Enhancing gravitational waveform models through dynamic calibration

TL;DR

This paper introduces a method to dynamically enhance gravitational waveform models by combining accurate numerical data with faster approximate models, improving their accuracy across parameter space for gravitational wave analysis.

Contribution

The study presents an automated approach to update and combine waveform models, enabling more accurate and comprehensive gravitational wave templates by integrating numerical relativity data.

Findings

Enriched basis models outperform previous models in accuracy.

Combining PhenomB and PhenomD models corrects previous shortcomings.

Method facilitates dynamic integration of numerical relativity data.

Abstract

Strategies to model the inspiral, merger and ringdown gravitational waveform of coalescing binaries are restricted in parameter space by the coverage of available numerical-relativity simulations. When more numerical waveforms become available, substantial efforts to manually (re-)calibrate models are required. The aim of this study is to overcome these limitations. We explore a method to combine the information of two waveform models: an accurate, but computationally expensive target model, and a fast but less accurate approximate model. In an automatic process we systematically update the basis representation of the approximate model using information from the target model and call the new model as the enriched basis. This new model can be evaluated anywhere in the parameter space jointly covered by either the approximate or target model, and the enriched basis model is considerably more accurate in regions where the sparse target signals were available. Here we show a proof-of-concept construction of signals from non-precessing, spinning black-hole binaries based on the phenomenological waveform family. We show that obvious shortcomings of the previous PhenomB being the approximate model in the region of unequal masses and unequal spins can be corrected by combining its basis with interpolated projection coefficients derived from the more recent and accurate PhenomD as the target model. Our success in building such a model constitutes an major step towards dynamically combining numerical relativity data and analytical waveform models in the computationally demanding analysis of LIGO and Virgo data.