Complexity analysis of gravitational waves

Abstract

The recent observations of gravitational waves during the coalescence of black holes has confirmed various predictions of general relativity. However, some concerns have been raised that these putative waves are artifacts. We have analyzed the ten confident gravitational waves detected in Observation Run 1 (O1) and Observation Run 2 (O2) and the gravitational wave emitted by the single known example of a neutron star coalescence, all of which have been verified by the Advanced Laser Interferometer Gravitational Wave Observatory and/or the Advanced Virgo Collaboration. On three occasions both Advanced LIGO and Advanced Virgo detected the same event (Abbott et al., 2019).Whitening of the Hanford data (Gravitational Wave Open Science Center) was employed in this study to reduce colored non-transient noise, but template matching was not employed. In each of the ten events mentioned above we verified the presence of such waves by the application of non-stationary data analysis via the Morlet wavelet. We examined the strain data before, during and after each wave. In each case we employed seven measures of complexity or information before, and during each wave and analyzed these data using doubly multivariate MANOVA, the appropriate statistical test for repeated measures (pre-wave vs wave). In addition we compared the strain data before and after the putative event to see if there was persistence of any effect of the event. In each case the information content was significantly elevated during the wave compared to similar time samples taken before the wave. No persistent effects on the strain data were noted after the event. We conclude our analysis is consistent with the actual detection of gravitational waves by the LIGO and Virgo Collaboration. Our results suggest black hole coalescence generates mathematically complex gravitational waves which may elucidate aspects of such coalescence.