Navigating the noise: exploring black hole systems using time-series analysis
Navigating the noise: exploring black hole systems using time-series analysis
Understanding variability in accreting systems offers one of the most powerful tools for probing the innermost regions of compact objects — regions otherwise inaccessible to direct imaging. Variable signals encode information about everything from the geometry to the dynamics of the innermost flows. Time-domain astrophysics comes with immense potential, but it also brings many challenges. In this thesis, I aim to show the possibilities of time domain studies in equal measure to their caveats and pitfalls. I will demonstrate traditional techniques in the Fourier domain to brand new machine learning routines. Then I will show how no technique in this field is obsolete; they all build upon each other. This work focuses on publicly available survey data, illustrating the immense potential of what we already have and preparing us for what is to come. In a world where data is coming so fast, we might need to ask, can we keep up?
University of Southampton
Ward, Madeleine-Mai
1f668011-f593-4659-b2aa-f382e6fe4d5f
1 January 2026
Ward, Madeleine-Mai
1f668011-f593-4659-b2aa-f382e6fe4d5f
Middleton, Matthew
f91b89d9-fd2e-42ec-aa99-1249f08a52ad
Ward, Madeleine-Mai
(2026)
Navigating the noise: exploring black hole systems using time-series analysis.
University of Southampton, Doctoral Thesis, 259pp.
Record type:
Thesis
(Doctoral)
Abstract
Understanding variability in accreting systems offers one of the most powerful tools for probing the innermost regions of compact objects — regions otherwise inaccessible to direct imaging. Variable signals encode information about everything from the geometry to the dynamics of the innermost flows. Time-domain astrophysics comes with immense potential, but it also brings many challenges. In this thesis, I aim to show the possibilities of time domain studies in equal measure to their caveats and pitfalls. I will demonstrate traditional techniques in the Fourier domain to brand new machine learning routines. Then I will show how no technique in this field is obsolete; they all build upon each other. This work focuses on publicly available survey data, illustrating the immense potential of what we already have and preparing us for what is to come. In a world where data is coming so fast, we might need to ask, can we keep up?
Text
MWard_FinalSubmission_pdfA3
Text
Final-thesis-submission-Examination-Miss-Madeleine-Mai-Ward
Restricted to Repository staff only
More information
Published date: 1 January 2026
Identifiers
Local EPrints ID: 507879
URI: http://eprints.soton.ac.uk/id/eprint/507879
PURE UUID: 79e56edc-4bd9-46a3-821c-e027c34bd76d
Catalogue record
Date deposited: 07 Jan 2026 15:29
Last modified: 08 Jan 2026 03:10
Export record
Contributors
Author:
Madeleine-Mai Ward
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics