Gravitational-Wave Inference For Noise PSD Jumps Across Data Gaps

Abstract

Long-duration gravitational-wave (GW) inference inevitably encounters data gaps, which are often accompanied by abrupt, non-stationary changes in the detector noise power spectral density (PSD). Conventional practice typically analyzes the pre-gap and post-gap data separately, causing avoidable information loss and potentially degrading parameter inference. We build on Bayesian gap augmentation in the Wilson--Daubechies--Meyer (WDM) time--frequency domain and focus on a practically important fai…