Thursday

Development of the photometric catalogue, and all-sky surveys, is driven by increasing etendue (collecting area times field of view). Improving etendue, as with the Rubin Observatory and LSST, enables a ground-based survey to have a much fainter completeness limit, with significantly more efficient on-sky telescope time; however, Rubin will begin to challenge current observational techniques. Chief amongst these is the catalogue cross-match, the merging of two datasets to leverage increased wavelength coverage or dynamic range for scientific pursuit. The crowding of these incredibly faint observations, both in the Galactic plane but extra-galactic for longer wavelengths, mean that traditional methods for combining datasets will fail.

In this talk I will discuss the methodology we developed to overcome this, generalising what we dubbed the "Astrometric Uncertainty Function" -- the probability density function describing the likelihood of a sky object having some "true" sky coordinates given information on where we measured it to be. I will describe the necessary methodology behind our first major extension to the AUF, vital for WISE but also very important for LSST's crowded fields, in which blended objects perturb measured centroids of brighter sources in a photometric catalogue. These center-of-light tugs can overwhelm on-sky separations between objects in two catalogues, otherwise assumed to solely be controlled by the centroid uncertainty of the objects. This additional, but necessary, "perturbation" component to the AUF demands non-analytic solutions to questions like "what is the probability that these two sources are counterparts, given the separation between them", which has a huge impact on the computational efficiency of any potential cross-match scheme. Finally, I will briefly discuss our efforts to extend the AUF to include the effect of unknown proper motions, another important aspect of catalogue matching for LSST science due to a lack of Gaia proper motions at such faint magnitudes.