A flexible, GPU-accelerated approach for the joint characterization of LISA instrumental noise and Stochastic Gravitational Wave Backgrounds
Alessandro Santini, Martina Muratore, Jonathan Gair, Olaf Hartwig · 2025
Read A flexible, GPU-accelerated approach for the joint characterization of LISA instrumental noise and Stochastic Gravitational Wave Backgrounds by Alessandro Santini, Martina Muratore, Jonathan Gair, Olaf Hartwig (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-06300.txt.
Introduction
Alessandro Santini, Martina Muratore, Jonathan Gair, and Olaf Hartwig build a flexible ear for LISA, the coming space gravitational wave observatory, in a 2025 paper. LISA must characterize its own instrument noise while hunting a faint stochastic background of gravitational waves. Assuming the exact spectral shapes in advance risks fooling the search.
Their method represents both noise and signal spectra with splines, letting the shapes bend as the data demand. A reversible jump algorithm fits the components together and decides how much flexibility each scenario needs. Graphics processors carry the computational load. Tests on simulated LISA data under varied assumptions showed how the added freedom affects recovery of signal and noise. The authors discuss when a detection claim would be credible. Letting the spectra speak, they argue, beats trusting templates in a brand-new instrument.
Catalog record
| CREATOR | Alessandro Santini, Martina Muratore, Jonathan Gair, Olaf Hartwig |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-06300
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A flexible, GPU-accelerated approach for the joint characterization of LISA instrumental noise and Stochastic Gravitational Wave Backgrounds Alessandro Santini, Martina Muratore, Jonathan Gair, Olaf Hartwig Abstract LISA data analysis represents one of the most challenging tasks ahead for gravitational-wave (GW) astronomy. Characterizing the instrument’s noise properties while fitting for all the other detectable sources is a key requirement of any robust inference pipeline. Noise estimation will also play a crucial role in searches and parameter estimation of cosmological and astrophysical stochastic signals. Previous studies have tackled this topic by assuming perfect knowledge of the spectral shape of the instrumental noise and of different possible types of GW Stochastic Backgrounds (SGWBs), usually resorting to parametrized templates. Recently, various works that employ template-agnostic methods have been presented. In this work, we take an additional step further, introducing flexible spectral shapes in both the instrumental noise and the stochastic signals. We account for the lack of knowledge of the exact shape of the individual contributions to the overall power spectral density by using splines to represent arbitrary perturbations of the noise and signal spectral densities. We implement a data-driven Reversible Jump MCMC algorithm to fit different components simultaneously and to infer the level of flexibility required under different scenarios. We test this approach on simulated LISA data produced under different assumptions. We investigate the impact of this increased flexibility on the reconstruction of both the injected signal and the noise level, and we discuss the prospects for claiming a successful SGWB detection. I Introduction The Laser Interferometer Space Antenna (LISA) Colpi et al. 2024 is a joint ESA-NASA space-based Gravitational-Wave (GW) detector that is expected to be launched in the mid-2030s. Operating in the millihertz region of the frequency spectrum, it will enable the observation of sources in a completely unexplored band, complementing ground-based GW detectors Abbott et al. 2019 ; Abbott et al. 2021 ; Abbott et al. 2024 ; Abbott et al. 2023 and Pulsar Timing Arrays Antoniadis et al. 2024 ; Afzal et al. 2023 ; Reardon et al. 2023 ; Xu et al. 2023 observations. Over the mission duration, LISA is expected to observe a wide range of individually resolvable sources, including massive black-hole binary (MBHB) mergers Klein et al. 2016 , Galactic white dwarf binaries (WDBs) Korol et al. 2020 , extreme mass ratio inspirals (EMRIs) Babak et al. 2017 , and stellar-mass black-hole binaries (SBHBs) in their early inspiral phase Gerosa et al. 2019 ; Buscicchio et al. 2021 ; Toubiana et al. 2022 . In addition, LISA will be sensitive to stochastic gravitational-wave backgrounds (SGWBs) of astrophysical or cosmological origin. Astrophysical stochastic backgrounds are generated by the incoherent superposition of GW signals from a