arXiv subjects

Martin J. Hardcastle

Publications and source records attributed to Martin J. Hardcastle.

At least 37 records · Page 2Linked to original sources

arXiv2023

Numerical modelling of the lobes of radio galaxies -- Paper V: Universal Pressure Profile cluster atmospheres

We present relativistic magnetohydrodynamic modelling of jets running into hydrostatic, spherically symmetric cluster atmospheres. For the first time in a numerical simulation, we present model cluster atmospheres based upon the Universal Pressure Profile (UPP), incorporating a temperature profile for a typical self-similar atmosphere described by only one parameter - $M_{500}$. We explore a comprehensive range of realistic atmospheres and jet powers and derive dynamic, energetic and polarimetric data which provide insight into what we should expect of future high-resolution studies of AGN outflows. From the simulated synchrotron emission maps which include Doppler beaming we find sidedness distributions that agree well with observations. We replicated a number of findings from our previous work, such as higher power jets inflating larger aspect-ratio lobes and the cluster environment impacting the distribution of energy between the lobe and shocked regions. Comparing UPP and $β$-profiles we find that the cluster model chosen results in a different morphology for the resultant lobes with the UPP more able to clear lobe material from the core; and that these different atmospheres influence the ratio between the various forms of energy in the fully developed lobes. This work also highlights the key role played by Kelvin-Helmholtz (KH) instabilities in the formation of realistic lobe aspect-ratios. Our simulations point to the need for additional lobe-widening mechanisms at high jet powers, for example jet precession. Given that the UPP is our most representative general cluster atmosphere, these numerical simulations represent the most realistic models yet for spherically symmetric atmospheres.

astro-ph.HE↗

arXiv2023

Spectral age distribution for radio-loud active galaxies in the XMM-LSS field

Jets of energetic particles, as seen in FR type-I and FR type-II sources, ejected from the center of Radio-Loud AGN affect the sources surrounding intracluster medium/intergalactic medium. Placing constraints on the age of such sources is important in order to measure the jet powers and determine the effects on feedback. To evaluate the age of these sources using spectral age models, we require high-resolution multi-wavelength data. The new sensitive and high-resolution MIGHTEE survey of the XMM-LSS field along with data from the Low Frequency Array (LOFAR) and the Giant Metrewave Radio Telescope (GMRT) provide data taken at different frequencies with similar resolution, which enables us to determine the spectral age distribution for radio loud AGN in the survey field. In this study we present a sample of 28 radio galaxies with their best fitting spectral age distribution analyzed using the Jaffe-Perola (JP) model on a pixel-by-pixel basis. Fits are generally good and objects in our sample show maximum ages within the range of 2.8 Myr to 115 Myr with a median of 8.71 Myr. High-resolution maps over a range of frequencies are required to observe detailed age distributions for small sources and high-sensitivity maps will be needed in order to observe fainter extended emission. We do not observe any correlation between the total physical size of the sources and their age and we speculate both dynamical models and the approach to spectral age analysis may need some modification to account for our observations.

