National Astronomy Meeting Poster Exhibition

Interest in Astronomy has increased during the pandemic as people are seeking diverse interests at home. Since 2014 the Brecon Beacons Observatory (BBO) has been encouraging public interest in astronomy via observing sessions, public talks and outreach, online learning and courses and cooperation with local astronomical societies. This talk will report on our progress and achievements at the BBO and how focus on dark sky initiatives benefits the Brecon Beacons national park and the public.

The DESI survey will collect data on a large number of galaxies. We shall present our work on techniques to maximise the physical data retrieval for the purpose of galaxy evolution analysis. We focus on the use of machine learning algorithms as a means of intelligent noise reduction, primarily auto encoder methods. In addition, we make use of the latent space representations produced by such methods to automate the selection of similar spectra for stacking.

In anticipation of the DESI SV data we first train our models using SDSS spectra with high signal to noise, adding artificial noise to create a training set. Once the SV data are available we shall look into both training the model on the data, and also at transfer learning between the two datasets.

There is still contention to the origins and drivers behind the observed correlations between supermassive black holes (SMBHs) and their host galaxies. These relations are fundamental to our understanding of galaxy evolution. Whilst there is evidence that SMBHs grow via secular processes, major mergers have often been cited as the main instigators of SMBH, galaxy growth and the drivers of observed SMBH-galaxy correlations. However, recent studies of SMBHs within disk-dominated galaxies with growth histories free of significant mergers have found contradictory results on whether these black holes are consistent with the SMBH-stellar mass relation. We combine three samples of active disk-dominated galaxies, which have different selection methods and biases, and interrogate them in the context of a model that assumes we are likely observing the system’s first instances of significant SMBH growth. The samples, which span a combined redshift range of 0.02 z 0.25, a stellar mass range of 8.9 log M*/M_Sun 11.4 and a black hole mass range of 5.0 log M_BH/M_Sun 8.9 give consistent results within this framework, despite their significantly different selection methods and sampling of different parts of the relevant parameter space. Our combined analysis resolves the apparent contradiction between these three studies under a straightforward framework of merger-free black hole growth. This suggests that merger-free SMBH growth onto the well-established MBH-M* relation can precede major merger events

The levels of heavy elements in stars are the product of enhancement by previous stellar generations, and the distribution of this metallicity among the population contains clues to the process by which a galaxy formed. We present analysis from our most recent paper, in which we measure metallicity distributions for a large sample of spiral galaxies, using high quality spectral data obtained by the SDSS-IV MaNGA survey. We find that high-mass spirals generically show a deficit of low-metallicity stars, implying that gas accretion is common in such galaxies. This is behaviour similar to what has classically been observed in the Milky Way. By contrast, low-mass spirals exhibit metallicity distributions that would be expected if such systems evolved as closed boxes. This distinction can be understood from the differing timescales for star formation in galaxies of differing masses. Furthermore, we also discuss our ongoing project, in which we endeavour to investigate the interplay between stellar and gas-phase metallicity in the same sample of spiral galaxies. More specifically, we determine both the stellar and gas-phase metallicity histories of these galaxies, in addition to their star formation histories. It is our hope that measuring the progression of these properties over cosmic time will allow us to unravel the cosmic evolution histories of spiral galaxies, as well as furthering our understanding of how the timescale of star formation varies according to stellar masses of such galaxies.

Our understanding of the sources that produce gravitational waves hinges on the ability to perform Bayesian inference on the physical parameters that describe them. However, for certain waveforms and systems, this is computationally expensive. In this talk, I will present our new nested sampling algorithm that incorporates machine learning to improve the efficiency. Our new algorithm called nessai is applicable to general problems. In this work, we focus on the application to compact binary coalescence and achieve a factor of two improvement compared to current methods in use by the LVK Collaboration.

The plasmasphere plays a key role in magnetospheric dynamics, influencing the dynamics of the radiation belt, ULF wave profiles and magnetospheric convection. Conjugate observations of the undulations on the plasmapause and at the equatorward edge of the auroral oval (He et al., 2020) imply that these two boundaries are co-located. Using global plasmapause identifications from extreme ultraviolet observations, we compare the mapped ionospheric footpoints of the plasmapause identifications with the equatorward auroral boundary in all local time sectors and under different levels of geomagnetic activity. The results of this analysis show that the ionospheric footpoint of the plasmapause maps closely to the equatorward boundary of the auroral oval in the nightside sectors statistically. However, in the dayside local time sectors, there is a large (~10°) statistical offset between the two boundaries implying that the processes that limit the equatorward extent of the aurora differ on the dayside and nightside of the Earth. The upcoming SMILE mission will provide an excellent suite of both in-situ and remote sensing observations of the aurora and magnetospheric plasma environments to explore in more detail the difference in the agreement between the plasmapause and the equatorward auroral boundary in different local time sectors.

