Search for: All records

In this work, we extend our recently developed multifidelity emulation technique to the simulated Lyman-α forest flux power spectrum. Multifidelity emulation allows interpolation of simulation outputs between cosmological parameters using many cheap low-fidelity simulations and a few expensive high-fidelity simulations. Using a test suite of small-box (30 Mpc h−1) simulations, we show that multifidelity emulation is able to reproduce the Lyman-α forest flux power spectrum well, achieving an average accuracy when compared to a test suite of $0.8\, {\rm {per\ cent}}$. We further show that it has a substantially increased accuracy over single-fidelity emulators, constructed using either the high- or low-fidelity simulations only. In particular, it allows the extension of an existing simulation suite to smaller scales and higher redshifts.

 

Broad-band tomography statistically extracts the redshift distribution of frequency dependent emission from the cross-correlation of intensity maps with a reference catalog of galaxy tracers. We make forecasts for the performance of future all-sky UV experiments doing broad-band tomography. We consider the Cosmological Advanced Survey Telescope for Optical-UV Research (castor) and the Spectro-Photometer for the History of the Universe, Epoch of Reionization, and Ices Explorer (SPHEREx). The dominant uncertainty is from variability in the photometric zero-point, which scales with limiting magnitude and mirror size. With this scaling and assuming a galaxy number density characteristic of future spectroscopic data sets, we find that castor measures the UV background SED 2–10 times better than existing data. The applicable redshift range will expand from the current z < 1 to z ≈ 0–3 with castor and z = 5–9 with SPHEREx. We show that castor can provide competitive constraints on the EBL monopole to those available from galaxy number counts and direct measurement techniques. At high redshift especially, these results will help understand galaxy formation and reionization. Our modelling code and chains are publicly available.

 

We introduce the Astrid  simulation, a large-scale cosmological hydrodynamic simulation in a $250 \, h^{-1}\mathrm{Mpc}$ box with 2 × 55003 particles. Astrid contains a large number of high redshift galaxies, which can be compared to future survey data, and resolves galaxies in haloes more massive than $2\times 10^9 \, \mathrm{M}_{\odot }$. Astrid  has been run from z = 99 to 3. As a particular focus is modelling the high redshift Universe, it contains models for inhomogeneous hydrogen and helium reionization, baryon relative velocities and massive neutrinos, as well as supernova and AGN feedback. The black hole model includes mergers driven by dynamical friction rather than repositioning. We briefly summarize the implemented models, and the technical choices we took when developing the simulation code. We validate the model, showing good agreement with observed ultraviolet luminosity functions, galaxy stellar mass functions and specific star formation rates (SFRs). We show that the redshift at which a given galaxy underwent hydrogen reionization has a large effect on the halo gas fraction. Finally, at z = 6, haloes with $M \sim 2\times 10^9 \, \mathrm{M}_{\odot }$ which have been reionized have an SFR 1.5 times greater than those which have not yet been reionized.

 

Extremely red quasars (ERQs) are an interesting sample of quasars in the Baryon Oscillation Spectroscopic Sample (BOSS) in the redshift range of 2.0–3.4 and have extreme red colours of i − W3 ≥ 4.6. Core ERQs have strong C  iv emission lines with rest equivalent width of ≥100 Å. Many core ERQs also have C  iv line profiles with peculiar boxy shapes which distinguish them from normal blue quasars. We show, using a combination of kernel density estimation and local outlier factor analyses on a space of the i − W3 colour, C iv rest equivalent width and line kurtosis, that core ERQs likely represent a separate population rather than a smooth transition between normal blue quasars and the quasars in the tail of the colour-REW distribution. We apply our analyses to find new criteria for selecting ERQs in this 3D parameter space. Our final selection produces 133 quasars, which are three times more likely to have a visually verified C iv broad absorption line feature than the previous core ERQ sample. We further show that our newly selected sample are extreme objects in the intersection of the WISE AGN catalogue with the MILLIQUAS quasar catalogue in the colour–colour space of (W1 − W2, W2 − W3). This paper validates an improved selection method for red quasars which can be applied to future data sets such as the quasar catalogue from the Dark Energy Spectroscopic Instrument.

 

We examine massive black hole (MBH) mergers and their associated gravitational wave signals from the large-volume cosmological simulation Astrid . Astrid includes galaxy formation and black hole models recently updated with an MBH seed population between 3 × 104h−1M⊙ and 3 × 105h−1M⊙ and a sub-grid dynamical friction (DF) model to follow the MBH dynamics down to 1.5 ckpc h−1. We calculate the initial eccentricities of MBH orbits directly from the simulation at kpc-scales, and find orbital eccentricities above 0.7 for most MBH pairs before the numerical merger. After approximating unresolved evolution on scales below ${\sim 200\, \text{pc}}$, we find that the in-simulation DF on large scales accounts for more than half of the total orbital decay time ($\sim 500\, \text{Myr}$) due to DF. The binary hardening time is an order of magnitude longer than the DF time, especially for the seed-mass binaries (MBH < 2Mseed). As a result, only $\lesssim 20{{\rm per \,cent}}$ of seed MBH pairs merge at z > 3 after considering both unresolved DF evolution and binary hardening. These z > 3 seed-mass mergers are hosted in a biased population of galaxies with the highest stellar masses of $\gt 10^9\, {\rm M}_\odot$. With the higher initial eccentricity prediction from Astrid , we estimate an expected merger rate of 0.3−0.7 per year from the z > 3 MBH population. This is a factor of ∼7 higher than the prediction using the circular orbit assumption. The Laser Interferometer Space Antenna events are expected at a similar rate, and comprise $\gtrsim 60\,{\rm{per\,cent}}$ seed-seed mergers, $\sim 30\,{\rm{per\,cent}}$ involving only one seed-mass MBH, and $\sim 10\,{\rm{per\,cent}}$ mergers of non-seed MBHs.