astro-ph.GA↗

arXiv2023

Finding AGN remnant candidates based on radio morphology with machine learning

Remnant radio galaxies represent the dying phase of radio-loud active galactic nuclei (AGN). Large samples of remnant radio galaxies are important for quantifying the radio galaxy life cycle. The remnants of radio-loud AGN can be identified in radio sky surveys based on their spectral index, or, complementary, through visual inspection based on their radio morphology. However, this is extremely time-consuming when applied to the new large and sensitive radio surveys. Here we aim to reduce the amount of visual inspection required to find AGN remnants based on their morphology, through supervised machine learning trained on an existing sample of remnant candidates. For a dataset of 4107 radio sources, with angular sizes larger than 60 arcsec, from the LOw Frequency ARray (LOFAR) Two-Metre Sky Survey second data release (LoTSS-DR2), we started with 151 radio sources that were visually classified as 'AGN remnant candidate'. We derived a wide range of morphological features for all radio sources from their corresponding Stokes-I images: from simple source catalogue-derived properties, to clustered Haralick-features, and self-organising map (SOM) derived morphological features. We trained a random forest classifier to separate the 'AGN remnant candidates' from the not yet inspected sources. The SOM-derived features and the total to peak flux ratio of a source are shown to be most salient to the classifier. We estimate that $31\pm5\%$ of sources with positive predictions from our classifier will be labelled 'AGN remnant candidates' upon visual inspection, while we estimate the upper bound of the $95\%$ confidence interval for 'AGN remnant candidates' in the negative predictions at $8\%$. Visual inspection of just the positive predictions reduces the number of radio sources requiring visual inspection by $73\%$.

astro-ph.GA↗

arXiv2023

CosmoDRAGoN simulations -- I. Dynamics and observable signatures of radio jets in cosmological environments

We present the Cosmological Double Radio Active Galactic Nuclei (CosmoDRAGoN) project: a large suite of simulated AGN jets in cosmological environments. These environments sample the intra-cluster media of galaxy clusters that form in cosmological smooth particle hydrodynamics (SPH) simulations, which we then use as inputs for grid-based hydrodynamic simulations of radio jets. Initially conical jets are injected with a range of jet powers, speeds (both relativistic and non-relativistic), and opening angles; we follow their collimation and propagation on scales of tens to hundreds of kiloparsecs, and calculate spatially-resolved synthetic radio spectra in post-processing. In this paper, we present a technical overview of the project, and key early science results from six representative simulations which produce radio sources with both core- (Fanaroff-Riley Type I) and edge-brightened (Fanaroff-Riley Type II) radio morphologies. Our simulations highlight the importance of accurate representation of both jets and environments for radio morphology, radio spectra, and feedback the jets provide to their surroundings.

astro-ph.HE↗

arXiv2023

New mechanisms for forming multiple hotspots in radio jets

Hotspots of radio galaxies are regions of shock-driven particle acceleration. Multiple hotspots have long been identified as potential indicators of jet movement or precession. Two frequent explanations describe a secondary hotspot as either the location of a prior jet termination point, or a deflected backflow-driven shock: the so-called Dentist's Drill and Splatter Spot models. We created high-resolution simulations of precessing jets with a range of parameters. In addition to the existing mechanisms, our results show three additional mechanisms for multiple hotspot formation: (1) the splitting of a large terminal hotspots into passive and active components; (2) jet stream splitting resulting in two active hotspots; (3) dynamic multiple hotspot complexes that form as a result of jet termination in a turbulent cocoon, linked here to rapid precession. We show that these distinct types of multiple hotspots are difficult to differentiate in synthetic radio maps, particularly hotspot complexes which can easily be mistaken for the jet itself. We discuss the implication for hypothesised binary supermassive black hole systems where jet precession is a key component of the morphology, and show a selection of potential precession candidates found using the LOFAR Two-Metre Sky Survey Data Release 2 (LoTSS DR2).

