Title: Ordering the confusion: A study of the impact of lens models on gravitational-wave lensing detection capabilities

Abstract: When traveling from their source to the observer, gravitational waves can get deflected by massive objects along their travel path. When the lens is massive enough and the source aligns closely with the line-of-sight to the lens, the wave undergoes strong lensing, leading to several images with the same frequency evolution. These images are separated in time, magnified, and can undergo an overall phase shift. Searches for strongly-lensed gravitational waves are already ongoing. In essence, such searches look for events originating from the same sky location and having the same detector frame parameters. However, the agreement between these quantities can also happen by chance, leading to an important confusion background. To reduce the overlap between background and foreground, one can include lensing models, enforcing the lensing parameters to also be consistent with our expectations. In principle, such models should decrease the confusion background. However, when doing realistic searches, one does not know which model is the correct one. The use of an incorrect model could lead to the non-detection of genuinely lensed events. In this work, we investigate under realistic conditions how one can identify lensed events in the unlensed background. We focus on the impact of the addition of a model for the lens density profile and investigate the effect of potential errors in the modelling. We show that it is extremely difficult to identify lensing confidently without the addition of a lens model. We also show that slight variations in the profile of the lens model are tolerable but a model with incorrect assumptions about the underlying lens population causes significant errors compared to assuming no lens model at all. We also suggest some strategies to improve the confidence in the detection of strong lensing for gravitational waves.