 

In this work, we establish and test methods for implementing dynamical friction (DF) for massive black hole pairs that form in large volume cosmological hydrodynamical simulations that include galaxy formation and black hole growth. We verify our models and parameters both for individual black hole dynamics and for the black hole population in cosmological volumes. Using our model of DF from collisionless particles, black holes can effectively sink close to the galaxy centre, provided that the black hole’s dynamical mass is at least twice that of the lowest mass resolution particles in the simulation. Gas drag also plays a role in assisting the black holes’ orbital decay, but it is typically less effective than that from collisionless particles, especially after the first billion years of the black hole’s evolution. DF from gas becomes less than $1{{\ \rm per\ cent}}$ of DF from collisionless particles for BH masses >107 M⊙. Using our best DF model, we calculate the merger rate down to z = 1.1 using an Lbox = 35 Mpc h−1 simulation box. We predict ∼2 mergers per year for z > 1.1 peaking at z ∼ 2. These merger rates are within the range obtained in previous work using similar resolution hydrodynamical simulations. We show that the rate is enhanced by factor of ∼2 when DF is taken into account in the simulations compared to the no-DF run. This is due to ${\gt}40{{\ \rm per\ cent}}$ more black holes reaching the centre of their host halo when DF is added.

 

Cosmological simulations are indispensable for understanding our Universe, from the creation of the cosmic web to the formation of galaxies and their central black holes. This vast dynamic range incurs large computational costs, demanding sacrifice of either resolution or size and often both. We build a deep neural network to enhance low-resolution dark-matter simulations, generating superresolution realizations that agree remarkably well with authentic high-resolution counterparts on their statistical properties and are orders-of-magnitude faster. It readily applies to larger volumes and generalizes to rare objects not present in the training data. Our study shows that deep learning and cosmological simulations can be a powerful combination to model the structure formation of our Universe over its full dynamic range.

 

Reionization leads to large spatial fluctuations in the intergalactic temperature that can persist well after its completion. We study the imprints of such fluctuations on the $z$ ∼ 5 Ly α forest flux power spectrum using a set of radiation-hydrodynamic simulations that model different reionization scenarios. We find that large-scale coherent temperature fluctuations bring ${\sim}20\text{--}60{{\ \rm per\ cent}}$ extra power at k ∼ 0.002 s km−1, with the largest enhancements in the models where reionization is extended or ends the latest. On smaller scales (k ≳ 0.1 s km−1), we find that temperature fluctuations suppress power by ${\lesssim}10{{\ \rm per\ cent}}$. We find that the shape of the power spectrum is mostly sensitive to the reionization mid-point rather than temperature fluctuations from reionization’s patchiness. However, for all of our models with reionization mid-points of $z$ ≤ 8 ($z$ ≤ 12), the shape differences are ${\lesssim}20{{\ \rm per\ cent}}$ (${\lesssim}40{{\ \rm per\ cent}}$) because of a surprisingly well-matched cancellation between thermal broadening and pressure smoothing that occurs for realistic thermal histories. We also consider fluctuations in the ultraviolet background, finding their impact on the power spectrum to be much smaller than temperature fluctuations at k ≳ 0.01 s km−1. Furthermore, we compare our models to power spectrum measurements, finding that none of our models with reionization mid-points of $z$ < 8 is strongly preferred over another and that all of our models with mid-points of $z$ ≥ 8 are excluded at 2.5σ. Future measurements may be able to distinguish between viable reionization models if they can be performed at lower k or, alternatively, if the error bars on the high-k power can be reduced by a factor of 1.5.

 

Abstract We present methods for emulating the matter power spectrum by combining information from cosmological N-body simulations at different resolutions. An emulator allows estimation of simulation output by interpolating across the parameter space of a limited number of simulations. We present the first implementation in cosmology of multi-fidelity emulation, where many low-resolution simulations are combined with a few high-resolution simulations to achieve an increased emulation accuracy. The power spectrum’s dependence on cosmology is learned from the low-resolution simulations, which are in turn calibrated using high-resolution simulations. We show that our multi-fidelity emulator predicts high-fidelity counterparts to percent-level relative accuracy when using only 3 high-fidelity simulations and outperforms a single-fidelity emulator that uses 11 simulations, although we do not attempt to produce a converged emulator with high absolute accuracy. With a fixed number of high-fidelity training simulations, we show that our multi-fidelity emulator is ≃ 100 times better than a single-fidelity emulator at k ≤ 2 hMpc−1, and ≃ 20 times better at 3 ≤ k < 6.4 hMpc−1. Multi-fidelity emulation is fast to train, using only a simple modification to standard Gaussian processes. Our proposed emulator shows a new way to predict non-linear scales by fusing simulations from different fidelities.