arXiv subjects

Thomas D. Kitching

Publications and source records attributed to Thomas D. Kitching.

At least 19 recordsLinked to original sources

arXiv2022

Posterior sampling for inverse imaging problems on the sphere in seismology and cosmology

Inverse problems defined on the sphere arise in many fields, including seismology and cosmology where problems are defined on the globe and the cosmic sphere. These are generally high-dimensional and computationally very complex and, as a result, sampling the posterior of spherical inverse problems is a challenging task. In this work, we describe a framework that leverages a proximal Markov chain Monte Carlo (MCMC) algorithm to efficiently sample the high-dimensional space of spherical inverse problems with a sparsity-promoting wavelet prior. We detail the modifications needed for the algorithm to be applied to spherical problems, and give special consideration to the crucial forward modelling step which contains spherical harmonic transforms that are computationally expensive. By sampling the posterior, our framework allows for full and flexible uncertainty quantification, something which is not possible with other methods based on, for example, convex optimisation. We demonstrate our framework in practice on full-sky cosmological mass-mapping and on a common problem in global seismic tomography. We find that our approach is potentially useful at moderate resolutions, such as those of interest in seismology. Our framework is generally limited by resolution requirements, such as those required for astrophysical applications, due to the poor scaling of the complexity of spherical harmonic transforms with resolution. A new Python package, pxmcmc, containing the proximal MCMC sampler, measurement operators, wavelet transforms and sparse priors is made publicly available.

physics.data-an↗

arXiv2021

Mapping dark matter on the celestial sphere with weak gravitational lensing

Convergence maps of the integrated matter distribution are a key science result from weak gravitational lensing surveys. To date, recovering convergence maps has been performed using a planar approximation of the celestial sphere. However, with the increasing area of sky covered by dark energy experiments, such as Euclid, the Large Synoptic Survey Telescope (LSST), and the Wide Field Infrared Survey Telescope (WFIRST), this assumption will no longer be valid. We recover convergence fields on the celestial sphere using an extension of the Kaiser-Squires estimator to the spherical setting. Through simulations we study the error introduced by planar approximations. Moreover, we examine how best to recover convergence maps in the planar setting, considering a variety of different projections and defining the local rotations that are required when projecting spin fields such as cosmic shear. For the sky coverages typical of future surveys, errors introduced by projection effects can be of order tens of percent, exceeding 50% in some cases. The stereographic projection, which is conformal and so preserves local angles, is the most effective planar projection. In any case, these errors can be avoided entirely by recovering convergence fields directly on the celestial sphere. We apply the spherical Kaiser-Squires mass-mapping method presented to the public Dark Energy Survey (DES) science verification data to recover convergence maps directly on the celestial sphere.

astro-ph.CO↗

arXiv2021

Towards Machine Learning-Based Meta-Studies: Applications to Cosmological Parameters

We develop a new model for automatic extraction of reported measurement values from the astrophysical literature, utilising modern Natural Language Processing techniques. We use this model to extract measurements present in the abstracts of the approximately 248,000 astrophysics articles from the arXiv repository, yielding a database containing over 231,000 astrophysical numerical measurements. Furthermore, we present an online interface (Numerical Atlas) to allow users to query and explore this database, based on parameter names and symbolic representations, and download the resulting datasets for their own research uses. To illustrate potential use cases we then collect values for nine different cosmological parameters using this tool. From these results we can clearly observe the historical trends in the reported values of these quantities over the past two decades, and see the impacts of landmark publications on our understanding of cosmology.

astro-ph.IM↗

arXiv2021

Cosmological Parameter Biases from Doppler-Shifted Weak Lensing in Stage IV Experiments

The advent of Stage IV weak lensing surveys will open up a new era in precision cosmology. These experiments will offer more than an order-of-magnitude leap in precision over existing surveys, and we must ensure that the accuracy of our theory matches this. Accordingly, it is necessary to explicitly evaluate the impact of the theoretical assumptions made in current analyses on upcoming surveys. One effect typically neglected in present analyses is the Doppler-shift of the measured source comoving distances. Using Fisher matrices, we calculate the biases on the cosmological parameter values inferred from a Euclid-like survey, if the correction for this Doppler-shift is omitted. We find that this Doppler-shift can be safely neglected for Stage IV surveys. The code used in this investigation is made publicly available.

