DCDuncan Campbell

Duncan Campbell

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Duncan Campbell has a total of 51 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Halotools: A New Release Adding Intrinsic Alignments to Halo-Based Methods

    by Nicholas Van Alfen, Duncan Campbell, Andrew Hearin, Jonathan Blazek


  • Galaxies and haloes on graph neural networks: Deep generative modelling scalar and vector quantities for intrinsic alignment

    by Yesukhei Jagvaral, François Lanusse, Sukhdeep Singh, Rachel Mandelbaum, Siamak Ravanbakhsh, Duncan Campbell

    In order to prepare for the upcoming wide-field cosmological surveys, large simulations of the Universe with realistic galaxy populations are required. In particular, the tendency of galaxies to naturally align towards overdensities, an effect called intrinsic alignments (IA), can be a major source of systematics in the weak lensing analysis. As the details of galaxy formation and evolution relevant to IA cannot be simulated in practice on such volumes, we propose as an alternative a Deep Generative Model. This model is trained on the IllustrisTNG-100 simulation and is capable of sampling the orientations of a population of galaxies so as to recover the correct alignments. In our approach, we model the cosmic web as a set of graphs, where the graphs are constructed for each halo, and galaxy orientations as a signal on those graphs. The generative model is implemented on a Generative Adversarial Network architecture and uses specifically designed Graph-Convolutional Networks sensitive to the relative 3D positions of the vertices. Given (sub)halo masses and tidal fields, the model is able to learn and predict scalar features such as galaxy and dark matter subhalo shapes; and more importantly, vector features such as the 3D orientation of the major axis of the ellipsoid and the complex 2D ellipticities. For correlations of 3D orientations the model is in good quantitative agreement with the measured values from the simulation, except for at very small and transition scales. For correlations of 2D ellipticities, the model is in good quantitative agreement with the measured values from the simulation on all scales. Additionally, the model is able to capture the dependence of IA on mass, morphological type, and central/satellite type.


  • Open Access

    Probabilistic model for dynamic galaxy decomposition

    by Yesukhei Jagvaral, Duncan Campbell, Rachel Mandelbaum, Markus Michael Rau

    In the era of precision cosmology and ever-improving cosmological simulations, a better understanding of different galaxy components such as bulges and discs will give us new insight into galactic formation and evolution. Based on the fact that the stellar populations of the constituent components of galaxies differ by their dynamical properties, we develop two simple models for galaxy decomposition using the TNG100 cosmological hydrodynamical simulation from the IllustrisTNG project. The first model uses a single dynamical parameter and can distinguish four components: thin disc, thick disc, counter-rotating disc, and bulge. The second model uses one more dynamical parameter, was defined in a probabilistic manner, and distinguishes two components: bulge and disc. We demonstrate the improved robustness of these models compared to a widely used method in literature involving cuts on the circularity parameter. The number fraction of disc-dominated galaxies at a given stellar mass obtained by our models agrees well with observations for masses exceeding log10(M*/M⊙) = 10. The galaxies classified as bulge-dominated by the second model are mostly red; however, the population classified as disc-dominated contains significant number of red galaxies alongside the blue population. The contributions of the different galaxy components to the total stellar mass budget exhibits similar trends with stellar mass compared to the observational data, although there is a quantitative disagreement at high and low masses. The Sérsic indices and half-mass radii for the bulge and disc components agree well with those of real galaxies.


  • Generating synthetic cosmological data with GalSampler

    by Andrew Hearin, Danila Korytov, Eve Kovacs, Andrew Benson, Han Aung, Christopher Bradshaw, Duncan Campbell, (The LSST Dark Energy Science Collaboration)

    As part of the effort to meet the needs of the Large Synoptic Survey Telescope Dark Energy Science Collaboration (LSST DESC) for accurate, realistically complex mock galaxy catalogues, we have developed galsampler, an open-source python package that assists in generating large volumes of synthetic cosmological data. The key idea behind galsampler is to recast hydrodynamical simulations and semi-analytic models as physically motivated galaxy libraries. galsampler populates a new, larger volume halo catalogue with galaxies drawn from the baseline library; by using weighted sampling guided by empirical modelling techniques, galsampler inherits statistical accuracy from the empirical model and physically motivated complexity from the baseline library. We have recently used galsampler to produce the cosmoDC2 extragalactic catalogue made for the LSST DESC Data Challenge 2. Using cosmoDC2 as a guiding example, we outline how galsampler can continue to support ongoing and near-future galaxy surveys such as the Dark Energy Survey, the Dark Energy Spectroscopic Instrument, WFIRST, and Euclid.


