Pulsar timing array analysis for black hole backgrounds

An astrophysical population of supermassive black hole binaries is thought to be the strongest source of gravitational waves in the frequency range covered by pulsar timing arrays (PTAs). A potential cause for concern is that the standard cross-correlation method used in PTA data analysis assumes that the signals are isotropically distributed and Gaussian random, while the signals from a black hole population are likely to be anisotropic and deterministic. Here we show that while the conventional analysis is not optimal for detecting signals from black hole binaries, the technique still works as the standard Hellings–Downs correlation curve turns out to hold for point sources. Moreover, the small effective number of signal samples blurs the distinction between Gaussian and deterministic signals. Possible improvements to the standard cross-correlation analysis that account for the anisotropy of the signal are discussed.