Dr Harry Desmond

Abstract:

f ( R ) is a paradigmatic modified gravity theory that typifies extensions to General Relativity with new light degrees of freedom and hence screened fifth forces between masses. These forces produce observable signatures in galaxy morphology, caused by a violation of the weak equivalence principle due to a differential impact of screening among galaxies’ mass components. We compile statistical datasets of two morphological indicators—offsets between stars and gas in galaxies and warping of stellar disks—and use them to constrain the strength and range of a thin-shell-screened fifth force. This is achieved by applying a comprehensive set of upgrades to past work [H. Desmond et al., Phys. Rev. D 98, 064015 (2018); H. Desmond et al., Phys. Rev. D 98, 083010 (2018) ]: we construct a robust galaxy-by-galaxy Bayesian forward model for the morphological signals, including full propagation of uncertainties in the input quantities and marginalization over an empirical model describing astrophysical noise. Employing more stringent data quality cuts than previously we find no evidence for a screened fifth force of any strength Δ G / G N in the Compton wavelength range 0.3–8 Mpc, setting a 1 σ bound of Δ G / G N < 0.8 at λ C = 0.3     Mpc that strengthens to Δ G / G N < 3 × 10 − 5 at λ C = 8     Mpc . These are the tightest bounds to date beyond the Solar System by over an order of magnitude. For the Hu-Sawicki model of f ( R ) with n = 1 we require a background scalar field value f R 0 < 1.4 × 10 − 8 , forcing practically all astrophysical objects to be screened. We conclude that this model can have no relevance to astrophysics or cosmology.

Abstract:

We use the mass discrepancy–acceleration relation (the correlation between the ratio of total-to-visible mass and acceleration in galaxies; MDAR) to test the galaxy–halo connection. We analyse the MDAR using a set of 16 statistics that quantify its four most important features: shape, scatter, the presence of a ‘characteristic acceleration scale’, and the correlation of its residuals with other galaxy properties. We construct an empirical framework for the galaxy– halo connection inLCDMto generate predictions for these statistics, starting with conventional correlations (halo abundance matching;AM)and introducing more where required. Comparing to the SPARC data, we find that: (1) the approximate shape of the MDAR is readily reproduced by AM, and there is no evidence that the acceleration at which dark matter becomes negligible has less spread in the data than in AM mocks; (2) even under conservative assumptions, AM significantly overpredicts the scatter in the relation and its normalization at low acceleration, and furthermore positions dark matter too close to galaxies’ centres on average; (3) the MDAR affords 2σ evidence for an anticorrelation of galaxy size and Hubble type with halo mass or concentration at fixed stellar mass. Our analysis lays the groundwork for a bottom-up determination of the galaxy–halo connection from relations such as the MDAR, provides concrete statistical tests for specific galaxy formationmodels, and brings into sharper focus the relative evidence accorded by galaxy kinematics to LCDM and modified gravity alternatives.

Abstract:

Context. Current and future large-scale structure surveys aim to constrain the neutrino mass and the equation of state of dark energy. To do this efficiently, rapid yet accurate evaluation of the matter power spectrum in the presence of these effects is essential. Aims. We aim to construct accurate and interpretable symbolic approximations of the linear and nonlinear matter power spectra as a function of cosmological parameters in extended ΛCDM models that contain massive neutrinos and nonconstant equations of state for dark energy. This constitutes an extension of the SYREN-HALOFIT emulators to incorporate these two effects, which we call SYREN-NEW (SYmbolic-Regression-ENhanced power spectrum emulator with NEutrinos and W0−wa). We also wish to obtain a simple approximation of the derived parameter, σ8, as a function of the cosmological parameters for these models. Methods. We utilizedd symbolic regression to efficiently search through candidate analytic expressions to approximate the various quantities of interest. Our results for the linear power spectrum are designed to emulate CLASS, whereas for the nonlinear case we aim to match the results of EUCLIDEMULATOR2. We compared our results to existing emulators and N-body simulations. Results. Our analytic emulators for σ8, and the linear and nonlinear power spectra achieve root mean squared errors of 0.1%, 0.3%, and 1.3%, respectively, across a wide range of cosmological parameters, redshifts and wavenumbers. The error on the nonlinear power spectrum is reduced by approximately a factor of 2 when considering observationally plausible dark energy models and neutrino masses. We verify that emulator-related discrepancies are subdominant compared to observational errors and other modeling uncertainties when computing shear power spectra for LSST-like surveys. Our expressions have similar accuracy to existing (numerical) emulators, but are at least an order of magnitude faster, both on a CPU and a GPU. Conclusions. Our work greatly improves the accuracy, speed, and applicability range of current symbolic approximations of the linear and nonlinear matter power spectra. These now cover the same range of cosmological models as many numerical emulators with similar accuracy, but are much faster and more interpretable. We provide publicly available code for all symbolic approximations found.