  • How to optimally constrain galaxy assembly bias: supplement projected correlation functions with count-in-cells statistics

    by Kuan Wang, Yao-Yuan Mao, Andrew R Zentner, Frank C van Den Bosch, Johannes U Lange, Chad M Schafer, Antonia Sierra Villarreal, Andrew P Hearin, Duncan Campbell

    Most models for the statistical connection between galaxies and their haloes ignore the possibility that galaxy properties may be correlated with halo properties other than halo mass, a phenomenon known as galaxy assembly bias. And yet, it is known that such correlations can lead to systematic errors in the interpretation of survey data that are analysed using traditional halo occupation models. At present, the degree to which galaxy assembly bias may be present in the real Universe, and the best strategies for constraining it remain uncertain. We study the ability of several observables to constrain galaxy assembly bias from redshift survey data using the decorated halo occupation distribution (dHOD), an empirical model of the galaxy–halo connection that incorporates assembly bias. We cover an expansive set of observables, including the projected two-point correlation function $w$p(rp), the galaxy–galaxy lensing signal ΔΣ(rp), the void probability function VPF(r), the distributions of counts-in-cylinders P(NCIC), and counts-in-annuli P(NCIA), and the distribution of the ratio of counts in cylinders of different sizes P(N2/N5). We find that despite the frequent use of the combination $w$p(rp) + ΔΣ(rp) in interpreting galaxy data, the count statistics, P(NCIC) and P(NCIA), are generally more efficient in constraining galaxy assembly bias when combined with $w$p(rp). Constraints based upon $w$p(rp) and ΔΣ(rp) share common degeneracy directions in the parameter space, while combinations of $w$p(rp) with the count statistics are more complementary. Therefore, we strongly suggest that count statistics should be used to complement the canonical observables in future studies of the galaxy–halo connection.


  • Halo occupation distribution (HOD) modelling of high redshift galaxies using the BlueTides simulation

    by Aklant K Bhowmick, Duncan Campbell, Tiziana Di Matteo, Yu Feng

    We construct halo occupation distribution (HOD) models of high redshift ($$z$$ ≳ 7.5) galaxies with M* > 108 M⊙ h−1 using the BlueTides hydrodynamic simulation suite, with a particular emphasis on modelling the small scale / 1-halo clustering ($$0.01\lesssim r \lesssim 1 h^{-1}\rm {Mpc}$$). Similar to low redshift studies, we find that the central and satellite mean HODs (〈Ncen〉 and 〈Nsat〉) can be modelled by a smoothed step function and a power law, respectively. The number density of satellite galaxies is however significantly suppressed compared to low redshift (satellite fractions drop from $${\sim } 50 {{\ \rm per\ cent}}$$ at $$z$$ = 0 to $${\lesssim } 10 {{\ \rm per\ cent}}$$ at $$z$$ = 7.5). The mean number of satellites 〈Nsat〉 < 1 for halo masses below 3 × 1011 M⊙ h−1 (a rare halo at these redshifts). For the radial number density profiles, satellites with 108 ≲ M* ≲ 109 M⊙ h−1 in haloes with MH ≳ 3 × 1011 M⊙ h−1 are consistent with NFW (Navarro, Frenk & White profile with concentrations csat ∼ 10−40). Within haloes of mass MH ≲ 3 × 1011 M⊙  h−1, satellites exhibit a power-law profile with slope of −3. Because these haloes dominate the small-scale clustering, the resulting 1-halo term is steeper than predicted using standard NFW profiles. Using this power-law profile for satellites, we can successfully reproduce the small-scale clustering exhibited by BlueTides galaxies using HOD modelling. We predict the highest probability of detecting satellites at $$z$$ > 7.5 is around centrals of M* ∼ 3 × 1010 M⊙ h−1 (with M* ≳ a few 107 M⊙ h−1). This should be achievable with the James Webb Space Telescope.


  • The galaxy clustering crisis in abundance matching

    by Duncan Campbell, Frank C van Den Bosch, Nikhil Padmanabhan, Yao-Yuan Mao, Andrew R Zentner, Johannes U Lange, Fangzhou Jiang, Antonia Sierra Villarreal

    Galaxy clustering on small scales is significantly underpredicted by sub-halo abundance matching (SHAM) models that populate (sub-)haloes with galaxies based on peak halo mass, Mpeak. SHAM models based on the peak maximum circular velocity, Vpeak, have had much better success. The primary reason for Mpeak-based models fail is the relatively low abundance of satellite galaxies produced in these models compared to those based on Vpeak. Despite success in predicting clustering, a simple Vpeak-based SHAM model results in predictions for galaxy growth that are at odds with observations. We evaluate three possible remedies that could ‘save’ mass-based SHAM: (1) SHAM models require a significant population of ‘orphan’ galaxies as a result of artificial disruption/merging of sub-haloes in modern high-resolution dark matter simulations; (2) satellites must grow significantly after their accretion; and (3) stellar mass is significantly affected by halo assembly history. No solution is entirely satisfactory. However, regardless of the particulars, we show that popular SHAM models based on Mpeak cannot be complete physical models as presented. Either Vpeak truly is a better predictor of stellar mass at z ∼ 0 and it remains to be seen how the correlation between stellar mass and Vpeak comes about, or SHAM models are missing vital component(s) that significantly affect galaxy clustering.


