Johnston, Harry; (2020) The alignments and clustering of galaxies in wide-area photometric galaxy surveys. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Thesis__final.pdf - Accepted Version Download (16MB) | Preview

Abstract

The upcoming decades will see the transformation of our understanding of the Universe. New experiments, soon to come online, aim to observe the evolution of large-scale structures traced by billions of galaxies, throughout much of cosmic time. The volume of data soon to be at our disposal comes with responsibility; with unprecedented levels of statistical power, we must identify and control sources of systematic error in our analyses to an ever greater degree, lest we waste newfound precision upon inaccurate inferences. This thesis explores the impacts of such errors upon our understanding of galaxy data, and proposes methods for their mitigation. First, I detail my study of the intrinsic alignments of galaxies (Ch. 2). The study of weak cosmological lensing (or `cosmic shear') posits that the distribution of intrinsic galaxy shapes should be random, and thus that we can learn about the Universe by attributing shape correlations to the effects of gravitational lensing by the large-scale structure. However, we observe different galaxies to be intrinsically aligned with structure in complex ways; violating the assumption of randomness and forming the primary astrophysical systematic for cosmic shear analyses. Using a unique set of highly-complete, spectroscopic data, I directly measure and model the projected 3D galaxy intrinsic alignments and clustering, revealing new complexity regarding the spiral/elliptical, central/satellite nature of galaxies, before forecasting the benefits of my data-driven priors for intrinsic alignment models in future cosmic shear work. The physical galaxy distribution is another powerful probe of the Universe, however, measurements of galaxy clustering must contend with spatially non-uniform observing conditions. If poor conditions result in systematic failures to detect objects, the observed clustering will not represent the true galaxy density field. I describe (in Ch. 3) a method of mitigation for such biases, centred around the retrieval of systematic density modes from galaxy data using self-organising maps (SOMs). Creating random galaxy catalogues which mimic and thereby subtract the systematic density trends, I demonstrate the accurate recovery of clustering signals from realistically deprecated synthetic data. I go on to present the first photometric angular clustering measurement from the Kilo Degree Survey, made robust by our corrective randoms. Studies of 3D, or projected, clustering and intrinsic alignments are typically limited to spectroscopic, rather than photometric, data; accurate redshifts are necessary to isolate objects in the radial dimension. The Physics of the Accelerating Universe Survey bridges this gap, offering greater depths at a small cost to redshift accuracy, by observing in 40 optical narrow-bands. In Ch. 4, I derive principled random galaxy catalogues, capable of reproducing the galaxies' redshift distribution sans structure, and doing so robustly for arbitrary galaxy sample selections. With these randoms, I explore the projected 3D clustering and intrinsic alignments of these data, finding quite remarkable support for the conclusions of my previous work (Ch 2), and extending the study of intrinsic alignments to yet fainter objects and smaller scales.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as JSONCSVXML

Archive Staff Only

View Item