The existence of Gravitational Waves from binary black holes is one of the most interesting predictions of General Relativity. These ripples in space-time should be visible to ground-based gravitational wave detectors worldwide in the next few years. One such detector, the Laser Interferometer Gravitational-wave Observatory (LIGO) is in the process of being upgraded to its Advanced sensitivity which should make gravitational wave detections routine. Even so, the signals that LIGO will detect will be faint compared to the detector noise, and so accurate waveform templates are crucial. In this thesis, we present a detailed analysis of the accuracy of hybrid gravitational waveforms. Hybrids are created by stitching a long post-Newtonian inspiral to the late inspiral, merger, and ringdown produced by numerical relativity simulations. We begin our investigation with a study of the systematic errors in the numerical waveform, and errors due to hybridization and choice of detector noise. For current NR waveforms, the largest source of error comes from the unknown high-order terms in the post-Newtonian waveform, which we first explore for equal-mass, non-spinning binaries, and also for unequal-mass, non-spinning binaries. We then consider the potential reduction in hybrid errors if these higher-order terms were known. Finally, we investigate the possibility of using hybrid waveforms as a detection template bank and integrating NR+PN hybrids into the LIGO detection pipeline.

The existence of Gravitational Waves from binary black holes is one of the most interesting predictions of General Relativity. These ripples in space-time should be visible to ground-based gravitational wave detectors worldwide in the next few years. One such detector, the Laser Interferometer Gravitational-wave Observatory (LIGO) is in the process of being upgraded to its Advanced sensitivity which should make gravitational wave detections routine. Even so, the signals that LIGO will detect will be faint compared to the detector noise, and so accurate waveform templates are crucial. In this thesis, we present a detailed analysis of the accuracy of hybrid gravitational waveforms. Hybrids are created by stitching a long post-Newtonian inspiral to the late inspiral, merger, and ringdown produced by numerical relativity simulations. We begin our investigation with a study of the systematic errors in the numerical waveform, and errors due to hybridization and choice of detector noise. For current NR waveforms, the largest source of error comes from the unknown high-order terms in the post-Newtonian waveform, which we first explore for equal-mass, non-spinning binaries, and also for unequal-mass, non-spinning binaries. We then consider the potential reduction in hybrid errors if these higher-order terms were known. Finally, we investigate the possibility of using hybrid waveforms as a detection template bank and integrating NR+PN hybrids into the LIGO detection pipeline.