astro-ph.CO↗

arXiv2021

Sparse Bayesian mass-mapping with uncertainties: hypothesis testing of structure

A crucial aspect of mass-mapping, via weak lensing, is quantification of the uncertainty introduced during the reconstruction process. Properly accounting for these errors has been largely ignored to date. We present a new method to reconstruct maximum a posteriori (MAP) convergence maps by formulating an unconstrained Bayesian inference problem with Laplace-type l1-norm sparsity-promoting priors, which we solve via convex optimization. Approaching mass-mapping in this manner allows us to exploit recent developments in probability concentration theory to infer theoretically conservative uncertainties for our MAP reconstructions, without relying on assumptions of Gaussianity. For the first time these methods allow us to perform hypothesis testing of structure, from which it is possible to distinguish between physical objects and artifacts of the reconstruction. Here we present this new formalism, demonstrate the method on simulations, before applying the developed formalism to two observational datasets of the Abel-520 cluster. Initial reconstructions of the Abel-520 catalogs reported the detection of an anomalous 'dark core' -- an over dense region with no optical counterpart -- which was taken to be evidence for self-interacting dark-matter. In our Bayesian framework it is found that neither Abel-520 dataset can conclusively determine the physicality of such dark cores at 99% confidence. However, in both cases the recovered MAP estimators are consistent with both sets of data.

astro-ph.CO↗

arXiv2021

Spherical Bayesian mass-mapping with uncertainties: full sky observations on the celestial sphere

To date weak gravitational lensing surveys have typically been restricted to small fields of view, such that the $\textit{flat-sky approximation}$ has been sufficiently satisfied. However, with Stage IV surveys ($\textit{e.g. LSST}$ and $\textit{Euclid}$) imminent, extending mass-mapping techniques to the sphere is a fundamental necessity. As such, we extend the sparse hierarchical Bayesian mass-mapping formalism presented in previous work to the spherical sky. For the first time, this allows us to construct $\textit{maximum a posteriori}$ spherical weak lensing dark-matter mass-maps, with principled Bayesian uncertainties, without imposing or assuming Gaussianty. We solve the spherical mass-mapping inverse problem in the analysis setting adopting a sparsity promoting Laplace-type wavelet prior, though this theoretical framework supports all log-concave posteriors. Our spherical mass-mapping formalism facilitates principled statistical interpretation of reconstructions. We apply our framework to convergence reconstruction on high resolution N-body simulations with pseudo-Euclid masking, polluted with a variety of realistic noise levels, and show a significant increase in reconstruction fidelity compared to standard approaches. Furthermore we perform the largest joint reconstruction to date of the majority of publicly available shear observational datasets (combining DESY1, KiDS450 and CFHTLens) and find that our formalism recovers a convergence map with significantly enhanced small-scale detail. Within our Bayesian framework we validate, in a statistically rigorous manner, the community's intuition regarding the need to smooth spherical Kaiser-Squires estimates to provide physically meaningful convergence maps. Such approaches cannot reveal the small-scale physical structures that we recover within our framework.

astro-ph.CO↗

arXiv2021

Sparse Bayesian mass-mapping with uncertainties: peak statistics and feature locations

Weak lensing convergence maps - upon which higher order statistics can be calculated - can be recovered from observations of the shear field by solving the lensing inverse problem. For typical surveys this inverse problem is ill-posed (often seriously) leading to substantial uncertainty on the recovered convergence maps. In this paper we propose novel methods for quantifying the Bayesian uncertainty in the location of recovered features and the uncertainty in the cumulative peak statistic - the peak count as a function of signal to noise ratio (SNR). We adopt the sparse hierarchical Bayesian mass-mapping framework developed in previous work, which provides robust reconstructions and principled statistical interpretation of reconstructed convergence maps without the need to assume or impose Gaussianity. We demonstrate our uncertainty quantification techniques on both Bolshoi N-body (cluster scale) and Buzzard V-1.6 (large scale structure) N-body simulations. For the first time, this methodology allows one to recover approximate Bayesian upper and lower limits on the cumulative peak statistic at well defined confidence levels.

