Pluto’s Surface Mapping Using Unsupervised Learning from Near-infrared Observations of LEISA/Ralph

Abstract

We map the surface of Pluto using an unsupervised machine-learning technique using the near-infrared observations of the LEISA/Ralph instrument on board NASA’s New Horizons spacecraft. The principal-component-reduced Gaussian mixture model was implemented to investigate the geographic distribution of the surface units across the dwarf planet. We also present the likelihood of each surface unit at the image pixel level. Average I/F spectra of each unit were analyzed—in terms of the position and strengths of absorption bands of abundant volatiles such as N2, CH4, and CO and nonvolatile H2O—to connect the unit to surface composition, geology, and geographic location. The distribution of surface units shows a latitudinal pattern with distinct surface compositions of volatiles—consistent with the existing literature. However, previous mapping efforts were based primarily on compositional analysis using spectral indices (indicators) or implementation of complex radiative transfer models, which need (prior) expert knowledge, label data, or optical constants of representative end-members. We prove that an application of unsupervised learning in this instance renders a satisfactory result in mapping the spatial distribution of ice compositions without any prior information or label data. Thus, such an application is specifically advantageous for a planetary surface mapping when label data are poorly constrained or completely unknown, because an understanding of surface material distribution is vital for volatile transport modeling at the planetary scale. We emphasize that the unsupervised learning used in this study has wide applicability and can be expanded to other planetary bodies of the solar system for mapping surface material distribution.

Constraining Thermal Emission of Pluto’s Haze from Infrared Rotational Lightcurves

“…Due to limited information, we follow the approach of Lellouch et al (2011) and divide Pluto's surface into three component units based on the albedo map and assume the same emissivity within each unit. However, recent principal component analysis using the LEISA observations has revealed several more ice units (Schmitt et al 2017;Gabasova et al 2021;Emran et al 2023). Different mixtures of ice could have varying porosity and grain size, which would affect emissivity (Labed & Stoll 1991;.…”

“…We associate such composition units: dark H 2 O ice/tholin mix (unit 3), CH 4 ice of intermediate albedo (unit 2), and bright N 2 ice (unit 1) with three albedo units in Figure 1(b) based on the histogram of Pluto's updated albedo distribution from New Horizons: unit 3 of darkest ices (0-0.36), unit 2 of intermediate ices (0.36-0.82), and unit 1 of brightest ices (0.82-1.0). The divided units (Figure 1(c)) appear to represent an approximate distribution of icy components identified by LEISA (Grundy et al 2016;Schmitt et al 2017;Gabasova et al 2021;Emran et al 2023), although multiple ice mixtures show a much more complex pattern in the infrared LEISA maps. In contrast, Charon's surface (Figure 1(d)) is relatively uniform, and we use only one emissivity value.…”

Supervised and unsupervised learning of (1+1) -dimensional even-offspring branching annihilating random walks

“…Over the past decade, machine learning (ML) has shown great potential in physics research owing to its incomparable capabilities for complex data processing and feature extraction [1]. Various concepts regarding ML have been developed to solve vexing problems in high-energy physics [2,3], astronomy [4,5], quantum information [6,7],and molecular dynamics [8].…”

“…Supervised learning is mainly used to identify or classify the phases of matter [3,18], where the input data must be labeled. Unsupervised learning does not require input labels and is more suitable for clustering and reduction of dimensions [5,19,20]. However, the transfer learning method updates the algorithm structure in the existing experience and generalizes the scope of application of the model [21].…”

Discrepancy in Grain Size Estimation of H2O Ice in the Outer Solar System

“…Neptune's moon Triton, a trans-Neptunian object (TNO), is also assumed to host amorphous and crystalline water ice (Cruikshank et al 2000). Although the presence of crystalline H 2 O ice has been reported on Pluto's surface (Cook et al 2019;Emran et al 2023), its largest satellite Charon may host both amorphous and crystalline H 2 O phases (Dalle Ore et al 2018). Molecules of amorphous H 2 O ice have been detected on icy dust grains in dense interstellar clouds (e.g., Herbst 2001).…”