Title:

The impact of baryonic astrophysics and intrinsic alignments on dark energy and neutrino constraints in cosmology

In the early 2020s, cosmology will enter an era of unprecedented precision when the next generation of large scale structure surveys begin receiving data. As a result, it is expected that stronger constraints on major features of cosmology like dark energy and massive neutrinos are forthcoming. The former could help to favour an explanation for the present accelerated expansion of the Universe, while the latter has the potential to enhance our understanding of a prominent intersection between particle physics and cosmology. However, there are many systematics that must be accounted for in parameter forecasts. One of the most prominent theoretical cases is the influence of baryonic astrophysics on large scale structure (e.g., AGN and supernova feedback, adiabatic contraction etc.) and the effect that marginalising over a limited theoretical understanding of the associated phenomena will have on forecasts. This question is one of the core concerns of this thesis. It will be applied respectively in Chapters 3, 4 and 5 to constraints on dynamical dark energy, the neutrino mass sum and a possible coupling between dark energy and dark matter. While forecasts are the primary focus here, much of this work has implications for parameter inference more broadly, and could be used to inform the direction that model building or simulation development should take in pursuit of the goal of more accurate parameter constraints. The key statistics used here to probe the growth of structure derive from the power spectrum of matter overdensities. This permits the use of both weak gravitational lensing of light from background sources by foreground objects, and galaxy clustering. The former requires accounting for the intrinsic alignments of ellipticities and shears, a systematic examined in depth for its impact on forecasts. This thesis presents Fisher analyses using weak lensing and galaxy clustering probes for parameter forecasts for a Euclidlike survey. The approach here to modelling the baryonic phenomena is to adopt a generic treatment of their global redistribution of the dark matter content in haloes, via energy transfer to their surroundings. Different baryonic effects are separated into three general but distinct categories: large scale adiabatic contraction caused by radiative cooling; high impact energy transfer from specific, localised sources; and smallscale effects that manifest as inner halo cores. I introduce the inner halo core through analytic modelling. A central tenet of this work is the use of analytic modelling, rather than numerical simulations, in capturing the relevant physics so as to circumvent computational expenses and underlying systematics associated with the latter, while retaining the useful physical insight offered by the former. Employing a maximum likelihood method, matter and weak lensing power spectra are varied around a fiducial cosmology given by Planck Collaboration et al. (2016b). For a Euclidlike survey covering 15000 sq. deg. of sky, measuring 10 redshift bins in the range 0 < z < 2 , the w0  wa dark energy Figure of Merit is shown to experience a ~40% degradation due to the combination of baryon effects. This thesis presents a detailed analysis of the relative dark energy and baryon sensitivities over the range of available lensing modes. Ultimately, it is found that the application of cosmic microwave background (CMB) priors alleviate the baryon impact for individual errors on the dark energy parameters but the relative degradation to the Figure of Merit for the parameter space remains. A similar approach is used to address the question of whether Stage IV surveys, whilst accounting for baryons and intrinsic alignments, can make a positive detection of the neutrino mass and whether they can distinguish between the normal and inverted hierarchy of mass eigenstates. Combined forecasts from weak lensing, the CMB and galaxy clustering preclude a meaningful distinction of hierarchies but do achieve a positive detection of the mass sum, overcoming the significant degradation by factors of ~2 to that arise when marginalising over baryons for weak lensing alone. These results could be improved upon with future CMB priors on the spectral index and information from neutrinoless double beta decay to achieve a 2σ distinction of the hierarchies. The effect of intrinsic alignments on forecasts is shown to be minimal, with constraints even experiencing mild improvements due to information from the intrinsic alignment signal. Finally, this work explores the prospects for using large scale structure to constrain the strength of a possible coupling between dark matter and a dark energy scalar field. While the growth of structure in the linear regime in this model has been wellexplored, the nonlinear regime is more challenging. Forecasts have been made using polynomial fits to power spectra directly. However, this thesis presents a more physically motivated approach to the problem. I show that a single parameter, the halo virial density, is responsible for describing most of the impact of the coupling on small scales. By computing the changes to the virial density in this model, a fit can be found that allows for a full physicallymotivated halo model. This allows Fisher forecasts to be made for the coupling strength, including an assessment of the impact of baryons. Degradations to the coupling strength constraint of ~20% due to this systematic are found. While CMB and galaxy clustering priors notably improve the absolute errors without marginalising over baryons, when this systematic is accounted for these priors provide little improvement and the relative degradation increases. Taken as a whole, this thesis provides a comprehensive analysis of the impact of baryons and intrinsic alignments on constraints for a wide range of cosmological phenomena responsible for the accelerated expansion, the neutrino mass hierarchy and beyond ΛCDM physics coupling dark matter and dark energy. Through modified and improved approaches to halo modelling, this work demonstrates which of these phenomena are subject to the most severe degeneracies with baryonic effects. This rigorous analysis, grounded in empirically motivated parameterisations, is designed to inform optimal mitigation strategies to minimise the impact on forecasts. In turn, this provides a wide scope for making future improvements to modelling and simulations that can advance efforts to constrain cosmology still further.
