Surface age of the asteroids Itokawa and Eros by machine learning.

Introduction: Near-Earth asteroids (NEAs) like Itokawa and Eros provide invaluable insights into the collisional and surface evolution of the inner solar system. These bodies, classified as S-type asteroids, are particularly abundant in the inner asteroid belt and are believed to retain signatures of collisional events that have shaped their surfaces over time. Once exposed, these surfaces undergo space weathering (SW), a gradual alteration process caused primarily by solar wind irradiation and micrometeorite bombardment. Since the progression of SW is time-dependent, it offers a means of estimating the surface exposure age of an asteroid following a resurfacing event. Determining the SW age of NEAs like Itokawa and Eros can thus provide critical constraints on their geologic history, regolith dynamics, and collisional evolution.Aim: This study aims to estimate the surface exposure ages of Itokawa and Eros by analyzing their reflectance spectra using a machine learning approach that captures the spectral alterations due to SW.Methods: We employed an ensemble machine learning model, trained on laboratory reflectance spectra of irradiated silicate samples. These samples, comprising olivine, pyroxene, olivine-pyroxene mixtures, and chondritic meteorites, simulate the mineralogy typically found on S-type asteroid surfaces. Spectral data were sourced from the Reflectance Experiment Laboratory (RELAB), published literature, and direct contributions from authors.The model’s inputs included reflectance spectra and corresponding SW conditions (micrometeorite impact dose and solar wind flux), while the output was the exposure time at 1 AU. Itokawa's surface-resolved spectral data were acquired by the Near Infrared Spectrometer (NIRS) aboard the Hayabusa spacecraft (Abe et al. 2011). Eros data were obtained by the Near-Infrared Spectrometer (NIS) on the NEAR Shoemaker spacecraft (Warren et al. 1997). All asteroid data were retrieved from the NASA Planetary Data System (PDS). The spectra were interpolated from 820–2080 nm range at 20 nm intervals for Itokawa and 820–2360 nm, at 20 nm intervals for Eros, as required by the model. The model predictions were subsequently corrected for heliocentric distance to reflect actual surface ages.Results: The surface age estimates for Itokawa range from approximately 2 × 103 to 2 × 109 years (Fig 1). Both solar wind irradiation and micrometeorite impacts contributed to surface alteration, though solar wind effects were found to be more dominant. In contrast, Eros shows evidence of much older surfaces, with estimated surface ages ranging from 4 × 10⁸ to 2 × 10⁹ years (Fig 2), largely driven by micrometeorite impacts. However, the spectral resolution of the Eros dataset was notably lower than that of Itokawa, introducing greater uncertainty in the model’s predictions for Eros.Discussion and conclusion: The relatively young surface ages of Itokawa (2.8 × 103  years) align well with findings from the Hayabusa sample return mission, and other studies, which revealed that the dominating SW agent is solar wind irradiation (Hiroi et al. 2006, Matsumoto et al. 2016, Keller et al. 2016, Burges et al. 2021, Sunho et al. 2022). This supports the hypothesis that Itokawa has undergone frequent resurfacing events, possibly due to regolith migration (Miyamoto eta al. 2007), tidal resurfacing (Binzel et al. 2010), or by seismic shaking (Tsuchiyama et al. 2011). However, our study also reveals that certain regions of Itokawa (e.g., Arcoona) exhibit mature surface ages, dominated by micrometeorite impacts, reaching up to 2 × 10⁹ years, suggesting localized areas of minimal resurfacing and long-term exposure. The mature surface ages for Eros, consistent with previous spectral studies and surface morphology analyses, suggest a mature surface (Mahlke et al. 2022, Korda et al. 2023). Eros appears to lack significant recent resurfacing activity, allowing prolonged micrometeorite bombardment to dominate its SW history. Relatively younger ages can be observed in crater regions. This may be due to the material movement on the crater slope (Mantz et al. 2004) or it may reflect the cratering event age.The contrasting surface ages and dominant SW processes between Itokawa and Eros underscore the role of asteroid size, orbital dynamics, and regolith properties in governing SW rates. While smaller asteroids like Itokawa are dynamically active and frequently refreshed, larger bodies like Eros tend to preserve ancient surface features. Additionally, Itokawa’s contact binary origin may contribute to more recent reshaping events, which could explain the presence of localized areas with younger surface ages. Our results demonstrate the utility of machine learning in decoding the complex interplay between spectral alteration and exposure history. Fig 1. Predicted surface age for asteroid Itokawa. Fig 2. Predicted surface age for asteroid Eros. Abe et al. 2011 Data set informationBinzel et al. 2010 DOI 10.1038/nature08709Burges et al. 2021 DOI 10.1111/maps.13692Hiroi et al. 2006 DOI 10.1038/nature05073Keller et al. 2016 https://ntrs.nasa.gov/citations/20160002375Korda et al. 2023 DOI 10.1051/0004-6361/202346290Mahlke et al. 2022 DOI 10.1051/0004-6361/202243587Matsumoto et al.2016 DOI 10.1016/j.icarus.2015.05.001Miyamoto eta al. 2007 DOI 10.1126/science.1134390Sunho et al. 2022 DOI 10.1051/0004-6361/202244326Tsuchiyama et al. 2011 DOI 10.1126/science.1207794Warren et al. 1997 DOI 10.1023/A:1005015719887