A data-driven approach to multi-ring basin identification on Mercury

Multi-ring impact basins represent some of the oldest and most degraded large-scale structures on terrestrial planetary bodies, making their identification and characterization particularly challenging. Only a few well-preserved examples are known, such as the Orientale basin on the Moon, commonly regarded as the archetype of multi-ring basins. On Mercury, several multi-ring basins were initially proposed based on Mariner 10 imagery (Spudis & Guest, 1988); however, most of these candidates were not confirmed by subsequent analyses using MESSENGER data (e.g., Fassett et al., 2012; Orgel et al., 2020), highlighting the difficulty of recognizing ancient, highly modified basin architectures. Here we present a semi-automatic workflow aimed at the systematic characterization of multi-ring basins on Mercury. The workflow combines manual structural mapping with quantitative, data-driven analyses and consists of four main steps: (1) construction of a structural map of tectonic features; (2) determination of the basin center using concentric deviation analysis (Karagoz et al., 2024); (3) estimation of the multi-ring geometry through a newly developed tool that analyzes the radial distribution of mapped structures using one-dimensional kernel density estimation (KDE). In this step, dominant concentric rings are identified as statistically robust density maxima obtained with a Gaussian kernel and an objectively defined Silverman bandwidth, while ring uncertainty is quantified through the interquartile range (IQR) of associated structures; and (4) comparison of the inferred ring geometry with the basin’s median radial topographic profile, derived from 360 azimuthally distributed radial profiles, to assess geometric and morphological consistency. We apply this workflow to two basins of different confidence levels. For the Orientale basin on the Moon, the method identifies three concentric rings corresponding to the Inner Rook Ring, Outer Rook Ring, and Cordillera Ring, consistent with previous studies (Spudis et al., 2013). For the Andal–Coleridge basin on Mercury, a probable multi-ring basin, the workflow retrieves a four-ring geometry that broadly coincides with rings II–V proposed by Spudis & Guest (1988). These results demonstrate that the combined use of structural mapping, KDE-based ring detection, and radial profile analysis provides a robust and reproducible framework for investigating degraded multi-ring basins. Future work will apply this workflow to additional candidate basins on Mercury to reassess their multi-ring nature and improve constraints on the planet’s early impact and tectonic history.

Acknowledgements: We gratefully acknowledge funding from the Italian Space Agency (ASI) under ASI-INAF agreement 2024-18-HH.0.