Surrogate Models of Gravitational Waves from Numerical Relativity Simulations of Binary Black Hole Mergers

The advanced LIGO detectors have made multiple detections of gravitational waves from the mergers of binary black hole systems, bringing us into the era of gravitational wave astronomy. From such gravitational wave detections, we can put constraints on deviations from general relativity (GR), as well as measure the masses and spins of the black holes involved in the mergers. Such measurements require knowledge of the gravitational waveforms predicted by GR for all relevant masses and spins. Numerical relativity (NR) simulations are now sufficiently robust that we can accurately simulate binary black hole mergers and obtain the waveform for all but the most extreme parameters, but they are too computationally expensive for a dense coverage of the parameter space. NR surrogate models rapidly and accurately interpolate the waveforms from a set of NR simulations over a subset of parameter space. Using the Spectral Einstein Code (SpEC), we have built several NR surrogate models for various subsets of the parameter space, culminating in a model which includes all 7 intrinsic parameter dimensions. The surrogate model waveforms are nearly as accurate as NR waveforms, and can be evaluated in milliseconds whereas a single NR simulation can take weeks.