  • Brightest galaxies as halo centre tracers in SDSS DR7

    by Johannes U. Lange, Frank C. van Den Bosch, Andrew Hearin, Duncan Campbell, Andrew R. Zentner, Antonia Sierra Villarreal, Yao-Yuan Mao

    Determining the positions of halo centres in large-scale structure surveys is crucial for many cosmological studies. A common assumption is that halo centres correspond to the location of their brightest member galaxies. In this paper, we study the dynamics of brightest galaxies with respect to other halo members in the Sloan Digital Sky Survey DR7. Specifically, we look at the line-of-sight velocity and spatial offsets between brightest galaxies and their neighbours. We compare those to detailed mock catalogues, constructed from high-resolution, dark-matter-only N-body simulations, in which it is assumed that satellite galaxies trace dark matter subhaloes. This allows us to place constraints on the fraction fBNC of haloes in which the brightest galaxy is not the central. Compared to previous studies, we explicitly take into account the unrelaxed state of the host haloes, velocity offsets of halo cores and correlations between fBNC and the satellite occupation. We find that fBNC strongly decreases with the luminosity of the brightest galaxy and increases with the mass of the host halo. Overall, in the halo mass range 1013–1014.5 h− 1M⊙ we find fBNC ∼ 30 per cent, in good agreement with a previous study by Skibba et al. We discuss the implications of these findings for studies inferring the galaxy–halo connection from satellite kinematics, models of the conditional luminosity function and galaxy formation in general.


  • Open Access

    The immitigable nature of assembly bias: the impact of halo definition on assembly bias

    by Antonia Sierra Villarreal, Andrew R. Zentner, Yao-Yuan Mao, Chris W. Purcell, Frank C. van Den Bosch, Benedikt Diemer, Johannes U. Lange, Kuan Wang, Duncan Campbell

    Dark matter halo clustering depends not only on halo mass, but also on other properties such as concentration and shape. This phenomenon is known broadly as assembly bias. We explore the dependence of assembly bias on halo definition, parametrized by spherical overdensity parameter, Δ. We summarize the strength of concentration-, shape-, and spin-dependent halo clustering as a function of halo mass and halo definition. Concentration-dependent clustering depends strongly on mass at all Δ. For conventional halo definitions (Δ ∼ 200 − 600 m), concentration-dependent clustering at low mass is driven by a population of haloes that is altered through interactions with neighbouring haloes. Concentration-dependent clustering can be greatly reduced through a mass-dependent halo definition with Δ ∼ 20 − 40 m for haloes with M200 m ≲ 1012 h−1M⊙. Smaller Δ implies larger radii and mitigates assembly bias at low mass by subsuming altered, so-called backsplash haloes into now larger host haloes. At higher masses (M200 m ≳ 1013 h−1M⊙) larger overdensities, Δ ≳ 600 m, are necessary. Shape- and spin-dependent clustering are significant for all halo definitions that we explore and exhibit a relatively weaker mass dependence. Generally, both the strength and the sense of assembly bias depend on halo definition, varying significantly even among common definitions. We identify no halo definition that mitigates all manifestations of assembly bias. A halo definition that mitigates assembly bias based on one halo property (e.g. concentration) must be mass dependent. The halo definitions that best mitigate concentration-dependent halo clustering do not coincide with the expected average splashback radii at fixed halo mass.


  • Introducing decorated HODs: modelling assembly bias in the galaxy–halo connection

    by Andrew P. Hearin, Andrew R. Zentner, Frank C. van Den Bosch, Duncan Campbell, Erik Tollerud

    The connection between galaxies and dark matter haloes is often inferred from data using probabilistic models, such as the halo occupation distribution (HOD). Conventional HOD formulations assume that only halo mass governs the galaxy–halo connection. Violations of this assumption, known as galaxy assembly bias, threaten the HOD programme. We introduce decorated HODs, a new, flexible class of models designed to account for assembly bias. Decorated HODs minimally expand the parameter space and maximize the independence between traditional and novel HOD parameters. We use decorated HODs to quantify the influence of assembly bias on clustering and lensing statistics. For SDSS-like samples, the impact of assembly bias on galaxy clustering can be as large as a factor of 2 on r ∼ 200 kpc scales and ∼15 per cent in the linear regime. Assembly bias can either enhance or diminish clustering on large scales, but generally increases clustering on scales r ≲ 1 Mpc. We performed our calculations with halotools, an open-source, community-driven python package for studying the galaxy–halo connection (http://halotools.readthedocs.org). We conclude by describing the use of decorated HODs to treat assembly bias in otherwise conventional likelihood analyses.