I'm a lecturer in cosmology at Queen Mary University of London, with interests in theory and observations of large-scale structure (especially with optical and radio telescopes), the CMB, and general relativistic effects.
I completed my DPhil (PhD) in Astrophysics at the University of Oxford in 2013, followed by a postdoc at the Institute of Theoretical Astrophysics in Oslo (til 2015), a NASA Postdoctoral Program (NPP) Fellowship at JPL/Caltech (til 2017), and a postdoc at UC Berkeley in the Radio Astronomy Lab and Berkeley Center for Cosmological Physics (til 2018). My undergrad was at the University of Manchester, from 2006-2010.
I currently supervise project students on the SPA6776 and SPA6913 courses.
I am not currently teaching, but I do have open PhD projects.
My research covers the intersection of theoretical and observational cosmology. I am interested in what inhomogeneities can tell us about dark energy, and how novel observables and statistical tools can be used to make inferences about the cosmos on the largest scales. Research topics include:
- Theory and observations of large-scale structure, especially with optical and radio (21cm) surveys like LSST and SKA.
- Data analysis techniques for 21cm intensity mapping surveys, esp. with HERA, SKA, and MeerKAT.
- Secondary anisotropies and spectral distortions of the CMB as cosmological probes.
- General Relativistic effects on matter inhomogeneities and light propagation.
- Bayesian inference, stochastic processes, and computational physics.
I am currently offering PhD projects in theoretical and observational cosmology, in collaboration with other members of the Astronomy Unit and my collaborators in various other institutions.
This is not an exhaustive list and I would be happy to discuss other project possibilities.
Unveiling Cosmic Dawn and the Epoch of Reionisation with HERA
While our knowledge of very early and very late cosmic times has come on leaps and bounds over the past couple of decades, we still know comparatively little about the intervening period, when the first stars and galaxies formed and began to light up the Universe. This era can be broadly broken up into the Dark Ages, when neutral gas filled the cosmos; Cosmic Dawn, when the first stars began to shine; and the Epoch of Reionisation, when radiation from stars and galaxies ionised the neutral gas, leaving the Universe in much the same ionisation state as it is today.
The first stars and galaxies were rare objects, making them hard to find at such immense distances, while other forms of emission that might be easier to see were not yet in operation. As such, our understanding of this period remains very patchy, as it is hard to probe observationally. Fortunately, the neutral hydrogen gas that pervaded the Universe at this time gives off a faint glow at radio wavelengths, called 21cm emission. While this is hard to detect -- it is hidden behind the very bright radio emission of our own galaxy -- a number of experiments are currently in operation that hope to find and characterise the 21cm signal in order to finally understanding this most mysterious period in the Universe's history.
The Hydrogen Epoch of Reionisation Array (HERA), operating from the Karoo desert in South Africa, is one of the most powerful such experiments. Built as a large array of many close-packed radio dishes, HERA is already taking data, and in principle already has the sensitivity to put the best limits on fluctuations in the neutral hydrogen distribution during Cosmic Dawn and the EoR to date. Extracting this signal from the bright galactic emission and correcting for the imperfect response of the instrument to the signal remains extremely challenging however.
The aim of this project is to contribute to the development and application of advanced statistical methods to the analysis of existing and future HERA data. Key problems include:
- Efficiently and precisely estimating the power spectrum of 21cm fluctuations, in the presence of bright foreground emission.
- Correcting for "leakage" due to cuts in the data, imposed by excising radio frequency interference.
- Analysing the stability of the results to various analysis assumptions, by rigorously applying null tests and feeding simulated data through the analysis pipeline.
- Accurately estimating the covariance matrix of the data, which is needed to optimally extract the faint 21cm signal.
- Model the physical EoR and Cosmic Dawn signal in a computationally-efficient way.
This work will be carried out in collaboration with other members of the HERA Collaboration based in Berkeley, McGill, U. Western Cape, JPL, U. Penn, and U. Washington. You will gain skills in some or all of the following: analysis and interpretation of radio astronomical data, the application of high-dimensional inference techniques to large volumes of data, the application of structured and unsupervised machine learning techniques to detect spurious features in data, and the development of fast parallel data analysis codes to be run on high-performance computers. You will also gain experience in project management, transferable data science skills, and working in a mid-sized international collaboration.
Mapping the cosmic horizon with next-generation surveys
On the distance scales probed by current galaxy surveys, General Relativistic effects are only mildly important, and most of the interpretation of the data can be done within the context of Newtonian gravity. Forthcoming surveys like LSST, SKA, and Euclid are much bigger and deeper than their predecessors though, and so will be able to probe much larger scales -- approaching the size of the "cosmic horizon" -- where GR effects are important and must be taken into account.
Excitingly, the observable signatures of these effects should contain valuable information about some of the most fundamental questions in cosmology and theoretical physics -- the nature of gravity, the cause of cosmic acceleration, and how inflation set the initial conditions of the Universe we see today. By measuring them, we can test fundamental physics on truly cosmological scales.
The aim of this project is to put together a suite of tools for making practical measurements of GR effects on the very largest observable distance scales in the Universe. Because of an effect called cosmic variance, these measurements can only be made if data from several different large-scale surveys can be combined in a precise, statistically-robust way, with excellent rejection of systematic effects that contaminate the data. Only a few preliminary studies have attempted to use this "multi-tracer" method so far, and substantial advances in modelling and statistical analysis methods are urgently needed to make it usable in future surveys.
This project will develop the most important components of this method, with a view to using it on simulations, precursor datasets, and eventually the large surveys themselves. The work will include:
- Performing theoretical calculations to understand how the large-scale signal changes in different cosmological scenarios;
- Building realistic simulations of the signal, incorporating models of how galaxies trace the matter distribution;
- Simulations of systematic errors in the data caused by astrophysical contamination and the instrumental responses of optical and radio telescopes;
- Development of new statistical analysis methods that can clean out systematic effects and recover the true signal with sufficient precision.
Through this project, you will gain skills and experience in a range of important topics across cosmology and astrophysics: performing calculations using cosmological perturbation theory; running and analysing N-body simulations; applying galaxy-halo models to simulations and data; simulating data produced by optical and radio telescopes; working with Bayesian statistical analysis methods; and writing efficient parallel computer codes. You will also gain experience working in large scientific collaborations, where you will develop project management and communication skills, and have the opportunity to attend international meetings and work with international collaborators in the US, South Africa, and Italy.
I make most of the code used in my research public. You can find a full listing on my website, including links to the code, but here are a few highlights:
- RadioFisher: Fisher forecasting code for 21cm intensity mapping
- LSST Core Cosmology Library: General-purpose code with validated predictions for cosmological observables
- HERA PSpec: Optimal quadratic power spectrum estimation code for redundant arrays
- Bubble: Background predictions for exact, spherically-symmetric inhomogeneous GR solutions
- Commander 2: Full-sky CMB Gibbs sampler based on multi-grid solver