Abstract

Contributed Talk - Splinter EScience

Tuesday, 16 September 2025, 14:19

Quantifying Model-Observation Similarity: Implications for Stellar Flyby Models

Marco Bischoff, Prof. Dr. Susanne Pfalzner
Forschungszentrum Jülich

Trans-Neptunian Objects (TNOs) are among the most pristine bodies in the Solar System. Their orbital dynamics serve as crucial tests for competing hypotheses concerning the formation of the Solar System (e.g. the flyby model). By assessing how well different models reproduce observed TNO orbits, we hope to identify the most likely scenario that shaped the early Solar System. With a significant increase in TNO discoveries anticipated as new instrumentation such as the LSST comes online, automating these model comparison tests has become essential. However, astronomical data are subject to numerous detection biases, making direct comparisons between model predictions and observational data unreliable. To address this challenge, we have trained a classifier to estimate the likelihood of each observed or simulated object belonging to a hypothesised TNO population. We then compared the resulting groups using clustering metrics to provide an overall similarity score for each set of simulations and observations. In this talk, we present a modular data pipeline designed for performing these model tests. The code base is open source to promote transparency, and all pipeline components are readily exchangeable. Our underlying software framework offers solutions for dataset management, standardisation, and pipeline construction, and can be easily adapted to test various formation hypotheses.