astro-ph.HE↗

arXiv2023

Measuring the giant radio galaxy length distribution with the LoTSS

Radio galaxies are luminous structures created by the jets of supermassive black holes, and consist of atomic nuclei, relativistic electrons, and magnetic fields. In exceptional cases, radio galaxies attain cosmological, megaparsec extents - and thus turn into giants. Giants embody the most extreme known mechanism through which galaxies can impact the Cosmic Web around them. The triggers of giant growth remain a mystery. Excitingly, new sensitive low-frequency sky surveys hold promise to change this situation. In this work, we perform a precision measurement of the distribution of giant growth's central dynamical quantity: total length. We first construct a statistical geometric framework for radio galaxies that is both rigorous and practical. We then search the LOFAR Two-metre Sky Survey for giants, discovering 2060 previously unknown specimina: more than have been found in all preceding literature combined. Spectacular discoveries include the longest giant hosted by an elliptical galaxy, the longest giant hosted by a spiral galaxy, and 13 giants with an angular length larger than that of the full Moon. By combining theory and observations - and carefully forward modelling selection effects - we infer that giant radio galaxy lengths are well described by a Pareto distribution with tail index $-3.5 \pm 0.5$. This finding is a new observational constraint for models and simulations of radio galaxy growth. In addition, for the first time, we determine the comoving number density of giants, $5 \pm 2\ (100\ \mathrm{Mpc})^{-3}$, and the volume-filling fraction of giant radio galaxy lobes in clusters and filaments, $5\substack{+8\\-2}\cdot 10^{-6}$. We conclude that giants are truly rare - not only from an observational perspective, but also from a cosmological one. At any moment in time, most clusters and filaments - the building blocks of the modern Cosmic Web - do not harbour giants.

astro-ph.GA↗

arXiv2022

An intergalactic medium temperature from a giant radio galaxy

The warm-hot intergalactic medium (warm-hot IGM, or WHIM) pervades the filaments of the Cosmic Web and harbours half of the Universe's baryons. The WHIM's thermodynamic properties are notoriously hard to measure. Here we estimate a galaxy group - WHIM boundary temperature using a new method. In particular, we use a radio image of the giant radio galaxy (giant RG, or GRG) created by NGC 6185, a massive nearby spiral. We analyse this extraordinary object with a Bayesian 3D lobe model and deduce an equipartition pressure $P_\mathrm{eq} = 6 \cdot 10^{-16}\ \mathrm{Pa}$ -- among the lowest found in RGs yet. Using an X-ray-based statistical conversion for Fanaroff-Riley II RGs, we find a true lobe pressure $P = 1.5\substack{+1.7\\-0.4}\cdot 10^{-15}\ \mathrm{Pa}$. Cosmic Web reconstructions, group catalogues, and MHD simulations furthermore imply an $\mathrm{Mpc}$-scale IGM density $1 + δ_\mathrm{IGM} = 40\substack{+30\\-10}$. The buoyantly rising lobes are crushed by the IGM at their inner side, where an approximate balance between IGM and lobe pressure occurs: $P_\mathrm{IGM} \approx P$. The ideal gas law then suggests an IGM temperature $T_\mathrm{IGM} = 11\substack{+12\\-5} \cdot 10^6\ \mathrm{K}$, or $k_\mathrm{B}T_\mathrm{IGM} = 0.9\substack{+1.0\\-0.4}\ \mathrm{keV}$, at the virial radius -- consistent with X-ray-derived temperatures of similarly massive groups. Interestingly, the method is not performing at its limit: in principle, estimates $T_\mathrm{IGM} \sim 4 \cdot 10^6\ \mathrm{K}$ are already possible -- rivalling the lowest X-ray measurements available. The technique's future scope extends from galaxy group outskirts to the WHIM. In conclusion, we demonstrate that observations of GRGs in Cosmic Web filaments are finally sensitive enough to probe the thermodynamics of galaxy groups and beyond.

astro-ph.GA↗

arXiv2022

Radio source-component association for the LOFAR Two-metre Sky Survey with region-based convolutional neural networks

