2025A&A...699A.273H

Context. Open clusters have long been used as tracers of Galactic structure. However, without a selection function to describe the completeness of the cluster census, it is difficult to quantitatively interpret their distribution.
Aims. We create a method to empirically determine the selection function of a Galactic cluster catalogue. We test it by investigating the completeness of the cluster census in the outer Milky Way, where old and young clusters exhibit different spatial distributions.
Methods. We develop a method to generate realistic mock clusters as a function of their parameters, in addition to accounting for Gaia's selection function and astrometric errors. We then inject mock clusters into Gaia DR3 data, and attempt to recover them in a blind search using HDBSCAN.
Results. We find that the main parameters influencing cluster detectability are mass, extinction, and distance. Age also plays an important role, making older clusters harder to detect due to their fainter luminosity function. High proper motions also improve detectability. After correcting for these selection effects, we find that old clusters are 2.97 ± 0.11 times more common at a Galactocentric radius of 13 kpc than in the solar neighbourhood - despite positive detection biases in their favour, such as hotter orbits or a higher scale height.
Conclusions. The larger fraction of older clusters in the outer Galaxy cannot be explained by an observational bias, and must be a physical property of the Milky Way: young outer-disc clusters are not forming in the outer Galaxy, or at least not with sufficient masses to be identified as clusters in Gaia DR3. We predict that in this region, more old clusters than young ones remain to be discovered. The current presence of old, massive outer-disc clusters could be explained by radial heating and migration, or alternatively by a lower cluster destruction rate in the anticentre.