large number of individual sources. Astrophysical populations that could produce detectable stochastic backgrounds for LISA include Galactic Karnesis et al. 2021 and extragalactic Farmer and Phinney 2003 WDBs, SBHBs Babak et al. 2023 and EMRIs Barack and Cutler 2004 ; Bonetti and Sesana 2020 ; Pozzoli et al. 2023 . The WDB background in particular is expected to dominate over the expected instrumental noise Team 2018 between \sim 0.5 and \sim 3 mHz. SGWBs in the LISA band could also be of cosmological origin Caprini 2015 . Possible cosmological sources include First Order Phase Transitions (FOPTs), which are expected in many theories that extend physics beyond the Standard Model Caprini et al. 2020 ; Caprini et al. 2019 ; Caprini et al. 2016 , Cosmic Strings Auclair et al. 2020 , primordial fluctuations in the early universe Ricciardone 2017 , and Primordial Black Holes (PBHs) Bartolo et al. 2019 ; Auclair et al. 2023 . The wide variety of sources populating the LISA band represents an exciting and unique opportunity to explore the GW Universe with unprecedented precision. However, it also poses significant data analysis challenges. Besides the instrumental noise and stochastic backgrounds, which will always be present in data we will receive from the instrument, many individually resolvable sources will be long-lived and will remain in band for months or years. This scenario will require a global analysis of the data, where the parameters of all the transient and stochastic components will have to be estimated simultaneously. The development of the necessary data analysis pipeline, referred to in the LISA community as the “global fit”, is currently an active area of research. Early implementations are described in Refs. Littenberg and Cornish 2023 ; Katz et al. 2024 ; Deng et al. 2025 ; Strub et al. 2024 . In this work, we focus on the task of simultaneously characterizing the instrumental noise and possible SGWBs present in the data. Previous studies have tackled this problem by using relatively simple parametrized templates with 2-12 parameters to describe the spectral shape of the instrumental noise and of the different SGWBs considered (see, e.g., Refs. Boileau et al. 2023 ; Adams and Cornish 2014 ; Hartwig et al. 2023 ). Nevertheless, we expect that the instrumental noise entering the LISA data streams will be significantly more complex and hard to fully model analytically. The LISA-Pathfinder (LPF) mission Armano et al. 2016 ; Armano et al. 2024 has already shown a similar behaviour, with a significant fraction of the total measured noise that could not be explained by on-board measurements of individual noise components. As the goal of LPF was to test the behavior of (part of) the key LISA hardware, we can already expect the same to happen for LISA, which is a significantly more complex system than LPF. Therefore, we must be prepared to account for uncertainties in the noise models we will ultimately use in data analysis pipelines, which will come at the cost of affecting the detection and characterisation of SGWBs Muratore et al. 2024 . Moreover, the different SGWBs described are expected to have different spectral shapes. Fitting them would require the inclusion of models for all of them in our pipelines, which is not only impractical but would also require a complete SGWB template bank. Since our knowledge of the exact spectral shape of the different SGWBs is limited, we cannot even compile a complete catalogue of all the possible sources and associated spectral shapes. For this reason, robust and flexible signal reconstruction algorithms are of pivotal importance to successfully extract the information carried by these persistent signals. Some previous works have explored possible solutions to the two problems. In Refs. Caprini et al. 2019 ; Flauger et al. 2021 , the authors propose an automated pipeline based on a binned power-law signal reconstruction, where the frequency band is divided into bins. In each bin, the data is fitted with a power-law signal plus a model of the instrumental noise. The latter is parametrized and common across all the bins. Conversely, in Ref. Baghi et al. 2023 , the authors propose a flexible model for the instrumental noise, where the spectral shape is described by a set of splines, maintaining a parametrized description of the SGWB signals. Finally, in Ref. Pozzoli et al. 2024 , the authors propose a flexible model for both the instrumental noise and the SGWB signals, where their weakly-parametric models rely on expectation values of Gaussian processes Rasmussen and Williams 2005 . In all these