Radio loud active galactic nuclei (RLAGNs) are often morphologically complex objects that can consist of multiple, spatially separated, components. Astronomers often rely on visual inspection to resolve radio component association. However, applying visual inspection to all the hundreds of thousands of well-resolved RLAGNs that appear in the images from the Low Frequency Array (LOFAR) Two-metre Sky Survey (LoTSS) at $144$ MHz, is a daunting, time-consuming process, even with extensive manpower. Using a machine learning approach, we aim to automate the radio component association of large ($> 15$ arcsec) radio components. We turned the association problem into a classification problem and trained an adapted Fast region-based convolutional neural network to mimic the expert annotations from the first LoTSS data release. We implemented a rotation data augmentation to reduce overfitting and simplify the component association by removing unresolved radio sources that are likely unrelated to the large and bright radio components that we consider using predictions from an existing gradient boosting classifier. For large ($> 15$ arcsec) and bright ($> 10$ mJy) radio components in the LoTSS first data release, our model provides the same associations for $85.3\%\pm0.6$ of the cases as those derived when astronomers perform the association manually. When the association is done through public crowd-sourced efforts, a result similar to that of our model is attained. Our method is able to efficiently carry out manual radio-component association for huge radio surveys and can serve as a basis for either automated radio morphology classification or automated optical host identification. This opens up an avenue to study the completeness and reliability of samples of radio sources with extended, complex morphologies.

astro-ph.IM↗

arXiv2022

A numerical study of the interplay between Fermi acceleration mechanisms in radio lobes of FR-II radio galaxies

Context: Radio-loud AGNs are thought to possess various sites of particle acceleration, which gives rise to the observed non-thermal spectra. Stochastic turbulent acceleration (STA) and diffusive shock acceleration (DSA) are commonly cited as potential sources of high-energy particles in weakly magnetized environments. Together, these acceleration processes and various radiative losses determine the emission characteristics of these extra-galactic radio sources. Aims: The purpose of this research is to investigate the dynamical interplay between the STA and DSA in the radio lobes of FR-II radio galaxies, as well as the manner in which these acceleration mechanisms, along with a variety of radiative losses, collectively shape the emission features seen in these extra-galactic sources. Methods: A phenomenologically motivated model of STA is considered and subsequently employed on a magneto-hydrodynamically simulated radio lobe through a novel hybrid Eulerian-Lagrangian framework. Results: STA gives rise to a curved particle spectrum that is morphologically different from the usual shock-accelerated spectrum. As a consequence of this structural difference in the underlying particle energy spectrum, various multi-wavelength features arise in the spectral energy distribution of the radio lobe. Additionally, we observe enhanced diffuse X-ray emission from radio lobes for cases where STA is taken into account in addition to DSA.

astro-ph.HE↗

arXiv2022

Faraday tomography of LoTSS-DR2 data: I. Faraday moments in the high-latitude outer Galaxy and revealing Loop III in polarisation

Observations of synchrotron emission at low radio frequencies reveal a labyrinth of polarised Galactic structures. However, the explanation for the wealth of structures remains uncertain due to the complex interactions between the interstellar medium and the magnetic field. A multi-tracer approach to the analysis of large sky areas is needed. This paper aims to use polarimetric images from the LOFAR Two metre Sky Survey (LoTSS) to produce the biggest mosaic of polarised emission in the northern sky at low radio frequencies (150 MHz) to date. The large area this mosaic covers allows for detailed morphological and statistical studies of polarised structures in the high-latitude outer Galaxy, including the well-known Loop III region. We produced a 3100 square degree Faraday tomographic cube using a rotation measure synthesis tool. We calculated the statistical moments of Faraday spectra and compared them with data sets at higher frequencies (1.4 GHz) and with a map of a rotation measure derived from extragalactic sources. The mosaic is dominated by polarised emission connected to Loop III. Additionally, the mosaic reveals an abundance of other morphological structures, mainly {narrow and extended} depolarisation canals, which are found to be ubiquitous. We find a correlation between the map of an extragalactic rotation measure and the LoTSS first Faraday moment image. The ratio of the two deviates from a simple model of a Burn slab (Burn 1966) along the line of sight, which highlights the high level of complexity in the magnetoionic medium that can be studied at these frequencies.