astro-ph.CO↗

arXiv2021

Sparse Bayesian mass-mapping with uncertainties: local credible intervals

Until recently mass-mapping techniques for weak gravitational lensing convergence reconstruction have lacked a principled statistical framework upon which to quantify reconstruction uncertainties, without making strong assumptions of Gaussianity. In previous work we presented a sparse hierarchical Bayesian formalism for convergence reconstruction that addresses this shortcoming. Here, we draw on the concept of local credible intervals (cf. Bayesian error bars) as an extension of the uncertainty quantification techniques previously detailed. These uncertainty quantification techniques are benchmarked against those recovered via Px-MALA - a state of the art proximal Markov Chain Monte Carlo (MCMC) algorithm. We find that typically our recovered uncertainties are everywhere conservative, of similar magnitude and highly correlated (Pearson correlation coefficient $\geq 0.85$) with those recovered via Px-MALA. Moreover, we demonstrate an increase in computational efficiency of $\mathcal{O}(10^6)$ when using our sparse Bayesian approach over MCMC techniques. This computational saving is critical for the application of Bayesian uncertainty quantification to large-scale stage IV surveys such as LSST and Euclid.

astro-ph.CO↗

arXiv2021

Group-scale intrinsic galaxy alignments in the Illustris-TNG and MassiveBlack-II simulations

We study the alignments of satellite galaxies, and their anisotropic distribution, with respect to location and orientation of their host central galaxy in MassiveBlack-II and IllustrisTNG simulations. We find that: the shape of the satellite system in halos of mass ($> 10^{13}h^{-1}M_{\odot}$) is well aligned with the shape of the central galaxy at $z=0.06$ with the mean alignment between the major axes being $\sim Δθ= 12^{\circ}$ when compared to a uniform random distribution; that satellite galaxies tend to be anisotropically distributed along the major axis of the central galaxy with a stronger alignment in halos of higher mass or luminosity; and that the satellite distribution is more anisotropic for central galaxies with lower star formation rate, which are spheroidal, and for red central galaxies.Radially we find that satellites tend to be distributed along the major axis of the shape of the stellar component of central galaxies at smaller scales and the dark matter component on larger scales. We find that the dependence of satellite anisotropy on central galaxy properties and the radial distance is similar in both the simulations with a larger amplitude in MassiveBlack-II. The orientation of satellite galaxies tends to point toward the location of the central galaxy at small scales and this correlation decreases with increasing distance, and the amplitude of satellite alignment is higher in high mass halos. However, the projected ellipticities do not exhibit a scale-dependent radial alignment, as has been seen in some observational measurements.

astro-ph.CO↗

arXiv2020

Accounting for object detection bias in weak gravitational lensing studies

Weak lensing by large-scale structure is a powerful probe of cosmology if the apparent alignments in the shapes of distant galaxies can be accurately measured. Most studies have therefore focused on improving the fidelity of the shape measurements themselves, but the preceding step of object detection has been largely ignored. In this paper we study the impact of object detection for a Euclid-like survey and show that it leads to biases that exceed requirements for the next generation of cosmic shear surveys. In realistic scenarios, blending of galaxies is an important source of detection bias. We find that MetaDetection is able to account for blending, leading to average multiplicative biases that meet requirements for Stage IV surveys, provided a sufficiently accurate model for the point spread function is available. Further work is needed to estimate the performance for actual surveys. Combined with sufficiently realistic image simulations, this provides a viable way forward towards accurate shear estimates for Stage IV surveys.

astro-ph.CO↗

arXiv2020

Accessing the high-$\ell$ frontier under the Reduced Shear Approximation with $k$-cut Cosmic Shear

The precision of Stage IV cosmic shear surveys will enable us to probe smaller physical scales than ever before, however, model uncertainties from baryonic physics and non-linear structure formation will become a significant concern. The $k$-cut method -- applying a redshift-dependent $\ell$-cut after making the Bernardeau-Nishimichi-Taruya transform -- can reduce sensitivity to baryonic physics; allowing Stage IV surveys to include information from increasingly higher $\ell$-modes. Here we address the question of whether it can also mitigate the impact of making the reduced shear approximation; which is also important in the high-$κ$, small-scale regime. The standard procedure for relaxing this approximation requires the repeated evaluation of the convergence bispectrum, and consequently can be prohibitively computationally expensive when included in Monte Carlo analyses. We find that the $k$-cut cosmic shear procedure suppresses the $w_0w_a$CDM cosmological parameter biases expected from the reduced shear approximation for Stage IV experiments, when $\ell$-modes up to $5000$ are probed. The maximum cut required for biases from the reduced shear approximation to be below the threshold of significance is at $k = 5.37 \, h{\rm Mpc}^{-1}$. With this cut, the predicted $1σ$ constraints increase, relative to the case where the correction is directly computed, by less than $10\%$ for all parameters. This represents a significant improvement in constraints compared to the more conservative case where only $\ell$-modes up to 1500 are probed, and no $k$-cut is used. We also repeat this analysis for a hypothetical, comparable kinematic weak lensing survey. The key parts of code used for this analysis are made publicly available.

