Surface Stability Predictions of Sublimating Ice-Covered Worlds via the Optimization of Surrogate Neural Networks: Application to Europa

Abstract

The morphology of the surface ice on Europa remains unknown at the meter scale. Hobley et al. (2018, https://doi.org/10.1038/s41561-018-0235-0) suggested the possible presence of penitentes—bladed ice structures that form on Earth—on the surface of Europa, based on analogy to Earth. Characterized as instability features, penitentes grow due to differential sublimation erosion and ablation of a snowfield. Previous studies have provided contrasting results when examining the stability of the terrain on Europa from isolated radiative heat transfer and molecular transport mechanisms. We analyze these mechanisms simultaneously, leveraging Physics-Informed Neural Networks to identify physical conditions and suitable terrains that could potentially host centimeter-to meter-scale ridges and penitentes on Europa. We find that for the specific condition of a pre-defined sinusoidal ice morphology, values of ice thermal inertia near the lower limit of physically plausible values for low density, pure water ice may lead to the growth of features such as penitentes. Conversely, higher values of thermal inertia may lead to erosion of morphologies and convergence toward a flat surface. Our results can help inform the interpretation of NASA Europa Clipper and ESA Jupiter Icy Moons Explorer (JUICE) observations, and help guide landing site selection for future missions to the surface of Europa.