astro-ph.GA↗

arXiv2022

A machine learning classifier for LOFAR radio galaxy cross-matching techniques

New-generation radio telescopes like LOFAR are conducting extensive sky surveys, detecting millions of sources. To maximise the scientific value of these surveys, radio source components must be properly associated into physical sources before being cross-matched with their optical/infrared counterparts. In this paper, we use machine learning to identify those radio sources for which either source association is required or statistical cross-matching to optical/infrared catalogues is unreliable. We train a binary classifier using manual annotations from the LOFAR Two-metre Sky Survey (LoTSS). We find that, compared to a classification model based on just the radio source parameters, the addition of features of the nearest-neighbour radio sources, the potential optical host galaxy, and the radio source composition in terms of Gaussian components, all improve model performance. Our best model, a gradient boosting classifier, achieves an accuracy of 95 per cent on a balanced dataset and 96 per cent on the whole (unbalanced) sample after optimising the classification threshold. Unsurprisingly, the classifier performs best on small, unresolved radio sources, reaching almost 99 per cent accuracy for sources smaller than 15 arcsec, but still achieves 70 per cent accuracy on resolved sources. It flags 68 per cent more sources than required as needing visual inspection, but this is still fewer than the manually-developed decision tree used in LoTSS, while also having a lower rate of wrongly accepted sources for statistical analysis. The results have an immediate practical application for cross-matching the next LoTSS data releases and can be generalised to other radio surveys.

astro-ph.IM↗

arXiv2022

The discovery of a radio galaxy of at least 5 Mpc

We discover what is in projection the largest known structure of galactic origin: a giant radio galaxy with a projected proper length of $4.99 \pm 0.04\ \mathrm{Mpc}$. The source, named Alcyoneus, was first identified in low-resolution LOFAR Two-metre Sky Survey images from which angularly compact sources had been removed. Being an extreme example in its class, Alcyoneus could shed light on the main mechanisms that drive radio galaxy growth. We find that - beyond geometry - Alcyoneus and its host galaxy appear suspiciously ordinary: the total low-frequency luminosity density, stellar mass and supermassive black hole mass are all lower than, though similar to, those of the medial giant radio galaxy (percentiles $45 \pm 3\%$, $25 \pm 9 \%$ and $23 \pm 11 \%$, respectively). The source resides in a filament of the Cosmic Web, with which it might have significant thermodynamic interaction. At $5 \cdot 10^{-16}\ \mathrm{Pa}$, the pressures in the lobes are the lowest hitherto found, and Alcyoneus therefore represents one of the most promising radio galaxies yet to probe the warm-hot intergalactic medium.

astro-ph.GA↗

arXiv2021

Radio Morphology of Red Geysers

We present 150 MHz, 1.4 GHz, and 3 GHz radio imaging (LoTSS, FIRST and VLASS) and spatially resolved ionized gas characteristics (SDSS IV-MaNGA) for 140 local ($z<0.1$) early-type "red geyser" galaxies. These galaxies have low star formation activity (SFR $\sim \rm 0.01\ M_{\odot} yr^{-1}$), but show unique extended patterns in spatially-resolved emission line maps that have been interpreted as large-scale ionized winds driven by active galactic nuclei (AGN). In this work we confirm that red geysers host low-luminosity radio sources ($\rm L_{1.4GHz} \sim 10^{22} W Hz^{-1}$). Out of 42 radio-detected red geysers, 32 are spatially resolved in LoTSS and FIRST, with radio sizes varying between $\sim 5-25$ kpc. Three sources have radio sizes exceeding 40 kpc. A majority display a compact radio morphology and are consistent with either low-power compact radio sources ("FR0" galaxies) or "radio-quiet quasars". They may be powered by small-scale AGN-driven jets which remain unresolved at the current $5"$ resolution of radio data. The extended radio sources, not belonging to the "compact" morphological class, exhibit steeper spectra with a median spectral index of $-0.67$ indicating the dominance of lobed components. The red geysers hosting extended radio sources also have the lowest specific star formation rates, suggesting they either have a greater impact on the surrounding interstellar medium or are found in more massive halos on average. The degree of alignment of the ionized wind cone and the extended radio features are either 0$^{\circ}$ or 90$^{\circ}$, indicating possible interaction between the interstellar medium and the central radio AGN.

