Diamonds in the Rough: Characterizing the Effect of Noise Transients on Gravitational-Wave Data

During its first two observing runs, LIGO detected gravitational waves from the merging of binary black holes and neutron stars. The collaboration operates two detectors and has multiple methods used to locate signals in the detector data, but this presentation will focus on one search known as PyCBC, which identifies potential signals and produces triggers to flag them. While it finds signals, it also sometimes flags noise transients. The LIGO detector characterization group works to distinguish these transients, called glitches, from real signals. Glitches can be separated into categories, as is done in the GravitySpy web-based citizen science project. In my research, I analyzed the PyCBC triggers produced by GravitySpy glitch categories. I first compared glitches to real signal models. I then identified the astrophysical system that would produce a signal most closely resembling a glitch of a given type and plotted the characteristics of the systems to look for commonalities in the triggers produced by each glitch class. I lastly produced plots showing the probability that a certain glitch type will produce a trigger resembling a signal of certain characteristics. Such numbers will help LIGO determine whether a trigger is a glitch or a real signal.