astro-ph.CO↗

arXiv2020

Post-Limber Weak Lensing Bispectrum, Reduced Shear Correction, and Magnification Bias Correction

The significant increase in precision that will be achieved by Stage IV cosmic shear surveys means that several currently used theoretical approximations may cease to be valid. An additional layer of complexity arises from the fact that many of these approximations are interdependent; the procedure to correct for one involves making another. Two such approximations that must be relaxed for upcoming experiments are the reduced shear approximation and the effect of neglecting magnification bias. Accomplishing this involves the calculation of the convergence bispectrum; typically subject to the Limber approximation. In this work, we compute the post-Limber convergence bispectrum, and the post-Limber reduced shear and magnification bias corrections to the angular power spectrum for a Euclid-like survey. We find that the Limber approximation significantly overestimates the bispectrum when any side of the bispectrum triangle, $\ell_i<60$. However, the resulting changes in the reduced shear and magnification bias corrections are well below the sample variance for $\ell\leq5000$. We also compute a worst-case scenario for the additional biases on $w_0w_a$CDM cosmological parameters that result from the difference between the post-Limber and Limber approximated forms of the corrections. These further demonstrate that the reduced shear and magnification bias corrections can safely be treated under the Limber approximation for upcoming surveys.

astro-ph.CO↗

arXiv2020

Massive neutrinos and degeneracies in Lyman-alpha forest simulations

Using a suite of hydrodynamical simulations with cold dark matter, baryons, and neutrinos, we present a detailed study of the effect of massive neutrinos on the 1-D and 3-D flux power spectra of the Lyman-$α$ (Ly$α$) forest. The presence of massive neutrinos in cosmology induces a scale- and time-dependent suppression of structure formation that is strongest on small scales. Measuring this suppression is a key method for inferring neutrino masses from cosmological data, and is one of the main goals of ongoing and future surveys like eBOSS, DES, LSST, Euclid or DESI. The clustering in the Ly$α$ forest traces the quasi-linear power at late times and on small scales. In combination with observations of the cosmic microwave background, the forest therefore provides some of the tightest constraints on the sum of the neutrino masses. However there is a well-known degeneracy between $Σm_ν$ and the amplitude of perturbations in the linear matter power spectrum. We study the corresponding degeneracy in the 1-D flux power spectrum of the Ly$α$ forest, and for the first time also study this degeneracy in the 3-D flux power spectrum. We show that the non-linear effects of massive neutrinos on the Ly$α$ forest, beyond the effect of linear power amplitude suppression, are negligible, and this degeneracy persists in the Ly$α$ forest observables to a high precision. We discuss the implications of this degeneracy for choosing parametrisations of the Ly$α$ forest for cosmological analysis.

astro-ph.CO↗

arXiv2019

Towards Machine-assisted Meta-Studies: The Hubble Constant

We present an approach for automatic extraction of measured values from the astrophysical literature, using the Hubble constant for our pilot study. Our rules-based model -- a classical technique in natural language processing -- has successfully extracted 298 measurements of the Hubble constant, with uncertainties, from the 208,541 available arXiv astrophysics papers. We have also created an artificial neural network classifier to identify papers in arXiv which report novel measurements. From the analysis of our results we find that reporting measurements with uncertainties and the correct units is critical information when distinguishing novel measurements in free text. Our results correctly highlight the current tension for measurements of the Hubble constant and recover the $3.5σ$ discrepancy -- demonstrating that the tool presented in this paper is useful for meta-studies of astrophysical measurements from a large number of publications.

