Waveform systematics in identifying strongly gravitationally lensed gravitational waves: Posterior overlap method

TL;DR

This study investigates how waveform model choices affect the detection of strongly lensed gravitational wave events, ensuring robustness of statistical tests amid waveform uncertainties using the posterior overlap method.

Contribution

It provides the first systematic analysis of waveform systematics in identifying lensed GW candidates, comparing Bayes factors across multiple waveform models in GW data.

Findings

Waveform choice causes significant variation in Bayes factors for some event pairs.

Likely no missed lensed events due to waveform choice from O1 to O3.

Parameter estimation with additional waveforms helps understand differences in results.

Abstract

Gravitational lensing has been extensively observed for electromagnetic signals, but not yet for gravitational waves (GWs). Detecting lensed GWs will have many astrophysical and cosmological applications, and becomes more feasible as the sensitivity of the ground-based detectors improves. One of the missing ingredients to robustly identify lensed GWs is to ensure that the statistical tests used are robust under the choice of underlying waveform models. We present the first systematic study of possible waveform systematics in identifying candidates for strongly lensed GW event pairs, focusing on the posterior overlap method. To this end, we compare Bayes factors from all posteriors using different waveforms included in GWTC data releases from the first three observing runs (O1-O3). We find that waveform choice yields a wide spread of Bayes factors in some cases. However, it is likely that no event pairs from O1 to O3 were missed due to waveform choice. We also perform parameter estimation with additional waveforms for interesting cases, to understand the observed differences. We also briefly explore if computing the overlap from different runs for the same event can be a useful metric for waveform systematics or sampler issues, independent of the lensing scenario.