2014ApJ...789..156M

We present a model for merger-driven evolution of the mass function for massive galaxies and their central supermassive black holes at late times. We discuss the current observational evidence in favor of merger-driven massive galaxy evolution during this epoch, and demonstrate that the observed evolution of the mass function can be reproduced by evolving an initial mass function under the assumption of negligible star formation. We calculate the stochastic gravitational wave signal from the resulting black hole binary mergers in the low redshift universe (z ≤ 1) implied by this model, and find that this population has a signal-to-noise ratio 2xto 5xlarger than previous estimates for pulsar timing arrays, with a {2σ, 3σ} lower limit within this model of h_c(f = 1/yr) = {1.1x10^–15, 6.8x10^–16}. The strength of this signal is sufficient to make it detectable with high probability under conservative assumptions within the next several years. A principle reason that this result is larger than previous estimates is our use of a recent recalibration of the black hole-stellar mass correlation for the brightest cluster galaxies, which increases our estimate by a factor of ∼2 relative to past results. For cases where a galaxy merger fails to lead to a black hole merger, we estimate the probability for a given number of satellite black holes to remain within a massive host galaxy, and interpret the result in light of ULX observations. We find that in rare cases, wandering supermassive black holes may be bright enough to appear as ULXs.