Localizing coalescing massive black hole binaries with gravitational waves

Massive black hole binary coalescences are prime targets for space-based gravitational wave (GW) observatories such as LISA. GW measurements can localize the position of a coalescing binary on the sky to an ellipse with a major axis of a few tens of arcminutes to a few degrees, depending on source redshift, and a minor axis which is 2-4 times smaller. Neglecting weak gravitational lensing, the GWs would also determine the source's luminosity distance to better than percent accuracy for close sources, degrading to several percent for more distant sources. Weak lensing cannot, in fact, be neglected and is expected to limit the accuracy with which distances can be fixed to errors no less than a few percent. Assuming a well-measured cosmology, the source's redshift could be inferred with similar accuracy. GWs alone can thus pinpoint a binary to a three-dimensional pixel'' which can help guide searches for the hosts of these events. We examine the time evolution of this pixel, studying it at merger and at several intervals before merger. One day before merger, the major axis of the error ellipse is typically larger than its final value by a factor of similar to 1.5-6. The minor axis is larger by a factor of similar to 2-9, and, neglecting lensing, the error in the luminosity distance is larger by a factor of similar to 1.5-7. This large change over a short period of time is due to spin-induced precession, which is strongest in the final days before merger. The evolution is slower as we go back further in time. For z = 1, we find that GWs will localize a coalescing binary to within similar to 10 deg(2) as early as a month prior to merger and determine distance (and hence redshift) to several percent.