Distal Rays and Secondary Craters of the Tycho Impact Event Revealed Through Deep Learning

Abstract

The nature of secondary impact craters is a matter of ongoing debate, with important implications for understanding the impact cratering process, and for calibration of crater chronology. Rays generated by primary impact craters are characterized by secondary craters, providing identifiable terrains to study the size, morphology and clustering properties of secondary impact processes. The prominent ray system of the young Tycho crater has been extensively studied on the Moon since the Apollo era. In this work, we study the entire crater population of a quadrangle that extends to the northeast of Tycho for ∼3,000 km. We relied on a bespoke application of YOLOLens, a state‐of‐the‐art deep learning model specifically optimized for high‐resolution crater detection. We generated a data set of ∼7.1 × 10 ⁶ craters more than 50 m in diameter when integrated with an existing crater database for completeness. We then used a spatial cluster analysis method to identify clustered (potential secondary) craters. Our analysis identifies rays and clusters associated with Tycho in agreement with previous maps. We identified several distal elongated clusters that belong to the Tycho ray system that have not been previously identified. Also, we interpret additional non‐radial clusters as plausible products of projectiles from Tycho ejected at high angles (∼74°–75°). Our results show that secondary craters <200 m dominate (∼71%) the crater population within the rays and clusters of Tycho. This work shows that the products from cluster analysis applied to large lunar crater data sets hold insightful information about the size and distribution of potential secondary craters.