a genuine signature of how the solar system assembled.

How PhAst Can be Useful for Rubin/LSST

The Rubin Observatory will record billions of serendipitous asteroid observations, the great majority of them sparse. Methods that can only use dense, targeted light curves will characterise a tiny fraction of what Rubin sees. PhAst is a step towards the automated, survey-scale pipeline the field will need to keep characterisation from falling even further behind discovery.

I presented PhAst at the 2024 LA Planetary Defense and Asteroid Exploration mini-Conference, the AGU 2024 Fall Meeting, and the Rubin/LSST Community Workshop. Here is the talk:

▶ My presentation begins at 3:34:45. Open the video at that point on YouTube.

▶ My presentation begins at 1:10:39. Open the video at that point on YouTube.

References

• AGU 2024 abstract (NASA ADS): https://ui.adsabs.harvard.edu/abs/2024AGUFM.P01...09N/abstract
• Presentation slides, LA Planetary Defense mini-Conference 2024 (PDF): https://hotpoprobot.com/wp-content/uploads/2024/06/iiaa-1.pdf
• ISEF 2024 project PHYS035: https://isef.net/project/phys035-accelerated-asteroid-characterization-using-phast