A Taxonomic Study of Asteroid Families from KMTNet-SAAO Multi-band Photometry

TL;DR

This study uses multi-band photometry and machine learning to classify over 2000 main-belt asteroids into taxonomic types, analyzing family distributions and homogeneity, and exploring implications for asteroid origins.

Contribution

It introduces a probabilistic taxonomy method combining color data with machine learning, applied to a large asteroid sample, revealing family compositions and potential parent body characteristics.

Findings

>85% of targets classified into main asteroid complexes

High taxonomic homogeneity in some families like Themis and Koronis

No correlation between rotation periods and family membership

Abstract

We present here multi-band photometry for over 2000 Main-belt asteroids. For each target we report the probabilistic taxonomy using the measured V-R and V-I colors in combination with a machine-learning generated decision surface in color-color space. Through this method we classify >85% of our targets as one the four main Bus-DeMeo complexes: S-, C-, X-, or D-type. Roughly one third of our targets have a known associated dynamic family with 69 families represented in our data. Within uncertainty our results show no discernible difference in taxonomic distribution between family members and non-family members. Nine of the 69 families represented in our observed sample had 20 or more members present and therefore we investigate the taxonomy of these families in more detail and find excellent agreement with literature. Out of these 9 well-sampled families, our data show that the Themis, Koronis, Hygiea, Massalia, and Eunomia families display a high degree of taxonomic homogeneity and that the Vesta, Flora, Nysa-Polana, and Eos families show a significant level of mixture in taxonomies. Using the taxonomic purity and the degree of dispersion in observed colors for each of the 9 well-sampled collisional families we also speculate which of these families potentially originated from a differentiated parent body and/or is a family with a possible undetermined nested family. Additionally, we obtained sufficient photometric data for 433 of our targets to extract reliable rotation periods and observe no obvious correlation between rotation properties and family membership.