Numerical-relativity surrogate modeling with nearly extremal black-hole spins

Numerical relativity (NR) simulations of binary black hole (BBH) systems provide the most accurate gravitational wave predictions, but at a high computational cost -- especially when holes have nearly extremal spins (i.e. near theoretical upper limit) or very unequal masses. Recently, technique Reduced Order Modeling (ROM) has enabled construction surrogate models trained on an existing set NR waveforms. Surrogate enable rapid computation waves emitted by BBHs. Typically these are used for interpolation to compute waveforms BBHs with mass ratios and within bounds training set. Because so technically challenging, almost always rely sets only moderate spins. In this paper, we explore how well can extrapolate includes For simplicity, focus one-dimensional equal masses equal, aligned We assess performance higher spin magnitudes calculating mismatches between extrapolated model waveforms, differences measurements remnant black-hole mass, testing improves as extends find that while extrapolation in case is viable current detector sensitivities, next-generation detectors should use extend