The coalescence of massive black holes is one of the primary sources of gravitational waves (GWs) for LISA. Measurements of the GWs can localize the source on the sky to an ellipse with a major axis of a few tens of arcminutes to a few degrees, depen
ding on source redshift, and a minor axis which is 2--4 times smaller. The distance (and thus an approximate redshift) can be determined to better than a per cent for the closest sources we consider, although weak lensing degrades this performance. It will be of great interest to search this three-dimensional `pixel for an electromagnetic counterpart to the GW event. The presence of a counterpart allows unique studies which combine electromagnetic and GW information, especially if the counterpart is found prior to final merger of the holes. To understand the feasibility of early counterpart detection, we calculate the evolution of the GW pixel with time. We find that the greatest improvement in pixel size occurs in the final day before merger, when spin precession effects are maximal. The source can be localized to within 10 square degrees as early as a month before merger at $z = 1$; for higher redshifts, this accuracy is only possible in the last few days.
We investigate the capability of LISA to measure the sky position of equal-mass, nonspinning black hole binaries, combining for the first time the entire inspiral-merger-ringdown signal, the effect of the LISA orbits, and the complete three-channel L
ISA response. We consider an ensemble of systems near the peak of LISAs sensitivity band, with total rest mass of 2times10^6 Modot, a redshift of z = 1, and randomly chosen orientations and sky positions. We find median sky localization errors of approximately sim3 arcminutes. This is comparable to the field of view of powerful electromagnetic telescopes, such as the James Webb Space Telescope, that could be used to search for electromagnetic signals associated with merging massive black holes. We investigate the way in which parameter errors decrease with measurement time, focusing specifically on the additional information provided during the merger-ringdown segment of the signal. We find that this information improves all parameter estimates directly, rather than through diminishing correlations with any subset of well- determined parameters. Although we have employed the baseline LISA design for this study, many of our conclusions regarding the information provided by mergers will be applicable to alternative mission designs as well.
