Efficient Detection-level Linking of Trans-Neptunian Objects via a Spherical-orbit Basis

Abstract

Discovering solar system objects in time-domain survey data requires linking: determining which detections observed at different times were produced by the same moving object. In wide-field surveys, this association problem is challenging because detection catalogs are dominated by artifacts, variable sources, and other unrelated transients, creating a large combinatorial search over possible links. The problem is compounded when the cadence is sparse or the objects are faint, such that detections of a given object are not guaranteed to form same-night tracklets. We present an algorithm for linking single-opposition detections of solar system objects, such as trans-Neptunian objects (TNOs), that operates directly on individual detections rather than first forming tracklets. The method searches over trial orbits in a spherical-orbit parameterization, propagates detections to a common epoch, and identifies candidate objects as coherent clusters in the transformed coordinates. We demonstrate the pipeline on 12 months of Dark Energy Survey (DES) Supernova X3 field observations, a region with Legacy Survey of Space and Time (LSST)–like cadence. The method achieves a linking efficiency of better than 98% on implanted synthetic objects and successfully recovers all nine known linkable TNOs in the field, while discovering 17 new TNOs.