astro-ph.IM↗

arXiv2019

Cosmic Shear: Inference from Forward Models

Density-estimation likelihood-free inference (DELFI) has recently been proposed as an efficient method for simulation-based cosmological parameter inference. Compared to the standard likelihood-based Markov Chain Monte Carlo (MCMC) approach, DELFI has several advantages: it is highly parallelizable, there is no need to assume a possibly incorrect functional form for the likelihood and complicated effects (e.g the mask and detector systematics) are easier to handle with forward models. In light of this, we present two DELFI pipelines to perform weak lensing parameter inference with lognormal realizations of the tomographic shear field -- using the C_l summary statistic. The first pipeline accounts for the non-Gaussianities of the shear field, intrinsic alignments and photometric-redshift error. We validate that it is accurate enough for Stage III experiments and estimate that O(1000) simulations are needed to perform inference on Stage IV data. By comparing the second DELFI pipeline, which makes no assumption about the functional form of the likelihood, with the standard MCMC approach, which assumes a Gaussian likelihood, we test the impact of the Gaussian likelihood approximation in the MCMC analysis. We find it has a negligible impact on Stage IV parameter constraints. Our pipeline is a step towards seamlessly propagating all data-processing, instrumental, theoretical and astrophysical systematics through to the final parameter constraints.

astro-ph.CO↗

arXiv2019

Non-parametric Cosmology with Cosmic Shear

We present a method to measure the growth of structure and the background geometry of the Universe -- with no a priori assumption about the underlying cosmological model. Using Canada-France-Hawaii Lensing Survey (CFHTLenS) shear data we simultaneously reconstruct the lensing amplitude, the linear intrinsic alignment amplitude, the redshift evolving matter power spectrum, P(k,z), and the co-moving distance, r(z). We find that lensing predominately constrains a single global power spectrum amplitude and several co-moving distance bins. Our approach can localise precise scales and redshifts where Lambda-Cold Dark Matter (LCDM) fails -- if any. We find that below z = 0.4, the measured co-moving distance r (z) is higher than that expected from the Planck LCDM cosmology by ~1.5 sigma, while at higher redshifts, our reconstruction is fully consistent. To validate our reconstruction, we compare LCDM parameter constraints from the standard cosmic shear likelihood analysis to those found by fitting to the non-parametric information and we find good agreement.

astro-ph.CO↗

arXiv2019

Covariant polarized radiative transfer on cosmological scales for investigating large-scale magnetic field structures

Polarization of radiation is a powerful tool to study cosmic magnetism and analysis of polarization can be used as a diagnostic tool for large-scale structures. In this paper, we present a solid theoretical foundation for using polarized light to investigate large-scale magnetic field structures: the cosmological polarized radiative transfer (CPRT) formulation. The CPRT formulation is fully covariant. It accounts for cosmological and relativistic effects in a self-consistent manner and explicitly treats Faraday rotation, as well as Faraday conversion, emission, and absorption processes. The formulation is derived from the first principles of conservation of phase-space volume and photon number. Without loss of generality, we consider a flat Friedmann-Robertson-Walker (FRW) space-time metric and construct the corresponding polarized radiative transfer equations. We propose an all-sky CPRT calculation algorithm, based on a ray-tracing method, which allows cosmological simulation results to be incorporated and, thereby, model templates of polarization maps to be constructed. Such maps will be crucial in our interpretation of polarized data, such as those to be collected by the Square Kilometer Array (SKA). We describe several tests which are used for verifying the code and demonstrate applications in the study of the polarization signatures in different distributions of electron number density and magnetic fields. We present a pencil-beam CPRT calculation and an all-sky calculation, using a simulated galaxy cluster or a model magnetized universe obtained from GCMHD+ simulations as the respective input structures. The implications on large-scale magnetic field studies are discussed; remarks on the standard methods using rotation measure are highlighted.

astro-ph.CO↗

arXiv2018

k-cut Cosmic Shear: Tunable Power Spectrum Sensitivity to Test Gravity

If left unchecked modeling uncertainties at small scales, due to poorly understood baryonic physics and non-linear structure formation, will significantly bias Stage IV cosmic shear two-point statistic parameter constraints. While it is perhaps possible to run N-body or hydrodynamical simulations to determine the impact of these effects this approach is computationally expensive; especially to test a large number of theories of gravity. Instead we propose directly removing sensitivity to small-scale structure from the lensing spectrum, creating a statistic that is robust to these uncertainties. We do this by taking a redshift-dependent l-cut after applying the Bernardeau-Nishimichi-Taruya (BNT) nulling scheme. This reorganizes the information in the lensing spectrum to make the relationship between the angular scale, l, and the structure scale, k, much clearer compared to standard cosmic shear power spectra -- for which no direct relationship exists. We quantify the effectiveness of this method at removing sensitivity to small scales and compute the predicted Fisher error on the dark energy equation of state, w0, for different k-cuts in the matter power spectrum.

astro-ph.CO↗