astro-ph.GA↗

arXiv2021

Characterising the extended morphologies of BL Lacs at 144 MHz with LOFAR

We present a morphological and spectral study of a sample of 99 BL Lacs using the LOFAR Two-Metre Sky Survey Second Data Release (LDR2). Extended emission has been identified at gigahertz frequencies around BL Lacs, but with LDR2 it is now possible to systematically study their morphologies at 144 MHz, where more diffuse emission is expected. LDR2 reveals the presence of extended radio structures around 66/99 of the BL Lac nuclei, with angular extents ranging up to 115 arcseconds, corresponding to spatial extents of 410 kpc. The extended emission is likely to be both unbeamed diffuse emission and beamed emission associated with relativistic bulk motion in jets. The spatial extents and luminosities of the extended emission are consistent with the AGN unification scheme where BL Lacs correspond to low-excitation radio galaxies with the jet axis aligned along the line-of-sight. While extended emission is detected around the majority of BL Lacs, the median 144-1400 MHz spectral index and core dominance at 144 MHz indicate that the core component contributes ~42% on average to the total low-frequency flux density. A stronger correlation was found between the 144 MHz core flux density and the gamma-ray photon flux (r = 0.69) compared to the 144 MHz extended flux density and the gamma-ray photon flux (r = 0.42). This suggests that the radio-to-gamma-ray connection weakens at low radio frequencies because the population of particles that give rise to the gamma-ray flux are distinct from the electrons producing the diffuse synchrotron emission associated with spatially-extended features.

astro-ph.HE↗

arXiv2021

Unmasking the history of 3C 293 with LOFAR sub-arcsecond imaging

Active galactic nuclei (AGNs) show episodic activity, evident in galaxies that exhibit restarted radio jets. These restarted jets can interact with their environment, leaving signatures on the radio spectral energy distribution. Tracing these signatures requires resolved spectral index measurements over a broad frequency range including low frequencies. We present such a study for the radio galaxy 3C 293. Using the International LOFAR telescope (ILT) we probed spatial scales as fine as ~0.2" at 144 MHz, and to constrain the spectrum we combined these data with Multi-Element Radio Linked Interferometer Network (MERLIN) and Very Large Array (VLA) archival data. In the inner lobes (~2 kpc), we detect the presence of a spectral turnover that peaks at ~225 MHz and is most likely caused by free-free absorption from the rich surrounding medium. We confirm that these inner lobes are part of a jet-dominated young radio source (spectral age $\lesssim$0.17 Myr), which is strongly interacting with the rich interstellar medium (ISM) of the host galaxy. The outer lobes (~100 kpc) have a spectral index of $α$~0.6-0.8 from 144-4850 MHz with a remarkably uniform spatial distribution and only mild spectral curvature ($Δα\lesssim$ 0.2). We propose that intermittent fuelling and jet flow disruptions are powering the mechanisms that keep the spectral index in the outer lobes from steepening and maintain the spatial uniformity of the spectral index. Overall, it appears that 3C 293 has gone through multiple (two to three) epochs of activity. This study adds 3C 293 to the new sub-group of restarted galaxies with short interruption time periods. This is the first time a spatially resolved study simultaneously studies a young source as well as the older outer lobes at such low frequencies. This illustrates the potential of the ILT to expand such studies to a larger sample of radio galaxies.

astro-ph.GA↗

arXiv2021

Can the Local Bubble explain the radio background?

