When can we claim the detection of a quasinormal mode in a black hole ringdown?

How many quasinormal modes have we detected in black hole ringdown gravitational waves? How many will we detect in future events? To answer these questions, we need to clarify what we mean by a "detection" of a quasinormal mode. In this talk, I will discuss a list of criteria for claiming detection of a quasinormal mode, such as a high Bayes factor, the non-zero nature of the mode amplitude, and the self consistency of its complex amplitude over different time windows. The criteria on the list can be ordered by their strength, with quasinormal modes expected to pass stronger criteria when the signal-to-noise ratio is higher. I motivate the need for stronger and stronger criteria by showing that using only the weaker criteria could lead to false positive detections. I argue that there is not a simple clear cut definition of a "detection" because the number of criteria that are required to be met before a detection can be claimed will be an arbitrary choice. However, by checking how many of the detection criteria a quasinormal mode has satisfied, we can quantify the strength of a detection. I examine detection claims of different quasinormal modes in various gravitational wave events and clarify the criteria that have been (or have not been) met by each of them. Finally, I discuss how the list of criteria can be applied for forecasts or analyses of future gravitational wave events with a loud ringdown.