Magnetic fields are critical for galaxy evolution and star formation, yet the origins of cosmic magnetic fields are still poorly constrained. One way to study the origin of magnetic fields is to observe galaxy clusters. The linearly polarized emission coming from host and background galaxies give us important information about the cluster such as the magnetic field strength, radial profile and magnetic field power spectrum. Although the rotation of the polarization angle as a function of wavelength along the line of sight produced by the ionized-magnetized gas in the cluster can be calculated using the Faraday Rotation Measurement Synthesis technique, resolution plays an important role in this problem. Poorer resolution produces a beam-depolarization effect which will encompass several Rotation Measure (RM) sources across the beam. This effect can be mitigated by using a long-baseline radio-interferometer such as e-MERLIN. In this talk, I will describe the processing pipeline used for e-MERLIN L-band data to constrain magnetic field profiles in a sample of Abell clusters, show some early results from the e-MERLIN study, and describe how we intend to use these observations to investigate their properties and magnetic fields.

Deep learning relies on finding meaningful representations of data. These representations are particularly important for galaxy morphology tasks, where complex images are difficult to interpret directly. We argue that the recently-created Galaxy Zoo DECaLS model, trained to accurately answer every Galaxy Zoo question simultaneously, has learned a meaningful representation of morphology that is useful for new tasks. We exploit this to provide several open-source tools for investigating large galaxy samples. These are aimed at researchers hoping to exploit deep learning approaches for their own challenges but without the capacity for citizen-science-scale labeling.

These tools are; a similarity search web interface, to identify galaxies of similar morphology to a query galaxy; an active learning anomaly detection algorithm (extending astronomaly), to identify the most interesting anomalies to a particular researcher; and a morphology transfer learning Python package, to build classifiers from only a few hundred labelled examples.

We develop and demonstrate the performance of our tools using 911,442 galaxies imaged by DECaLS. This includes producing the first large-scale catalogue of ring galaxies, identified using transfer learning and 212 examples tagged by volunteers on the Galaxy Zoo forum.

We introduce zeus, a well-tested Python implementation of the Ensemble Slice Sampling (ESS) method for Bayesian parameter inference. ESS is a novel Markov chain Monte Carlo (MCMC) algorithm specifically designed to tackle the computational challenges posed by modern astronomical and cosmological analyses. In particular, the method requires no hand-tuning of any hyper-parameters, its performance is insensitive to linear correlations and it can scale up to 1000s of CPUs without any extra effort. Furthermore, its locally adaptive nature allows to sample efficiently even when strong non-linear correlations are present. Lastly, the method achieves a high performance even in strongly multimodal distributions in high dimensions. Compared to emcee, a popular MCMC sampler, zeus's efficiency is an order of magnitude higher in a cosmological analysis of Baryon Acoustic Oscillations.

I will present a data-based, Machine Learning analysis aimed at identifying which physical properties are mostly connected with the position of local star forming galaxies in the classical diagnostic 'BPT' diagrams.

Exploiting the huge statistics available from spectroscopic surveys in the local Universe like the SDSS and MaNGA, I have defined a framework in which the dispersion of galaxies in the BPT diagrams and, in particular, their deviation from the local sequence best-fit, can be described by means of the relative variation in different observational properties compared to the average value retained by the bulk of the galaxies along the sequence. Artificial Neural Networks and Random Forest Trees are implemented to both classify whether galaxies lie above or below the sequence and to predict the exact distance/offset from the sequence itself. We achieve a high accuracy on the test sample in both classification and regression tasks (AUC>95%, RMSE~0.025 ), with no clear overfitting. Moreover, different approaches are implemented to rank the parameters in terms of how much informative they are for the models. We show that the nitrogen-over-oxygen abundance ratio (N/O) and the ionisation parameter (U) are the most predictive parameters in the [N II]-BPT, whereas features related to the star-forming state of galaxies perform better in the [S II]-BPT. However, we also show that both the performances and relative importance of each feature change as we consider different regions within the diagrams.

These models represent also a valuable benchmark for high redshift galaxy samples, in order to assess to what extent the physics that shape the local BPT diagrams is the same causing the offset seen in high-z sources or, instead, whether a different framework or even different physical mechanisms need to be involved.