The ARCADE 2 balloon bolometer along with a number of other instruments have detected what appears to be a radio synchrotron background at frequencies below about 3 GHz. Neither extragalactic radio sources nor diffuse Galactic emission can currently account for this finding. We use the locally measured Cosmic ray electron population, demodulated for effects of the Solar wind, and other observational constraints combined with a turbulent magnetic field model to predict the radio synchrotron emission for the Local Bubble. We find that the spectral index of the modelled radio emission is roughly consistent with the radio background. Our model can approximately reproduce the observed antenna temperatures for a mean magnetic field strength B between 3-5 nT. We argue that this would not violate observational constraints from pulsar measurements. However, the curvature in the predicted spectrum would mean that other, so far unknown sources would have to contribute below 100 MHz. Also, the magnetic energy density would then dominate over thermal and cosmic ray electron energy density, likely causing an inverse magnetic cascade with large variations of the radio emission in different sky directions as well as high polarisation. We argue that this disagrees with several observations and thus that the magnetic field is probably much lower, quite possibly limited by equipartition with the energy density in relativistic or thermal particles (B = 0.2-0.6 nT). In the latter case, we predict a contribution of the Local Bubble to the unexplained radio background at most at the per cent level.

astro-ph.HE↗

arXiv2020

A low frequency study of linear polarization in radio galaxies

Radio galaxies are linearly polarized -- an important property that allows us to infer the properties of the magnetic field of the source and its environment. However at low frequencies, Faraday rotation substantially depolarizes the emission, meaning that comparatively few polarized radio galaxies are known at low frequencies. Using the LOFAR Two Metre Sky Survey at 150 MHz and at 20 arcsec resolution, we select 342 radio galaxies brighter than 50 mJy and larger than 100 arcsec in angular size, of which 67 are polarized (18 per cent detection fraction). These are predominantly Fanaroff Riley type II (FR-II) sources. The detection fraction increases with total flux density, and exceeds 50 per cent for sources brighter than 1 Jy. We compare the sources in our sample detected by LOFAR to those also detected in NVSS at 1400 MHz, and find that our selection bias toward bright radio galaxies drives a tendency for sources depolarized between 1400 and 150 MHz to have flatter spectra over that frequency range than those that remain polarized at 150 MHz. By comparing observed rotation measures with an analytic model we find that we are preferentially sensitive to sources in low mass environments. We also infer that sources with one polarized hotspot are inclined by a small angle to the line of sight, while sources with hotspots in both lobes lie in the plane of the sky. We conclude that low frequency polarization in radio galaxies is related to a combination of environment, flux density and jet orientation.

astro-ph.GA↗

arXiv2020

Unveiling the rarest morphologies of the LOFAR Two-metre Sky Survey radio source population with self-organised maps

The Low Frequency Array (LOFAR) Two-metre Sky Survey (LoTSS) is a low-frequency radio continuum survey of the Northern sky at an unparalleled resolution and sensitivity. In order to fully exploit this huge dataset and those produced by the Square Kilometre Array in the next decade, automated methods in machine learning and data-mining will be increasingly essential both for morphological classifications and for identifying optical counterparts to the radio sources. Using self-organising maps (SOMs), a form of unsupervised machine learning, we created a dimensionality reduction of the radio morphologies for the $\sim$25k extended radio continuum sources in the LoTSS first data release, which is only $\sim$2 percent of the final LoTSS survey. We made use of \textsc{PINK}, a code which extends the SOM algorithm with rotation and flipping invariance, increasing its suitability and effectiveness for training on astronomical sources. After training, the SOMs can be used for a wide range of science exploitation and we present an illustration of their potential by finding an arbitrary number of morphologically rare sources in our training data (424 square degrees) and subsequently in an area of the sky ($\sim$5300 square degrees) outside the training data. Objects found in this way span a wide range of morphological and physical categories: extended jets of radio active galactic nuclei, diffuse cluster haloes and relics, and nearby spiral galaxies. Finally, to enable accessible, interactive, and intuitive data exploration, we showcase the LOFAR-PyBDSF Visualisation Tool, which allows users to explore the LoTSS dataset through the trained SOMs.

astro-ph.IM↗