Solar, Stellar and Time-domain Astrophysics

Supervisory Team: Klaas Wiersema, James Collett, William Cooper

The last few years have seen an exponential growth in the number of orbiting satellites and space debris, mostly because of the build-up of mega-constellations of communication satellites (e.g. Starlink), anti-satellite weapon tests and accidental collisions. This has greatly increased the importance of accurate monitoring of orbiting objects, not just in low-Earth orbits but also in geostationary orbits. We astronomers have the best equipment: big telescopes with very sensitive instruments, suited for highly detailed studies of these objects using novel techniques. This allows us to turn these objects from a nuisance into useful tools, to study things like: the physics of reflection and absorption of light by objects in the near-Earth environment; the properties of dust particles in the upper atmosphere; the magnetic field properties far from Earth; the interaction of radiation and high-energy particles with solid surfaces, and the long-term orbital and spin evolution of (irregularly shaped) debris.

In this project, you will obtain and use high-cadence data of light reflected by orbiting satellites and debris, using large astronomical telescopes with specialist astronomical instrumentation. This project will mainly focus on high-speed polarisation measurements, using data obtained through several large surveys currently led by the project supervisor. Through these polarisation data, combined with simultaneous multi-colour lightcurves, we will measure the orientation of reflecting satellites (because the polarisation is a function of the angle of incidence) and the spin-state evolution of inactive (dead) satellites and debris. We will use these data to characterise the effects of sunlight on the long-term evolution of tumbling debris, and compare these measurements with those of near-Earth asteroids. In addition, you will use the light reflected by satellites to measure properties of dust particles in the upper atmosphere, which are valuable input into exoplanet atmosphere models. Lastly, you will be able to take advantage of new developments in polarisation detectors and new widefield astronomical facilities to help reduce the negative impact of satellites and debris on astronomical data, as well as to turn them into useful astronomical tools.

Supervisory Team: Jan Forbrich, Mykola Gordovskyy, Mike Kuhn

Radio stars are coming back, this time in colour! New and upgraded observational capabilities are changing our view of the radio emission of stars, and of young stars in particular. In spite of their early evolutionary stage, they already show intense high-energy processes, as evidenced by both X-ray and nonthermal radio emission, and they also show collimated radio jets. In an ongoing multi-year project, we are using the best radio astronomy observatories in the cm- and mm-wavelength range for observations of the famous Orion Nebula Cluster (ONC), one of the richest young and nearby star clusters. We would like to understand better the impact of high-energy processes on mass accretion, that is the process by which stars gain their mass, and also on protoplanetary disks, where they can influence the initial conditions of planet formation and young planets. In a major next step, we are conducting an ALMA high-priority project in September 2025, when we will obtain the first simultaneous time-series observations of the ONC using both ALMA and the VLA. Having found evidence for large radio flares in the ONC, we will now obtain, for the first time, radio spectral index time series, where an spectral index is akin to a radio ‘colour’. Analysis of detailed so-called dynamic radio spectra is a standard method in solar physics, and our goal is to establish connections between these radio spectral index time series in the ONC and our understanding of the space weather produced by our present-day Sun, which has shown similarly enhanced levels of high-energy processes in its youth. Among other science, this project will also produce the first systematic survey of radio stars in the ONC in the newly accessible ALMA band 1 (35-50 GHz). This project could be either on the observational side (radio astronomy time domain data analysis) and/or on the computational modelling side, depending on your interests.

Supervisory Team: Peter Scicluna and Phil Lucas

At the end of their lives, stars like the Sun eject their outer layers and in the process return the products of the nucleosynthesis that has been happening in their cores. We observe this mass loss through observations of dust and gas emission in the infrared and sub-mm wavelength ranges. This project focuses on understanding mass-loss processes in asymptotic giant branch (AGB) stars, a key phase in stellar evolution that significantly influences the chemical evolution of the Milky Way.

You will work with data from the Nearby Evolved Stars Survey (NESS), which is observing a volume-limited sample of ~800 Galactic AGB stars. You will have the chance to work with observations such as molecular line and dust emission in the sub-millimeter, optical and near-infrared imaging and spectroscopy of the stars, time-series data on stellar variability, or mid-infrared spectra detailing dust composition, depending on your interests. By interpreting these observations with a combination of computational modelling, machine learning and Bayesian statistics, you will explore how AGB stars’ mass loss is influenced by the other properties of the stars and, in turn, the evolution and distribution of dust, gas, and chemical abundances across our galaxy.

In addition, you will have the opportunity to propose and develop future observational projects to expand on the insights generated in this work. This project is ideal for students interested in stellar evolution, observational astrophysics, and statistical analysis, seeking to unravel the complex interactions between dying stars and the galactic ecosystem.

Supervisory Team: Peter Scicluna, Ashley Spindler, Will Alston

Astronomical data is often riddled with noise, artefacts, and contaminating signals that can obscure scientific insights. Conventional methods tackle these issues by sequentially removing each contaminant, a process that can lead to compounded uncertainties and potential inaccuracies. This Ph.D. project seeks to develop new software tools based on variational inference – a form of Bayesian machine learning – to process data through a forward-modelling approach.

Rather than removing noise sources retrospectively, the student will explore techniques to construct a comprehensive model of observational data that accounts for all noise sources simultaneously, guided by reasonable priors. This forward-model approach enables the simultaneous estimation of multiple measurement effects, yielding more accurate uncertainties for the final scientific images and spectra. For complex data, this has the potential to significantly improve the final products.

The project will focus on interpreting observations of dusty evolved stars, which contribute significantly to the cosmic dust budget, using data from a variety of telescopes and instruments. This work will be well-suited to students interested in developing expertise in advanced statistical techniques, machine learning applications in astrophysics, and observational studies of stellar evolution.

