[Paper Review] Data formats for numerical relativity waves

P. Ajith, M. Boyle+12 more|ORCA Online Research @Cardiff (Cardiff University)|Sep 2, 2007

Pulsars and Gravitational Waves Research22 citations

TL;DR

This paper proposes a standardized, extensible data format for exchanging numerical relativity waveforms, particularly gravitational waveforms from compact binary systems, to enhance collaboration between numerical relativity and gravitational-wave data analysis communities. It defines a human-readable metadata file with key-value pairs and a structured data format for time-series waveforms, enabling seamless integration with LVC software tools and supporting future extensions for neutron star binaries and stellar collapse simulations.

ABSTRACT

This document proposes data formats to exchange numerical relativity results, in particular gravitational waveforms. The primary goal is to further the interaction between gravitational-wave source modeling groups and the gravitational-wave data-analysis community. We present a simple and extendable format which is applicable to various kinds of gravitational wave sources including binaries of compact objects and systems undergoing gravitational collapse, but is nevertheless sufficiently general to be useful for other purposes.

Motivation & Objective

• To establish a common data format for sharing numerical relativity results, especially gravitational waveforms, across research groups.
• To improve collaboration between numerical relativity researchers and gravitational-wave data analysts by enabling seamless integration of NR waveforms into data analysis pipelines.
• To define a flexible, extensible, and human-readable metadata format that captures essential simulation parameters and provenance.
• To ensure compatibility with existing LIGO/Virgo Collaboration (LVC) data analysis software, particularly the LSC Algorithms Library in ANSI C99.
• To support future extensions for diverse sources such as neutron star binaries and stellar gravitational collapse.

Proposed method

• Define a metadata file in plain text using key = value pairs organized in sections, storing simulation metadata such as mass ratio, spin parameters, code used, authors, and publication links.
• Specify a three-column data format for waveforms: {t, h_+, h_x}, with time t in equally spaced intervals, suitable for time-domain analysis and LVC software processing.
• Use a post-Newtonian-inspired parameterization via v = (Mπf)^(1/3) to relate time-domain waveforms to frequency-domain amplitudes.
• Provide explicit expressions for waveform multipoles H_{ℓm} in terms of orbital phase Φ, mass ratio η, symmetric mass ratio, and spin parameters χ_s, χ_a, δ.
• Derive the Fourier-domain amplitude via the stationary phase approximation: ~A_ℓm = A_ℓm * sqrt(2π / (m * d²Φ/dt²)) = A_ℓm * sqrt(2πM / (3m v² * dv/dt)).
• Include spin-dependent contributions to waveforms using equations from Will & Wiseman (1996) and Blanchet et al. (2008), with proper normalization and sign corrections.

Experimental results

Research questions

• RQ1 How can numerical relativity waveforms be consistently and unambiguously shared across different research groups and data analysis pipelines?
• RQ2 What metadata structure is both human-readable and machine-processable for describing NR simulation parameters and provenance?
• RQ3 How can time-series gravitational wave data be formatted to ensure compatibility with the LSC Algorithms Library and LVC data analysis tools?
• RQ4 What is the minimal yet sufficient set of physical parameters needed to reconstruct waveforms from metadata and data files?
• RQ5 How can the format be extended to support non-spinning, spinning, and non-binary sources such as neutron star binaries and stellar collapse?

Key findings

• The proposed metadata format uses simple key = value pairs in sections, enabling clear documentation of simulation parameters, authors, codes, and publications.
• The waveform data format uses three columns: time t, h_+, and h_x, with time equally spaced, ensuring compatibility with standard time-series processing tools.
• The format supports the inclusion of spin effects through explicit expressions for H_{2,2}, H_{2,1}, and H_{3,2} multipoles in terms of v, η, χ_s, χ_a, and δ.
• The Fourier-domain amplitude is derived via the stationary phase approximation, with ~A_ℓm = A_ℓm * sqrt(2πM / (3m v² * dv/dt)), enabling efficient template matching.
• The format is designed to be extensible, with clear guidelines for adding new data types such as black hole spin evolution or density profiles in collapse simulations.
• The format has been adopted for the NINJA project and is aligned with the LSC Algorithms Library, ensuring immediate usability in current LVC data analysis workflows.