Sky marginalization in black hole spectroscopy and tests of the area theorem

Direct observation of gravitational waves from binary black hole (BBH) mergers has made it possible to test the laws of black hole thermodynamics using real astrophysical sources. These tests rely on accurate and unbiased parameter estimates from the pre and postmerger portions of a signal. Due to numerical complications, previous analyses have fixed the sky location and coalescence time when independently estimating the parameters of the pre and postmerger signal. Here we overcome the numerical complications and present a novel method of marginalizing over sky location and coalescence time. Doing so, we find that it is not possible to model only the pre or postmerger portions of the signal while marginalizing over timing uncertainty. We surmount this problem by simultaneously yet independently modeling the pre and postmerger signal, with only the sky location and coalescence time being shared between the models. This allows us to marginalize over all parameters. We use our method to measure the change in area $\Delta A_{\rm measured} = A_f - A_i$ between the final and initial black holes in the BBH merger GW150914. To measure the final black hole's area $A_f$ we do an analysis using quasinormal modes (QNMs) to model the postmerger signal, and another analysis using the postmerger portion of an inspiral-merger-ringdown (IMR) template. We find excellent agreement with expectations from general relativity. The Hawking area theorem (which states that $A_f \geq A_i$) is confirmed to $95.4\%$ and $99.5\%$ confidence using the QNM and IMR postmerger models, respectively. Both models yield $\Delta A_{\rm measured} / \Delta A_{\rm expected} \sim 1$, where $\Delta A_{\rm expected}$ is the expected change in area derived from fits to numerical relativity simulations.