Supervisory Team:  Peter Scicluna, Ashley Spindler, Will Alston

With the exponential growth in astronomical research, thousands of new papers are published each year, containing valuable data and new astrophysical insights. Unfortunately, critical information – e.g. estimates of physical constants – is often embedded within text and tables rather than added to searchable databases, and as a result remains inaccessible to automated queries. This adds extra labour to future research by requiring someone (typically students) to find and read papers and manually extract data, which also carries significant risk that important results may be overlooked. This project aims to address this challenge by exploring the use of large language models (LLMs) and retrieval-augmented generation techniques to automatically extract and organise this data in structured formats.

You will work on developing models and tools capable of interpreting scientific text, tables, and figures in astronomical publications, extracting structured data to populate databases of astrophysical constants and parameters. You will explore the benefits of different natural language processing techniques and frameworks, and the cost-benefit tradeoffs of e.g. larger models or more extensive pre-processing. By transforming previously inaccessible information into structured, searchable formats, this project will support astronomers by streamlining access to knowledge locked in the literature.

This project is ideal for students interested in natural language processing, AI-driven information retrieval, and advancing scientific databases. The work has the ultimate goal of creating tools that can automatically build and curate astrophysical databases, to democratise access to data.

Supervisory Team: Chiaki Kobayashi, Federico Sestito

Soon after the Big Bang, only very light elements (H, He, Li, Be, and B) can be produced, and carbon and heavier elements are all formed in stars and ejected by supernovae. Many elements (from carbon to uranium) have been observed in millions of stars in the Milky Way and its satellite galaxies with spectroscopic ‘galactic archaeology' surveys. Some of the element production sites are also observed as gravitational wave events such as the neutron star merger in 2017. The student will study the origin of elements in the Universe by comparing computational simulations of galaxies to these observational data, then predict the gravitational wave events for future missions in space (LISA) and on Moon. Our hydrodynamical simulation code already includes basic physics such as star formation and supernova feedback, and thus it is possible to compare with the observed elemental abundances in the Milky Way and its satellite galaxies. The student will update the code (written in c) to include the detailed effects from binary stars such as the neutron star and black hole mergers (for the first time in the world!), and use local LINUX cluster and national supercomputer facility (DiRAC).

Supervisory Team: Philip Lucas, Gabriella Zsidi and Michael Kuhn

The VISTA Variables in the Via Lactea survey (VVV) and its extension (VVVX) are the first serious exploration of variable stars in the Milky Way at infrared wavelengths. The survey ran for 13 years, led by Prof Lucas at Hertfordshire and Prof Minniti in Chile, and produced light curves for about 1 billion stars. It used the 4-m VISTA telescope in Chile.This project is about eruptive variable protostars, which flare up brightly due to a sudden huge increase in the rate of flow of matter through the star-forming disc and on to the star, for reasons not yet understood. VVV has dominated this field in recent years, finding larger numbers of eruptive YSOs for the first time. This project would extend the work to the new VVVX survey fields to complete sample of nearby eruptive YSOs in relatively nearby star clusters and enable a first statistical analysis of the population. Associated follow up work includes spectroscopy with the Very Large Telescope in Chile and photometry with the Infrared Survey Survey Facility in South Africa.

Supervisory Team: Philip Lucas, Michael Kuhn and Gabriella Zsidi

The variability of Young Stellar Objects (YSOs) has been studied at optical wavelengths for many decades. However, infrared observations are needed to study the earlier stages of pre-main sequence evolution, i.e. the protostars, for which large samples have only recently been identified. The SPICY catalogue, led by Michael Kuhn, provides the best large sample of protostars whilst the VISTA Variables in the Via Lactea survey (VVV, led by Philip Lucas) provides long duration near infrared light curves (~10 yr) for these stars. The near infrared data are supplemented by mid-infrared light curves from the WISE satellite. Infrared variability in YSOs is caused by changes in accretion rate and variable extinction within the star system, neither of which are well understood processes. Variability happens on a range of timescales but it has not been studied on timescales longer than 2 or 3 years, save for the more dramatic eruptive events. This project will be the first investigation of the full range of types of infrared variability on long timescales and it will also employ a much larger sample than previous studies. By combining VVV, WISE and optical light curves where possible, this project aims to characterise the different types of variability in YSOs and make progress in understanding their origin. The project may also involve follow up spectroscopy to better understand the nature of these systems.

Supervisory Team: Mykola Gordovskyy, Jan Forbrich, James Geach

We live in the heliosphere - a volume of space dominated by the Sun. The Sun is active, it changes, constantly producing sunspots, flares and coronal mass ejections. Solar flares - the most energetic explosive events in the solar system - are one of the main features of solar activity. They result in major perturbations in the solar atmosphere, or corona, and heliosphere. Understanding solar activity and solar flares is key to understanding how the Sun and other magnetically-active stars work. In addition, solar flares can explain the behaviour of hot magnetised plasma in laboratory devices such as tokamaks.

In this project you would investigate solar energetic particles - electrons and ions travelling at fractions of the speed of light, which dominate the primary energy release in flares. Some of these particles precipitate in the corona, producing bright radio, X-ray and gamma-ray radiations, but some escape into the heliosphere, affecting space weather. This project will use the combination of magnetohydrodynamic and kinetic simulations to create models of particle acceleration and transport in individual solar flares matching multi-wavelength and in situ observations from the Solar Orbiter and Parker Solar Probe missions, as well as the LOFAR radio-telescope. The ultimate goal will be to develop a pipeline for nearly-real-time modelling of energetic particles in flares and forecasting their properties in the solar corona and inner heliosphere. Although the project will be primarily theoretical/computational, it will also involve analysis of observational data using state-of-the-art techniques, including machine learning and neural network algorithms.