A synthetic Centaur generation pipeline and other support for a big-data Centaur search

Abstract

We complete three supporting components of a search for Centaurs, a type of minor planet, in Pan-STARRS1 telescope data. First, we build a software pipeline which generates a high-fidelity synthetic Centaur population and injects it into Pan-STARRS1 data. It includes two non-trivial methods for speeding up the simulation which make the project’s computation tractable. The pipeline has such a high degree of accuracy that the observations it simulates are indistinguishable from real data. A synthetic population of 1,000,000 objects is used as a control in the survey for two purposes: to determine an optimal set of Centaur-searching parameters and to measure the observational and algorithmic biases intrinsic to the otherwise opaquely-complex Centaur search. The methods developed for the pipeline, as well as much of the code, are fully general to any minor planet search, and are well-suited to be used in archival or upcoming large-scale, all-sky surveys, especially LSST. Second, we develop and test techniques to reduce Pan-STARRS1 data by 94% before the search by removing detections that are likely astrophysical false positives (stars and galaxies) or spurious detections, another necessary step to make the survey computationally tractable. Finally, while testing different choices of parameters for the search, we identify a previously-undiscovered problem with Pytrax, the Minor Planet Center’s asteroid-finding software.